FACTORS AND PITFALLS INFLUENCING THE DETECTION OF BACTERIAL KIDNEY DISEASE By Carolyn A . Schulz A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Fisheries and Wildlife Doctor of Philosophy 2014 ABSTRACT FACTORS AND PITFALLS INFLUENCING THE DETECTION OF BACTERIAL KIDNEY DISEASE By Carolyn A . Schulz Bacterial kidney disease (BKD) has been a significant problem for over 50 years in the Great Lakes basin (GLB) where it has been associated with large - scale fish mortalities and chronic infections in feral and propagated fish. The most commonly used and widely accepted method for detection of R. salmoninarum is a sandwich enzyme - linked immunosorbent assay (ELISA) on lethally colle cted kidney and spleen samples. While ELISA is relatively fast and more efficient than other serological assays, there are some concerns regarding what could be overlooked as a result of conducting a single assay on specific samples. Herein, I describe s everal factors that could affect the detection of BKD in feral and experimentally infected fish that are often overlooked upon assay or sample choices. Using data from the past decade, a decline in the overall presence of BKD in feral and hatchery - raised Chinook salmon ( Oncorhynchus tshawytscha ) (CHS), coho salmon ( O. kisutch ), and steelhead ( O. mykiss ) was observed. This also coincided with the implementation of enhanced biosecurity measures at state hatcheries and gamete - collection. Moreover, to potent ially reduce the dependency on lethally collected samples for regular screening, the detection of R. salmoninarum in other sample types, such as mucus, blood, and a urine/feces mixture, was evaluated. All sample types were collected from experimentally in fected CHS and compared for their efficiency to detect the presence of R. salmoninarum by standard bacterial culture techniques, nested polymerase chain reaction (nPCR), and sandwich ELISA. It was found that the urine/feces mixture was the preferred non - l ethal sample, and that a combination of assays and samples greatly increased the likelihood of detecting R. salmoninarum . Not only does collecting several samples enhance the detection of R. salmoninarum , but the presence of R. salmoninarum in the urine/f eces mixture suggests that the bacterium may utilize the anal opening as a portal of entry. To determine if anti - R. salmoninarum antibodies are elicited during an infection, and to establish a protocol for their detection, a single - dilution indirect - ELISA was modified and assessed for its usefulness in detecting and semi - quantifying levels of elicited antibodies. The ability of the indirect - ELISA to detect antibodies was evaluated in several groups of experimentally infected rainbow trout and feral Oncorh ynchus spp. The antibody response observed in the experimentally infected fish suggested their response was relatively short - lived. On the contrary, data from the spawning adult fish provided evidence of elevated antibodies, implying that they have been e xposed multiple times to R. salmoninarum while existing in the GLB. A thorough analysis of R. salmoninarum infections was also conducted on over 600 CHS returning to spawn at several GLB gamete - collecting weirs to assess infection levels, shedding, and po tential disease progression. Results suggested that female fish could be more susceptible to R. salmoninarum infection, but males may still play a role in shedding of the bacterium. Low levels of circulating anti - R. salmoninarum antibodies observed in th e late spawning CHS, along with high intensities of infection, are evidence that fish returning to spawn later in the season are more heavily infected with R. salmoninarum than those early in the season. To this end, the detection of R. salmoninarum appea rs to depend upon several factors (i.e., type of sample and assay, time of collection, and fish age) that should be taken into consideration before deciding how to diagnose BKD in a particular fish population. iv This dissertation is dedicated to m y two loves, Jay and Cohen . v ACKNOWLEDGEMENTS I would first like to give my sincere thank s to my major advisor, Dr. Mohamed Faisal, for his patience, first and foremost, as well as the guidance , wisdom , and life lessons that he has shared with me these past 10 years. I will never forget our first meeting at Lake St. Clair , and I owe my desire to educate myself as far as I have to him and his unwavering encouragement, for he saw something in me that I never did . Without his careful direction du ring my graduate career , I would not have been able to be as successful as I was. I thank him for his confidence in me, his support , and his direction , and I will always consider him my mentor and my friend . In addition, I would like to genuinely than k my committee member s: Dr. Travis Brenden, Dr. Kim Scribner, and Dr. Amber Peters . I am extremely grateful for contributions and assistance they provided to my dissertation , as well as their advice and support throughout this process . Many thanks are due to past and current members of the Michigan State University - Aquatic Animal Health Laboratory for their invaluable assis tance throughout my research, as well as their everlasting friendship. Specifically, I would like to thank Mic helle Gunn, Dan Bjorkland, Dr. Thomas Loch, Dr. Andrew Winters, Dr. Robert Kim, Dr. Elena Millard , Danielle Van Vliet, Elizabeth Throckmorton, Isaac Standish, and so many more! My research efforts would not have been possible without the funding and supp ort of the United States Fish and Wildlife Service (Grant #USFWS 30181AG013) and the Michigan Department of Natural Resources for providing fish used in the experimental challenges in this study. vi Lastly, with my deepest gratitude, I would like to thank my family. To my parents, Paul and Robin, and my siblings, Jennifer, Juan, Emma, and Jack, thank you for being so supportive, even when it meant not seeing me (or hearing from me) for long periods of time. You are the best family I could ask for and I lo ok forward to spending more time with you all. And most of all, thanks to the love of my life, my partner in adventure , my fiancé Jay, who I have needed by my side more than anyone throughout this process. Words cannot describe how excited I am to start my life with you. vii TABLE OF CONTENTS L . xii . xv I 1 REF ERENCES .. 5 Chapter 1 7 8 . 10 . 2.3. Isolation, culture, and cultural chara . 13 . 15 2.8. Molecular . 17 17 . . 17 .. 3.1.2. Int .. 19 .. . .. 3.3.1. Proc . . 3.3.2. Effect of BKD on host immun . 3.3 .. 3.3.3.3. Effect of estuarine and salt - . 26 . 3.4.1. Geographical distributio .. . .. 28 . . . 29 3.4.3.2. Horizontal transmiss .. 0 . . . . .. 31 3.4.3.4. Fish as possible vectors a . viii 33 . 34 .. 34 . . 35 .. .. 36 4.2. Isolation and bacteriological identificat .. 4.3. Antigen - 37 . . 4.3.3. Enzyme linked immunosorbent a ssay (ELISA 4.3.4. Immunohistochemistry (IHC . . 40 .. 6. Discrepancies among diagnostic test . . . . 48 7.4.1. Reducing the risk of BKD int .. APPENDIX . . REFERENCES . 53 Chapter 2 Epidemiological investigation of Renibacterium salmoninarum in three Oncorhynchus spp. in . 1 . 72 .. 3.2. . 77 77 3.2.2. Pre - stocking fingerlin ... 3.4. Q - ELISA . 4.1. Renibacterium salmoninarum infection prevalence and intensity in salmon broodstocks . ix 4.2. Renibacterium salmoninarum infection prevalence and intensity in propagated pre - 4.3. Renibacterium salmoninarum in gametes of spawning broods .89 4.4. Relationship between R. salmoninarum prevale . 92 A PPENDIX .. 99 R EFERENCES 112 Chapter 3 The Use of Non - lethal Samples for the Detection of Renibacterium salmoninarum in Chinook salmon, Oncorhynchus tshawytscha (Walb . . 121 3.2. Challenge by immersion bath 3.3. Challenge by intraperitoneal inj 3.5. Bacterial culture and isolatio .. .. 3.8. Q - .. . . 4.1. Detection in non - lethally collecte 4.3. Comparison of R. salmoninarum detectio 4.4. Comparison of diagn ostic assays am .. 4.5. Population disease composition and esti . A R EFERENCES .. 157 Chapter 4 Evidence of anti - Renibacterium salmoninarum Antibodies in Naturally and Experimentally Infected Oncorhynchus 4 . 5 6 .. 7 3.1. Determination of anti - R. salmoninarum antibodies in experimentally infected .. 1 70 3.2. Protection associated with anti - R. salmoninarum 2 17 3 x 3.4. Assessment of anti - R. salmoninarum antibodies in th ree Oncorhynchus spp. returning to spawn at gamete collecting f a 4 3.5. Modified single dilution indirect ELISA to detect and quantitate R. salmoninarum 4 3.6. Stat . 6 . 6 6 4.2. Antibody production in experimentally 17 7 4.3. Detection of antibodies in adult Oncorhynchus 8 4.4. Potential protective role of circulating binding antibodies to challenge with live R. salmoninarum 9 80 A PPENDIX 4 REFERENCES .. 19 4 Chapter 5 Efficacy of Current Testing Procedures of Spawning Chinook Salmon ( Oncorhynchus tshaw ytscha ) in Minimizing the Introduction of Renibacterium salmoninarum into Michigan . .. 9 .. 200 20 1 .. 4 4 5 3.3. Bacterial culture and isolation .. 5 .. 6 . 7 3.6. Q - .. 8 .. 9 . 10 . 1 4.1. Prevalence and semi - quantitation of R. salmoninarum by Q - 1 4.2. Detection of R. salmoninarum a 2 4.2.1. Results of Q - . 3 .. 3 4.2.3. Results of bacterial cult . 4 4.2.4. Detection of circulat ing antibodies against R. salmoninarum by 5 5 . 7 A . . 22 3 R EFERENCES .. 23 3 xi Chapter 6 .. 7 . 8 2. Future Researc 1 xii L ist of Tables Table 1.1. Summary of the morphological and biochemical characteristics of Renibacterium salmoninarum (Eissa, 2005; Austin and Austin, 2007). *API - ZYM is a bacterial enzyme based assay used for the specific identification of dif Table 2.1. Enhanced biosecurity measures to control Bacterial Kidney Disease that have been implemented at Michigan Department of Natural Resources gamete - collec ting weirs and 100 Table 2.2. The number of spawning Oncorhynchus spp. analyzed for the presence of R. salmoninarum antigens from 2001 to 2010. Hinchenbrooke coho salmon were not collected Table 2.3. The number of propagated Oncorhynchus spp. reared in state fish hatcheries and tested for Renibacterium salmoninarum antigens prior to stocking. State fish hatchery facil ities included the Platte River State Fish Hatchery (PRSFH), the Thompson State Fish Hatchery (TSFH), and the Wolf Lake State Fish Hatchery (WLSFH). *Species/strains were not reared at 2 Table 2.4. Listing and description of models fit to the R. salmoninarum prevalence and shedding data. Analysis column indicates models fit only to propagated pre - stocking fingerling prevalence . 3 Table 3.1. The prevalence of Renibacterium salmoninarum (with the number of positive fish) detected from the lethally (kidney and spleen homogenate) and non - lethally (combined, uro - fecal, blood, and mucus) collected samples from each sampling p eriod (days post - infection) and injection and immersion challenge Table 3.2. Diagnostic performance of the lethal and no n - lethal samples (uro - fecal, blood, and mucus) from the injection and immersion challenges, and a combination of the challenges . The accuracy and sensitivity (with 95% confidence intervals), and the area under the curve (AUC) values are reported. *AUC va lues that have acceptable (0.7 - 0.8), excellent (0.8 - 0.9), outstanding (0.91 - 0.99), or perfect (1.00) discrimination. AUC values < 0.6 are no better than chance alone xiii Table 3.3. Diagnostic performance of all of the lethally and non - lethally (uro - fecal, blood, and mucus) collected samples from the combined injection and immersion challenges. The prevalence (with the number of positive fish), accuracy and s ensitivity (with 95% confidence intervals), and the area under the curve (AUC) values are reported as detected by bacterial culture, nested polymerase chain reaction (PCR), semi - quantitative enzyme - linked immunosorbent assay (ELISA), and by all assays combined. *AUC values that have acceptable (0.7 - 0.8) o r excellent (0.8 - 0.9) discrimination. AUC values < 0.6 are no better than chance alone. Table 3.4. The prevalence of Renibacterium salmoninarum (with the number of positive fish) detected from leth al (kidney and spleen homogenate) samples only (L+/UF - ), uro - fecal samples only (L - /UF+), and both lethal and uro - fecal samples (L+/UF+) using results from ELISA for the lethal samples and from nPCR for the uro - fecal samples. Samples were collected from t he injection challenge, immersion challenge, and a combination of the injection and immersion challenge Table 3.5. Diagnostic performance of lethally collected samples (LETH), non - lethally collected sam ples (NON), and a combined uro - fecal and kidney/spleen sample (UF/KS) from the combined injection (sub - acute) and immersion (chronic) challenges at varying percentages in a fish population. Lethal and non - lethal samples were tested by all assays (bacteria l culture, nested polymerase chain reaction (nPCR), and the semi - quantitative enzyme - linked immunosorbent assay (ELISA), while the uro - fecal and kidney/spleen sample was tested by nPCR and ELISA, respectively . The accuracy, sensitivity, and area under the curve (AUC) values are reported. *AUC values of acceptable (0.7 - 0.8), excellent (0.8 - 0.9), outstanding (0.91 - 0.99), or perfect (1.00) discrimination. AUC values < 0.6 are no better than chance alone Ta ble 3.6. The estimated samples size needed for testing to conclude with 95% confidence that disease is not present in a population of 10,000 fish was calculated at a minimum expected prevalence (MEP) of 5 and 10% for lethally collected samples (LETH), non - lethally collected samples (NON), and a combined uro - fecal and kidney/spleen sample (UF/KS) from the combined injection (sub - acute) and immersion (chronic) challenges at varying percentages in a fish population. Lethal and non - lethal samples were tested b y all assays (bacterial culture, nested polymerase chain reaction (nPCR), and the semi - quantitative enzyme - linked immunosorbent assay (ELISA), while the uro - fecal and kidney/spleen sample was tested by nPCR and ELISA, respectively Table 4.1. Summary of the presence of anti - Renibacterium salmoninarum antibodies, p57 antigen, and the bacterium in blood and kidney and spleen samples from experimentally and naturally infected Oncorhynchus spp. An indirect enzyme - link ed immunosorbent assay (ELISA) xiv was used to determine the number of fish producing antibodies (No. Ab+ fish) with the resultant mean optical density (OD) value. Active infections with R. salmoninarum were assessed by bacterial culture [No. K/S+ fish cultur e)] and sandwich ELISA [No. K/S+ fish (ELISA)]. The intensity of infection as determined by ELISA was designated as low (L), medium (M), or high (H). ND = not determined Table 4.2. Summary of the survival of naï ve non - immunized (NPn) and naïve immunized (NPi) rainbow trout ( Oncorhynchus mykiss ) in the weeks post - infection with Renibacterium salmoninarum . An indirect enzyme - linked immunosorbent assay (ELISA) was used to determine the number of fish producing anti bodies (No. Ab+ fish) with the resultant mean optical density (OD) value. Active infections with R. salmoninarum were assessed by bacterial culture [No. K/S+ fish culture)] and sandwich ELISA [No. K/S+ fish (ELISA)]. The intensity of infection as determi ned by ELISA was designated as low (L), medium (M), or high (H) Table 5.1. Diagnostic disease stages were determined based on R. salmoninarum recovered from a kidney and spleen homogenate and the presence of circulating antibodies as dete cted by agglutination. Nested polymerase chain reaction (nPCR), culture, semi - quantitative enzyme - linked immunosorbent assay (Q - ELISA), and agglutination (Agglut.) were used to determine disease stages 223 T able 5.2. The prevalence of Renibacterium salmoninarum detected in the blood, a kidney and spleen homogenate, and reproductive fluids from male ( ) and female ( ) Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at the Little Manistee River Weir. Detection of R. salmoninarum as circulating antibodies in the blood was determined by an agglutination assay (Agglut.), while the prevalence of R. salmoninarum in the kidney and spleen homogenate (k/s) was detected by bacterial culture, nested polymerase chain reaction (nPCR), and semi - quantitative enzyme - linked immunosorbent assay (Q - ELISA). Occurrence of R. salmoninarum in the reproductive f luids (gam) was also determined by Q - ELISA. ND = not done . xv List of Figures Figure 2.1. The Michigan Department of Natural Resources state fish hatcheries and gamete - collecting weirs where Chinook ( Oncorhynchus tshawytscha ) and coho sal mon ( O. kisutch ) and steelhead ( O. mykiss ) were collected from 2001 to 2010: Little Manistee River Weir (44°11'51.66"N, 86°11'38.99"W), Platte River Weir and State Fish Hatchery (44°39'48.88"N, 85°56'13.20"W), Swan River Weir (45°24'10.09"N, 83°44'5.52"W ) , Thompson State Fish Hatchery (45°57'16.07"N, 86°15'29.36"W), and Wolf Lake State Fish Hatchery (42°17'40.14"N, 85°47'2.29"W). Figure 2.2. The prevalence of R. salmoninarum in Chinook salmon ( Oncorhynchus tshawytscha ) broodstock from the Little Manistee River Weir (LMRW - CHS) and the Swan River Weir (SRW - CHS), Hinchenbrooke coho salmon ( O. kisutch ) broodstock (HB - COS) and Michigan - adapted coho salmon broodstock (MI - COS) from the Platte River Weir, and steelh ead ( O. mykiss ) broodstock from the Little Manistee River Weir (LMRW - STT) from 2001 to 2010. F - J: The low, medium, and high intensity levels of infection of R. salmoninarum in LMRW - CHS, SRW - CHS, HB - COS, MI - COS, and LMRW - STT from 2001 to 2010. The lines re present the logistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote . 107 Figure 2.3. The prevalence of R. salmoninar um in Chinook salmon ( Oncorhynchus tshawytscha ) pre - stocking fingerlings propagated at the Wolf Lake State Fish Hatchery (WLSFH), the Platte River State Fish Hatchery (PRSFH), and the Thompson State Fish Hatchery (TSFH) from 2002 to 2010. Fingerlings are the progeny of broodstock spawned at the Little Manistee River Weir (LMRW) and the Swan River Weir (SRW). F - J: The low, medium, and high intensity levels of infection of R. salmoninarum in LMRW - CHS from WLSFH and PRSFH and SRW - CHS from the WLSFH, PRSFH, a nd TSFH from 2002 to 2010. The lines represent the logistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote 108 Figu re 2.4. The prevalence of R. salmoninarum in the Hinchenbrooke strain (HB - COS) of coho salmon ( Oncorhynchus kisutch ) pre - stocking fingerlings and the Michigan - adapted strain of coho salmon (MI - COS) propagated at the Platte River State Fish Hatchery (PRSFH) from 2003 to 2010. C - D: The low, medium, and high intensity levels of infection of R. salmoninarum in HB - COS and MI - COS propagated at the PRSFH from 2003 to 2010. The lines represent the logistic regression predicted prevalence and intensity levels of in fection, while open circles and triangles denote the observed prevalence and intensity xvi Figure 2.5. The prevalence of R. salmoninarum in steelhead ( Oncorhynchus mykiss ) pre - stocking fingerlings (LMRW - STT) propagated at th e Wolf Lake State Fish Hatchery (WLSFH) and the Thompson State Fish Hatchery (TSFH) from 2005 to 2010. C - D: The low, medium, and high intensity levels of infection of R. salmoninarum in LMRW - STT propagated at the WLSFH and TSFH from 2005 to 2010. The line s represent the logistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote the observed prevalence .. Figure 2.6. The prevalen ce of male and female fish that may be shedding R. salmoninarum and are not shedding R. salmoninarum for (A) Chinook salmon ( Oncorhynchus tshawytscha ) broodstock from the Little Manistee River Weir; (B) Chinook salmon broodstock from the Swan River Weir; ( C) Hinchenbrooke coho salmon ( O. kisutch ) broodstock from the Platte River Weir; (D) Michigan - adapted coho salmon broodstock from the Platte River Weir; (F) and steelhead ( O. mykiss ) broodstock from the Little Manistee River Weir. The lines represent the l ogistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote the observed prevalence and intensity Figure 3.1. Characteristic bacterial kidney disease clinical signs observed in the injection challenge included (A) petechial hemorrhage throughout the body of the fish (B) gill pallor, (C) corneal opacity with associated hemorrhage, (D) false membranes overlying the liver, and (E) granulomatous - like lesions in the kidney Figure 3.2. Characteristic bacterial kidney disease clinical signs observed in the immersion challenge included (A) petechial hemorrhages in the caudal peduncle region, as well as between the fin rays (black arrow), and (B) exophthalmia and associated hemorrhage Figure 4.1. The Michigan Department of Natural Resources gamete - collecting weirs where Chinook ( Oncorhynchus tshawytscha ), coho salmon ( O. kisutch ), and steelhead ( O. mykiss ) were collected in 2009 and 2 013: the Little Manistee River Weir (Chinook salmon and steelhead) and the Platte River Weir (coho salmon) Figure 4.2. The prevalence of Renibacterium salmoninarum in the kidney/spleen tissue samples of A) 8 month ol d naïve rainbow trout ( Oncorhynchus mykiss ), and B) age cohort rainbow trout that have received two doses of R. salmoninarum bacterin per os at each sampling period (weeks post - infection with live bacteria). Renibacterium salmoninarum and its p57 antigen were detected by the sandwich enzyme - linked immunosorbent assay, with the infection intensity expressed as the proportion of fish exhibiting low, medium, and high levels of infection (n=5 per sampling periods; n=45 at week 26 p.i.). The average number of R . salmoninarum colony forming units (CFUs) isolated from kidney is depicted as a line graph xvii Figure 4.3. The prevalence of Renibacterium salmoninarum in the kidney/spleen tissue samples of A) 11 month old rainbow trout that survived an infection with li ve R. salmoninarum 16 weeks prior to being used in this experiment B) age cohort survived rainbow trout that have received two doses of R. salmoninarum bacterin per os at each sampling period (weeks post - infection with live bacteria). Renibacterium salmon inarum and its p57 antigen were detected by the sandwich enzyme - linked immunosorbent assay, with the infection intensity expressed as the proportion of fish exhibiting low, medium, and high levels of infection (n=5 per sampling period; n=44 - 45 at week 26 p .i.). The average number of R. salmoninarum colony forming units (CFUs) isolated from kidney is depicted as a line graph . Figure 4.4. The mean antibody response (± SE) of A) 8 month old naïve rainbow trout ( Oncorhynchus mykis s ), and B) age cohort rainbow trout that have received two doses of a R. salmoninarum bacterin per os at each sampling period (weeks post - infection with live bacteria). The average optical density (OD) value was used to evaluate the production of anti - R . salmoninarum antibodies in fish. A separate group of naïve fish were used to determine the positive - negative threshold (dashed line), which was the average OD value plus two standard deviations (i.e., 0.094) Figure 4.5. The mean antibody response (± SE) of A) 11 month old rainbow trout that survived an infection with live R. salmoninarum 16 weeks prior to being used in this experiment B) age cohort survived rainbow trout that have received two doses of a R. s almoninarum bacterin per os at each sampling period (weeks post - infection with live bacteria). The average optical density (OD) value was used to evaluate the production of anti - R. salmoninarum antibodies in fish. A separate group of naïve fish were used to determine the positive - negative threshold, which was the average OD value plus two standard deviations (i.e., 0.094) Figure 4.6. The mean circulating antibody levels of Chinook salmon ( Oncorhynchus tshawytscha ), steelhead ( O. mykiss ), a nd coho salmon ( O. kisutch ) returning to spawn at the Little Manistee River Weir and the Platte River Weir. The data are presented as box and whisker plots, where the central box contains the interquartile range and the median is represented as a horizont The average optical density (OD) value was used to evaluate the production of anti - R. salmoninar um antibodies in fish. A separate group of naïve fish were used to determine the positive - negative threshold (dashed line), which was the average OD value plus two standard deviations (i.e., 0.094) Figure 4.7. The survival probability of 11 month old naïve rainbow trout ( Oncorhynchus mykiss ) (NPn) and age cohort rainbow trout that have received one dose of R. salmoninarum bacterin per os (NPi) in the weeks post - infection with live Renibacterium salmoninarum xviii Figure 5.1. The Michigan Department of Natural Resources gamete - collecting weirs where Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn were collected in September and October of 2005: Little Manistee River Weir, Medusa Creek Weir, a nd Boardman River Weir (Lake Michigan watershed), and Swan River Weir (Lake Huron watershed) Figure 5.2. The prevalence and intensity (high, medium, low) of Renibacterium salmoninarum detected by the quantitative enzyme - linked immunosorbent assay in a kidney and spleen homogenate from A) all Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at Little Manistee River Weir (LMRW), Medusa Creek Weir (MCW), Boardman River Weir (BRW), and the Swan River Weir (SRW), and B) male (M) and f emale (F) Chinook salmon returning to spawn at LMRW, MCW, BRW, and SRW. Figure 5.3. The prevalence and intensity (high, medium, low) of Renibacterium salmoninarum detected by the quantitative enzyme - linked immunosorbent assay in reproductive fluids from A) all Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at Little Manistee River Weir (LMRW), Medusa Creek Weir (MCW), and the Swan River Weir (SRW), and B) male (M) and female (F) Chinook salmon returning t o Figure 5.4. The proportion of male and female Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at the Little Manistee River Weir (LMRW), Medusa Creek Weir (MCW), and Swan River Weir (SRW) that may be s hedding Renibacterium salmoninarum in their reproductive fluids (kid+/gam+, kid - /gam+) and are not shedding R. salmoninarum (kid - /gam - , kid+/gam - ), as detected by the quantitative enzyme - linked immunosorbent assay Figure 5.5. The prevalence and intensity (high, medium, low) of Renibacterium salmoninarum detected by the quantitative enzyme - linked immunosorbent assay in a kidney and spleen homogenate from A) all Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at the Little Manistee Ri ver Weir during the pre - season, early, and late sampling periods, and B) male (M) and female (F) Chinook salmon during the same sampling periods Figure 5.6. The prevalence and intensity (high, medium, low) of Renibacterium salmoninarum d etected by the quantitative enzyme - linked immunosorbent assay in reproductive fluids from A) all Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at the Little Manistee River Weir during the pre - season, early, and late sampling periods, and B) male (M) and female (F) Chinook salmon during the same sampling periods Figure 5.7. Proportion of Chinook salmon ( Oncorhynchus tshawytscha ) collected from Little Manistee River Weir in each disease stage (1 - 6) during the early and late sampling periods, xix including A) all Chinook salmon, B) female Chinook salmon, and C) male Chinook salmon. Disease stages were determined based on results from nested polymerase chain reaction (nPCR), bacterial culture, semi - quantitative enzyme - lin ked immunosorbent assay (Q - ELISA), and agglutination (Agglut). Stage 1 represents the onset of infection (nPCR+, culture - , Q - ELISA - , Agglut - ), stage 2 represents a settled infection (nPCR+, culture+, Q - ELISA+/ - , Agglut - ), stage 3 represents an active infe ction (nPCR+, culture+, Q - ELISA+, Agglut - ), stage 4 represents the beginning of remission (nPCR+, culture - , Q - ELISA+, Agglut - ), stage 5 represents advanced remission (nPCR - , culture - , Q - ELISA+, Agglut+), and stage 6 represents recovering from an infection (nPCR - , culture - , Q - ELISA - 1 I ntroduction Bacterial kidney disease (BKD), caused by the Gram - positive bacterium Renibacterium salmoninarum , is a systemic disease that causes chronic infections and m ortalities in wild and aqua - cultured salmonid fish populations worldwide ( Paterson et al., 1979; Evelyn et al., 1981; Holey et al., 1998). Renibacterium salmoninarum is an obligate intracellular pathogen that invades phagocytic cells (e.g., macrophages), immune system, allowing it to proliferate (Gutenberger et al., 1997). Most notably, R. salmoninarum secretes excessive amounts of a water - soluble cell surface 57 kilodalton heavily glycosylated protein (p57) that enables the bacterium to spread easily throughout the fish host. These unique characteristics of R. salmoninarum permit the bacterium to survive within the fish host relatively undetected for long periods of time, resulting in chronic infections. While R. salmoninarum has been detected in non - salmonids, it has a predilection for the kidneys of many salmonid species (MacLean and Yoder, 1970; Mitchum et al., 1979; Traxler and Bell, 1988). In the Great Lakes basin (GLB), salmonids play a critical r ole in the ecosystem, as they are one of the top predators and are a large part of a $7 billion per year fishing industry (Tsehaye et al., 201 4 ). Since the 1970s, the Michigan Department of Natural Resources (MDNR) has been aiming to reduce the presence a nd effects of BKD in the Great Lakes salmon populations through regular health inspections and multiple prevention and control measures. Routine disease screening is an effective way to detect R. salmoninarum in a population, yet the stage of infection a fish is experiencing and the subsequent positive or negative result depends on the detection method that is used. For example, a sandwich enzyme - linked immunosorbent 2 assay (ELISA) is a common method used for detection of R. salmoninarum in a kidney and sp leen sample. ELISAs are preferred by many professionals because they can test many samples at once in a relatively short period of time. However, as most BKD ELISAs test for the presence of the p57 antigen, a positive result usually indicates an active o r a recent infection, but does not elaborate on the status of infection. Using multiple diagnostic assays that inform on other characteristics of the bacterium (i.e., detection of nucleic acids, viable bacteria, or specific antibodies) and the resultant s tage of infection could influence the management and treatment decisions made by fishery and hatchery biologists. Furthermore, as t he kidneys and spleen of fish remove circulating bacterial antigens from the host, they are commonly used to provide indicati on of an infection (Bruno and Poppe, 1996). However, little is known regarding the use of other tissues (e.g., mucus, blood, feces) for detection of the pathogen. However , fish must be euthanized in order to collect the necessary kidney and spleen sample s to detect R . salmoninarum . Not only could tissues such as mucus, blood, or feces reduce the need to sacrifice fish, but detection of R . salmoninarum in these tissues could also contribute to a better understanding of how the bacterium spreads in a fish host after infection. By using a single diagnostic assay (e.g., ELISA, PCR, or bacterial culture), it is likely that the full range of infections (onset, established, active, recovering, and remission) will not be detected, thus potentially resulting in an underestimated prevalence of disease, which can impact treatment decisions for the fish population. Furthermore, as the sandwich ELISA detects the presence of the p57 antigen found on the outer surface of R. salmoninarum , it reveals whether a fish has been exposed to the bacterium, but not if the fish is producing antibodies. It 3 would be highly beneficial to know if the fish was mounting an immune response , and therefore would not require treatment with antibiotics. Lastly, sacrificing fish to collect a kidney and spleen sample for BKD testing can be extremely difficult for valuable broodstocks and endangered species. To this end, the overall objective of this study was to evaluate the various shortcomings and assets of currently used and new diagnost ic assays and sample types for the detection of R. salmoninarum . Firstly, a thorough review of literature regarding the history, characteristics, geographical and host distribution, diagnostic methods for detection, and prevention and control measures for BKD is given in Chapter 1 . Chapter 2 of this study evaluated the implementation of several enhanced biosecurity measures conducted by the state of Michigan at their hatcheries and gamete - collection weirs. Substantial declines in the overall prevalence o f BKD were detected in each of the examined fish stocks, with differences in the prevalence of BKD observed in the different species and strains. In Chapter 3, in addition to kidney and spleen samples, other samples types (mucus, blood, and a urine/feces mixture) were also tested by molecular, serological, and standard bacterial techniques to assess their ability to detect BKD, taking the effects of exposure route and the resultant disease course into consideration. Additionally, the number of fish that w ould be required to sample non - lethally was investigated. Analyses suggested that the urine/feces mixture had the best potential to detect R. salmoninarum compared to the mucus or blood. Detection was further enhanced when the urine/feces mixture and the kidney and spleen sample were tested in conjunction. 4 In Chapter 4, a single - dilution indirect ELISA was evaluated with the aim of detecting anti - R. salmoninarum antibodies in experimentally infected and feral Oncorhynchus spp. All groups of experimental ly infected and feral fish were found to produce detectable levels of antibodies, although there were minimal differences among them. While the feral fish produced elevated levels of antibodies, the observed antibody response in experimentally infected fi sh was short - lived and long - term elevated levels of antibody production were not observed. Chapter 5 investigated the prevalence and intensity of R. salmoninarum infections in male and female Chinook salmon returning to spawn at multiple gamete - collect ing weirs in Michigan using several detection methods. Also, potential diagnostic patterns indicative of a progressive R. salmoninarum were considered. A greater prevalence of BKD infection was observed at the Swan River Weir, with female Chinook salmon appearing to be more susceptible to R. salmoninarum than males. Lastly, evidence of disease progression was observed, with earlier stages and more intense infections occurring later in the spawning run. Lastly, the overall conclusions and recommendations for future research are presented in Chapter 6. This research highlights that despite the encouraging decline of BKD in the GLB over the past decade, there is still much we need to learn to maximize what is achieved with routine health inspections. Opti mal opportunities to improve management and treatment decisions are being overlooked and neglected. More consideration needs to be taken with regards to how the chosen assay and sample type can affect the diagnosis of disease. 5 R EFERENCES 6 REFERENCES Bruno, D.W., and Poppe, T.T. 1996. A Colour Atlas of Salmonid Diseases. Academic Press Ltd, London, 149 pp. Evelyn, T.P.T., Ketcheson, J.E., and Prosperi - Porta, L. 1981. The clinical significance of immunofluorescence - based diagnoses o f bacterial kidney disease carriers. Fish Pathol. 15: 293 - 300. Gutenberger, S.K., Duimstra, J.R., Rohovec, J.S., and Fryer, J.L. 1997. Intracellular survival of Renibacterium salmoninarum in trout mononuclear phagocytes. Dis. Aquat. Org. 28: 93 - 106. Hole y, M.E., Elliot, R.F., Marcquenski, S.V., Hnath, J.G., and Smith, K.D. 1998. Chinook salmon epizootics in Lake Michigan: possible contributing factors and management implications. J. Aquat. Anim. Health 10: 202 - 210. MacLean, D.G., and Yoder, W.G. 1970. Ki dney disease among Michigan salmon in 1967. Prog. Fish - Cult. 32: 26 - 30. Mitchum, D.L., Sherman, L.E., and Baxter, G.T. 1979. Bacterial kidney disease in feral populations of brook trout ( Salvelinus fontinalis ), brown trout ( Salmo trutta ), and rainbow trou t ( Salmo gairdneri ). J. Fish. Res. Board Can. 36: 1370 - 1376. Paterson, W.D., Gallant, C., Desautels, D., and Marshall, L. 1979. Detection of bacterial kidney disease in wild salmonids in the Margaree River system and adjacent waters using an indirect fluo rescent antibody technique. J. Fish. Res. Board Can. 36: 1464 - 1468. Traxler, G., and Bell, G. 1988. Pathogens associated with impounded Pacific herring Clupea harengus pallasi , with emphasis on viral erythrocytic necrosis (VEN) and atypical Aeromonas salm onicida . Dis. Aquat. Org. 5: 93 - 100. Tsehaye , I., Jones, M., Bence, J.R., Brenden, T.O., Madenjian, C.P., and Warner, D.M. 2014 . A multispecies statistical age - structured model to assess predator - prey balance: application to an intensively managed Lake Mi chigan pelagic fish community . Can. J. Fish. Aquat. Sci. 71 : 627 - 644 . 7 Chapter 1 Literature R eview 8 1. Historical b ackground Renibacterium salmoninarum is the causative agent of Bacterial Kidney Disease (BKD), a serious bacterial d isease of salmonid fish species worldwide. Bacterial Kidney Disease was first described in 1930 in Atlantic salmon ( Salmo salar ) from the river s Dee and Spey in Aberdeenshire, Scotland (Mackie et al., 1933; Smith, 1964). The disease was initially referred to as Dee Disease, and has also been known as Kidney Disease, Salmonid Kidney Disease, and Corynebacterial Kidney Disease (Fryer and Sanders, 1981). G ranulomatous lesions, which are the characteristic clinical sign of BKD , were observed in the spleen and other int ernal organs of Atlantic salmon, along with a small gram - positive diplo bacillus (Mackie et al., 1933). Shortly after the initial discovery of BKD in Scotland, a similar bacterial disease was described in brook trout ( Salvelinus fontinalis ) , brow n trout ( Salmo trutta ) , and rainbow trout ( Oncorhynchus mykiss ) at a Massachusetts state fish hatchery (Belding and Merrill, 1935). Several external symptoms were observed, including exophthalmia, raised, fluid - filled blebs, abscesses, and reddening at th e base of pectoral fins (Belding and Merrill, 1935). Granulomatous lesions similar to those described by Mackie et al. (1933) were noted in the kidneys, as well as in the livers and , to a lesser degree, in the spleens, along with excess fluid in the perit oneal and pericardial cavities, and hemorrhagic ovaries and testes (Belding and Merrill, 1935). Again, small, short, gram - positive, diplobacilli bacteria were recovered from the lesions associated with diseased fish (Belding and Merrill, 1935). Subseque ntly, BKD was discovered in British Columbia, Canada in 1937 (Evelyn, 1986), followed by detections in the Pacific Northwest in the lat e 1940s, where a n unknown k idney 9 disease infecting sockeye salmon ( O. nerka ), Chinook salmon ( O. tshawytscha ), and coho s almon ( O. kisutch ) from state fish hatcheries in California, Oregon, and Washington was described , which was later confirmed to be BKD (Rucker et al. , 1951; Earp et al., 1953). B acterial kidney disease then spread to the Laurentian Great Lakes Basin (LGLB ) in the 1950s with the introduction of salmonids and their eggs from the Pacific Northwest (Allison, 1958) , and has since been detected in several wild and aqua - culture d salmonid and coregonid species throughout the LGLB (MacLean and Yoder, 1970; Holey e t al., 1998; Beyerle and Hnath, 2002; Jonas et al., 2002; Faisal et al., 2012). It was first found in Michigan in 1955 in brook trout from two state fish hatcheries (Allison, 1958). Furthermore, reports have linked the disease to mortalities in wild salm onids from Canada (Paterson et al., 1979; Evelyn et al., 1981), and mass fish kills in the LGLB (Holey et al., 1998). In 1968, BKD was detected in farmed Atlantic salmon from Iceland (Helgason, 1985), and then became widespread in several salmonid specie s in Europe , including Denmark (Lorenzen et al., 1997), Germany (Hoffman et al., 1984), Finland (Rimaila - Pärnänen, 2002), France (DeKinkelin, 1974), Italy (Ghittino et al., 1977), Norway ( Jansson , 2002 ), Poland (Kozin´ska et al., 2001), Spain (Martínez - Mil lán, 1977), Sweden (Jansson et al., 1996), former Yugoslavia (Fijan, 1977 ), North America (United States of America and Canada; Souter et al., 1987), and Japan (Kimura and Awakura, 1977). The disease has also spread to Chile (Sanders and Barros, 1986) and it is widely accepted that BKD is prevalent in all parts of the world where wi ld or cultured salmonids exist . 10 2 . The pa thogen 2.1. Nomenclat ure and current classification Initially, the causative bacterium of BKD was suggested to be a member of the g enus Corynebacterium , based on the Gram stain properties and morphology (Ordal and Earp, 1956; Smith, 1964). However, further taxonomic analyses revealed several d eviations from typical Corynebacterium characteristics, including the absence of mycolic aci d, differences in the sugar and amino acid compositions of the peptidoglycan cell wall layer, as well as a different guanine plus cyto sine (G + C) content of DNA (Sanders and Fryer, 1980). As such, the authors pro posed that this bacterium form s a single s pecies in the new genus Renibacterium and they identified the bacterium as Renibacterium salmoninarum (Sanders and Fryer, 1980) . Furthermore, sequencing of the 16S ribosomal ribonucleic acid (rRNA) from R. salmoninarum and additional evaluation of the G + C content has placed the organism in the order Actinomycetales and family Micrococcaceae (Banner et al., 1991; Gutenberger et al., 1991). Arthrobacter and Micrococcus spp. are the closest relatives to R. salmoninarum (Holt et al., 2001). 2. 2 . Cell morph ology Renibacterium salmoninarum is a short coccobacilli (0.3 - 1.5 by 1 . 0 - 1. 5 µm), Gram - positive, non - sporing, non - capsulated, non - motile, and non - acid fast bacterium that is arranged singly, in pairs (diplo) and rarely as short chains (Sanders and Fryer, 1980 ; Fryer and Lannan, 11 1993 ). Renibacterium salmoninarum consists of two regions: a central region with lightly stained filaments ( generally regarded to represent DNA) and a peripheral region filled with small, electron dense ribosomes (Young and Chapma n, 1978). 2.3. Isolation, culture, and cultural characteristics Renibacterium salmoninarum is a slow - growing , fastidious organism which has been difficult to isolate on ordinary bacterial media (Sanders and Fryer, 1980). Earp et al. (1953) initially c u ltured the bacterium from infected kidney tissues on an artificial medium that consisted of fish extract, glucose, yeast extract, and meat infusion in agar. The authors achieved limited growth with the first appearance of colonies after more than two week s of incubation (Earp et al., 1953). Also, w hen the same authors used minced chick embryo tissues to 0.1% L - cysteine hydrochloride (HCl) has further enhanced the growth of R. salmoninarum upon primary isolation (Ordal and Earp, 1956). Moreover , Ordal and Earp (1956) developed the Kidney Disease Medium 1 (KDM1) w hich consisted of: tryptose 1.0 %, beef extract 0.3%, NaCl 0.5%, yeast extract 0 .05%, L - cysteine - hydrochloride 0.1%, human blood 20 %, and agar 1.5 %. This medium is also referred to as Cysteine Blood Agar Medium ( Ordal and Earp, 1956). Additionally, Wolf and Dunbar (1959) achieved fair growth on Mueller - Hinton medium, also supplement ed with cysteine . K idney D isease M edium 1 was later modified by Evelyn (1977) by replacing the human blood, tryptose, and beef extract with 20% fetal bovine serum and peptone, which was then 12 designated as KDM2. Furthermore, to reduce the time needed for primary isolation, which can be up to 6 weeks, Evelyn et al. (1989) added 25 µ l of heavy inoculum of R. salmoninarum , which can accelerate the growth in primary cultures. E velyn et al. (1990) were able to again improve upon this method and achieve more consistent growth of the primary culture by replacing the nurse culture with 25 µ l of filter - sterilized R. salmoninarum spent medium. The major drawbacks of KDM2 medium, howe ver, a re the high cost and presence of serum proteins, which hinder the identification of proteins of bacterial origin. There are a number of serum - free media for R. salmoninarum growth that have been developed. Embley et al. (1982) developed a serum - free, semi - defined growth medium that supported secondary, but not primary, growth of R. salmoninarum . Also, Daly and Stevenson (1985) formulated the Charcoal Agar Medium in which they substituted activated charcoal for serum. Starliper et al. (1998) co mpared the performance of 13 serum - free media and 1 serum - supplemented media for the growth of R. salmoninarum isolates and found that there were no significant differences among the 14 medium formations when mean cell counts were compared after 10, 20, an d 30 days of incubation. Additionally, Austin et al. (1983) incorporated four antibiotics (cycloheximide, D - cycloserine, oxolonic acid, and polymyxin B) into the KDM2 medium to control growth of other bacterial species. Austin et al. (1983) also reduced the volume of fetal bovine serum from 20% to 10% and designated the medium Selective KDM (SKDM; Austin et al., 1983). These modifications significantly reduced bacterial contaminants, facilitating the growth of R. salmoninarum from clinical and environmen tal samples (Austin et al., 1983). L astly, Faisal et al. (2010) proposed modifying the SKDM to 13 incorporate 1% spent medium into the agar to enhance the growth of R. salmoninarum colonies, shorten the period of incubation, and minimize the growth of contam inating bacteria. This growth medium is known as Modified Kidney Disease Medium (MKDM) and is the only one to incorporate all of the improvements of the other media, cysteine, spent nurse medium, and antibiotics, into one medium (Faisal et al., 2010). Renibacterium salmoninarum does require cysteine as an ingredient in the growth medium and grows best at 15°C, slowly at 5 or 22°C, and does not grow at all above 37°C (Smith, 1964). While colonies can develop as early as 14 days post - inoculation, up to 19 weeks may be required for isolation of very low numbers of bacteria (Benediktsdóttir et al., 1991). Renibacterium salmoninarum colonies are creamy (non - pigmented), shiny, smooth, round, raised, entire, and 1 - 2 mm in diameter (Austin and Austin, 2007). Additionally, on cysteine supplemented solid media, old colonies (i.e., > 12 weeks) can appear extremely granular due to crystallization of cysteine (Austin and Austin, 2007). 2. 4 . Preservation of cultures Several methods have been used to preserve di fferent species of the order Actinomycetales , including Streptomyces , Actinomyces , and Renibacterium species. For long term preservation, methods such as lyop hilization (Hopwood and Ferguson, 1969) and storage under liquid nitrogen (Pridham and Hesseltine , 1975) have been successfully used. Bacterial cells can also be preser ved in diluted glycerol (10 - 20% v/v) and frozen at - 20°C ; however, repeated t hawing and freezing cycles can affect the stability and viability of the cell (Wellington 14 and Williams, 197 9). To overcome this drawback , Feltham et al. (1978) stored bacteria on glass beads in 10% (v/v) glycerol at - 76°C. The glass beads allowed removal of small samples without thawing the entire culture, which was advantageous for long - term preservation (We llington and Williams, 1979). Also, smaller aliquots can be utilized to prevent repeated thawing and freezing of larger samples. Preservation of small inocula of R. salmoninarum in KDM2 (Evelyn , 1977) or peptone saline (Starliper et al., 1997) and storag e at - 80°C were also successfully used. 2. 5 . Biochemical characteristics Renibacterium salmoninarum is cytochrome oxidase negative, catalase positive, proteolytic , and dependent upon the presence of L - cysteine - HCl for growth i n artificial media (Sanders and Fryer, 1980 ; Austin and Austin, 2007 ). Interestingly, R. salmoninarum isolates from different sources are homogeneous in their biochemical characteristics (Austin et al., 1983; Goodfellow et al., 1985; Bruno and Munro, 1986a), but the result s for a g iven test can vary depending upon the testing system used. Thus, R. salmoninarum is negative for the gelatinase and DNase reactions by the API - ZYM system (Goodfellow et al., 1985), but was positive by standard methods (Bruno and Munro, 1986a). The organi sm is also known to be - hemolytic on media supplemented with blood (Bruno and Munro, 1986a). Additionally, R. salmoninarum can liquefy gel atin, degrade Tween (20 - 60), hy drolyze casein , and is negative for esculin hydrolysis, DNase, urease, nitrate reduction, phosphatase, methy l red, indole test, and the carboh ydrate utilization test (Table 1 .1 ). 15 2. 