This is to certify that the thesis entitled Piliated Escherichia Goli and Porcine Enteric Disease in Michigan presented by Mark G . Evans has been accepted towards fulfillment of the requirements for M.S . Pathology degree in wax Major professor Date Max 12, 1983 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution MSU LlBRARlES RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. DO! NOT CIRCIULATE PILIATED ESCHERICHIA COLI AND PORCINE ENTERIC DISEASE IN MICHIGAN BY Mark G. Evans A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Pathology 1983 /37-/7VO ABSTRACT PILIATED ESCHERICHIA COLI AND PORCINE ENTERIC DISEASE IN MICHIGAN BY Mark G. Evans A study of 125 neonatal pigs was performed to determine the prevalence of pilus antigens of enterotoxigenic Escherichia coli in Michigan swine herds. Live, diarrheic pigs under 2 weeks of age submitted for study were euthana- tized and frozen sections of ileum were subjected to an indirect fluorescent antibody technique to detect the 3 known pilus antigens of g; ggli known to affect young swine (K88, K99, and 987PL. Ten centimeter ileal sections were used to determine numbers of lactose-fermenting bacteria. Of 52 pigs in which pili of §;.22li were demonstrated, 14 had K88 (27%), 23 had K99 (42%), 13 had 987P (25%) and 2 had K88 and K99 simultaneously (4%L. Numbers of lactose— fermenting bacteria were significantly higher “VWL05) in pigs with piliated g; ggli than in pigs without piliated g; ggli. significant cause of enteric disease in Michigan neonatal swine, and that the K99 pilus antigen is most frequently encountered. DEDICATION To my family ii ACKNOWLEDGMENTS My sincere thanks go to Dr. Glenn L. Waxler for his patient guidance during the course of this research project. His assistance was much appreciated. I wish to thank Dr. Allan Trapp from the Department of Pathology and Dr. John P. Newman from the Department of Microbiology for having served on my guidance committee. Their assistance has been most helpful. I am indebted to several individuals for their support: Cheryl Assaff for her clerical assistance; Irene Brett, Paul Carlson, and Becky Davis for their technical help; and John Allen for his hard work with this project. iii TABLE OF CONTENTS LIST OF TABLES O O O O O O O O O O O O O O O O O O O O LIS‘P OF FIGURES O O O O O O O O O O O O O O O O O O 0 INTRODUCTION 0 O O O O O O O O O O O O O O O O O O O 0 LITERATURE REVIEW 0 O O 0 O O O O O O O O O O 0 O O 0 Clinical Signs . . . . . . . . . . . . . . . . Lesions . . . . . . . . . . . . . . . . . . . Pathogenesis . . . . . . . . . . . . . . . . . Enterotoxins . . . . . . . . . . . . . Adhesion . . . . . . . . . . . . . . . The K88 Pilus . . . . . . . . The K99 Pilus . . . . . . . . The 987P Pilus . . . . . . . . Unidentified Adhesins . . . . Diagnostic Techniques for Pilus Identification Vaccination with Pili . . . . . . . . . . . . MATERIALS AND METHODS O O O O O O O O O O O O C O O 0 RESULTS Immunofluorescence Procedures . . . . . . . . Control Animals . . . . . . . . . . . Bacteriologic Monitoring . . . Control Inocula . . . . . . . Quantitation of E; coli from Control Inocula . . . Determination of Clinical Signs . . . . . . . . . . . Necropsy Procedures . . . . . Preparation of Antibodies . . . . . . Preparation of Tissues for Indirect Fluorescent Antibody Procedure . . Procurement of Field Specimens . . . . . . . . Quantitation of Lactose-Fermenting Bacteria . . . . . . . . . . . . . Association Index . . . . . . . . . . O O O O O O O O O O O O O O O O O O O O O O 0 Control Animals . . . . . . . . . . . . . . . Clinical Signs . . . . . . . . . . . . Gross Lesions . . . . . . . . . . . . Microsc0pic Lesions . . . . . . . . . iv Page . vi .vii 28 3O 31 31 .31 32 DISCUSSI SUMMARY LIST OF VITA . . APPENDIX Immunofluorescence . . . . . . . . . Bacteriology . . . . . . . . . . . Monitoring Germfree Isolators Quantitation of Control Inocula . . . . . . . . . Field Specimens . . . . . . . . . Clinical Signs . . . . . . Gross Lesions . . . . . . Microscopic Lesions . . . . Prevalence of Pilus Antigens Quantitation of Lactose-Fermenting Bacteria . . . . . . . . . . . . Association Indices . . . . . . . . O O O O O O O O 0 ON 0 O O O O O O O O O O O O O O O O O O O 0 REFERENCES 0 O O O O O O O O O O O O O O 0 Preparation of Antibodies . . . . . . . . . Data from Fluorescent Antibody Determination Page .33 .34 .34 . 34 , 35 . 35 , 35 . 36 36 O .39 .40 .44 .50 .52 .62 .63 .63 .64 Table 10 11 Al A2 LIST OF TABLES Page Escherichia coli strains and pilus antigens . . . Number and distribution of pigs receiving 1 ml of trypticase soybroth (TSB) cultures . . . . . Results of indirect fluorescent antibody test on control animals . . . . . . . . . . . . . . . . Number of E; coli from control inocula . . . . . . Distribution of pigs submitted with a history Of diarrhea O O O O O O O O O O O O O O O O O O 0 Number of pigs, diagnoses made, and E; coli pili identified in pigs submitted with a history Of diarrhea O O O O O O O O O O O O O O O O O O O Numbers of pigs with diagnoses other than piliated E; coli infection in pigs submitted with a history of diarrhea . . . . . . . . . . . Numbers of herds in which pili of _E_:_._ coli were identified 0 O O O O O O O O O O O O O O O O O 0 Prevalence of piliated E; coli in swine herds and pigs in this study . . . . . . . . . . . Numbers of lactose-fermenting bacteria per lO-cm loop of ileum in pigs with and without piliated E..- S-O-il O O O O O O O O O O O O O 0 O 0 Association indices of bacterial adhesion in pigs with and without piliated E; coli . . . . . . Distribution of rabbits injected with purified pili O O O O O O O O I O O O O O O O I O O O O O 0 Data from fluorescent antibody determinations . . Vi 24 24 33 34 35 37 38 38 39 39 40 63 64 Figure LIST OF FIGURES Photomicrograph of a fluorescent antibody- stained section of ileum from a control pig administered E; coli bearing K99. Posi- tive, specific fluorescence for the K99 pilus antigen of E; coli is located along the tips and lateral surfaces of villi. Eosinophils in lamina prepria have non- specific autofluorescence .. . . .. . . Photomicrograph of a fluorescent antibody- stained section of ileum from a control pig inoculated with trypticase soy broth only. Positive, Specific fluorescence is lacking. Eosinophils in the lamina propria have nonspecific autofluorescence .. . .. . . Photomicrograph of Giemsa-stained section of ileum from a pig submitted with a his- tory of diarrhea. Rod-shaped bacteria are adherent to the tips and lateral surfaces of villi .. . .. . .. . .. . .. . .. Vii Page 42 42 , , 43 INTRODUCTION Neonatal diarrhea in pigs is one of the most common causes of morbidity and mortality in swine herds. In Michigan, enteric diseases of young swine are considered to be a major obstacle toward increased profitability of swine production. Enteropathogenic Escherichia coli are responsi- ble for a substantial portion of the economic losses due to enteric disease. Diarrheal disease caused by enteropathogenic B4 3911 (enteric colibacillosis) has long been recognized as a pro- blem in Michigan, yet only recently have definitive diagnos- tic procedures been developed. The demonstration of certain adherence-promoting, hair-like organelles.(called pili or pilus antigens) on the bacterial surface of ngli and the positive correlation of these structures with enterotoxi— genicity has contributed to a better elucidation of the pathogenesis of enteric colibacillosis. Subsequently, vac— cines which hold the promise for a reduction of the inci- dence and severity of this disease in swine have recently been developed. A determination of the prevalence of pi- liated g; 9911 in Michigan neonatal swine will permit such vaccines to be of maximal value. 1 The objectives of this study were: 1) To evaluate the usefulness of an indirect fluores- cent antibody technique to identify K88, K99, and 987P pilus antigens of Escherichia coli in the ileum of neonatal pigs submitted with a history of enteric disease. 2) To compare, by quantitative methods, the number of lactose-fermenting bacteria in a lo-cm length of ileum in pigs with and without enteric colibacillosis. 3) To determine the prevalence of g; coli with each of the 3 pilus antigens in neonates in Michigan swine herds having enteric disease. LITERATURE REVIEW Pathogenic Escherichia coli have been incriminated as a cause of porcine enteric disease since 1899 (Jensen, 1948). Several reviews of this disease in pigs (Barnum 33 31., 1967; Nielsen__t___l_., 1968; Kohler, 1972) and other species (Tennant, 1971) have been published. There is a general consensus among investigators that diarrheal disease caused by pathogenic §;.22li is one of the most economically impor- tant diseases in pigs and other species (Sojka, 1965). Not all serotypes of Escherichia coli cause enteric disease, as some are a component of the normal flora of the intestinal tract of man and animals. The organism is a gram-negative, non-sporeforming rod of the family Enterobac- teriaceae and is 2-3 pm in length and 0.6 um in width. The bacteria may be motile or nonmotile and are classified as lactose fermenters. Illness caused by pathogenic strains of £5,991; is widespread due to the ubiquitous nature of patho- genic strains of the organism. The term enteric colibacil- losis has been used to describe the intestinal disease caused by pathogenic E; coli (Wilson, 1981). Clinical Signs Enteric colibacillosis in neonatal pigs is most commonly seen as an acute enteritis. The disease has been referred to by several names, including white scours, baby pig scours, coliform enteritis, diarrhea neonatorum, and neonatal colibacillary disease (Leman, 1970). The pigs affected by the disease are often less than.4ldays of age. However, the disease may strike at any time before weaning. Some pigs become ill as soon as 12 hours after birth (Dunne, 1975). Post-weaning enteric colibacillosis does occur, but death losses are fewer in this age group. Neonatal enteric colibacillosis ix: pigs is characterized by variable morbidity. Not all litters will be affected, nor will all pigs in an affected litter show signs of illness (Barnum _e__t_ 31., 1967). Mortality is also variable depending upon the age of the animal, Dunne and Bennett (1970) reported a mortality rate of approximately 70% in pigs 3 days of age or less, but mortality rarely reaches 100% (Leman, 1970). The clinical signs of enteric colibacillosis in young pigs include the passage of whitish-yellow, watery feces. This is often followed by dehydration. Vomition is usually absent but may be present in severe outbreaks (Wilson, 1981). Anorexia may be apparent. However, many suckling pigs having diarrhea continue to nurse. The hair coat may appear roughened, and the perineum may become irritated due to the continual presence of fluid feces (Dunne, 1975). Lesions There are no pathognomonic lesions, either gross or microsc0pic, in pigs with enteric colibacillosis. The gross lesions of this disease may include mild to severe inflammatory reactions. Occasionally, no inflammatory changes are present. The stomach is sometimes partially filled with curdled milk but lacks gross lesions. The intestines are usually distended and contain visible amounts of yellowish to grayish contents, often with excessive mucus. Hyperemia and edema