x i «V . . . ‘ , V . . v. ., $2 m. ,. . . , . , , g3? .t . ‘ . . . V . ‘ ‘ . . . Hi. ‘ . ‘ . .. . . nth. u. A .X mi ..MflL ..«w. town»... :1... 1 .. . 1:3...“ ......T. shivfa 1.451 ‘11.... .4. i .6: 1.14.3... .... ..., :i I: ...}. $191.! . .- vs: T... .. .. ’ E.§ ”A E I. = A1: .143. I 3.1.- ..q! x .ux .-..d. ,‘.r.. .. . ‘... l.. . v. A, .. V . .I .. €1.11... . ‘ ..v. .N‘. . : _. . x ‘ K a _ . ... a . 2.. u .. .h 1. A Van’s; x. .. ,1; .1 z... «.4 . 5, . .. t ‘ . , ..S)‘. , . hwmfikfi. .mmnwfiun. , ti..2._ , in? am: «8 Wm. “.2 ....fifimamuwfi mil»? .W¢.aer¢wfum ., “53%me . Ag «39%, 53%.... .21.. x . . .. 3.. w .... ,3... . .3; . ‘ .. , . . .. fl : . . . . I I e A , I . I ...i: \ 1.5.. T118338 This is to certify that the dissertation entitled lGF-l AND LEPTIN HORMONAL INFUSIONS ALTER THE CELL PROLIFERATION AND TRANSCRIPTIONAL PROFILE OF THE PREPUBERTAL BOVINE MAMMARY GLAND presented by Brett Eric Etchebame has been accepted towards fulfillment of the requirements for the Ph. D. in Animal Science Mfg/(UL: FQlajor Professor’s‘Signature JFFULAVI) Z ng / Date MSU is an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State University - ..-v-.-.-u-n-0—o-o—-o-o-.--—c-o-u-o-n-o-a--—vc- — PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 2/05 p:/C|RCIDateDue.indd-p.1 IGF-l AND LEPTIN HORMONAL INFUSIONS ALTER THE CELL PROLIFERATION AND TRANSCRIPTIONAL PROFILE OF THE PREPUBERTAL BOVINE MAMMARY GLAND By Brett Eric Etchebarne A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Animal Science 2006 Higfiei devetc. result a likely p ABSTRACT lGF-l AND LEPTIN HORMONAL INFUSIONS ALTER THE CELL PROLIFERATION AND TRANSCRIPTIONAL PROFILE OF THE PREPUBERTAL BOVINE MAMMARY GLAND by Brett Eric Etchebarne High-energy diets promoting body growth rates >1 kgld impair mammary development in prepubertal dairy heifers. Biological mechanisms to explain this result are not well understood, but insulin-like growth factor-I (lGF-l) and Ieptin likely play a role. Adipocytes produce the protein leptin, and leptin concentrations increase with increased fat deposition in the body and mammary gland. lGF-l stimulates and leptin inhibits proliferation of mammary epithelial cells (MEC) in vitm and in vivo in cattle, in contrast to human and mouse MEC, which proliferate in response to Ieptin. We hypothesize that Ieptin inhibits the mammogenic action of lGF-I, and diets that promote rapid prepubertal body growth inhibit mammary development if fat deposition is increased. Our objectives were to elucidate the effects of lGF-l and leptin infusion on cell cycling in MEC, and identify key genes controlling the interaction of these two hormones in bovine parenchya. A functional genomics approach was used to analyze mammary tissue collected from six prepubertal Holstein heifers after 7 d of intramammary hormone infusions of lGF-l, Ieptin, lGF-l plus Ieptin, or saline control. Mammary parenchyma tissue was collected, fixed, and immunostained for proliferation markers (BrdU and Ki-67), and an apoptotic marker (caspase-3). Infusion of lGF-l increased the percentage of BrdU-labelled MEC in the S-phase of the cell cycle 52% (Silva, 2002). Leptin infusion decreased the percentage of BrdU-labeled MEC 48% in lGF-l-treated quarters, and 19% in control quarters (Silva, 2002). No differences were observed in the total cells within the cell cycle (by Ki—67 immunostaining). A numerically-insignificant increase in caspase-3 immunostaining was seen in lGF-l and leptin treated quarters relative to control. We conclude that increased concentrations of lGF-l promote, and leptin inhibit, MEC proliferation by increasing the progression of cells from the G1 into the S- phase of the cell cycle. Gene expression profiles of total parenchyma mRNA from each quarter of the same six animals (Silva, 2002) were examined using bovine-specific microarrays. mRNA expression of genes known to mediate cell cycling and proliferation within specific pathways in the mammary gland, including the JAK/STAT pathway, Raisaf pathway, PI3K pathway, and Bad/Bcl pathway were altered by lGF-l and leptin treatments. Interestingly, addition of leptin to lGF-l treated quarters increased the mRNA of one of these genes, the suppressor of cytokine signaling (SOCS)-3, 23-fold relative to lGF-l-infused quarters. Assuming protein changes follow mRNA expression, increased 30083, which signals within the JAK/STAT cell proliferation pathway, could possibly explain the inhibition of lGF-l-induced mammary epithelial cell proliferation. Overall, this study has demonstrated that lGF-l and leptin intramammary infusions in prepubertal heifers alter the cell cycling and mRNA profiles within the mammary gland to affect MEC proliferation. Copywrite by, Brett E. Etchebarne 2006 ACKNOWLEDGMENTS Very special thanks go out to my advisor, Dr. Mike VandeHaar, for all of his help over the course of my research in all aspects of the experience. I know that I will look back at working in the VandeHaar lab as one of the most enjoyable and exciting times of my life, and I appreciate all of my time as a member of this group. Dr. VandeHaar's seemingly endless patience in the face of trial and tribulation has always amazed me and inspired me to forge ahead and fight the good fight. I thank my committee Dr. Mike Allen, Dr. Dale Romsos, Dr. Paul Coussens, and Dr. Miriam Weber Nielsen for their time, advice, criticism and patience throughout my research. Their guidance and help in shaping my graduate program has undoubtedly been essential to my intellectual and professional development and will continue to serve me well throughout my career. My unorthodox and erratic approaches to the PhD. process undoubtedly unnerved them all at times, but in the end I believe that things worked out fine. Heartfelt thanks also go to Mr. Jim Liesman not only for his help in data analysis and statistics, which were essential to my research, but also for encouragement and advice during rough times. I will always cherish the time and personal and scientific mentorship that he was able to provide me over the years. I thank my graduate student colleagues, especially Mr. Kevin Harvatine, Mr. Bill Nobis, Dr. Laurie Davis Rincker, Dr. Mike Rincker, Mr. Barry Bradford, and Dr. Betina Lew for all the times we shared in work collaboration, scientific discussions, and enthusiasm. indmC Colvzr knows techn: Encouw this we, My research could not have been completed without the help of many individuals at the Center for Animal Functional Genomics, especially Mr. Chris Colvin, Mrs. Xioaning Ren, and Mrs. Sue Sipkovsky who were very helpful and knowledgable in the use of microarray and quantitative real-time PCR technologies. A big thanks to the MSU Dairy Group, Dr. VandeHaar, Dr. Allen, Dr. Dave Beede, and Dr. Herb Bucholtz for many an entertaining and lively lunchtime roundtable discussion. Finally, I thank my parents and family for their constant support and encouragement through the good times and the bad. Without their love none of this would ever have been possible. vi USTC USTC USTC CHAP~ REflE CHAPT DEVEL ABST. INTRO TABLE OF CONTENTS LIST OF TABLES .................................................................................................. x LIST OF FIGURES .............................................................................................. xii LIST OF ABBREVIATIONS ................................................................................ xiv CHAPTER 1 Statement of Problem ............................................................................. 1 REVIEW OF LITERATURE Introduction .............................................................................................. 3 Mammary gland development ................................................................. 4 Nutrition and mammogenesis .................................................................. 6 Physiological basis for effect of feeding level on mammary gland development .......................................................... 9 Key mammary gland growth factors ...................................................... 12 Growth hormone and IGF-I .................................................................... 12 Signaling in the IGF-I family ................................................................... 17 Biological actions of Ieptin ..................................................................... 20 Leptin receptor ....................................................................................... 21 Suppressor of Cytokine Signaling (SOCS) Proteins .............................. 23 CHAPTER 2 DEVELOPMENT OF THE BOVINE METABOLISM MICROARRAY ................... 35 ABSTRACT ......................................................................................................... 35 INTRODUCTION ................................................................................................. 36 Nutrition and functional genomics .......................................................... 36 Developing a microarray ........................................................................ 37 Nutritional genomics and animal production diseases ........................... 42 Dietary factors influence gene expression ............................................. 43 Nutrition and genome interactions ......................................................... 44 Cattle metabolic disease and nutritional genomics ................................ 46 RESULTS ........................................................................................................... 47 Development of the bovine metabolism (BMET) microarray .................. 47 CONCLUSIONS .................................................................................................. 