6 . Antibiotic susceptibility Renibacterium salmoninarum isolates are sensitive to carbenicillin, cephaloridine, chloramphenicol, erythromycin, novobiocin, rifampicin, streptomycin, sulfamerazine, and tetracycline ( Wolf and Dunbar, 1959; Austin and Rodgers, 1980; Austin, 1985; Goodfellow et al., 1985). Renibacterium salmoninarum is also sensitive to enrofloxacin (Hsu et al., 1994), tiamulin, cefazolin (Bandin et al., 1991), and azit hromycin (Rathbone et al., 1999). Furthermore, the organism is resistant to D - cycloserine, oxolonic acid (4 µg/ml), polymyxin B, and cycloheximide (Wolf and Dunbar, 1959; Goodfellow et al., 1985). It should be noted that while some studies have reported sensitivity of R. salmoninarum to chloramphenicol, it has also been shown to be not as effective (Austin, 1985). 2.7. Antigenic characteristics and virulence factors Renibacterium salmoninarum is an intracellular , obligate pathogen that is able to invad e all types of fish cells, in particular, the phagocytic cells ( i.e., macrophages; Gutenberger et al., 1997; Ellis, 1999). The cap ability of R. salmoninarum to invade phagocytes or other cells depends up on certain virulence factor s (Gutenberger et al., 19 97; Ellis, 1999; Piganelli et al., 1999). It has been demonstrated that R. salmoninarum secretes a number of extracellular products (ECP) that possess proteolytic, hemolytic, and DNA degradation activities in vitro (Austin and Rodgers, 1980; Bruno and Mun ro, 1986a). Furthermore, a 65 kiloDalton ( kDa ) R. salmoninarum zinc metalloprotease - like protein has been extracted from R. salmoninarum ECP 16 that possesses hemolytic activities against a number of fish and mammalian erythrocytes. The encoding gene of the R. salmoninarum ECP with hemolytic activity was designated as hly (Grayson et al., 1995). Mo st notably , R . salmoninarum secretes a water - soluble, heat - stable, hydrophobic cell surface 57 kDa protein (p57 ) that is believed to be the major virulence determ inant of this bacterium (Getchell et al., 1985). In vitro , purified p57 exhibited both hemolytic (Daly and Stevenson, 1990) and leuco - agglutinating (Wiens and Kaattari, 1991) properties . Moreover, R. salmoninarum is able to secrete excessive amounts of t he p57 protein, enabling the bacteria to easily spread throughout the host (Evenden et al., 1993). Also, Hamel (2001) reported that R . salmoninarum isolates differed in their pathogenicity to salmonids, a finding that correlated positively with the amount of surface - associated p57. The agglutinating properties of R. salmoninarum also contribute to its pathogenicity. C hallenge of susceptible fish with auto - agglutinating strains of R. salmoninarum caused significantly higher mortality than non - auto - aggluti nating strains (Daly and Stevenson, 1990; Additionally, s oluble R. salmoninarum surface proteins possess immunosuppressive action against the salmonid specific antibody response (Turaga et al., 1987), which was attributed not only to th e p57 protein, but also to a 22 kDa surface protein (Fredriksen et al., 1997). Starliper et al. (1997) compared a number of strains of R. salmoninarum isolated from Chinook and coho salmon from the Pacific Northwest and Great Lakes regions in North America for virulence. Th os e authors found that virulence differed significantly among the isolates and concluded that isolates retrieved from Lake Michigan weirs were the most virulent compared to the Pacific Northwest (Starliper et al., 1997). Interest ingly, while strains of R. salmoninarum from Lake Michigan are considered to be more virulent, 17 Chinook salmon from Lake Michigan are more resistant to R. salmoninarum infection s than Chinook salmon from Green River, Washington (Purcell et al., 2008). 2. 8 . Molecular and genetic diversity Altho ugh the biochemical uniformity and phylogenetic homology of R. salmoninarum strains are fairly similar ( Bruno and Munro, 1986a; Gutenberger et al., 1991), a minimal molecular diversity was detected among strains isol ated from different parts of the world (Alexander et al., 2001). Using polymerase chain reaction (PCR) amplification of length polymorphisms in the tR NA intergenic spacer regions , Alexander et al. (2001) succeeded in differentiati ng among isolates of R. s almoninarum from several farms in the United Kingdom . Moreover, genetic diversity was also detected among 40 North American isolates using the mul tilocus enzyme electrophoresis assay (Starliper, 1996). T he highest genetic diversity was detected in strain s isolated from Chinook and coho salmon returning to spawn at the Little Manistee River weir in Michigan (Starliper, 1996). 3 . The d isease 3. 1 . Disease course Even though BKD develops slowly in host organisms, there are several factors that can influe nce the progress ion of BKD, including water temperature (Sanders et al., 1978; Fryer and 18 Sanders, 1981; Bullock and Herman, 1988), host factors (Evenden et al., 1993), and R. salmoninarum strain virulence (Starliper et al., 1997) , which will be discussed l ater . Moreover, various disease signs may be overt or more obscure in individuals, depending on the extent of the infection. 3.1.1. External signs Affected fish can manifest behavioral changes, as well as a wide range of external lesions that might var y according to the age and species of the fish affected and the virulence of the R. salmoninarum strain (Fryer and Sanders, 1981; Bullock and Herman, 1988; Evenden et al., 1993). Superficial blebs (or blisters) of the skin, exophthalmia, e rratic swimming behavior, deep abscesses all over the body surface, and cavitations in muscles have been reported in affected fish (Belding and Merrill, 1935; Smith, 1964; Fryer and Sanders, 1981; Bullock and Herman, 1988 ; Evenden et al, 1993 ). The blebs and cavitations m ay contain a white to yellowish or hemorrhagic fluid (Bullock and Herman, 1988). P etechial hemorrhages in muscles and fins and ascitic fluid in the peritoneal cavity have also been reported (Belding and Merrill, 1935; Earp et al., 1953; Evelyn, 1993). I n very rare cases, the external signs of the disease in coho and Chinook salmon might only be manifested by exophthalmia , associated with the accumulation of infective fluid containing large amounts of the bacteria, pus, and necrotic tissue in the enlarged eyes (Bullock and Herman, 1988 ). 19 3.1.2. Internal lesions The kidneys of affected fish usually exhibit white foci that contain leukocytes, bacteria, and host cell debris and can be swollen (Fryer and Sanders, 1981). In advanced cases, the spleen may i ncrease in size , the liver can appear very pale in color , and the kidney s are mostly grayish in color (Wood and Yasutake, 1956; Fryer and Sanders, 1981 ; Evenden et al., 1993 ). The most characteristic clinical lesions associated with BKD are the presence o f scattered nodules of various sizes over the surface of the kidneys, spleen, and liver (Belding and Merrill, 1935; Snieszko and Griffin, 1955; Klontz, 1983). In some instances , petechial hemorrhages were noticed in the muscles lining the peritoneum , ofte n associated with the accumulation of ascitic fluid (Ferguson, 1989). An opaque pseudo membrane cover ing the internal organs has also been reported, especially in fish maintained at a temperature below 9 ° C (Snieszko and Griffin, 1955; Bell, 1961; Fryer and Sanders, 1981). The pseudomembrane consists of fibrin and leukocytes (Smith, 1964). Similar membranes occur in trout at higher temperatures (12 - 13 ° C; Bullock and Herman, 1988). Hemorrhages with a white or yellow viscous fluid in the hindgut and petechi al hemorrhages were often found in the peritoneum of infected Atlantic salmon (Smith, 1964). 3.1.3. Histopathology The initial histopathological description indicated that the kidney s w ere the major organ affected by R. salmoninarum infection (Belding a nd Merrill, 1935) . I nfected brook trout and 20 brown trout demonstrated microscopic lesions in the kidney s , and to a lesser extent, in the liver and spleen (Belding and Merrill, 1935) . Lesions are typically chronic in nature , with multiple granulomas that r esemble those observed in mammalian tuberculosis (Snieszko and Griffin, 1955; Wood and Yasutake, 1956). Fibrotic lesions have also been noticed in the kidney s , spleen, liver, and intestine of infected fish , with proliferating fibro blasts forming distinct nodules that then coalesce to form large masses of affected tissues (Wood and Yasutake, 1956). It is believed that the granulomas are formed as a result of macrophage activation, followed by their adherence to each other , forming giant cells (Secombes, 1 985 ). The giant cells and activated macrophages can release large amounts of lytic enzymes into the surrounding tissues , leading to necrosis at the central part of the granuloma (Bruno, 1986; Jansson, 2002). Interestingly, R. salmoninarum c an occur extra cellularly or intracellularly in the granulomas or necrotic foci (Bruno, 1986; Bullock and Herman, 1988). In the kidney s , t he hematopoietic tissue of the anterior kidney s appears to be affected initial ly, followed by widespread damage to the excretory p or tion of the kidney s (Wood and Yasutake, 1956; Jansson, 2002 ). Kidney pathology can include hypercellularity of the glomeruli, occlusion of lamentous or granular deposits, and presence of eosinophilic granules in proximal tubules (Young and Chapman, 1978 ; Sami et al., 1992). Massive myocarditis , meningitis, and encephalitis were also recorded in some salmonids (Wood and Yasutake, 1956; Speare, 1997) . In the liver, histopathological changes take the form of granulomatous nodules in the connective tissue stroma between the cords of the hepatic cells ( Wood and Yasutake, 1956). 21 3.2. Host s usceptibility There are a number of reports demonstrating that salmonid species can differ in their susceptibility to BKD. Generally, Chinook salmon, coho salmon, and domestic Atlantic salmon are considered to be the most susceptible to BKD and are more likely to experience mortality due to the disease. Additionally, lake trout ( Salvelinus namaycush ) , rainbow trout [ steelhead ] , brook trout, lake whitef ish ( Coregonus clupeaformis ) , and bloater ( C. hoyi ) are only somewhat susceptible to BKD and may not necessarily experience mortality upon becoming infected (Starliper et al., 1997; Jonas et al., 2002; Hay, 2003; Nuhfer et al., 2005). In addition to diff erences between species, different strains of the same species have also been observed with differential susceptibility to BKD. For example, coho salmon of three different transferrin genotypes (AA, AC, and CC) differed in resistance to experimental infec tion with R. salmoninarum (Suzumoto et al., 1977). Furthermore , three populations of Chino ok salmon from different rivers demonstrated various mortality rates when exposed to R. salmoninarum in an experimental infection (Beacham and Evelyn, 1992). Winter et al. (1980) reported similar results in c oho salmon and steelhead . Further, Belding and Merrill (1935) reported that brook trout were more susceptible to R. salmoninarum infection than the rainbow trout when experimentally infected. In this context, M itchum and Sherman (1981) also reported that brook trout were more susceptible to natural BKD infections than rainbow trout and brown trout. 22 3. 3 . Pathogenesis and immunity 3.3.1. Process of infection and pathogenesis Renibacterium salmoninarum has the ability to induce uptake by non - phagocytic cells and can also survive ingestion, provid ing a possible means of entry into the host via the gastrointestinal tract ( Balfry et al., 1996; Evelyn, 1996; Flaño et al., 1996 ) . The gills are a less likely portal of entry for R. salmoninarum , as it has been dem onstrate d that R. salmoninarum is not internalized by healthy rainbow trout gills in vitro (McIntosh et al., 2000) . Uptake of R . salmoninarum through the eggs and reproductive fluids from parent s to offsprin g has been demonstrated in several vertical transmission studies ( Evelyn et al., 1984; Evelyn et al., 1986a, b; Bruno and Munro, 1986b). Renibacterium salmoninarum is considered to spread through blood and also through intracellular habitation and replica tion in macrophages (Gutenberger et al., 1997; Ellis, 1999). Even though R. salmoninarum is a slow - growing organism, once it is established in the circulatory system, it can reach levels of 10 9 cells/g in spleen and kidney tissues before initiation of fis h mortality (Evelyn, 1996). With most pathogens, o psonization of the pathogen by an antibody and/or complement usually limits the ability of the pathogen to cause harm; however, with R. salmoninarum , opsonization actually increases its ability to surviv e and replicat e within phagocytes (Bandin et al., 1995). T o survive and replicate, R. salmoninarum needs to acquire nutrients from the host. In the absence of iron, R. salmoninarum may produce iron reductase, which makes bound iron more available for bacte rial uptake (Grayson et al., 1995). 23 Renibacterium salmoninarum also produces large amounts of the p57 antigen (Wiens and Kaattari, 1989), both in serum and intracellularly. The quantity it produces can neutralize the vast majority of antibodies that may be evoked in response to infection ; therefore, t hese antibody - p57 complexes may remain in tissue and contribute to tissue destructive hypersensitivity and result in granulomas (Bruno, 1986; Sami et al., 1992). Moreover, t he p57 protein has immunosuppressi ve and tissue destructive properties. It can agglutinate salmon leukocytes and suppress antibody production against unrelated antigens in vitro , thereby further reducing the protective capabilities of the host (Turaga et al., 1987; Wiens and Kaattari, 199 1 ) . The p57 is a potent inhibitor of the phagocyte respiratory burse response , which typically plays a critical role in degrading bacteria (Campos - Perez et al., 1997) . Furthermore, p57 could potentially decrease the bactericidal activity of juvenile Chin ook salmon macrophages against another important fish pathogenic bacterium , Aeromonas salmonicida (Siegel and Congleton, 1997). Lastly, Senson and Stevenson (1999) suggested that p57 and its breakdown products might form a protective layer around R. salmo ninarum cells. Cell surface associated p57 and its breakdown products could effectively block highly immunogenic areas of the bacterial cell Indeed, b acterial cells stripped o f p57 induced a stronger immune response than those not stripped of p57 in an experimental challenge (Wood and Kaattari, 1996). 24 3.3.2. Effect of BKD on host immune response Grayson et al. (2002) studied the immunosuppressive effect of R. salmoninaru m in vi vo and in vi tr o . Within the in vitro assay, macrophages showed a rapid inflammatory response in which the expression of inducible cyclooxygenase, ind ucible nitric oxide synthase , interleukin - 1 (IL - , and major histocompatibility complex class II (MHC II) were enhanced (Grayson et al., 2002). Additionally, at first tumor necrosis factor - (TNF - ) expression was greatly reduced , but then later increased (Grayson et al., 2002) . In the in vivo study, intraperitoneal (i.p.) injection of R. salmonina rum DNA vaccine constructs reduced the expression of IL - 1 , Cox - 2, and MHC II , but stimulated TNF - (Grayson et al., 2002) . T he authors concluded that the p57 suppresses the host immune response and proposed that the chronic granulomatous reaction is due to the prolonged stimulation of TNF - (Grayson et al., 2002) . Furthermore, t he p57 possess es immunosuppressive abilities against the salmonid - specific antibody response (Turaga et al., 1987), tissue destructive properties (Bruno, 1986), and is also capabl e of agglutinating salmon leukocytes (Wiens and Kaattari, 1999). Aside from its opsonizing action, antibodies can interact directly with free p57, creating immune complexes that accumulate within the tissue and cause hypersensitivity reactions, which can result in granulomas and tissue damage (Bruno, 1986). Macrophage activating factor (MAF) - activated macrophages can effectively kill R. salmoninarum cells , and the proliferation and action of T - cells in activating macrophages may be the primary successful immune response against R. salmoninarum (Secombes, 1985; Hardie et al., 1996) . However, low temperatures can frequently suppress the production of MAF in immature helper T - cells, 25 thereby reducing the potential immune response against R. salmoninarum (Sieg el and Congleton, 1997). 3.3.3. Environmental factors 3.3.3.1. Effect of diet Research has suggested that the prevalence and severity of BKD might be part ial ly associated with particular environmental and dietary factors. Diets formulated of gluten as opposed to cottonseed meal have resulted in higher BKD prevalence in several hatcheries in the state of Washington (Wood, 1974 ). Additionally, Wedemeyer and Ross (1973) demonstrated that while the incidence of a BKD infection was similar in fish fed gl uten and cottonseed diets, the non - specific stress of the infection was more severe in the gluten group, as reflected by the increased ascorbate depletion, and could result in elevated mortalities. While Sakai et al. (1986) concluded that vitamins had no effect on BKD prevalence , Paterson et al. (1981) indicated that Vitamin A, zinc, and iron levels can be significantly reduced in BKD - infected fish . Furthermore, subsequent feeding trials demonstrated a lower incidence of BKD in fish fed diets high in trac e elements ( iron , copper , magnesium , c o balt , iodine, and fluorine ) or low in calcium ( Paterson et al., 1981 ). Lastly, Woodall and LaRoche (1964) suggested that iodine insufficiency could also be responsible for increased BKD incidence in juvenile Chinook salmon. L imited food availability , in addition to other factors, may also be related to increased BKD infections (Holey et al., 1998). The Chinook salmon epizootics that occurred in Lake 26 Michigan in the late 1980s were attributed to a lack of a food sour ce (i.e., alewives) and a co - infection of R. salmoninarum and the acanthocephalan Echinorhynchus salmonis (Holey et al., 1998). The lack of an abundant food source (i.e., alewives), which could not satisfy the energy demands of Chinook salmon, coupled wit h the presence of E. salmonis in the intestinal tract, likely resulted in nutritional stress that made the fish vulner able to BKD (Holey et al., 1998) . 3.3.3.2. Effect of temperature BKD has been reported to occur over a wide range of water temperatur es (Belding and Merrill, 1935; Earp et al., 1953; Fryer and Sanders, 1981; Bullock and Herman, 1988). Sanders et al. (1978) reported the occurrence of an earlier time - to - death (21 - 34 days versus 61 - 70 days) for experimentally infected juvenile coho salmon , sockeye salmon, and steelhead that were held at 15 - 20 ° C compared to 6.7 ° C . Similarly , Wood ( 1974 ) reported that mortalities due to BKD in hatcheries occurred 30 - 35 days post exposure at temperatures above 11 ° C , but took 60 - 90 days at 7.2 - 10°C. Moreover , Earp et al. (1953) detected seasonal trends occurring with BKD epizootics in Washington state hatcheries. M ost of the epizootics occurred during the autumn and winter, under conditions of declining water temperatures; however, the greatest mortalit ies w ere associated with periods of the highest water temperatures. Also, it was noted that during periods of low water temperatures, the disease produced a slow steady death rate (Earp et al., 1953) . 27 3.3.3.3. Effect of estuarine and salt - water environments Despite the fact that BKD occurs mainly in freshwater, significant infections also occur in saltwater (Banner et al., 1983). Reports have demonstrated that deaths continued in Chinook, coho, and pink salmon ( Oncorhynchus gorbuscha ) stocks after movement to saltwater - rearing ponds (Earp et al., 1953; Bell, 1961). Additionally, Frantsi et al. (1975) reported that R. salmoninarum impaired the ability of Atlantic salmon smolts to acclimate to saltwater and caused subsequent reduction in their ocean survival . Renibacterium salmoninarum can also persist in individuals after spending prolonged time periods in saltwater (Ellis et al., 1978). Furthermore, Fryer and Sanders (1981) indicated that BKD was thought to be the main cause of death among coho salmon smo lts released from the Siletz H atchery in Oregon. The se authors reported that the majority of deaths occurred between two and four months after the fish entered saltwater. Additionally, after migration into saltwater, fish continued to die due to BKD at a n accelerated rate (Fryer and Sanders, 1981). Also, BKD infection s can affect the ability of fish to acclimate to seawater and can result in death (Mesa et al., 1999). Further, Price and Schreck (2003) experimentally assessed the effect of BKD on saltwat er preference of juvenile spring Chinook s almon and concluded that there wa s a significant negative relationship between mean infection level and saltwater preference. Their results demonstrated that the higher the level of BKD - infection, the less likely Chinook salmon were to migrate into saltwater or they took a much longer time to migrate permanently into saltwater ( Price and Schreck, 2003 ). This behavior may increase the risk of avian predation for Chinook salmon migrating out to the ocean (Price and Schreck, 2003) . 28 3.4. Epizootiology 3.4. 1 . Geographical distribution Bacterial kidney disease has been reported worldwide, nearly everywhere that susceptible salmonid populations are present (Fryer and Sanders, 1981; Klontz, 1983). The disease is common ly reported in cultured salmonid species from North America, Europe, Japan, and South America (Fryer and Sanders, 1981; Bullock and Herman, 1988). B acterial kidney disease has also been observed in a wide range of wild (Pippy, 1969; Evelyn et al., 1973; E llis et al., 1978; Paterson et al., 1979; Mitchum and Sherman, 1981) and feral salmonid populations from North America (Elliott and Pascho, 1991; Sanders et al., 1992; Holey et al., 1998; Jonas et al., 2002; Faisal et al., 2012). The geographic range of B KD includes Canada, Chile, Denmark, England, France, Finland, Germany, Iceland, Italy, Japan, Norway, Poland, Scotland, Spa in, Sweden, Turkey, United States, former Yugoslavia (Bullock and Herman, 1988 ; Lorenzen et al., 1997; Jansson, 2002 ). BKD was presu mptively diagnosed and reported in Australia in the early 1970s in farmed Chinook salmon; however, further work identified the syndrome to be nocard iosis (Humphrey et al., 1987). 3.4. 2 . Host range Renibacterium salmoninarum has been detected in Chinook salmon (Holey et al., 1998), coho salmon (MacLean and Yoder, 1970), brown trout, brook trout, rainbow trout (Belding and 29 Merrill, 1935; Mitchum et al., 1979), Pacific salmon, Atlantic salmon, lake trout (Awakura, 1978), grayling ( Thymallus thymallus ) (Ket tler et al., 1986 ), lake whitefish and bloater (Jonas et al., 2002) , and whitefish ( C lupeaformis lavretus ) in Finland ( Rimaila - Pärnänen , 2002). The organism has also been detected in absence of disease in few non - salmonid species such as flathead ( Platyce phalus indicus ), Pacific hake ( Merluccius productus ), and Pacific herring ( Glupea pallasi pallasi ) (Traxler and Bell, 1988 ; Kent et al., 1998 ). Renibacterium salmoninarum antigen has also been detected in Japanese sculpin ( Cottus japonicas ) and Japanese s callops ( Patinopecten yessoensis ) (Sakai and Kobayashi, 1992). Recently, the organism has been isolated for the first time from the adult parasitic stage of Lake Ontario sea lamprey ( Petromyzon marinus ) ( Eissa et al., 2006 ). Experime ntal infections have shown that sablefish ( Anoplopoma fimbria ) are susceptible to R. salmoninarum , resulting in death in some instances, and can harbor the pathogen for up to 165 days post - in fection, demonstrating their potential to act as reservoir hosts (Bell et al., 1990) . Also , a natural outbreak of BKD in ayu ( Plecoglossus altivelis ) was documented in Japan, where they were cultured closely wi th mas u salmon ( Oncorhynchus masou ) and were found to be more susceptible (Nagai and Iida, 2002). 3.4.3. Disease transmission 3. 4.3.1. Source of infection Renibacterium salmoninarum can survive in feces, saltwater, freshwater, and pond sediments for different durations of time. Renibacterium salmoninarum excreted in the fece s 30 of clinically diseased trout can survive for u p to on e week in feces and two weeks in sterile seawater (Balfry et al., 1996). The organism can also survive in non - sterile freshwater and pond sediments for up to 21 days (Austin and Rayment, 1985). Thus, the oral - fecal route of horizontal transmission may co ntribute significantly to the increasing prevalence of BKD in salmonids. 3.4.3.2. Horizontal transmission Several studies have demonstrated the ability of R. salmoninarum to be horizontally transmitted among fish. Renibacterium salmoninarum possesses a powerful capability of inducing uptake by tissue cells including the epithelial lining of the gastrointestinal tract ( Bruno, 1986; Evelyn, 1996; Flañ o et al., 1996). Infection is thereby likely to occur whe n sufficient numbers of bacteria are present wit hin the immediate vicinity of an aquatic environment. The o ral - fecal route of infection can also occur in net pens by ingestion of contaminated feces during feeding (Balfry et al., 1996). Waterborne infection has been shown to o ccur through the gills, ey es, lesions, and wound s (Evenden et al., 1993). Renibacterium salmoninarum wa s al so transmitted by feeding fish infected or inefficiently pasteurized fish offal or fish flesh (Wood, 1974; Fryer and Sanders, 1981). Thus, uptake of R. salmoninarum through the intestinal wall is a likely pathway of infection (Jansson, 2002). Horizontal transmission can also occur between wild and stocked hatchery trout in natural systems (Mitchum and Sherman, 1981). It has also been shown that coded wire tags, which are used extensively for identification and management of anadromous salmonid populations in the Pacific Northwest, 31 enhance the horizontal transmission of R. salmoninarum ( Elliott and Pascho , 2001). Tagged fish demonstrated typical BKD lesions near the site o f implantation. Furthermore, the authors suggested that the procedures might promote transmission of the pathoge contaminated tagging needles (Elliott et al., 2001). 3.4.3.3. Vertical transmission Numerous studies have been conducted in the last several decades to study the role of vertical transmission of R. salmoninarum from mother to offspring v ia eggs. Allison (1958) was the first to report the development of BKD in offspring hatched at a facility where BKD had never been detected. Interestingly , the offspring were transferred as eggs from a hatchery where the disease had been endemic for many years (Allison, 1958). Bullock et al. (1978) demonstrated transmission of R. salmoninarum from parental broodstock to their progeny via the eggs. In some instances, the bacterium is believed to be located intra - ovum, within the perivitteline membrane of the egg, where it has been shown to be protected from surface disinfectants (Bruno and Munro, 1986 b ; Evelyn et al., 1986a, b; Evelyn, 1993). Coelomic fluid infected with high bacterial counts has also been shown to be an important source of infection for the egg (Evelyn, 1993). Moreover, intra - ovum infections can also occur prior to ovulation and directly from the ovarian tissue (Evelyn, 1993). Contrary to the strong evidence for vertical transmission of BKD via female parental broodstock, there is lit tle documentation in support of male contribution to vertical transmission. Evelyn et al. (1986b) reporte d that male coho salmon and steelhead did not play 32 a significant role in vertical transmission in experimental infections and Klontz (1983) submitted that the male does not contribute to vertical transmission either . However, Eissa et al. (2007) demonstrated similar levels of R. salmoninarum infection in milt and ovarian fluid from naturally infected brook trout and suggested that males do contribute t o vertical transmission of BKD. Moreover, at certain gamete - collecting weirs in Michigan, Faisal et al. (2012) reported a higher incidence of R. salmoninarum infection in milt than ovarian fluid from naturally infected salmon, implying that males could be contributing to the vertical transmission of R. salmoninarum , although to what extent this is occurring is not known. 3.4.3.4. Fish as possible vectors and carriers Although there is enough satisfactory data indicating that R. salmoninarum is an obliga te intracellular pat hogen of salmonid fishes , and the reservoir hosts and carrier s of infection are other infected salmonids (Wood and Yasutake, 1956; Fryer and Sanders, 1981; Klontz, 1983; Bullock and Herman, 1988), there are still other indications regar ding the possibility that non - salmonids can also act as a reservoir s or possible vector s for the bacterium . There are a f ew non - salmonid species that are able to contract the infection naturally or experimentally , and in turn , they have the potential to b ecome carriers , play ing an important role in transmission of the disease to salmonid species by cohabitation. For example, Pacific herring living in net pens with R. salmoninarum - infected coho salmon were also infected (P a clibare et al., 1988) , as well as ayu living in close proximity to masu salmon in Japan (Nagai and Iida, 2002). Also, Pacific herring (Traxler and Bell, 1988 ), sablefish (Bell et al., 1990), 33 common shiner ( Notropis cornutus ) (Hicks et al., 1986), and fathead minnow ( Pimephales promelas ) (Hicks et al., 1986) were able to contract infection by i.p. injection of R. salmoninarum . The organism was also dete cted in moribund Pacific hakes (Kent et al., 1998). In addition, Japanese sculpin and flathead were also reported as possible vectors for the disease (Sakai and Kobayashi, 1992). More recently, the de te ction of R. salmoninarum from the adult parasitic stage of the sea lamprey suggest s that it may also play a role in disease transmission ( Eissa et al., 2006 ). 3.4.3.5. Possible vectors othe r than fish A limited number of st udies have been conducted investigating t he assumption that animals other than fish can act as possible vectors for the transmission of R. salmoninarum to salmon id fish species. T he Japanese scallop has been reported as a possible vector for R. salmoninarum transmission to coho salmon pen - raised in the neighboring seawater (Sakai and Kobayashi, 1992). 3.4.3.6. Reservoirs While c linically infected, sub clinically infected, or latent carrier salmonids are the main res ervoir s of infection (Klontz, 1983; Richards et al., 1985; Bullock and Herman, 1988) , b acterial laden - feces and R. salmoninarum rich pond sediment can also act as reservoir s of infection 34 (Austin and Rayment, 1985; Balfry et al., 1996). In addition, ineffi ciently pasteurized infected salmon viscera used as feed are a confirmed reservoir of infection ( Wood , 1974). 4. Diagnosis of BKD 4. 1 . Type of sample used for pathogen detection The current ly accepted methods for detecting many bacterial fish pathogens , including Renibacterium salmoninarum , are outlined by the American Fisheries Society - Fish Health Section (AFS - FHS) Blue Book (201 2 ) and the World Organisation for Animal Health (OIE) Manual of Diagnostic Tests for Aquatic Animals (2012). While these met hods require the use of kidney and spleen samples, there are multiple tissues that can be used for R. salmoninarum detection. 4.1.1. Kidney s and spleen The kidney s and spleen contain haematopoietic tissue and are major sites for blood filtration. Furth ermore, they remove circulating bacterial antigens from the host and commonly provide indication of an infection (Bruno and Poppe, 1996). The kidney s and spleen are the recommended tissues used for detection of R. salmoninarum in salmonids by the AFS - FHS (201 2 ) and the OIE (2012). Since the first detection of BKD, there have been numerous studies demonstrating the detection capability of the kidney s and spleen ( Pippy, 1969; Bullock and Stuckey, 1975; Mitchum and Sherman, 1981; Pascho et al., 1991; Gudmund sdóttir et al., 35 1993; Chase and Pascho , 1998; Faisal et al., 2012). While the kidney s and spleen are widely accepted as suitable sample types for BKD detection, they do require that the fish host be euthanized prior to collection. 4.1.2. Reproductive flu id s It is well established that R. salmoninarum is vertically transmitted intra - ovum and in ovarian fluid ( Evelyn et al., 1986 a ). Due to this mode of transmission, R. salmoninarum is frequently detected in eggs and ovarian fluid from various salmonid sp ecies (Pascho et al., 1987; Pascho, et al., 1998; Rhodes et al., 1998 ; Eissa et al., 2007; Faisal et al, 2012 ) . Additionally, while the extent to which they contribute to vertical transmission is unknown, R. salmoninarum has also been detected in milt fro m brook trout, Chinook salmon, coho salmon, and steelhead (Eissa et al., 2007; Faisal et al., 2012). Detection of R. salmoninarum in ovarian fluid has become an important management strategy to prevent vertical transmission (Pascho et al., 1991 ; Faisal et al., 2012 ), yet ovarian fluid may not always be readily available from each fish and can only be obtained from females. 4.1.3. Blood Blood contains specific antibodies that have been produced to contest a particular pathogen (Alcorn and Pascho, 2000). As such, the presence of a bacterium can be determined from the blood; additionally, the antibody response can also be measured to determine the 36 degree of infection (Alcorn and Pascho, 2000). Renibacterium salmoninarum has been detected fr om the blood of brook trout, brown trout, coho salmon, rainbow trout , and sockeye salmon ( Bullock and Snieszko , 1969; Pascho et al. , 1987; Rhodes et al. , 1998; Alcorn and Pascho , 2000) , although its capability to be detect R. salmoninarum in a manner comparable to kidney and spleen samples has yet to be determined . 4.1.4. Mucus A s an active part of the fish immune system, t he mucus is well known to be a vital barrier fo r various pathogens (Hjelmeland et al., 1983) . I t also has a protective role in fish immunity and contains proteolytic enzymes, lymphocytes, antibodies, and lysozyme, which are important components of the fish immune response ( Ourth, 1980; Hjelmeland et al. , 1983; St. Louis - Cormier et al., 1984; Ellis, 2001 ). While R. salmoninarum has been detected in skin lesions from rainbow trout and coho salmon, it has not been detected from so lely the mucus (Hoffmann et al., 1984). However, due to its role in the immune response, it is likely that R. salmoninarum would be detected from the mucus layer. 4.1.5. Ur ine and/or feces Though the horizontal transmission of R. salmoninarum is still not fully understood, it has been suggested that R. salmoninarum can be transmitted via the feces of infected fish ( Austin and Rayment, 1985; Fryer and Lannan , 1993 ; Balfry e t al., 1996 ). Renibacterium 37 salmoninarum has been detected from within the intestine of Chinook salmon and rainbow trout (Bullock et al., 1978), and from the feces of brown trout, brook trout, coho salmon, and rainbow trout (Bu llock et al., 1980; Mitchum and Sherman , 1981). Additionally, Bruno (1986) demonstrated the presence of R. salmoninarum within the endothelial cells of the urinary tract of experimentally infected rainbow trout and Atlantic salmon, suggesting that R. salmoninarum could be passed to the external environment via the urinary tract. 4. 2 . Isolation and bacteriological identification of the agent As discussed previously, a number of culture media have been successfully used for the primary isolation of R. salmoninarum from clinically i nfected fish. Among these media , cysteine blood agar (Ordal and Earp, 1956), KDM2 (Evelyn, 1977 ), S KDM (Austin et al., 1983 ), charcoal agar medium (Daly and Stevenson, 1985) , and MKDM (Faisal et al., 2010) are used with varying degrees of success. The mo st common drawback of bacterial culture is the slow growing nature of R. salmoninarum , which can require up to 12 weeks to achieve bacterial growth. The optimal incubation temperature for the isolation of R. salmoninarum on culture media is 15°C and t he o rganism is differentiated from other Gram - positive bacteria using the morpho - chemotaxonomic features described by Sanders and Fryer (1980). 38 4. 3 . Antigen - antibody reactions Several diagnostic assays have been developed over the past 40 years to detect the presence of R. salmoninarum - specific antigens and antibodies, including the agglutination test, various fluorescent antibody tests, enzyme linked immunosorbent assays, and an immunohistochemistry test. 4.3.1. Agglutination test Although easy and ra pid to perform, the agglutination test requires that bacteria are first cultured, which conveys no advantage if compared with that of other diagnostic methods. To develop a coagglutination test to detect R. salmoninarum in kidney tissues, Kimura and Yoshi mizu (1981) used S taphylococci - specifically sensitized with antibod ies against R. salmoninarum with limited success. 4.3.2. Immunofluorescence Direct and indirect fluorescent antibody tests (FAT) have commonly been used to detect R. salmoninarum in infe cted tissues , including fixed and paraffin - embedded tissues. Bullock and Stuckey (1975) were the first to describe the indirect fluorescent technique (IFAT) to visualize R. salmoninarum cells in tissues of infected fish. They concluded that IFAT is more sensitive than Gram stain ing and can detect the bacteria in subclinical infections. Several 39 methods to quantify R. salmoninarum utilizing FAT tests have been used, including a subjective scoring of fluorescence intensity (1+ to 4+) in tissue smears (Bullo ck et al., 1980). In a later procedure, bacteria are immobilized on filter - paper grids and titers expressed as cells per unit of tissue or ovarian fluid (Elliot t and Barila, 1987). Elliot t and McKibben (1997) compared two fluorescent antibody technique s , membrane filtration FAT ( MF - FAT ) and s mear - FAT ( S - FAT) for detection of R. salmoninarum in ovarian fluid from naturally infected Chinook salmon. They reported greater sensitivity of MF - FAT compared to the S - FAT and concluded that MF - FAT was preferable fo r detection of low numbers of bacteria. Cross reactivity of other bacterial species with antisera prepared against R. salmoninarum has been reported (Bullock et al., 1980; Austin et al., 1985; Brown et al., 1995 ), thus the inclusion o f control material fr om R. salmoninarum - positive fish is necessary for comparison of cell morphology and staining properties of bacteria in test and control samples (Elliott and McKibben, 1997). Inter - laboratory comparisons reve aled that FAT reproducibility can be poor when u sed in detection of very low levels of infection (Armstrong et al., 1989). 4.3.3. Enzyme linked immunosorbent assay (ELISA) A do uble antibody sandwich ELISA, also known as quantitative ELISA (Q - ELISA), can provide an indication of th e real prevalence of BKD in a fish population due to it s ability to determine both prevalence and intensity of the infection (Pascho et al., 1998). The procedures are fairly standardized by the studies of Pascho and Mulcahy (1987) and Pascho et al. (1991). A positive thresh old can be computed fo r Q - ELISA results interpretation (Meyers et al., 1993; 40 Pascho et al., 1998). The positive - negative cutoff absorbance for the kidney homogenate was determined as 0.10 , with the following antigen level categories for positive kidney sa mples: low (0.10 to 0.19), medium (0.20 - 0.99), and high (1.00 or more ) ( Pascho et al., 1998). Hsu et al. (1991) developed an improved monoclonal antibody based ELISA assay for detection of the p57 protein of R. salmoninarum . The assay is both specific and sensitive for detection of soluble R. salmoninarum antigen at concentrations as low as 50 - 100 ng/ml. 4.3.4. Immunohistochemistry (IHC) Immunohistochemistry has the advantage of simultaneously visualizing R. salmoninarum and the tissue alteration caused by the infection (Jansson et al., 1991; Evensen et al., 1994). Immunohistochemistry has been used to detect natural and experimental infections of BKD. For example, using in situ IHC, Lorenzen et al. (1997) reported the first detection of R. salmoninaru m in rainbow trout in Denmark. Evensen et al. (1994) detected the organism in situ by using IHC in paraffin embedded tissue specimens from Atlantic salmon , using monoclonal antibodies specific for the R. salmoninarum p57 protein. However, it has been rep orted that prolonged preservation of ti ssue samples in formalin has deleterious effect s on antigen detection and retrieval in immunohistochemical assays (Evensen et al., 1994). 41 4. 4 . Polymerase chain reaction Polymerase chain reaction has been success fully used to detect very low rates of R. salmoninarum infection. For example, R. salmoninarum DNA was detected within individ ual Chinook salmon eggs with a sensitivity of 2 bacterial cells/egg (Brown et al., 1994 ). A nested PCR (nPCR) developed by Chase and Pascho (1998) amplif ies a 320 - base pair ( bp ) fragment of the gene encoding the p57 protein . The se authors also recorded no specific fragment amplification when other fish bacterial pathogens were used for templates for the nPCR (Chase and Pascho, 199 8). The sensitivity of the nested method increased one hundred - fold compared to the conventional PCR method (Pascho et al., 1998). Pascho et al. (1998) also compared the sensitivities of nPCR, ELISA, and FAT assays in the detection of R. salmoninarum in kidneys of infected Chinook salmon . They concluded that nPCR showed the highest sensitivity , followed by ELISA , and then FAT (Pascho et al., 1998) . Pascho et al. (1998) also reported that nPCR detected R. salmoninarum in 100% of the tested ovarian fluid samples , concluding that nPCR was the most accurate and sensitive method for detection of R. salmoninarum . Hong et al. (2002) designed a pair of specific primer s for nested amplification of 501 bp and 314 bp DNA fragments of the sequence coding p57 of R. salmoninarum and also recorded no specific fragment amplification when other principal fish bacterial pathogens were used as templat es . However, Miriam et al. (1997) have cautioned that PCR positive samples may contain some proportion of dead R. salmonina rum with a detectable level of DNA, implying that kidney tissues containing non - culturable R. salmoninarum (i.e., no live bacterial cells) can be falsely positive when tested with nPCR. 42 In addition to the nested PCR, Halaihel et al. ( 2009 ) established a q uantitative reverse transcriptase - PCR (RT - qPCR ) to detect R. salmoninarum in kidney tissue samples. The RT - qPCR technique is able to detect low numbers of viable bacterial m essenger RNA , impl ying a higher capacity of detecting chronically infected animals . Additionally, Chase et al. ( 2006 ) developed and ass essed a quantitative polymerase chain reaction (qPCR) assay for the detection and enumeration of R. salmoninarum , allowing scientists to determine the level of intensity of an infection . The qPCR ampli fies a 69 - b p region of the gene encodi ng the major soluble antigen o f R. salmoninarum , and consistently detected as few as 5 R. salmoninarum cells / reaction in kidney tissue (Chase et al., 2006) . 5. Differential d iagnosis External manifestations of BKD a re non - pathognomonic, but the course of the disease and the granulomatous nature of the kidney le sions may provide presumptive identificat ions. The disease can be differentiated from other kidney diseases of chronic progression including proliferative kid ney disease, manifested as lymphoid hyperplasia in response to the myxozoan parasite Tetracapsula bryosalmonae (Clifton - Hadley et al., 1984), nephrocalcinosis, which is calcium deposits in the kidney (Peddie, 2004), and pseudo - kidney disease , caused by the bacterium Carnobacterium piscicola (Ross and Toth, 1974 ). Differentiation is mainly based on observation and detection of the organism or its antigens using immunofluorescence, IHC, ELISA, or PCR. In the case of nephrocalcinosis, differentiation is main ly based on bacteriological assessment to initially determine the presence of the bacterium ; however on - farm examination 43 of lesion consistency can help to discriminate between these conditions as BKD lesions are soft , whilst those caused by nephrocalcinosi s have a gritty texture (Peddie, 2004). Renibacterium salmoninarum can be differentiated from other bacteria in the coryneform group , which includes the genera of Listeria , Erysipelothrix , Corynebacterium , Actinomyces , Celullomonas , Curtobacterium , Arthro bacter , and Brevibacterium , by cell wall composition and G + C contents of DNA ( Stuart and Welshimer, 1974; Sanders and Fryer, 1980). In particular, even though R. salmoninarum share s certain characteristics with Actinomyces pyogenes (formerly Corynebacte rium pyogenes ), they differ in a number of other characteristics. Actinomyces pyogenes is facultatively anaerobic, catalase negative, and produce s acid from carbohydrates (Holt et al., 2001). Also, t he genus Renibacterium can be separated from pathogenic Corynebacteria and Caseobacter by the presence of lysine in the cell wall and the absence of mycolic acids (Crombach, 1978) . Caseobacter is further differentiated by a mol % G + C of 60 - 67 and Celullomonas contains the diamino acid ornithine in its cell wall peptidoglycan and has a mol % G + C ranging from 65 - 72 (Crombach, 1978). Interestingly , some of the cory neform groups of bacteria have overlapping characteristics and phylogenetic homology. Among this group of bacteria , a peptidoglycan cell wall co ntaining lysine occurs primarily in the Arthrobacter and Brevibacterium genera (Holt et al., 2001 ). DNA homology studies also show a close relationship between several species in these two genera. However, these bacteria have usually been isolated from t he environment, are chemo - organotrophic, show a progression of morphological changes during the growth cycle , and have a mol % G + C above 60 (Holt et al., 2001). Interestingly, all of these characteristics are distinctly different from that of Renibacter ium . 44 6 . Discrepancies a mong d iagnostic t ests Frequently, when multiple diagnostic assays are used to determine a R. salmoninarum infection in the same sample, discrepancies among the results can occur , thus making diagnosis difficult (Cipriano et al., 19 85; Pascho et al., 1987; Sakai et al., 198 6 ; Griffiths et al., 1991; Gudmundsdóttir et al., 1993; Meyers et al, 1993; White et al., 1995; Faisal and Eissa, 2009; Nance et al., 201 0; Elliott et al., 2013 ). However, as each of the methods targets different components of the bacterium , the inconsistencies are not entirely unexpected . M olecular assays like PCR are designed to detect the presence of DNA, which does not distinguish if the pathogen is alive or dead (Pascho et al., 1998). Alternatively, propagat ing the bacterium in vitro require s the presence of viable bacteria at high numbers to ensure isolation (Miriam et al., 1997). Also, commonly used serological technique s , such as ELISA, detect antigens secreted by the pathogen (Pascho et al., 1998). Faisa l and Eissa (2009) suggested that the disagreement in results among assays may reflect different phases of a R. salmoninarum infection at the time of sampling. In their study, the se authors documented six patterns of testing results , with each of the patt erns representing a p otential stage along the course of a natural R. salmoninarum infection (initial, established, well - established, recovery, advanced recovery, and no exposure/eliminated) . The authors also proposed that by studying the patterns of infec tion, the course of BKD infection s in a particular population could be determined . Furthermore, Nance et al. (2010) also suggested using inconsistent results from diagnostic tests to infer upon the state of infection of individual fish. 45 7 . Control Method s 7. 1 . Chemotherapy Since the early 1950s , a relatively large number of chemotherapeutics have been intensively tested in vivo and in vitro for efficacy in treating BKD. Rucker et al. (1951) was first to use antimicrobial agents against clinical BKD an d their results showed a definite decrease in mortalities when sulfadiazine was incorporated into fish diets. Although treatment d oes not completely cure clinically sick fish, sulfamerazine can reduce BKD mortalities alone and when combined with sulfaguan idine and sulfadiazine (Allison, 1958). Wolf and Dunbar (1959) tested 34 therapeutic agents , including erythromycin, thiocyanate, and sulfamerazine , on 16 strains of R. salmoninarum using the disk method for drug sensitivity screening , followed by in vivo feeding trials with experimentally infected fish. They concluded that erythromycin fed at the rate of at least 100 mg / kg of fish for 21 days gave the best results (Wolf and Dunbar, 1959) . Generally, due to the occurrence of the bacterium intracellularly as well as extracellularly, these treatmen t s only suppressed the systemic spread of the organism and induced partial relief (Amos, 1977). Intramuscular (i.m.) and i.p. administration of sulfonamide drugs significantly reduced prespawning mortality among Chinook salmon broodstocks being h e ld prior to spawning (Amend and Fryer, 1968). However, sulfonamides administered by i.m. or i.p. routes often produced sterile abscesses at the injection site in adults , and can induce mortalities and teratogenicity with in their progeny (Amos, 1977). 