may be seen in the small and large intestines, and petechial and ecchymotic hemorrhages may be present throughout the gastrointestinal tract. Villous atrOphy, usually associated with coronavirus infection, is occasionally seen with colibacillosis (Kenworthy and Allen, 1966; Moon 35 a1, 1970). Microsc0pically, the changes seen in colibacillosis in conventionally-reared pigs are variable. Smith and Jones (1963) reported no inflammatory changes in the intestinal tract in pigs with enteric colibacillosis, while Dunne and Bennett (1970) noted enlargement of goblet cells and vacuolation of absorptive cells. Edema in the lamina propria and neutrOphilic infiltration into villi have been noted (Christie and Waxler, 1972). A layer of adherent rod- shaped bacteria is often seen contiguous to the tips and lateral surfaces of small intestinal villi (Wilson, 1981). Pathogenesis Many recent advances have contributed to our understanding of enteric colibacillosis as it occurs in swine. In the past, scientists had great difficulty explaining why some serotypes of E; 9911 caused disease while others did not. However, the identification and characterization of certain virulence factors of some E; ggli have been a beginning in the quest for a more complete understanding of the pathogenesis of this disease. Enterotoxins One virulence factor of enterOpathogenic g; 991i is its ability to produce enterotoxins. The production of enterotoxins was demonstrated by Skerman 35 31. (1972) to be mediated by plasmids. These enterotoxins are generally classified according to their thermal stability, and this cflassification is generally in accord with different cellular activities of these toxin types (Evans 35 Elar 1973a). Smith and Halls (1967) demonstrated that certain E; 9911 produced a heat-stable (ST) enterotoxin while Gyles and Barnum (1969) demonstrated a heat-labile enterotoxin (LT). The heat-stable toxin was found to resist temperatures of 1000 C for 15 minutes; heat-labile toxin became inactivated at 60° C for 15 minutes. Enterotoxigenic E; 991i may produce either LT, ST, or both classes of enterotoxins (Gyles, 1970). .14 I - u . —' 5-. . '... J1 ‘ ..._‘ - , . ..- .2 iq :1": (J "J ;. _ ‘....' n-k' .. —.------I--- l J _ rib .. ~ 5- 'h' .. . .I 513-- -'- .n"' hil' _.Jg'.lJuL’ 1:51.; '.‘ 1' '4'. ‘3. _' . . J" 910“ .. '.\.— .$-.:-‘_ ...:._ at. C. .. 'i.‘ .1. '_. - _, L- .‘ . . .I] t:_ 4: _ IL'. .'. . .- - ' I ~‘- ~- . . l n ‘ u i {I} ‘-‘ '1. -'I _ . . .| ._ . . L)... I 1 . L— .. . . i . 1“,“— .i'. 1. _ .- . ‘. .. \ ! '.'..t:a '.'- 1'...“ l1; .2. 9 Ink).- .JC-l'. . '\.|N L...A:. I 1m- _-.. L‘ ._ . .. ‘5‘}: :. ”:1- .- .. J -a.'.1 ".;.' .an _ .' .. .- L.‘-|. _ "3H _ “1 ”v .1th- - . --.. -.i . E.» - . .. . .2 . . . . ‘ ~.. \ ...L .‘J....'.'. ._, -.- a -. .L :.;...J'._;.It_ n A-‘4 .l I am“: i. -.:LII.L o J hiIlLi. .- 1.1 -.'-|1».'I.'JJUJ. . -..' __._..___._.....- ._.. . .5 . ”misc. .. .: . - ~ . I - — - LJ 1;: . .. _ . .i . . - '. --. 1-”. - ____.- it: . ., - I - - I 4 . l4 . \ J . . I. ~. ‘ . .. . o \ i The antigenic prOperties of LT and ST differ. The LT has been shown to have antigenic behavior (Finkelstein 31 31., 1975), and Kohler (1978) has demonstrated that antibodies made against Ifl3‘were helpful in preventing disease. The ST is nonantigenic (Smith and Halls, 1967). However, pigs develop resistance with age to some, but not all, strains of enterotoxigenicr§1.gg11 that produce only heat-stable enterotoxin (Moon and Whipp, 1970). Whipp £1 31. (1980) concluded that these strains may produce an inhibitor which interferes with the cellular secretory reSponse of older pigs to ST. The structure of LT has been investigated. It is comprised of a single polypeptide chain with a molecular weight of 100,000 Daltons (Evans gt 21., 1976) and is related to the cholera toxin in its antigenicity (Gyles, 1974; Klipstein and Engert, 1977). The effect of LT in the intestinal epithelial cell is to increase levels of membrane-bound adenylate cyclase with a subsequent increase in the levels of intracellular cyclic adenosine 3'JV- monophosphate (cAMP). This causes a net fluid and electrolyte secretion by the epithelial cells of the small intestine, and the disease produced is therefore defined mechanistically as a secretory diarrhea (Moon, 1978a). Other nonintestinal cells, such as HeLa cell monolayers used in cell culture, have also been found to be sensitive to LT (Keusch and Donta, 1975). The mechanism of cAMP-mediated hypersecretion probably accounts for the diarrhea in other diseases such as salmonellosis and human cholera, both of which involve toxin production by bacteria (Moon, 1978a). Recently; Newsome 21.11.(1978) suggested that LT-induced intestinal secretion may be mediated by changes in the cyclic adenosine 3',5'-mon0phosphate/cyclic guanosine 3',5'- monophosphate (cGMP) ratio. The ST toxin of E1 9911 has not been fully characterized, and occasional reports indicate that different types of ST exist (Guerrant _1 __1., 1975; Steiner 31 21., 1972). Newsome §_t_ a1. (1978) reported on 2 subtypes of ST from §1H9211 strain P16 and designated them as STA and STB. The STA component was soluble in methanol, active in the 1-3 day old piglet, but inactive in the weaned pig. STB was insoluble in methanol, inactive in the 1-3 day old piglet, but active in the weaned pig. Available reports indicate that ST may induce diarrhea by a mechanism other than via adenylate cyclase (Gianella, 1977; Hamilton 31 21., 1978). Hughes 35 a1. (1978) and Newsome 31 21. (1978) concluded that either an increase in cGMP concentration or a decrease in the ratio of cAMP/cGMP may be the mechanism of action of STA, but little is known about the mechanism for the effects of STB. Kapitany 31 31. (1979) isolated an ST from porcine enterotoxigenic E1 9911 and designated it as ST-1261. This toxin had distinct similarities in chemical composition to the ST-43l isolated from a different strain of porcine enterotoxigenic E; coli by Alderate and Robertson (1978). However, it was difficult to make comparisons of these enterotoxins to the STA and STB previously described. The purification of heat-stable enterotoxins of 31 3311 has apparently been complicated by the complexity of the various growth media employed for enterotoxin production (Alderate and Robertson, 1978). Furthermore, the apparent heterogeneity of ST may be based on the sensitivity of the various test systems used for its purification (Burgess 33 '31., 1978). There are several methods availabLe to detect enterotoxins of E_._ 3311. An 13 3133 assay for heat-stable and heat-labile enterotoxins is the gut 100p test in which ligated segments of small intestine of an anesthetized pig are injected intraluminally with a: test inoculum. Distention of the loop after 24 hours indicates a positive reaction for enterotoxin (Smith and Halls, 1968). This test is most sensitive and reproducible when isolates are tested in the same species from which they were originally obtained. More recent methods for the detection of l_?._._ 3311 enterotoxins have been developed. Tissue cultures of Y-l adrenal cells (Donta 33 33,, 1974) or Chinese hamster ovary (CI-IO) cells (Guerrant 33 31., 1974) are used to detect LT. ST does not react in these tissue culture techniques. An 13 3133 technique for detection of LT is the rabbit skin test. This test is based on the effects of LT on vascular permeability (Evans 33 31., 1973b). 10 Several immunologic methods are used to detect LT. A lysis inhibition test (Evans and Evans, 1977), a radioimmunoassay technique (Greenberg 33 31., 1977), and an enzyme-linked immunosorbent assay (ELISA) have been described (Yolken 33 31., 1977). The lack of antigenic properties of ST make it inappropriate for detection by immunologic methods. Heat-stable enterotoxin may be detected by the infant mouse test, in which a test inoculum is injected into the stomach of the young mouse, and the subsequent fluid secretion into the intestines is measured. Of the 2 kinds of ST provisionally recognized, only STA is detected by the infant mouse test; the ligated gut loop technique in the pig is used to detect STB (Newsome 33 31., 1978L. Adhesion A second virulence attribute of enterotoxigenic 33 3311 is the ability to adhere to intestinal epithelial cell surfaces. It is the preperty of adherence to those cell surfaces that permits the microorganism to resist the peristaltic movement of ingesta and therefore avoid being washed out of the small intestine. In general, those bacterial organelles that promote adhesion to cell surfaces are defined as adhesins. These adhesins allow the 333311 to reproduce in great numbers, a process known as colonization (Nagy 33 33,, 1976). The antigenic structures responsible for adherence: and «colonization of enterotoxigenic:§3_coli in many species of animals have been 11 identified by several researchers (Burrows 33 31., 1976; Nagy 33 31., 1977; firskov and flrskov, 1966; firskov 33 31., 1975; Smith and Linggood, 1971). In each case, the adhesins were found to be hair-like projections, designated as pili or fimbriae, on the surface of the bacterium. The presence of pili is associated with ability of the organism to produce enterotoxin in most field isolates of enterotoxigenic 33 3311 (Isaacson, 1981). Brinton (1959) was the first to use the term pili (Latin for hairs) to describe the nonflagellar structures projecting from the bacterial surface. Duguid (1955) used the term fimbriae (Latin for fringe) to identify the same appendages. This class of pili is not involved in the transfer of DNA between bacteria, a function of conjugal or sex pili. (Other classes of pili are known as type I soma- tic,‘ or ”common", pili. These have no role in conjugation and are coded for by chromosomal genes. Common pili are hollow fibers 7 nm in diameter and are 500-2000 nm in length. A given bacterial cell may have 50-400 per bac- terium (Brinton, 1959, 1967). Common pili enhance upon bacteria many behaviors lacking in the nonpiliated phase. These behaviors include surface translocation, enhanced growth in marginal oxygen concentrations, tight colonial association of bacterial growth, the ability to agglutinate guinea pig erythrocytes, and the ability to adhere tolcell surfaces (Brinton, 1978). The hemagglutinating and adher- ence properties are eliminated by the monosaccharide 12 D-mannose. This observation has lead to the term "mannose- sensitive hemagglutination and adhesion? (MSHA) and is a characteristic of adhesion mediated by type I somatic pili (Duguid 33 31., 1955). Type I piliated §_._ 3311 may be distinguished from nonpiliated organisms by colonial mor- phology (Salif and Gotschlich, 1977a, 1977b). However, the role of common pili in the colonization of enterotoxigenic 33 3311 has not been proven (Isaacson, 1980). The K88 Pilus. firskov and firskov (1961) isolated 33 3311 from piglets with diarrhea and described a new pilus antigen designated as K88. It was designated as a capsular (K) antigen due to its serologic behavior during serotyping procedures. However, it was later recognized that the pili were protein antigens rather than polysaccharides (as reviewed by Moon, 1978b). 