50 ACKNOWLEDGMENTS ...................................................................................... 51 WEBSITE REFERENCES .................................................................................. 52 CHAPTER 3 VALIDATION OF A METABOLISM-SPECIFIC LONG-OLIGONUCLEOTIDE MICROARRAY FOR TRANSCRIPTION PROF ILING IN CATTLE ...................... 59 ABSTRACT ......................................................................................................... 59 INTRODUCTION ................................................................................................. 60 MATERIALS AND METHODS ............................................................................ 60 vii Animals .................................................................................................. 60 Sample collection and tissue processing ............................................... 61 RNA isolation procedure ........................................................................ 61 BMET microarray experimental design .................................................. 63 Preparation of labeled cDNA for BMET hybridization ............................ 64 Quantitative real-time PCR validation of microarray gene expression changes ................................................................... 66 Statistical analysis of microarray data .................................................... 68 RESULTS ........................................................................................................... 70 Identification of genes differentially expressed in mammary and liver of lactating cows ................................................. 70 Identification of genes differentially expressed in adipose and mammary of lactating cows ............................................ 70 Identification of genes differentially expressed in adipose and liver of lactating cows ..................................................... 71 Q-RT-PCR confirms IGF-I induced expression changes in 7 of 8 genes significantly altered in microarray analysis ..................... 72 DISCUSSION ...................................................................................................... 73 Tissue specificity of expression ............................................................. 73 CHAPTER 4 IGF-I AND LEPTIN HORMONAL INFUSIONS ALTER THE CELL PROLIFERATION AND TRANSCRIPTIONAL PROFILE OF THE PREPUBERTAL BOVINE MAMMARY GLAND .................................... 89 ABSTRACT ......................................................................................................... 89 INTRODUCTION ................................................................................................. 91 MATERIALS AND METHODS ............................................................................ 93 Animals .................................................................................................. 93 Infusion procedure ................................................................................. 94 Sample collection and tissue processing ............................................... 95 BrdU immunohistochemistry .................................................................. 96 Ki-67 immunohistochemistry .................................................................. 96 Caspase—3 immunohistochemistry ......................................................... 97 Infusion study statistical analysis ........................................................... 98 RNA isolation procedure ........................................................................ 98 BOTL4 microarray experimental design .............................................. 100 BMET microarray experimental design ................................................ 101 Preparation of labeled cDNA for BOTL4 hybridization ......................... 103 Preparation of labeled cDNA for BMET hybridization .......................... 104 Statistical analysis of microarray data .................................................. 105 Quantitative RT-PCR validation of microarray gene expression changes ................................................................. 107 RESULTS ......................................................................................................... 1 10 Intramammary infusion of oLeptin decreased BrdU-labeling of mammary epithelial cells ........................................ 110 viii Cell cycle commitment is unaffected by treatment ............................... 110 Apoptosis in prepubertal mammary epithelial cells is low .................... 111 BOTL4 microarray SAL vs. IGF differentially expressed genes ........... 111 BOTL4 microarray IGF vs. IGF+LEP differentially expressed genes... 1 12 BOTL4 microarray SAL vs. LEP differentially expressed genes .......... 112 BMET microarray SAL vs. IGF differentially expressed genes ............ 113 BMET microarray IGF vs. IGF+LEP differentially expressed genes ....114 Q-RT-PCR confirms IGF-l-induced expression changes in 4 of 21 genes significantly altered by microarray analysis ................ 115 Q-RT-PCR analysis reveals potential IGF-I and Ieptin signaling pathway interaction ............................................................ 116 DISCUSSION .................................................................................................... 117 Microarray gene expression study ....................................................... 125 CHAPTER 5 SUMMARY AND CONCLUSIONS .................................................................... 169 CHAPTER 6 FUTURE RESEARCH ....................................................................................... 179 BIBLIOGRAPHY ............................................................................................... 182 APPENDIX ........................................................................................................ 205 ix LIST OF TABLES CHAPTER THREE Table 1. Liver vs. Mammary Tissue GenMAPP Analysis .................................... 77 Table 2. Adipose vs. Mammary Tissue GenMAPP Analysis ............................... 79 Table 3. Adipose vs. Liver Tissue GenMAPP Analysis ....................................... 83 Table 4. Gene specific-Q-RT-PCR Primer Pairs ................................................. 87 Table 5. Tissue comparison Q-RT—PCR Results ................................................. 88 CHAPTER FOUR Table 1. Ki-67 and BrdU labeling comparison ................................................... 152 Table 2. Genes altered by IGF-I versus SAL treatment detected by BOTL4 microarray analysis ................................................................................... 153 Table 3. Genes altered by IGF-LEP versus IGF-I treatment detected by BOTL4 microarray analysis ................................................................................... 155 Table 4. Genes altered by LEP versus SAL treatment detected by BOTL4 microarray analysis ................................................................................... 157 Table 5. Genes altered by IGF-l versus SAL Treatment detected by BMET microarray analysis ................................................................................... 159 Table 6. Genes altered by IGF-LEP versus IGF-I treatment detected by BMET microarray analysis ................................................................................... 161 Table 7. Gene specific Q-RT-PCR Primer Pairs ............................................... 164 Table 8. Microarray versus Q-RT-PCR analysis ............................................... 167 APPENDIX Supplementary Table 1. Significant BOTL4 microarray genes .......................... 220 Supplementary Table 2. Significant BMET microarray genes: SAL vs. IGF-I ....................................................................... 230 Supplementary Table 3. Significant BMET microarray genes: IGF vs. IGF+LEP ......................................................... 236 xi LIST OF FIGURES CHAPTER 1 Figure 1. The growth hormone/Insulin-like growth factor-I (GHIIGF-I) axis and the bovine mammary gland ......................................... 32 Figure 2. Insulin-like growth factor I receptor (IGF-IR) signal transduction pathways ...................................................................... 33 Figure 3. Leptin receptor signaling pathways ...................................................... 34 CHAPTER 2 Figure 1. Spotted DNA Arrays ............................................................................. 54 Figure 2. Affymetrix Arrays .................................................................................. 55 Figure 3. Microarray experimental designs for two-color arrays .......................... 56 Figure 4. Nutritional genomics model .................................................................. 57 Figure 5. Metabolic gene database assembly ..................................................... 58 CHAPTER 3 Figure 1. Tissue comparison microarray loop design .......................................... 76 CHAPTER 4 Figure 1. Representative mammary parenchyma sections immunostained for A) Bromodeoxyuridine (BrdU), B) Proliferating cell nuclear antigen (Ki-67) and C) Caspase-3 ......................................................................... 133 Figure 2. Effects of intramammary infusion of ovine Ieptin and IGF-l on bromodeoxyuridine- (BrdU) labeling of mammary epithelial cells in prepubertal heifers .................................................................................... 136 Figure 3. Effects of intramammary infusion of ovine Ieptin and IGF-l on proliferating cell nuclear antigen (Ki-67) labeling in prepubertal heifers ................................................................................ 