46 In an attempt to reduce or prevent vertical transmission of BKD, salmo n eggs can be water hardened for one hour in two ppm erythromycin (Amos, 1977). However, in this process, erythromycin was rapidly eliminated from the eg gs and dropped below detectable level s in the eggs within 24 hours after water hardening (Evelyn et al., 1986a , b ). Interestingly, erythromycin remains in the eggs of injected females for up to 60 days before spawning (Evelyn et al., 1986 a, b ; Moffitt, 19 91). It is believed that erythromycin residues in side the eggs assist in preventing vertical transmission of R. salmoninarum from parents to their offspring (Lee and Evelyn, 1994). Detectable amounts of erythromycin often remain in the perfused tissues o f both juvenile and adult salmon long after they are no longer de tected in the plasma and muscle, which possibly contributes to the efficacy of erythromycin against the slow growing R. salmoninarum (Moffitt, 1991; Haukenes and Moffitt, 1999) . Feeding eryt hromycin can also efficiently reduce mortalities of infected hatchery raised salmonids , with a dose of 200 mg/kg body weight for 21 days being the most effective ( Wolf and Dunbar, 1959; Austin, 1985; Moffitt and Bjornn, 1989; Moffitt, 1992). Monthly subcu taneous injections of adult female Pacific salmon with 11 mg/kg erythromycin can also result in reduced pre - spawning mortality due to BKD (Klontz, 1983). Erythromycin is only available as an Investigational New Animal Drug through the United States Food and Drug Administration (Moffitt, 1992 ). Austin (1985) e valuated the efficacy of more than 70 antimicrobial compounds both in vivo and in vitro and found that penicillin G, kitasamycin, erythromycin, spiramycin, and clindamycin were useful for reducing ear ly clinical BKD cases , while lincomycin, rifampicin, and cephradine were effective prophylactically , but had limited use therapeutically. Hsu et al. (1994) tested the efficacy of enrofloxacin in treating BKD in vitro and in vivo and they concluded 47 that hi gh bioavailability, low mini mal inhibition concentrations , and large volume distribution of the antibiotic make it a potential candidate for use as an effective therapeutic against BKD. 7. 2 . Adult segregation Broodstock segregation is a more practical method for reducing the prevalence and levels of R. salmoninarum in hatchery - reared salmon (Pascho et al., 1991) and for increasing survival during their downriver migration and entry into seawater (Pascho et al., 1993; Elliott et al., 1995). This procedu re aims to interrupt the vertical transmission of R. salmoninarum by destroying eggs from parental brood fish that exhibit clinical signs of BKD , or that test positive with a high titer against R. salmoninarum antigens. Along with stricter hygienic measur es, Faisal et al. (2012) documented a drastic decline in BKD prevalence in Michigan hatcheries operated by the Michigan Department of Natural Resources after the implementation of broodstock screening and culling. The method is used successfully in a numb er of U.S. states and Canadian provinces, such as Washington, Idaho, Wisconsin, and Ontario. 7. 3 . Eradication Due to the complicated nature of BKD and its threats to fisheries, Hoskins et al. (1976) recommended complete destruction of infected stocks an d disinfection of the holding facilities to achieve complete eradication of the disease. However, due to the widespread occurrence of 48 R. salmoninarum , t his procedure is cons idered by fisheries managers to be impractical (Sanders and Fryer, 1980). Eradica tion can still be of value in single fish farms or hatcheries that receive their water supply from a specific patho gen - free source (European Commission, 1999). Eradi c ation procedures should be followed by standard cleaning and disinfection procedures. Al though some trials have been attempted to eradicate BKD from fish farmed in open waters (e.g., sea and lake cages) or from farms and hatcheries with water supplies from rivers, results were very discouraging (European Commission, 1999) . After eradication procedures have been applied in the fish farm s and hatcheries, restocking should only utilize certified BKD - free stocks. Restocking should be followed by two inspections and laboratory examinations per year for a total period of two years before the facil - ). 7. 4 . Prophylaxis 7.4. 1 . Reducing the risk of BKD introduction Special attention should be paid to prevent the introduction of infected fish or their gametes (Evelyn et al., 1984; Yoshimiz u, 1996). This can only be achieved through prior examination and quarantine. Speci fic efforts should be made to restrict the movement of vehicle s associated with the facility , restrict visitors to the facility, and to utilize separate nets, buckets, and brushes for each raceway and building associated with the facility. Repopulation 49 must be accompanied with certification by a competent authority declaring that the fish or eggs are specific pathogen free. 7.4.2. Vaccination In the last three decades, vaccination against BKD has achieved different levels of success. Paterson et al. (1981) reported that an inactivated suspension of R. salmoninarum mixed 1:1 with adjuvant administered by i.p. injection reduced the level of infection of R. salmon inarum in yearling salmon, but did not completely elim in ate the infection. McCarthy et al. (1984) examined the use of a pH - lysed bacterin in rainbow trout, delivered via immersion and i .p. injection, and while it provided some protective immunity, the aut hors acknowledged that more detailed diagnostic testing techniques needed to be used to more accurately ensure that vaccinated fish were free from infection. Sakai et al. (1993) found that although vaccination evoked specific antibodies, these antibodies did not endow fish with any significant protection. Wood and Kaattari (1996) evaluated a formalin - killed R. salmoninarum vaccine that contained the p57 protein (p57+), and a vaccine which lacked most of the p57 protein (p57 - ), in Chinook salmon. The se au thors were able to determine that the p57 - vaccine produced antibody titers five times higher than the p57+ vaccine, demonstrating an enhanced antibody response. Piganelli et al. (1999) demonstrated that oral administration of R. salmoninarum expressing low levels of cell associated p57, resulted in an extension of the mean time to death after challenge and they concluded that the protection was not due to humoral antibodies . This conclusion supported earlier histopathological indications of an involvem ent of the cell 50 mediated immune response in recovery, due to intracellular survival and the composition of inflammatory cells in connection with signs of regression (Munro and Bruno, 1988). Rhodes et al. (2004) presented DNA adjuvants and whole bacterial cell vaccines against R. salmoninarum that were tested in Chinook salmon fingerlings. These authors concluded that whole cell vaccines of either a nonpathogenic Arthrobacter spp. or an attenuated R. salmoninarum strain produced limited protection against acute intraperitoneal challenge with virulent R. salmoninarum . They also concluded that the addition of either synthetic oligodeoxynucleotides or purified R. salmoninarum genomic DNA as adjuvants did not increase protection; however, a combination of both whole cell vaccines significantly increased survival among fish naturally infected with R. salmoninarum . Also, the surviving fish treated with the combination vaccine exhibited reduced levels of bacterial antigens in the kidney. T here have also been sev eral other studies examining the use of the nonpathogenic Arthrobacter spp. vaccine (known commercially as Renogen® ) , all with varying and inconsistent results (Burnley et al., 2010; Alcorn et al ., 2005; Salonius et al., 2005 ). 51 APPENDIX 52 Table 1.1. Summary of the morphological and biochemical characteristics of Renibacterium salmoninarum (Eissa, 2005; Austin and Austin, 2007). *API - ZYM is a bacterial enzyme based assay used for the specific identification of different bacteria. T est Criteria Notes Agar hydrolysis - Amylase - Arginine hydrolysis - Bile solubility - Butyrate esterase - Caprylate esterase - Carbohydrate utilization - Casein hydrolysis + Catalase + Cytochrome oxidase - DNase + ( - ) By API - ZYM* Esc ulin hydrolysis - Gelatin liquefaction + ( - ) By API - ZYM* Gram stain + Hemolytic activity - hemolytic Complete clearance zone around bacteria Indole test - Methyl red - Nitrate reduction - Oxidase - PAS (Periodic Acid Schiff) stain + Phosphatase - Trypsin + Tween - 20, 40, and 60 hydrolysis + Tween - 80 hydrolysis - Urease - Ze ihl - Nielsen (Acid Fast) stain - Non - acid fast 53 REFERENCES 54 REFERENCES Alcorn, S.W., and Pascho, R.J. 2000. Single - dilution enzyme - linked immunosorbent assay for quantification of antigen - specific salmonid antibody. J. Vet. Diagn . Invest. 12: 245 - 252. Alcorn, S., Murray, A.L., Pascho, R.J., and Varney, J. 2005. A cohabitation challenge to compare the efficacies of vaccines for bacterial kidney disease (BKD) in Chinook salmon Oncorhynchus tshawytscha . Dis. Aquat. Org. 63(2): 151 - 1 60. Alexander, S.M., Grayson, T.H., Chambers, E.M., Coopers, L.F., Barker, G.A., and Gilpin, M.L. 2001. Variation in the spacer regions separating tRNA genes in Renibacterium salmoninarum distinguishes recent clinical isolates from the same location. J. C lin. Microbiol. 39: 119 - 128. Allison, L.N. 1958. Multiple sulfa therapy of kidney disease among brook trout. Prog. Fish - Cult. 20: 66 - 68. Amend, D.F., and Fryer, J.L. 1968. The administration of sulfonamide drugs to adult salmon. Prog. Fish - Cult. 30: 168 - 172. American Fisheries Society - Fish Health Section (AFS - FHS). 2010. FHS blue book: suggested procedures for the detection and identification of certain finfish and shellfish pathogens, 2010 edition. AFS - FHS, Bethesda, Maryland. Amos, K.H. 1977. The con trol of bacterial kidney disease in spring Chinook salmon. M.S. thesis, University of Idaho, Moscow. 20pp. Armstrong, R.D., Evelyn, T.P.T., Martin, S.W., Dorwand, W., and Ferguson, H.W. 1989. Erythromycin levels in eggs and alevins derived from spawning b roodstock Chinook salmon ( Oncorhynchus tshawytscha ) injected with the drug. Dis. Aquat. Org. 6: 33 - 36. Austin, B., and Austin, D.A. 2007. Bacterial Fish Pathogens, Diseases in Farmed and Wild Fish. 4 th Edition. Springer - Praxis, Chichester, United Kingdom. Austin, B., and Rayment, J.N. 1985. Epizootiology of Renibacterium salmoninarum , the causal agent of bacterial kidney disease in salmonid fish. J. Fish Dis. 8: 505 - 509. Austin, B., and Rodgers, C.J. 1980. Diversity among strains causing bacterial kidney disease in salmonid fish. Curr. Microbiol. 3: 231 - 235. Austin, B., Embley, T.M., and Goodfellow, M. 1983. Selective isolation of Renibacterium salmoninarum . FEMS Microbiol Lett. 17: 111 - 114. 55 Austin, B., Bucke, D., Feist, S., and Rayment, J.N. 1985. A fa lse positive reaction in the fluorescent antibody test for Renibacterium salmoninarum Eur. Assoc. Fish Pathol. 5: 89. Austin, B. 1985. Evaluation of antimicrobial compounds for the control of Bacterial Kidney Disease in rainbow trout, Salmo gairdneri Richardson. J. Fish Dis. 8: 209 - 220. Awakura, T. 1978. Bacterial kidney disease in mature salmonids. Fish Pathol. 13: 53 - 54. Balfry, S.K., Albright, L.J., and Evelyn, T.P.T. 1996. Horizontal transfer of Renibacterium salmo ninarum among farmed salmonids via the fecal - oral route. Dis. Aquat. Org. 25: 63 - 69. Bandin, I., Santos, Y., Bruno, D.W., Raynard, R.S., Toranzo, A.E., and Barja, J.L. 1991. Lack of biological activities in the extracellular products of Renibacterium salm oninarum . Can. J. Fish. Aquat. Sci. 48: 421 - 425. Bandin, I., Rivas, C., Santos, Y., Secombes, C.J., Barja, J.L., and Ellis, A.E. 1995. Effects of serum factors on the survival of Renibacterium salmoninarum within rainbow trout macrophages. Dis. Aquat. Org . 23: 221 - 227. Banner, C.R., Rohovec, J.S., and Fryer, J.L. 1983. Renibacterium salmoninarum as a cause of mortality among Chinook salmon in salt water. J. World. Maricult. Soci. 14: 236 - 239. Banner, C.R., Rohovec, J.S., and Fryer, J.L. 1991. A new value for mol percent guanine+cytosine of DNA for the salmonid fish pathogen Renibacterium salmoninarum . FEMS Microbiol. Lett. 79: 57 - 59. Beacham, T.D., and Evelyn, T.P.T. 1992. Genetic variation in disease resistance and growth of Chinook, coho, and chum salm on with respect to vibriosis, furunculosis, and bacterial kidney disease. Trans. Am. Fish. Soc. 12(1): 456 - 485. Belding, D.L., and Merrill, B. 1935. A preliminary report upon a hatchery disease of the salmonidae. Trans. Am. Fish. Soc. 65(1): 76 - 84. Bell, G.R. 1961. Two epidemics of apparent kidney disease in cultured pink salmon ( Oncorhynchus gorbuscha ). J. Fish. Res. Board Can. 18: 559 - 562. Bell, G.R., Hoffmann, R.W., and Brown, L.L. 1990. Pathology of experimental infections of the sablefish Anoplopoma fimbria (Pallas), with Renibacterium salmoninarum , the agent of bacterial kidney disease in salmonids. J. Fish Dis. 13: 355 - 367. 5 6 Benediktsdóttir, E., Helgason, S., and Gudmundsdóttir, S. 1991. Incubation time for the cultivation of Renibacterium salmon inarum from Atlantic salmon, Salmo salar L., broodfish. J. Fish Dis. 14(1): 97 - 102. Beyerle, J., and Hnath, J.G., 2002. History of fish health inspections: state of Michigan 1970 1999. In: Michigan Department of Natural Resources Technical Report 2002 2, Lansing, Michigan. Brown, L.L., Iwama, G.K., Evelyn, T.P.T., Nelson, W.S., and Levine, R.P. 1994. Use of PCR to detect DNA from Renibacterium salmoninarum within individual salmon eggs. Dis. Aquat. Org. 18: 165 - 171. Brown, L.L., Evelyn, T.P.T., Iwama, G. K., Nelson, W.S., and Levine, R.P. 1995. Bacterial species other than Renibacterium salmoninarum cross react with antisera against R. salmoninarum , but are negative for the p57 gene of R. salmoninarum as detected by the polymerase chain reaction (PCR). Dis . Aquat. Org. 21: 227 - 231. Bruno, D.W. 1986. Histopathology of bacterial kidney disease in laboratory infected rainbow trout, Salmo gairdneri Richardson, and Atlantic salmon, Salmo salar L., with reference to naturally infected fish. J. Fish Dis. 9: 523 - 5 37. Bruno, D.W., and Munro, A.L.S. 1986a. Uniformity in the biochemical properties of Renibacterium salmoninarum isolates obtained from several sources. FEMS Microbiol. Lett. 33: 247 - 250. Bruno, D.W., and Munro, A.L.S. 1986b. Observation on Renibacterium salmoninarum and the salmonid egg. Dis. Aquat. Org. 1: 83 - 87. Bruno, D.W., and Poppe, T.T. 1996. A Colour Atlas of Salmonid Diseases. Academic Press Ltd, London, 149 pp. Bullock, G.L., and Herman, R.L. 1988. Bacterial kidney disease of salmonid fishes c aused by Renibacterium salmoninarum . Fish Disease Leaflet 78, U.S. Fish and Wildlife Service, Washington, D.C. Bullock, G.L., and Snieszko S.F. 1969. Bacteria in blood and kidney of apparently healthy hatchery trout. Trans. Am. Fish. Soc. 98: 268 - 271. Bu llock, G.L., and Stuckey, H.M. 1975. Fluorescent antibody identification and detection of the Corynebacterium causing the kidney disease of salmonids. J. Fish. Res. Board Can. 32(11): 2224 - 2227. 57 Bullock, G.L., Stuckey, H.M., and Chen, P.K. 1974. Coryneba cterial kidney disease of salmonids: growth and serological studies on the causative bacterium. App. Microbiol. 28(5): 811 - 814. Bullock, G.L., Stuckey, H.M., and Mulcahy, D. 1978. Corynebacterial kidney disease: egg transmission following iodophore disinf ection. Fish Health News 7: 51 - 52. Bullock, G.L., Griffin, B.R., and Stuckey, H.M. 1980. Detection of Corynebacterium salmonius by direct fluorescent antibody tests. Can. J. Fish. Aquat. Sci. 37: 719 - 721. Burnley, T.A., Stryhn, H., Burnley, H.J., and Ham mel, K.L. 2010. Randomized clinical field trial of a bacterial kidney disease vaccine in Atlantic salmon, Salmo salar L. J. Fish Dis. 33(7): 545 - 557. Campos - Perez, J.J., Ellis, A.E., and Secombes, C.J. 1997. Investigation of factors influencing the abilit y of Renibacterium salmoninarum to stimulate rainbow trout macrophage respiratory burst activity. Fish and Shellfish Immunol. 7(8): 555 - 566. Chase, D.M., and Pascho, R.J. 1998. Development of a nested polymerase chain reaction for amplification of a seque nce of the p57 gene of Renibacterium salmoninarum that provides a highly sensitive method for detection of the bacterium in salmonid kidney. Dis. Aquat. Org. 34: 223 - 229. Chase, D.M., Elliott, D.G., and Pascho, R.J. 2006. Detection and quantification of R enibacterium salmoninarum DNA in salmonid tissues by real - time quantitative polymerase chain reaction analysis. J. Vet. Diagn. Invest. 18(4): 375 - 380. Cipriano, R.C., Starliper, C.E., and Schachte, J.H. 1985. Comparative sensitivities of diagnostic proced ures used to detect bacterial kidney disease in salmonid fishes. J. Wildl. Dis. 21(2): 144 - 148. Clifton - Hadley, R.S., Bucke, D., and Richards, R.H. 1984. Proliferative kidney disease of salmonid fish: a review. J. Fish Dis. 7: 363 - 377. Crombach, W.H.J. 1 978. Caseobacter polymorphus gen. nov., sp. nov., a coryneform bacterium from cheese. Int. J. Syst. Bacteriol. 28: 354 - 366. Daly, J.G., and Stevenson, R.M. 1985. Charcoal agar, a new growth medium for the fish disease bacterium Renibacterium salmoninarum . Appl. Environment. Microbiol. 50: 868 - 871. Daly, J.G., and Stevenson, R.M. 1990. Characterization of the Renibacterium salmoninarum haemagglutinin. J. Gen. Microbiol. 136: 949 - 953. DeKinkelin, P. 1974. Corynébactériose des salmonidés: première observati on en France. Bull. Franc. Pisc. 258: 3 - 8. 58 Earp, B.J., Ellis, C.H., and Ordal, E.J. 1953. Kidney disease in young salmon. Special Report Ser. No. 1, Department of Fisheries, Washington State, 73 pp. Eissa, A.E., Elsayed, E.E., McDonald, R., and Faisal, M . 2006. First record of Renibacterium salmoninarum in the sea lamprey ( Petromyzon marinus ). J. Wildl. Dis. 42(3): 556 - 560. Eissa, A.E., Elsayed, E.E., and Faisal, M. 2007. Prevalence and shedding of Renibacterium salmoninarum in brook trout ( Salvelinus fo ntinalis ) in Michigan. Nature and Science 5(1): 9 - 17. Elliott, D.G., and Barila, T.Y. 1987. Membrane filtration - fluorescent antibody staining procedures for detecting and quantifying Renibacterium salmoninarum in coelomic fluid of Chinook salmon ( Oncorhyn chus tshawytscha ). Can. J. Fish. Aquat. Sci. 44: 206 - 210. Elliott, D.G., and McKibben, C.L. 1997. Comparison of two fluorescent antibody techniques (FATs) for detection and quantification of Renibacterium salmoninarum in coelomic fluid of spawning Chinoo k salmon Oncorhynchus tshawytscha . Dis. Aquat. Org. 30: 37 - 43. Elliott, D.G., and Pascho, R.J. 1991. Juvenile fish transportation: impact of bacterial kidney disease on survival of spring/summer Chinook salmon stocks. Annual Report, 1989. Prepared by the U.S. Fish and Wildlife Service, Seattle, Wash., for the U.S. Army Corps of Engineers, Walla Walla, Washington. Elliott, D.G., Pascho, R.J., and Palmisano, A.N. 1995. Brood stock segregation for the control of bacterial kidney disease can affect mortality of progeny Chinook salmon ( Oncorhynchus tshawytscha ) in seawater. Aquacult. 132: 133 - 144. Elliott, D.G., and Pascho, R.J. 2001. Evidence that coded - wire - tagging procedures can enhance transmission of Renibacterium salmoninarum in Chinook salmon. J. Aquat. Anim. Health 13(3): 181 - 193. validation testing of selected immunological and molecular Renibacterium salmoninarum diagnostic assays by comparison with quanti tative bacteriological culture. J. Fish Dis. 36: 779 - 809. Ellis, R.W., Novotny, A.J., and Harrell, L.W. 1978. Case report of kidney disease in wild Chinook salmon ( Oncorhynchus tshawytscha ) in the sea. J. Wildl. Dis. 14: 121 - 123. Ellis, A.E. 1999. Immuni ty to bacteria in fish. Fish and Shellfish Immunol. 9: 291 - 308. Embley, T.M., Goodfellow, M., and Austin, B. 1982. A semi - defined growth medium for Renibacterium salmoninarum . FEMS Microbiol. Lett. 14: 299 - 301. 59 European Commission. 1999. Bacterial kidney disease. Report of the Scientific Committee on Animal Health and Animal Welfare. European Commission, Health and Consumer Protection, Brussels, pp. 1 - 44. Evelyn, T.P.T. 1977. An improved growth medium for the kidney disease bacterium and some notes on us ing the medium. Bull. Off. Int. Epiz. 87: 511 - 513. Evelyn, T.P.T., Hoskins, G.E., and Bell, G.R. 1973. First record of bacterial kidney disease in apparently wild salmonids in British Columbia. J. Fish. Res. Board. Can. 30: 1578 - 1580. Evelyn, T.P.T., Ket cheson, J.E., and Prosperi - Porta, L. 1981. The clinical significance of immunofluorescence - based diagnoses of bacterial kidney disease carriers. Fish Pathol. 15: 293 - 300. Evelyn, T.P.T., Ketcheson, J.E., and Prosperi - Porta, L. 1984. Further evidence for t he presence of Renibacterium salmoninarum in salmonid eggs and for the failure of povidone - iodine to reduce the intra - ovum infection rate in the water. J. Fish Dis. 7: 173 - 182. Evelyn, T.P.T., Prosperi - Porta, L., and Ketcheson, J.E. 1986a. Experimental in tra - ovum infection of salmonid eggs with Renibacterium salmoninarum and vertical transmission of the pathogen with such eggs despite their treatment with erythromycin. Dis. Aquat. Org. 1: 197 - 202. Evelyn, T.P.T., Ketcheson, J.E., and Prosperi - Porta, L. 19 86b. Use of erythromycin as a means of preventing vertical transmission of Renibacterium salmoninarum . Dis. Aquat. Org. 2: 7 - 11. Evelyn, T.P.T. 1986. A historical note on bacterial kidney disease (BKD) in Canada. Fish Health Section, American Fisheries So ciety Newsletter 14(4): 6. Evelyn, T.P.T., Bell, G.R., Prosperi - Porta, L., and Ketcheson, J.E. 1989. A simple technique for accelerating the growth of the kidney disease bacterium Renibacterium salmoninarum on a commonly used culture medium (KDM2). Dis. A quat. Org. 7: 231 - 234. Evelyn, T.P.T., Prosperi - Porta, L., and Ketcheson, J.E. 1990. Two new techniques for obtaining consistent results when growing Renibacterium salmoninarum on KDM2 culture medium. Dis. Aquat. Org. 9: 209 - 212. Evelyn, T.P.T. 1993. Bac terial kidney disease BKD. In: Inglis, V., Roberts R.J., and Bromage, N.R. (Eds.). Bacterial Diseases of Fish. Blackwell Scientific Publications, Oxford, United Kingdom, pp. 177 - 195. Evelyn, T.P.T. 1996. Infection and disease. In: Iwama, G., and Nakanis hi, T. (Eds.), The fish immune system. Academic Press, San Diego, United States. pp. 339 - 366. 60 Evenden, A.J., Grayson, T.H., Gilpin, M.L., and Munn, C.B. 1993. Renibacterium salmoninarum and bacterial kidney disease the unfinished jigsaw. Ann. Rev. Fish Dis. 3: 87 - 104. Evensen, O., Dale, O.B., and Nilsen, A. 1994. Immunohistochemical identification of Renibacterium salmoninarum by monoclonal antibodies in paraffin - embedded tissues of Atlantic salmon ( Salmo salar L.), using paired immunoenzyme and paired immunofluorescence techniques. J. Vet. Diagn. Invest. 6: 48 - 55. Faisal, M., and Eissa, A.E. 2009. Diagnostic testing patterns of Renibacterium salmoninarum in spawning salmonid stocks in Michigan. J. Wildl. Dis. 45(2): 447 - 456. Faisal, M., Eissa, A.E., and Starliper, C.E. 2010. Recovery of Renibacterium salmoninarum from naturally infected salmonine stocks in Michigan using a modified culture protocol. J. Adv. Res. 1(1): 95 - 102. Faisal, M., Schulz, C., Eissa, A., Brenden, T., Winters, A., Whelan, G., Wo lgamood, M., Eisch, E., and VanAmberg, J. 2012. Epidemiological investigation of Renibacterium salmoninarum in three Oncorhynchus spp. in Michigan from 2001 to 2010. Prev. Vet. Med. 107(3 - 4): 260 - 274. Feltham, R.K.A., Power, A.K., Pell, P.A., and Sneath, P.H.A. 1978. A simple method for storage of bacteria at - 76 degree C. J. App. Bact. 44: 313 - 316. Ferguson, H.W. 1989. Systemic pathology of fish. Iowa State University Press. Ames, Iowa, United States. 263 pp. Fijan, N. 1977. Corynebacteriosis (Dee disea se, kidney disease) of salmonids in Yugoslavia Bull. Off. Int. Epiz. 87: 509. Flaño, K., Kaattari, S.L., Razquin, B., and Villena, A.J. 1996. Histopathology of the thymus of coho salmon Oncorhynchus kisutch experimentally infected with Renibacterium salmo ninarum . Dis. Aquat. Org. 26: 11 - 18. Frantsi, C.J., Flewelling, T.C., and Tidswell, K.G. 1975. Investigations on corynbacterial kidney diseases and Diplostomulum sp. (eye - fluke) at Margaree hatchery, 1972 - 1973. Technical Report Serial No. Mar/T - 75 - 9. Rese arch Development Branch, Fish and Marine Services, Department of Environment. Halifax, Nova Scotia, Canada. 30 pp. Fredriksen, A. Endresen, C., and Wergeland, H.I. 1997. Immunosuppressive effect of a low molecular weight surface protein from Renibacterium salmoninarum on lymphocytes from Atlantic salmon ( Salmo salar L.). Fish and Shellfish Immunol. 7: 273 - 282. Fryer, J.L., and Lannan, C.N. 1993. The history and current status of Renibacterium salmoninarum isolates by antigenic analysis. Fish Pathol. 20: 1 49 - 159. 61 Fryer, J.L., and Sanders, J.E. 1981. Bacterial kidney disease of salmonid fish. Ann. Rev. Microbiol. 35: 273 - 298. Getchell, R.G., Rohovec, J.S., and Fryer, J.L. 1985. Comparison of Renibacterium salmoninarum isolates by antigenic analysis. Fish P athol. 20: 149 - 159. Ghittino, P., Andruetto, S., and Vigliani, E. 1977. Recent findings in trout kidney pathology in Italy. Bull. Off. Int. Epiz. 87: 491 - 493. Goodfellow, M., Embley, T.M., and Austin, B. 1985. Numerical taxonomy and emended description o f Renibacterium salmoninarum . J. Gen. Microbiol. 131: 2739 - 2752. Grayson, T.H., Coopers, L.F., Wrathmell, A.B., Roper, J., Evenden, A.J., and Gilpin, M.L. 2002. Host responses to Renibacterium salmoninarum and specific components of the pathogen reveal th e mechanisms of immune suppression and activation. J. Immunol. 106: 273 - 283. Grayson, T.H., Bruno, D.W., Evenden, A.J., Gilpin, M.L., and Munn, C.B. 1995. Iron acquisition by Renibacterium salmoninarum ; contribution by iron reductase. Dis. Aquat. Org. 22: 157 - 162. Griffiths, S.G., Olivier, G., Fildes, J., and Lynch, W.H. 1991. Comparison of western blot, direct fluorescent antibody and drop - plate culture methods for the detection of Renibacterium salmoninarum in Atlantic salmon ( Salmo salar L.). Aquacult. 97(2): 117 - 129. Gudmundsdóttir, S., Benediktsdóttir, E., and Helgason, S. 1993. Detection of Renibacterium salmoninarum ELISA and isolation on selective medium. J. Fish Dis. 16(3): 185 - 195. Gutenberger, S.K., Giovannoni, S.J., Field, K.G., Fryer, J.L., and Rohovec, J.S. 1991. A phylogenetic comparison of the 16S rRNA sequence of the fish pathogen, Renibacterium salmoninarum , to Gram - positive bacteria. FEMS Microbiol. Lett. 77: 151 - 156. Gutenberger, S.K., Duimstra, J.R., Rohovec, J.S., and Fryer, J.L. 1997. Intracellular survival of Renibacterium salmoninarum in trout mononuclear phagocytes. Dis. Aquat. Org. 28: 93 - 106. Halaihel, N., Vendrell, D., Ruiz - Zarzuela, I., de Blas, I., Al onso, J.L., Gironés, O., Pérez, T., and Muzquiz, J.L. 2009. A new real time PCR - based assay for diagnosing Renibacterium salmoninarum in rainbow trout ( Oncorhynchus mykiss ) and comparison with other techniques. J. Microbiol. Meth. 76(1): 75 - 80. Hamel, O. 2001. The dynamics and effects of bacterial kidney disease in Snake River spring Chinook salmon ( Oncorhynchus tshawytscha ). Ph.D. Dissertation. University of Washington. Washington, D.C. USA. 62 Hardie, L.J., Ellis, A.E., and Secombes, C.J. 1996. In vitro ac tivation of rainbow trout macrophages stimulates inhibition of Renibacterium salmoninarum growth concomitant with augmented generation of respiratory burst products. Dis. Aquat. Org. 27: 175 - 183. Haukenes, A.H., and Moffitt, G.M. 1999. Concentrations of e rythromycin in maturing salmon following intraperitoneal injection of two drug formulations. J. Aquat. Anim. Health 11: 61 - 67. Hay, R. 2003. Little Manistee River weir steelhead egg - take report, 1967 - 92. Michigan Department of Natural Resources Fisheries Division. Technical Report 96 - 1. 17 pp. Helgason, S. 1985. Diseases of salmonids on fish farms in Iceland. In: Symposium Handbook, Akvakultur - Miljøproblem Nordforsk, Miljøvårdsserien Publikation 2: 213 - 217. Hicks, B.D., Daly, J.G., and Ostland, V.E. 1986 . Experimental infection of minnows with bacterial kidney disease bacterium Renibacterium salmoninarum . Abstract, Third Annual Meeting, Aquaculture Association of Canada, July 1986, Guelph, Ontario. Hjelmeland, K., Christie, M., and Raa, J. 1983. Skin muc us protease from rainbow trout, Salmo gairdneri Richardson, and its biological significance. J. Fish Biol. 23: 13 - 22. Hoffman, R., Popp, W., and van de Graaff, S. 1984. Atypical BKD predominantly causing ocular and skin lesions. Bull. Eur. Assoc. Fish Pat hol. 4: 7 - 9. Holey, M.E., Elliot, R.F., Marcquenski, S.V., Hnath, J.G., and Smith, K.D. 1998. Chinook salmon epizootics in Lake Michigan: possible contributing factors and management implications. J. Aquat. Anim. Health 10: 202 - 210. Holt, J.G., Krieg, N .R., Sneath, P.H.A., Staley, J.T., and Williams, S.T. 2001. Genus Renibacterium . th edition. The Williams and Wilkins Co., Baltimore, pp. 567 - 568. Hong, I., Longying, G., Xiujie, S., and Yulin, J. 2002. Detection of Renibacterium salmoninarum by nested - PCR. J. Fish. China/Shuichan Xuebao. 26(5): 453 - 458. Hopwood, D.A., and Ferguson, H.M. 1969. A rapid method for lyophilizing Streptomyces culture. J. Appl. Bact. 32: 434 - 436. Hoskins, G.E., Bell, G.R., an d Evelyn, T.P.T. 1976. The occurrence, distribution, and significance of infectious disease and neoplasms observed in fish in the Pacific Region up to the end of 1974. Fish. Mar. Serv. Res. Dev. Tech. Rep. 609: 37. Hsu, H. - M., Bowser, P.R., and Schachte J r., J.H. 1991. Development and evaluation of a monoclonal - antibody - based enzyme - linked immunosorbent assay for the diagnosis of Renibacterium salmoninarum infection. J. Aquat. Anim. Health 3: 168 - 175. 63 Hsu, H., Wooster, G.A., and Bowser, P.R. 1994. Efficacy of enrofloxacin for the treatment of salmonids with bacterial kidney disease, caused by Renibacterium salmoninarum . J. Aquat. Anim. Health. 6(3): 220 - 223. Humphrey, J.D., Lancaster, C.E., Gudkovs, N., and Copland, J.W. 1987. The disease status in Austral ia of Australian salmonids: bacterial and bacterial diseases. J. Fish Dis. 10: 403 - 410. Jansson, E., Hongslo, T., Lindberg, R., Ljungberg, O., and Svensson, B.M. 1991. Detection of Renibacterium salmoninarum and Yersinia ruckeri by the peroxidase - antipero xidase immunohistochemical technique in melanin - containing cells of fish tissues. J. Fish Dis. 14: 689 - 692. Jansson, E., Torbjörn, H., Höglund, J., and Ljungberg, O. 1996. Comparative evaluation of bacterial culture and two ELISA techniques for the detect ion of Renibacterium salmoninarum antigens in salmonid kidney tissues. Dis. Aquat. Organ. 27: 197 - 206. Jansson, E. 2002. Bacterial kidney disease in salmonid fish: development of methods to assess immune functions in salmonid fish during infection by Reni bacterium salmoninarum . Doctoral Thesis, University of Upsala, Sweden. Jonas, J.L., Schneeberger, P.J., Clapp, D.F., Wolgamood, M., Wright, G., and Lasee, B. 2002. Presence of the BKD - causing bacterium Renibacterium salmoninarum in lake whitefish and bloa ters in the Laurentian Great Lakes. Arch. Hydrobiol. Spec. Issues Advanc. Limnol. 57: 447 - 452. Kent, M.L., Traxler, G.S., Kieser, D., Richard, J., Dawe, S.C., Shaw, R.W., Prosperi - Porta, G., Ketcheson, J., and Evelyn, T.P.T. 1998. Survey of salmonid patho gens in ocean - caught fishes in British Columbia, Canada. J. Aquat. Anim. Health 10(2): 211 - 219. Kettler, S., Pfeilputzien, C., and Hoffman, R. 1986. Infection of grayling (Thymallus) with the agent of bacterial kidney disease (BKD). Bull. Eur. Assoc. Fish Pathol. 6: 69 - 71. Kimura, T., and Awakura, T. 1977. Bacterial kidney disease of salmonid: First observation in Japan. Bull. Jpn. Soc. Sci. Fish. 43: 143 - 150. Kimura, T., and Yoshimizu, M. 1981. Rapid methods for detection of bacterial kidney disease of salmonid (BKD) by coagglutination of antibody sensitized protein A - containing Staphylococci. Bull. Jap. Soc. Sci. Fish. 47(9): 1173 - 1183. Klontz, G.W. 1983. Bacterial kidney disease in salmonids: an overview. In: Anderson, D.P., Dorson, M., and Dubourget, P.H. (Eds.). Antigens of Fish Pathogens: Development and Production for Vaccines and Serodiagnostics. Collection Foundation Marcel Merieux, Lyons, France, pp. 177 - 200. 64 Kozin´ska, A., Mazur, W., and Paszowksa, K. 2001. Cases of bacterial kidney disease in brook trout in Poland. Medycyna Weterynaryjna 57(4): 273 - 275. Lee, E.G - H., and Evelyn, T.P.T. 1994. Prevention of vertical transmission of the bacterial kidney disease agent Renibacterium salmoninarum by broodstock injection with erythromycin. Dis. Aquat. Org. 18: 1 - 4. Lorenzen, E., Olesen, N.J., Korsholm, H., Heuer, O.E., and Evensen, O.E. 1997. First demonstration of Renibacterium salmoninarum /BKD in Denmark. Bull. Eur. Assoc. Fish Pathol. 17(3 - 4): 140 - 144. Mackie, T.J., Arkwright, J.A., Pryce - Tannatt, T.E., Mottram, J.C., Johnston, W.D., and Menzies, W.J. 1933. Second Report of the Furunculosis Committee. HMSO, Edinburgh, 81 pp. MacLean, D.G., and Yoder, W.G. 1970. Kidney disease among Michigan salmon in 1967. Prog. Fish - Cult. 32: 26 - 30. Martínez - Mil lán, L. 1977. New diseases in salmon culture in Spain. Bull. Off. Int. Epiz. 87: 515 - 516. McCarthy, D.H., Croy, T.R., and Amend, D.F. 1984. Immunisation of rainbow trout, Salmo gairdneri Richardson, against bacterial kidney disease: preliminary efficacy e valuation. J. Fish Dis. 7: 65 - 71. McIntosh, D., Austin, B., Flano, E., Villena, A., Martinez - Pereda, J.A., and Tarazona J. 2000. Lack of uptake of Renibacterium salmoninarum by gill epithelia of rainbow trout. J. Fish Biol. 56: 1053 - 1061. Mesa, M.G., Mau le, A.G., Poe, T.P., and Schreck, C.B. 1999. Influence of bacterial kidney disease on smoltification in salmonids: is it a case of double jeopardy? Aquacult. 174: 25 - 41. Meyers, T.R., Short, S., Farrington, C., Lipson, K., Geiger, H.J., and Gates, R. 1993 . Establishment of a negative - positive threshold optical density value for the enzyme - linked immunosorbent assay (ELISA) to detect soluble antigen of Renibacterium salmoninarum in Alaskan Pacific salmon. Dis. Aquat. Org. 16 (3): 191 - 197. Miriam, A., Griff iths, S.G., Lovely, J.E., and Lynch, W.H. 1997. PCR and probe - PCR assays to monitor broodstock Atlantic salmon ( Salmo salar L.) ovarian fluid and kidney tissues for the presence of DNA of the fish pathogen Renibacterium salmoninarum . J. Clin. Microbiol. 35 : 1322 - 1326. Mitchum, D.L., and Sherman, L.E. 1981. Transmission of bacterial kidney disease from wild to stocked hatchery trout. Can. J. Fish. Aquat. Sci. 38: 547 - 551. 65 Mitchum, D.L., Sherman, L.E., and Baxter, G.T. 1979. Bacterial kidney disease in fera l populations of brook trout ( Salvelinus fontinalis ), brown trout ( Salmo trutta ), and rainbow trout ( Salmo gairdneri ). J. Fish. Res. Board Can. 36: 1370 - 1376. Moffitt, C.M., and Bjornn, T.C. 1989. Protection of Chinook salmon smolts with oral doses of ery thromycin against acute challenges of Renibacterium salmoninarum . J. Aquat. Anim. Health 1: 227 - 232. Moffitt, C.M. 1991. Oral and injectable applications of erythromycin in salmonid fish culture. Vet. Human Toxicol. 3: 49 - 53. Moffitt, C.M. 1992. Survival of juvenile Chinook salmon challenged with Renibacterium salmoninarum and administered oral doses of erythromycin thiocyanate for different durations. J. Aquat. Anim. Health 4: 119 - 125. Munro, A.L.S., and Bruno, D.W. 1988. Vaccination against bacterial k idney disease. In: Ellis, A.E. (ed) Fish Vaccination, Academic Press, London, England, pp. 125 - 134. Nagai, T., and Iida, Y. 2002. Occurrence of bacterial kidney disease in cultured ayu. Fish Pathology 37(2): 77 - 81. Nance, S.L., Riederer, M., Zubkowski, T ., Trudel, M., and Rhodes, L.D. 2010. Interpreting dual ELISA and qPCR data for bacterial kidney disease of salmonids. Dis. Aquat. Org. 91: 113 - 119. Nuhfer, A. 2006. Evaluation of brown trout and steelhead competitive interactions in Hunt Creek Michigan. In: Michigan Department of Natural Resources Study Performance Report #230654: October 1, 2005 to September 30, 2006, Lansing, Michigan. Chinook salmon Oncorhyn chus tshawytscha exposed to virulent and attenuated Renibacterium salmoninarum strains. Dis. Aquat. Org. 43: 199 - 209. Ordal, E.J., and Earp, B.J. 1956. Cultivation and transmission of the etiological agent of bacterial kidney disease. Proc. Soc. Exp. Biol . Med. 92: 85 - 88. Ourth, D.D. 1980. Secretory IgM, lysozyme and lymphocytes in the skin mucus of the channel catfish ( Ictalurus punctatus ). Develop. Comp. Immun. 4: 65 - 74. Paclibare, J.O., Albright, L.J., and Evelyn, T.P.T. 1988. Investigations on the oc currence of the kidney disease bacterium Renibacterium salmoninarum in non - salmonids on selected farms in British Columbia. Bull. Aquacult. Assoc. Can. 88: 113 - 115. 66 Pascho, R.J., and Mulcahy, D. 1987. Enzyme - linked immunosorbent assay for a soluble antige n of Renibacterium salmoninarum , the causative agent of salmonid bacterial kidney disease. Can. J. Fish. Aquat. Sci. 44: 183 191. Pascho, R.J., Elliott, D.G., Mallett, R.W., and Mulcahy, D. 1987. Comparison of five techniques for the detection of Renibact erium salmoninarum in adult coho salmon. Tran. Am. Fish. Soc. 116(6): 882 - 890. Pascho, R.J., Elliott, D.G., and Streufert, J.M. 1991. Brood stock segregation of spring Chinook salmon Oncorhynchus tshawytscha by use of the enzyme - linked immunosorbent assay (ELISA) and the fluorescent antibody technique (FAT) affects the prevalence and levels of Renibacterium salmoninarum infection in progeny. Dis. Aquat. Org. 12: 25 40. Pascho, R.J., Elliott, D.G., and Achord, S. 1993. Monitoring of the in - river migration of smolts from two groups of spring Chinook salmon, Oncorhynchus tshawytscha (Walbaum), with different profiles of Renibacterium salmoninarum infection. Aquacult. Fish. Management 24: 163 - 169. Pascho, R.J., Chase, D., and McKibben, C.L. 1998. Comparison o f the membrane - filtration fluorescent antibody test, the enzyme - linked immunosorbent assay, and the polymerase chain reaction to detect Renibacterium salmoninarum in salmonid ovarian fluid. J. Vet. Diagn. Invest. 10: 60 66. Paterson, W.D., Lall, S.P., and Desautels, D. 1981. Studies on bacterial kidney disease in Atlantic salmon ( Salmo salar ) in Canada. Fish Pathol. 15: 283 - 292. Paterson, W.D., Gallant, C., Desautels, D., and Marshall, L. 1979. Detection of bacterial kidney disease in wild salmonids in th e Margaree River system and adjacent waters using an indirect fluorescent antibody technique. J. Fish. Res. Board Can. 36: 1464 - 1468. Peddie, S. 2004. Nephrocalcinosis. Fish farmers 27(3): 37. Piganelli, J.D., Wiens, G.D., Xhang, J.A., Christensen, J.M., and Kaattari, S.L. 1999. Evaluation of a whole cell p57 - vaccine against Renibacterium salmoninarum . Dis. Aquat. Org. 36: 37 - 44. Pippy, J.H.C. 1969. Kidney disease in juvenile Atlantic salmon ( Salmo salar ) in the Margaree River. J. Fish. Res. Board Can. 26: 2535 - 2537. Price, C.S., and Schreck, C.B. 2003. Stress and saltwater - entry behavior of juvenile Chinook salmon ( Oncorhynchus tshawytscha ): conflicts in physiological motivation. Can. J. Fish. Aquat. Sci. 60: 910 - 918. Pridham, T.G., and Hesseltine, C. W. 1975. Culture collections and patent depositions. Adv. Appl. Microbiol. 19: 1 - 23. 67 Purcell, M.K, Murray, A.L., Elz, A., Park, L.K., Marcquenski, S.V., Winton, J.R., Alcorn, S.W., Pascho, R.J., and Elliott, D.G. 2008. Decreased mortality of Lake Michigan Chinook salmon after bacterial kidney disease challenge: evidence for pathogen - driven selection? J. Aquat. Anim. Health 20(4): 225 - 235. Rathbone, C.K., Harrell, L.W., Peterson, M.E., Poysky, F.T., and Strom, M.S. 1999. Preliminary observations on the eff icacy of azithromycin for chemotherapy of bacterial kidney disease. Western Fish Disease Workshop, Fish Health Section/American Fisheries Society Annual Meeting, Twin Falls, Idaho. Rhodes, L.D., Nilsson, W.B., and Strom, M.S. 1998. Sensitive detection of Renibacterium salmoninarum in whole fry, blood, and other tissues of Pacific salmon by reverse transcription - polymerase chain reaction. Mol. Mar. Biol. Biotech. 7(4): 270 - 279. Rhodes, L.D., Rathbone, C.K., Corbett, S.C., Harrell, L.W., and Strom M.S. 2004 . Efficacy of cellular vaccines and genetic adjuvants against bacterial kidney disease in Chinook salmon ( Oncorhynchus tshawytscha ). Fish Shellfish Immunol. 16(4): 461 - 474. Richards, R.H., Roberts, R.J., and Schlotfeldt, H. - J. 1985. Backterielle erkrankun gen der knochenfische. In: Roberts, R.J., and Schlotfeldt, H. - J. (Eds.), Grundlagen der Fischpathologie, Verlag Paul Parey, Hamburg, Germany, pp. 174 - 207. Rimaila - Pärnänen, E. 2002. First case of bacterial kidney disease (BKD) in whitefish ( Coregonus lava retus ) in Finland. Bull. Eur. Assoc. Fish Pathol. 22(6): 403 - 404. Ross, A.J., and Toth, R.J. 1974. Lactobacillus a new fish pathogen? Prog. Fish - Cult. 36: 191. Rucker, R.R., Bernier, A.F., Whipple, W.J., and Burrows, R.E. 1951. Sulfadiazine for kidney disease. Prog. Fish - Cult. 13(3): 135 - 137. Sakai, M.S., and Kobayashi, M. 1992. Detection of Renibacterium salmoninarum the causative agent of bacterial kidney disease in salmonid fish, from pen - cultured coho salmon. Appl. Environment. Microbiol. 58: 1061 - 1063. Sakai, D.K., Nagata, M., Iwami, T., Koide, N., Tamiya, Y., Ito, Y., and Atoda, M. 1986. Attempt to control BKD by dietary modification and erythromycin chemotherapy in hatchery - reared masou salmon Oncorhynchus masou Brevoort. Bull. Jap. Soc. Sci. Fi sh. 52: 61 - 67. Sakai, M., Atsuta, S., and Kobayashi, M. 1993. The immune response of rainbow trout ( Oncorhynchus mykiss ) injected with five Renibacterium salmoninarum bacterins. Aquacult. 113: 11 - 18. 68 Salonius, K., Siderakis, C., MacKinnon, A.M., and Grif fiths, S.G. 2005. Use of Arthrobacter davidanieli as a live vaccine against Renibacterium salmoninarum and Piscirickettsia salmonis in salmonids. Develop. Biol. 121: 189 - 197. Sami, S., Fischter - Scherl, T., Hoffman, R.W., and Pfeil - Putzien, C. 1992. Immun e complex mediated glomerulonephritis associated with bacterial kidney disease in the rainbow trout ( Oncorhynchus mykiss ). Vet. Pathol. 29: 169 - 174. Sanders, J.E., and Barros, J.R. 1986. Evidence by the fluorescent antibody test for the occurrence of Reni bacterium salmoninarum among fish in Chile. J. Wildl. Dis. 22: 255 - 257. Sanders, J.E., and Fryer, J.L. 1980. Renibacterium salmoninarum gen. nov., sp. nov., the causative agent of bacterial kidney disease in salmonid fishes. Intern. J. System. Bacteriol. 30: 496 - 502. Sanders, J.E., Pilcher, K.S., and Fryer, J.L. 1978. Relation of water temperature to bacterial kidney disease in coho salmon ( Oncorhynchus kisutch ), sockeye salmon ( O. nerka ), and steelhead trout ( Salmo gairdneri ). J. Fish. Res. Board Can. 35 : 8 - 11. Sanders, J.E., Long, J.J., Arakawa, C.K., Bartholomew, J.L., and Rohovec, J.S. 1992. Prevalence of Renibacterium salmoninarum among downstream - migrating salmonids in the Columbia River. J. Aquat. Anim. Health 4: 72 - 75. Secombes, C.J. 1985. The in vitro formation of teleost multinucleate giant cells. J. Fish Dis. 8: 461 - 464. Senson, P.R., and Stevenson, R.M. 1999. Production of the 57 kDa major surface antigen by a non - agglutinating strain of the fish pathogen Renibacterium salmoninarum . Dis. Aqua t. Org. 38(1): 23 - 31. Siegel, D.C., and Congleton, J.L. 1997. Bactericidal activity of juvenile Chinook salmon macrophages against Aeromonas salmonicida after exposure to live or heat - killed Renibacterium salmoninarum or to soluble proteins produced by R. salmoninarum . J. Aquat. Anim. Health 9: 180 - 189. Smith, I.W. 1964. The occurrence and pathology of Dee disease. Fresh. Salmon Fish. Res. 34: 3 - 12. Snieszko, S.F., and Griffin, P.J. 1955. Kidney disease in brook trout and its treatment. Prog. Fish - Cult. 17: 3 - 13. Souter, B.W., Dwilow, A.G., and Knight, K. 1987. Renibacterium salmoninarum in wild Arctic char Salvelinus alpinus and lake trout S. namaycush from the Northwest Territories, Canada. Dis. Aquat. Org. 3: 151 - 154. 69 Speare, D.J. 1997. Differences in patterns of meningoencephalitis due to bacterial kidney disease in farmed Atlantic and Chinook salmon. Res. Vet. Sci. 62: 79 - 80. St. Louis - Cormier, E.A., Osterland, D.K., and Anderson, P.D. 1984. Evidence for a cutaneous secretory immune system in the ra inbow trout ( Salmo gairdneri ). Develop. Comp. Immunol. 8: 71 - 80. Starliper, C.E., Smith, D.R., and Shatzer, T. 1997. Virulence of Renibacterium salmoninarum to salmonids. J. Aquat. Anim. Health 9(1): 1 - 7. Starliper, C.E., William, B.S., and Jay, M. 1998. Performance of serum - free broth media for growth of Renibacterium salmoninarum . Dis. Aquat. Org. 34: 21 - 26. Starliper, C.E. 1996. Genetic diversity of North American isolates of Renibacterium salmoninarum . Dis. Aquat. Org. 27: 207 - 213. Stuart, S.E., and Welshimer, H.J. 1974. Taxonomic reexamination of Listeria pirie and transfer of Listeria grayi and Listeria murrayi to a new genus Murraya . Int. J. Syst. Bacteriol. 24: 177 - 185. Suzumoto, B.K., Schreck, C.B., and McIntyre, J.D. 1977. Relative resistances of three transferrin genotypes of coho salmon ( Oncorhynchus kisutch ) and their hematological responses to bacterial kidney disease. J. Fish. Res. Board Can. 34(1): 1 - 8. Traxler, G., and Bell, G. 1988. Pathogens associated with impounded Pacific herring C lupea harengus pallasi , with emphasis on viral erythrocytic necrosis (VEN) and atypical Aeromonas salmonicida . Dis. Aquat. Org. 5: 93 - 100. Turaga, P.S., Wiens, G.D., and Kaattari, S.L. 1987. Analysis of Renibacterium salmoninarum antigen production in sit u . Fish Pathol. 22: 209 - 214. Wedemeyer, G.A., and Ross, A.J. 1973. Nutritional factors in the biochemical pathology of corynebacterial kidney disease in the coho salmon ( Oncorhynchus kisutch ). J. Fish. Res. Board Can. 30: 296 - 298. Wellington, E.M.H., and Williams, S.T. 1979. Preservation of actinomycetes inoculum in frozen glycerol. Microbios. Lett. 6: 151 - 157. White, M.R., Wu, C.C., and Albregts, S.R. 1995. Comparison of diagnostic tests for bacterial kidney disease in juvenile steelhead trout ( Oncorhyn chus mykiss ). J. Vet. Diagn. Invest. 7(4): 494 - 499. Wiens, G.D., and Kaattari, S.L. 1989. Monoclonal antibody analysis of common surface protein(s) of Renibacterium salmoninarum . Fish Pathol. 24: 1 - 7. 70 Wiens, G.D., and Kaattari, S.L. 1991. Monoclonal anti body characterization of a leukoagglutinin produced by Renibacterium salmoninarum . Infect. Immunol. 59: 631 - 637. Wiens, G.D., and Kaattari, S.L. 1999. Bacterial kidney disease ( Renibacterium salmoninarum ). In: Woo, P.T.K., and Bruno, D.W. (eds.), Fish Dis eases and Disorders, CABI Publishing, New York, U.S.A. pp. 269 - 302. Winter, G.W., Schreck, C.B., and McIntyre, J.D. 1980. Resistance of different stocks and transferrin genotypes of coho salmon, Oncorhynchus kisutch , and steelhead trout, Salmo gairdneri , to bacterial kidney disease and vibriosis. Fish. Bull. 77 (4): 795 802. Wolf, K., and Dunbar, C.E. 1959. Tests of 34 therapeutic agents for control of kidney disease in trout. Trans. Am. Fish. Soc. 88: 117 - 124. Wood, J.W. 1974. Diseases of Pacific salmon , their prevention and treatment. Olympia, Washington: Dep. Fish. Hatchery Div., 82 pp. 2 nd edition. Wood, P.A., and Kaattari, S.L. 1996. Enhanced immunogenicity of Renibacterium salmoninarum in Chinook salmon after removal of the bacterial cell surface - a ssociated 57 kDa protein. Dis. Aquat. Org. 25: 71 - 79. Wood, E.M., and Yasutake, W.T. 1956. Histopathology of kidney disease in fish. Am. J. Pathol. 32: 845 - 857. Woodall, A.N., and LaRoche, G. 1964. Nutrition of salmonid fishes. 11. Iodide requirements of Chinook salmon. J. Nutr. 82: 475 - 482. World Organisation for Animal Health (OIE). 2012. Manual of Diagnostic Tests for Aquatic Animals, 7th edition. Office International des Epizooties, Paris. Yoshimizu, M. 1996. Disease problems of salmonid fish in Jap an caused by international trade. - 549. Young, C.L., and Chapman, G.B. 1978. Ultrastructural aspects of the causative agent and renal histopathology of bacterial kidney disease in brook trout ( Sa lvelinus fontinalis ). J. Fish. Res. Board Can. 35(9): 1234 - 1248. 71 Chapter 2 Epidemiological investigation of Renibacterium salmoninarum in three Oncorhynchus spp. in Michigan from 2001 to 2010 Faisal, M., Schulz, C., Eissa, A., Brenden, T., Win ters, A., Whelan, G., Wolgamood, M., Eisch, E., and VanAmberg, J. 2012. Epidemiological investigation of Renibacterium salmoninarum in three Oncorhynchus spp. in Michigan from 2001 to 2010. Prev . Vet . Med . 