'The protein has a high molecular weight which can be separated into apparently identical subunits of approximately 25,000 Daltons each (Mooi and DeGraaf, 1979). flrskov 33 31. (1964) demonstrated that K88 existed in two distinguishable forms, designated as K88ab and K88ac, but Stirm 33 31. (1967a) were the first to isolate and partially purify K88ab. Recently, a: new variant, provisionally designated as K88ad, has been described by Guinee and Jansen (1979). These subtypes of the K88 antigen are serologic variants which have minor differences in their amino acid compositions. Each variant of K88 can be detected by immunodiffusion and immunoelectrophoretic l3 techniques (Mooi and DeGraaf, 1979). Stirm 33 31. (1967b) found that K88 was morphologically distinguishable from type I somatic pili in electron micrographs. The K88 antigen appeared as fine fibers on the bacterial surface. These pili were 100-1500 nm long and 7- 11 nm wide. flrskov and flrskov (1966) found that K88 pili were plasmid-mediated and that nutrient requirements 13 31333 for K88 expression were fastidious. The K88 antigen was expressed after growth on solid medium at 37° C but not at 18° C. Stirm 33 31. (1967b) demonstrated that entero- toxigenic 3; 3311 bearing the K88 pilus had the ability to agglutinate guinea pig red blood cells in the absence and presence of D-mannose. Smith and Linggood (1971) demonstrated that K88 was a virulence factor of K88-positive strains of 33 3311 that cause diarrhea in neonatal pigs. These strains manifested adhesive behavior to intestinal epithelial cells, permitting attachment and colonization in the small intestine. It was further demonstrated that K88-negative mutants of K88-posi- tive strains were nonenteropathogenic because they could not attach to enterocytes and therefore not colonize the intes- tine, despite being comparable in their abilities to produce enterotoxin. Jones and Rutter (1972) demonstrated that K88 antisera inhibited the 13 3133 attachment of the K88 pilus and con- cluded that antibodies directed against K88 could prevent its adherence in) porcine intestinal epithelium. This l4 discovery has been an impetus for the development of veterinary vaccines against enteric disease caused by pi- liated 313311. Smith and Linggood (1971) also found that 33 3311 bearing K88 did not attach to the intestinal epithelium of all nonvaccinated piglets. Rutter 33 31. (1975) described 2 different phenotypes in pigs based on the ability of K88 positive 33 3311 to adhere. The "adhesive” phenotype was presumed to possess a receptor for the K88 pilus that pro— moted preferential attachment to it. These pigs were more susceptible to colibacillosis from 33 3311 bearing this pilus than the "nonadhesive" phenotypes which lacked the receptor. Gibbons 33 31. (1977) demonstrated that the receptor for K88 is inherited in a simple Mendelian fashion, with l locus and 2 alleles, and the dominant allele is expressed as the receptor for K88 on the microvillous sur- face. The homozygous recessive pig lacks this receptor. Sellwood 33 31. (1975) have described an 13 31333 assay in which the interaction between K88-bearing 33 3311 and the intestinal brush borders was directly observed. Sellwood (1980) has recently demonstrated a technique which allows measurements of the K88/receptor interaction. Attempts to further define the chemical properties of the receptor uti- lizing the ability of soluble K88 pilus antigen to aggluti- nate guinea pig erythrocytes have been made. Gibbons 33 31. (1975) concluded that hemagglutination can be inhibited by some glyc0protein moieties, especially the galactosyl residue. However, much remains to be learned about the chemistry of the receptor for K88. Recently, Bijlsma 33 31. (1982) described a further subdivision within the group of "adhesive” animals by demon- strating different provisional phenotypes in pigs distin- guishable with the 3 variants of the K88 antigen, using the brush border technique of Sellwood 33 31. (1975). Pigs of phenotype A were susceptible to adherence of all three K88 variants; pigs of phenotype B were susceptible to K88ab and K88ac; pigs of phenotype C were susceptible to K88ab and K88ad; pigs of phenotype D were exclusively susceptible to K88ad; and pigs of phenotype E were resistant to adhesion by any of the K88 variants. It is of interest that the K88 antigen has been shown to be a‘virulence factor in only 1 species, the pig (Moon, 1978b). The K99 Pilus. Smith and Linggood (1972) described another antigen of 33 3311 and designated it as KCO’ They demonstrated that it was a plasmid-mediated antigen and that when the plasmid was removed from enterotoxigenic lamb strains the organism was not diarrheogenic. These re- searchers further demonstrated that KCO-positive strains colonized the small intestine in high numbers and speculated that KCO in lamb and calf enterotoxigenic 3: 3311 strains may serve a similar function as K88. KCO later beCame officially designated as K99 (flrskov 33 1., 1975) and was speculated to have a structure resembling pili. 16 The K99 pilus antigen was found to be similar to K88 in many respects. Burrows 33 31. (1976) demonstrated that K99, like K88, was expressed after cultivation on solid agar at 37°’C but not at 18° C and that 33 3311 bearing K99 caused mannose-resistant agglutination of ovine erythrocytes. ,brskov 33 31. (1975) confirmed that genes for K99, like K88, were found in plasmids. Both K88 and K99 have a similar appearance on electron micrographs (Isaacson, 1977). The K99 pili were purified by Isaacson (1977). It is characterized by rod-like structures that averaged 8.4 nm in diameter and 130 nm in length. The antigen was composed of 2 subunits that were mainly protein with a minor lipid component. The major subunit had a molecular weight of 22,500 Daltons, and the minor subunit had a molecular weight of 29,500 Daltons. Isaacson (1977) confirmed that K99 had the structure of pili. The nutrient requirements 13 31333 that promote expression of K99 have been explored. Guin‘e'e 33 31. (1976) reported that K99 could be more easily demonstrated when strains were cultured on a buffered medium containing a minimal amount (HE casein (Minca medium). The supplementation of this medium by 1% IsoVitalexa gave improved results (Guinée 33 31.,1977). Isaacson 33 31. (1978b) further improved this method by growing K99 strains (M1 Minca-IsoVitaleX agar. Cultures that tested negatively aBaltimore Biological Laboratory, Baltimore, MD. 17 for K99 by slide agglutination were then passaged daily for 4 days in trypticase soy broth with vigorous shaking. Strains were then regrown on Minca-IsoVitaleX agar. Unlike the K88 antigen, K99 has been associated with enterotoxigenic strains of 33 3311 in lambs and calves (arskov 33 31,, 1975). (Moon 33 31. (1977) detected K99 from enterotoxigenic 33 3311 isolated from piglets and demonstrated that calf strains of 3.33311 bearing K99 were enteropathogenic for pigs. Little is known about the intestinal receptor for K99 in the pig, lamb, or calf. Data suggest, however, that pilus receptors exist in the small intestine that recognize only a single pilus type (Isaacson t _1., 1978c). Moon 33 31. (1979) have speculated that all neonatal pigs have intestinal receptors for K99 indicating that there is no inheritable resistance in pigs to infection with K99-bearing 33 coli. However, this has not been rigorously established. The 987P Pilus. A third pilus antigen of & coli has been described by Isaacson 33 31, (1978c). Many enterotoxi- genic 33 3311 strains isolated from piglets with diarrhea lacked K88 and K99, yet were clearly enteropathogenic. These strains adhered to porcine enterocytes 13 3133 and 13 31333 and induced diarrhea when given orally to piglets. One such strain,§3 3311 987, has been found to adhere by a class of pili that are different from both K88 and K99. Escherichia coli 987 pili had similar dimensions to common pili in electron micrographs, yet they differed 18 biochemically and antigenically. Moreover,tnflike common pili, 33 3311 987 pili did not hemagglutinate guinea pig erythrocytes (Isaacson 33 31., 1978c; Nagy 33 31., 1977). The 987P pili have been detected in calf, lamb, and pig strains of E3311, but they are a virulence factbr only in the pig (as reviewed by Moon, 1978b). It is not known whether the genes that code for 987P pili are located in plasmid or chromosomal nucleoproteins (Moon 33 31., 1979). Recently, 2 enterotoxigenic strains of 33 3311 from different sources were both found to express 987P and K88 pilus antigens at different times. ‘This indicates that some _E_.3311 occur naturally with the potential to produce more than 1 of the pili that promote colonization (Schneider and To, 1982). It would be helpful to determine if strains which produce more than 1 pilus antigen produce them both 13 3133 and if they both promote colonization of such strains. Little is known about the nature of the intestinal receptor for the pilus of 1°33 3311 987 in the pig. Dean and Isaacson (1982) used an 13 31333 laboratory model to identify and isolate a 987P pilus-specific receptor- containing fraction from small intestinal epithelium and brush borders isolated from adult female rabbits. Interestingly, the receptor-containing fraction was not found in neonatal rabbits. Moon 33 31. (1979) have suggested that, as in the case of receptors for K99, all neonatal pigs have intestinal receptors for 987P. 19 Unidentified Adhesins. Recently, Amad-Masalmeh 33 31. (1982) described 3 strains of enterotoxigenic 33 3311 that adhered, colonized, and caused diarrhea in pig intestine but which did not produce K88, K99, or 987P antigens. Such strains were designated as 3P“ strains. These workers con- cluded that 3P" enterotoxigenic _E_._ 3311 produce previously uncharacterized adhesins and demonstrated that certain strains did produce pili but that these pili apparently did not promote adherence or colonization. Apparently, the adherence of 3P’ strains to enterocytes is mediated by adhesins that are not of a pilus nature. Morris 33 31. (1982) documented a new filamentous antigen, provisionally designated as F41, which was produced by a strain of 3:3 3311 infective for calves that was a K99- negative mutant of the K99 reference strain B41. Adhesiveness was demonstrated both 13 3133 and 13 31333, and the antigen was shown to have mannose-resistant hemagglutination properties. While these features point to the possibility of F41 being a new pilus antigen of enterotoxigenic 15:3 3311, experiments are necessary to clarify the role of F41 in the pathogenesis of diarrheal disease. Moon (1978b) has suggested evidence for previously uncharacterized pilus antigens. These strains have the ability to colonize intestines of neonatal pigs and yet lack K88, K99, or 987P. Such 33 coli may yield new pilus antigens. However, Francis (1982) has indicated that the u' . 'o: . a ‘ - . - n .— 111... .' ‘ .\I‘\'II. I . .- 1 .9 n. ., . -.'I'.'-..;|'... . .1 .l -'u I.‘ ' - l. - " -- .-- '-JJ.‘. _ ,. '. .' .JL . ..-. l- u -. _ .'I 1'-“ . . -.. a. v. - . a z." ‘1. .— I ’ I ‘ ' - ' I ..- -I ‘§l- ‘l- L . ' . . . L'InL': '. . . . I I . . - " . . 'l.’ .k . :. . _ . _. . - ”J .-. 1 ..-&:-t--' - "-t . - I .'- .J‘l -: ‘._.'... J-Il. . . .14 n- | . . . . t . 1'1..- v.’ . ..E—.‘.'-- .- .._._. _ .1..L. - 14.5w ..'-. .Ls .-'-' . -, L'il . . r '. -‘J L\: LII-I 431L141.) .' .‘- -' - .'-‘ \- LL.) \ .2\'. 3 ' 4L. . . ' . . . ... . ., l .. _ | . no.1. L: ' . =-.u. -‘ . . L' . . . - .‘lliZ-L- . 