137 Figure 4. Caspase-3 immunostaining indicates apoptosis xii ‘ in the prepubertal mammary gland ............................................................ 138 Figure 5. Effects of ovine leptin and IGF-I on the percentage of cells committed to the S-phase of the cell cycle ................................................ 139 Figure 6. Genes found to be significantly altered by treatment in mammary tissue from six prepubertal bovine heifers ............................ 140 Figure 7. A schematic of IGF-I and Ieptin intracellular signal transduction ........ 144 Figure 8. Genes differentially expressed in the SAL vs. IGF treatment comparison are significantly overrepresented in the TGF-beta signaling pathway ........................................................... 145 Figure 9. Genes differentially expressed in the SAL vs. IGF treatment comparison are significantly overrepresented in the insulin signaling pathway ................................................................. 146 Figure 10. Genes differentially expressed in the SAL vs. IGF treatment comparison are significantly overrepresented in the G1 to S Cell Cycle Control Pathway ................................................ 147 Figure 11. Genes differentially expressed in the IGF vs. IGF-LEP treatment comparison are significantly overrepresented in the TGF-beta signaling pathway ........................................................... 148 Figure 12. Genes differentially expressed in the IGF vs. IGF-LEP treatment comparison are significantly overrepresented within the insulin signaling pathway. ........................................................ 150 xiii 14-3-3 eta ANOVA BDNF BLAST BMET BOTL BrdU BW CASP8 cDNA ChREBP CISH CNS ‘ CNTF CT Cy3 Cy5 DEPC DMSO DNA dNTP dUTP E2 ERK ExPASy FABP FADS FBS FDR FRAP1 GALT GAPDH GH GHRH GO HGF IFNA2 IGF+LEP IGFBP IGF-l IGF-II IGF-IR IL IR LIST OF ABBREVIATIONS 14-3-3 protein eta Analysis of variance Brain derived neurotrophic factor Basic local alignment search tool Bovine metabolism Bovine total leukocyte Bromodeoxyuridine Body weight Caspase 8 Complementary DNA Carbohydrate responsive element binding protein Cytokine-inducible SH2 domain containing protein Central nervous system Ciliary neurotrophic factor Cycles to threshold Cytidine 3, green fluorescent dye Cytidine 5, red fluorescent dye Diethyl pyrocarbonate Dimethyl sulfoxide Deoxyribonucleic acid 2'-deoxynucleoside 5'-triphosphate 2'-Deoxyuridine 5'-Triphosphate Estradiol, estrogen Extracellular signal-regulated kinase Expert Protein Analysis System Fatty acid binding protein Fatty acid desaturase Fetal bovine serum False discovery rate F K506 binding protein 12-rapamycin associated protein 1 Lactose-1-phosphate uridyltransferase Glyceraldehyde phosphate dehydrogenase Growth hormone Growth hormone releasing hormone Gene Ontology Hematopoietic growth factor Interferon alpha 2 IGF-l plus Ieptin Insulin-like growth factor binding protein Insulin-like growth factor-I Insulin-like growth factor-ll Insulin-like growth factor receptor Interleukin Insulin receptor xiv IRS1 JAK KEGG Ki-67 LEP MAPK MEC MGE MM mRNA NCBI NF-kappa B OB-R OB-Rb oLepfin OMIM PCR PF KL PI3K PIGF PKB PKU PM PPAR PRL PTP Q-RT—PCR RefSeq RELA RNA RPL-19 RXR SAL 800 $008 SREBP STAT TGF TIGR TNF TUNEL Insulin receptor substrate 1 Janus kinase Kyoto Encyclopedia of Genes and Genomes Proliferating cell nuclear antigen Ki-67 Saline plus Ieptin Mitogen activated protein kinase Mammary epithelial cells Mammary gland extract Mismatch probe Messenger RNA National Center for Biotechnology Information Nuclear factor kappa B Leptin receptor Leptin receptor, long form Recombinant ovine Ieptin Online Mendelian Inheritance in Man Pyruvate dehydrogenase kinase 1 Phosphofructokinase, liver isoform Phosphoinositide 3'-kinase Phosphatidylinositol glycan, class F Protein kinase B Phenylketonuria Perfect match probe Peroxisome proliferator-activated receptor Prolactin Protein tyrosine phosphatase Quantitative real-time reverse-transcriptase PCR NCBI reference sequence NF-kappa B p65 subunit Ribonucleic acid Ribosomal protien L-19 Retinoic acid receptor Saline Stearoyl CoA desaturase Suppressor of cytokine signaling Sterol regulatory element binding protein Signal transducer and activator of transcription Transforming growth factor The Institute for Genome Research Tumor necrosis factor Terminal deoxynucleotidyltransferase-mediated dUTP nick-end labeling XV CHAPTER 1 Statement of Problem Feeding high-energy diets in the prepubertal growth period in dairy cattle impairs mammary gland development. Biological mechanisms to explain this result are not well understood, but insulin-like growth factor-I (IGF-I) and Ieptin likely play a role. Adipocytes produce the protein Ieptin, and Ieptin concentrations increase in response to increased fat deposition in the body and mammary gland. IGF-I stimulates and leptin inhibits proliferation of bovine mammary epithelial cells in vitro and in vivo. My objectives were the following: 1) To elucidate the effects of IGF-I and Ieptin infusion on cell cycling in mammary epithelial cells using immunohistochemical labeling of cell cycle markers. 2) Identify key genes controlling the interaction of these two hormones in bovine mammary tissue employing a functional genomics approach. 3) Determine if Ieptin alters lGF-l-induced expression of genes involved in the major IGF-I signaling pathways using Q-RT—PCR. Crosstalk between hormones is a critical area of research geared toward the understanding of the endocrine feedback loop regulating body energy stores. “Adiposity signals”, which are putative molecular substrates present in the circulation that reveal information about body condition to the brain in order to help regulate energy intake, could possibly be present in the body in the form of the hormones insulin and Ieptin. Both of these hormones are known to circulate in the body in proportion to body fat mass in the physiologically normal mammalian organism, and can act upon the CNS to inhibit food intake through their ability to signal within the brain (5:61). The suggested physiological and intracellular signaling interaction between insulin and leptin at the level of the CNS has subsequently been extended to other specific tissues, including the mammary gland. It has been well established that high energy diets increase fat deposition within peripheral tissue stores, including the mammary gland (293). Local concentrations of IGF-I and leptin are known to be increased in the mammary gland with increased energy intake (242). However, although IGF-I has been shown to have a stimulatory effect on mammary epithelial cell proliferation, mammary gland development is inhibited by high energy diets (275;278). Thus, I believe that a hereto unknown factor is increased in concentration with the feeding of high energy diets which interferes with the proliferative action of IGF-I on mammogenesis. It is here proposed that Ieptin, a hormone produced by adipocytes, could potentially mediate the inhibition of IGF- l-stimulated mammary parenchymal development through crosstalk between the IGF-I and Ieptin receptors. Evidence is presented supporting 80083 as a potential mediator of the negative feedback on IGF-IR activation promoted by Ieptin intracellular signaling. REVIEW OF LITERATURE Introduction Mammary gland development in the prepubertal phase of growth in the dairy heifer is crucial to determining future milk production (267). Genetic selection for higher milk yield in dairy cattle is correlated to growth capacity of heifers and milk yield of cows which gives rise to a positive relationship between milk yield and body weight at calving (141). This relationship suggests a positive relationship between rate of gain of heifers in the prepubertal stage and subsequent milk yield potential. However, numerous studies have established that high energy intake and subsequent high rate of gain can lead to reduced mammary development and impaired future milk yield in dairy cattle. This window of time for mammary developmental sensitivity to energy intake is isolated to the prepubertal phase, with post-pubertal and pregnancy periods being unaffected. Milk yield capacity is largely influenced by the number of mammary epithelial cells (MEC) present at puberty, which will form the foundation for the future milk producing. cell population (306). The negative relationship between prepubertal growth rate and milk yield capacity suggests that nutritional regimes can influence the MEC population and have a negative effect on mammary development. Numerous hormones act to guide mammary development including estrogen, progesterone, growth hormone (GH), insulin-like growth factor-I (IGF-I), and leptin (107;165;194;278). Studies have shown that GH levels are altered by nutrient availability and that impaired mammary development may be due to alterations in GH action (236;237). This review aims to investigate the effect of energy intake on mammary gland development and discuss potential physiological regulators mediating this effect on future milk production potential in dairy cows. Mammary gland development Postnatal bovine mammary gland development can be divided into four distinct stages of development as first described by Tucker (307). These include periods between birth and conception (the prepubertal phase), between conception and the first parturition, during the first lactation, and the dry period prior to a subsequent pregnancy. Of critical importance to future milk yield considerations is the prepubertal phase in which mammary development is most dramatically affected by energy intake (267). This phase of growth prior to puberty will be the area of focus for this review. Development of the mammary gland and its internal structures is referred to as