107: 260 - 274. Reproduced and repri nted with permission by Elsevier B.V. 72 1. Abstract Bacterial kidney disease (BKD) has caused mortalities and chronic infections in wild and farm - raised salmonids throughout the world. In the Laurentian Great Lakes of North America, BKD was associa ted with several large - scale mortality events of Oncorhynchus spp. throughout the 1980s and 1990s. In response to these mortality events, the state of Michigan implemented several enhanced biosecurity measures to limit the occurrence of BKD in state - opera ted hatcheries and gamete - collection weirs. The objectives of this study were to assess if infection levels (prevalence and intensity) of Renibacterium salmoninarum, the causative agent of BKD, have changed in broodstock and pre - stocking fingerlings of th ree feral Oncorhynchus spp. [Chinook salmon ( O. tshawytscha ), coho salmon ( O. kisutch ), and steelhead ( O. mykiss )] over a decade, following the implementation of the enhanced biosecurity measures. Between 2001 and 2010, a total of 3,530 broodstock salmoni ds collected from lakes Huron and Michigan tributaries during spawning runs and 4,294 propagated pre - stocking salmonid fingerlings collected from three state of Michigan fish hatcheries were tested for the presence of R. salmoninarum antigens using the enz yme - linked immunosorbent assay. Substantial declines in the overall prevalence of the bacterium were detected in each of the examined broodstocks. Most propagated pre - stocking fingerlings also exhibited substantial declines in R. salmoninarum prevalence. Prevalence was typically higher in Chinook salmon from Lake Michigan than from Lake Huron; prevalence was also generally higher in the Hinchenbrooke strain of coho salmon than in the Michigan - adapted strain. For most strains and stocks examined, intensi ty of R. salmoninarum infection was found to have declined. Although there were declines in the 73 potential for shedding the bacteria for both male and female Chinook and coho salmon, overall shedding rates were generally low (<15%) except for Hinchenbrooke coho salmon strain, which had shedding prevalences in excess of 50% at the beginning of the study. This study provides evidence that enhanced biosecurity measures at culture facilities and collection sites are capable of severely curtailing disease infec tion in wild populations even at the scale of Lake Michigan fisheries. 2. Introduction The Laurentian Great Lakes (LGL) support a diverse fish community, including Chinook salmon ( Oncorhynchus tshawytscha ), coho salmon ( O. kisutch ), and steelhead ( O. my kiss ; the migratory strain of rainbow trout), which are both recreationally and ecologically important. Even though these three species have supported valuable sport fisheries for a number of decades, they are not native to the LGL and were initially intr oduced to exert predatory pressure on and reduce densities of non - native pelagic prey fishes (mainly alewives, Alosa pseudoharengus ), and further expand LGL sportfishing opportunities (Keller et al., 1990; Holey et al., 1998; Hansen and Holey, 2002). The stocking attempts proved to be successful and high quality recreational fisheries quickly became established (Hansen and Holey, 2002; Tanner and Tody, 2002). The successful introduction of Chinook salmon, coho salmon, and steelhead to the LGL and the pop ular sport fisheries that resulted, led the State of Michigan to develop a Pacific salmonid rearing program in several state - 74 Chinook salmon, coho salmon, and steelhead are propagated annually using egg a nd milt samples collected at gamete - collection weirs operated by the Michigan Department of Natural hatcheries until they are between 6 and 18 months of age (dep ending on the species) and are stocked at several locations in the LGL, as well as in a number of LGL streams in Michigan Among the pathogens that Oncorhynchus spp. are susceptible to is the Gram - positive diplobacillus Renibact erium salmoninarum , the causative agent of bacterial kidney disease (BKD), which is transmitted both horizontally and vertically and hence extremely difficult to control. BKD can take the form of chronic and acute infections and is characterized by the fo rmation of granulomas in kidneys and other visceral organs (Fryer and Sanders, 1981). The disease was first documented in Michigan hatcheries in 1955 (Allison, 1958) and has since spread to all of the Great Lakes, as well as several inland lakes and river s in the LGL region (Allis on, 1958; MacLean and Yoder, 19 7 0 ; Holey et al., 1998; Beyerle and Hnath, 2002; Eissa et al., 2006; Nuhfer, 2006; Faisal et al., 2010). BKD has been associated with several mortality events of Oncorhynchus spp. in the LGL (MacLe an and Yoder, 197 0 ; Holey et al., 1998). MacLean and Yoder (197 0 ) documented the presence of R. salmoninarum in dead coho salmon from lakes Michigan and Superior. During the period from 1988 to 1992, annual mortality events of Chinook salmon in Lake Mich igan occurred, which was attributed at least in part to a BKD epidemic (Holey et al., 1998). These die - offs had drastic effects on the Lake Michigan Chinook salmon fishery, with the recreational 75 fishery yield declining approximate ly four - fold between the mid and late 1980s (Hansen and Holey, 2002). In response to the BKD epidemic, the MDNR initiated a number of enhanced biosecurity practices at state - operated hatcheries and gamete - collection weirs to limit the occurrence and spread of BKD (Table 2.1). T he MDNR expanded its biosecurity practices to include clinical inspections, culling, egg disinfection, hardening the eggs in water containing the antibiotic erythromycin, regular screening of propagated fish, and treatment with antibiotics, such as erythro mycin. While erythromycin is not yet fully approved by the U.S. Food and Drug Administration to treat BKD, limited use of the antibiotic is allowed as an investigational new animal drug exemption. The objectives of this study were to assess if infection levels (prevalence and intensity) of R. salmoninarum in broodstock and pre - stocking fingerlings from three feral Oncorhynchus spp. during the decade 2001 2010, have changed with the enhanced biosecurity practices. An additional objective of the study was to assess the role of shedding R. salmoninarum in the gametes of the broodstock on the overall prevalence and intensity of R. salmoninarum in pre - stocking fingerlings. 76 3. Materials and Methods 3.1. Fish collection Between 2001 and 2010, a tota l of 3,530 feral, spawning Chinook and coho salmon and steelhead were collected from MDNR gamete - collection weirs in Michigan, USA (Figure 2.1; Table 2.2). Ages of the fish ranged from approximately 3 5 years. For this study, the five species/stocks/stra ins of fish were designated based on the location of collection (or source of gametes) and fish species or strain: Chinook salmon from the Little Manistee River Weir (LMRW - CHS), Chinook salmon from the Swan River Weir (SRW - CHS), the Hinchenbrooke strain of coho salmon from the Platte River Weir (HB - COS), the Michigan - adapted strain of coho salmon from the Platte River Weir (MI - COS), and steelhead from the Little Manistee River Weir (LMRW - STT; Table 2.2). Chinook and coho salmon were sampled during the mont hs of September and October of every year, while steelhead were sampled in April of every year. Overall, male and female fish were collected in roughly equal proportions, although sex ratios did vary across the sampling periods. In addition to the brood stock, 4,294 propagated pre - stocking fingerlings were evaluated for R. salmoninarum (Table 2.3). Fish were between 6 and 18 months of age at time of sampling. Chinook salmon fingerlings were propagated at the Platte River State Fish Hatchery (PRSFH), the Thompson State Fish Hatchery (TSFH), and the Wolf Lake State Fish Hatchery (WLSFH). However, prevalence declines were not assessed for LMRW - CHS at the TSFH because there was only one year of data collection. Coho salmon fingerlings were only propagated a t 77 the PRSFH. Steelhead fingerlings were propagated at the TSFH and WLSFH (Figure 2.1). The environmental conditions at the hatcheries (water temperature, dissolved oxygen levels, fish densities, etc.) were monitored and remained optimal for fish rearing, while reducing the risk of disease. The number of fish sampled varied depending on the availability of fish returning to spawn at the gamete - collection weirs and the number of fish available from MDNR hatcheries (Tables 2.2 and 2.3). Additionally, sacr ificing large groups of fish (i.e., at least 60 fish) is required to provide a 95% confidence of detecting the pathogen, with an assumed minimum incidence of 5% of the disease (Hnath, 1993; Ossiander and Wedemeyer, 1973). A description of the enhanced bio security measures implemented by the MDNR at gamete - collecting weirs and hatchery facilities is provided in Table 2.1. 3.2. Sample collection 3.2.1. Broodstock Length, weight, and sex of all fish collected by MDNR personnel at gamete collection weir s were recorded. Each fish was also thoroughly examined both externally and internally for clinical signs of BKD or other diseases. The gross pathology that was observed included exophthalmia, ascites, granulomas in the kidneys, and swelling and congesti on of hematopoietic organs. Samples (<5 g) of kidneys and spleens were removed in the field and stored on ice in whirlpaks (VWR International, West Chester, PA), as recommended by the 78 American Fisheries Society Bluebook (2012) and the World Organization f or Animal Health (OIE, 2012), and frozen at - 20°C until processing. Additionally, approximately 5 ml of ovarian fluid or milt were collected from the same feral fish, stored on ice in 15 ml centrifuge tubes (Denville Scientific, Inc., Metuchen, NJ), and f rozen at - 20°C once returned to the laboratory. Samples were frozen at - 20°C for no longer than 30 days before being processed by laboratory personnel. 3.2.2. Pre - stocking fingerlings Propagated fish were collected within the hatcheries and euthanize d with an overdose (250 mg/L) of Tricaine Methanesulfonate (MS - 222, Argent Chemicals, Redmond, WA). Lengths and weights of each fish were recorded. Each fingerling was examined both externally and internally for clinical signs of BKD infection. The gros s pathology that was observed included exophthalmia, ascites, granulomas in the kidneys, and swelling and congestion of hematopoietic organs. The entire kidney and a sample of spleen were aseptically removed, stored in whirlpaks, and frozen at - 20°C until processing. Samples were frozen at - 20°C for no longer than 30 days before being processed by laboratory personnel. 3.3. Sample processing Solution (HBSS, Sigma Aldric h, St. Louis, MO) and homogenized on high speed for 2 min with a 79 Biomaster Stomacher (Wolf Laboratories Limited, Pocklington, York, UK) as described by Faisal et al. (2009). Gamete samples were also diluted 1:4 (w:v) with HBSS and vortexed on high speed f or approximately 30 s. Each homogenized kidney and spleen tissue sample and gamete sample were then tested for the presence of R. salmoninarum using the quantitative enzyme - linked immunosorbent assay (Q - ELISA). 3.4. Q - ELISA The general Q - ELISA protoco l outlined in Pascho and Mulcahy (1987), with modifications recommended by Gudmundsdóttir et al. (1993) and Olea et al. (1993), was used to assess R. salmoninarum antigens in all sampled fishes. Homogenized spleen, kidney, and gamete samples (250 µl) were aliquoted into 1.5 - ml microcentrifuge tubes (DOT Scientific, Inc., Burton, MI) containing 250 µl of Phosphate Buffered Saline with Tween - 20 (PBS - T20; Sigma) and 5% goat serum (Sigma) and 50 µl of CitriSolv (Fisher Scientific, Pittsburgh, PA). The purpose of the CitriSolv solvent was to dissolve and remove liquids from the aqueous supernatant (Gudmundsdóttir et al., 1993), while the introduction of 5% goat serum was to increase the sensitivity of the assay (Olea et al., 1993). Additionally, Pascho and Mul cahy (1987) demonstrated the excellent specificity of the procedure by showing that the Q - ELISA did not react with antigens from 11 species of bacteria, including the common fish pathogens Aeromonas salmonicida , Vibrio anguillarum , and Yersinia ruckeri . S amples were vortexed for approximately 10 s, heated at 100°C for 15 min, and then centrifuged at 14,000 rpm for 10 min. The aqueous supernatant of each sample was used for Q - ELISA testing. The positive negative 80 cut - off absorbance for the samples was 0.10 (Meyers et al., 1993). Samples that tested positive were assigned the following antigen levels: low (0.10 0.199), medium (0.20 0.999), and high two negative co ntrols (a negative fish kidney and spleen sample and a dilution buffer) and two positive controls (a positive kidney and spleen sample and the standards supplied with the kit). Broodstock specimens were separated into one of four infection categories bas ed upon Q - ELISA results from the kidney/spleen and gamete samples: (1) individuals that were negative for R. salmoninarum individuals that were positive for R. salmoninarum in the kidney and spleen sample and R. salmoninarum in th that were positive for R. salmoninarum in the kidney and spleen sample and the gamete sample (KS+/G+). Individuals that were positive for R. salmoninarum in the gam ete sample were considered to be potential shedders of the bacterium. 3.5. Data analyses Logistic regression was used to evaluate how R. salmoninarum prevalence in broodstock and propagated pre - stocking fingerlings had changed over time. Logistic reg ression was also used to evaluate how prevalence of broodstock shedding had changed over time. For each response variable, a series of models that differed with respect to model intercepts and slopes were fit to observed infection or shedder data. A comp lete listing and description of the 81 models is presented in Table 2.4. For assessing R. salmoninarum prevalence in broodstock, the most complex model had species/stock/strain - specific intercepts and slopes. For assessing R. salmoninarum prevalence in pre - stocking fingerlings, the most complex model had species/stock/strain/hatchery - specific intercepts and slopes. For assessing shedding p revalence in broodstock, the most complex model had species/stock/strain/sex - specific intercepts and slopes. For each f itted model, Akaike information criterion (AIC) was calculated to evaluate model goodness of fit (a low AIC value indicates a parsimonious model that provides a good fit to the data; Burnham and Anderson, 2002). Changes in rates of R. salmoninarum infect ion intensities over time for both broodstock and propagated pre - stocking fingerlings were evaluated using multinomial logistic regression. Multinomial logistic regress ion was also used to assess changes in the different shedding models rather than cumulative logit models because generally the data did not meet the proportional odd s assumption necessary for a cumulative logit model (Agresti, 2007). When fitting the multinomial logistic regression models to infection intensities, we used the negative infection category (Q - ELISA < 0.10) as the reference category. When fitting the mu ltinomial category. Thus, the multinomial regression models for the infection intensity levels fit models that evaluated how the log - odds of the low, medium, and high in fection intensity levels changed relative to that of non - infected category. For the shedding data, the multinomial logistic regression models evaluated how the log - categories chan d ing category. Unlike the model selection 82 process that was used when fitting the R. salmoninarum prevalence data, intensity and shedding category models were fit individually to each species/stock/strain (broodstock intensity), species/stock/strain/hatcher y (propagated pre - stocking fingerling intensity), and species/stock/strain/sex (broodstock shedding) combination to simplify analysis and facilitate interpretation. All logistic and multinomial logistic regression models were fit in SAS using PROC GLIMMIX (SAS Institute Inc., 2010). To determine if there was any association in R. salmoninarum prevalence between broodstock and progeny, we conducted Pearson correlation analyses on the calculated broodstock and pre - stocking fingerling prevalences. Correlat ion analyses were only conducted on Chinook salmon and coho salmon as there was insufficient data for steel - head for this analysis. For Chinook salmon, there was a 1 year time lag between the broodstock and pre - stocking fingerling comparisons as this spec ies is reared as spring fingerling then released. Thus, Chinook salmon broodstock prevalence in 2007 was compared to pre - stocking fingerling prevalence in 2008. For Coho salmon, there was a 2 - year time lag as fish of this species as this species is reare d as yearlings then released. Thus, coho salmon brood stock prevalence in 2007 was compared to pre - stocking fingerling prevalence in 2009. Correlation analyses were conducted in SAS using PROC CORR (SAS Institute Inc., 2010). 83 4. Results 4.1. Reni bacterium salmoninarum infection prevalence and intensity in salmon broodstocks When data were combined for the five stocks, the overall prevalence of R. salmoninarum was 22.1% (SE = 0.7%). Coho salmon had the greatest overall prevalence of R. salmonin arum at 34.0% (SE = 1.3%), followed by Chinook salmon at 18.0% (SE = 0.9%) and steelhead at 3.8% (SE = 0.9%). For individual stocks, HB - COS had the greatest overall prevalence of R. salmoninarum at 48.2% (SE = 2.3%), followed by LMRW - CHS at 27.3% (SE = 1. 6%), MI - COS at 26.2% (SE = 1.5%), SRW - CHS at 11.7% (SE = 0.9%), and LMRW - STT at 3.8% (SE = 0.9%). The decline of R. salmoninarum prevalence in all five stocks was supported by the calculated AIC values, which showed that the models with the poorest fit w ere those that assumed prevalences had remained constant over time. These models had much greater AIC values than models that allowed prevalence to change over time. For all fitted models where infection was assumed to change over time, the estimated slo pe parameters describing the linear (on a logit scale) change in infection prevalence per year ranged from - 0.678 to - 0.901. In all cases, these slopes were significantly different from 0 at P - values less than 0.0001, providing strong indication that R. s almoninarum rate of infection had indeed declined over time. The four best - performing models had nearly equal AIC values, suggesting that each of these models would be almost equally useful for describing declines of R. salmoninarum infection data. The m odel with the lowest AIC value had species/stock/strain - specific intercepts and species - 84 specific slopes, but it performed only slightly better than the model with species - specific intercepts and species/stock/strain - specific slopes. The next best performi ng model had species/stock/strain - specific intercepts and a common slope, but it again performed only slightly better than the model with a common intercept and species/stock/strain - specific slopes. Additionally, the model with species/stock/strain - specif ic intercepts and slopes had an AIC difference of within 3, suggesting that there was at least some support for this model based on the observed data. Because there were several models with at least some support for being the best model based on observed R. salmoninarum prevalence data, we chose to use AIC model averaging based on AIC weights to calculate a weighted - average of model parameters (intercepts and slopes) from those models. Only parameter estimates from those models with AIC differences of 3. 0 or smaller were included in the model averaging (Burnham and Anderson, 2002). The model - averaged logistic regression slopes that were calculated for each of the species/stocks/strain combinations were - 0.718 (SE = 0.055) for LMRW - CHS, - 0.719 (SE = 0.058 ) for SRW - CHS, - 0.804 (SE = 0.063) for HB - COS, - 0.796 (SE = 0.058) for MI - COS, and - 0.775 (SE = 0.141) for LMRW - STT. Based on these model - averaged slopes, R. salmoninarum prevalence was predicted to have declined at a rate of approximately 51% per year fo r LMRW - CHS (Figure 2.2A) and SRW - CHS (Figure 2.2B). The predicted prevalence of R. salmoninarum infection declined at a rate of approximately 55% per year for HB - COS (Figure 2.2C) and MI - COS (Figure 2.2D). Lastly, the R. salmoninarum predicted prevalence declined at a rate of approximately 54% per year for LMRW - STT (Figure 2.2E). Regardless of sampling year, HB - COS had the greatest predicted R. salmoninarum prevalence of the species/stocks, strains, followed by 85 LMRW - CHS, and MI - COS. SRW - CHS and LMRW - STT had approximately equal R. salmoninarum prevalences during those years where prevalence data were available for both stocks (Figure 2.2A - E). Infection rates of the different intensity levels generally decreased for all species relative to non - infected f ish (Figure 2.2F - J), the only exceptions being medium intensity infection rate for SRW - CHS (Figure 2.2G) and medium and high infection rates for LMRW - STT (Figure 2.2J). The multinomial log - odds for having a low, medium, or high infection intensity level v ersus not being infected declined by 3.14 - 10.80 log - odds units per year depending on the stock and intensity level. In the case of LMRW - CHS, predicted low, medium, and high intensity infection prevalences declined to less than 0.1% during the sampling per iod (Figure 2.2F). For SRW - CHS, the predicted low and high intensity infection prevalences also declined to less than 0.1% during this same time period (Figure 2.2G). There were also declines in the predicted low, medium, and high intensity infection pre valences for HB - COS (Figure 2.2H). For MI - COS, the predicted low and medium intensity infection rates declined to less than 1% in 2010 (Figure 2.2I), while the high intensity infection prevalences for MI - COS initially increased 2003; however, since then, the high infection prevalences have declined to a rate of 1.2% in 2010 (Figure 2.2I). For LMRW - STT, low intensity infection prevalences also declined throughout the study period (Figure 2.2J). 86 4.2. Renibacterium salmoninarum infection prevalence and intensity in propagated pre - stocking salmon fingerlings When data were combined for the five stocks over the entire study period, the overall prevalence of R. salmoninarum was 15.0% (SE = 0.5%). For individual species, coho salmon had the greatest ove rall prevalence of R. salmoninarum at 20.0% (SE = 1.3%), followed by Chinook salmon at 18.1% (SE = 0.8%) and steelhead at 6.7% (SE = 0.7%). For individual stocks, HB - COS had the greatest overall prevalence of R. salmoninarum at 28.1% (SE = 2.4%), followed by LMRW - CHS at 24.2% (SE = 1.3%), MI - COS at 14.6% (SE = 1.5%), SRW - CHS at 11.4% (SE = 1.0%), and LMRW - STT at 6.7% (SE = 0.7%). When categorized by hatchery, LMRW - CHS WLSFH had an overall prevalence of 32.4% (SE = 2.1%), whereas LMRW - CHS PRSFH had an over all prevalence of 19.4% (SE = 1.7%) and LMRW - CHS TSFH had an overall prevalence of 0% (SE = 0.0%). Conversely, SRW - CHS WLSFH had an overall prevalence of 13.3% (SE = 2.2%), whereas SRW - CHS PRSFH had an overall prevalence of 18.3% (SE = 2.1%) and SRW - CHS T SFH had an overall prevalence of 4.0% (SE = 0.9%). Steelhead propagated at the TSFH and WLSFH had overall prevalences of 7.9% (SE = 0.01%) and 5.8% (SE = 0.01%), respectively. Like the broodstock analysis, based on calculated AIC values, the models with the poorest fit to observed R. salmoninarum infection data for the propagated pre - stocking fingerlings were those that assumed infection rates had remained constant over time. Unlike the results from the broodstock analysis, however, for the propagated p re - stocking fingerlings there was a single model that was picked by the AIC model selection criteria as having vastly superior performance compared to the other models. The model with the lowest AIC value had 87 species/ stock/strain/hatchery specific model intercepts and slopes. Therefore, the best performing model demonstrated that there was a considerable difference among the species and strains of fish, as well as the hatcheries where the fish were propagated. The next best performing model had stocks/h atchery specific intercepts and species/stock specific slopes, but this AIC difference for this model was greater than 90, indicating that there was very little empirical support for this model (Burnham and Anderson, 2002). Analysis demonstrated that the R. salmoninarum infection rate in propagated fish had indeed declined over time, since for most of the stock/hatchery combinations, the logistic regression slopes were negative and were significantly different from 0 at P - values < 0.0001 (Figure 2.3A - E). The only exceptions to this were for LMRW - CHS from the TSFH for which a slope could not be calculated because there was only 1 year of data collected from this hatchery, and for SRW - CHS from the TSFH which had a positive slope indicating that R. salmonina rum infection was increasing over time (Figure 2.3E). For the other stock/hatchery combinations, estimated slopes equaled - 1.295 (SE = 0.135) for LMRW - CHS PRSFH, - 1.241 (SE = 0.116) for LMRW - CHS WLSFH, - 1.325 (SE = 0.176) for SRW - CHS PRSFH, - 1.196 (SE = 0 .252) for SRW - CHS WLSFH, - 0.745 (SE = 0.111) for HB - COS PSFH, - 0.730 (SE = 0.096) for MI - COS PRSFH, - 2.079 (SE = 0.000) for LMRW - STT WLSFH, and - 0.317 (SE = 0.0967) for LMRW - STT TSFH. Based on these estimated slopes, R. salmoninarum prevalences in propag ated pre - stocking fingerlings were predicted to decline by between 52% to 87% per year for each of these species/stock/strain/hatchery combinations. In terms of predicted R. salmoninarum prevalences, prevalence of R. salmoninarum for LMRW - CHS from the WLS FH was predicted to have declined to less than 0.1% in 2010 (Figure 2.3A). LMRW - CHS from the PRSFH also 88 experienced a major decline in R. salmoninarum prevalence (Figure 2.3B). For SRW - CHS from the WLSFH, prevalences were predicted to have declined to 1. 4% (Figure 2.3C), whereas the predicted prevalence of R. salmoninarum in SRW - CHS from the PRSFH declined to less than 0.1% (Figure 2.3D). In terms of infection intensity, there were significant declines in rates of infection at all intensity levels using non - infected as the baseline for LMRW - CHS at both PRSFH and WLSFH (Figure 2.3F - J). For LMRW - CHS at the WLFSH, predicted rates of infection at low and medium intensity levels initially increased from 2002 to 2003; however, since 2003, the predicted preval ence of low and medium intensity infection levels has declined to less than 0.1% (Figure 2.3F). Predicted prevalence of high intensity infection levels for LMRW - CHS at the WLSFH declined overall from 2002 to 2010 (Figure 2.3F). The predicted rates of inf ection at low, medium, and high intensities for LMRW - CHS from the PRSFH declined to less than 0.1% (Figure 2.3G). For SRW - CHS, significant declines in rates of infection for low intensity levels were detected at the WLSFH (Figure 2.3H), as well as for low and medium intensity levels at the PRSFH (Figure 2.3I). Additionally, the predicted prevalence for medium intensity levels of infection for SRW - CHS at the PRSFH declined to less than 0.1% (Figure 2.3I). For SRW - CHS from the TSFH, there was a significant increase in prevalence at medium intensity levels from 2002 to 2010 (Figure 2.3J). Coho salmon fingerlings propagated at the PRSFH also saw a substantial decline in the prevalence of R. salmoninarum infections from 2003 to 2010. For HB - COS and MI - COS f rom the PRSFH, prevalence was predicted to have declined throughout the study period (Figure 2.4A and 2.4B). In regards to infection intensity for coho salmon, significant decreases in low rates 89 of infection were found for both Hinchenbrooke and Michigan - adapted strains (Figure 2.4C and 2.4D). Also, a significant decrease in medium rate of infection was found for the MI - COS (Figure 2.4D). Moreover, the prevalence infection rates of R. salmoninarum in steelhead pre - stocking fingerlings also decreased con siderably from 2005 to 2010. For LMRW - STT from the WLSFH, predicted prevalence declined to less than 0.1% in 2010 (Figure 2.5A); while for LMRW - STT from the TSFH the predicted prevalence declined to 4.0% in 2010 (Figure 2.5B). For LMRW - STT, significant d ecreases in prevalence at low intensity levels were observed for both WLSFH and TSFH (Figure 2.5C and 2.5D). 4.3. Renibacterium salmoninarum in gametes of spawning broodstock Across all species/stocks/strains throughout the study period, the overall p revalence of broodstock that had R. salmoninarum antigens in ovarian fluid and milt was 6.2% (SE = 0.5%). By species, the overall prevalence of brood - stock with positive gametes was 11.1% (SE = 1.1%) for coho salmon, 4.3% (SE = 0.6%) for Chinook salmon, a nd 0.7% (SE = 0.5%) for steelhead. By species/stock/strain, the over - all prevalence of gamete positive broodstock was 3.6% (SE = 0.8%) for LMRW - CHS, 4.9% (SE = 0.8%) for SRW - CHS, 23.6% (SE = 2.4%) for HB - COS, and 4.3% (SE = 0.9%) for MI - COS. When calcula ted by sex, the overall prevalence of broodstock that were shedding the bacteria in gametes was 6.5% (SE = 0.7%) for females and 5.9% (SE = 0.7%) for males. 90 Based on the calculated AIC values, there were two models that had some support based on observed shedding data. The model with the lowest AIC value had species/stock/strain/sex - specific intercepts and slopes. The next best performing model had species/stock/strain/sex - specific model intercepts and species/stock/strain - specific slopes. As with the broodstock prevalence analysis, we used model averagi ng based on AIC weights to aver age the parameter estimates from these two models. The calculated model - averaged slopes equaled - 0.515 (SE = 0.236) for female LMRW - CHS, - 0.983 (SE = 0.237) for male LMRW - CHS, - 0.384 (SE = 0.208) for female SRW - CHS, - 0.093 (SE = 0.179) for male SRW - CHS, - 2.165 (SE = 0.001) for female HB - COS, - 2.132 (SE = 0.025) for male HB - COS, - 0.379 (SE = 0.123) for female MI - COS, - 0.430 (SE = 0.213) for male MI - COS, 0.088 (SE = 0.113) fo r female LMRW - STT, and 0.210 (SE = 0.018) for male LMRW - STT. Based on model - averaged slopes, gamete shedding prevalences were predicted to have declined by between 40% and 63% per year for LMRW - CHS (Figure 2.6A) and 9% and 32% per year for SRW - CHS (Figure 2.6B). Additionally, the predicted decline in gamete shedding prevalence for HB - COS was approximately 88% per year (Figure 2.6C) and between 32% and 35% per year for MI - COS (Figure 2.6D). For LMRW - STT, shedding prevalence was predicted to have increased by between 9% and 23% (Figure 2.6E). In terms of predicted gamete shed ding prevalences, for most species/stocks/strains shedding prevalence in gametes was generally less than 15% throughout the course of the study. The one exception to this was HB - COS where both females and males had gamete shedding prevalences of between 50% and 60% at the beginning of this study. However, by 2007, gamete shedding prevalences of both sexes had declined to less than 1% (Figure 2.6C). Although there were clear differen ces between sexes in predicted shedding prevalences for 91 some of the examined broodstock, results were inconsistent as to whether males or females had higher shedding prevalences. In terms of the different shedding categories, significant declines in shedd detected for some of the species/stock/strain/sex combinations. For HB - COS, significant LMRW - CHS , SRW - CHS, and MI - were detected, while for male LMRW - CHS, SRW - CHS, and MI - COS, only significant declines in - STT, no significant declines in shedding rates for any of the shedding categories were detected. The largest predicted changes in shedding - CHS, which were predicted to have declined from between 40% and 60% in 2004 to less than 3% by 2007 (Figure 2.6A). The other large predicted change in shedding prevalences were in the KS+/G+ category for female and male HB - COS, which were predicted to have declined from between 40% and 55% in 2004 to less than 5% in 2005 (Figure 2.6C). 4.4. Relationship between R. salmoninarum prevalence in broodstock and progeny Based on the correlation analyses conducted, no significant association between broodstock and progeny prevalences in Chinook salmon (r = 0.220; P - value = 0.517) was detect ed. There was, however, a statistically significant positive association found between broodstock and progeny prevalences in coho salmon (r = 0.890; P - value = 0.0073). A strong 92 positive, albeit not statistically significant, associati on was also found be tween brood stock and progeny prevalences in steelhead (r = 0.697; P - value = 0.0549). 5. Discussion The findings of this study demonstrate that the intensity and prevalence of R. salmoninarum infections have decreased during the course of the last deca de; not only in the feral broodstock, but also in hatchery settings. This correlated with the implementation of enhanced biosecurity practices in the state fish hatcheries. In this study, we opted to use the Q - ELISA as the only diagnostic tool to assess R. salmoninarum presence in fish tissues. In a previous study, it was demonstrated that Q - ELISA values were commensurate with disease progression in feral stocks (Faisal and Eissa, 2009). There are several likely reasons why R. salmoninarum has rapidly declined in an area where it has been endemic for over half a century. As fish are exposed to R. salmoninarum , it is possible that the overall population may have shifted to individuals with heightened resistance to the pathogen (Purcell et al., 2008). E arlier studies on coho salmon have shown that resistance to BKD is linked to the transferrin gene (Suzumoto et al., 1977; Winter et al., 1980). more likely to die from a R. salmoninarum infection than coho salmon with the transferrin survive, and produce offspring that are also less likely to become infected with R. salm oninarum , thereby reducing the prevalence of the disease over time. 93 An additional factor that may explain the minimal presence of R. salmoninarum in state fish hatcheries is the improved screening process for signs of diseases that was initiated at MDNR egg - collection weirs in the 1990s but expanded in the early 2000s. Once gametes were removed from adult Oncorhynchus spp., the fish were examined externally (presence of ulcers, furuncles, lesions, etc.) and internally (pale or swollen organs, granulomas, hemorrhages, etc.) by trained MDNR staff and experienced fish health professionals for signs of disease. If fish were suspected to be harboring R. salmoninarum , the gametes were not used for production. By using this method, fish affected by the acute f orm of BKD were removed from the brood - stock population, in addition to those that developed the chronic form of BKD (characterized by the formation of renal granulomas); thereby reducing the potential amount of R. salmoninarum that would be passed on to t he progeny. Elliott et al. (1995) found significant differences in clinical BKD signs in Chinook salmon progeny from parent broodstock with a low prevalence of R. salmoninarum infection (low - BKD), which would mimic a chronic infection, when compared to a broodstock with a high prevalence of infection (high - BKD), which would be similar to an acute infection. Compared to the fish in the low - BKD group, a higher proportion of fish in high - BKD group had evidence of organ and tissue pathology, such as exophthal mia, corneal opacity, pale and/or frayed gills, fin erosion, swollen or mottled kidneys, enlarged spleens, and abnormal livers. While it is possible that the fish with low intensities of infection will not be detected by a screening method such as visual observation of the disease when compared to highly infected fish, fish with the low intensity of infection pose less of a risk of passing R. salmoninarum to their offspring. Similar to Elliott et al. (1995), Pascho et al. (1991) found that within the prog eny that had positive ELISA results, most of the low - BKD group had low 94 intensity infections, while the majority of the fish in the high - BKD group had high intensity infections. In this context, a long - term study (1993 - 2005) of R. salmoninarum infections in Chinook salmon in Idaho hatcheries has demonstrated the effectiveness of establishing a program such as screening the broodstock by an ELISA method and then culling based on the results, and continuing it on a yearly basis (Munson et al., 2010). It wa s found that the ELISA - based (as outlined in Munson et al., 2010) broodstock screening program reduced the prevalence, the intensity of infection, and mortality rates due to R. salmoninarum in Chinook salmon juveniles and broodstock. Our findings in Michi gan corroborate with those of Munson et al. (2010) in Idaho. Combining broodstock culling with the more stringent biosecurity measures implemented at the MDNR hatcheries as of 2002, have minimized, not only the vertical transmission from broodstocks, but also limited potential for horizontal transmission of R. salmoninarum . As recommended by Danner and Merrill (2006), each of the six state fish hatcheries utilizes separate nets, brushes, and buckets for each of the raceways that are cleaned and disinfect ed on a regular basis. Additionally, disinfecting footbaths and mats are placed at the entrance to all facilities to reduce the possible cross - contamination between facilities. Another method of disease prevention that the MDNR implemented in 2002 is th e use of the antibiotic erythromycin as a therapeutic treatment for hatchery salmon and trout. As it is extra - label use to use erythromycin as a therapeutic, the MDNR used it under the supervision of a veterinarian. Renibacterium salmoninarum is known to be susceptible to exposure to 95 erythromycin (Stoffregen et al., 1996). While MDNR uses erythromycin baths, Evelyn et al. (1986a) and Lee and Evelyn (1994) demonstrated that the intramuscular injection of erythromycin into broodstock fish can minimize the vertical transmission of R. salmoninarum into the ova. Additionally, Evelyn et al. (1986a) showed the antibiotic persisted within the eggs after attempts were made to leach it out; suggesting that it reduces the initial vertical transmission from parent t o offspring; and potentially lowers the risk of horizontal transmission. It is likely that the erythromycin therapeutic baths that MDNR implemented on all egg lots acted in a similar fashion and further contributed to minimizing the potentials of vertical transmission. Unfortunately, as the MDNR could not risk the loss of valuable propagated fish, it was not possible to include a negative control (i.e., a hatchery with no newly implemented enhanced biosecurity measures) for the sake of comparison for this study. Our finding of differences in Renibacterium salmoninarum prevalence among the three Oncorhynchus species matches the results of Starliper et al. (1997) and Beacham and Evelyn (1992). Starliper et al. (1997) demonstrated that strains of R. salmon inarum from coho salmon have the potential to be more harmful than strains of the bacteria from Chinook salmon. In six out of eight different salmonid hosts, a strain of R. salmoninarum from coho salmon from Manistee, Michigan was found to be more virulen t than a strain from Chinook salmon from Manistee, Michigan (Starliper et al., 1997). Also, Starliper et al. (1997) found that coho and Chinook salmon were more susceptible to R. salmoninarum than rainbow (steelhead) trout. Furthermore, in a study by Bea cham and Evelyn (1992), juvenile coho and Chinook salmon were infected with R. salmoninarum to better understand how the bacterium affected mortality rates, the mean time to death, and growth rates of the three species. It was concluded that 96 although they had a longer time to death, coho salmon had a higher percent of mortality than Chinook salmon (Beacham and Evelyn, 1992). A possible contributory factor that may help explain the greater prevalence of R. salmoninarum in coho salmon in this study is that broodstock of this species were only collected from streams in the Lake Michigan watershed, which overall had a greater prevalence of R. salmoninarum compared to fish collected from Lake Huron tributary streams. The generally greater prevalence rate of R . salmoninarum in HB coho salmon versus MI coho salmon suggests that the strains may differ in their susceptibility to BKD. The studies of Withler and Evelyn (1990) documented that such variations in disease susceptibility can exist between strains of coh o salmon. These investigators exposed two strains of coho salmon from British Columbia to R. salmoninarum to determine the likelihood of resistance to R. salmoninarum , as determined by survivability and time to death. The Kitimat River strain of coho sal mon had greater survival and a longer time to death when compared to the Robertson Creek strain of coho salmon. In addition to our study, several other studies have documented low occurrences of R. salmoninarum in O. mykiss , at least when compared to occ urrence in brook trout, brown trout, Chinook salmon, and coho salmon (Bullock et al., 1971; Mitchum and Sherman, 1981; Mitchum et al., 1979; Hsu et al., 1991; Sakai et al., 1991; Jansson et al., 1996; Starliper et al., 1997). Additionally, Mitchum and She rman (1981) found that O. mykiss had the lowest mortalities and the least severe clinical signs of BKD in a study investigating the horizontal transmission of R. salmoninarum from infected wild brook trout to newly stocked hatchery - raised brook trout, brow n trout, and rainbow trout. Hsu et al. (1991) also found that steelhead in Lake Ontario had 97 lower prevalences of R. salmoninarum and also lower detectable antigen levels than coho or Chinook salmon among fish returning to the Salmon River Fish Hatchery (A ltmar, New York) based on monoclonal - antibody - based ELISA. The prevalence of R. salmoninarum shedding in this study was low, with the exception of the HB coho salmon, which may be attributable to the finding that HB coho salmon had the highest overall pre valence of R. salmoninarum in the kidney and spleen of the parental broodstock. The prevalence of shedding R. salmoninarum in coho salmon broodstocks in this study corresponded to the prevalence of the pathogen in the progeny. For example, in 2004, 72.5 % of HB coho salmon were positive for R. salmoninarum and 52.5% of them were capable of shedding the bacteria. As a result, 43.3% of their progeny were positive for the pathogen. Interestingly, in 2005, the prevalence of R. salmoninarum in broodstock dec reased to 18.3% with 11.7% of them shedding the bacteria. Consequently, none of the progeny were positive for R. salmoninarum . While vertical transmission clearly plays a role in the infection of the progeny, there are several other factors that can aff ect the prevalence of the disease as well, such as density of the fish in the raceway, environmental conditions (water temperature, dissolved oxygen levels, etc.), and biosecurity measures at the facility. It is clear that female salmonids contribute to vertical transmission by having R. salmoninarum - infected ovarian fluid, but the role of male salmonids in vertical transmission is less understood. The shedding in this study occurred fairly equally between male and female fish; however, in the case of SR W - CHS and LMRW - STT, the males had a somewhat higher shedding prevalence than the females. Evelyn et al. (1986b) concluded that male coho salmon and steelhead did not play a significant role in vertical transmission as a result of their studies 98 examining i nfection rates in eggs fertilized with infected or non - infected milt. Based on our findings, it is at least plausible that male SRW - CHS and LMRW - STT could be contributing to the vertical transmission of R. salmoninarum , although the extent this is occurri ng is not known. The overall decline of R. salmoninarum from 2001 to 2010 in the three feral broodstock and propagated fish stocks has shown that in addition to a possible heightened genetic resistance, preventative measures such as an improved screening process, broodstock culling, and enhanced biosecurity measures can be successful in reducing the prevalence of a pathogen in hatcheries and perhaps in returning broodstock from the Great Lakes. The broad decline in the prevalence of R. salmoninarum in th e various fish species and stocks in this study is most likely due to a combination of improved visual inspections and culling conducted at the weirs, implementation of increased biosecurity measures at the hatcheries, reduced rearing stress, and iodophor and erythromycin disinfection of eggs. While the decline of R. salmoninarum in these three salmonid species is promising, BKD continues to be a potential problem in the LGL basin. Lake whitefish (Coregonus clupeaformis), which are in a closely related su bfamily to salmonids (i.e., Coregoninae), have been shown to heavily infected with R. salmoninarum . Recently, Faisal et al. (2010) documented the presence of R. salmoninarum in approximately 66% of the lake whitefish populations sampled in northern Lake H uron by the Q - ELISA method described above, with predominant clinical signs of infection. This high infection prevalence in another susceptible fish species attests for the continuous strong presence of R. salmoninarum in the LGL ecosystem. It further st rengthens the finding of this study that the decline of the bacterial presence in Oncorhynchus species is the result of disease management measures undertaken in Great Lakes salmonid gamete collection weirs and state fish hatcheries. 