3.. - 2h 1., '. _ _. :.. _, ‘- . .-‘. ‘- ‘},5,: Jul-3 11.. - . . , , . ,.. , , _. U“ ._.|. , - .E'J.‘ 20 vast majority of piliated 33 coli found in a prevalence study in north central United States bore K88, K99, or 987P pili. Diagnostic Techniques for Pilus Identification several techniques. An enzyme-linked immunosorbent assay (ELISA) has been used to detect K99 in calf feces and K99 antibody in serum and colostrum of calves (Ellens 33 31., 1978). These results indicate that K99-positive E. coli infections can be diagnosed serologically. Mills 33 31. (1982) recently described an ELISA for the detection of K88 pili. A slide agglutination test for detection of K99 has been reported by Guinée 33 31. (1976). 'When Minca or Minca- IsoVitaleX medium was employed, the K99 antigen was more readily detected (Guine’e 33 31., 1977). A modification of this technique utilizing trypticase soy broth gave improved results (Isaacson 33 33?, 1978b). Probably the most suitable technique for the demonstration of pilus antigens is the indirect fluorescent antibody technique. Several workers (Moon, 1978b; Isaacson _3 _1., 1978a, 1978b; Moon 33 31., 1978, 1980) recommended immunofluorescent staining of bacteria in frozen ileal sections using specific K88, K99, or 987P antisera for a sensitive and simple technique for the detection of these pilus antigens. 21 Vaccination with Pili Because enteric colibacillosis in pdglets usually occurs during the neonatal period, the disease is controlled by immunization of pregnant sows to stimulate colostral antibodies which protect piglets that suckle such dams (Rutter 33 31., 1976). Controlled experiments have been done using purified pili as parenteral vaccines with subsequent challenge of suckling piglets with strains of enterotoxigenic 33 3311 bearing a homologous pilus antigen. Protection correlated with the presence of anti-pilus antibodies in the colostrum and milk of immunized sows.(Nagy 33 31., 1978; Rutter 33 31., 1976). In other experiments, piglets born to control and immunized dams were then challenged with a strain of enterotoxigenic 33 3311 bearing the homologous pilus antigen, but different cell wa11.(O) and flagellar (H) antigens from those of the strain from which the vaccine was prepared. 'The results again indicated that vaccine effectiveness was determined by the presence of anti-pilus antibodies (Morgan 33 31., 1978). These studies concluded that up to 80 or 90% of _E_._ 3311 diarrhea in pigs can be prevented by use of a trivalent pilus vaccine con- taining purified K88, K99, and 987P pili. MATERIALS AND METHODS Immunofluorescence Procedures Control Animals Gnotobiotic pigs were obtained from a Hampshire sow using a technique described by Waxler, _3 31. (1966). Epidural anesthesia was induced by injection of 25 m1 of b given at the lumbosacral 2.5% procaine hydrochloride articulation, followed by 5 ml of tranquilizer given intramuscularlyc. The sow was positioned in right lateral recumbency. The left flank was surgically prepared and sprayed with a sterile surgical adhesived, and the plastic isolator was positioned over the area. The surgeon cut through the vinyl floor of the isola- tor unit and the sow's skin with a cautery scalpel unit. The remaining layers of the abdominal wall were incised with scissors. Each pig was removed from the uterus by making a separate incision in the uterine wall. ‘Umbilical clampse were placed on the umbilical cord of each pig prior to severing the cord. The pigs were then aseptically bEpidural, Haver-Lockhart, Kansas City, MO. cSparine, Wyeth Laboratories, Philadelphia, PA. dVi-Drape, Parke-Davis, Detroit, MI; e"Double Grip" disposable cord-clamp, Hollister, Inc., Chicago, IL. 22 23 transferred into a plastic holding isolator. The litter consisted of 8 live pigs, 7 of which were used in this study. The 7 pigs were allocated to 4 separate isolators, with 2 pigs in 3 isolators and l pig in a fourth isolator. Equipment for feeding and for microbiologic determination was contained in each isolator. A sufficient quantity of a commercial sterilized liquid dietf was aseptically transfer- red into each isolator, and each pig was given 3 to 4 oz. of liquid diet 3 times a day. Each pig was maintained in an individual cage in sterile plastic rearing isolators at a room temperature of 30-32°C. Bacteriologic Monitoring. Swabs were taken from each isolator prior to the pigs'«exposure to the test inoculum. Specimens of rectal contents, waste material from the cages, and contents of feeding trays were taken. Material was streaked on blood agar9 plates and inoculated into thiogly- collate brothg. The media were incubated both aerobically and anaerobically at 25°, 37°, and 55°C. Control Inocula. Cultures of 3 strains of 33 coli of known pilus type growing on trypticase soy agar were ob- tained from Dr. R. E. Isaacson, School of Public Health, University of Michigan, Ann Arbor (Table 1). ‘The strains were transferred to trypticase soy broth and incubated aero- bically overnight at 37°C. fSPF-Lac, Borden Inc., Norfolk, VA. 9Difco Laboratories, Detroit, MI. “.m— 24 Table l - Escherichia coli strains and pilus antigens. Strain Pilus Antigen 1248 K88 431 K99 987 987P Three milliliters of the broth culture of each strain were placed in individual 5 ml sterile glass ampules and heat sealed. .An ampule containing only trypticase soy broth with no bacteria was also prepared in this manner. The ampules were aseptically transferred into the appropriate isolators just prior to exposure of the pigs. At 7 days of age, each pig was orally inoculated with 1 ml trypticase soy broth culture (Table 2). Table 2 - Number and distribution of pigs receiving 1 ml of trypticase soy broth (TSB) cultures. Isolator Number Number of Pigs Inoculum l l TSB only 2 2 Strain 987 in TSB 3 2 Strain 1248 in TSB 4 2 Strain 431 in TSB 25 Quantitation 33 33 coli from Control Inocula. One milliliter of each inoculum given to the gnotobiotic pigs was serially diluted in 9 m1 of phOSphate-buffered saline. Eleven 10-fold serial dilutions were made. One milliliter of each dilution from 10"4 to 10"8 was cultured using MacConkey agar and a standard pour plate technique. All plates were incubated overnight at 37°C. Because dilutions at 10-1, 10-2, and 10"3 would probably yield growth that was too numerous to count, cultures of those dilutions were not made. Dilutions in which 30 to 300 colonies could be coun- ted were used in the calculation of the number of 33 3311. Determination 33 Clinical Signs. Pigs were observed 3 times daily for clinical signs of anorexia, dehydration, diarrhea, and vomiting throughout the duration of the experiment. Necropsy Procedures. The pigs were euthanatized 24 hours after exposure to the inocula by administering 2.0 m1 of sodium pentobarbitolh into the anterior vena cava. Pigs were then necropsied in dorsal recumbancy by incising and caudally reflecting the sternum and ventrum. Sections of duodenum, jejunum, and ileum were frozen in OCTi mounting medium, and other sections of these same tissues were fixed in 10% neutral buffered formalin. Formalin-fixed intestinal sections were stained with hematoxylin and eosin as well as RHaver-Lockhart Laboratories, Kansas City, MO. 1Tissue-Tek II OJLT. Compound, Lab-Tek Products Division, Miles Laboratories, Inc., Naperville, IL. 26 Giemsa stains. Sections of kidney, liver, lung, lymph node and spleen were also formalin-fixed and stained with hema- toxylin and eosin. Preparation 33 Antibodies The monospecific anti-pilus antibodies were prepared in this laboratory prior to the beginning of this study. “The method of preparation is summarized in the Appendix. Preparation 33 Tissues for Indirect Fluorescent Antibody Procedure The gnotobiotic pigs served as controls to determine the monospecificity of each anti-pilus antibody preparation. Sections of ileum 5 cm in length were taken adjacent to the ileocecal junction. Sections were ligated at 1 end with suture material, and the embedding medium was introduced into the lumen of the intestine until it was mildly dis- tended. The open end was then ligated. The tissue Was placed in a container with dry ice to harden the embedding medium. Care was taken to not allow direct contact of the tissue with the dry ice. (Sections of embedded ileum were then cut into cross sections 3-5 mm in length. Selected sections were placed on a cork disc 2.2 cm in diameter and 0.3 cm in thickness. Additional embedding medium was added to cover the cross sections. Tissues were then placed into small labelled plastic bags, sealed, and stored at -70°C. When sections of frozen, embedded tissue were utilized for the indirect fluorescent antibody procedure, they were attached onto metal cylinders. Six-micron thick cross 27 sections were cut using a cryostatj. Sections were mounted on glass slides, heat-fixed at 37°C in a moist chamber, and fixed in acetone for 15 minutes at room temperature. Slides were allowed to dry, and individual tissue sections were overlaid with the anti-pilus antibodies made in the rabbit. Eight ileal sections from each animal were used. Two sec- tions had anti-K88 antibody placed over thenn ‘two sections had anti-K99 antibody placed over them; two sections had anti-987P antibody placed over them. The remaining 2 sec- tions had normal rabbit serum placed over them. All antisera were used at a dilution of 1:100 in phosphate buffered saline (PBS) at pH 7.2. Slides were incubated in a moist chamber for 30 minutes at 37°C followed by 2 rinses in PBS at pH 7.2. The second antibody was a commercially available anti-rabbit IgG fluorescein conjugate made in the goatk'l. It was used at a dilution of 1:16 in PBS at pH 7L2. Slides were incubated in a moist chamber at 37°C as before. ‘Two PBS rinses were made. 'The slides were dried and coverslipped with a 1:9 mixture of PBS and glycerol. The mounted specimens were examined on a Zeiss Photomicroscope IIIm with immunofluorescence capacity. .All tissues were examined the same day they were processed. A JSouth London Electrical Equipment Co., Ltd., London, Englafld. Miles-Yeda Ltd., Kiryat Weizman, Rehovot, Israel. lAntibodies Incorporated, Davis, CA. mCarl Zeiss, Oberkocken, West Germany. 