mammogenesis. While mammogenesis is recognized as being most extensive and dramatic during pregnancy, mammary development really begins in the fetus and proceeds beyond initiation of lactation (306). One remarkable feature of the mammary gland is that it can repeatedly undergo periods of growth and functional differentiation during pregnancy and upon initiation of lactation, and regression during the dry period (307). The mammary gland contains both parenchymal and extraparenchymal structures. The parenchymal tissue is composed of epithelial structures such as alveoli and ducts, and associated stromal connective tissue (8). The stromal tissue provides a support system surrounding the epithelium containing cellular components such as fibroblasts, endothelial cells associated with blood vessels, and leukocytes localized within the tissue, and non-cellular components including collagen and other connective tissue proteins (8). The extraparenchymal region of the gland includes a considerable white adipose tissue region called the “fat pad’ which is large during early phases of development and largely disappears during pregnancy (8). A critical period of mammogenesis exists between 3-9 months of age in the growing dairy heifer. During this period the mammary gland grows at a faster rate compared with the rest of the body (allometric growth) (282). Puberty in well managed heifers occurs between 7-9 months of age, and after reaching puberty, the mammary growth rate decelerates and grows at a rate equal to that of other tissues (isometric growth) after a few estrous cycles (153:282). Quantitative measures of the total numbers of parenchymal cells in the mammary gland and thus mammary development include deoxyribonucleic acid (DNA) concentration in the parenchymal tissue (8). The amount of DNA in the gland has been shown to increase 1.6 times faster than body weight between birth and 2 months of age. Between 5-9 months of age this measurement increases to 3.5 times faster than that of body weight (BW), and from 9-12 months of age this value returns to 1.5 (282). The allometric growth phase brings about rapid growth of the mammary fat pad and ducts, but no alveoli are yet formed. This phase is tied to reproductive development of the animal, and ovariectomy in the first weeks of life almost completely abolishes mammary development (236:313). Upon initiation of puberty, the mammary glands of heifers weigh 2-3 kg of which only 0.5-1 kg is parenchymal tissue (269). This parenchymal tissue contains approximately 10- 20% epithelial cells, 40-50% stromal tissue, and 30-40% fat cells (8). The number of alveolar structures in the mammary gland is a basic element limiting milk production (305). Milk yield during lactation will be determined by the number of epithelial cells present and by the secretory activity per cell (106;167). Of primary importance to an animal’s lactation potential is cell number, because cell activity is generally not a limiting factor (201). Correlation between total mammary cell numbers and milk yield is generally high, with estimates ranging from 0.50 to 0.85 (304). It thus becomes clear that to maximize milk production potential, the development of an adequate prepubertal ductal network is of great importance to establish subsequent alveolar growth and differentiation prior to lactation. Nutrition and mammogenesis Herman and Ragsdale were among the first to observe that mammary development was stunted in cows raised on a high energy diets (137). Soon after, this assertion was confirmed by Swanson (293), Little and Kay (185), and Harrison et al. (131). Of particular importance to studies of mammogenesis were the observations that glands from cows raised on high energy diets differed in both size and shape from those of cows raised under moderate feeding regimes (1 37). Upon examination of a number of studies focusing on the effects of nutrition during the prepubertal phase of mammary development, it becomes clear that this period was especially critical to the subsequent milk yield capacity of heifers. Numerous separate experiments have confirmed that high energy feeding levels during the prepubertal period reduce mammary development (196;265;266). To investigate the influence of plane of nutrition on mammary development in Holstein heifers, Sejrsen et al. (1982) fed pre- and postpubertal heifers a ration containing 60% concentrate and 40% forage in restricted or ad Iibitum amounts (267). They found that ad libitum feeding during the prepubertal period lowered parenchyma weights 23% and parenchyma DNA concentration 32% compared with restricted feeding. Parenchyma composed 38% of the total mammary gland mass in heifers receiving the restricted feeding level compared with 23% of total mass of the heifers fed ad libitum. Feeding different diets after puberty, however, did not alter parenchymal growth (266;267). The high intensity feeding after puberty tended to increase the amount of adipose tissue, but total fat-free parenchyma was only 3% lower than for heifers fed the restricted diet (267). Two studies comparing the effect of feeding level from 300-450 kg of body weight in unbred heifers (267) and from 300 kg to 3 months prior to calving (104) showed a lack of effect of feeding level on mammary growth in the postpubertal period. Studies examining the effect of diet in the early prepubertal phase on mammary gland development are equivocal. While a study by Sejrsen et al. (1998) suggests that mammary growth is unaffected by plane of nutrition prior to 90 kg live body weight (268), Brown al. (2005) reported that increasing energy and protein intake from 2 to 8 wk of age increased mammary parenchymal mass, DNA and RNA of Holstein heifer calves (37). This study suggests that increasing protein and energy intake in Holstein heifer calves prior to the 3-9 months of age allometric growth phase can increase the rate of development of mammary parenchyma without detrimental effects on potential future milk yield. Specific factors present within a diet type fed at different levels have also been examined to determine effects on mammary development. In one study, alteration of the forage to concentrate ratio in two isoenergetic diets did not alter mammary gland growth (263). However, Capuco et al. (1995) reported that a high energy corn-based diet produced a more dramatic decrease in mammary development than a high energy alfalfa-based diet (43). Here it was suggested that the higher protein supply in the alfalfa-based diet could partially compensate for the negative effects of the high energy level on mammary development. Similarly, Radcliff et al. (1997) proposed that the minor negative effect of feeding a high energy diet on mammary development seen in their study may have been related to dietary protein levels (242). The authors suggested that high protein levels in high energy diets may provide a protective effect against inhibition of mammary gland development. The importance of ruminal protein degradability was investigated by Mantysaari et al. (1995) and the negative effect of feeding level on prepubertal mammogenesis was unaffected by the amount of by—pass protein present in the diet (196). In a study in which heifers were fed a high energy diet at three levels of protein (14, 16 and 19% of diet), those heifers on a low protein diet that attained puberty early had a 33% greater impairment of mammary parenchymal development compared to those fed a high protein diet (324). This study showed once again that while dietary protein does not have a major effect on mammary development of rapidly grown prepubertal heifers, feeding low-protein diets increases the risk of impaired mammary development when heifers are fed for rapid growth and attain puberty early. Dietary fat has also been shown to have effects on prepubertal mammary growth in ruminants. Sheep reared on diets higher in polyunsaturated fat had increased mammary parenchymal development (200). In a recent study in which diets differing in Iinoleic acid concentrations at a moderate energy level (~800 gld) were fed to growing heifers over the first 12 months of age, a tendency for increased mammary parenchymal development at 12 months of age on the lower linoleic acid concentration diet was observed which did not translate to differences in first lactation milk yield, likely because of the small magnitude of difference between diets (299). Physiological basis for effect of feeding level on mammary gland development Bovine somatotropin (bST) or growth hormone (GH), is a growth factor that is elevated in prepubertal dairy heifers fed restricted diets compared with those fed a high plane of nutrition (143;265;266). In an early study monitoring mammary development at increased dietary energy levels, prepubertal and postpubertal heifers fed either an ad libitum or 60% of ad libitum diet had different responses for average daily gain and mammary parenchymal development. Parenchymal weight was significantly lower in the ad Iibitum fed prepubertal group (265). In this study restricted heifers gained 613 gld, the ad Iibitum group gained nearly double this at 1218 g/d over the allometric growth period (265). The feeding level in this study did not affect mammary development of postpubertal heifers and mammary parenchyma fat versus protein composition was unaffected in all treatments. Concentrations of prolactin, insulin, and glucocorticoids were higher in blood in the ad Iibitum fed group for both pre- and postpubertal heifers (265). Interestingly, GH was increased in the restricted feeding prepubertal heifer group compared with the ad libitum fed group. No difference in GH levels was detected in the postpubertal groups. The amount of mammary parenchymal tissue was significantly positively