99 A PPENDIX 100 Table 2.1. Enhanced biosecurity measures to control Bacterial Kidney Disease that have been implemented at Michigan Departmen t of Natural Resources gamete - collecting weirs and hatchery facilities. Biosecurity measure Description Clinical inspectio n Examination for disease signs such as hemorrhages, exophthalmia, congested internal organs, and granulomas Culling Euthanasia of any individuals exhibiting the above signs Egg disinfection Disinfecting the external surface of the eggs to reduce the amo unt of bacteria Hardening eggs in erythromycin - laden water Water hardening the eggs (a necessary step for fertilization) in the antibiotic erythromycin, which Renibacterium salmoninarum is susceptible to Regular screening Frequent testing of propagated fish to determine if infection prevalence or mortality exceeded 0.05% Antibiotic treatment If prevalence indeed exceeded 0.05%, treatment with antibiotics under the investigational new animal drug exemption (INAD) , as chosen by the antibiotic disc diffusi on test 101 Table 2.2. The number of spawning Oncorhynchus spp. analyzed for the presence of R. salmoninarum antigens from 2001 to 2010. Hinchenbrooke coho salmon were not collected after 2007. ND = no data collected. Chinook salmon Coho salmon St eelhead Little Manistee River Weir (LMRW - CHS) Swan River Weir (SRW - CHS) Michigan - adapted (MI - COS) Hinchenbrooke (HB - COS) Little Manistee River Weir (LMRW - STT) Year # of # of TOTAL # of # of TOTAL # of # of TOTAL # of # of TOTAL # of # of TOTAL 2001 30 30 60 ND ND ND 19 19 38 25 24 49 ND ND ND 2002 30 30 60 30 29 59 83 82 165 28 28 56 ND ND ND 2003 30 30 60 17 17 34 30 30 60 24 24 48 ND ND ND 2004 60 60 120 560 0 560 58 59 117 60 60 120 30 30 60 2005 30 30 60 30 30 60 31 30 61 31 30 61 31 31 62 2006 30 30 60 30 30 60 30 30 60 30 30 60 30 30 60 2007 30 30 60 30 30 60 30 30 60 30 30 60 30 30 60 2008 30 30 60 30 30 60 30 30 60 ND ND ND 30 30 60 2009 50 50 100 50 50 100 50 50 100 ND ND ND 30 30 60 2010 50 50 100 50 50 100 50 50 100 ND ND ND 30 30 60 TOTAL 372 371 740 827 266 1093 411 410 821 228 226 454 211 211 422 102 Table 2.3. The number of propagated Oncorhynchus spp. reared in state fish hatcheries and tested for Renibacterium salmoninarum antigens prior to stocking. St ate fish hatchery facilities included the Platte River State Fish Hatchery (PRSFH), the Thompson State Fish Hatchery (TSFH), and the Wolf Lake State Fish Hatchery (WLSFH). *Species/strains were not reared at these locations at that year. Year Chinook s almon Coho salmon Steelhead Little Manistee River Weir Swan River Weir Platte River Weir Little Manistee River Weir PRSFH TSFH WLSFH PRSFH TSFH WLSFH HB - PRSFH MI - PRSFH TSFH WLSFH 2002 65 * 60 * * * * * * * 2003 60 * 60 60 60 * 60 60 * * 2004 60 * 60 60 60 60 60 60 * * 2005 60 * 60 60 60 60 120 120 60 60 2006 60 * 60 60 39 60 60 60 30 92 2007 60 * 60 60 60 60 59 60 119 159 2008 60 58 11 * * * * 60 141 137 2009 60 * 60 60 60 * * 60 126 120 2010 60 * 60 * 59 * * 60 120 139 TOTAL 545 58 491 360 39 8 240 359 540 596 707 103 Table 2.4. Listing and description of models fit to the R. salmoninarum prevalence and shedding data. Analysis column indicates models fit only to propagated pre - stocking fingerling prevalence or broodstock gamete shedding dat a. Model Description Analysis Intercept Common intercept; no change over time All Intercept + time Common intercept; common change over time All Intercept + species × time Common intercept; species specific change over time All Intercept + SS × time C ommon intercept; species/stock/strain specific change over time All Intercept + SSH × time Common intercept; species/stock/strain/ hatchery specific change over time Propagated pre - stocking fingerlings Intercept + sex × time Common intercept; sex specifi c change over time Broodstock shedding Intercept + SX Common intercept; species/sex specific change over time Broodstock shedding Intercept + SSX × time Common intercept; species/stock/strain/sex specific change over time Broodstock shedding Species Spe cies specific intercept; no change over time All Species + time Species specific intercept; common change over time All Species + species × time Species specific intercept; species specific change over time All Species + SS × time Species specific inter cept; species/stock/strain specific change over time All Species + SSH × time Species specific intercept; species/stock/strain/ hatchery specific changes over time Propagated pre - stocking fingerlings Species + sex × time Species specific intercept; sex s pecific change over time Broodstock shedding Species + SX × time Species specific intercept; species/sex specific change over time Broodstock shedding Species + SSX × time Species specific intercept; species/stock/strain/ sex specific change over time Br oodstock shedding SS Species/stock/strain specific intercept; no change over time All SS + time Species/stock/strain specific intercept; common change over time All 104 Model Description Analysis SS + species × time Species/stock/strai n specific intercept; species specific change over time All SS + SS × time Species/stock/strain specific intercept; Species/stock/strain specific change over time All SS + SSH × time Species/stock/strain specific intercept; species/ stock/strain/hatcher y specific change over time Propagated pre - stocking fingerlings SS + sex × time Species/stock/strain specific intercept; sex specific change over time Broodstock shedding SS + SX × time Species/stock/strain intercept; species/sex specific change over tim e Broodstock shedding SS + SSX × time Species/stock/strain specific intercept; species/stock/strain/sex specific change over time Broodstock shedding SSH Species/stock/strain/hatchery specific intercept; no change over time Propagated pre - stocking finger lings SSH + time Species/stock/strain/hatchery specific intercept; common change over time Propagated pre - stocking fingerlings SSH + species × time Species/stock/strain/hatchery specific intercept; species specific change over time Propagated pre - stockin g fingerlings SSH + SS × time Species/stock/strain/hatchery specific intercept; species/stock/strain specific change over time Propagated pre - stocking fingerlings SSH + SSH × time Species/stock/strain/hatchery specific intercept; species/stock/strain/hat chery specific change over time Propagated pre - stocking fingerlings Sex Sex specific intercept; no change over time Broodstock shedding Sex + time Sex specific intercept; common change over time Broodstock shedding Sex + species × time Sex specific inte rcept; species specific change over time Broodstock shedding Sex + SS × time Sex specific intercept; species/stock/strain specific change over time Broodstock shedding Sex + sex × time Sex specific intercept; sex specific change over time Broodstock shed ding 105 Model Description Analysis Sex + SX × time Sex specific intercept; species/sex specific change over time Broodstock shedding Sex + SSX × time Sex specific intercept; species/stock/strain/sex specific change over time Broodstock shedding SX Species/sex specific intercept; no change over time Broodstock shedding SX + time Species/sex specific intercept; common change over time Broodstock shedding SX + species × time Species/sex specific intercept; species specific change over ti me Broodstock shedding SX + SS × time Species/sex specific intercept; species/stock/strain specific change over time Broodstock shedding SX + sex × time Species/sex specific intercept; sex specific change over time Broodstock shedding SX + SX × time Spe cies/sex specific intercept; species/specific change over time Broodstock shedding SX + SSX × time Species/sex specific intercept; species/stock/strain/sex specific change over time Broodstock shedding SSX Species/stock/strain/sex specific intercept; no change over time Broodstock shedding SSX + time Species/stock/strain/sex specific intercept; common change over time Broodstock shedding SSX + species × time Species/stock/strain/sex specific intercept; species specific change over time Broodstock sheddi ng SSX + SS × time Species/stock/strain/sex specific intercept; species/stock/strain specific change over time Broodstock shedding SSX + sex × time Species/stock/strain/sex specific intercept; sex specific change over time Broodstock shedding SSX + SX × time Species/stock/strain/sex specific intercept; species/sex specific change over time Broodstock shedding SSX + SSX × time Species/stock/strain/sex specific intercept; species/stock/strain/sex specific change over time Broodstock shedding 106 Figure 2.1. The Michigan Department of Natural Resources state fish hatcheries and gamete - collecting weirs where Chinook ( Oncorhynchus tshawytscha ) and coho salmon ( O. kisutch ) and steelhead ( O. mykiss ) were collected from 2001 to 2010: Little Manistee River Wei r (44°11'51.66"N, 86°11'38.99"W), Platte River Weir and State Fish Hatchery (44°39'48.88"N, 85°56'13.20"W), Swan River Weir (45°24'10.09"N, 83°44'5.52"W ), Thompson State Fish Hatchery (45°57'16.07"N, 86°15'29.36"W), and Wolf Lake State Fish Hatchery (42°1 7'40.14"N, 85°47'2.29"W). 107 Figure 2.2. The prevalence of R. salmoninarum in Chinook salmon ( Oncorhynchus tshawytscha ) broodstock from the Little Manistee River Weir (LMRW - CHS) and the Swan River Weir (SRW - CHS), Hinchenbrooke coho salmon ( O. kisutch ) broodstock (HB - COS) and Michigan - adapted coho salmon broodstock (MI - COS) from the Platte River Weir, and steelhead ( O. mykiss ) broodstock from the Little Manistee River Weir (LMRW - STT) from 2001 to 2010. F - J: The low, medium, and high intensity levels of infection of R. salmoninarum in LMRW - CHS, SRW - CHS, HB - COS, MI - COS, and LMRW - STT from 2001 to 2010. The lines represent the logistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote the observed prev alence and intensity levels of infection. 108 Figure 2.3. The prevalence of R. salmoninarum in Chinook salmon ( Oncorhynchus tshawytscha ) pre - stocking fingerlings propagated at the Wolf Lake State Fish Hatchery (WLSFH), the Platte River State Fish Hatchery ( PRSFH), and the Thompson State Fish Hatchery (TSFH) from 2002 to 2010. Fingerlings are the progeny of broodstock spawned at the Little Manistee River Weir (LMRW) and the Swan River Weir (SRW). F - J: The low, medium, and high intensity levels of infection of R. salmoninarum in LMRW - CHS from WLSFH and PRSFH and SRW - CHS from the WLSFH, PRSFH, and TSFH from 2002 to 2010. The lines represent the logistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote t he observed prevalence and intensity levels of infection. 109 Figure 2.4. The prevalence of R. salmoninarum in the Hinchenbrooke strain (HB - COS) of coho salmon ( Oncorhynchus kisutch ) pre - stocking fingerlings and the Michigan - adapted strain of coho salmon (M I - COS) propagated at the Platte River State Fish Hatchery (PRSFH) from 2003 to 2010. C - D: The low, medium, and high intensity levels of infection of R. salmoninarum in HB - COS and MI - COS propagated at the PRSFH from 2003 to 2010. The lines represent the lo gistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote the observed prevalence and intensity levels of infection. 110 Figure 2.5. The prevalence of R. salmoninarum in steelhead ( Oncorhynch us mykiss ) pre - stocking fingerlings (LMRW - STT) propagated at the Wolf Lake State Fish Hatchery (WLSFH) and the Thompson State Fish Hatchery (TSFH) from 2005 to 2010. C - D: The low, medium, and high intensity levels of infection of R. salmoninarum in LMRW - S TT propagated at the WLSFH and TSFH from 2005 to 2010. The lines represent the logistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote the observed prevalence and intensity levels of infection. 111 Figure 2.6. The prevalence of male and female fish that may be shedding R. salmoninarum and are not shedding R. salmoninarum for (A) Chinook salmon ( Oncorhynchus tshawytscha ) broodstock from the Little Manistee River Weir; (B) Chinook salmon bro odstock from the Swan River Weir; (C) Hinchenbrooke coho salmon ( O. kisutch ) broodstock from the Platte River Weir; (D) Michigan - adapted coho salmon broodstock from the Platte River Weir; (F) and steelhead ( O. mykiss ) broodstock from the Little Manistee Ri ver Weir. The lines represent the logistic regression predicted prevalence and intensity levels of infection, while open circles and triangles denote the observed prevalence and intensity levels of infection. 112 REFERENCES 113 R EFERENCES Agresti, A., 2007. An Introduction to Categorical Data Analysis, second edition. John Wiley & Sons, Inc., Hoboken, New Jersey. Allison, L.N. 1958. Multiple sulfa therapy of kidney disease among brook trout. Prog. Fish - Cult. 20: 66 - 68. American Fisheries Society - Fish Health Section, 2010. FHS blue book: suggested procedures for the detection and identification of certain finfish and shellfish pathogens, 2010 edition. AFS - FHS, Bethesda, MA. Beacham, T.D., and Evelyn, T.P.T. 1992. Genetic variation in disease resistance and growth of Chinook, coho, and chum salmon with respect to vibriosis, furunculosis, and bacterial kidney disease. Trans. Am. Fish. Soc. 121: 456 - 485. Beyerle, J., and Hnath, J.G. 2002. History of fish health inspections: state of Michigan 1970 - 1999. Michigan Department of Natural Resources Technical Report 2002 - 2. Lansing, Michigan. Bullock, G.L., Conroy, D., and Snieszko, S.F. 1971. Bacterial diseases of fishes: Corynebacterial diseases. In: Snieszko, S.F., Axelrod, H.R ., (Eds.), Diseases of Fishes. T.F.H. Publications, Inc., Neptune, New Jersey, pp. 94 - 104. Burnham, K.P., and Anderson, D.R. 2002. Model selection and multi - model inference: a practical information - theoretic approach, second edition. Springer Science + Bu siness Media, LLC, New York. Danner, G.R., and Merrill, P. 2006. Disinfectants, disinfection, and biosecurity in aquaculture. In: Eradication of Aquatic Animal Disease. Blackwell Publishing, Ames, Iowa, pp. 91 - 128. higan Department of Natural Resources, Fisheries Special Report 32, Ann Arbor, Michigan. Eissa, A.E., Elsayed, E.E., McDonald, R., and Faisal, M. 2006. First record of Renibacterium salmoninarum in the sea lamprey ( Petromyzon marinus ). J. Wildl. Dis. 42: 3, 556 - 560. Elliott, D.G., Pascho, R.J., and Palmisano, A.N. 1995. Brood stock segregation for the control of bacterial kidney disease can affect mortality of progeny Chinook salmon ( Oncorhynchus tshawytscha ) in seawater. Aquaculture 132: 133 - 144. 114 Evelyn , T.P.T., Ketcheson, J.E., and Prosperi - Porta, L. 1986a. Use of erythromycin as a means of preventing vertical transmission of Renibacterium salmoninarum . Dis. Aquat. Org. 2: 7 - 11. Evelyn, T.P.T., Prosperi - Porta, L., and Ketcheson, J.E. 1986b. Experimental intra - ovum infection of salmonid eggs with Renibacterium salmoninarum and vertical transmission of the pathogen with such eggs despite their treatment with erythromycin. Dis. Aquat. Org. 1: 197 - 202. Faisal, M., and Eissa, A.E. 2009. Diagnostic testing patterns of Renibacterium salmoninarum in spawning salmonid stocks in Michigan. J. Wildl. Dis. 45(2): 447 - 456. Fa isal, M., Loch, T.P., Brenden, T.O., Eissa, A.E., Ebener, M.P., Wright, G.M., and Jones, M.L. 2010. Assessment of Renibacterium salmoninarum infections in four lake whitefish ( Coregonus clupeaformis ) stocks from northern Lakes Huron and Michigan. J. Great Lakes Res. 36(sp1): 29 - 37. Fryer, J.L., and Sanders, J.E. 1981. Bacterial kidney disease of salmonid fish. Ann. Rev. Microbiol. 35: 273 - 298. Gudmundsdottir, S., Benediktsdottir, E., and Helgason, S. 1993. Detection of Renibacterium salmoninarum in salmon id kidney samples: a comparison of results using double - sandwich ELISA and isolation on selective medium. J. Fish Dis. 16: 185 - 195. Hansen, M.J., and Holey, M.E. 2002. Ecological factors affecting the sustainability of Chinook and coho salmon populations in the Great Lakes, especially Lake Michigan. In: Lynch, K.D., Jones, M.L., Taylor, W.W., (Eds.), Sustaining North American Salmon: Perspectives Across Regions and Disciplines, American Fisheries Society, Bethesda, Maryland, pp. 155 - 180. Hnath, J.G. 1993. Great Lakes fish disease control policy and model program (super cedes September 1985 edition). Great Lakes Fish Commission Special Publication. 93 - 1, 1 - 38. Holey, M.E., Elliot, R.F., Marcquenski, S.V., Hnath, J.G., and Smith, K.D., 1998. Chinook salmon epizootics in Lake Michigan: possible contributing factors and management implications. J. Aquat. Anim. Health 10: 202 - 210. Hsu, H - M., Bowser, P.R., and Schachte, Jr., J.H. 1991. Development and evaluation of a monoclonal - antibody - based enzyme - linked immu nosorbent assay for the diagnosis of Renibacterium salmoninarum infection. J. Aquat. Anim. Health 3: 168 - 175. Jansson, E., Torbjörn, H., Höglund, J., and Ljungberg, O. 1996. Comparative evaluation of bacterial culture and two ELISA techniques for the dete ction of Renibacterium salmoninarum antigens in salmonid kidney tissues. Dis. Aquat. Organ. 27: 197 - 206. Keller, M., Smith, K.D., and Rybicki, R.W. 1990. Review of salmon and trout management in Lake Michigan. Fisheries Special Report No. 14. Michigan Dep artment of Natural Resources. Lansing, Michigan. 115 Lee, E. G - H., and Evelyn, T.P.T. 1994. Prevention of vertical transmission of the bacterial kidney disease agent Renibacterium salmoninarum by broodstock injection with erythromycin. Dis. Aquat. Org. 18: 1 - 4 . Ma cLean, D.G., and Yoder, W.G. 19 7 0 . Kidney disease among Michigan salmon in 1967. Prog. Fish - Cult. 32: 26 - 30. Meyers, T.R., Short, S., Farrington, C., Lipson, K., Geiger, H.J., and Gates, R. 1993. Establishment of a negative - positive threshold optical density value for the enzyme - linked immunosorbent assay (ELISA) to detect soluble antigen of Renibacterium salmoninarum in Alaskan Pacific salmon. Dis. Aquat. Org. 16(3): 191 - 197. Mitchum, D.L., Sherman, L.E., and Baxter, G.T. 1979. Bacterial kidney dise ase in feral populations of brook trout ( Salvelinus fontinalis ), brown trout ( Salmo trutta ), and rainbow trout ( Salmo gairdneri ). J. Fish. Res. Board Can. 36: 1370 - 1376. Mitchum, D.L., and Sherman, L.E. 1981. Transmission of bacterial kidney disease from wild to stocked hatchery trout. Can. J. Fish. Aquat. Sci. 38: 547 - 551. Moffitt, C.M., and Bjornn, T.C. 1989. Protection of Chinook salmon smolts with oral doses of erythromycin against acute challenges of Renibacterium salmoninarum . J. Aquat. Anim. Health 1(3): 227 - 232. Munson, A.D., Elliott, D.G., and Johnson, K. 2010. Management of bacterial kidney disease in Chinook salmon hatcheries based on broodstock testing by enzyme - linked immunosorbent assay: a multiyear study. N. Am. J. Fish. Manage. 30: 940 - 955 . Nuhfer, A. 2006. Evaluation of brown trout and steelhead competitive interactions in Hunt Creek, Michigan. Michigan Department of Natural Resources Study Performance Report #230654: October 1, 2005 to September 30, 2006. Lansing, Michigan. Office Inter national des Epizooties, 2012. Manual of diagnostic tests for aquatic animals, 5 th edition. Office International des Epizooties, Paris. Olea, I., Bruno, D.W., and Hastings, T.S. 1993. Detection of Renibacterium salmoninarum in naturally infected Atlantic salmon, Salmo salar L., and rainbow trout, Oncorhynchus mykiss (Walbaum) using an enzyme - linked immunosorbent assay. Aquaculture 116: 99 - 110. Ossiander, F.J., and Wedemeyer, G. 1973. Computer program for sample sizes required to determine disease incidenc e in fish populations. J. Fish. Res. Board Can. 30: 1383 - 1384. Pascho, R.J., and Mulcahy, D. 1987. Enzyme - linked immunosorbent assay for a soluble antigen of Renibacterium salmoninarum , the causative agent of salmonid bacterial kidney disease. Can. J. Fis h. Aquat. Sci. 44: 183 - 191. 116 Pascho, R.J., Elliott, D.G., and Streufert, J.M. 1991. Brood stock segregation of spring Chinook salmon Oncorhynchus tshawytscha by use of the enzyme - linked immunosorbent assay (ELISA) and the fluorescent antibody technique (FAT ) affects the prevalence and levels of Renibacterium salmoninarum infection in progeny. Dis. Aquat. Org. 12: 25 - 40. Pascho, R.J., Chase, D., and McKibben, C.L. 1998. Comparison of the membrane - filtration fluorescent antibody test, the enzyme - linked immuno sorbent assay, and the polymerase chain reaction to detect Renibacterium salmoninarum in salmonid ovarian fluid. J. Vet. Diagn. Invest. 10: 60 - 66. Purcell, M.K., Murray, A.L., Elz, A., Park, L.K., Marcquenski, S.V., Winton, J.R., Alcorn, S.W., Pascho, R.J ., and Elliott, D.G . 2008. Decreased mortality of Lake Michigan Chinook salmon after bacterial kidney disease challenge: evidence for pathogen - driven selection? J. Aquat. Anim. Health 20(4): 225 - 35. Sakai, M., Atsuta, S., and Kobayashi, M. 1991. Susceptib ility of five salmonid fishes to Renibacterium salmoninarum . Fish Pathol. 26: 159 - 160. Starliper, C.E., Smith, D.R., and Shatzer, T. 1997. Virulence of Renibacterium salmoninarum to salmonids. J. Aquat. Anim. Health 9(1): 1 - 7. Stoffregen, D.A., Bowser, P.R., and Babish, J.G. 1996. Antibacterial chemotherapeutants for finfish aquaculture: a synopsis of laboratory and field efficacy and safety studies. J. Aquat. Anim. Health 8(3): 1 81 - 207. Suzumoto, B.K., Schreck, C.B., and McIntyre, J.D. 1977. Relative resistances of three transferrin genotypes of coho salmon ( Oncorhynchus kisutch ) and their hematological responses to bacterial kidney disease. J. Fish. Res. Board Can. 34(1): 1 - 8. Tanner, H.A., and Tody, W.H. 2002. History of the Great Lakes salmon fishery: a Michigan perspective. In: Lynch, K.D., Jones, M.L., Taylor, W.W., (Eds.), Sustaining North American Salmon: Perspectives Across Regions and Disciplines, American Fisheries Soci ety, Bethesda, Maryland, pp. 139 - 154. Winter, G.W., Schreck, C.B., and McIntyre, J.D. 1980. Resistance of different stocks and transferrin genotypes of coho salmon, Oncorhynchus kisutch , and steelhead trout, Salmo gairdneri , to bacterial kidney disease an d vibriosis. Fish. Bull. 77(4): 795 - 802. Withler, R.E., Evelyn, T.P.T., 1990. Genetic variation in resistance to bacterial kidney disease within and between two strains of coho salmon from British Columbia. Trans. Am. Fish. Soc. 119, 1003 - 1009. 117 Chapt er 3 The Use of Non - lethal Samples for the Detection of Renibacterium salmoninarum in Chinook salmon, Oncorhynchus tshawytscha (Walbaum ) 118 1. Abstract Bacterial kidney disease, caused by Renibacterium salmoninarum , threatens salmonid populations in the Northern hemisphere. Most fishery regulatory authorities require continuous monitoring of the disease in hatcheries and spawning runs. As per the diagnostic protocols of the World Organisation for Animal Health and the American Fisheri es Society - Fish Health Section, lethal sampling of visceral organs is used from a set number of fish, which depends on the assumed disease prevalence. Non - lethal sampling would be a preferable alternative, especially in the case of valuable broodstock and endangered species. In this study, non - lethal sampling methods were evaluated for their ability to detect R. salmoninarum in experimentally infected Chinook salmon ( Oncorhynchus tshawytscha ). Non - lethal (e.g., blood, mucus, and a urine/feces mixture) an d lethal (e.g., a kidney and spleen homogenate) samples were collected from 178 Chinook salmon that were experimentally infected with R. salmoninarum . Renibacterium salmoninarum was detected in all samples by culture on modified kidney disease medium, nes ted polymerase chain reaction (nPCR), and quantitative enzyme - linked immunosorbent assay (ELISA) , although detection depended on sampling period . The sensitivity, specificity, and accuracy of the lethal and non - lethal samples in detecting the presence of R. salmoninarum were calculated using receiver operator characteristic (ROC) analyses using the assumption that all infected fish were positive for R. salmoninarum . Non - lethal samples did detect R. salmoninarum ; however, the level of sensitivity and accur acy depended upon the exposure route and the subsequent disease course. ROC analyses revealed that the uro - fecal sample has the greatest potential for non - 119 lethal sampling compared to mucus and blood samples. Also, combining the nPCR and ELISA data from t he lethal samples with the uro - fecal samples has the potential to be the best strategy for detecting R. salmoninarum prevalence and intensity in a population . 2. Introduction Bacterial pathogens frequently threaten t he health of both wild and aqua cul tured fishes worldwide (Meyer, 1991; reviewed in Austi n and Austin, 2007). T o prevent the spread of bacterial infections in aquaculture facilities or for stocking purposes in public waters, health inspections are required to effectively manage disease out breaks within infected fish populations ( World Organisation for Animal Health (WOAH ), 2013). The current methods for detecting many bacterial fish pathogens, as outlined by the American Fisheries Society - Fish Health Section (AFS - FHS) Blue Book (2012) and the Manual of Diagnostic Tests for Aquatic Animals ( WOAH , 2012), require the use of lethal sampling techniques, and therefore, the sacrifice of individual fish. In order to provide a 95% confidence of detecting a pathogen at an assumed minimum incidence o f 5%, at least 60 fish must be sampled lethally from the overall population, or with a minimum incidence of 10%, at least 30 fish (Ossiander and Wedemeyer, 1973; Hnath, 1993; AFS - FHS Blue Book, 2012). In some instances, lethal sampling of more than 60 fis h is required because it increases the likelihood of detecting the pathogen (Fenichel et al., 2008). Sacrificing this number of fish from a threatened or endangered species or from a broodstock population containing irreplaceable individuals with unique t raits is often impractical (Powell et al., 2005). This matter is complicated by the fact that results of each of 120 the diagnostic assays commonly used in BKD diagnosis depend primarily on the disease progression and course (Faisal and Eissa, 2009; Schulz, Ch apter 5). Several studies on fish pathogens have generated encouraging results regarding the use of non - lethal sampling in disease diagnosis. For example, Yersinia ruckeri , the causative agent of Enteric Redmouth Disease, has been detected from fecal mate rial and blood collected non - lethally (Rodgers, 1992; Altinok et al., 2001 - blood), and Aeromonas salmonicida , the etiological agent of furunculosis was recovered non - lethally from the gills, mucus, and blood of infected fish ( Benediktsd ó ttir and Helgason, 1990; Cipriano et al., 1996; Cipriano et al., 1997; Klinger et al., 2003). In addition, viral pathogens have also been detected in non - lethally collected tissues, such as Infectious Pancreatic Necrosis Virus from the pectoral fin (Bowers et al., 2008) and Infectious Salmon Anemia Virus from the blood of infected fish (Giray et al., 2005). Renibacterium salmoninarum , the causative agent of Bacterial Kidney Disease (BKD), is unique in terms of its pathogenesis. BKD can run an acute course with mortalities occurring, or more often, it can run a subclinical disease course with a low infection prevalence and intensity. The strong host granulomatous reaction against R. salmoninarum sequesters the bacteria, and unless the tissues containing the bacteria are co llected, the bacterium presence may be undetected (MacLean and Yoder, 19 7 0 ; DeKinkelin, 1974; Kimura and Awakura, 1977; Hoffmann et al., 1984; Sanders and Barros, 1986; Holey et al., 1998). Although both horizontal and vertical routes have been identified for R. salmoninarum transmission, details pertaining to the spread of the bacterium within the fish body remains largely unknown, despite genuine research efforts. This fact constitutes an obstacle for the proper diagnosis of the bacterium, particularly in latently infected fish. 121 To this end, this study was designed to 1) determine if R. salmoninarum could be detected in samples that are collected non - lethally (blood, mucus, and uro - fecal samples), 2) evaluate how exposure route and the resultant disease course impacts the detection of R. salmoninarum in all sample types, 3) compare the efficacy of non - lethally collected samples to those collected lethally (i.e., spleen and kidney samples), 4) determine the best combination of sample type and diagnostic m ethod to detect R. salmoninarum , and lastly 5) determine the number of fish that would be required for non - lethal testing, while still maintaining the efficacy of lethal testing, with an assumed disease prevalence of 5 and 10%. 3. Materials and methods 3 .1. Source of fish Five hundred Chinook salmon (~8 months old) were acquired from the Michigan Department of Natural Resources (MDNR) Wolf Lake State Fish Hatchery (Van Buren County) and transferred to the University Research Containment Facility at Michi gan State University for experimental challenges. A sub - sample of the fish (n=60) were sacrificed and screened for the presence of R. salmoninarum . All fish were negative for R. salmoninarum by the quantitative enzyme - linked immunosorbent assay (Q - ELISA) as outlined in Pascho and Mulcahy (1987), with minor modifications (see description below). Fish were fed to satiation twice a day with BioTrout 2.0 mm pellets (Bio - Oregon, Westbrook, Maine) and were held in a 600 L continuous 122 flow - through tank with chil led freshwater at an average temperature of 12 ± 2.0 ° C, until needed for experimental infections. 3.2. Challenge by immersion bath To mimic the natural route of infection, fish were exposed to R. salmoninarum via an immersion bath. Cryo - preserved R. s almoninarum (ATCC #33209) was revived on modified kidney disease medium (MKDM; Faisal et al., 2010), incubated at 15 ° C for 14 d, purity verified, and then a single colony was inoculated into a 7 ml aliquot of MKDM broth (x4) and incubated at 15 ° C for 7 d. Twenty µ l from each of the four broth cultures were then sub - cultured onto trypticase soy agar (TSA) and MKDM to verify purity, and the remaining broth was added to 1900 ml of fresh MKDM broth and incubated at 15 ° C on a Thermolyne Nuova stir plate (Thermo Fischer Scientific, Inc., Waltham, MA) at approximately 50 rpm in a 2 L Celstir Spinner Flask (Wheaton, Millville, NJ). After 14 days of incubation, the broth culture was centrifuged in a Hermle Z382K centrifuge (Labnet International, Inc., Woodbridge, N J) at 4300 rpm for 10 min, the supernatant was discarded, and the bacterial pellet was re - suspended in 0.85% saline solution, which was repeated a total of 3 times. The number of colony forming units per ml (CFU ml - 1 ) within the R.salmoninarum - 0.85% salin e suspension was then determined using 10 - fold serial dilutions and plate counts via drop culture in triplicate. One third of the R. salmoninarum suspension were then added to 3.5 L of autoclaved tank water supplemented with 0.85% NaCl (total final volume of 4 L) for a final concentration of ~2.3 x 10 7 CFU ml - 1 . Each of the three suspensions was placed into a separate, sterile 5 gallon bucket that was heavily 123 aerated, to which 42 Chinook salmon (12.6 g ± 4.4 g; 10.7 cm ± 1.5cm) were added. Each bucket wa s then covered and fish were immersion exposed for one hour. One bucket of 42 fish served as a negative control, with fish being immersed in the same fashion in a sterile saline suspension. Fish and the R. salmoninarum suspension were then poured into th eir empty respective 74 L tanks ( 12 ± 2.0 ° C ) and water for flow - through system (1.09 L min - 1 ) was resumed. All tanks were monitored daily for mortalities. 3.3. Challenge by intraperitoneal injection To generate an acute course of BKD, fish were intrap eritoneally (i.p.) injected with R. salmoninarum . Renibacterium salmoninarum was revived and cultured as described above, with the exception that the bacterial suspension was grown in 1 L of MKDM broth. After 14 days of incubation, the broth culture was centrifuged and washed as described above, and the remaining pellet was re - suspended into 101 ml of sterile saline, for a final concentration of 2.1 x 10 10 cfu ml - 1 . Fish (61 .1 g ± 4.8 g; 18.8 cm ± 0.6 cm ) were anesthetized with 100 mg L - 1 of sodium bicar bonate - buffered tricaine methanesulfonate (MS - 222; Argent Chemical Laboratories, Inc., Redmond, Washington) for approximately 10 - 15 sec, and were then i.p. injected with 200 µ l of the bacterial suspension and revived by placing them in freshwater in their respective 225 L tanks, with a flow rate of 1.62 L min - 1 and water temperature of 12 ± 2.0 °C . This was done in accordance with the Michigan State University International Animal Care and Use Committee (Animal Use Form # 02/10 - 013 - 00). Three replicates of 25 fish were 124 infected in this fashion. One tank of 25 fish served as a negative control, whereby fish were injected with 200 µ l of sterile saline. All tanks were monitored daily for mortalities. 3.4. Sampling procedures Fish challenged with R. salm oninarum via immersion were sampled every 21 days (n=21), culminating in a total of 6 sampling periods for a total of 115 fish sampled from July to October 2010. Fish that were challenged via i.p. injection with R. salmoninarum were sampled every 7 days ( n=15) over 5 sampling periods, for a total of 63 fish sampled during May and June 2011. Prior to non - lethal sample collection, fish were anesthetized with a dose of 100 mg L - 1 of sodium bicarbonate - buffered MS - 222 (Argent). Once anesthetized, mucus, bloo d, and a urine - - Mucus was collected by gently running an individual plastic cover slip in a posterior direction from the left pectoral fin to the caudal fin. Mucu s was then placed into a sterile 1.5 ml microcentrifuge tube and frozen at - 80°C until diagnostic assays were performed. Mucus was tested for the presence of R. salmoninarum by culture, nested polymerase chain reaction (nPCR), and Q - ELISA (see description s below). Blood samples were collected by venipuncture of the caudal vein using a sterile needle and syringe, which was subsequently stored in a 1.5 ml microcentrifuge tube at - 80°C until diagnostic assays were performed. Prior to cryopreservation, 5 µl o f blood was inoculated onto MKDM agar plates (see description below). Blood was tested for the presence of R. salmoninarum by culture and nPCR. 125 A urine - feces mixture (referred to as uro - fecal) was obtained by gently pressing the ventral abdomen in an ante rior to posterior direction and collecting the mixture in a sterile 1.5 ml microcentrifuge tube. Any urine - feces mixture that was expressed was collected in the microcentrifuge tube and stored at - 80°C until assayed. The uro - fecal samples were tested for R. salmoninarum by culture, nPCR, and Q - ELISA. After non - lethal sample collection was completed, fish were euthanized with a lethal dose of 250 mg L - 1 of sodium bicarbonate - buffered MS - 222 (Argent). A thorough internal and external examination for gross s igns of disease was performed. Kidney and spleen samples (referred to as kidney/spleen or lethal samples) were collected with individual sterile forceps and scissors, placed in individual 1.5 ml microcentrifuge tubes, and stored at - 80°C until diagnostic assays were performed. Lethal samples were tested for R. salmoninarum by culture, nPCR, and Q - ELISA. 3.5. Bacterial culture and isolation Renibacterium salmoninarum isolation was performed by streaking 1 µl of the target tissue onto MKDM agar plates us ing sterile disposable inoculating loops and incubating the plates under aerobic conditions at 15 ° C for a total of 42 days. Five microliters of blood was inoculated directly onto MKDM plates at the time of necropsy. The mucus, uro - fecal, and kidney/splee n samples were diluted 1:10 (weight/volume) in sterile phosphate buffered saline (PBS; pH 7.2) and suspended by repeated expulsion through a sterile pipette. Each tissue suspension was then diluted via 10 - fold serial dilutions and 10 µ l of each dilution (e.g., 126 undiluted to 10 - 8 ) was dispensed onto the MKDM plates. Plates were incubated at 15 ° C and were observed for typical R. salmoninarum growth every 7 days for a total of 6 weeks. Colonies were observed under a dissecting microscope for typical R. salm oninarum morphological characteristics: convex, cream - colored, round, and smooth (Austin and Austin, 2007). Colonies that fit this criterion were assessed by additional biochemical testing, which included Gram stain, cytochrome oxidase, and catalase react ion, which are key biochemical tests used to identify R. salmoninarum (Sanders and Fryer, 1980). Colonies that were Gram positive, cytochrome oxidase negative, and catalase positive were then confirmed molecularly via nPCR (see below). 3.6. Extraction o f DNA The DNA from blood and mucus samples was extracted using a DNeasy® Blood and Tissue Kit (Qiagen, Inc., Valencia, CA) according to the manufacturer instructions for nucleated blood samples. The DNA from kidney/spleen samples were also extracted usi ng the DNeasy® Blood and Tissue Kit (Qiagen), but the protocol for animal tissue was followed. DNA from the uro - fecal samples was extracted using the QIAamp® DNA Stool Mini Kit (Qiagen), according to the protocol of the manufacturers. In all cases, the pr e - treatment for gram - positive bacteria (including the lysis buffer) was applied to all of the sample types. After extraction, the DNA was quantified with the Qubit ® Fluorometer (Life Technologies, Grand Island, NY) and then diluted to a 20 ng µl - 1 concent ration. 127 3.7. Nested PCR The nPCR method and primers recommended by Pascho et al. (1998) were used initially, annealing temperature was changed to 60 ° C for DNA extr acted from pure bacterial cultures and kidney/spleen samples (based upon optimization experiments). Additionally, the total reaction volume was reduced from 50 µ l to 25 µ l, consisting of 1 µ l each of template DNA (20 ng total), forward primer (10 µmol), a nd reverse primer (10 µmol), as well as 12.5 µ l of GoTaq® Green Master Mix (Promega Corp., Madison, WI) and 9.5 µ l of nuclease - free water. The controls were composed of a PCR mixture containing water instead of DNA template (negative control) and DNA from a pure culture of R. salmoninarum ATCC #33209 strain (positive control). Five microliters of the nPCR products and controls were mixed with 1 µ l of SYBR Green II RNA Gel Stain (Cambrex Bio Science Rockland, Inc., East Rutherford, NJ) and loaded into a ge l consisting of 2% Ultra Pure TM agarose (Invitrogen, Grand Island, NY). Each electrophoresis gel included 5 µ l of a 1 Kb plus ladder (Invitrogen) mixed with 1 µ l of 6X gel loading dye (New England Bio Labs, Inc., Ipswich, MA). Gels were run at 100 v for 35 minutes in 1X Tris - Acetate Buffer (Sigma - Aldrich Corp., St. Louis, MO) in a Gel XL Ultra V - 2 electrophoresis box (Labnet) and a Sub - Cell GT electrophoresis box (Bio Rad Laboratories, Inc., Hercules, CA). Gels were visualized with a Canon G10 camera and UV Trans - Illuminator. Samples were considered positive for R. salmoninarum when a 320 - bp band was present (Pascho et al. 1998). Prior to conducting nPCR on the blood, mucus, and uro - fecal samples, representative samples that were known to be R. salmonin arum - positive were first analyzed to assess the 128 suitability of the protocol described above. Subsequently, it was determined that DNA extracted from the different tissue types had different optimal annealing temperatures. While the nPCR for the kidney/sp leen samples was performed as described above, the annealing temperature for the blood samples was modified to 58.5 ° C, the mucus samples were modified to 59 ° C, and the uro - fecal samples were modified to 56 ° C. 3.8. Q - ELISA The general Q - ELISA protocol outlined in Pascho and Mulcahy (1987), with modifications recommended by Gudmundsdóttir et al. (1993) and Olea et al. (1993), was used to assess R. salmoninarum antigens in the mucus, uro - fecal, and kidney/spleen samples. Due to the small amount of tissue available from each fish, its need to be used in several diagnostic assays, and a minimum volume of diluted tissue required for each assay, different dilution factors were determined for the various sample types, as described below. Prior to the Q - ELISA procedure, Sigma) and stomached on high speed for two minutes with the Biomaster Stomacher (Wolf Laboratories Limited, Pocklington, York, UK). An aliquot of 250 µl of stomached kidney/spleen samples were dispensed into 1.5 ml microcentrifuge tubes containing 250 µl of phosphate buffered saline with Tween - 20 (PBS - T20; Sigma) and 5% goat serum (Sigma) and 50 µl of CitriSolv (Fisher Scientific, Pittsburgh, PA). The purpose of the CitriSolv solvent is to dissolve and remove lipids from the aqueous supernatant (Gudmundsdóttir et al., 1993), while the introduction of 5% goat serum increases the sensitivity of the assay (Olea et al., 1993). Samples 129 were vortexed for appr oximately 10 sec, heated at 100°C for 15 min, and then centrifuged at 14,000 rpm for 10 min. The aqueous supernatant of each sample was used for Q - ELISA testing. Mucus samples were diluted 1:4 (w:v) with sterile HBSS and were vortexed thoroughly to ensur e a homogeneous mixture. The mucus samples were then aliquoted (250 µl) into 1.5 ml microcentrifuge tubes containing 250 µl of PBS - T20 with 5% goat serum and 50 µl of CitriSolv and processed as described above. The uro - fecal samples were diluted 1:100 (w :v) with sterile HBSS, vortexed to homogenize the mixture, and then subsequently centrifuged at 14,000 rpm for 5 min to pellet any inhibiting factors. The supernatant from the uro - fecal samples were then aliquoted (250 µl) into 1.5 ml microcentrifuge tube s containing 250 µl of PBS - T20 with 5% goat serum and 50 µl of CitriSolv and processed as described above. The positive - negative cutoff absorbance for the samples was 0.10. Samples that tested positive were assigned the following antigen levels: low (0. 10 - 0.199), medium (0.20 - 0.999), and included two negative controls, a negative fish tissue sample and a dilution buffer; and two positive controls, a positive fish ti ssue sample and a R. salmoninarum positive control (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD). 3.9. Statistical analyses Receiver operating characteristic (ROC) analyses were used to assess the analytical sensitivity, specificity, and a ccuracy of the non - lethal samples to predict the presence of R. salmoninarum . ROC analyses compared the results of the non - 130 (TP), and true negatives (TN), which were then used to determine sensitivity, specificity and accuracy (Zhu et al., 2010). All calculations were performed in SAS 9.3 (SAS Institute, Inc., Cary, NC). In both challenges, fish were exposed to or injected with virule nt R. salmoninarum ; positive. Sensitivity was considered the proportion of true positives that were correctly identified by the non - lethal samples and was calcula ted by dividing the number of true positive assessments by all of the positive assessments: Sensitivity is often an indicator of how capable a test is at correctly detecting a pathogen; therefore, tests with high s ensitivity capture most of the positive outcomes and are usually ideal to screen for a pathogen (Zhu et al., 2010). The specificity was also calculated, which was the proportion of true negatives that were correctly identified by the non - lethal samples, and was calculated by dividing the number of true negative assessments by all of the negative assessments: Specificity is an indicator of how well a diagnostic test identifies the normal, non - diseased conditions; therefore, tests with high specificity imply that a test is highly reliable for identifying true negatives. 131 Accuracy was the proportion of true results (positive and negative) that were correctly identified by the non - lethal samples and was calculated by dividing the number of correct assessment by all of the assessments: Accuracy is a measure of the reliability of a diagnostic test, combining the sensitivity and specificity to demonstrate how well it identifies the true positives and negatives. However, a test with high sensitivity and specificity does no t always have a high accuracy rate, especially when diagnosing rare diseases (Zhu et al., 2010). In addition to sensitivity, specificity, and accuracy, the area under the curve (AUC) metric was also determined. Area under the curve values are typically ca lculated by plotting the true positive rate against the false positive rate for different cut - off points of the diagnostic test; however, in binary tests, where there is only one cut - off point (positive or negative), the AUC is calculated by adding the sen sitivity and specificity together and dividing by two (Cantor and Kattan, 2000). In this study, the AUC measures the ability of the non - lethal samples to correctly identify individuals with and without disease. Area under the curve values that are 1.0 im ply that the diagnostic test has perfect discrimination (i.e., always correctly identifies true discrimination, 0.9 - - no better than chance alone (Hosmer and Lemeshow, 2000; Lalkhen and McCluskey, 2008; Zhu et al., 2010). The accuracy, sensitivity, specificity and AUC values were calculated for the lethal and non - lethal samples from the imme rsion challenge, the injection challenge, and also for the two 132 challenges combined together. Additionally, for the immersion and injection challenge, the accuracy, sensitivity, specificity, and AUC values were determined for each particular non - lethal tis sue overall, and during the various stages of disease progression. An individual sample was considered positive for R. salmoninarum if it was positive by at least one of the diagnostic tests (e.g., culture, ELISA, and/or nPCR). Using the assumption tha t a naturally infected population would be comprised of fish in various stages of disease progression (chronic and sub - acute), the number of non - lethal samples that would be necessary to collect from a population of wild fish to detect BKD was estimated. The immersion and i.p. injection challenge data was used to create a simulated population that would be randomly composed of individuals from both challenges, but with specified numbers of both disease courses (e.g., 100% sub - acute, or 40% sub - acute and 6 0% chronic). To create each simulated population, stratified random sampling was used, using PROC SURVEYSELECT in SAS to randomly draw a specified percentage of individuals from each exposure regime, while maintaining a consistent representation across da ys of exposure. The simulated population was then used to calculate sensitivity, specificity and accuracy values. The simulated populations were designed to represent the range of possible scenarios of sub - acute and chronic composition in a populatio n. Analysis began with 100% sub - acute, and then incrementally decreased the sub - acute cases by 10% while simultaneously incrementally increasing the chronic cases by 10% until 100% chronic was reached. For each scenario, experimental data was resampled u sing the stratified random sampling technique 50 times and sensitivity, specificity and accuracy values were calculated each time. The average sensitivity, specificity, and accuracy for 50 iterations for each scenario was used, and FreeCalc 133 2.0 (AusVet An imal Health Services, Lyon, France) estimated the sample size for a population 10,000 fish, with type I and II error levels set at 0.05 and a minimum estimated prevalence (MEP) of 5 and 10%. An MEP of 5 and 10% was chosen based on data reflecting the curr ent and historical prevalence of R. salmoninarum in fish populations in Michigan (Schulz, Chapter 2). FreeCalc 2.0 calculates the number of individuals that must be tested in order to provide evidence, at a specified level of confidence, that disease is n ot present (Cameron and Baldock, 1998; Cameron, 2001). Therefore, for this study, the number of non - lethal samples required for testing to obtain the same 95% level of confidence that lethal samples afford was estimated. 4. Results 4.1. Detection in n on - lethally collected samples Renibacterium salmoninarum was detected in both the lethal and non - lethal samples recovered from fish that were exposed to R. salmoninarum by the immersion and i.p. injection methods (Table 3.1). The uro - fecal samples had t he highest prevalence of R. salmoninarum in both of the challenges (Table 3.1). In the injection challenge, there was a greater prevalence of R. salmoninarum in the blood samples compared to the mucus samples; conversely, the mucus samples had a slightly higher detection of R. salmoninarum than the blood samples in the immersion challenge (Table 3.1). 134 4.2. Effect of exposure route In the injection challenge, the fish developed a sub - acute course of BKD, with some of the challenged fish surviving through the end of the observation period. As early as day 1 post - infection (p.i.), characteristic clinical signs of BKD were observed (Figure 3.1A - E). Disease signs included diffuse petechial hemorrhages covering the body of the fish (Figure 3.1A), severe gill pallor (Figure 3.1B), corneal opacity, frequently associated with ocular hemorrhage (Figure 3.1C), whitish, false membranes covering the liver (Figure 3.1D), and whitish, granulomatous - like lesions in the spleen and kidney (Figure 3.1E). Differences were noted throughout the study period, with fewer clinical signs being observed during early sampling periods. For example, on days 1 and 8 p.i., ocular hemorrhage and few external petechial hemorrhages were noted, along with mild heart and liver pallor, but only in 13 - 27% of the fish. However, on day 15 p.i., the formation of ascites was first observed, as well as heart and liver pallor, white granulomatous - like lesions in the kidney, and whitish, false membranes covering the spleen in 73% of the fish. On days 22 and 29 p.i., gill pallor, petechial hemorrhages throughout the external body, exophthalmia, ascites, liver and heart pallor, white granulomatous - like lesions in the spleen and kidney, and false membranes covering the liver were observed in 100% of the fish. Outside of the fish euthanized for regular sampling, mortalities started occurring as early as day 9 p.i. with a total of 7 mortalities occurring throughout the 29 - day study period. Mortalities also exhibited the external and internal clinical signs described above. There were no clinical signs observed in the control (i.e., non - infected) fish. 135 At each sampling period in the injection challenge, R. salmoninarum was detected in all of the individual lethal samples, for a prevalence of 100%. H owever, depending on the non - lethal tissue, there was more variation in the presence and detection of R. salmoninarum at the sampling periods (Table 3.1). The uro - fecal samples were the most consistent non - lethal sample for R. salmoninarum detection, alth ough there was a decrease in the prevalence on day 29 p.i. (Table 3.1). The blood samples regularly detected R. salmoninarum throughout most of the study at a prevalence > 67% and similar to the uro - fecal samples, there was also a decline in prevalence on day 29 p.i. (Table 3.1). Detection of R. salmoninarum in the mucus samples occurred at a more gradual rate than the other samples; however, at day 15 p.i., the prevalence increased substantially and remained consistent (Table 3.1). Based on clinical sign s and laboratory findings, the immersion - challenged fish exhibited an extremely mild course of a chronic, or sub - chronic, nature. On the contrary of the injection - challenged fish, characteristic clinical signs of BKD were seldom seen (Figure 3.2A - B). When present, disease signs included exophthalmia, sometimes associated with ocular hemorrhage (Figure 3.2A), diffuse petechial hemorrhages in the caudal peduncle region, occasionally associated with hemorrhages between the fin rays (Figure 3.2B), swollen sple ens, generalized visceral edema, darkened kidneys, and swollen and pale livers. Additionally, there was one mortality that occurred on day 33 p.i. throughout the 105 - day study period. The mortality exhibited the external and internal clinical signs descr ibed above. There were no clinical signs observed in the control (i.e., non - infected) fish. Consequently, R. salmoninarum was detected at a lower prevalence in the lethal and non - lethal samples collected in the immersion challenge compared to the inject ion challenge 136 (Table 3.1). While the detection of R. salmoninarum progressively increased in the lethal samples at several of the sampling periods in the immersion challenge, there was more inconsistency in the detection of R. salmoninarum in the non - leth al samples (Table 3.1). Renibacterium salmoninarum was detected in the uro - fecal samples on days 1 and 63 p.i., in the mucus samples on day 21 p.i., and in the blood samples on days 1 and 84 p.i. (Table 3.1). However, it should be noted that R. salmonina rum was observed in the uro - fecal samples at much higher prevalences than it was detected in the blood and mucus samples. 