28 specimen was considered positive for pilus antigens if specific, apple-green-colored fluorescence was noted on the tips and lateral surfaces of villi. Procurement 33 Field Specimens The animals utilized for this study were pigs submitted to the Animal Health Diagnostic Laboratory at Michigan State University by Michigan producers between 6-30-81 and 12-31- 82. Only live pigs 2 weeks of age or younger with clinical evidence of diarrhea were used in this study. Pigs were euthanatized with T-61n administered intravenously. Necropsy was performed immediately following death to mini- mize autolytic postmortem changes. Ileal sections from these pigs were Subjected to the indirect fluorescent anti- body technique previously described, and the results were used to determine a prevalance of the K88, K99, and 987P pilus antigens of 33 3311 in Michigan neonatal pigs. Quantitation 33 Lactose-Fermenting Bacteria A 10 cm section of ileum was obtained from each pig included in this study. The ileal segment was ligated at each end with suture material. fThe tissue was placed in a 0.3% solution of peptone water° to give a total volume of 30 m1. This segment in peptone water was then blended at high nT-6l Euthanasia Solution, National Laboratories Corp, American Hoechst Corporation, Somerville, NJ. acto-Peptone, Difco, Detroit, MI. .4" 29 speed for 1 minute in a Sorvall Omni-MixerPL One milliliter of the resultant slurry was pipetted into 9 ml of peptone water, and eight lO-fold serial dilutions were made in triplicate» One milliliter of each dilution was cultured on MacConkey agar using a standard pour plate technique with incubation overnight at 37°C. An average colony count for each dilution was determined by counting viable colonies using a digital colony counterq, adding the counts together for that dilution, and dividing by 3. Those dilutions in which the average count was between 30 and 300 were used in the calculation of the number of lactose—fermenting bacteria according to the following formula: Con = l/lO"a x 30 x b where Con is the original concentration of the numbers of lactose-fermenting bacteria per 10 cm of ileum, 10"at is the dilution from which the inoculum was taken, 30 is the total volume of peptone water into which the 10 cm segment was homogenized, and b represents the average number of colonies counted for that dilution. An unpaired "t" test was used to compare numbers of viable lactose-fermenting bacteria in pigs with and without piliated 33 3311 in their ilea (Steel and Torrie, 1960). pIvan Sorvall, Inc., Newtown, CT. qLab-Line Digimatic Colony Counter, Lab—Line, Inc., Melrose Park, IL. 30 Association Index The degree of association between piliated 33 3311 and intestinal epithelium was determined morphologically by a modified association index as described by Bertschinger 33 31. (1972). Giemsa-stained cross sections of ileum were examined microscopically with the 40x objective. The association index was determined by 2 criteria. The first criterion was the number of bacteria seen at the base of villi, and this evaluation was given a value from 1 to 5. A value of 1 represented no bacteria at the villous base and a value of 5 represented maximal numbers of bac- teria in this location. The second criterion was based on the tendency of these bacteria to be contiguous to epithe— lial cells, and was graded similarly from 1 to 5. The association index for an ileal segment was obtained by multiplying the values obtained from those 2 criteria. An association index value of 25 indicated maximal numbers of bacteria at the villous base with a strong tendency to be contiguous to the apical portions of epithelial cells. However, an index of 1 indicated that bacteria were lacking at the villous base and that bacteria were not contiguous to the intestinal epithelium. A Sign test was used to deter- mine statistical significance between pigs with and without piliated 33 coli in their ilea (Steel and Torrie, 1960). RESULTS Control Animals Clinical Signs All gnotobiotic pigs appeared to be healthy prior to the administration of the control inocula. Within 24 hours after giving the inocula, pigs in isolators 2, 3, and 4 had evidence of diarrhea, and the pigs in isolator 3 had died. The pig in isolator I appeared normal. A summary of the pilus antigens of 33 3311 to which the pigs were exposed is seen in Table 3. Gross Lesions The pigs in isolators 1, 2, and 4 were euthanatized 24 hours after exposure. NecrOpsies were performed on all pigs. The pig in isolator 1 had a cyst on 1 lobe of the liver and mild edema of the spiral colon. Intestinal con- tents were normal. Pigs from isolators 2, 3, and 4 had watery contents in the small and large intestines. .Also the wall of the small and large intestines appeared thinner than normal. Pigs from isolator 3 both had hyperemia of the serosal surface of the small intestine. 31 32 Microscopic Lesions Sections of kidney, liver, and lung from the pig in isolator 1 were normal microscopically; However, sections of lymph node and spleen appeared decreased in cellularity. Sections of duodenum, jejunum, and ileum were normal in microsc0pic appearance, but the jejunum and ileum had vacuo- lated cytoplasm in their epithelial cells. The pigs in isolators 2 and 4 had similar microsc0pic lesions. The kidneys, lungs, lymph nodes, and livers of these pigs were normal microscopically. The Spleen appeared hypocellular in l of the pigs from isolator 2. No bacterial colonies were seen in Giemsa-stained sections of duodenum, but the jejunum had occasional rod-shaped organisms adherent to tips of villi. Vacuolation of the cytoplasm was also noted in the jejunum. The ileal epithelium was charac- terized by cytoplasmic vacuolation and greater numbers of adherent rod-shaped bacteria than seen in the jejunum. The pigs in isolator 3 had a normal microscopic appearance in the kidneys, livers, lungs, and lymph nodes. The spleen appeared congested in 1 of these pigs and normal in the other; 'The gastrointestinal tract had autolysis in the duodenum, jejunum, and ileum, and cytOplasmic vacuola- tion was noted in the epithelial cells of those three levels. On Giemsa-stained sections, the duodenum and ileum had the most rod-shaped bacteria adherent to epithelial cell surfaces, and the jejunum had the fewest. Some of the villi 33 in sections of ileum from these pigs appeared shorter than normal. Immunofluorescence Ileal sections of control animals were used to determine the specificity of the indirect fluorescent antibody technique for demonstrating pilus antigens. Sections that had apple-green-colored fluorescence along the tips and lateral surfaces of the villi were considered positive for pilus antigens (Figure 1). Sections that lacked such fluorescence were considered negative for the 3 pilus antigens (Figure 2). In many sections, autofluores- cent eosinophils were seen in the lamina prOpria (Figures 1 and 2). All 3 pilus antigens had a similar degree and distribution of fluorescence. The results of this test are summarized in Table 3. Table 3 - Results (ME indirect fluorescent antibody test on control animals. Pigs Test Result Isolator Exposed No. To K88 K99 987P Neg l Unexposed - - - + 2 987P — - + - 3 K88 + - - - 4 K99 - + - - All— 34 Bacteriology Monitoring Germfree Isolators. No bacterial growth from isolators 1 and 3 could be demonstrated. However, isolators 2 and 4 yielded a gram positive coccus, which may have been a contaminant. This was probably caused by an undetected break in the isolator unit. This was not judged to alter the results of this experiment. Quantitation 33 Control Inocula. One milliliter of each broth culture of 33 coli given to gnotobiotic pigs contained approximately 109 33 coli. The control inoculum administered to the pigs in isolator 1 yielded no bacterial growth (Table 4). Table 4 - Numbers of 33 coli from control inocula. Strain of 33 3311 Numbers Isolator in control inoculum of 33 3311 1 Control 0 2 987 ga 3 1248 9 4 431 9 aData expressed as mean loglo. 35 Field Specimens A total of 125 pigs from 55 Michigan swine herds were submitted to the Animal Health Diagnostic Laboratory between 6-30-81 and 12-31-82. disease (Table 5). Table 5 - Distribution of pigs submitted with a history of diarrhea. Each herd had a history of diarrheal No. of herds No. of pigs No. of pigs No. of pigs represented examined with diarrhea 55 125 119 Clinical Signs Of the 125 pigs in this study, 119 had clinical evi- dence of diarrhea. The feces were often yellowish in color and watery in consistency. Such pigs usually had sunken eyes and increased skin turgor as evidence of dehydration. Vomiting was rare. Gross Lesions The gross lesions of pigs submitted with a history of diarrhea varied considerably; A reddening of the skin in the perianal area was common because of constant irritation due to watery feces. The small and large intestines were often hyperemia on the serosal surface. Several pigs had edema of the mesentery of the spiral colon. Mesenteric without diarrhea _ LIII— 36 lymph nodes were occasionally swollen. However, none of these lesions could be considered pathognomonic for 1 etio- logic agent. MicroscoPic Lesions Observation of microscopic lesions for this study was limited to sections of ilea from pigs submitted with a history of diarrhea. The natune of the ileal lesions usually varied depending upon the etiologic agents isolated from the various tissue specimens routinely examined. For example, pigs with coronavirus infection were often observed to have blunted, atrophic villi in certain sections of ileum. Those pigs in which a piliated 33 3311 was demonstrated had ilea in which rod-shaped bacteria could often be seen adhering to the tips and lateral surfaces of villi. This was best seen by using a Giemsa stain.(Figure 3). Prevalence 33 Pilus Antigens The K99 pilus antigen was the most frequently encoun- tered in this study (23 of 125), and the K88 (14 of 125) and 987P (13 of 125) pili were encountered less frequently (Table 6). 1312 of the 125 pigs examined, both K88 and K99 pilus antigens were present simultaneously. 13145 speci- mens, a piliated §_._ 3311 was judged to be the only entero— pathogen responsible for diarrheal disease. In 7 other cases, a piliated 33 coli concomittant with another . ' - n. ' - .~.-.- .I'J: ‘. .' I. _'| .a . . IL. 2| .1: '. ’- -. ... - " I run.- . J 0' I. "' '- .. -- --——-l--——. . - ~ I l . . . J l '-'- .. - . :‘I. .‘_ , . r v 551 .l I ._ .I .. VD:— _: . . . ' ' 4. . . I . . . n. .- n . u a _ u . 1 J I 37 enteropathogen was responsible for enteric disease. In 37 specimens, pathogens of the gastrointestinal tract other than piliated §_._' gall were judged to be the cause of diar- rhea. Of these 37, coronavirus was demonstrated in 18 pigs, coccidiosis was diagnosed in 3 pigs, and other causes of disease were felt to be present in 16 pigs. This category included septicemias due to Streptococcus Spa, milk produc- tion problems in the sow, or nutritional problems in the herd in question. In the remaining 36 pigs examined, no enteropathogen could be definitively identified to be the cause of intestinal disease. (Table 7). Table 6 - Numbers of pigs, diagnoses made, and E; coli pili identified in pigs submitted with a history of diarrhea. Diagnoses made Pili identified Colibacillosis only K88 K99 987P K8 g K9 45 ll 21 12 1 Mixed infections with colibacillosis 7 13,2c 2b 1C la Total 52 14 23 13 2 aColibacillosis and rotavirus. Colibacillosis and coronavirus. CColibacillosis and salmonellosis. 38 Table 7 - Numbers of pigs with diagnoses other than piliated E; coli infection submitted with a history of diarrhea. Coronavirus Coccidiosis Othera Undetermined 18 3 16 36 aIncludes septicemias, milk production problems, or nutritional imbalances. The number of Michigan swine herds with piliated g; gglg_as a cause of enteric disease was 27 of 55 herds exa- mined as determined by the indirect fluorescent antibody technique. The numbers of herds in which pili of g; 221i were identified are summarized in Table 8. The total number of individual pigs testing positively for piliated g; ggli was 52 of 125 pigs examined. On a herd basis, this represented a prevalence of 49%, and on an individual basis the prevalence was 42% (Table 9). Table 8 - Numbers of herds in which pili of E_. 39$} were identified. K88 K99 987P K88 & K99 K88 & 987P 7 11 5 2a 2b aBoth pilus antigens were found simultaneously in two pigs rom different herds. Pilus antigens were not found simultaneously but were found in different submissions from these 2 herds. 