correlated with GH levels. Additionally, it was also found to be negatively correlated with mammary adipose tissue concentration. It was therefore hypothesized that GH may be a major factor in the control of mammary development during the prepubertal period (265). These data were supported by Capuco et al. (1995) in a study starting heifers at a body weight of 175 kg on diets set to gain 0.725 or 0.95 kgld until puberty (43). In the high rate of gain group, GH was significantly lower and mammary parenchymal tissue was reduced compared to the lower rate of gain group. It was then hypothesized that GH may be a limiting factor in mammogenesis of rapidly gaining heifers and studies then began to focus on supplemental exogenous bST (43). Daily GH injections in heifers from 8 to 15.6 mo of age increased the amount of mammary parenchyma and decreased extraparenchymal tissue compared to control (264). Interestingly, however, this GH administration does not seem to increase milk yield during the first lactation (123:211). Studies focusing on the manipulation of heifer growth rates through nutrition during certain prepubertal 10 developmental periods either with or without GH administration have shown that the effects of diet and GH are not additive (45;228). In a study in which GH was administered to heifers on a high energy diet, the hormone mitigated the negative effects on milk production by the high plane of nutrition and reduced age at first calving (240). Mammary gland development is under the control of complex hormonal mechanisms involving numerous factors, and many facets of this process are not clearly understood. Though certain aspects of developmental processes are often species specific, much of the study of mammary gland endocrinology and development still finds its roots in the classical rat studies (194). Based on these early studies it has become clear that certain hormones are involved in specific phases of mammary development (107). A number of hormones are required for allometric ductal growth and end bud formation during the peripubertal period in rats. These hormones were largely determined using adrenalectomized, ovariectomized and hypophysectomized rats to identify estrogen, adrenal corticoid and somatotropin as basal factors needed for normal development (194). Studies by Kleinberg and colleagues (165) in the rat repopularized and extended the findings of Lyons. Their research showed that both lactogenic (human GH) and non-lactogenic (rat GH) GHs were more potent than prolactins (PRLs) in furthering the process of mammary development. Using GH receptor and PRL receptor mutations it was shown that binding of the GH receptor mediated mammary gland differentiation, and development was mediated by the actions of estradiol, PRL and placental lactogen by PRL receptor activation 11 (101). More recently the importance of other hormones including the growth factors epidermal growth factor, platelet-derived growth factor, transforming growth factors, fibroblast growth factors, mammary derived growth factors, and the insulin-like growth factors (IGF) have come to the forefront as possible mediators of parenchymal development (107;234;274). Key Mammary Gland Growth factors Growth Hormone and IGF-I Bovine somatotropin (GH) and insulin-like growth factor-I (IGF-l) are the most studied effectors of bovine mammary gland development and numerous experiments have focused on elucidating the mechanism by which GH mediates its effects. Growth hormone is both required for prepubertal development and is affected by feeding level (265). High feeding levels have been shown to depress circulating GH levels, yet GH is positively correlated with mammary development (265) and bST administration increases mammary growth (264). Some debate exists, however, about whether GH itself has a direct action in mammary tissue. It does not appear that somatotropin binds to its receptor in the prepubertal mammary gland (202), and GH does not stimulate mammary epithelial cells in vitro though GH-receptor mRNA is present (133). ' Insulin-like growth factor-I is likely the mediator of GH effects on mammary growth in ruminants (Figure 1). IGF-l is widely expressed across tissues and is a peptide of 70 amino acids with a molecular weight of 7.6 kDa. IGF-l is known to elicit numerous biologic actions on cell growth and death, linear growth, 12 embryonic and postnatal growth, ovarian function, and protein and carbohydrate metabolism (8;154;325). Bovine, human and porcine IGF-l are identical in amino acid sequence, and murine IGF-l differs by 5 amino acid residues (325). Three related ligands, insulin, IGF-I, and IGF-II make up the IGF family. Growth hormone treatment stimulates the liver to produce IGF-l, which then enters the circulation and can act on peripheral tissues (253). In addition to this endocrine action, IGF-I can be produced locally within a tissue (126;192;210) and thus may act in this autocrine and/or paracrine fashion in the mammary gland (203). Therefore, the presence of locally acting IGF-l within peripheral tissues could cooperate with pituitary-derived GH to stimulate a biological response. In addition, liver-derived IGF-l can serve as an endocrine feedback regulator of GH secretion, which may also act to stimulate peripheral tissue IGF-l production (178;193;283). Both IGF-I and IGF-ll receptors are present in the bovine mammary gland (40;80;125) and IGF-l has been shown to act locally in the mammary gland to induce mammary cell proliferation both in vitro at physiological concentrations found in blood and serum (236), and in vivo (275) in a study in which supraphysiological levels of IGF-I were directly infused into the mammary gland via the teat canal. Somewhat paradoxically, however, IGF-l levels are increased, not decreased like GH at high levels of feeding. These results are thus apparently conflicting with in vivo data in which mammary development is positively correlated with serum GH and negatively correlated with serum IGF-l (267). 13 IGF-I and II are found in the circulation and extracellular fluids. They can be conjugated to any one of a number of different IGF binding proteins (IGFBP) (56), or bound to binding proteins together with a ternary complex (30). Free IGF- I in the circulation has a half-life of only 10 minutes, which can be extended to 30 to 90 minutes when associated with binding proteins and to more than 12 hours when bound to the ternary complex (30). The IGF binding proteins have been proposed as mediators of the paradoxical in vivo and in vitro effects of IGF-I on mammary development. However, binding protein levels present in heifers on elevated feeding regimes do not seem sufficiently different to account for the decreased mitogenic capacity of IGF-I. It has thus been suggested that a decrease in sensitivity of the mammary gland to IGF-l is the reason for reduced mammary growth at high levels of feeding (267). In vitro results support this hypothesis. Mammary explants removed from heifers fed at a high level show a lower mitogenic response to IGF-l than tissue from heifers on a moderate feeding level (238). In addition, neither IGF-receptor mRNA nor IGF-l binding to its receptor were reduced in these tissues, and mammary gland extracts (MGE) from heifers fed at a high level of feeding were unable to stimulate cell proliferation at the level of MGE from heifers at a moderate level of feeding (320). This action suggested that an additional factor(s) present in the extracts may be interfering with IGF-l signaling and thus reducing the mitogenic response (267). The addition of IGF-I in vitro induces mammary cells to produce their own binding proteins (202). The IGF binding proteins are high affinity, soluble carrier proteins that circulate in blood and extracellular fluids and appear to control IGF blood 14 transport and association with the IGF surface receptors (57). One such binding protein, IGFBP-3 has been suggested as a possible mediator of the reduced effect of IGF-I at high levels of feeding in dairy heifers. IGFBP-3 levels were found to be higher in MGE from heifers on the high feeding level. Mammary explant and organoid culture studies also support this assertion, because addition of IGFBP-3 inhibited the IGF-l mitogenic response in vitro (238:320). While these data support the hypothesis that IGFBP-3 may be the factor inhibiting mammary growth at high levels of intake, no mechanism to date has been implicated in the regulation of IGFBP-3 levels. Other factors have recently been suggested as potential mediators, including locally produced growth factors and cytokines, including TGF-beta (238) and leptin (276:278). Regulation of mammary ductal development is largely regulated by IGF-l (163). Using the rodent model, Ruan and colleagues (1992) showed that localized IGF-l administration in the mammary gland of estrogen-treated hypophysectomized and ovariectomized rats was sufficient to stimulate ductal development (250). A critical requirement for estrogen (estradiol, E2) has been demonstrated in the analysis of GH and IGF-l action in the rodent and bovine mammary gland within the IGF-l axis (9;248). In addition, Kleinberg et al., (2000) demonstrated in rodents that when various elements of the IGF-l axis were deleted, IGF-I and the IGF-l receptor (IGF-R) were essential to normal mammary development (164). In addition, neither estradiol nor GH alone or in combination had any effect on prepubertal mammary development in lGF-l-knockout mice (164). Normal development of the ductal system in this lGF-I-knockout was 15 restored following administration of exogenous estrogen and IGF-l in this model (249). Much evidence indicates that a local effect of GH exists on IGF-l action on development in the mammary gland