4.3. Comparison of R. salmoninarum detection in all samples In the injection - challenged fish, R. salmoninarum was detected at a p revalence of 100% in the lethal and the combined non - lethal samples, resulting in an AUC value of 1.00 for each, indicating that in a sub - acute infection, the collective non - lethal samples are equally effective as the lethal samples when detecting R. salmo ninarum (Table 3.2). Individually, the uro - fecal absence) of R. salmoninarum (Table 3.2). T he respective sensitivity and accuracy for the mucus samples were low Similar to the injection challenge, the uro - fecal samples from the immersion challenge had the best success at detecting BKD, with the highest AUC value of all the non - lethal sa mples (Table 3.2). I nfections in blood and mucus samples were detected at very low prevalences, yielding very low sensitivities and poor AUC values (Table 3.2). Conversely, R. salmoninarum 137 was detected more frequently in the lethal samples and resulted i abilities (Table 3.2). In a natural infection, it is more likely that fish will exhibit multiple stages of infection and a population will be comprised of fish in various stages of disease progression (Faisal and Eissa, 2009; Nance et al., 2010). Therefore, results from the injection and immersion challenge were combined to represent a population equally comprised of both disease stages (sub - acute and chronic). Based on these results, the sensitivities of the non - lethal sampl es (combined and individual) were distinctly lower the sensitivity of the lethal sample, suggesting that the non - lethal samples are not as effective at identifying positive samples as the lethal samples (Table 3.2). Due to the lower sensitivities, the acc uracy of each tissue type was also lower than would be desired; however, the AUC value for the uro - fecal sample was in the middle of the The specificity for all sample types was 1.00, implying that there were no false posi tives and that each sample type correctly identified the true negative samples 100% of the time. However, this is an artifact of the assumption that all infected fish were positive; therefore, specificity was chosen to not be reported in this study. 4.4 . Comparison of diagnostic assays among sample types Interestingly, there were clear differences in the prevalence, sensitivity, accuracy, and AUC values of the various samples when analyzed by the assay that was used (ELISA, nPCR, or culture). The preva lence of R. salmoninarum in the lethal samples was the highest when tested 138 by ELISA, followed by nPCR, and culture (Table 3.3). Furthermore, the lethal samples were in discrimination (Table 3.3). When all of the non - lethal samples were combined together, samples tested by nPCR yielded the highest prevalence; however, culture, ELISA, and nPCR all yielded identical accuracy, sensitivity, and AUC values (Table 3.3). A dditionally, the AUC values correctly identify diseased and non - diseased individuals (Table 3.3). Similarly, a comparable AUC value can also be achieved if only t he uro - fecal sample is collected and tested by nPCR (Table 3.3), which eliminates the need for the blood and mucus sample to also be analyzed. Interestingly, ELISA resulted in the highest accuracy, sensitivity, and AUC values for the blood and mucus sampl es (Table 3.3). Based on the encouraging detection of R. salmoninarum in the uro - fecal samples in this study, when tested by all assays (culture, ELISA, and nPCR) and by nPCR alone, this sample was selected for further analysis to evaluate its potential use in conjunction with the traditional lethal sample. Specifically, when nPCR results from the uro - fecal sample and ELISA results from the lethal sample were combined, a relatively high level of agreement was observed (Table 3.4). In the injection chall enge, R. salmoninarum was detected concurrently in both sample types in most of the fish (Table 3.4). In the immersion challenge, there was a greater disagreement between the lethal and uro - fecal sample, with a greater proportion of R. salmoninarum being detected in the lethal sample only (Table 3.4). Moreover, if all of the R. salmoninarum - positive lethal and uro - fecal samples were combined (L+/UF - , L - /UF+, L+/UF+), the overall detection of 139 R. salmoninarum increased substantially in the immersion challen ge, from 69% in the lethal samples and 33% in the non - lethal samples to 81% overall (Table 3.4). 4.5. Population disease composition and estimated sampling size While it is likely that fish will exhibit several disease stages in a naturally infected fis h population, it is unlikely that the population will be comprised of equally distributed chronic and sub - acute infections. Therefore, when the sensitivity, accuracy, and AUC values were calculated for populations composed of varying infection levels, the results were more diverse (Table 3.5). If non - lethal samples were collected from a population of fish that was comprised of entirely sub - acute infections, the non - lethal sample s would be highly sensitive and accurate, with perfect discrimination (Table 3 .5). However, as the percentage of fish suffering from a chronic disease course progressive ly increases, the sensitivity , accuracy, and discriminatory abilities of the non - lethal samples decreases (Table 3.5). Once the population is comprised of 90% chro nic infections, the discrimination of the non - lethal samples is no longer acceptable (Table 3.5). Similarly, if lethal samples were collected from a population of fish comprised of sub - acute infections, they would also be highly sensitive and accurate, wi th perfect discrimination (Table 3.5). However, as the percentage of fish suffering from a chronic disease course increases, the accuracy and sensitivity decline, but the discriminatory abilities are ely chronic infections (Table 3.5). Most interestingly, if a uro - fecal sample and a lethal sample were collected from a similar population, 140 and tested by nPCR and ELISA, respectively, the highest accuracy, sensitivity, and AUC values were observed regardl ess of the disease composition of the population (Table 3.5). T he number of samples (e.g., lethal, combined non - lethal, and a combined uro - fecal and lethal sample) that would need to be collected to determine if BKD was present in a wild population was also estimated. If a population consisted of only sub - acute infections, it would require collecting lethal, non - lethal, and combined uro - fecal and lethal samples from 59 fish and 29 fish, with a MEP of 5% and 10%, respectively (Table 3.6). Furthermore, t he number of samples required to be collected increased as the percentage of chronic infections increased, regardless of sample type or MEP (Table 3.6). However, noticeable differences were observed in the number of fish that would need to be sampled leth ally and those that would have an additional uro - fecal sample collected (Table 3.6). Specifically, even though an additional sample is required, fewer fish would need to be tested if a lethal and uro - fecal sample were analyzed together compared to only a lethal sample (Table 3.6). 5. Discussion Indeed, R. salmoninarum was detected in all of the non - lethally collected samples ; however, differences were observed among sample types and sampling periods . Renibacterium salmoninarum was detected most frequ ently from the urine/feces mixture, followed by the blood, and lastly, the mucus. This is in agreement with other studies that have documented the presence of R. salmoninarum in fish intestines and feces (Bullock et al., 1978; Bullock et al., 1980; Mitchu m and Sherman, 1981; Austin and Rayment, 1985; Balfry et al., 1996). Balfry et al. 141 (1996) detected R. salmoninarum in Chinook salmon feces and demonstrated fecal - oral transmission of the bacterium. It was suggested by Austin and Rayment (1985), who recov ered R. salmoninarum from the feces of rainbow trout, that R. salmoninarum may have an affinity for organic, particulate matter. In this study, in 14 of the 16 fish where R. salmoninarum was concurrently isolated from both the uro - fecal sample and the kid ney/spleen homogenate, the CFU g - 1 of tissue was greater in the uro - fecal sample. In addition, there were seven samples where R. salmoninarum was recovered from the uro - fecal samples, but not the kidney/spleen homogenate. Interestingly, Mitchum and Sherm an (1981) also found that R. salmoninarum was more readily detectable in the feces of several salmonids species compared to the kidneys. Some researchers have attributed the presence of R. salmoninarum in the feces of infected salmonids to the accumulatio n of macrophages containing R. salmoninarum in the gut associated lymphoid tissue (GALT) in the intestine. Renibacterium salmoninarum has a predilection for the macrophages of the fish host, and has been shown to live and replicate within them (Gutenberge r et al., 1997; Wiens and Kaattari, 1999). Therefore, the accumulation of macrophages containing R. salmoninarum in the GALT is a possible explanation for the high detection rate of R. salmoninarum in the uro - fecal samples. Also, since R. salmoninarum exi sts in the kidneys, some bacteria can travel in the urine outside of the fish host. In other words, the high prevalence of R. salmoninarum in uro - fecal samples is likely due to the shedding of the bacterium in both the digestive and urinary systems. Taki ng into consideration that R. salmoninarum was detected in all of the uro - fecal samples on the first sampling period of the immersion challenge, yet there was a considerably lower prevalence in later sampling periods, it is also possible that R. salmoninar um utilizes the 142 anal opening as a portal of entry, which could have resulted in the higher initial prevalence of the bacterium. Also, once an infection is established, increased rates of R. salmoninarum being shed with the feces could lead to increased ho rizontal transmission and result in high prevalences of R. salmoninarum , as was seen in the i.p. injection challenge. In this context, the high prevalence of R. salmoninarum in the feces and resultant shedding could also contribute to horizontal transmiss ion via a fish host ingesting the bacterium. Furthermore, as an infection progresses and the fish reduce their feeding, less fecal material would be available to shed the bacterium, resulting in a decline in the prevalence of R. salmoninarum in the feces, as was demonstrated on the last day of the injection challenge. Thus, findings from this study suggest that R. salmoninarum is readily detected in a urine/feces mixture from fish, which could help establish how frequently fish are being exposed to R. sal moninarum . Even though the mucus sample was the least effective at detecting R. salmoninarum in this study, it is the first report of R. salmoninarum being recovered from the mucus layer of a fish. Hoffman et al. (1984) isolated R. salmoninarum from skin lesions from brown trout, coho salmon, and rainbow trout, but not from the actual mucus. Mucus is first and foremost a physical barrier to pathogens, as it is constantly being produced and sloughed from the surface, physically trapping and removing the ba Yet it also has a protective role in fish immunity and contains proteolytic enzymes, lymphocytes, antibodies, and lysozyme, which are important components of the fish immune response (Ourth, 19 80; Hjelmeland et al., 1983; St. Louis - Cormier et al., 1984; Ellis, 2001). It was interesting to note, however, that the R. salmoninarum infection prevalence in mucus collected from immersion - challenged fish was much lower than that of injection - challenge d 143 fish. One potential explanation is that injection - challenged fish consistently had a higher infection prevalence in their uro - fecal samples than did immersion - challenged fish, possibly indicating higher loads of bacteria being shed into the water that c ould result in more bacteria becoming imbedded within the mucus layer. Renibacterium salmoninarum was also detected, albeit to a lesser degree, in the blood samples collected in this study. The level of prevalence depended on the exposure route, with a hi gher observed prevalence in injection infected fish. This finding lends support that a state of bacteremia is longstanding in the sub - acute course of the disease. Despite the fact that fish in the immersion challenge were exposed to a higher total number of bacteria in the water compared to the injection challenge, injection - challenged fish developed a sub - acute course of BKD, exhibiting clear clinical signs, while the immersion - challenged fish exhibited a milder course of BKD, with very few clinical sign s. Granulomatous - like lesions in the kidney were noticed in several fish from multiple sampling periods in the injection group, but were not observed in the immersion group. The differences noted in the immersion and i.p. injection challenges were most l ikely due to the nature of their exposure to the bacteria. Intraperitoneal injection delivers the bacteria directly into the fish, bypassing its external defense mechanisms; whereas an immersion challenge does not forego these important components of the immune system. Clearly, the exposure route and resultant disease course impact the detection capabilities of lethal and non - lethal samples. Renibacterium salmoninarum was detected at much higher levels in the samples collected from fish exhibiting a sub - acute disease course than a chronic course. In the sub - acute fish, R. salmoninarum was able to spread throughout the fish 144 to all of the organs that were tested (kidney, spleen, intestine, blood, and mucus) and was detected more consistently. Conversely, the chronic nature of the immersion challenge impeded the ability to consistently detect R. salmoninarum from the non - lethally collected samples. In all three scenarios (sub - acute, chronic, or a combined infection), the uro - fecal sample was the ideal ca ndidate for non - lethal disease detection. W hile testing a lethal sample (i.e., kidney and spleen homogenate) is the current method to determine if a fish is infected (AFS - FHS, 2012; WOAH, 2012), there was some concern as to whether they reflected the tr ue disease status and they were not chosen to be - lethal samples were compared. In both the immersion and i.p. injection challenge, fish were exposed to or injected with virulent R. salmoninarum ; therefore, I contend that R. salmoninarum should have been detected in all of the lethal samples. In fact, on day 1 and 63 p.i. in the immersion challenge, there were 17 fish where R. salmoninarum was detected in the uro - fecal sample, but not the lethal sample; sig nifying that the lethal samples alone missed almost 10% of the positive fish. However, it is important to point out that the lethal samples had the best ability to determine true positive and negative samples, followed by the uro - fecal samples. The assay s used in this study (culture, ELISA, and nPCR) have been previously shown to yield variable results when used on the same sample (Cipriano et al., 1985; Pascho et al., 1987; White et al., 1995; Faisal and Eissa, 2009; Schulz, Chapter 2 ); therefore, how th e diagnostic assay affected the efficacy of the non - lethal sample was evaluated. Interestingly, nPCR was determined to be the best assay to correctly identify positive and negative fish in the uro - fecal samples. The nPCR that was used in this study is hi ghly specific for R. salmoninarum , and only 145 requires a small amount of DNA be present for detection (Chase and Pascho, 1998). The high detection of R. salmoninarum by nPCR in the uro - fecal samples could be due to a continuous shedding of the bacterium int o the surrounding environment. Conversely, ELISA was better suited for detecting R. salmoninarum in the blood, mucus, and lethal samples than nPCR, which is one of the most widely used assays for detection of R. salmoninarum (Pascho and Mulcahy, 1987; Pa scho et al., 1987; Pascho et al., 1991; White et al., 1995; Jansson et al., 1996; Bruno, 2004; Schulz, Chapter 2). The detection rate of non - lethal samples was greatly enhanced by combining the collection and testing of both the lethal and uro - fecal sampl es tested by ELISA and nPCR, respectively. The lethal sample is comprised of a sample of the kidneys and spleen, organs which frequently remove circulating bacterial antigens from the host and provide an indication of a septicemic infection (Bruno and Pop pe, 1996). Also, persistence of R. salmoninarum in the sediment and water could be reflected in the uro - fecal sample, indicating whether fish are actively shedding the bacterium and re - infecting other individuals. Therefore, collection of both a kidney/s pleen homogenate and uro - fecal sample could shed light on the presence of R. salmoninarum in the infected host, as well as its presence in the surrounding environment. Additional knowledge regarding the extent of an R. salmoninarum infection in an aquacul ture facility would assist biologists in making better management and treatment decisions. Moreover, with the current WOAH (2012) and AFS - FHS (2012) approved sampling methods, the only additional sample to be tested would be the uro - fecal sample, which is relatively inexpensive and easy to collect. Additionally, while nPCR would be an additional assay to perform, there are several well - defined conventional and nested PCR procedures in existence 146 (Brown et al., 1994; Chase and Pascho, 1998; Hong et al., 200 2) that can be adapted and readily carried out in most laboratories. While the combined non - lethal samples did not yield very high sensitivities from immersion - challenged fish, it was still of interest to evaluate their potential use in a wild populatio n. As mentioned previously, the effectiveness of the non - lethal samples depended on the disease course of infected individuals. It is difficult to assess disease course in wild populations, although it is most likely that the wild fish populations will h ave multiple disease courses occurring simultaneously (Faisal and Eissa, 2009; Nance et al., 2010; Elliott et al., 2013), therefore the combined results from the injection and immersion challenge are the most relevant for a wild population. Interestingly, the most promising results indicated that collecting a lethal and uro - fecal sample could result in fewer fish being required for testing as a population consisted of more chronic infections. The addition of the uro - fecal sample resulted in higher diagnos tic sensitivities, increasing the likelihood of detecting more true positives and less false negatives. A s such, it is suggested that a urine/feces mixture should be collected in addition to the traditional kidney/spleen sample to improve the overall dete ction of R. salmoninarum and potentially reduce the number of fish needed for sampling. N on - lethal sampling may still be a viable alternative in certain scenarios. For example, the Endangered Species Act of 1973 prohibits harmful actions to any endangered or threatened animal species, including species that need to undergo disease testing. It is recommended by the United States Fish and Wildlife Service that when lethal sampling of endangered and threatened species is prohibited, non - lethal sampling techn iques should be considered (Heil, 2004). Similar to endangered species, biologists and managers may not be able to sacrifice 147 valuable broodstock populations for lethal disease testing either. In these situations, even though more fish may need to be samp led wit h lesser diagnostic sensitivity , it may be a viable alternative. 148 A PPENDIX 149 Table 3. 1. The prevalence of Renibacterium salmoninarum (with the number of positive fish) detected from the lethally (kidney and spleen hom ogenate) and non - lethally (combined, uro - fecal, blood, and mucus) collected samples from each sampling period (days post - infection) and injection and immersion challenge. Challenge Days post - infection Prevalence Lethal Combined non - lethal Uro - fecal Blood Mucus Injection 1 100% (15) 100% (15) 100% (15) 100% (15) 0% (0) 8 100% (15) 100% (15) 100% (15) 73% (11) 13% (2) 15 100% (15) 100% (15) 100% (15) 93% ( 14) 87% (13) 22 100% (15) 100% (15) 100% (15) 73% (11) 100% (15) 29 100% (3) 100% (3) 33% (1) 67% (2) 100% (3) Total 100% (63) 100% (63) 97% (61) 84% (53) 52% (33) Immersion 1 24% (5) 100% (21) 100% (21) 5% (1) 0% (0) 21 62% (13) 24% (5) 0% (0) 0 % (0) 24% (5) 42 76% (16) 0% (0) 0% (0) 0% (0) 0% (0) 63 91% (19) 52% (11) 52% (11) 0% (0) 0% (0) 84 62% (13) 5% (1) 0% (0) 5% (1) 0% (0) 105 93% (13) 0% (0) 0% (0) 0% (0) 0% (0) Total 69% (79) 33% (38) 28% (32) 2% (2) 4% (5) Combined Total 8 0% (142) 57% (101) 52% (93) 31% (55) 21% (38) 150 Table 3 .2 . Diagnostic performance of the lethal and non - lethal samples (uro - fecal, blood, and mucus) from the injection and immersion challenges, and a combination of the challenges. The accuracy an d sensitivity (with 95% confidence intervals), and the area under the curve (AUC) values are reported. *AUC values that have acceptable (0.7 - 0.8), excellent (0.8 - 0.9), outstanding (0.91 - 0.99), or perfect (1.00) discrimination. AUC values < 0.6 are no bet ter than chance alone. Challenge Sample Accuracy Sensitivity AUC Injection Lethal 1.00 (0.96 - 1.00) 1.00 (0.94 - 1.00) 1.00* Non - lethal 1.00 (0.96 - 1.00) 1.00 (0.94 - 1.00) 1.00* Uro - fecal 0.98 (0.92 - 1.00) 0.97 (0.89 - 1.00) 0.99* Blood 0.89 (0.81 - 0.95) 0. 84 (0.73 - 0.92) 0.92* Mucus 0.67 (0.57 - 0.77) 0.52 (0.39 - 0.65) 0.76* Immersion Lethal 0.72 (0.64 - 0.79) 0.65 (0.56 - 0.74) 0.83* Non - lethal 0.48 (0.40 - 0.57) 0.33 (0.25 - 0.42) 0.67 Uro - fecal 0.44 (0.36 - 0.53) 0.28 (0.20 - 0.37) 0.64 Blood 0.24 (0.18 - 0.32) 0.02 (0.002 - 0.06) 0.51 Mucus 0.26 (0.19 - 0.34) 0.04 (0.01 - 0.10) 0.52 Combined Lethal 0.82 (0.77 - 0.87) 0.77 (0.70 - 0.83) 0.89* Non - lethal 0.67 (0.61 - 0.73) 0.57 (0.49 - 0.64) 0.79* Uro - fecal 0.64 (0.58 - 0.70) 0.52 (0.45 - 0.60) 0.76* Blood 0.48 (0.41 - 0.54 ) 0.31 (0.24 - 0.38) 0.66 Mucus 0.40 (0.34 - 0.47) 0.21 (0.16 - 0.28) 0.61 151 Table 3 . 3. Diagnostic performance of all of the lethally and non - lethally (uro - fecal, blood, and mucus) collected samples from the combined injection and immersion challenges. The prevalence (with the number of positi ve fish), accuracy and sensitivity (with 95% confidence intervals), and the area under the curve (AUC) values are reported as detected by bacterial culture, nested polymerase chain reaction (PCR), semi - quantitative enzyme - linked immunosorbent assay (ELISA), and by all assays combined. *AUC values that have acceptable (0.7 - 0.8) or excellent (0.8 - 0.9) discrimination. AUC values < 0.6 are no better than chance alone. Sample Type Assay Prevalence Accurac y Sensitivity AUC Non - lethal Combined 57% (101) 0.44 (0.38 - 0.50) 0.51 (0.43 - 0.58) 0.63 ELISA 35% (63) 0.67 (0.61 - 0.73) 0.57 (0.49 - 0.64) 0.79* nPCR 53% (95) 0.67 (0.61 - 0.73) 0.57 (0.49 - 0.64) 0.79* Culture 31% (56) 0.67 (0.61 - 0.73) 0.57 (0.49 - 0.64) 0. 79* Uro - fecal Combined 52% (93) 0.37 (0.30 - 0.43) 0.34 (0.27 - 0.42) 0.61 ELISA 29% (52) 0.46 (0.40 - 0.53) 0.29 (0.23 - 0.36) 0.65 nPCR 52% (93) 0.64 (0.57 - 0.70) 0.52 (0.44 - 0.60) 0.76* Culture 13% (23) 0.34 (0.28 - 0.41) 0.13 (0.08 - 0.19) 0.57 Blood Combine d 31% (55) 0.31 (0.25 - 0.37) 0.20 (0.14 - 0.26) 0.43 ELISA 25% (45) 0.35 (0.29 - 0.42) 0.15 (0.10 - 0.21) 0.58 nPCR 11% (20) 0.33 (0.27 - 0.39) 0.11 (0.07 - 0.17) 0.56 Culture 28% (49) 0.45 (0.39 - 0.52) 0.28 (0.21 - 0.35) 0.64 Mucus Combined 21% (38) 0.29 (0.24 - 0 .36) 0.14 (0.09 - 0.20) 0.46 ELISA 15% (26) 0.43 (0.37 - 0.50) 0.25 (0.19 - 0.32) 0.63 nPCR 16% (28) 0.36 (0.30 - 0.43) 0.15 (0.11 - 0.22) 0.58 Culture 6% (10) 0.29 (0.23 - 0.35) 0.06 (0.03 - 0.10) 0.53 Lethal Combined 80% (142) 0.50 (0.43 - 0.56) 0.57 (0.49 - 0.64) 0.43 ELISA 79% (140) 0.84 (0.79 - 0.88) 0.79 (0.72 - 0.84) 0.90* nPCR 34% (61) 0.50 (0.44 - 0.57) 0.34 (0.27 - 0.42) 0.67 Culture 21% (38) 0.40 (0.34 - 0.47) 0.21 (0.16 - 0.28) 0.61 152 Table 3.4. The prevalence of Renibacterium salmoninarum (with the nu mber of positive fish) detected from lethal (kidney and spleen homogenate) samples only (L+/UF - ), uro - fecal samples only (L - /UF+), and both lethal and uro - fecal samples (L+/UF+) using results from ELISA for the lethal samples and from nPCR for the uro - feca l samples. Samples were collected from the injection challenge, immersion challenge, and a combination of the injection and immersion challenge. Challenge L+/UF - L - /UF+ L+/UF+ Total Injection 3% (2) 0% 97% (61) 100% (63) Immersion 53% (61) 15% (17) 13% (15) 81% (93) Combined 35% (63) 10% (17) 43% (76) 88% (156) 153 Table 3 . 5. Diagnostic performance of lethally collected samples (LETH), non - lethally collected samples (NON), and a combined uro - fecal and kidney/spleen sample (UF/KS) from the combined inje ction (sub - acute) and immersion (chronic) challenges at varying percentages in a fish population. Lethal and non - lethal samples were tested by all assays (bacterial culture, nested polymerase chain reaction (nPCR), and the semi - quantitative enzyme - linked immunosorbent assay (ELISA), while the uro - fecal and kidney/spleen sample was tested by nPCR and ELISA, respectively. The accuracy, sensitivity, and area under the curve (AUC) values are repo rted. *AUC values of acceptable (0.7 - 0.8), excellent (0.8 - 0.9), outstanding (0.91 - 0.99), or perfect (1.00) discrimination. AUC values < 0.6 are no better than chance alone. Sub - acute % Chronic % Accuracy Sensitivity AUC LETH NON UF/KS LETH NON UF/KS LETH NON UF/KS 100 100 1.00 0.99 1.00 1.00 0.99 1.00 1.00* 1.00 * 1.00* 90 10 0.96 0.91 0.98 0.95 0.89 0.97 0.98* 0.95* 0.99* 80 20 0.93 0.84 0.96 0.90 0.79 0.94 0.95* 0.90* 0.97* 70 30 0.90 0.78 0.95 0.87 0.70 0.93 0.94* 0.85* 0.97* 60 40 0.87 0.72 0.92 0.83 0.63 0.90 0.92* 0.82* 0.95* 50 50 0.85 0.67 0.91 0.80 0 .57 0.89 0.90* 0.79* 0.95* 40 60 0.82 0.62 0.90 0.77 0.50 0.86 0.89* 0.75* 0.93* 30 70 0.81 0.59 0.89 0.75 0.46 0.85 0.88* 0.73* 0.93* 20 80 0.79 0.55 0.88 0.73 0.41 0.84 0.87* 0.71* 0.92* 10 90 0.77 0.51 0.87 0.71 0.36 0.83 0.86* 0.69 0.92* 0 100 0.7 6 0.48 0.86 0.69 0.33 0.82 0.85* 0.67 0.91* 154 Table 3.6. The estimated samples size needed for testing to conclude with 95% confidence that disease is not present in a population of 10,000 fish was calculated at a minimum expected prevalence (MEP ) of 5 and 10% for lethally collected samples (LETH), non - lethally collected samples (NON), and a combined uro - fecal and kidney/spleen sample (UF/KS) from the combined injection (sub - acute) and immersion (chronic) challenges at varying percentages in a fis h population. Lethal and non - lethal samples were tested by all assays (bacterial culture, nested polymerase chain reaction (nPCR), and the semi - quantitative enzyme - linked immunosorbent assay (ELISA), while the uro - fecal and kidney/spleen sample was tested by nPCR and ELISA, respectively. Sub - acute % Chronic % 5% MEP 10% MEP LETH NON UF/KS LETH NON UF/KS 100 100 59 59 59 29 29 29 90 10 62 66 61 30 33 30 80 20 66 75 63 32 37 31 70 30 68 85 63 33 42 31 60 40 71 94 66 35 47 32 50 50 74 104 66 36 52 33 40 60 77 119 69 38 59 34 30 70 79 129 69 39 64 34 20 80 81 145 70 40 72 35 10 90 83 165 71 41 80 35 0 100 86 120 72 42 90 35 155 Figure 3.1. Characteristic bacterial kidney disease clinical signs observed in the injection chall enge included (A) petechial hemorrhage throughout the body of the fish (B) gill pallor, (C) corneal opacity with associated hemorrhage, (D) false membranes overlying the liver, and (E) granulomatous - like lesions in the kidney. 1 56 Figure 3.2. Characte ristic bacterial kidney disease clinical signs observed in the immersion challenge included (A) petechial hemorrhages in the caudal peduncle region, as well as between the fin rays (black arrow), and (B) exophthalmia and associated hemorrhage. 157 R EFERENCES 158 R EFERENCES Altinok, I., Grizzle, J.M., and Liu, Z. 2001. Detection of Yersinia ruckeri in rainbow trout blood by use of the polymerase chain reaction. Dis. Aquat. Org. 44: 29 - 34. American Fisheries Society - Fish Health Section. 2012. FHS blue book: suggested procedures for the detection and identification of certain finfish and shellfish pathogens, 2012 edition. AFS - FHS, Bethesda, Maryland. Austin, B., and Austin, D.A. 2007. Bacterial Fish Pathogens: Disease in Farmed and Wild Fish, 4th edition. Springer - Praxis, Chichester, United Kingdom. Austin, B., and Rayment, J.N. 1985. Epizootiology of Renibacterium salmoninarum , the causal agent of bacterial kidney disease in salmonid fish. J. Fish Dis. 8: 505 - 509. Balfry, S.K ., Albright, L.J., and Evelyn, T.P.T. 1996. Horizontal transfer of Renibacterium salmoninarum among farmed salmonids via the fecal - oral route. Dis. Aquat. Org. 25: 63 - 69. Benediktsd ó ttir, E., and Helgason, S. 1990. The isolation of Aeromonas salmonicida s ubsp. achromogenes from the gills of salmonid fish. Bull. Eur. Assoc. Fish Pathol. 10: 104 - 105. Bowers, R.M., LaPatra, S.E., and Dhar, A.K. 2008. Detection and quantitation of infectious pancreatic necrosis virus by real - time reverse transcriptase - polymer ase chain reaction using lethal and non - lethal tissue sampling. J. Virol. Methods 147(2): 226 - 234. Brown, L.L., Iwama, G.K., Evelyn, T.P.T., Nelson, W.S., and Levine, R.P. 1994. Use of the polymerase chain reaction (PCR) to detect DNA from Renibacterium s almoninarum within individual salmonid eggs. Dis. Aquat. Org. 18: 165 - 171. Bruno, D.W., and Poppe, T. 1996. A colour atlas of salmonid diseases. Academic Press, London, England. Bullock, G.L., Stuckey, H.M., and Mulcahy, D. 1978. Corynebacterial kidney d isease: egg transmission following iodophore disinfection. Fish Health News 7: 51 - 52. Bullock, G.L., Griffin, B.R., and Stuckey, H.M. 1980. Detection of Corynebacterium salmoninus by direct fluorescent antibody test. Can. J. Fish. Aquat. Sci. 37: 719 - 721. Cameron, A.R. 2001. FreeCalc software Version 2. http://www.ausvet.com.au/content.php?page=software#freecalc 159 Cameron, A.R., and Baldock, F.C. 1998. A new probability formul a for surveys to substantiate freedom from disease. Prev. Vet. Med. 34(1): 1 - 17. Cantor, S.B., and Kattan, M.W. 2000. Determining the area under the ROC curve for a binary diagnostic test. Med. Decis. Making 20: 468 - 470. Chase, D.M., and Pascho, R.J., 19 98. Development of a nested polymerase chain reaction for amplification of a sequence of the p57 gene of Renibacterium salmoninarum that provides a highly sensitive method for detection of the bacterium in salmonid kidney. Dis. Aquat. Org. 34(3): 223 - 229. Cipriano, R.C., Starliper, C.E., and Schachte, J.H. 1985. Comparative sensitivities of diagnostic procedures used to detect bacterial kidney disease in salmonid fishes. J. Wildl. Dis. 21(2): 144 - 148. Cipriano, R.C., Ford, L.A., Teska, J.D., Schachte, J.H ., Petrie, C., Novak, B.M., and Flint D.E. 1996. Use of non - lethal procedures to detect and monitor Aeromonas salmonicida in potentially endangered or threatened populations of migrating and post - spawning salmon. Dis. Aquat. Org. 27: 233 - 236. Cipriano, R. C., Ford, L.A., Smith, D.R., Schachte, J.H., and Petrie, C.J. 1997. Differences in detection of Aeromonas salmonicida in covertly infected salmonid fishes by the stress - inducible furunculosis test and culture - based assays. J. Aquat. Anim. Health 9: 108 - 113 . DeKinkelin, P. 1974. Corynébactériose des salmonidés: première observation en France. Bull. Franc. Pisc. 258: 3 - 8. Elliott, D.G., Applegate, L.J., Murray, A.L., Purcell, M.K., and McKibben, C.L. 2013. Bench - top validation testing of selected immunologi cal and molecular Renibacterium salmoninarum diagnostic assays by comparison with quantitative bacteriological culture. J. Fish Dis. 36: 779 - 809. Ellis, A.E. 2001. Innate host defense mechanisms of fish against viruses and bacteria. Dev. Comp. Immunol. 25 : 827 - 839. Faisal, M., and Eissa, A.E. 2009. Diagnostic testing patterns of Renibacterium salmoninarum in spawning salmonid stocks in Michigan. J. Wildl. Dis. 45(2): 447 - 456. Faisal, M., Eissa, A.E., and Starliper, C.E. 2010. Recovery of Renibacterium s almoninarum from naturally infected salmonine stocks in Michigan using a modified culture protocol. J. Advanc. Res. 1: 95 - 102. 160 Fenichel, E.P., Tsao, J.I., Jones, M., and Hickling, G.J. 2008. Fish pathogen screening and its influence on the likelihood of a ccidental pathogen introduction during fish translocations. J. Aquat. Anim. Health 20: 19 - 28. Giray, C., Opitz, H.M., MacLean, S., and Bouchard, D. 2005. Comparison of lethal versus non - lethal sample sources for the detection of infectious salmon anemia v irus (ISAV). Dis. Aquat. Org. 66(3): 181 - 185. Gudmundsd ó ttir, S., Benediktsd ó ttir, E., and Helgason, S. 1993. Detection of Renibacterium salmoninarum in salmonid kidney samples: a comparison of results using double - sandwich ELISA and isolation on selectiv e medium. J. Fish Dis. 16: 185 - 195. Gutenberger, S.K., Duimstra, J.R., Rohovec, J.S., and Fryer, J.L. 1997. Intracellular survival of Renibacterium salmoninarum in trout mononuclear phagocytes. Dis. Aquat. Org. 28: 93 - 106. Heil, N.P. 2004. Chapter 2: Sam ple Collection and Submission. In: United States Fisheries and Wildlife Service Handbook of Aquatic Animal Health Procedures and Protocols Volume 2: National Wildlife Fish Health Survey Laboratory Procedures Manual, 2 nd edition. Hjelmeland, K., Christie, M., and Raa, J. 1983. Skin mucus protease from rainbow trout, Salmo gairdneri Richardson, and its biological significance. J. Fish Biol. 23: 13 - 22. Hnath, J.G. 1993. Great Lakes fish disease control policy and model program (ed. by J.G. Hnath), pp. 1 - 38. Great Lakes Fish Commission Special Publication, Ann Arbor, Michigan. - 1991. Michigan Department of Natural Resources Technical Report 91 - 13. Lansing, Michigan. Hoffmann, R., Popp, W., and van de Graaff, S. 1984. Atypical BKD predominantly causing ocular and skin lesions. Bull. Eur. Assoc. Fish Pathol. 4: 7 - 9. Holey, M.E., Elliot, R.F., Marcquenski, S.V., Hnath, J.G., and Smith, K.D. 1998. Chinook salmon epizootics in Lake Michigan: possible contributing factors and management implications. J. Aquat. Anim . Health 10: 202 - 210. Hong, I., Longying, G., Xiujie, S., and Yulin, J. 2002. Detection of Renibacterium salmoninarum by nested - PCR. J. Fish. China/Shuichan Xuebao. 26(5): 453 - 458. Hosmer, D.W., and Lemeshow, S. 2000. Chapter 5: Assessing the Fit of the Model. In: Hosmer, D.W., Lemeshow, S. (Eds.), Applied Logistic Regression. John Wiley & Sons, Inc., New York, New York, pp. 143 - 202. Jansson, E., Torbjörn, H., Höglund, J., and Ljungberg, O. 1996. Comparative evaluation of bacterial culture and two ELISA techniques for the detection of Renibacterium salmoninarum antigens in salmonid kidney tissues. Dis. Aquat. Organ. 27: 197 - 206. 161 Kimura, T., and Awakura, T. 1977. Bacterial kidney disease of salmonid: first observation in Japan. Bull. Jap. Soc. Sci. Fish. 43(2): 143 - 150. Klinger, R., Francis - Floyd, R., Riggs, A., and Reed, P. 2003. Use of blood culture as a non - lethal method for isolating bacteria from fish. J. Zoo Wildl. Med. 34: 206 - 207. Lalkhen, A.G., and McCluskey, A. 2008. Clinical tests: sensitivit y and specificity. Contin. Educ. Anesth. Crit. Care Pain 8(6): 221 - 223. MacLean, D.G., and Yoder, W.G. 197 0 . Kidney disease among Michigan salmon in 1967. Prog. Fish - Cult. 32: 26 - 30. Meyer, F.P. 1991. Aquaculture disease and health management. J. Anim. S ci. 69: 4201 - 4208. Meyers, T.R., Short, S., Farrington, C., Lipson, K., Geiger, H.J., and Gates, R. 1993. Comparison of the enzyme - linked immunosorbent assay (ELISA) and the fluorescent antibody test (FAT) for measuring the prevalences and levels of Renib acterium salmoninarum in wild and hatchery stocks of salmonid fishes in Alaska, USA. Dis. Aquat. Org. 16: 181 - 189. Mitchum, D.L., and Sherman, L.E. 1981. Transmission of bacterial kidney disease from wild to stocked hatchery trout. Can. J. Fish. Aquat. Sc i. 38: 547 - 551. Nance, S.L., Riederer, M., Zubkowski, T., Trudel, M., and Rhodes, L.D. 2010. Interpreting dual ELISA and qPCR data for bacterial kidney disease of salmonids. Dis. Aquat. Org. 91: 113 - 119. Olea, I., Bruno, D.W., and Hastings, T.S. 1993. D etection of Renibacterium salmoninarum in naturally infected Atlantic salmon, Salmo salar L., and rainbow trout, Oncorhynchus mykiss (Walbaum) using an enzyme - linked immunosorbent assay. Aquaculture 116: 99 - 110. Ossiander, F.J., and Wedemeyer, G. 1973. Co mputer program for sample sizes required to determine disease incidence in fish populations. J. Fish. Res. Board Can. 30: 1383 - 1384. Ourth, D.D. 1980. Secretory IgM, lysozyme, and lymphocytes in the skin mucus of the channel catfish, Ictalurus punctatus . Dev. Comp. Immunol. 4: 65 - 74. Pascho, R.J., and Mulcahy, D. 1987. Enzyme - linked immunosorbent assay for a soluble antigen of Renibacterium salmoninarum , the causative agent of salmonid bacterial kidney disease. Can. J. Fish. Aquat. Sci. 44: 183 - 191. Pasc ho, R.J., Elliott, D.G., Mallett, R.W., and Mulcahy, D. 1987. Comparison of five techniques for the detection of Renibacterium salmoninarum in adult coho salmon. Tran. Am. Fish. Soc. 116(6): 882 - 890. 162 Pascho, R.J., Elliott, D.G., and Streufert, J.M. 1991. Brood stock segregation of spring Chinook salmon Oncorhynchus tshawytscha by use of the enzyme - linked immunosorbent assay (ELISA) and the fluorescent antibody technique (FAT) affects the prevalence and levels of Renibacterium salmoninarum infection in prog eny. Dis. Aquat. Org. 12: 25 - 40. Pascho, R.J., Chase, D., and McKibben, C.L. 1998. Comparison of the membrane - filtration fluorescent antibody test, the enzyme - linked immunosorbent assay, and the polymerase chain reaction to detect Renibacterium salmoninar um in salmonid ovarian fluid. J. Vet. Diagn. Invest. 10: 60 - 66. Powell, M., Overturf, K., Hogge, C., and Johnson, K. 2005. Detection of Renibacterium salmoninarum in Chinook salmon, Oncorhynchus tshawytscha (Walbaum), using quantitative PCR. J. Fish Dis. 28: 615 - 622. Rodgers, C.J. 1992. Development of a selective - differential medium for the isolation of Yersinia ruckeri and its application in epidemiological studies. J. Fish Dis. 15: 243 - 254. Sanders, J.E., and Barros, J. 1986. Evidence by the fluoresce nt antibody test for the occurrence of Renibacterium salmoninarum among salmonid fish in Chile. J. Wildl. Dis. 22(2): 255 - 257. Sanders, J.E., and Fryer, J.L. 1980. Renibacterium salmoninarum gen. nov., sp. nov., the causative agent of bacterial kidney dis ease in salmonid fishes. Int. J. Syst. Bacteriol. 30: 496 - 502. St. Louis - Cormier, E.A., Osterland, C.K., and Anderson, P.D. 1984. Evidence for a cutaneous secretory immune system in rainbow trout ( Salmo gairdner i). Dev. Comp. Immunol. 8: 71 - 80. White, M. R., Wu, C.C., and Albregts, S.R. 1995. Comparison of diagnostic tests for bacterial kidney disease in juvenile steelhead trout ( Oncorhynchus mykiss ). J. Vet. Diagn. Invest. 7(4): 494 - 499. Wiens, G.D., and Kaattari, S.L. 1999. Bacterial kidney disease ( Ren ibacterium salmoninarum ). In: Woo, P.T.K., Bruno, D.W. (Eds.), Fish Diseases and Disorders: Viral, Bacterial and Fungal Infections, Volume 3. New York: CABI Publishing, pp. 269 - 301. World Organization for Animal Health. 2012. Manual of Diagnostic Tests fo r Aquatic Animals, 7th edition. Office International des Epizooties, Paris. World Organization for Animal Health. 2013. Chapter 1: Notification of Diseases and Epidemiological Information. In: Aquatic Animal Health Code. Office International des Epizootie s, Paris. 163 Zhu, W., Zeng, N., and Wang, N. 2010. Sensitivity, specificity, accuracy, associated confidence interval and ROC analysis with practice SAS® implementations. In: NorthEast SAS® Users Group Conference Proceedings , November, 14 - 17, 2010, Baltimore , Maryland. 164 Chapter 4 Evidence of anti - Renibacterium salmoninarum A ntibodies in N aturally and E xperimentally I nfected Oncorhynchus spp. 165 1. Abstract Bacterial kidney disease, caused by Renibacterium salmoninarum , is a long - standing pr oblem among wild and aqua - cultured salmonids in the Great Lakes basin and Pacific Northwest. Several assays have been developed for the detection of R. salmoninarum in fish tissues, including the enzyme - linked immunosorbent assay (ELISA). In this study, a single - dilution indirect - ELISA was assessed with the aim to detect anti - R. salmoninarum antibodies in experimentally infected juvenile rainbow trout ( Oncorhynchus mykiss ) and adult Chinook salmon ( O. tshawytscha ), coho salmon ( O. kisutch ), and steelhead. Over a 26 wk period, antibody levels were measured every 3 wks in multiple groups of naïve (n=169) and non - naïve (n=171) rainbow trout that were injected intraperitoneally (i.p.) with live R. salmoninarum . Antibody production and mean time - to - death was determined over a 33 wk period in an additional group of i.p. injected naïve rainbow trout (n=46). Lastly, feral Chinook salmon (n=60), coho salmon (n=60), and steelhead (n=40) returning to spawn at Lake Michigan weirs were also assessed for production of anti - R. salmoninarum antibodies. The intensity of the infection in the kidneys and spleen in all fish groups was detected using standard bacterial culture techniques and a sandwich ELISA. All groups of experimentally infected and feral fish were found to produce detectable levels of antibodies, although there were minimal differences among them. The observed antibody response in experimentally infected fish appeared to be short - lived and long - term elevated levels of antibody production were not observed. Despite low infections of R. salmoninarum in the kidneys and spleen, feral fish still produced elevated levels of antibodies, suggesting that despite a currently low prevalence of BKD in the Great 166 Lakes basin, exposure to R. salmoninarum is occurring and resulting in the production of circulating antibodies. 2. Introduction Significant losses associated with Renibacterium salmoninarum , the causative agent of bacterial kidney disease (BKD), have been documented in infected salmonids from both marine ( Ellis et al., 1978; Paterson et al., 1979; Banner et al., 1983; Meyer et al., 1999) and freshwater systems (MacLean and Yoder, 1970; Mitchum et al., 1979; Schulz, Chapter 2) worldwide. In particular, BKD poses a significant risk for wild and aqua - cultured salmonids in the Great Lakes basin (GLB) and Pacific Northwest (PNW), where the pathogen is persistent, often causing clinical episodes (Mitchum and Sherman, 1981; Hnath and Zischke, 1991; Maule et al., 1996; Schulz, Chapter 2). Bacterial kidney disease is particularly difficult to control due to its ability to be transmitted by both horizontal and vertical routes, thus routine health monitoring of broodstock and their offspring, with culling of infected fish, is one of the most important methods for comb atting BKD (Schulz, Chapter 2). Several assays have been developed for the detection of R. salmoninarum in fish tissues, eggs, and ovarian fluid by bacterial culture (Evelyn, 1977; Austin et al., 1983; Faisal et al., 2010a), molecular assays (Brown et al. , 1994; Chase and Pascho, 1998; Chase et al., 2006), and immunological methods such as Western blot, immunofluorescence techniques, and sandwich enzyme - linked immunosorbent assay (ELISA) (Pascho and Mulcahy, 1987; Griffiths et al., 1991; Hsu et al., 1991; Jansson et al., 1996). On the contrary, the detection and quantitation of 167 circulating antibodies against R. salmoninarum has received little attention, and has been, for the most part, based on experimental infections. For example, Evelyn (1971) develop ed an agglutination assay to determine antibody titers in experimentally infected sockeye salmon ( Oncorhynchus nerka ). However, it was not determined if the detected antibodies conferred any protective immunity in this experiment. Later, Olivier et al. ( 1992) used Western blot to monitor the production of antibodies against R. salmoninarum in experimentally infected Atlantic salmon. With the expansion of monoclonal antibody production, several ELISA based assays have been developed to detect whole R. s almoninarum bacteria, its soluble antigen, or specific anti - R. salmoninarum antibodies circulating in sera of surviving fish (Pascho and Mulcahy, 1987; Bartholomew et al., 1991; Hsu et al., 1991; Rockey et al., 1991). Other than experimentally - infected fi sh, antibody assessment in cultured, feral, or wild salmonids has seldom been used. However, Jansson and Ljungberg (1998) did detect an elevated antibody response to R. salmoninarum in farm - raised Atlantic salmon ( Salmo salar ) and rainbow trout ( Oncorhync hus mykiss ) using an indirect ELISA, leading the authors to suggest this method may be valuable for monitoring stocks of fish that are known to have likely been exposed to R. salmoninarum or survived a BKD episode. Using an ELISA - based assay to assess p roduction of R. salmoninarum antibodies has gone through several phases of development. To semi - quantitate levels of antibodies being produced in a fish with ELISA, several dilutions of each serum sample are usually analyzed to calculate units of antibody or to arrive at an endpoint dilution (Wood and Kaattari, 1996). However, this method can increase the cost of the reagents and limit the number of fish 168 assessed on each ELISA run (Alcorn and Pascho, 2000). Therefore, a single - dilution antibody ELISA was developed with the aim to increase the number of samples being tested on each ELISA microplate, thereby decreasing the cost per sample (Alcorn and Pascho, 2000). The method required the establishment of a standard curve with a positive serum sample of kn own titer to function as a reference for the test samples (Alcorn and Pascho, 2000). Additionally, typical antibody quantitation by ELISA requires antibodies from the test sera to bind to an antigen that has been coated to the microplate wells for several hours before running the assay (Arkoosh and Kaattari, 1990). The main soluble antigen from R. salmoninarum consists primarily of a highly glycosylated protein with a molecular weight of ~57 kilodaltons (kDa), known as p57 (Getchell et al., 1985). Coatin g the wells with the overwhelmingly glycosylated p57 only can mask the detection and quantitation of fish antibodies directed against the bacteria themselves, and not their soluble antigen. This led Wood and Kaattari (1996) to develop an ELISA protocol fo r antibody detection with whole R. salmoninarum bacteria to coat the wells, which has lowered the costs associated with soluble antigen preparation. The GLB has been plagued with Renibacterium salmoninarum since the 1950s with serious mortality of hatcher y - raised salmonids frequently observed, a matter that necessitated the implementation of stringent control measures including broodstock culling, egg disinfection, and regular screening (Hnath, 1993). It is clear that the measures 7 implemented basinwide h ave caused a dramatic reduction in the prevalence and intensity of R. salmoninarum infections (Schulz, Chapter 2) to the extent that BKD has become an endemic disease of less concern in Great Lakes (GL) Oncorhynchus species. While the reason for R. salmon inarum decline has been mainly attributed to the stringent biosecurity measures and 169 egg disinfection program, others have demonstrated that Chinook salmon from the Great Lakes are less susceptible and have decreased BKD - associated mortality upon experiment al infection compared to their species and age cohort from the PNW (Purcell et al., 2008). The previous authors attributed this to pathogen - driven selection; when the GL Chinook salmon population was subjected to the mass fish mortalities in the late 1980 s, the most resistant fish survived, resulting in a more resistant population overall (Purcell et al., 2008). However, it was unclear if the decline in BKD incidence in Oncorhynchus spp. is due to inherent resistance, or that the pathogen presence in the GLB has substantially diminished. The study performed by Faisal et al. (2010b), however, demonstrated that R. salmoninarum continues to be a threat to GL salmonids. The authors found high prevalence (~65%) and intensity of R. salmoninarum infection, as w ell as clinical BKD signs, in four naturally occurring (i.e., no artificial propagation and stocking) lake whitefish stocks in lakes Michigan and Huron. While the data of Faisal et al. (2010b) confirmed the continuous presence of R. salmoninarum in the Gr eat Lakes watershed at levels sufficient to cause BKD in susceptible salmonids, it also raised the important question as to why wild Oncorhynchus species, sharing the same area with heavily infected lake whitefish, have not succumbed again to R. salmoninar um and developed clinical BKD. To this end, this study was initiated to develop an ELISA - based assay that is capable of detecting and quantitating circulating antibodies against R. salmoninarum in experimentally infected juvenile fish, as well as adult On corhynchus spp. returning to spawn at gamete - collecting facilities. The modified ELISA assay developed in the course of this study enabled the assessment of the serological status in Oncorhynchus species exposed to R. salmoninarum under both laboratory an d natural field conditions. Additionally, it was also tested if the 170 presence of circulating anti - R. salmoninarum antibodies can confer protection to fish experimentally exposed to this serious pathogen. 