39 Table 9 - Prevalence of piliated E; coli in swine herds and pigs in this study. Herds with Pigs with piliated E; coli piliated E; coli 27/55a (49%) 52/125a (42%) aNumber positive/number tested. Quantitation of Lactose-Fermenting Bacteria Pigs in which piliated §;.22ll were identified had a significantly higher number of lactose-fermenting bacteria present in the ileum (p<0.05) than did pigs in which piliated g5 coli were not identified (Table 10). Bio- chemical tests on selected colonies from 10 pigs revealed g; ggli; however, bioassays for enterotoxigenicity were not performed. Table 10 - Numbers of lactose-fermenting bacteria in lO-cm loops of ilea in pigs with and without piliated E; coli. No. of pigs Meana S.E.M.b p0 value With piliated E; coli 52 7.4 0.3 Without piliated P<0.05 E; coli 73 5.6 0.3 aData expressed as mean loglo/lo cm of ileal segment. Standard error of the mean. CProbability of no difference between the means. 40 Association Indices In those pigs in which a pilus antigen of .E_. coli was identified, the mean association index was lZJL. In those pigs in which no pilus antigens were identified, the mean association index was 1.9. Statisitical analysis by the sign test demonstrated a significant difference between these two groups (Table 11). Table ll - Association indices of bacterial adhesion in pigs with and without piliated E; coli. Mean Range Mode P valuea With piliated E; coli 12.1 2-25 12 P<0.05 Without piliated E; coli 1.9 1-20 1 aProbability of no difference between the means. .4.”— 41 Figure l. Photomicrograph of a fluorescent antibody- stained section of ileum from a control pig administered E; coli bearing K99. Positive, specific fluorescence for the K99 pilus antigen is located along the tips and lateral surfaces of villi. EosinOphils in the lamina propria have nonspecific autofluorescence (160X). Figure 2. Photomicrograph of a fluorescent antibody- stained section of ileum from a control pig inoculated with trypticase soy broth only. Positive, specific fluorescence is lacking. Eosinophils in the lamina propria have nonspecific autofluorescence (160X). 42 Figure l 43 Figure 3. Photomicrograph of a section of ileum from a pig submitted with a history of diarrhea. Rod-shaped bacteria (arrows) are adherent to the tips and lateral surfaces of villi (Giemsa stain, lOOOX). DISCUSSION The increased vacuolation of the cytOplasm in sections of the intestinal epithelium was believed to be caused by therlack.of normal epithelial cell turnover, often seen in the gnotobiotic pig in which normal gut flora is absent (Moon, 1972). The hypocellularity of certain lymphoid organs may be due to lack of antigenic stimulation which may be seen in the young animal. These changes were felt to be consistent with the gnotobiotic environment in which these pigs were maintained. The indirect fluorescent antibody technique applied to intestinal sections proved to be a rapid and effective method for the diagnosis of colibacillosis caused by piliated Escherichia coli in pigs. Other workers have demonstrated that this technique is preferable to the slide agglutination test (Moon, 1978b; Isaacson gt al., 1978b) due to the fact that piliated g; ggli enter a nonpiliated phase when cultured in some kinds of artificial media (flrskov and flrskov, 1966; firskov__t__l_., 1975). The indirect fluores- cent antibody method detects the organisms in their piliated phase and is, therefore, a more reliable diagnostic tool. The results of this study indicated that piliated E; coli are responsible for diarrheal disease in Michigan neonatal pigs. The frequency with which the 3 pilus 44 45 antigens were found was approximately equal. However, the K99 pilus was encountered slightly more often than K88 or 987P. frhis is in contrast with results from another porcine study encompassing the north central United States in which the K88 pilus antigen was encountered most frequently (Francis, 1982). This study included pigs with diarrhea that were younger and older than 2 weeks of age, and this fact alone may have been responsible for the K88 pilus being most commonly found. It has been speculated that K88 is more comnmnxin.the weanling pig, whereas 987P tends to be found in pigs less than 1 week of age (Francis, 1982). Perhaps the fact that the Michigan study excluded pigs beyond 2 weeks of age accounted for the lower than expected numbers of pigs with _E_:_._ coli bearing the K88 pilus. It is of interest that 10 of the 13 pigs with 987P in our study were 1 week of age or less. The fact that the K99 pilus antigen was most frequently encountered in this study may be due to the relatively small number of pigs available for this study. Perhaps if more pigs with colibacillosis were available for pilus antigen determination, a different prevalence pattern may have emerged. Another possible explanation may be that many of the pigs examined were from small farms in which other livestock, including cattle, were present. Because E; coli bearing K99 is known to be infective for cattle, perhaps the presence of cattle and pigs on the same premises is repon— sible for our observation.that K99 was the most commonly . -..=._.. - -- ..-.. . .- . .'r. - " .. -.4:'.L_ - _ . . . . . '_ ' . ' . . ' '-‘ ( u I. I. :4 I -.~- _ -. '1 “:6": I ..I'. - . _ -. . . in ." . ' ._ . .- _ ‘ _. ' .I. (:1 ..I .‘i, u . - . 'V . I a . - - I. a - 'lUn .' ll.— . ; - - 1' - - . .._.. .; . .' L . _- ..:-:~ . x l I ' I I . . I-. . - . \'.. I...'II- ..f E ..' . . . - 51- :".u.... u'. .I . .-.. \ . . . . . - . I I I 1 l‘ . . - .. - . .L--'. I "'- . . . '- -". .-.'. . . . . . . . ‘ -:I ll . - I I I I I II I . I l . .' I ' -. . . . . .‘ .I -: . " ...: -. I .1. -. 1:": . '- 'I""J. . .I ..u ‘.-\- iJf“ -. .'.‘ .. . .'Ii'I-. :1 . . ' l:-' a . . . l - I . l - ' I 46 encountered pilus antigen. Nevertheless, the results of this and other studies (Francis, 1982; Moon, 1978b; flrskov _5 31., 1961) justify the use of trivalent vaccines to prevent the intestinal colonization in neonatal pigs of g; ggli bearing K88, K99, or 987P. In diarrheic pigs from the same farm there was a tendency for the same pilus antigen to be detected from all pigs. For example, a single submission of 5 pigs from the same owner yielded _E_. 9911 bearing K99 (See Appendix). An- other submission of 3 pigs from a different farm all had E; coli with 987P. 'This same observation tended to be true for pigs submitted within the same litter. Pigs with other enteropathogens (rotavirus, coronavirus, or Salmonella sp.) concomittant with piliated g; 991$ were not necessarily colonized with one particular pilus type. The K88, K99, or 987P pili were found simultaneously with other infectious agents known to cause enteric disease in the pig. The distributions of E; coli with each of the 3 pili were nearly the same in mixed infec- tions as in those infections involving exclusively g; 291}. This observation is in agreement with a study by Francis (1982). Those specimens in which 2 pilus antigens were present in the same sections of ileum are of particular interest. Apparently, more than 1 strain of piliated E; cell may colonize the small intestine simultaneously. These specimens probably represent 2 different strains of 47 enterotoxigenic g; 3911, each bearing a different pilus antigen rather than 1 strain of the organism with 2 dif- ferent pili on each bacterium (Moon, 1982). However, it has been reported by Schneider and To (1982) that 2 enterotoxi- genic strains of E; coli were both found to express 2 dif- ferent pilus antigens at different times. ‘This indicates that some field isolates of E; coli have the potential to produce more than 1 type of pilus antigen that promotes intestinal colonization, but it has not been rigorously established if such a strain produces 2 pilus antigens simultaneously. In pigs in which piliated g; 9911 were identified, the numbers of bacteria able to ferment lactose were sometimes lower than expected. Some workers have quantitatively defined intestinal, colonization in tflua pig by enterotoxigenic g; 9911 (ETEC) as 108 ETEC per 10 cm segment of ileum (Moon 3;: 31., 1979). The fact that total numbers of lactose-fermenters per 10 cm of ileum in pigs harboring g; 9913 with pilus antigens was sometimes less than 108 may have been associated with the administration of oral anti- bacterial agents to pigs prior to submission to the Animal Health Diagnostic Laboratory. This same factor also could cause a false negative fluorescent antibody result if num- bers of piliated E_. coli were substantially diminished in those areas of the ileum from which sections were taken for fluorescent antibody determinations. 48 Some pigs that tested negatively for pilus antigens gave a higher than expected association index value. Such cases may represent g; 3911 with pilus antigens that are previously uncharacterized. These antigens would not be detected by our fluorescent antibody technique. (Evidence for new, uncharacterized pili of g; 9911 is increasing (Moon, 1978b; Morris 35 gg,, 1982). Adhesins other than pili may also account for these results, but bacterial adherence by mechanisms other than pili has not been well elucidated. bearing K88 tend to colonize both the upper and lower small intestine, whereas §_._ 9911 with K99 or 987P colonize only the lower half of the small intestine. The observations in our control pigs concurred with this, as the pigs exposed to g; 9911 bearing K88 (strain 1248) were observed to have rod- shaped bacteria adherent to the sides and tips of villi when Giemsa-stained ileal and duodenal sections were viewed microscopically; However, duodenal sections from those pigs administered E; coli bearing K99 or 987P pili lacked such adherent bacteria. Some workers have associated more virulence with those strains of g; ggli bearing the K88 pilus than with those bearing K99 or 987P (Moon, 1982). The fact that our only deaths among control animals were those administered K88- bearing g; 2911 supports this contention. These animals died less than 18 hours after administration of control inocula. Wilson (1982) recently demonstrated that some 3; 49 coli with K88 pili produced both LT and ST, but strains with either K99 or 987P produced only ST. The observations that enterotoxigenic g; 9911 bearing the K88 pilus produce 2 types of enterotoxins and colonize a greater length of the small intestine may confer increased virulence to those strains and may explain the deaths in our control animals. It may also explain the lower prevalence of K88 in this study, as only pigs submitted alive were used for pilus antigen determination. Perhaps if dead pigs with histories of diarrhea had been included, K88 may have occurred more frequently. More studies correlating virulence with toxin type and with type of pili expressed need to be conducted. 