of peripubertal rodents (164). Growth hormone first binds to GH receptors in the stromal tissue of the gland which induces production of IGF-I, which then promotes the development of terminal end buds (9). ln ruminants, evidence for a similar action exists (108;144;240;271). In sheep, IGF-I expression is predominantly found in the stromal tissue of the mammary parenchyma which is likely the mediator of allometric growth of parenchyma tissue (144). The fetal ovine mammary gland has also been examined for IGF-l production, and in situ hybridization indicated that IGF-l production is relegated to stromal cells (108). Similar evidence for action of IGF-I promoting prepubertal mammary gland development in ruminants has been shown using in vitro bovine mammary epithelial cell culture (133271). In vivo, Radcliff and colleagues (2000) showed that administration of exogenous GH increases both serum IGF-l levels and prepubertal mammary gland development (240). In addition, Collier et al. (1993) infused IGF-l for 10 days via the teat canal of pregnant beef heifers and noted a trend indicating a 17% increase in dry fat-free parenchyma mass without a corresponding increase in parenchyma DNA (59). IGF-l is secreted locally in the bovine mammary gland by stromal cells, but not epithelial cells (125). In addition, the IGF-l receptor is found abundantly on mammary epithelial cells (122;236). The presence of both a local IGF-l producing 16 source and target within the mammary parenchyma reinforces the role of locally- produced IGF-l as a mediator of ductal development. Some debate exists as to whether this locally produced IGF-l is responsible for its physiological action in the mammary gland, and Hodgkinson and colleagues (1991), have suggested that circulating IGF-l is the major source of IGF-I delivered to the mammary epithelium (140). Signaling in the IGF-l Family Normal development of the mammary gland is contingent upon the interactions of a wide range of signaling pathways activated in response to endocrine factors signaling in a tissue-specific manner. Prolactin, estrogen, GH and progesterone have all been known to be crucial hormones guiding mammary gland development for nearly half a century (215). The mechanisms by which a number of these factors elicit their actions, however, have only recently been discovered. Up to five different classes of membrane bound receptors may interact with the ligands of the IGF family. Two splice variants of the insulin receptor exist (IR- A and lR-B), the type I IGF receptor (IGF-IR), and two potential hybrid receptors based on the dimerization of one subunit of the IGF-IR with either of the two IR splice variants (14:111). The biological actions of insulin, IGF-I and IGF-ll are believed to be mediated through lR-A, IR-B, and IGF-IR, respectively (124;206;262). The IGF-IR is a transmembrane tyrosine kinase receptor with structural similarity to the IR that can bind IGF-I, and IGF-ll and insulin with 99% 17 and 99.8% lower affinities, respectively. The Type II IGF receptor is also known as the mannose-6-phosphate receptor. Many investigators believe that the receptor's role in IGF-ll function is to moderate IGF-ll action by mediating its endocytosis and degradation (172). A ligand-binding domain is present on each of the two extracellular alpha subunits of the IGF-IR. Each alpha subunit is linked to a beta subunit containing both the transmembrane domain and intracellular tyrosine kinase catalytic domain. Upon ligand binding, autophosphorylation on tyrosine residues of the catalytic domain propagate cytoplasmic signals (300). Stimulation of postnatal body growth is the predominant physiological action of IGF-I, this was shown by targeted mutations within the IGF-I and IGF-II genes, which disrupted body weight gains in growing mice (79;124;186). IGF-l’s action on cell proliferation is its most frequently studied action. IGF-I has been shown to be necessary for progression of cells through both the G1 and GZIM phases of the cell cycle (2;291). In addition, IGF-I has been shown to protect cells from death by its actions on apoptotic and necrotic pathways (63:309). Mammary cells are among the various cell types that have been shown to be protected from apoptosis by IGF-l. lGF-l-dependent signaling pathways dictating cell survival are well studied and defined and are highly cell-type specific (124). The effects of IGF-I on cell signaling in the mammary gland have recently been extensively outlined in a review by Hadsell and colleagues (124). Four distinct, but interacting, signaling pathways are now known to be engaged by the action of the IGF-IR in the mammary gland (Figure 2). Signaling 18 through the tyrosine phosphorylation of the scaffolding protein insulin receptor substrate 1 (IRS1) is known to enhance cell survival via IGF-I binding (223). The IRS-1 protein mediates IGF-l -dependent cell survival through activation of phosphatidylinositol 3’-kinase (PI3K) and the serine threonine kinase Akt (75). A second lGF-I-stimulated pathway acts to induce cell proliferation and involves activation of ras, raf, and mitogen activated protein kinase-1 (MAPK1) and MAPK3 (230;247). In this pathway, cell cycle progression is stimulated through the regulation of the cell cycle regulator cyclin D1 and the cyclin-dependent kinase inhibitor 13 (10). Another proliferative pathway has been shown to involve phosphorylation of beta-catenin, which is able to stimulate expression of cyclin D1 and the transcription factor c-myc (60;148;235). A fourth pathway promotes cell survival and is mediated through the interaction of IGF-IR with the 14-3-3 family of proteins (70:114). In this pathway, the serine threonine kinase raf1 translocates to the mitochondria, resulting in the phosphorylation and inactivation of the BCL2-antagonist of cell death, BAD (232). Although these pathway interactions are presented independently for simplicity, much evidence supports cross-talk among the different pathways that add additional levels of regulation, redundancy and amplification of signaling. Numerous other growth factor receptors are also known to activate these pathways, so a coordination of signaling between varieties of factors is also present depending on cell type and physiological context (124). In addition, much of the work done to elucidate these signaling pathways has used in vitro cell lines or transgenic animals. Thus, the complexity of signaling is likely much higher 19 when numerous interacting cell types are present in a wild-type animal. Much remains to be elucidated regarding how these lGF-lR-dependent signaling molecules interact in an intact animal to shape normal mammary gland development. Biological actions of Ieptin Leptin is a hormone secreted by fat cells (112;113:336) that mediates central hormonal and metabolic responses to the nutritional state of mammalian species. The prototypical Ieptin protein is a 16 kDa product of the murine obesity (ob) gene and orthologs have been identified in many mammals including sheep, pigs and cattle (91 :151 :180:243). This protein is highly conserved across species with the cattle amino acid sequence being 99% similar to that of sheep and 89% similar to that of mouse and human sequences (120:149:151:336). Circulating blood Ieptin concentrations vary with the amount of adipose tissue in the body, suggesting a role for Ieptin in signaling adiposity (41 :1 13:323). Blood leptin concentrations have been shown to increase with body fatness in cattle, sheep, rodents and humans (81 :93;1 12:1 13). In addition, acute alterations of energy intake have been shown to affect plasma leptin concentrations (149:177). Blood concentrations of leptin were shown to be high in subjects on a high energy diet independent of body fatness level, and fasting rapidly decreased blood leptin concentrations (28:51 :81 :197:207). Although several other tissue types, including the stomach, skeletal muscle, mammary gland and placenta 20 have been shown to secrete leptin in the body, adipocytes are believed to be the major source of leptin production (5:164). In humans and rodents, Ieptin secretion occurs in a pulsatile manner and is subject to diurnal changes (281). Pulsatile regulation of leptin secretion has also been observed in sheep, however, in both cows and sheep diurnal changes do not seem to be present (27:30:301). A number of hormones and factors have been shown to promote Ieptin secretion including insulin, glucocorticoids and cytokines. Testosterone, peroxisome proliferators-activator receptor (PPAR) gamma agonists, IF N- gamma, and sympathetic nervous activity mediated by catecholamines have been shown to inhibit Ieptin secretion (5:6:130:245;279). The pro-inflammatory cytokines tumor necrosis factor-alpha (T NF-alpha), interleukin-1-beta (IL-1 beta), and transforming growth factor-beta (TGF-beta) were able to stimulate leptin production from precursor preadipocytes in the absence of mature adipocytes (279). Growth hormone appears to have differing effects on Ieptin production in human and rat adipocytes, increasing synthesis in the former while decreasing it in the latter (6). Leptin Receptor The biological actions of leptin are elicited by the binding of leptin to the Ieptin receptor (OB-Rb). OB-Rb is the long isoform of the Ieptin receptor which contains a transmembrane intracellular tail that is required for endocrine function and signaling (32:34:62:100;116:175:176:208:296:297:334). Alternative splicing 21 of the primary LR transcript can result in the production of multiple isoforrns (OB- Ra to -e) that contain a common extracellular domain but lack transmembrane signaling capacity (100:175;176). OB-Rb is a member of the interleukin-6 family of class I cytokine receptors. Though they contain no