3. Materials and Methods 3.1. Determination of an ti - R. salmoninarum antibodies in experimentally infected rainbow trout Two groups of rainbow trout, acquired from Troutlodge, Inc. (Sumner, WA), were used in this study: 1) a group of rainbow trout (~11 months old; n=169) that survived a previous exposu re to R. salmoninarum n=171) (Table 4.1). The survivor fish group (SF) acclimated to laboratory conditions for seven weeks, and was then exposed to a low dose of R. salmoninarum [6.6 x 1 0 5 colony forming units (CFU)] via treated food pellets for one week. Nine weeks later, survivors from this challenge were further divided into two subgroups: a group of rainbow trout which were immunized with an oral BKD bacterin (SFi) at a dose of 0.38 µg of antigen fish - 1 day - 1 , and a group of non - immunized rainbow trout (SFn) that received sterile saline. This was followed by a second bacterin that was administered 16 weeks after the initial bacterin at a dose of 0.53 µg of antigen fish - 1 day - 1 . Two weeks after administration of the secondary bacterin, fish were intraperitoneally (i.p.) injected with R. salmoninarum as described below. Conversely, the naïve rainbow trout group (NF) acclimated to laboratory conditions for three weeks prior to being divided into similar subgroups. A portion of the NF fish group were 171 exposed to the same BKD oral bacterin at a dose of 0.085 µg of antigen fish - 1 day - 1 (NFi), while a non - immunized group received sterile saline (NFn). Sixteen weeks later, the NFi group r eceived a second bacterin dose of 0.16 µg of antigen fish - 1 day - 1 . Similarly, two weeks post - secondary bacterin, fish were i.p. injected with R. salmoninarum as described below. Prior to experimental challenge, R. salmoninarum (American Type Culture Coll ection #33209) was passed once through naïve rainbow trout to ensure infectivity. The recovered R. salmoninarum isolate was then inoculated onto modified kidney disease medium ( MKDM; Faisal et al., 2010a), incubated at 15 ° C for 14 d, purity verified, and then a single colony was inoculated into a 7 - ml aliquot of MKDM broth (x2) and incubated at 15 ° C for 7 d. Ten µ l from both of the broth cultures was then sub - cultured onto trypticase soy agar (TSA) and MKDM to verify purity, and the remaining broth was ad ded to 1000 ml of fresh MKDM broth and incubated at 15 ° C on a Thermolyne Nuova stir plate (Thermo Fischer Scientific, Inc., Waltham, MA) at approximately 50 rpm in a 2 L Celstir Spinner Flask (Wheaton, Millville, NJ). After 14 days of incubation, the brot h culture was centrifuged in a Hermle Z382K centrifuge (Labnet International, Inc., Woodbridge, NJ) at 4300 rpm for 10 min and the supernatant was discarded. The bacterial pellet was re - suspended in 0.85% saline solution, centrifuged as described above, a nd the supernatant was discarded. This was repeated three times. T he remaining pellet was re - suspended into 926 ml of sterile saline, for a final concentration of 3.7 x 10 10 cfu ml - 1 , which was used for i.p. injection. Fish were first anesthetized wit h 100 mg/L of sodium bicarbonate - buffered tricaine methanesulfonate (MS - 222; Argent Chemical Laboratories, Inc., Redmond, Washington) for approximately 10 - 15 sec, and were then i.p. injected with the bacterial suspension. Due to the 172 larger size of the SF g roup, survivor fish received 200 µl of the bacterial suspension, which is double the volume that was administered to the NF group (i.e, 100 µl). Infected NF and SF were then revived by placing them in freshwater in their 225 L and 74L tanks, respectively, and maintained at a water temperature of 12 ± 2.0 °C . Triplicate tanks containing 26 - 27 fish per tank were used for each of the four fish groups. This was done in accordance with the Michigan State University International Animal Care and Use Committee ( Animal Use Form #07/12 - 133 - 00). All tanks were monitored daily for mortalities. 3.2. Protection associated with anti - R. salmoninarum antibodies To determine if circulating antibodies confer protection to rainbow trout, an additional experiment was de signed with naïve rainbow trout (11 months old; n=46) divided into two subgroups: naïve rainbow trout immunized with the oral BKD bacterin (NPi) and a second group that was not immunized (NPn). The NPi group was exposed to a single dose of the BKD oral ba cterin (4.9 µg of antigen fish - 1 day - 1 ) 29 weeks post - acclimatization, while the NPn group received sterile saline. Four weeks after bacterin administration, both groups were i.p. injected with 500 µl of a 3.7 x 10 10 cfu ml - 1 R. salmoninarum bacterial sus pension, as described previously. Fish were then revived by placing them in freshwater in their respective 74 L tanks, with a water temperature of 12 ± 2.0 °C . Five replicate tanks containing 4 - 5 fish per tank were used for each group. This was also don e in accordance with the Michigan State University International Animal Care and Use Committee (Animal Use Form #07/12 - 133 - 00). All tanks 173 were monitored daily for mortalities. For each mortality, time to death (TTD) was recorded and samples were collecte d as described below. 3.3. Sample collection from experimentally infected fish Samples were collected every three weeks post - infection (p.i.) from the SFi, SFn, NFi, and NFn groups (n=5), while samples were only collected from moribund or freshly dead N Pi and NPn fish. Fish were euthanized with a lethal dose of 250 mg/L of sodium bicarbonate - buffered MS - 222 (Argent) prior to sample collection. Once euthanized, individual lengths and weights were recorded and a thorough internal and external examination for gross signs of disease was performed. Blood samples were collected by venipuncture of the caudal vein using a sterile needle and syringe and were allowed to clot overnight at 4°C. Blood samples were then centrifuged in a Denville 300D Centrifuge at 5000 rpm for 10 min to separate the serum from the plasma, the serum was removed, aliquoted, and stored at - 80°C until diagnostic assays were performed. Isolation of R. salmoninarum from the kidney was performed by streaking 10 µl of the kidney tissue ont o MKDM agar plates using sterile disposable inoculating loops, and incubating the plates under aerobic conditions at 15°C for a total of 42 days. The remaining kidney/spleen sample was collected with individual sterile forceps and scissors, placed in indi vidual whirlpaks, and stored at - 80°C until the sandwich enzyme - linked immunosorbent assay (ELISA) was performed. Kidney/spleen tissue processing and ELISA were performed as described in Chapter 2. 174 3.4. Assessment of anti - R. salmoninarum antibodies in th ree Oncorhynchus spp. returning to spawn at gamete collecting facilities During April, September, and October of 2009 and 2013, adult Chinook salmon (5.6 kg ± 2.5 kg), coho salmon (0.7 kg ± 2.6 kg), and steelhead (2.7 kg ± 5.6 kg) were collected from tw o Michigan Department of Natural Resources gamete - collecting weirs: Chinook salmon and steelhead from the Little Manistee River Weir (LMRW) and coho salmon from the Platte River Weir (PRW) (Figure 4.1; Table 4.1). Prior to sample collection, fish were eut hanized by immersion in CO 2 - laden water and manual blunt force trauma to the cranium. Blood was collected from the dorsal aorta at the caudal peduncle region and allowed to clot in sterile tubes overnight. Serum was separated as described above. Length, weight, and sex of all fish were recorded and external examinations were performed. Fish were externally disinfected with 70% ethanol and subjected to thorough internal clinical examinations. A kidney/spleen sample was collected as described above for E LISA analysis. 3.5. Modified single dilution indirect ELISA to detect and quantitate R. salmoninarum antibodies in fish sera The R. salmoninarum - specific antibody response of experimentally and naturally infected fish was quantitated by an indirect E LISA based on Wood and Kaattari (1996) and Alcorn and Pascho (2000), with modifications. Briefly, the surfaces of a microplate were coated with formalin - killed R. salmoninarum cells (500 µg ml - 1 ) and were incubated at 17°C overnight. 175 Plates were then blo cked with phosphate buffered saline (PBS) containing 1% goat serum (Sigma - Aldrich) for 1 hr at room temperature (RT). Test sera was diluted 1:10 in PBS containing 0.05% tween 20 (PBS - T20), added to the microplate, and incubated at RT for 2 hr. To help re duce non - specific binding, plates were washed three times with PBS - T20, followed by soaking the plates with PBS - T20 for three 10 min intervals. The presence of anti - R. salmoninarum activity was then detected using a purified anti - rainbow trout monoclonal antibody (CLF001AP, Cedarlane Labs, Burlington, NC) diluted 1:200 in PBS - T20 and was incubated at RT for 2 hr. After washing, an anti - mouse IgG - alkaline phosphatase antibody (Sigma - Aldrich), diluted 1:1000 in PBS - T20 containing 0.5% goat serum, was added and incubated at 37°C for 1 hr. Plates were washed again as previously described. Lastly, alkaline phosphatase substrate (Sigma - Aldrich) was added and the microplate was incubated in the dark at RT for 45 min. The plate was then read at 405 nm with a Bi oTek ELx808 (BioTek Instruments, Inc., Winooski, VT) plate reader. Each ELISA run included a previously recognized high titer serum sample diluted two - fold (1:10, 1:20, 1:40, 1:80) and a diluted negative serum sample (1:10). The positive - negative cut - o ff absorbance for the samples was the mean anti - R. salmoninarum antibody activity (plus two standard deviations) of 4 month old naïve rainbow trout (n=20) with no history of exposure to Renibacterium salmoninarum , as recommended by Alcorn and Pascho (2000) . 176 3.6. Statistical analyses Comparisons of antibody production among the fish groups were done using a one - way ANOVA and assessed using SAS statistical software, version 9.3 (SAS Institute Inc., Cary, NC). Survival statistics were calculated by the Kaplan - Meier analysis also using SAS 9.3. Differences between the probability of survival of NPn and NPi fish were determined by the logrank comparison, with statistical significance set at P 4. Results 4.1. Experimental infection Infection by R. salmoninarum and the development of BKD signs were observed in all four groups of the experimentally infected fish during the multiple sampling periods post - infection. Specifically, the intensity of the R. salmoninarum isolated from the kidneys (i.e. , average CFUs) in the NFn and NFi groups peaked at week 3 p.i. and then steadily declined in subsequent sampling periods, until the infection disappeared by week 15 p.i. (Figure 4.2). The p57 antigen was detected in the kidneys and spleen by sandwich ELI SA in all fish and continued its presence through the end of the observation period at week 26 p.i. (Figure 4.2). No trends were observed regarding the intensity of p57 between the two fish groups and among the sampling periods. As depicted in Figure 4.3 , a similar trend was noticed in the levels of p57 antigen, as well as the number of R. salmoninarum CFUs isolated from the kidneys, in SFn and 177 SFi groups. The maximum intensity of infection levels in the SFn and SFi groups, however, were considerably lowe r (e.g., 67 CFUs) than the NFn and NFi groups (e.g., 101 CFUs) (Figure 4.3). 4.2. Antibody production in experimentally infected fish Anti - R. salmoninarum antibodies were clearly detected in all four groups of experimentally infected fish, with a trend of increasing in levels starting as early as week 3 p.i. (Figures 4.4 and 4.5). Levels of antibodies averaged an OD value of 0.257 for all fish and reached a maximum OD value of 2.514. There were no clear differences among the production of antibodies i n the four fish groups and the small sample size at each sampling period p.i. did not allow the performance of statistical analysis. However, the experiment was terminated at week 26 p.i., thus increasing the sampl e size, allowing for statistical analysis to be performed. Analysis showed that the NFn group had a significantly higher antibody response than the NFi group ( F = 6.09 ; df = 1, 88; P = 0.0388), SFn group ( F = 10.25; df = 1, 85; P = 0.0126), and SFi group ( F = 6.90 ; df = 1, 87; P = 0.0303) at wee k 26 p.i. Measurable antibodies were first observed in the NFn and NFi groups at week 6 p.i. and peaked at week 15 p.i. (Figure 4.4). After week 15 p.i., the production of antibodies started to decline in both groups, which was consistent with the rate of infection observed in the kidneys and spleen (Figure 4.2). Antibodies were detected as early as week 3 p.i. in the SFn group of fish, while the antibody response of the SFi group (OD = 0.0916) was just below the positive - negative threshold (OD = 0.094 ) (Figure 4.5). The antibody response of the SFn group peaked between weeks 9 and 12 p.i. (Figure 4.5), which is also the period that followed their highest 178 rates of infection (Figure 4.3). Similarly, the time period (weeks 12 to 24 p.i.) when the antibo dy response of the SFi group was highest also coincided with the most consistent infection rates, as demonstrated by the prevalence of the R. salmoninarum p57 antigen in the kidneys (Figure 4.5). Additionally, the SFi group of fish produced elevated anti - R. salmoninarum antibodies well into weeks 18 and 24 p.i., contrary to the other three experimental groups (Figure 4.5). 4.3. Detection of antibodies in adult Oncorhynchus spp. Antibodies against R. salmoninarum were detected in sera of Chinook salmon, coho salmon, and steelhead returning to spawn at LMRW and PRW (Figure 4.6). These antibodies were detected in the majority of the Chinook salmon and steelhead, and to a lesser degree, in the coho salmon (Figure 4.6; Table 4.1). Overall, the distribution of antibody levels for most of the fish groups appeared to be positively skewed, with the larger proportion of anti - R. salmoninarum antibody - producing fish occurring in the upper quartile (Figure 4.6). Overall, Chinook salmon produced a wide range of con sistently similar levels of antibodies in both 2009 and 2013, which tended to be greater than the median antibody level (Figure 4.6). However, in 2009, the mean antibody response of female Chinook salmon was significantly greater than the male antibody re sponse ( F = 13.01 ; df = 1, 58; P = 0.006). Also, a larger range of antibody levels for male steelhead was observed in 2013 compared to 2009, with the most variation occurring in the upper quartile (Figure 4.6). Furthermore, the mean antibody response of male steelhead was significantly greater than female steelhead in 2013 ( F = 14.93 ; df = 1, 58; P = 0.0 0 02). In 179 2009 and 2013, considerably less variation was observed in the levels of antibodies detected in male coho salmon, with the exception of one indi vidual in 2009 with an OD value of 1.568 (Figure 4.6). Alternatively, a substantially wider range of antibody levels was observed in female coho salmon from 2013 than 2009, with the majority of antibody - producing fish occurring in the upper quartile (Figu re 4.6). Interestingly, the R. salmoninarum p57 antigen was only detected in two of the kidneys and spleens from Chinook salmon that were producing antibodies, at a medium and high intensity (Table 4.1). Similarly, one steelhead had a medium intensity in fection in the kidneys and spleen and was also producing antibodies (Table 4.1). However, while p57 was not detected in the kidneys and spleen of any coho salmon, it was detected at medium and low intensities in the reproductive fluids of one male and thr ee female fish (Table 4.1). Interestingly, none of the gamete - infected individuals were producing detectable levels of anti - R. salmoninarum antibodies (Table 4.1). 4.4. Potential protective role of circulating binding antibodies to challenge with live R. salmoninarum Antibodies were detected in the lethargic and dying fish in both NPn and NPi groups with the exception of week 30 p.i. (Table 4.2). Antibody levels in the NPn group reached up to 0.270 OD value, while antibodies detected in the NPi fis h group showed a much higher response, with OD values as high as 0.608 (Table 4.2). The antibody response in both fish was the highest at weeks 7 and 8 p.i. (Table 4.2), which was also when the mortalities diminished 180 (Figure 5.7). Similar to the NF and S F groups of fish, the R. salmoninarum p57 antigen was detected more frequently earlier in the challenge as well and was not detected at all after week 7 p.i. (Table 4.2). The Kaplan Meier analysis revealed that the difference in the overall survivabil ity of the NPn fish group was not significantly different from the NPi group ( 2 = 0.0524, df = 1, P = 0.8189). The mean time - to - death was longer for the NPi fish [22.6 weeks (SE ± 2.5019)] compared to the NPn fish [6.1 weeks (SE ± 0.2939)]; however, due to the small number of mortalities that occurred, this was not deemed signific ant. Also, over 60% of fish from both groups infected with R. salmoninarum survived until week 33 p.i. when the challenge was terminated (Table 4.2). 5. Discussion The indirect ELISA protocol used in this study consistently detected circulating anti - R. salmoninarum antibodies in the majority of tested fish sera. The modified protocol was based upon the two earlier studies of Wood and Kaattari (1996) and Alcorn and Pascho (2000). While the modified protocol utilized the same coating and blocking stra tegies developed by Wood and Kaattari (1996), a single dilution of each test serum (s recommended by Alcorn and Pascho (2000) was used in this protocol, as opposed to five dilutions of each test serum per assay used by Wood and Kaattari (1996). This modif ication enabled the testing of 40 samples on one ELISA microplate as opposed to ~8 samples/plate (Wood and Kaattari, 1996), without compromising the sensitivity and specificity of the assay. As such, the modified protocol reduces costs of reagents, potent ially reduces the error inherent to serial dilutions, and is relatively 181 uncomplicated to execute. Also, the Wood and Kaattari (1996) protocol used the biotinylated mouse anti - trout immunoglobulin monoclonal antibody 1 - 14, while the modified protocol uses a commercially available monoclonal antibody that is directed against the heavy chain of Oncorhynchus spp., a matter that allowed testing of rainbow trout/steelhead trout, Chinook salmon, and coho salmon. Further, the use of a tertiary alkaline phosphatas e - conjugated goat - anti - mouse antibody allowed more amplification and consequently better visualization of the reaction. Lastly, microplates from the modified ELISA were read a single time at 405 nm, compared to the kinetic read performed by Wood and Kaatt ari (1996) at the same wavelength, thereby saving 10 minutes per microplate. In this study, four experimentally infected groups were arbitrarily used to generate fish specific anti - R. salmoninarum circulating antibodies that could be used to test the m odified protocol. The antibodies were detected in the four fish groups as early as week 3 p.i. and lasted until the 26 th week p.i., when the observation period ended. However, in most of the fish groups, elevated levels of antibody production were short - lived and elevated long - term production of this particular fish antibody type did not appear to occur. Also, the maximum antibody responses were associated with either an ongoing or recovering R. salmoninarum infection, with declining production of antibo dies within 15 - 18 weeks p.i. Without extending the study period, it is unknown how long antibody production would persist , but it most likely would not have continued . Alcorn et al. (2005) observed a decline in the detection of anti - R. salmoninarum antib odies in experimentally infected Atlantic salmon after 11 weeks post infection, as did Jansson and Ljungberg (1998) after 8 weeks post - infection. To this end, the production of antibodies observed in experimental fish from this study is likely a transient 182 response to an infection with live R. salmoninarum . The extent to which protection is conferred by this antibody type appears to be minimal. Unexpectedly, antibody levels in spawning fish were considerably higher than in experimental fish. Indeed, th ere are a handful studies that have attempted to assess the presence of antibodies to R. salmoninarum in adult feral or wild salmonids. Bartholomew et al. (1991) detected anti - R. salmoninarum antibodies in naturally infected spring Chinook salmon and coho salmon using Western blot methods, describing the levels as low titers. Additionally, a weak antibody response against R. salmoninarum was observed in adult Atlantic salmon and rainbow trout from fish farms with a history of clinical BKD (Jansson and Lju ngberg, 1998). The authors hypothesized that the weak antibody response could have been due to lower water temperatures throughout their study (8°C) in the farms. The assay modified in this study detected antibodies against R. salmoninarum in feral Chi nook salmon, coho salmon, and steelhead that were returning to spawn at Lake Michigan weirs in 2009 and 2013 with relatively high levels. The extremely low presence of the p57 antigen coincides with the observed decline of BKD in the GLB (Schulz, Chapter 2). Despite the lack of detection of p57 in the kidneys, ~80% of the feral fish were producing detectable antibodies against R. salmoninarum at higher levels than the experimentally challenged fish. It is possible that the feral Chinook salmon, coho salm on, and steelhead are being exposed to R. salmoninarum while residing in the Great Lakes and are becoming re - infected with the pathogen due to its presence in other fish species (Faisal et al., 2010b). This data suggests that the spawning - stressed fish ar e prevailing against this serious pathogen as they find their way to their spawning grounds. Throughout this study, the ELISA - whole bacterium binding antibodies 183 were the only host defense mechanism measured. Fish are known to combat R. salmoninarum throu gh multiple immune molecules and phagocytic cells (Young and Chapman, 1978; Secombes, 1985; Hardie et al., 1996; Grayson et al., 2002), which can be more potent and long lasting compared to the antibodies measured in this study. 184 A PPENDIX 185 Table 4.1. Summary of the presence of anti - Renibacterium salmoninarum antibodies, p57 antigen, and the bacterium in blood and kidney and spleen samples from experimentally and naturally infected Oncorhynchus spp. An indirect enzyme - l inked immunosorbent assay (ELISA) was used to determine the number of fish producing antibodies (No. Ab+ fish) with the resultant mean optical density (OD) value. Active infections with R. salmoninarum were assessed by bacterial culture [No. K/S+ fish cul ture)] and sandwich ELISA [No. K/S+ fish (ELISA)]. The intensity of infection as determined by ELISA was designated as low (L), medium (M), or high (H). ND = not determined. Fish species No. fish sampled No. Ab+ fish Mean OD value No. K/S+ fish (ELISA) No. K/S+ fish (culture) Rainbow trout Naïve (NFn) 86 43/86 0.351 18/86 (4H - 7M - 7L) 17/86 Naïve, bacterin (NFi) 85 44/85 0.234 23/85 (2H - 12M - 9L) 16/85 Survivor (SFn) 85 41/82 0.188 17/85 (3H - 11M - 3L) 16/85 Survivor, bacterin (SFi ) 84 55/84 0.298 18/86 (5H - 8M - 5L) 23/84 Naïve (NPn) 22 10/22 0.125 7/22 (2H - 5M) 7/22 Naïve, bacterin (NPi) 23 14/24 0.198 4/24 (2H - 2M) 10/23 Chinook salmon Male 30 30/30 0.607 1/30 (M) ND Female 30 30/30 0.660 1/30 (H) ND Coh o salmon Male 20 7/20 0.203 0/20 ND Female 20 8/20 0.266 0/20 ND Steelhead Male 30 28/30 0.394 1/30 (1M) ND Female 30 26/30 0.297 0/30 ND 186 Table 4.2. Summary of the survival of naïve non - immunized (NPn) and naï ve immunized (NPi) rainbow trout ( Oncorhynchus mykiss ) in the weeks post - infection with Renibacterium salmoninarum . An indirect enzyme - linked immunosorbent assay (ELISA) was used to determine the number of fish producing antibodies (No. Ab+ fish) with the resultant mean optical density (OD) value. Active infections with R. salmoninarum were assessed by bacterial culture [No. K/S+ fish culture)] and sandwich ELISA [No. K/S+ fish (ELISA)]. The intensity of infection as determined by ELISA was designated as low (L), medium (M), or high (H). Weeks post - infection Group No. of mortalities No. Ab+ fish Mean OD value No. K/S+ fish (ELISA) No. K/S+ fish (culture) 4 NPn 5 1/5 0.076 5 (3M - 2H) 1/1 fish 5 NPn 2 1/2 0.102 2 (2M) 2/2 fish 7 NPn 1 1/1 0.270 0 1/1 fis h 33* NPn 14 7/14 0.135 0 3/14 fish 3 NPi 1 1/1 0.105 1 (H) 1/1 fish 4 NPi 2 1/2 0.080 2 (1M - 1H) 2/2 fish 6 NPi 2 1/2 0.158 0 2/2 fish 7 NPi 1 1/1 0.604 1 (1M) 1/2 fish 8 NPi 2 1/2 0.608 0 1/2 fish 30 NPi 1 0/1 0.061 0 1/1 fish 33* NPi 15 9 /15 0.151 0 2/15 fish *All remaining fish were euthanized on Week 33 and the challenge was terminated. 187 Figure 4.1. The Michigan Department of Natural Resources gamete - collecting weirs where Chinook ( Oncorhynchus tshawytscha ), coho salmon ( O. kisutch ), and steelhead ( O. mykiss ) were collected in 2009 and 2013: the Little Manistee River Weir (Chinook salmon and steelhead) and the Platte River Weir (coho salmon). 188 Figure 4.2. The prevalence of Renibacterium salmoninarum in the kidney/spleen tissue samples of A) 8 month old naïve rainbow trout ( Oncorhynchus mykiss ), and B) age cohort rainbow trout that have received two doses of R. salmoninarum bacterin per os at each sampling period (weeks post - infection with live bacteria). Renibacterium salmonin arum and its p57 antigen were detected by the sandwich enzyme - linked immunosorbent assay, with the infection intensity expressed as the proportion of fish exhibiting low, medium, and high levels of infection (n=5 per sampling periods; n=45 at week 26 p.i.) . The average number of R. salmoninarum colony forming units (CFUs) isolated from kidney is depicted as a line graph. 189 Figure 4.3. The prevalence of Renibacterium salmoninarum in the kidney/spleen tissue samples of A) 11 month old rainbow trout that s urvived an infection with live R. salmoninarum 16 weeks prior to being used in this experiment B) age cohort survived rainbow trout that have received two doses of R. salmoninarum bacterin per os at each sampling period (weeks post - infection with live bact eria). Renibacterium salmoninarum and its p57 antigen were detected by the sandwich enzyme - linked immunosorbent assay, with the infection intensity expressed as the proportion of fish exhibiting low, medium, and high levels of infection (n=5 per sampling period; n=44 - 45 at week 26 p.i.). The average number of R. salmoninarum colony forming units (CFUs) isolated from kidney is depicted as a line graph. 190 Figure 4.4 The mean antibody response (± SE) of A) 8 month old naïve rainbow trout ( Oncorhynchus myk iss ), and B) age cohort rainbow trout that have received two doses of a R. salmoninarum bacterin per os at each sampling period (weeks post - infection with live bacteria). The average optical density (OD) value was used to evaluate the production of anti - R. salmoninarum antibodies in fish. A separate group of naïve fish were used to determine the positive - negative threshold (dashed line), which was the average OD value plus two standard deviations (i.e., 0.094). 191 Figure 4.5. The mean antibody res ponse (± SE) of A) 11 month old rainbow trout that survived an infection with live R. salmoninarum 16 weeks prior to being used in this experiment B) age cohort survived rainbow trout that have received two doses of a R. salmoninarum bacterin per os at eac h sampling period (weeks post - infection with live bacteria). The average optical density (OD) value was used to evaluate the production of anti - R. salmoninarum antibodies in fish. A separate group of naïve fish were used to determine the positive - negativ e threshold, which was the average OD value plus two standard deviations (i.e., 0.094). 192 Figure 4.6. The mean circulating antibody levels of Chinook salmon ( Oncorhynchus tshawytscha ), steelhead ( O. mykiss ), and coho salmon ( O. kisutch ) returning to spawn at the Little Manistee River Weir and the Platte River Weir. The data are presented as box and whisker plots, where the central box contains the interquartile range and the median is represented as a horizontal bar, which divides the interquartile r The average optical density (OD) value was used to evaluate the production of anti - R. salmoninarum antibodies in fish. A separate group of naïve fish were used to determine the positive - negative threshold (dashed line), which was the average OD value plus two standard deviations (i.e., 0.094). 193 Figure 4 .7. The survival probability of 11 month old naïve rainbow trout ( Oncorhynchus mykiss ) (NPn) and age cohort rainbow trout that have received one dose of R. salmoninarum bacterin per os (NPi) in the weeks post - infection with live Renibacterium salmoninarum . 0.0 0.2 0.4 0.6 0.8 1.0 0 3 6 9 12 15 18 21 24 27 30 33 Survival Probability Weeks post - infection NPn NPi 194 R EFERENCES 195 R EFERENCES Alcorn, S. W., and Pascho, R.J. 2000. Single - dilution enzyme - linked immunosorbent assay for quantification of antigen - specific salmonid antibody. J. Vet. Diagn. Invest. 12: 245 - 252. Arkoosh, M.R., and Kaattari, S.L. 1990. Development of immunological memory in rainb ow trout ( Oncorhynchus mykiss ). I. An immunochemical and cellular analysis of the B cell response. Dev. Comp. Immunol. 15(4): 279 - 293. Austin, B., Embley, T.M., and Goodfellow, M. 1983. Selective isolation of Renibacterium salmoninarum . FEMS Microbiol Let t. 17: 111 - 114. Banner, C.R., Rohovec, J.S., and Fryer, J.L. 1983. Renibacterium salmoninarum as a cause of mortality among Chinook salmon in salt water. J. World Maric. Soc. 14: 236 - 239. Bartholomew, J. L., Arkoosh, M. R., and Rohovec, J. S. 1991. Demon stration of the specificity of the salmonid humoral response to Renibacterium salmoninarum with a monoclonal antibody against salmonid immunoglobulin. J. Aquat. Anim. Health 3 (4): 254 - 259. Brown, L.L., Iwama, G.K., Evelyn, T.P.T., Nelson, W.S., and Levine , R.P. 1994. Use of PCR to detect DNA from Renibacterium salmoninarum within individual salmon eggs. Dis. Aquat. Org. 18: 165 - 171. Chase, D.M., and Pascho, R.J. 1998. Development of a nested polymerase chain reaction for amplification of a sequence of the p57 gene of Renibacterium salmoninarum that provides a highly sensitive method for detection of the bacterium in salmonid kidney. Dis. Aquat. Org. 34: 223 - 229. Chase, D.M., Elliott, D.G., and Pascho, R.J. 2006. Detection and quantification of Renibacteri um salmoninarum DNA in salmonid tissues by real - time quantitative polymerase chain reaction analysis. J. Vet. Diagn. Invest. 18(4): 375 - 380. Ellis, R.W., Novotny, A.J., and Harrell, L.W. 1978. Case report of kidney disease in a wild Chinook salmon, Oncorh ynchus tshawytscha , in the sea. J. Wildl. Dis. 14: 121 - 123. Evelyn, T.P.T. 1971. The agglutinin response in sockeye salmon vaccinated intraperitoneally with a heat - killed preparation of the bacterium responsible for salmonid kidney disease. J. Wildl. Dis. 7:328 - 335. Evelyn, T.P.T. 1977. An improved growth medium for the kidney disease bacterium and some notes on using the medium. Bull. Off. Int. Epiz. 87: 511 - 513. 196 Faisal, M., Eissa, A.E., and Starliper, C.E.. 2010a. Recovery of Renibacterium salmoninarum from naturally infected salmonine stocks in Michigan using a modified culture protocol. J. Advanc. Res. 1: 95 - 102. Faisal, M., Loch, T.P., Brenden, T.O., Eissa, A.E., Ebener, M.P., Wright, G.M., and Jones, M.L. 2010b. Assessment of Renibacterium salmonin arum infections in four lake whitefish ( Coregonus clupeaformis ) stocks from northern Lakes Huron and Michigan. J. Great Lakes Res. 36: 29 - 37. Getchell, R.G., Rohovec, J.S., and Fryer, J.L. 1985. Comparison of Renibacterium salmoninarum isolates by antigen ic analysis. Fish Pathol. 20: 149 - 159. Grayson, T.H., Coopers, L.F., Wrathmell, A.B., Roper, J., Evenden, A.J., and Gilpin, M.L. 2002. Host responses to Renibacterium salmoninarum and specific components of the pathogen reveal the mechanisms of immune sup pression and activation. J. Immunol. 106: 273 - 283. Griffiths, S.G., Olivier, G., Fildes, J., and Lynch, W.H. 1991. Comparison of western blot, direct fluorescent antibody and drop - plate culture methods for the detection of Renibacterium salmoninarum in At lantic salmon ( Salmo salar L.). Aquacult. 97(2): 117 - 129. Hardie, L.J., Ellis, A.E., and Secombes, C.J. 1996. In vitro activation of rainbow trout macrophages stimulates inhibition of Renibacterium salmoninarum growth concomitant with augmented generation of respiratory burst products. Dis. Aquat. Org. 27: 175 - 183. Hnath, J.G., and Zischke, J. 1991. Fish health inspection history for Michigan state fish hatcheries and wild salmonid stocks for the period 1972 - 1991. Michigan Department of Natural Resources Fisheries Technical Report 91 - 13. Lansing, Michigan. Hsu, H. - M., Bowser, P.R., and Schachte Jr., J.H. 1991. Development and evaluation of a monoclonal - antibody - based enzyme - linked immunosorbent assay for the diagnosis of Renibacterium salmoninarum infecti on. J. Aquat. Anim. Health 3: 168 - 175. Jansson, E., and Ljungberg, O. 1998.. Detection of humoral antibodies to Renibacterium salmoninarum in rainbow trout Oncorhynchus mykiss and Atlantic salmon Salmo salar challenged by immersion and in naturally infect ed populations. Dis. Aquat. Org. 33 (2): 93 - 99. Jansson, E., Torbjörn, H., Höglund, J., and Ljungberg, O. 1996. Comparative evaluation of bacterial culture and two ELISA techniques for the detection of Renibacterium salmoninarum antigens in salmonid kidne y tissues. Dis. Aquat. Organ. 27: 197 - 206. Maclean, D.G., and Yoder, W.G. 1970. Kidney disease among Michigan salmon in 1967. Prog. Fish - Cult. 32 (1): 26 - 30. 197 Maule, A.G., Rondorf, D.W., Beeman, J., and Haner, P. 1996. Incidence of Renibacterium salmoninar um infections in juvenile hatchery spring Chinook salmon in the Columbia and Snake Rivers. J. Aquat. Anim. Health 8 (1): 37 - 46. Meyer, T.R., Thrower, F., Short, S., Lipson, K., Joyce, J., Farrington, C., and Doherty, S. 1999. Different prevalences of Renib acterium salmoninarum detected by ELISA in Alaskan Chinook salmon Oncorhynchus tshawytscha spawned from freshwater and seawater. Dis. Aquat. Org. 35: 101 - 105. Mitchum, D.L., Sherman, L.E., and Baxter, G.T. 1979. Bacterial kidney disease in feral populatio ns of brook trout ( Salvelinus fontinalis ), brown trout ( Salmo trutta ), and rainbow trout ( Salmo gairdneri ). J. Fish. Res. Board Can. 36: 1370 - 1376. Mitchum, D.L., and Sherman, L.E. 1981. Transmission of bacterial kidney disease from wild to stocked hatche ry trout. Can. J. Fish. Aquat. Sci. 38(5): 547 - 551. Munson, A.D., Elliott, D.G., and Johnson, K. 2010. Management of bacterial kidney disease in Chinook salmon hatcheries based on broodstock testing by enzyme - linked immunosorbent assay: a multilayer study . N. Am. J. Fish. Mana. 30(4): 940 - 955. Olivier, G., Griffiths, S.G., Fildes, J., and Lynch, W.H. 1992.. The use of Western blot and electroimmunotransfer blot assays to monitor bacterial kidney disease in experimentally challenged Atlantic salmon, Salmo salar L. J. Fish Dis 15(3): 229 - 241. Pascho, R.J., and Mulcahy, D. 1987. Enzyme - linked immunosorbent assay for a soluble antigen of Renibacterium salmoninarum , the causative agent of salmonid bacterial kidney disease. Can. J. Fish. Aquat. Sci. 44: 183 191 . Paterson, W.D., Gallant, C., Desautels, D., and Marshall, L. 1979. Detection of bacterial kidney disease in wild salmonids in the Margaree River system and adjacent waters using an indirect fluorescent antibody technique. J. Fish. Res. Board Can. 36: 14 64 - 1468. Purcell, M.K, Murray, A.L., Elz, A., Park, L.K., Marcquenski, S.V., Winton, J.R., Alcorn, S.W., Pascho, R.J., and Elliott, D.G. 2008. Decreased mortality of Lake Michigan Chinook salmon after bacterial kidney disease challenge: evidence for path ogen - driven selection? J. Aquat. Anim. Health 20(4): 225 - 235. Rockey, D. D., Gilkey, L. L., Wiens, G. D., and Kaattari, S. L. 1991. Monoclonal antibody - based analysis of the Renibacterium salmoninarum p57 protein in spawning chinook and coho salmon. J. Aq uat. Anim. Health 3 (1): 23 - 30. Secombes, C.J. 1985. The in vitro formation of teleost multinucleate giant cells. J. Fish Dis. 8: 461 - 464. 198 Wood, P.A., and Kaattari, S.L. 1996. Enhanced immunogenicity of Renibacterium salmoninarum in Chinook salmon after r emoval of the bacterial cell surface - associated 57 kDa protein. Dis. Aquat. Org. 25: 71 - 79. Young, C.L., and Chapman, G.B. 1978. Ultrastructural aspects of the causative agent and renal histopathology of bacterial kidney disease in brook trout ( Salvelinus fontinalis ). J. Fish. Res. Board Can. 35(9): 1234 - 1248. 199 Chapter 5 Efficacy of Current Testing Procedures of Spawning Chinook Salmon ( Oncorhynchus tshawytscha ) in Minimizing the Introduction of Renibac terium salmoninarum into Michigan Hatcheries 200 1. Abstract Bacterial kidney disease (BKD), caused by Renibacterium salmoninarum , has been associated with wide - scale mortalities in Chinook salmon ( Oncorhynchus mykiss ) from the Great Lakes basin, wh ich resulted in the adoption of stricter biosecurity practices at Michigan Department of Natural Resources (MDNR) gamete - collecting weirs. The purpose o f this study was to determine prevalence and intensity of R. salmoninarum in male and female Chinook sa lmon returning to spawn at four gamete - collecting weirs in Michigan: the Boardman River Weir, Little Manistee River Weir (LMRW), and Medusa Creek Weir (Lake Michigan watershed), and the Swan River Weir (Lake Huron watershed). Also, the extent to which inf ected fish could be shedding the bacterium within the gametes was determined, as well as if infected fish were exhibiting different diagnostic patterns indicative of a progressing R. salmoninarum infection. Nested polymerase chain reaction (nPCR), quantit ative enzyme - linked immunosorbent assay (Q - ELISA), and bacterial culture were used to assess prevalence of R. salmoninarum in a kidney and spleen homogenate and the gametes, while the presence of circulating antibodies was determined by agglutination. Amo ng the four weirs, SRW had the highest prevalence of R. salmoninarum , with female Chinook salmon appearing to be more susceptible to BKD infection than male Chinook salmon. Prevalence of R. salmoninarum in the gametes from both sexes was comparatively low , but there was evidence that males may contribute to shedding. Using the results from nPCR, Q - ELISA, culture, and agglutination, individuals were placed into one of six disease stages. Evidence of disease progression was observed at LMRW, with earlier d isease stages and more intense infections occurring later in the spawning run compared to earlier. It 201 is suggested that managers consider harvesting gametes used for propagation during the early spawning run, to reduce the influx of R. salmoninarum being introduced into hatcheries. 2. Introduction Renibacterium salmoninarum , the causative agent of bacterial kidney disease (BKD) in salmonids, is an obligate intracellular pathogen that is transmitted both horizontally (Mitchum and Sherman, 1981; Bell et al., 1984) and vertically (Evelyn et al., 1984, 1986) . In the 1980s, R. salmoninarum was associated with wide - scale mortalities in Chinook salmon ( Oncorhynchus mykiss ) from Lake Michigan (Holey et al., 1998). These unexpected BKD - epizootics highlighted t he lack of knowledge on the host defense mechanisms of Chinook salmon, particularly in its new habitat in the Great Lakes basin. The majority of knowledge regarding R. salmoninarum and Chinook salmon interactions has originated from studies performed in t he Pacific Northwest, where the fish alternate between marine and freshwater environments. However, in the Great Lakes basin, there is no access to a marine environment, thus resulting in different physiological conditions for Chinook salmon. To control BKD outbreaks and minimize the spread of R. salmoninarum in the Great Lakes basin, natural resources managers adopted a stringent procedure including a visual inspection of each spawning female and the subsequent culling of any gametes previously collecte d from spawning females with overt signs of clinical BKD. In addition, the kidneys of individual fish were routinely examined for the presence of R. salmoninarum soluble antigens using a monoclonal antibody based Field ELISA (FELISA). This rapid field te st allowed the 202 detection of fish whose kidneys were laden with R. salmoninarum antigen, resulting in the exclusion of their eggs from further incubation and propagation ( Beyerle and Hnath, 2002; Faisal and Hnath , 2005). The health inspection and the FELIS A were not performed on males due to the published studies of Klontz (1983) and Evelyn et al. (1986), which have minimized the role of male fish in vertical transmission of R. salmoninarum . However, more recently Eissa et al. (2007) demonstrated similar l evels of R. salmoninarum infection in milt and ovarian fluid from naturally infected brook trout and suggested that males could contribute to vertical transmission of BKD. Moreover, at certain gamete - collecting weirs in Michigan, Schulz (Chapter 2) report ed a higher incidence of R. salmoninarum infection in milt than ovarian fluid from naturally infected salmon, implying that males could be contributing to the vertical transmission of R. salmoninarum , although to what extent this is occurring is not known. Over the last three decades, a number of diagnostic assays have been developed to determine the presence of R. salmoninarum or its soluble antigens in infected fish tissues, such as culture on selective media, quantitative ELISA (Q - ELISA), and nested po lymerase chain reaction (nPCR) (Pascho and Elliott, 2004) . However, when multiple diagnostic tests are performed on the same sample, it is not uncommon for numerous discrepancies among findings to occur (Cipriano et al., 1985; Sakai et al., 1989; White et al., 1995; Jansson et al., 1996; Miriam et al., 1997; Pascho et al., 1998). This issue can often make findings, and the true infectious status of a fish, difficult to interpret. However, considering that each assay (culture, Q - ELISA, and nPCR) detects a different characteristic of the pathogen, the conflicting results could be related to the viability of R. salmoninarum . Recently, Faisal and Eissa (2009) analyzed the discrepancies among diagnostic assays in four salmonid stocks and suggested that the 203 ob served inconsistencies could be due to the phase of an R. salmoninarum infection at the time of sampling. For example, as nPCR detects the presence of bacterial DNA, a fish that was positive via nPCR only was suggested to be in an early infection (Faisal and Eissa, 2009). Conversely, Q - ELISA detects antigen released from the bacterium, thereby not requiring an active infection for detection; therefore, a fish that was positive by Q - ELISA only reflected a late, or recovering, infection (Faisal and Eissa, 2 009). Not only does this further the understanding of the pathogenicity of R. salmoninarum , but being aware of the current stage of infection occurring in a fish population could influence the control measures implemented by hatchery managers. The presen t study has been designed in order to better understand the epizootiology of R. salmoninarum collecting facilities and to determine how effective the current R. salmoninarum testing an d culling program is in minimizing the transmission of R. salmoninarum . Specifically, the objectives are 1) to determine the prevalence and intensity of R. salmoninarum in spawning xtent to which infected fish shed the bacterium with the gametes; and 3) to determine if infected fish are exhibiting different diagnostic patterns indicative of progressive R. salmoninarum infection. This novel approach will allow the elucidation of impo rtant aspects in the dynamics of R. salmoninarum infections in Chinook salmon, one of the most important fish species in the Great Lakes basin. 204 3. Material and Methods 3.1. Fish collection During September and October of 2005, returning Chinook salmo n spawners were collected from four gamete collecting weirs in Michigan: the Little Manistee River Weir (LMRW), the Medusa Creek Weir (MCW), and the Boardman River Weir (BRW), are all within the Lake Michigan watershed, while the Swan River Weir (SRW) resi ded within the Lake Huron watershed (Figure 5.1). Fish were collected from the LMRW on five separate occasions; September 20 (30 males, 30 females), October 4 (56 males, 30 females), October 5 (57 males, 22 females), October 6 (60 males, 44 females), and October 25 (55 males, 42 females). Fish were collected on October 14 from the BRW (30 males, 30 females) and the MCW (30 males, 30 females), while the MCW was also sampled on October 20 (36 males, 18 females). The SRW was sampled once on October 3 (30 ma les, 30 females). At the time of collection, the male Chinook salmon ranged in weight from 0.6 to 20 kg, with an average weight of 7.2 kg (± 3.8 kg) and ranged in length from 42 cm to 102.6 cm, with an average length of 75.2 cm (± 10.9 cm). Also, female Chinook salmon ranged in weight from 1.4 to 14.3 kg with an average weight of 6.4 kg (± 2.5 kg) and ranged in length from 53 to 92 cm, with an average length of 77.3 cm (± 7.0 cm). 205 3.2. Sample collection Prior to sample collection, Chinook salmon were euthanized by immersion in CO 2 - laden water and manual blunt force trauma to the cranium. Gametes were then harvested by Michigan Department of Natural Resources (MDNR) personnel for hatchery propagation, with a sub - sample collected in sterile 15 ml cent rifuge tubes (Denville Scientific Inc., Metuchen, NJ) for detection of R. salmoninarum with Q - ELISA (described below). Gamete samples were stored at - 20°C until processing. Blood was collected from the dorsal aorta at the caudal peduncle region and allo wed to clot in sterile tubes. Length, weight, and sex of all fish were recorded and external examinations were performed. Fish were externally disinfected with 70% ethanol and subjected to thorough internal clinical examinations and aseptic target tissue collection. Sterile individual forceps and scissors were used for each fish tissue collection. In addition to the gametes, samples (< 5 g) of kidneys and spleens were removed in the field and stored on ice in whirlpaks (VWR International, West Chester, PA), as recommended by the American Fisheries Society - Fish Health Section Bluebook (2012) and the World Organization for Animal Health (2012), and frozen at - 20°C until processing. Kidney and spleen homogenates were tested for the presence of R. salmonina rum by culture, nPCR, and Q - ELISA. 3.3. Bacterial culture and isolation Renibacterium salmoninarum isolation was performed by streaking 1 - 10 µl of the target tissue onto modified kidney disease medium (MKDM; Faisal et al., 2010) agar plates using 206 steril e disposable inoculating loops and incubating the plates under aerobic conditions at 15 ° C for a total of 42 days. The kidney/spleen samples were diluted 1:10 (weight/volume) in sterile phosphate buffered saline (PBS; pH 7.2) and suspended by repeated expu lsion through a sterile pipette. The suspension was then diluted via 10 - fold serial dilutions and 10 µ l of each dilution (e.g., undiluted to 10 - 8 ) was dispensed onto the MKDM plates. Plates were incubated at 15 ° C and were observed for typical R. salmoni narum growth every 7 days for a total of 6 weeks. Colonies were observed under a dissecting microscope for typical R. salmoninarum morphological characteristics: convex, cream - colored, round, and smooth (Austin and Austin, 2007). Colonies that fit this c riterion were assessed by additional biochemical testing, which included Gram stain, cytochrome oxidase, and catalase reaction, which are key biochemical tests used to identify R. salmoninarum (Sanders and Fryer, 1980). Colonies that were Gram positive, c ytochrome oxidase negative, and catalase positive were then confirmed molecularly via nPCR (see below). 3.4. Extraction of DNA The DNA extracted from the kidney and spleen samples was processed according to the DNeasy® Blood and Tissue Kit (Qiagen) ins tructions for animal tissue. Also, as recommended by the kit instructions, the pre - treatment for Gram - positive bacteria (including the lysis buffer) was applied to all of the samples. After extraction, the DNA was then quantified with the Qubit ® Fluorome ter (Life Technologies, Grand Island, NY) and diluted to a 20 ng/µl concentration. 