1. III— S UMMARY A study of 125 pigs 2 weeks of age or less with histories of diarrhea was performed to determine the prevalence of the 3 pilus antigens of Escherichia coli (K88, K99, and 987?). An indirect fluorescent antibody technique on frozen ileal sections was utilized. Similar sections of intestinal tissue from gnotobiotic pigs orally infected with pure cultures of £291; of known pilus type were used for controls. Sections of ileum 10 cm in length were used to determine total numbers of lactose-fermenting bacteria in pigs with and without piliated g; 9911 as determined by our indirect fluorescent antibody technique. The results of this study indicate that the indirect fluorescent antibody technique is a rapid and sensitive diagnostic tool for the detection of K88, K99, and 987P pilus antigens of E 3911. Pigs in which piliated _E_._ coli were identified had a significantly higher number of lactose-fermenting bacteria (P<£.05) in 10 cm sections of ileum than did pigs in which piliated E; 9911 were not detected. In addition, all 3 pilus antigens of E. coli were found in Michigan neonatal pigs with diarrheal disease. 'The K99 pilus antigen was encountered slightly more often than K88 or 987P. These findings justify the use of trivalent SO 51 pilus vaccines containing purified K88, K99, and 987P pilus antigens to decrease the incidence and severity of enteric disease caused by enterotoxigenic E; coli. BIBLIOGRAPHY B IBLIOGRAPHY Alderete JF, Robertson DC: Purification and chemical characterization of the heat stable enterotoxin produced by porcine strains of enterotoxigenic Escherichia coli. Infect Immun 19:1021-1030, 1978. Amad-Maselmeh M, Moon HW, Runnels, PL, Schneider RA: Pilus production, hemagglutination, and adhesion by porcine strains of enterotoxigenic Escherichia coli lacking K88, K99, and 987P antigens. Infect Immun 35:305-313, 1982. Barnum DA, Glantz PJ, Moon Hw: Colibacillosis. CIBA Vet Monogr Ser 2, 1967. Bertschinger HU, Moon HW, Whipp SC: Association of Escherichia coli with the small intestinal epithelium. I. Comparison of enteropathogenic and nonenteropathogenic porcine strains in pigs. Infect Immun 5:595-605, 1972. Bijlsma IGW, deNijs A, vanderMeer C, Frik JF: Different pig phenotypes affect adherence of Escherichia coli to jejunal brush borders by K88ab, K88ac, or K88ad antigen. Infect Immun 37:891-894. 1982. Brinton CC Jr: Nonflagellar appendages of bacteria. Nature 183:782-786, 1959. Brinton CC Jr: Contributions of pili to the specificity of the bacterial surface and a unitary hypothesis of conjugal infectious heredity. In The Specificity of Cell Surfaces, Edited by BJL.Davis and L. Warren, Prentice-Hall, Englewood Cliffs, New Jersey, pp. 37- 70, 1967. Brinton CC Jr: The piliation phase syndrome and the uses of purified pili in disease control. In Proceedings 2; the XIIth Joint UJL-Japan Conference on Cholera, National Institutes of Health, Bethesda, Maryland, pp. 33-70, 1978. 52 53 Burgess MN. Bywater RJ, Cowley CM, Mullan NA, Newsome PM: Biological evaluation of a methanol-soluble, heat- stable Escherichia coli enterotoxin in infant mice, pigs, rabbits and calves. Infect Immun 21:526-531. 1978. Burrows MR. Sellwood R, Gibson RA: Haemagglutinating and adhesive properties associated with the K99 antigen of bovine strains of Escherichia coli. J Gen Microbiol 96:269-275. 1976. Christie BR, Waxler GL: Experimental colibacillosis in gnotobiotic baby pigs. II. Pathology. CanJ Comp Med 37:271-280. 1972. Dean EA, Isaacson RE: In yitrg adhesion of piliated Escherichia coli to small intestinal villous epithe11a1 cells from rabbits and the identification of a soluble 987P pilus receptor-containing fraction. Infect Immun 36:1192-1198, 1982. Donta ST, Moon HW, Whipp SC: Detection of heat labile Escherichia coli enterotoxins with the use of adrenal cells in tissue culturen Science 183:334- 335.1974. Duguid JP, Smith IW, Dempster G, Edmunds PN: Non-flagellar filamentous appendages (”fimbriae”) and haemagglutinating activity in Bacterium coli. J Pathol Bacteriol 70:335-348. 1955. Dunne HW, Bennett PC: Colibacillosis and edema disease. In Diseases of Swing, Edited by H. W. Dunne, 3rd ed.. Iowa State University Press, Ames, Iowa, pp. 587- 616. 1970. Dunne HW: Colibacillosis and edema disease. In Diseases 2; Swing, Edited by H.W. Dunne and A.D. Leman, 4th ed.. Iowa State University Press, Ames, Iowa, pp. 650- 686. 1975. Ellens DJ, DeLeeuw PW, Rosemond H: The K99 antigen of Escherichia coli: Application of enzyme-linked immunosorbent assay (ELISA) for detection of the antigen in calf faeces and for titration of specific antibody. Proceedings Second International Sym- posium 9n Neonatal Diarrhea, University of Saskat— chewan, Saskatoon, Saskatchewan, Canada, pp. 57-72, 1978. Evans DG, Evans DJ Jr, Pierce NF: Differences in the response of rabbit small intestine to heat-labile and heat-stable enterotoxins of Escherichia coli. Infect. Immun. 7:873-880, 1973a. .‘dll I . A 1— , .. _ I . I 4 . . . I 1 . f) .c l b, I. . . CA In .0. O. I 4 . . l _ t .l . a r 54 Evans DJ, Evans DG, Gorbach. SL: Production of vascular permeability factor by enterotoxigenic Escherichia coli isolated from man. Infect Immun 8:725-730. 1973b. Evans DJ, Evans DG, Richardson SH, Gorbach SL: Purification of the polymyxin-released, heat labile enterotoxin of Escherichia coli. J Infect Dis 133:597-5102, 1976. Evans DJ, Evans DG: Inhibition of immune hemolysis: serological assay for the heat-labile enterotoxin of Escherichia coli. J. Clin. Microbiol. 5:100-105, 1977. Finkelstein RA, LaRue MK, Johnston DW, Vasil ML, Cho GJ, Jones.JR: IIsolation.and prOperties of heat labile enterotoxins from enterotoxigenic Escherichia coli. J Infect Dis 133:5120-5137. 1975. Francis DH: Prevalence of K88. K99, and 987P-positive Escherichia coli in piglets with colibacillosis. In 63rd Annual Meeting gf the Conference 9; Research Workers in Animal Disease, Chicago, Illinois, p. 43, 1982 (abstract). Gianella RA: Specificity of the suckling mouse assay for ST toxin and mode of action for the toxin. Gastroenterology 72:A-39/1062, 1977. Gibbons RA, Jones GW, Sellwood R: An attempt to identify the intestinal receptor for the K88 adhesin by means of a haemogglutination inhibition test using glyCOproteins and fractions from sow colostrum. J Gen Microbiol 86:228-240. 1975. Gibbons RA, Sellwood R, Burrows MR, Hunter, PA: Inheritance of resistance to neonatal E; coli diarrhea in the pig: examination of the genetic system. Theor Appl Genet 51:65-70. 1977. Greenberg HB, Sack DA, Rodriguez W, Sack RB, Wyatt RG, Kalica AR, Horswood RL, Chanock RM, Kapikian AZ: Microtiter solid-phase radioimmunoassay for detection of Escherichia 9911 heat-labile enterotoxin. Infect Immun 17:541-545. 1977. Guerrant RL, Brunton LL, Schnaitman JC, Rebhun LI, Gilman AG: Cyclic adenosine monophOSphase and alteration of Chinese hamster ovary cell morphology: a rapid, sensitive 13 vitro assay for the enterotoxins of Vibrio cholera and Escherichia coli. Infect Immun 10:320-327. 1974. 55 Guerrant RL, Moore RA. Kirschenfeld PM, Sands MA: Role of toxigenic and invasive bacteria in acute diarrhoea of childhood. New England J Med 293:567-573. 1975. Guinée PAM, Jansen WH, Agterberg CM: Detection of the K99 antigen by means of agglutination and immunoelectrophoresis in Escherichia coli isolates from calves and its correlation with enterotoxigenicity. Infect Immun 13:1369-1377. 1976. Guinée PAM, Veldkamp MJ, Jansen WH: Improved Minca medium for the detection of K99 antigen in calf enterotoxigenic strains of Escherichia coli. Infect Immun 15:676-678. 1977. Guiné’e PAM, Jansen WH: Behavior of Escherichia coli K antigens K88ab, K88ac, and K88ad in immunoelectrophoresis, double diffusion, and hemagglutination. Infect Immun 23:700-705. 1979. Gyles CL, Barnum DA: A heat labile enterotoxin from strains of Escherichia coli enteropathogenic for pigs. J Gyles CL: Heat-labile and heat-stable forms of enterotoxin from Escherichia coli strains enterotoxigenic for pigs. Ann NY Acad Sci 176:314-322, 1970. Gyles CL: Relationships among heat-labile enterotoxins of Escherichia coli and Vibrio cholerae. J Infect Dis 129:277-283, 1974. Hamilton DL, Johnson MR, Forsyth GW, Roe WE, Nielsen NO: The effect of cholera toxin and heat labile and heat stable Escherichia coli enterotoxin on cycliclAMP concentrations in small intestinal mucosa of pig and rabbit. Can J Comp Med 42:327-331. 1978. Hughes JM, Murad F, Chang B, Guerrant RL: Role of cyClic GMP in the action of heat stable enterotoxin of Escherichia coli. Nature 217:755-756. 1978. Isaacson RE: K99 surface antigen of Escherichia coli: purification and partial characterization. Infect Immun 15:272-279. 1977. ‘ Isaacson RE, Moon HW, Schneider RH: Distribution and virulence of Escherichia coli in the small intestine of calves with and without diarrhea. AmlJ Vet Res 39:1750-1755. 1978a. 56 Isaacson RE, Schneider RA, Moon HW: Escherichia coli in the small intestines of calves with and without diarrhea: identification of virulence factors. In Proceedings, Second International Symposium _c_>__n Neonatal Diarrhea, University of Saskatchewan, Saskatoon, Saskatchewan, Canada. pp. 17- 26. 1978b. Isaacson RE, Fusco PC, Brinton CC, Moon HW: _I_9_ y_i__t__r__g adhesion of Escherichia coli to porcine small intestinal epithelial cells: pili as adhesive factors. Infect Immun 21:392-397. 1978c. Isaacson RE: Pili of enterotoxigenic Escherichia coli. In Neonatal Diarrhea, University of Saskatchewan, Saskatoon, Saskatchewan, Canada, pp. 213- 226, 1980. Isaacson RE: Personal communication, 1981. Jensen CO: C.O. Jensen's selected papers, 1886-1908, Vol. 1. Ejnar Munksgaard, Copenhagen, 290, 1948. Jones GW, Rutter JM: Role of the K88 antigen in the pathogenesis of neonatal diarrhea caused by Escherichia coli in piglets. Infect Immun 6:918- 927. 1972. Kapitany RA, Forsyth GW, Scott A, McKenzie SF, Worthington RW: Isolation and partial characterization of two different heat-stable enterotoxins produced by bovine and porcine strains of enterotoxigenic Escherichia coli. Infect Immun 26:173-177. 1979. Kenworthy R, Allen WD: The significance of Escherichia coli to the young pig. J Comp Pathol 76:31. 1966. Keusch GT, Donta ST: Classification of enterotoxins on the bases of activity in cell culture. J Infect Dis 131:58-63. 1975. Klipstein FA, Engert RF: Immunological interrelationships between cholera toxin and heat-labile and heat- stable enterotoxin of coliform bacteria. Infect Immun 18:110-117. 1977. Kohler EM: Pathogenesis of neonatal enteric colibacillosis of pigs. J Am Vet Med Assoc 160:574-582, 1972. Kohler EM: Neonatal enteric colibacillosis of pigs and current research on immunization. J Am Vet Med Assoc 173(5):588-59l. 1978. 57 Ifimmn.AD: Swine health: common diseases affecting baby pigs - a research review. College of Veterinary Medicine. University of Illinois, Urbana-Champaign, Illinois, pp. 18-33, 1970. Mills KW, Phillips RM, Kelly BL, Baughman GL: Using enzyme- linked immunosorbent assay to detect Escherichia coli K88 pili antigens from clinical isolates. .Am J Vet Res 43:365-367. 1982. Mooi FR, deGraaf FK: Isolation and characterization of K88 antigens. FEMS Microbiol Lett 5:17-20. 1979. Moon HW, Whipp SC: Development of resistance with age by swine intestine to effects of enteropathogenic Escherichia coli. J Infect Dis 122:220-223. 1970. Moonlflh Nielsen NO, Kramer TT: Experimental enteric colibacillosis of the newborn pig: histopathology of the small intestine and changes in the plasma electrolytes. Am J Vet Res 31:103-112. 1970. Moon HW: Vacuolated villous epithelium of the small intes- tine of young pigs. Vet Pathol 9. 3-21. 1972. Moon HW, Nagy B, Isaacson RE, Orskov I: Occurrence of K99 antigen on Escherichia coli isolated from pigs and colonization of pig ileum by K99 positive enterotoxigenic E; 9911 from calves and pigs. Infect Immun 15:614-620. 