intrinsic enzymatic activity, cytokine receptors are able to transmit signals via noncovalently associated Janus kinase (JAK) family tyrosine kinases (Figure 3). Ligand binding to OB-Rb elicits the activation and tyrosine phosphorylation of Jak2 and the subsequent tyrosine phosphorylation of two tyrosine residues, Tyr985 and Tyr1138 on the OB-Rb (16:121:294). This then initiates the activation of intracellular signals by recruiting specific signaling proteins with phosphotyrosine binding domains leading to the subsequent activation of downstream signaling proteins along a number of pathways. One specific protein recruited by the phosphorylation of Tyr1138 of the OB-Rb is the SH2 domain-containing protein signal transducer and activator of transcription-3 (STAT3). Upon its phosphorylation and activation dimerized STAT3 then translocates to the nucleus and mediates gene transcription, including that of the suppressor of cytokine signaling, SOCS3, which is able to attenuate OB-Rb signaling in a feedback manner by interfering with further Jak2 activation (16:19:25:26:44:179:289;322). Crosstalk between hormones is a critical area of research geared toward the understanding of the endocrine feedback loop regulating body energy stores. “Adiposity signals”, which are putative molecular substrates present in the circulation that can reveal information about body condition to the brain in order to help regulate energy intake, could possibly be present in the body in the form 22 athe' stage mam." inhgbi‘. sugge and lei spec‘ hgh the m more he ian( was Pro The Vie 3L In of the hormones insulin and leptin (5). Both of these hormones are known to circulate in the body in proportion to body fat mass in the physiologically normal mammalian organism, and can act upon the central nervous system (CNS) to inhibit food intake through their ability to signal within the brain (5). The suggested physiological and intracellular signaling interaction between insulin and leptin at the level of the CNS has subsequently been extended to other specific tissues, including the mammary gland. It has been well established that high energy diets increase fat deposition within peripheral tissue stores, including the mammary gland. Local concentrations of IGF-I and leptin are known to be increased locally in the mammary gland with increased energy intake. However, although IGF-l has been shown to have a stimulatory effect on mammary epithelial cell proliferation, mammary gland development is inhibited by high energy diets. Thus, it is here hypothesized that a hereto unknown factor is increased in concentration with the feeding of high energy diets which interferes _ with the proliferative action of IGF-I on mammogenesis. Our laboratory has proposed that Ieptin, a hormone produced by adipocytes, could potentially mediate the inhibition of lGF-I-stimulated mammary parenchymal development via crosstalk between the IGF-I and Ieptin receptors. Evidence is presented supporting 80083 as a potential mediator of the negative feedback on IGF-IR activation promoted by Ieptin intracellular signaling. Suppressor of Cytokine Signaling (SOCS) Proteins 23 Will bm Dhc act C’l‘l Ph W hat DOI Recently a new family of cytokine-inducible proteins called the “suppressors of cytokine signaling (SOCS)” were discovered which function as negative regulators of signaling pathways involved in the cellular actions of numerous cytokines, growth factors, and hormones (310). The SOCS proteins share a homologous structure with an NH2-terminal region of variable length, a central SH2-domain, and a COOH-terrninal SOCS-box (310). While a number of SOCS genes are constitutively expressed in certain tissues, others have been shown to be highly regulated and act as temporal and spatial modulators of cytokine and growth hormone signaling in tissues such as the mammary gland. The JAK/STAT pathway is one specific signaling network that has been identified as being negatively regulated by the SOCS proteins. Both IGF-I and Ieptin are known to signal via the JAK/STAT pathway, which may indicate that the SOCS proteins may mediate signaling interactions and cross-talk between these two hormones which are both known to function locally within the mammary gland. Briefly, the binding of a cytokine or growth factor to its receptor triggers tyrosine phosphorylation and activation of receptor-associated JAKs. These kinases then act to phosphorylate multiple target proteins, including tyrosine residues in the cytoplasmic domains of the receptor, and receptor-associated STAT monomers. Phosphorylated STAT monomers form dimers that then translocate to the nucleus and bind to specific DNA target sequences, modulating gene transcription (31 5). Cytokines and growth factors have a plethora of biological effects and a great potency in most tissues, which necessitates a high degree of control to avoid 24 potentially damaging consequences. Soluble receptors, binding proteins, and counter-regulatory factors such as TGF-beta exist in the extracellular milieu to provide one level of control. A greater amount of control, however, may arise from the actions of intracellular molecules such as phosphatases to dephosphorylate and inactivate specific proteins (198). Similarly, the SOCS proteins, which are known to be induced by cytokine activation of the JAK/STAT signaling pathway, can act to bind specific proteins and downregulate pathway activation (315). Cytokine-inducible SH2 domain protein (CISH) was the first member of the SOCS family to be identified. CISH was found to be an immediate early-response gene whose expression was induced by several cytokines (331). The protein “suppressor of cytokine signaling 1” (80081) was the second member to be identified, discovered independently in three separate laboratories (96:213;290). Currently eight members of the SOCS family have been identified, SOCS1- SOCS7 and CISH. Based on database sequence mining at least 20 SOCS- related proteins exist in humans and mice, each containing a 40-residue C- terminal motif, the “SOCS box”, but differ in their individual types of protein- protein interaction motifs (138:290;315). SOCS genes vary in expression level across tissues, particularly SOCS1 and SOCS3. This expression is often not constitutive and can be rapidly regulated by various cytokines as well as GH, prolactin and leptin (315). The promoters of many SOCS genes contain STAT-responsive DNA elements, consistent with their expression dependence on the JAK/STAT signaling pathway (213:290). 25 However, JAK/STAT-independent regulation of SOCS expression has also been described for $0083 in neutrophils stimulated by lL-10 (49). Inhibition of the JAK/STAT signaling pathway by the SOCS proteins can arise from several distinct molecular interactions. This inhibition can work by SOCS proteins binding directly to the SH2 domain of the tyrosine phosphorylated cytokine receptors or by directly binding to the JAK tyrosine kinases which also interact with the receptors. For example, SOCS1 can act by directly binding to all four JAK molecules, and inhibiting their catalytic ability. The association between SOCS1 and JAK2 occurs at the SOCS1 SH2 domain, with the N-tenninal amino acid region providing kinase inhibitory activity (96:213;218:329). SOCS3 also binds to JAKs, but with lower affinity than SOCSl and a weaker inhibition of catalytic activity. SOCS3 may, however, be more effective at interacting with activated cytokine receptors. The SOCS3 protein has been shown to associate with the activated GH receptor to inhibit STAT5 signaling. Here the mode of SOCS action is not by prevention of the docking of STAT5 to its JAK receptor but still results in the inhibition of JAK2 activation (127). Various members of the SOCS family have been shown to have actions in the lactating rodent mammary gland (135). Expression levels of SOCS1-3 were examined in the postpartum mammary glands in a study by Helman and colleagues (135). While SOCS1 expression did not change over the course of 3 days, SOCSZ and SOCS3 expression gradually increased during the three days postpartum. It is likely that protein expression levels in the mammary gland were induced by an increase in circulating prolactin in response to suckling and that 26 SOCSS acts as a physiological inhibitor of prolactin (PRL) receptor signaling. PRL signaling is mediated through STAT5 and MAP kinase pathways and promotes lactogenesis (119:136). To further this work, Tam et al. (2001) showed that PRL is able to induce expression of SOCS1-3 and CISH in the mammary gland (295). In addition, it was shown that the mammary gland is rendered unresponsive to PRL following pup withdrawal, which is likely because of increased levels of SOCS3 (295). SOCS1 signaling has been implicated in the negative regulation of PRL action in the developing mammary gland of the mouse (181). While PRL has been intensely studied in relation to lactation and mammary growth, its role in mammary development is less clear (8). Species variation in the role of PRL in the process of mammogenesis has been demonstrated. In the virgin mouse, PRL induces the regression of fully elongated terminal end buds and promotes the appearance of ductal side branches (74). PRL-deficient mice have markedly reduced ductal branching with arrest of mammogenesis at puberty (142). STAT5a activation has been implicated in the PRL signaling cascade linking the transcriptional activator to lobulo-alveolar development during pregnancy and lactation (147) and STAT5a-null mice show a mammary phenotype very similar to PRLR+/- females, with impaired differentiation of lobqu-alveolar units and the inability to lactate (187:298). ln cattle, however, PRL seems to be less important, especially prior to lobqu-alveolar formation