207 3.5. Nested PCR The nPCR method and primers recommended by Pascho et al. (1998) were used initially, with minor modifications. To optimize the protocol to the laborato annealing temperature was changed to 60 ° C for DNA extracted from pure bacterial cultures and kidney/spleen tissues (based upon optimization experiments). Additionally, the total reaction volume was reduced from 50 µ l to 25 µ l, consist ing of 1 µ l each of template DNA (20 ng total), forward primer (10 µmol), and reverse primer (10 µmol), as well as 12.5 µ l of GoTaq® Green Master Mix (Promega Corp., Madison, WI) and 9.5 µ l of nuclease - free water. The controls were composed of a PCR mixtu re containing water instead of DNA template (negative control) and DNA from a pure culture of R. salmoninarum ATCC #33209 strain (positive control). Five microliters of the nPCR products and controls were mixed with 1 µ l of SYBR Green II RNA Gel Stain (Ca mbrex Bio Science Rockland, Inc., East Rutherford, NJ) and loaded into a gel consisting of 2% Ultra Pure TM agarose (Invitrogen, Grand Island, NY). Each electrophoresis gel included 5 µ l of a 1 Kb plus ladder (Invitrogen) mixed with 1 µ l of 6X gel loading dye (New England Bio Labs, Inc., Ipswich, MA). Gels were run at 100 v for 35 minutes in 1X Tris - Acetate Buffer (Sigma - Aldrich Corp., St. Louis, MO) in a Gel XL Ultra V - 2 electrophoresis box (Labnet) and a Sub - Cell GT electrophoresis box (Bio Rad Laborator ies, Inc., Hercules, CA). Gels were visualized with a Canon G10 camera and UV Trans - Illuminator. Samples were considered positive for R. salmoninarum when a 320 - bp band was present (Pascho et al., 1998). 208 3.6. Q - ELISA The general Q - ELISA protocol outl ined in Pascho and Mulcahy (1987), with modifications recommended by Gudmundsdóttir et al. (1993) and Olea et al. (1993), was used to assess R. salmoninarum antigens in the kidney/spleen samples. Prior to the Q - ELISA procedure, kidney/spleen tissue sample s from the immersion and injection challenge were diluted 1:8 naturally infected fish were diluted 1:4 with HBSS. The samples were then stomached on high speed for tw o minutes with the Biomaster Stomacher (Wolf Laboratories Limited, Pocklington, York, UK). An aliquot of 250 µl of stomached kidney/spleen samples were dispensed into 1.5ml microcentrifuge tubes containing 250 µl of phosphate buffered saline with Tween - 20 (PBS - T20; Sigma) and 5% goat serum (Sigma) and 50 µl of CitriSolv (Fisher Scientific, Pittsburgh, PA). The purpose of the CitriSolv solvent was to dissolve and remove liquids form the aqueous supernatant (Gudmundsdóttir et al., 1993), while the introduct ion of 5% goat serum was to increase sensitivity of the assay (Olea et al., 1993). Samples were vortexed for approximately 10 sec, heated at 100°C for 15 min, and then centrifuged at 14,000 rpm for 10 min. The aqueous supernatant of each sample was used fo r Q - ELISA testing. The positive - negative cutoff absorbance for the samples was 0.10. Samples that tested positive were assigned the following antigen levels: low (0.10 - 0.199), medium (0.20 - 0.999), and 998). Each assay included two negative controls, a negative fish tissue sample and a dilution buffer, and two 209 positive controls, a positive fish tissue sample and a R. salmoninarum positive control (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD ). 3.7. Agglutination assay The agglutination assay to detect and quantify Chinook salmon circulating antibodies against R. salmoninarum was performed according to the protocol originally described by Evelyn (1971), with few modifications. The antigen w as prepared from cryo - preserved R. salmoninarum (ATCC #33209) that was revived on MKDM. Following 5 - 7 days of incubation at 15°C, bacterial growth was centrifuged at 3000 rpm for 10 min, washed three times with sterile 0.9% NaCl saline. Washed bacteria w ere then re - suspended in saline at a concentration of 50 mg ml - 1 . Bacteria were then killed by heating the suspension to 62°C for 45 min in a water - bath. Standard two - fold serial dilutions of serum were made using sterile saline as the diluent. The ag glutinin test was performed in tubes with the heat - killed R. salmoninarum as the antigen and was adjusted to a concentration yielding an optical density of 1.25 at 420 mµ. A positive reaction was indicated by the formation of macroscopic clumping of the a ntigen following incubation with serum samples overnight at 4°C. A titer (expressed as the reciprocal of the highest serum dilution showing positive reaction) was determined for each tested serum sample. Sera showing no agglutination at a dilution of 1:2 were considered negative. 210 3.8. Statistical analys e s To assess potential shedding, individuals were separated into one of four infection categories based upon Q - ELISA results from the kidney and spleen homogenate and gamete sample: (1) individuals that were negative for R. salmoninarum in the kidney and spleen homogenate and the gamete sample (ks - /gam - ); (2) individuals that were positive for R. salmoninarum in the kidney and spleen homogenate and negative in the gamete sample (ks+/g - ); (3) individuals that were negative for R. salmoninarum in the kidney and spleen homogenate and positive in the gamete sample (ks - /gam+); and (4) individuals that were positive for R. salmoninarum in the kidney and spleen homogenate and the gamete sample (ks+/gam+). Indivi duals that were positive for R. salmoninarum in the gamete sample were considered to be potential shedders of the bacteria. To explore potential disease progression, samples were placed into one of six disease stages (Table 5.1) depending on the culture, Q - ELISA, nPCR, and agglutination results, based on the study of Faisal and Eissa (2009). Disease stage scores (1 - 6) were analyzed using a general linear model ( GLM ) based on a cumulative logistic link function. The GLM included the effects of period and gender and their interaction, as well as weight; other covariates such as length and condition factor were investigated but they were not statistically significant due to the high collinearity with weight. Estimated differences between different groups we re reported as odds ratios of being in lower numbered categories. SAS PROC GLIMMIX and SAS statistical software (SAS Institute, Inc., Cary, NC) was used for data analysis. Analyses of probabilities of co - infection in kidney and gametes and associations b etween the various disease stage 211 categories and the presence or absence of R. salmoninarum in the gametes were further investigated using contingency chi - disease stage categories were small. 4. Res ults 4.1. Prevalence and semi - quantitation of R. salmoninarum by Q - ELISA The Q - ELISA was performed on CHS returning to all four weirs (BRW, LMRW, MCW, and SRW). The prevalence of R. salmoninarum detected from the combined four weirs was 6.5% (43/660 fish). While not statistically significant (F = 1.71; df = 6, 654; P = 0.1161), SRW exhibited highest prevalence of R. salmoninarum at 10%, followed by MCW at 7%, BRW at 6.7%, and LMRW at 5.9% (Figure 5.2A). The highest intensity of an R. salmoninarum inf ection was observed in Chinook salmon collected from BRW, followed by LMRW, SRW, and MCW (Figure 5.2A). The combined R. salmoninarum prevalence in females from all four weirs was nearly double that of males (8.7% vs. 4.95%); however, this was not a signif icant difference in prevalence (F = 2.58; df = 1, 654; P = 0.1089). More specifically, female Chinook salmon from LMRW and MCW exhibited a higher prevalence of R. salmoninarum than males (Figure 5.2B). Conversely, male Chinook salmon collected from BRW a nd SRW were found to demonstrate a higher prevalence of R. salmoninarum than females (Figure 5.2B). In order to ascertain the extent to which infected fish were potentially shedding R. salmoninarum along with the gametes, Q - ELISA was also performed on o varian fluid and milt 212 from spawning Chinook salmon returning to LMRW, MCW, and SRW (Table 5.2, Figure 5.3A). The overall prevalence of R. salmoninarum detected in gametes from the three weirs was 4.3% (23/540 fish). By location, potential shedding was mo st prevalent in Chinook salmon from SRW (6.7%), followed by LMRW (4.0%) and MCW (3.7%) (Figure 5.3A). Overall, potential shedding within the gametes of females (6.0%) was twice as high in prevalence as males (3.1%), although this was not statistically sig 2 = 2.73; df = 1; P = 0.098). The increased prevalence of R. salmoninarum in females was likely due to fish collected from LMRW and MCW, where females had a higher prevalence of R. salmoninarum in their gametes (Figure 5.3B). To gain a more Q - ELISA results from the kidney and spleen sample and the gamete sample from LMRW, MCW, and SRW Chinook salmon were combined and placed into one of the four groups previously describ ed (Figure 5.4). The majority of the fish tested were not shedding the bacterium, with R. salmoninarum not detected in the gametes (kid+/gam - ) or in either sample (kid - /gam - ) (Figure 5.4). There was also no discernible difference in shedding between the male and female Chinook salmon (Figure 5.4). 4.2. Detection of R. salmoninarum at LMRW Little Manistee River Weir is the main gamete collection site for the MDNR and therefore was chosen for additional analysis. 213 4.2.1. Results of Q - ELISA For this assay, Chinook salmon were collected at three time periods: before the typical - y increase in the late run as opposed to the early run (Table 5.2, Figure 5.5A), which was also statistically 2 = 6.67 ; df = 1; P = 0.01). Additionally, the female Chinook salmon had a significantly higher prevalence of R. salmoninarum detec ted than the male individuals in the 2 = 6.86 ; df = 1; P = 0.009) (Figure 5.5B). Comparisons of the prevalence of an R. salmoninarum infection within the gametes of Chinook salmon from LMRW showed a pronounced increase in the later runs compar ed to earlier runs (Table 5.2; Figure 5.6A). More specifically, the observed increase in prevalence 2 = 6.40 ; df = 1; P = 0.011) (Figure 5.6A), which was due the increased prevalence of R. salm oninarum detected in the milt from 2 = 4.30 ; df = 1; P = 0.038) (Figure 5.6B). Generally, the prevalence and intensity of the R. salmoninarum infection was more prominent in the female individuals, but this trend was not statistically sig 2 = 3.16 ; df = 1; P = 0.076) (Figure 5.6B). 4.2.2. Results of nPCR Nested PCR was performed on early and late run samples only. The detection of R. salmoninarum in kidney and spleen homogenates via nPCR varied significantly according to 214 time of sampling (early run vs. late run) and sex. Overall, R. salmoninarum was detected from 23.2% of the fish sampled (85/366 fish), with 31.2% of the females (43/138 fish) and 18.4% of the males (42/228 fish) being infected (Table 5.2). Moreover, nPCR det ected R. salmoninarum at a significantly higher rate in the late run Chinook salmon compared to the early run 2 = 16.48 ; df = 1; P < 0.001) (Table 5.2). Additionally, nPCR detected a higher prevalence of R. salmoninarum in the kidneys of LMRW females than males, specifically in the 2 = 10.76 ; df = 1; P = 0.001) (Table 5.2). 4.2.3. Results of bacterial culture Bacterial culture was performed on early and late run samples only. Culture revealed the presence of R. salmoninarum in 11.9% of male fish (20/168 fish) and 12.9% of female fish (12/93 fish), for an overall prevalence of 12.3% (32/261 fish; Table 5.2). Colony forming units (CFUs) varied from 3 - 256 CFU g - 1 of kidney tissues in early run fish (mean = 2.7 CFU g - 1 ) to 5 - 912 CFU g - 1 in the late run fish (mean = 30 CFU g - 1 ). Over all, the prevalence of R. salmoninarum detected by culture was significantly higher in the late run Chinook salmon compared to the 2 = 16.03 ; df = 1; P < 0.001), which was due to the higher detection of R. salmoninarum in the male fish than fem 2 = 14.31 ; df = 1; P < 0.001) (Table 5.2). 215 4.2.4. Detection of circulating antibodies against R. salmoninarum by agglutination Agglutination was performed on early and late season samples only. The analysis revealed that ~40% of LMRW Chin ook salmon were producing antibodies (105/261 fish; Table 5.2). There was a larger number of early run Chinook salmon with measurable antibodies as 2 = 18.93 ; df = 1; P < 0.001) (Table 5.2). Additionally, male fish produced sign 2 = 4.92 ; df = 1; P = 0.027), with titers varying from 2 3 to 2 11 (Table 5.2). Of the early run Chinook salmon, the number of males with circulating antibodies were higher compared to their females c 2 = 4.26; df = 1; P = 0.039) (Table 5.2). 4.3. Disease stage analysis Grouping of CHS into potential stages of infection was performed on infected fish for which the agglutination, culture, nPCR, and Q - ELISA assays were performed, expanding o n the observations of Faisal and Eissa (2009). Fish were grouped into six stages based on positive results from the assays: Stage 1 represented the onset of an infection (nPCR+, culture - , Q - ELISA - , agglutination - ), Stage 2 signified a settled infection ( nPCR+, culture+, Q - ELISA+/ - , agglutination - ), individuals in Stage 3 had an active infection (nPCR+, culture+, Q - ELISA+, agglutination - ), Stage 4 represented the beginning of remission (nPCR+, culture - , Q - ELISA+, agglutination - ), individuals in Stage 5 wer e in advanced remission (nPCR - , culture - , Q - ELISA+, 216 agglutination+), and Stage 6 signified fish were recovering from an infection (nPCR - , culture - , Q - ELISA - , agglutination+) (Table 5.1). Chinook salmon from LMRW collected during the early and late run e xhibited different patterns of disease stages (Figure 5.7A). Early run Chinook salmon included more individual fish that possessed circulating antibodies against R. salmoninarum (Stage 6) and less in all other infection stages as compared to those individ uals in the late run (Figure 5.7A). Moreover, there was a substantial increase in the number of individuals in Stages 1 - 4 from the early to the late sampling period (Figure 5.7A). These observations were supported by the logistic regression analysis whic h revealed that the incidence of Chinook salmon at different stages of disease varied significantly as a function of the sampling period (F = 4.2; df = 1, 152; P = 0.042) and sex (F = 5.92; df = 1, 152; P = 0.016). There was also a significant interaction between the sex of Chinook salmon and the collection period (F = 6.22; df = 1, 152; P = 0.014). More specifically, female fish in the early sampling period were more likely to occur in higher disease stages than female fish in the late sampling period, w ith an odds ratio of 0.1437 (t = - 2.93; df = 152; P = 0.004) (Figure 5.7B). On the contrary, there was no difference in the occurrence of males in disease stages from early to late sampling periods significant period effect for males observed (t = - 0.44; df = 152; P = 0.663) (Figure 5.7C). In the early period, males were 5.14 times more likely to occur in a lower disease stage compared to females (t = 3.49; df = 152; P = 0.0006). However, no significant sex difference was observed in fish collected in th e late period (t = - 0.08; df = 152; P = 0.934). Most interestingly, analysis revealed that shedding of R. salmoninarum occurs primar ily from fish in Stage 3 . 217 Interestingly, when the non - infected fish (nPCR - , culture - , Q - ELISA - , agglutination - ) were incl uded in the analysis as an additional stage, there was no significant effect of sampling period, gender, or their interaction on disease stage (0.17 < p < 0.66). 5. Discussion Renibacterium salmoninarum was detected from Chinook salmon collected from all of the weirs, although the intensity of the infection depended upon location, as well as the sex of the fish. Interestingly, the highest overall prevalence of R. salmoninarum was documented at SRW in the Lake Huron watershed; yet it has been shown tha t in the past decade, SRW generally has a lower rate of infection of R. salmoninarum than Lake Michigan weirs (Chapter 2). Starting in 2002, the MDNR adopted enhanced biosecurity measures, which have severely reduced the intensity of R. salmoninarum infec tions in state - run hatcheries and at the gamete - collecting weirs (Chapter 2). It is possible that MDNR - raised fingerlings stocked into the Lake Huron watershed in previous years had a higher rate of infection, which did not diminish while the Chinook salm on matured in Lake Huron and was maintained upon returning to spawn. Furthermore, R. salmoninarum was detected in the kidney and spleen sample of female Chinook salmon at nearly double the rate of male Chinook salmon. It has been suggested that female fi sh play a greater role in the vertical transmission of R. salmoninarum than males (Evelyn et al., 1984; Evelyn et al., 1986; Bruno and Munro, 1986), but it is possible that female Chinook salmon are more susceptible to R. salmoninarum infection than males. Female Chinook salmon are more likely to be stressed from the spawning experience than males 218 (Carruth et al., 2000), which could contribute to a higher susceptibility to pathogens (e.g., R. salmoninarum ) than males. Gamete production, migration, and spa wning requires most of the energy stored in fish, thus weakening their immune system and making them more susceptible to pathogens. To fully examine this possibility, a sex susceptibility study of Chinook salmon to R. salmoninarum would need to be perform ed. Interestingly, most of the Chinook salmon tested were not shedding R. salmoninarum in their gametes and it was not found in their kidneys and spleen either. Renibacterium salmoninarum was detected from both the kidneys and spleen and gametes of few f ish, which were determined to be the potential shedders. Most of the fish shedding the bacterium were also experiencing an active infection (Stage 3), with viable R. salmoninarum , DNA, and antigen being detected. Fish occurring in Stage 3 were the most h eavily infected individuals, thus increasing the likelihood that they would shed the bacterium. Due to the heavy infection, Chinook salmon in Stage 3 were also more likely to die, which would contribute to the low number of fish shedding R. salmoninarum . Most surprisingly was that a small number of fish were shedding R. salmoninarum in their gametes, yet their kidneys and spleen were negative. It is possible that these fish have recovered systemically, yet remnants of R. salmoninarum , or its soluble anti gens, still remained in the milt or ovarian fluid. Unlike the kidney and spleen samples, R. salmoninarum was detected at fairly equal rates in the gametes from male and female Chinook salmon, suggesting that while female fish may be more susceptible to an R. salmoninarum infection, they may play similar roles in shedding and vertical transmission of the bacterium. It has long been known that female fish transmit R. salmoninarum vertically to their offspring, but this brings to light 219 that male fish may hav e a greater role than previously suggested. While other studies have recorded that male fish do not significantly contribute to vertical transmission (Klontz, 1983; Evelyn et al. 1986; Pascho and Elliott, 2004), more than 3% of males in this study were ca pable of shedding the bacterium in their milt, which has also been observed by our earlier studies (Eissa et al., 2007; Chapter 2). Therefore, it is recommended that the role of males in shedding be further evaluated and explored. Additionally, the great est potential for shedding of R. salmoninarum in gametes occurred at SRW. This is likely due to Chinook salmon from SRW also having the highest rate of R. salmoninarum infection in their kidneys and spleens. At SRW, most of the fish had concordance regar ding the R. salmoninarum infection in their gametes and their kidneys and spleen. Lastly, this study expands on the suggestion of Faisal and Eissa (2009) that discrepancy among diagnostic assay results is associated with the stage of infection that fish a re experiencing. Using nPCR, bacterial culture, and Q - ELISA, the authors proposed that fish could be placed into one of six disease stages. Stage 1 was an initial i nfection ( nPCR+, culture , Q - ELISA ), implying that there was a minimal amount of bacteria in the tissues collected, and only the DNA of R. salmoninarum was detected. Stage 2 occurred once the bacteria was viable and numerous enough to be cultured on selective medium, and has become established ( nPCR+, culture+, Q - ELISA ). Stage 3 occurred on ce the infection progressed to a systemic, well - established infection ( nPCR+, culture+, Q - ELISA+ ). Stage 4 implied that a fish host has started to recover from an infection, with only bacterial DNA and antigen still remaining ( nPCR+, culture , Q - ELISA+ ). Stage 5 occurred when only low levels of bacterial antigen remained and as a result, the fish was experiencing an advanced stage of recovery ( nPCR , culture , Q - ELISA+ ). Finally, 220 Stage 6 occurred in fish that have never been exposed to R. salmoninarum , w ere resistant to infection, or have eliminated the infection from their systems ( nPCR , culture , Q - ELISA ). While Faisal and Eissa (2009) laid the groundwork with this novel explanation of discrepancies among assays, this study expands on the concept by incorporating a fourth assay, agglutination. The addition of the agglutination assay allows for researchers to gain a greater understanding of the stages of infection, as it detects the production of circulating antibodies in a fish host. A higher amount of antibodies being produced by a fish host is indicative of a recently occurring or cleared infection, while lower amounts of antibodies implies that a fish has not been recently infected. Additionally, it was decided to exclude non - infected fish from t his analysis. At the time of sampling in 2005, the prevalence of BKD was declining in salmonid species returning to spawn at several Lake Michigan weirs, from a prevalence of 100% and 82% in 2001 and 2002 respectively, to 33% in 2005 (Chapter 2). As it w as expected that several of the fish would not be infected due to the declining prevalence, they would not provide worthwhile information to determining the stages of infection and were omitted. Moreover, to investigate potential disease progression o ver time, the present study also assessed the occurrence of disease stages in LMRW Chinook salmon at two time points. A general trend was observed of earlier disease stages, and more intense infections, occurring in the late run Chinook salmon compared to the early run. The majority of the early run fish were producing circulating antibodies and did not have an active R. salmoninarum infection. Furthermore, microbiological culture revealed that the intensity of the infections from early run fish was also less, as demonstrated by a lower average number of CFUs recovered from the early run Chinook salmon. On the contrary, the higher average number of CFUs from the late 221 run fish, along with a greater detection of R. salmoninarum antigen and DNA, indicates a more intense, active infection. It is possible that the early run Chinook salmon were healthier than fish in the later run, enabling them to return to spawn earlier. Chinook salmon that returned to spawn in the late sampling period exhibited signs of in creased settled and active infections, which could have delayed their spawning migration. This is not entirely unexpected as fish are likely to become more stressed and susceptible to disease as their spawning run continues. Alternatively, the individu als that returned to spawn during the early sampling period may have had a genetic advantage. While salmon have evolved to spawn at times that are ideal based on temperature regimes and other environmental factors, it has also been suggested that salmon a re spawning at earlier times as an indirect result of hatchery practices (Quinn et al., 2002). Factors that would typically select against early spawning are relaxed in a hatchery environment, where there is ample food availability and protection from pre dators (Quinn et al., 2002). Moreover, offspring from late spawners might not reach the required size in an appropriate period of time, resulting in selective culling or low survival rates after stocked (Quinn et al., 2002). Thus, the early run LMRW Chin ook salmon may have a better fitness than the late run individuals, resulting in a heightened resistance to disease as well. As the LMRW is the main source of gametes for state fish hatcheries in Michigan, it is evident that the time of gamete harvest can contribute to the R. salmoninarum load in donor fish, as well as in the potential shedding. Agencies should consider the time period at which gametes are harvested, as this can affect the rate of R. salmoninarum infection being introduced into the hatche ry system. Additionally, findings from this study suggests that using a single diagnostic assay to detect, and consequently cull, spawning CHS is not enough since 222 each of these assays measures a different target of the bacterium; however, performing all f our assays is an important way to evaluate the progress of R. salmoninarum infection in a particular stock. 223 A PPENDIX 224 Table 5.1. Diagnostic disease stages were determined based on R. salmoninarum recovered from a kidney and spleen homogenate and the presence of circulating antibodies as detected by agglutination. Nested polymerase chain reaction (nPCR), culture, semi - quantitative enzyme - linked immunosorbent assay (Q - ELISA), and agglutination (Agglut.) were used to determine disease stages. Disease Stage Methods of Detection 1 Potential explanation nPCR Culture Q - ELISA Agglut. 1 P N N N Onset of infection 2 P P P/N N Settled infection 3 P P P N Active infection and potentia l mortality 4 P N P (low) N Beginning of remission 5 N N P P Advanced remission 6 N N N P Recovery 1 P and N refer to positive and negative evidence for infection detected. 225 Table 5.2. The prevalence of Renibacterium salmoninarum detected in th e blood, a kidney and spleen homogenate, and reproductive fluids from male ( ) and female ( ) Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at the Little Manistee River Weir. Detection of R. salmoninarum as circulating antibodies in the blo od was determined by an agglutination assay (Agglut.), while the prevalence of R. salmoninarum in the kidney and spleen homogenate (k/s) was detected by bacterial culture, nested polymerase chain reaction (nPCR), and semi - quantitative enzyme - linked immunos orbent assay (Q - ELISA). Occurrence of R. salmoninarum in the reproductive fluids (gam) was also determined by Q - ELISA. ND = not done. Assay Male Female Combined Pre - Season Early Late Total Pre - Season Early Late Total Pre - Season Early Late Overall Agg lut. ND 55.8% (63/113) 23.6% (13/55) 45.2% (76/168) ND 38.5% (20/52) 22% (9/41) 31.2% (29/93) ND 50.3% (83/165) 22.9% (22/96) 40.2% (105/261) Culture ND 5.3% (6/113) 25.5% (14/55) 11.9% (20/168) ND 7.7% (4/52) 19.5% (8/41) 12.9% (12/93) ND 6.1% (10/165) 2 2.9% (22/96) 12.3% (32/261) nPCR ND 12.1% (21/173) 38.2% (21/55) 18.4% (42/228) ND 28.1% (27/96) 38.1% (16/42) 31.2% (43/138) ND 17.8% (48/269) 38.1% (37/97) 23.2% (85/366) Q - ELISA (k/s) 0% (0/30) 1.7% (3/173) 9.1% (5/55) 3.5% (8/228) 10% (3/30) 8.3% (8/ 96) 14.3% (6/42) 10.1% (14/138) 5% (3/60) 4.1% (11/269) 11.3% (11/97) 6.0% (22/366) Q - ELISA (gam) 0% (0/30) 1.7% (3/173) 7.3% (4/55) 3.0% (7/228) 0% (0/30) 5.2% (5/96) 11.9% (5/42) 7.2% (10/138) 0% (0/60) 3.0% (8/269) 9.3% (9/97) 4.6% (17/366) 226 Fi gure 5.1. The Michigan Department of Natural Resources gamete - collecting weirs where Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn were collected in September and October of 2005: Little Manistee River Weir, Medusa Creek Weir, and Boardman R iver Weir (Lake Michigan watershed), and Swan River Weir (Lake Huron watershed). 227 Figure 5.2. The prevalence and intensity (high, medium, low) of Renibacterium salmoninarum detected by the quantitative enzyme - linked immunosorbent assay in a kidney and spleen homogenate from A) all Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at Little Manistee River Weir (LMRW), Medusa Creek Weir (MCW), Boardman River Weir (BRW), and the Swan River Weir (SRW), and B) male (M) and female (F) Ch inook salmon returning to spawn at LMRW, MCW, BRW, and SRW. 228 Figure 5.3. The prevalence and intensity (high, medium, low) of Renibacterium salmoninarum detected by the quantitative enzyme - linked immunosorbent assay in reproductive fluids from A) all Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at Little Manistee River Weir (LMRW), Medusa Creek Weir (MCW), and the Swan River Weir (SRW), and B) male (M) and female (F) Chinook salmon returning to spawn at LMRW, MCW, and SRW. 229 Figure 5.4. The proportion of male and female Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at the Little Manistee River Weir (LMRW), Medusa Creek Weir (MCW), and Swan River Weir (SRW) that may be shedding Renibacterium salmoninarum in thei r reproductive fluids (kid+/gam+, kid - /gam+) and are not shedding R. salmoninarum (kid - /gam - , kid+/gam - ), as detected by the quantitative enzyme - linked immunosorbent assay. 0% 20% 40% 60% 80% 100% Male Female Male Female Male Female LMRW MCW SRW Proportion Location kid+/gam+ kid-/gam+ kid-/gam- kid+/gam- 230 Figure 5.5. The prevalence and intensity (high, medium, low ) of Renibacterium salmoninarum detected by the quantitative enzyme - linked immunosorbent assay in a kidney and spleen homogenate from A) all Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at the Little Manistee River Weir during the pre - seaso n, early, and late sampling periods, and B) male (M) and female (F) Chinook salmon during the same sampling periods. 231 Figure 5.6. The prevalence and intensity (high, medium, low) of Renibacterium salmoninarum detected by the quantitative enzym e - linked immunosorbent assay in reproductive fluids from A) all Chinook salmon ( Oncorhynchus tshawytscha ) returning to spawn at the Little Manistee River Weir during the pre - season, early, and late sampling periods, and B) male (M) and female (F) Chinook s almon during the same sampling periods. 232 Figure 5.7. Proportion of Chinook salmon ( Oncorhynchus tshawytscha ) collected from Little Manistee River Weir in each disease stage (1 - 6) during the early and late sampling periods, including A) all Chinook salmo n, B) female Chinook salmon, and C) male Chinook salmon. Disease stages were determined based on results from nested polymerase chain reaction (nPCR), bacterial culture, semi - quantitative enzyme - linked immunosorbent assay (Q - ELISA), and agglutination (Agg lut). Stage 1 represents the onset of infection (nPCR+, culture - , Q - ELISA - , Agglut - ), stage 2 represents a settled infection (nPCR+, culture+, Q - ELISA+/ - , Agglut - ), stage 3 represents an active infection (nPCR+, culture+, Q - ELISA+, Agglut - ), stage 4 repre sents the beginning of remission (nPCR+, culture - , Q - ELISA+, Agglut - ), stage 5 represents advanced remission (nPCR - , culture - , Q - ELISA+, Agglut+), and stage 6 represents recovering from an infection (nPCR - , culture - , Q - ELISA - , Agglut+). 233 R EFERENCES 234 R EFERENCES American Fisheries Society - Fish Health Section. 2012. FHS blue book: suggested procedures for the detection and identification of certain finfish and shellfish pathogens, 2010 edition. AFS - FHS, Bethesda, MA . Austin, B., and Austin, D.A. 2007. Bacterial Fish Pathogens, Diseases in Farmed and Wild Fish. 4 th Edition. Springer - Praxis, Chichester, United Kingdom. Bell, G.R., Higgs, D.A., and Traxler, G.S. 1984. The effect of dietary ascorbate, z i nc and manganes e on the development of experimentally induced bacterial kidney disease in sockeye salmon ( Oncorhynchus nerka ). Aquaculture 36: 293 - 301. Beyerle, J., and Hnath, J.G. 2002. History of fish health inspections, State of Michigan, 1970 - 1999. Michigan DNR Fis heries Technical Report 2002 - 2. Bruno, D.W., and Munro, A.L.S. 1986. Observation on Renibacterium salmoninarum and the salmonid egg. Dis. Aquat. Org. 1: 83 - 87. Carruth, L.L., Dores, R.M., Maldonado, T.A., Norris, D.O., Ruth, T., and Jones, R.E. 2000. Ele vation of plasma cortisol during the spawning migration of landlocked kokanee salmon ( Oncorhynchus nerka kennerlyi ). Comp. Biochem. Physiol. C. 127: 123 - 131. Cipriano, R.C., Starliper, C.E., and Schachte, J.H. 1985. Comparative sensitivities of diagnostic procedures used to detect bacterial kidney disease in salmonid fishes. J. Wildl. Dis. 21(2): 144 - 148. Eissa, A.E., Elsayed, E.E., and Faisal, M. 2007. Prevalence and shedding of Renibacterium salmoninarum in brook trout ( Salvelinus fontinalis ) in Michiga n. Nat. Sci. 5(1): 9 - 17. Evelyn, T.P.T. 1971. The agglutinin response in sockeye salmon vaccinated intraperitoneally with a heat - killed preparation of the bacterium responsible for salmonid kidney disease. J. Wildlife Dis. 7: 328 - 335. Evelyn, T. P. T., Ketcheson, J.E., and Prosperi - Porta, L. 1984. Further evidence for the presence of Renibacterium salmoninarum in salmonid eg gs and for the failure of povido ne - iodine to reduce the intra - ovum infection in water - hardened eggs. J. Fish Dis. 7: 173 - 182 Evely n, T.P.T., Prosperi - Porta, L., and Ketcheson, J.E. 1986. Experimental intra - ovum infection of salmonid eggs with Renibacterium salmoninarum and vertical transmission of the pathogen with such eggs despite their treatment with erythromycin. Dis. Aquat. Org. 1:197 - 202. 235 Faisal, M., and Eissa, A.E. 2009. Diagnostic testing patterns of Renibacterium salmoninarum in spawning salmonid stocks in Michigan. J. Wildl. Dis. 45(2): 447 - 456. Faisal, M., Eissa, A.E., and Starliper, C.E. 2010. Recovery of Renibacterium sa lmoninarum from naturally infected salmonine stocks in Michigan using a modified culture protocol. J. Advanc. Res. 1: 95 - 102. Faisal, M., and Hnath, J.G. 2005. Fish health and diseases issues in the Laurentian Great Lakes. In : R. C. Cipriano, I. S. Shchel kunov and M. Faisal (E ds.). Health and diseases of aquatic organisms: Bilateral perspectives . Michigan State University Press, East Lansing, Michigan, pp. 331 350. Gudmundsdóttir, S., Benediktsdóttir, E., and Helgason, S. 1993. Detection of Renibacterium salmoninarum ELISA and isolation on selective medium. J. Fish Dis. 16(3): 185 - 195. Holey, M.E., Elliot, R.F., Marcquenski, S.V., Hnath, J.G., and Smith, K.D. 1998. Chinook salmon epizootics in Lake Michigan: possible contributing factors and management implications. J. Aquat. Anim. Health 10: 202 - 210. Jansson, E., Torbjörn, H., Höglund, J., and Ljungberg, O. 1996. Comparative evaluation of bacterial culture and two ELISA techniques for the dete ction of Renibacterium salmoninarum antigens in salmonid kidney tissues. Dis. Aquat. Org. 27: 197 - 206. Klontz, G.W. 1983. Bacterial kidney disease in salmonids: an overview. In: Anderson, D.P., Dorson, M., and Dubourget, P.H. (Eds.). Antigens of Fish Path ogens: Development and Production for Vaccines and Serodiagnostics. Collection Foundation Marcel Merieux, Lyons, France, pp. 177 - 200. Meyers, T.R., Short, S., Farrington, C., Lipson, K., Geiger, H.J., and Gates, R. 1993. Establishment of a negative - positi ve threshold optical density value for the enzyme - linked immunosorbent assay (ELISA) to detect soluble antigen of Renibacterium salmoninarum in Alaskan Pacific salmon. Dis. Aquat. Org. 16(3): 191 - 197. Miriam, A., Griffiths, S.G., Lovely, J.E., and Lynch, W.H. 1997. PCR and probe - PCR assays to monitor broodstock Atlantic salmon ( Salmo salar L.) ovarian fluid and kidney tissues for the presence of DNA of the fish pathogen Renibacterium salmoninarum . J. Cl in. Microbiol. 35: 1322 - 1326. Mitchum, D. L., and Sherman, L. E. 1981. Transmission of bacterial kidney disease from wild to stocked hatchery trout. Can. J. Fish. Aquat. Sci. 38 : 547 - 551. Olea, I., Bruno, D.W., and Hastings, T.S. 1993. Detection of Renib acterium salmoninarum in naturally infected Atlantic salmon, Salmo salar L., and rainbow trout, Oncorhynchus mykiss (Walbaum) using an enzyme - linked immunosorbent assay. Aquaculture 116: 99 - 110. 236 Pascho, R.J., and Elliott, D.G. 2004. Bacterial Kidney Disea se In: American Fisheries Society - Fish Health Section Blue Book : Suggested Procedures for the Detection and Identification of Certain Finfish and Shellfish Pathogens, American Fisheries Society, Bethesda, Maryland. Pascho, R.J., and Mulcahy, D. 1987. E nzyme - linked immunosorbent assay for a soluble antigen of Renibacterium salmoninarum , the causative agent of salmonid bacterial kidney disease. Can. J. Fish. Aquat. Sci. 44: 183 - 191. Pascho, R.J., Chase, D., and McKibben, C.L. 1998. Comparison of the memb rane - filtration fluorescent antibody test, the enzyme - linked immunosorbent assay, and the polymerase chain reaction to detect Renibacterium salmoninarum in salmonid ovarian fluid. J. Vet. Diagn. Invest. 10: 60 - 66. Quinn, T.P., Peterson, J.A., Gallucci, V. F., Hershberger, W.K., and Brannon, E.L. 2002. Artificial selection and environmental change: countervailing factors affecting the timing of spawning by coho and Chinook salmon. Trans. Am. Fish. Soc. 131(4): 591 - 598. Sakai, M., S. Atsuta, and Kobayashi, M . 1989. Comparison of methods used to detect Renibacterium salmoninarum , the causative agent of bacterial kidney disease. J. Aquat. Anim. Health 1: 21 - 24. Sanders, J.E., and Fryer, J.L. 1980. Renibacterium salmoninarum gen. nov., sp. nov., the causative a gent of bacterial kidney disease in salmonid fishes. Intern. J. System. Bacteriol. 30: 496 - 502. White, M.R., Wu, C.C., and Albregts, S.R. 1995. Comparison of diagnostic tests for bacterial kidney disease in juvenile steelhead trout ( Oncorhynchus mykiss ). J. Vet. Diagn. Invest. 7(4): 494 - 499. World Organisation for Animal Health. 2012. Manual of diagnostic tests for aquatic animals, 5 th edition. Office International des Epizooties, Paris. 237 Chapter 6 Conclusions and Future Research 238 1 . Conclusions Considerable research has shown that the incidence of Bacterial Kidney Disease (BKD) in feral and propagated salmonid populations can be reduced and further controlled by employing several methods, including broodstock screening and culling , regular health inspections, and biosecurity measures at fish - rearing facilities. However, diagnosis of BKD can be inconsistent and greatly depends on the assay being used and the purpose of testing. In the Great Lakes basin (GLB), the prevalence of BKD has substantially declined over the past 10 years in Chinook salmon, coho salmon, and rainbow trout populations, making it even more imperative to continue to maintain the health of these economically and ecologically important populations. Continuing to monitor and control BKD infections, as well as developing new testing methods to enhance the detection of Renibacterium salmoninarum , is vital to maintain the health of the fish populations. The research presented in this dissertation emphasizes that the re is still much to learn regarding the detection of BKD. In Chapter Two, I describe how the prevalence of BKD has substantially declined over the past decade in Oncorhynchus spp. in the GLB. Historically, BKD has been associated with several mortality events of Oncorhynchus spp. in the GLB, and is known to cause chronic infections and mortalities in propagated and wild salmonids worldwide. I have shown how a drastic decline in the prevalence of BKD in wild and propagated Chinook salmon ( Oncorhynchus t shawytscha ), coho salmon ( O. kisutch ), and steelhead ( O. mykiss ) has coincided with the implementation of enhanced biosecurity measures at Michigan Department of Natural Resources (MDNR) gamete - collection weirs and hatchery facilities. It was found that 239 b iosecurity measures, such as selective culling, egg disinfection, separate nets for raceways, disinfecting footbaths, and regular screening of broodstock and pre - stocking fish, are important mechanisms for preventing and controlling the spread of R. salmon inarum in wild and enclosed environments. Over the ten - year period, it was also discovered that prevalence was typically higher in Chinook salmon from the Lake Michigan watershed than the Lake Huron watershed, which is also where BKD - associated large - scal e mortalities have occurred in the past. Within Chapter Three, I investigated the effects of exposure route and resultant disease course on the ability of non - lethally collected samples (blood, mucus, and a urine/feces mixture) to detect R. salmoninarum compared to the currently accepted method of a lethally collected kidney and spleen sample. While a urine/feces mixture was determined to have the greatest potential as an acceptable non - lethal surrogate, testing both a urine/feces mixture and a kidney a nd spleen sample was the best method for detecting R. salmoninarum prevalence and intensity in a population. This would greatly enhance the accuracy of disease detection, which would provide hatchery biologists with more reliable information to base their management decisions on. In Chapter Four, I tested a single - dilution indirect enzyme - linked immunosorbent assay with the purpose of detecting anti - R. salmoninarum antibodies in experimentally infected juvenile rainbow trout and feral adult Oncorhynchus s pp. The protocol used in this study was able to successfully detect antibodies in both the experimentally infected and feral fish. The antibody response observed in the experimental fish appeared to be of a short duration and did not confer protection wh en naïve fish were challenged with live R. salmoninarum . However, the antibody response of the feral fish was much higher, and also occurred in the absence of an 240 active R. salmoninarum infection in the kidneys and spleen of the fish. It is likely that th e feral fish are being exposed to R. salmoninarum on multiple occurrences, resulting in the continuous production of antibodies. This assay could be applied as a non - invasive method to determine if captive fish are being exposed to live R. salmoninarum . Within Chapter Five, I assessed the prevalence of BKD in male and female Chinook salmon returning to spawn at four MDNR gamete - collection weirs using multiple diagnostic assays, as well as the occurrence of several stages of disease. Not only was there a higher prevalence of BKD at the Swan River Weir, located in the Lake Huron watershed, but female Chinook salmon appeared to be more susceptible to R. salmoninarum infection than male Chinook salmon. While a higher prevalence of BKD in the Lake Huron wat ershed is contradictory to what was observed in Chapter 2, sampling for this study only occurred during 2005, which is encompassed by the study in Chapter 2, which transpired from 2001 to 2010. At the Little Manistee River Weir in the Lake Michigan waters hed, there was also evidence of disease progression, with more intense infections documented later in the spawning period rather than earlier. This is suggestive of healthier, better fit fish returning to spawn at earlier time points, while Chinook salmon returning to spawn late in the season are more heavily infected, resulting in a slower migration. As a result, it would likely be beneficial for fishery managers to consider harvesting gametes used for propagation earlier rather than later, to reduce the influx of R. salmoninarum being introduced into hatcheries. 24 1 2. Future Research The results presented in this dissertation have contributed substantially to further our understanding of R. salmoninarum epidemiology; particularly how the choice of diff erent tissues for sampling and diagnostic assays can impact the detection and spread of BKD in Oncorhynchus spp. in the Great Lakes basin. Furthermore, it has helped to establish an important foundation for non - lethal sampling and for deciding the appropr iate diagnostic assay. However, much more work is needed to ensure that the prevalence of BKD in GLB Oncorhynchus spp. continues to remain low. In particular, the prevalence and intensity of BKD should continue to be recorded on an annual basis, as any increases in the presence of BKD would be observed. Also, while it is very likely that the enhanced biosecurity measures implemented by the MDNR in 2002 were the leading reason as to why the prevalence of BKD declined so substantially, this should be val biosecurity measures, was not utilized as the MDNR could not manage to withhold the disease prevention methods from one of the hatcheries or gamete - collection weirs, which will likely continue to be a concern. Therefore, an alternative would be to mimic these conditions in an experimental challenge. This would aid in the determination of the true effectiveness of the enhanced biosecurity measures. Additionally, whil e this research evaluated the relationship between the prevalence of R. salmoninarum in broodstock and their resultant progeny, it would be meaningful to re - evaluate the same data, investigating the role of the wild environment on BKD infections in fish 242 re turning to spawn at the gamete - collection weirs. Supposing that fingerlings stocked into the GLB would return to spawn in 2 - 3 years, the prevalence and intensity of BKD in hatchery - raised pre - stocking fingerlings would be compared to adult fish returning to spawn 2 - 3 years later, with the assumption that these fish would be the same cohort. This analysis could allow for us to gain a better understanding of whether or not stocked hatchery fish maintain low BKD infections while in the wild environment. Fur thermore, it would be beneficial to monitor the prevalence in one - and two - year fish to further investigate the role of the natural environment in BKD infections. Of all the experiments performed in this dissertation, the concept of using non - lethal sampl es as an alternative for lethal sampling has recently become much more prevalent in the scientific community. The benefits of non - lethal sampling are numerous, but most important is that valuable fish species would not have to be sacrificed for disease te sting. While Chapter 4 did reveal that R. salmoninarum is detectable in non - lethally collected samples, future research should focus on mimicking a more natural challenge environment through the use of a cohabitation challenge. It was evident that the ex posure route and infectious dose affected the detection capabilities of the tissues; therefore, cohabitation challenges with varying infectious doses would reflect a more natural infection and would give a more accurate representation of the efficacy of no n - lethal samples. Also, although it can be costly, I would also recommend collecting samples more frequently, especially in a chronic disease course. Non - lethal samples collected from the chronic disease course challenge did not detect substantial levels of R. salmoninarum , which could be due to the timing of sample collection. If sampling occurred 243 more frequently, and if the study was extended for a longer period of time, we could increase the likelihood of detecting R. salmoninarum in tissues with low detection levels. The indirect ELISA used in this dissertation to detect anti - R. salmoninarum antibodies is an efficient test for screening Oncorhynchus spp. fish sera. Further research should focus on optimizing the current assay and investigating its use in other salmonid species. It would be beneficial to evaluate using R. salmoninarum with the p57 antigen removed as a coating preparation, as this antigen can sometimes mask the detection and quantitation of fish antibodies directed against the bacte ria themselves. Additionally, the feral fish that were tested in this study were a sub - sample of the available sera for antibody detection. A more thorough analysis of yearly serum samples from Oncorhynchus spp. during 2006 to 2013 would contribute to th e current knowledge of BKD prevalence. Furthermore, as Chinook salmon from the Little Manistee River Weir (LMRW) produced the highest levels of circulating anti - R. salmoninarum antibodies, Chinook salmon from the Swan River Weir in the Lake Huron watershe d should also be evaluated for antibodies against R. salmoninarum . The Swan River Weir generally has a lower incidence of BKD in Chinook salmon returning to the weir to spawn than LMRW, and as such, a lower antibody response could be expected as well. It would also be beneficial to continue to examine the progression of disease through the use of multiple diagnostic assays, as this can influence management decisions. The study performed in Chinook salmon from the Little Manistee River Weir should be repl icated over several consecutive years with more frequent sampling periods, to confirm the existence of the stages, as well as to better visualize trends in disease progression as only two time points were analyzed in this study. More complete knowledge of disease progression could influence when 244 the MDNR collects gametes from fish returning to spawn, as it would be advantageous to them to collect gametes during the time period when fish are least infected. As a result, this could reduce the amount of R. s almoninarum being brought into hatcheries through vertical transmission. Additionally, this analysis should be conducted at the other MDNR gamete - collection weirs in the GLB, as they are likely to experience different rates of BKD infection.