1977. Moonlflh Mechanisms in the pathogenesis of diarrhea: a review. J Am Vet Med Assoc 172(4):443-448. 1978a. Moon HW: Pili as protective antigens in vaccines for the control (HE enterotoxigenic Escherichia coli infections. In Proceedings, Second International Symposium on Neonatal Diarrhea, University of Saskatchewan, Saskatoon, Saskatchewan, Canada, pp. 393-405. 1978b. Moon HW, McClurkin AW, Isaacson RE, Pohlenz J, Skartvedt SM, Fillette KG, Baetz, AL: Pathogenic relationships of rotavirus, Escherichia coli and other agents in mixed infections in calves. J Am Vet Med Assoc 173:577-583. 1978. Moon HW, Isaacson RE. Pohlenz J: Mechanisms of association of enteropathogenic Escherichia cxfli. within intestinal epithelium. Am J Clin Nutr 32:119-127. 1979. 58 Moon HW, Kohler EM, Schneider RA, Whipp SC: Prevalence of pilus antigens, enterotoxin types, and enteropathogenicity among K88-negative enterotoxigenic Escherichia coli from neonatal pigs. Infect Immun 27:222-230. 1980. Moon HW: Personal communication, 1982. Morgan RL, Isaacson RE, Moon HW, Brinton CC, To CC: Immunization of suckling pigs against enterotoxigenic Escherichia coli-induced diarrheal disease by vaccinating dams with purified 987 or K99 pili: protection correlates with pilus homology of vaccine and challenge. Infect Immun 22:771-777. 1978. Morris JA, Thorns C, Scott AC, Sojka WJ, Wells GA: Adhesion in vitro and in vivo associated with an adhesive ant1gen (F41) produced by a K99 mutant of the reference strain Escherichia coli B41. Infectirmmun 36:1146-1153, 1982. Nagy B, Moon HW, Isaacson RE: Colonization of porcine small intestine by Escherichia coli: ileal colonization and adhesion by pig enteropathogens that lack K88 antigen and by some acapsular mutants. Infect Immun 13:1213-1220, 1976. Nagy B, Moon HW, Isaacson RE: Colonization of porcine intestine by enterotoxigenic Escherichia coli: selection of piliated forms in Vivo, adhesion of piliated forms to epithelial cells in vitro. and incidence of a pilus antigen among porcine enteropathogenic E; coli. Infect Immun 16:344-352. 1977. Nagy B. Moon HW, Isaacson RE, To CC, Brinton CC: Immunization of suckling pigs against enterotoxigenic Escherichia coli infection by vaccinating dams with purified pili. Infect Immun 21:269-274. 1978. NewsomeiPM, Burgess MN, Bywater RJ, Cowley CM, Mullan NA: Studies of the biological activities and mechanism of action of heat stable E_._ coli enterotoxins. In Proceedings, Second International Symposium 92 Neo- natal Diarrhea, University of Saskatchewan. Saska- toon, Saskatchewan, Canada, pp. 119-133, 1978. Nielsen NO, Moon HW, Roe WE: Enteric colibacillosis in swine. J Am Vet Med Assoc 153:1590-1606, 1968. ll— 59 Orskov I, Orskov F3 Sojka WJ, Leach JM: Simultaneous occurrence of E; coli B and L antigens in strains of diseased swine. Acta Pathol Microbiol Scand 53:404- 422, 1961. Orskov I, firskov F, Sojka WJ, Wittig W: K antigens K88ab(L) and K88ac(L) in E; 921i. Acta Pathol Microbiol Scand 62:439-447. 1964. Orskov I, Orskov F: Episome-carried surface antigen K88 of Egghggighlgplggll. I. Transmission of the determinant of the K88 antigen and influence on the transfer of chromosomal markers. J Bacteriol 91:69- 75, 1966. flrskov I, firskov F, Smith HW, Sojka WJ: The establishment of K99, a thermolabile transmissible Escherichia coli K antigen, previously called "KCO", possessed by calf and lamb enteropathogenic strains. Acta Pathol Microbiol Scand Sect Microbiol Immunol 83:31- 36. 1975. Rutter JM, Burrows MR, Sellwood R, Gibbons, RA: A genetic basis for resistance to enteric disease caused by E; coli. Nature 257:135-136. 1975. Rutter JM, Jones GW, Brown GTH, Burrows MR, Luther PD: Antibacterial activity in colostrum and milk associated with protection of piglets against enteric disease caused by K88-positive Escherichia coli. Infect. Immun. 13:667-676. 1976. Salit IE, Gotschlich EC: Hemagglutination by purified type I Escherichia £911 pili. J Exp Med 146:1169-1181, 1977a. Salit IE, Gotschlich EC: Type I Escherichia coli pili: characterization of binding to monkey kidney cells. J Exp Med 146:1182-1194, 1977b. Schneider RA, To SCM: Enterotoxigenic Escherichia coli strains that express K88 and 987P pilus antigens. Infect Immun 36:417-418. 1982. Sellwood R, Gibbons RA, Jones GW, Rutter JM: Adhesion of enteropathogenic Escherichia coli to pig intestinal brush borders: the existence of two pig phenotypes. J Med Microbiol 8:405-411, 1975. Sellwood R: The interaction of the K88 antigen with porcine intestinal epithelial cell brush borders. Biochemica et Biophysica Acta 632: 326- -335. 1980. 6O SKEUMHIFJ, Formal SB. Falkow S: Plasmid-associated enterotoxin production in a strain of E. coli isolated from humans. Infect Immun 5:622-6247 1972. Smith HW, Jones JFT: Observations on the alimentary tract and its bacterial flora in healthy and diseased pigs. J Path Bact 86:387-412. 1963. Smith HW, Halls S: Studies on Escherichia coli enterotoxin. J Bact Path 93:531-543. 1967. Smith HW, Halls S: The transmissible nature of the genetic factor in Escherichia coli that controls enterotoxin production. J Gen Microbiol 52:319-334. 1968. Smith HW, Linggood MA: Observations on the pathogenic properties of the K88, Hly and Ent plasmids of Escherichia coli with particular reference to porcine diarrhea. .J Med Microbiol 4:467-485. 1971. Smith HW, Linggood MA: Further observations on Escherichia coli enterotoxins with particular regard to those produced by atypical piglet strains used by calf and lamb strains: the transmissible nature of these enterotoxins and of a K antigen possessed by calf and lamb strains. J Med Microbiol 5:243-250, 1972. Sojka WJ: Escherichia coli in domestic animals and poultry. Commonwealth Agricultural Bureaux, Farnham Royal, Bucks.. England, 1965. Steel RGD, Torrie JM: Principles and Procedures _o_f_ Statistics, McGraw-Hill Book Co., Inc., New York, 1960. Stenun AL, Parker CW, Kipnis DM: Radioimmunoassay for cyclic nucleotides. J Biol Chem 247:1106-1113. 1972. Stirm S, Orskov F, flrskov I, Mansa B: Episome-carried surface antigen K88 of Escherichia coli. 11. Isolation and chemical analysis. J Bacteriol 93(2):731-739. 1967a. Stirm S, Orskov F, Orskov I, Birch-Anderson A: Episome- carried surface antigen K88 of Escherichia coli. III. Morphology. J Bacteriol 93:740-748. 1967b. Tennant B: Neonatal enteric infections caused by Escherichia coli. Ann NY Acad Sci 176, 1971. Waxler GL, Schmidt DA. Whitehair, CK: Technique for rearing gnotobiotic pigs. Amer J Vet Res 27:300-307. 1966. 61 Whipp SC, Moon HW, Argenzio RA, Robertson DC: Inhibition of response of swine jejunum to purified Egghggichia £911 heat-stable enterotoxin by homologous crude culture filtrates. In Proceedingsy Third International Symposium 92 Neonatal Diarrhea, University of Saskatchewan, Saskatoon, Saskatchewan, Canada, pp. 319-323. 1980. Wilson MR: Enteric colibacillosis. In Diseases 9f Swine, Edited by A.D. Leman, R.D. Glock, W.L. Mengeling, R.H.C. Penny, E. Scholl, and B. Straw. 5th ed.. Iowa State University Press, Ames, Iowa, pp. 471-477, 1981. WilmM1RA: An evaluation of serotype, toxin, and pili expressed by E; 9911 isolates from porcine colibacillosis. In gggg Annual Meetigg pf the Conference pf Research Workers in Animal Disease, Chicago, Illinois, p. 43. 1982 (abstract). Yolken RH, Greenberg HB, Merson MH, Sack RB, Kapikian AZ: Enzyme-linked immunosorbent assay for the detection of Escherichia coli heat-labile enterotoxin. J Clin Microbiol 6:439-444. 1977. VITA VITA The author was born in Berrien Springs, Michigan, on July 13. 1953. and completed primary and secondary education there. :He graduated from the College of Veterinary Medicine at Michigan State University in 1978 and returned to Berrien Springs to engage in private practice. In 1981, the author came to Michigan State University where he began a research project with Dr. Glenn L. Waxler on colibacillosis in neonatal pigs. The author is currently pursuing doctoral studies at Michigan State University. 62 APPENDIX 63 Preparation 9; Antibodies The monospecific anti-pilus antibodies were prepared using 8 New Zealand White rabbits. Preparations of purified pilus antigens K88. K99, and 987P were obtained from Dr. R. E. Isaacson, School of Public Health, University of Michigan, Ann Arbor. Rabbits were divided into 4 groups, with 1 group serving as noninjected controls (Table A1). Antigens were received as suspensions in varying concentrations. Each injected rabbit received a volume containing 250 119 of antigen subcutaneously with an equal volume of complete Freund's adjuvant. The antigen and adjuvant were emulsified prior to injection. Not more than 0.2 ml was given per injection site. Injections were repeated again 4 weeks later. All rabbits were bled 7-10 days after the second injection. Blood was centrifuged, and the serum was removed. Serum was filtered through a 0.4511 m filter and stored in 10 aliquots of 2 ml each in a freezer at -7o°c. Table A1 - Distribution of rabbits injected with purified pili. Rabbit number Pilus injected 1 None 2 None 3 K99 4 K99 5 K88 6 K88 7 987P 8 987P 64 Data from Fluorescent Antibody Determinations The data obtained in this study from individual pigs and herds are summarized in Table A2. Table A2 - Data from fluorescent antibody determinations AHDL Case No.3 Herd No. Pilus IdentifiedP 234992 1 None 235142 2 K99 235143 2 K99 235144 2 K99 235242 3 K88 235243 3 K88 235489 4 K88 235490 4 K88 235632 2 None 236152 5 K99 236153 5 K99 236363 6 None 236364 6 None 236564 7 987P 236734 8 K88 237315 9 None 237316 9 None 237317 10 None 238390 11 K99 238392 11 K99 238394 11 K99 239469 12 None 239471 12 None 240128 13 None 240507 14 None 241406 15 None 241407 15 None 242427 2 None 242635 16 None 242610 17 None 243773 18 K88 244002 19 None 245354 20 987P 245355 20 987P aEach Animal Health Diagnostic Laboratory case number repregents an individual pig. As determined by the indirect fluorescent antibody technique used in this study. AHDL Case No. 260588 260589 260590 260988 260989 261203 263533 263534 266191 266192 267032 267033 269555 269556 272951 272952 273169 275189 278087 278088 278089 278927 278928 278975 278981 278982 279616 279795 273796 279797 279798 279799 282606 282607 282997 282998 286367 286368 292502 292503 292504 292505 293536 293537 293583 295054 295721 295909 297708 297709 301362 Herd No. 65 Pilus Identified None None None None None K88 K88 K88 None None None None None None None None None None None None None None None None None None None K99 K99 K99 K99 K99 None None None None K99 K99 987P 987P 987P None K88 K88 None None 987P None K99 None None 66 IUHM; Case No. ngg N9; Pilus Identified 301363 40 None 301364 40 None 301365 40 None 301366 40 None 301367 40 K99 301368 40 None 301544 41 K88 & K99 301545 41 K88 302214 42 K88 302215 42 None 303812 43 None 304415 44 K99 304416 44 K99 308147 45 None 309594 43 None 309882 46 None 309883 46 None 309884 46 None 310925 47 None 310926 47 None 310927 47 ' None 311808 48 K88 312316 49 K99 315323 50 K99 315324 50 K99 320149 51 987P 320150 51 987P 320151 51 None 320779 52 K99 320780 52 None 320781 52 None 320782 52 None 320952 53 None 320953 53 None 321138 54 K88 & K99 327155 36 987P 327156 36 None 327677 55 987P 327678 55 987P 327679 55 987P 03056 2569 2 3