during pregnancy. The large natural fluctuations seen in blood PRL levels in response to environmental conditions do 27 not accompany changes in mammary gland development in growing calves, indicating that PRL concentrations have little impact on mammogenesis (8). In the pubertal, developing, rodent mammary gland, a critical role for SOCS1 in mammary cell proliferation and differentiation was demonstrated by Lindeman et al. (181). When the SOCS1 gene was knocked out in mice, it was shown that lobulo-alveolar development was accelerated during pregnancy thus establishing this protein as an inhibitor of PRL-signaling during normal development. Tonko- Geymayer et al., (2002) examined SOCS3 and CISH expression patterns in the mouse mammary gland over the course of pregnancy, lactation and involution, as well as in cultured mouse HC11 mammary epithelial cells treated with glucocorticoids (302). It was found that these SOCS protein family members inhibited PRL receptor-mediated activation of STAT5 in vitro, as expected and that glucocorticoids inhibited expression of both proteins (302). In addition, CISH and SOCS3 were differentially expressed in the mammary gland during different physiological states: CISH levels were high during the second half of pregnancy and SOCS3 expression was elevated during the first 12 days post conceptum. During involution, SOCS3 levels increased, whereas CISH levels decreased, lactation initiation coincided with high expression of both inhibitory factors (302). Epidermal growth factor was implicated as a possible mediator of SOCS expression in the mammary gland in this study in a manner independent of PRL signaling through STAT5 activation based on differential expression of SOCS3 and CISH in HC11 cells stimulated with either EGF or PRL (302). 28 The SOCS proteins have also been implicated in the control of insulin signaling in adipocytes in the mouse. CIS, SOCSZ and SOCS3 were transiently ncreased in adipose tissues taken from female mice stimulated with ovine PRL for 0-24 hours (183). It is believed that PRL-induced SOCS3 expression could clay a role in the negative regulation of insulin signaling in adipose tissue, as PRL exposure inhibited insulin-stimulated leptin production in these adipocytes. This inhibitory effect was accompanied by increased SOCSZ in adipose tissue of :he PRL-null transgenic mice, possibly indicating a mechanism for decreased eptin secretion. In addition, in the absence of insulin, PRL was shown to have no effect on Ieptin production (183). The SOCS proteins have been shown to interact directly with the activated nembers of the IGF family of receptors. Dey and colleagues were able to iemonstrate binding of the negative regulators of cytokine signaling 80082 (84) and SOCS3 (83) to the activated IGF-IR and IR using a yeast two-hybrid screen :0 detect binding of the receptor and SOCS substrate both in vitro and in intact :ells. It was also shown that in cells overexpressing IGF-IR and SOCS3, SOCS3 s phosphorylated on tyrosine residues after treatment of the cells with IGF-l, suggesting that SOCS3 is indeed a substrate for the receptor tyrosine kinase {83). A role for SOCS proteins in the inhibition of the GH tyrosine kinase receptor was also been demonstrated. In Chinese hamster ovary cells transfected with )Oth the GH receptor and SOCS cDNAs, SOCS1 and SOCSS were able to block 3H-induced transactivation of the GH-responsive serine protease inhibitor 2.1 gene promoter (1). 29 A role for the SOCS proteins in control of signaling by cytokines of the interleukin-6 (IL-6) family has been well established (78:256). Bjorbaek and colleagues (1999) showed that in addition to mediating SHP-2 binding and ERK activation during acute stimulation, a specific tyrosine residue of the Ieptin receptor mediates feedback inhibition of leptin receptor signaling by binding to activated Ieptin receptor-induced SOCS3 (23). Thus, the Ieptin receptor is able to autoregulate itself using the $0083 protein as a signaling inhibitor. In addition, using degenerate phosphopeptide libraries screened against recombinantly produced SH2 domains to determine the sequences of optimal phosphopeptide ligands, De Souza et al. (2002) predicted that SOCS3 would have a high affinity for the interleukin-6 family (78). This prediction was validated by binding SOCS3 to the phosphopeptides derived from the putative docking sites of the lL-6 family receptors gp130, leptin and erythropoietin. SOCS3 showed the highest affinity for the gp130 receptor, which has only one phosphopeptide docking site, and lower affinities for leptin and erythropoietin, which can compensate for the weaker affinity with the binding of SOCS3 to multiple individual sites on the receptor (78:99). Peripheral administration of leptin to ob/ob mice rapidly induced SOCS3 mRNA in the hypothalamus but had no effect on CIS, SOCS1 and $0082 (25). In mammalian cell lines, SOCS-3, but not CIS or SOCSZ, blocked Ieptin-induced signal transduction. Peripheral ciliary neurotrophic factor (CNTF) administration to ob/ob mice rapidly induced SOCS3 mRNA in areas of the hypothalamus that 30 are both overlapping and distinct from those in which SOCS3 was induced by Ieptin (24). 31 FIGURES Growth Hormone Releasing Hormone (GHRH) (+) v PITUITARY Growth Hormone (GH) GH (*0 LIVER OVARY (+1 I GF | (+) Estradiol V MAMMARY GLAND Leptin, I-I Insulin-like Growth Factor-l (IGF-l) Figure 1: The growth hormone/insulin-Iike growth factor-I (GHIIGF-l) axis and the bovine mammary gland. Estrogen, growth hormone (GH), and insulin-like growth factor-I (IGF-I) have stimulatory effects on mammary development. IGF-l is produced by the liver in response to GH and has a negative feedback effect on GH secretion by the pituitary gland. IGF-I also has a stimulatory effect within the mammary gland with an autocrine/paracrine mode of action. Recent evidence from Silva (2002) supports a role for Ieptin in the inhibition of IGF-I action in the mammary gland. Adapted from Hadsell et al. (124). 32 .Avms ..m So =33: Eat cognac/x .m-mc__mcm_w ociogo Lo 83933 n mwOOm .nm owns: mmm£c>m concern u 8me £03335: Co Lounges new 82695: .966 u FEM .ommci 5205 “329:8 comets. u ¥n_c=mcamoca u Xmi .F magmas» .2302 5:55 u 5m. 62582.95 new 5265 .3223 =8 :0 mfiotc mg meumfieE Sim. 9: 533 E “mic «$9553 BEE e2”. .gaifiun cote-.35.: i=2» EEO: .8332 .5308 £395 ox__.:__:n:_ "N 0.52.... «uuoz 32:3 =8 can E 20:80:35 :8 ...... 5.26 «scam 5305 v5 5.88.35 =8 338:. lo. €222.53... =00 AIE Tie.— 33 ‘ .Amm 5 28:3. was >95: Ea... 3384. .Amm 3 £30. 3 nofizuom 9m {<3 So Emoszsou 9m 533 .0 Son @3353 3.2% Qa<§ 82: seen .8658 5835.8”. .25 22$ 822...” 285326383 9: =8: Emcee 9 858.8. new :38. 2 23538 com—88:. EExo 8:: :_ Eflocou m Take new m-woow .Am TEA: 5 839.823 0585 £905 .3 cozsbocamocaou 3 new Am-woowv m-mc__mcm_m 9.325 be 288.33 5 8x83 m_ mc=mcm_m £83 .26 S 9:967. 55o. be 5.893% 3 Lo 5E3 589 noes 05 892... teams—w: cage. 5 Soon—mu m 3 Logo .mEmEmcooE 92 3:95 9:80 8:283 £33 .35: mucfi an. :9; 2 $80.53 9m 8:: con; 3552895 or: c_ 880.0% m_ m._.<.=w 99:05 2 359m «.533 .Loamzuom 95956 .Ehwis E28 a mm 93 E809 9150 on... .36: o>.S S coszocamoca c.2693 m._.zom 8589 £63 8:352. 2:253 .8382 cage.— 6 2:9“. £039.85... 230 5.22.... 1.4... 5.88.35 =8 E 88¢. _ 8888 5.8.. can... 34 CHAPTER 2 Design and application of a bovine metabolism long oligonucleotide microarray Abstract The Human Genome Project provided a catalyst for the development of extensive publicly available databases housing genome sequence information for many species of research interest, including production animals. When coupled with advancements in oligonucleotide synthesis, these bioinforrnatics resources provide a powerful tool for in-house generation of custom oligonucleotide microarrays designed to test specific hypotheses. Nutritional genomics is the integration of genomics into the nutritional sciences that has been used to elucidate the functions and interactions of genes and nutrients which can then be used to design diet-based therapeutic interventions to avoid disease manifestation. We have designed a bovine metabolism-focused microarray containing known genes using publicly available genomic internet database resources provided by NCBI, TIGR, KEGG, Swiss-Prot and BioCarta. 70mer oligonucleotides for each gene were designed and spotted on glass slides at sufficient replication to facilitate accurate detection of changes in gene expression. A microarray containing a robust collection of genes known to be involved in metabolism and regulation of metabolism will provide a better understanding of the etiology of common metabolic diseases in cattle. 35 Keywords: metabolism, gene expression, microarray, disease Introduction Nutrition and functional genomics Nutritional genomics is a new tool to study the interactions between dietary nutrients, environmental factors and cellular and genetic processes (156:209;292). Advances in a number of scientific fields have allowed for parallel assessment of expression of thousands of messenger RNAs (mRNA) to be performed in a single experiment. Oligonucleotide production has improved for both yield and quality of gene probes. Robotics technology now allows the arraying of hundreds or thousands of PCR amplified complementary DNA (cDNA) clones or genes at high density on derivatized glass slides or chip substrates. Genetics research has provided identification of a large fraction of the total gene products present in mammalian systems. Mathematical aspects of computational biology have made analysis of data from large scale genomics research increasingly possible. These advances will accelerate discovery of genes responsible for various nutritionally-related diseases and syndromes (18:1 18). A complete understanding of nutrition requires research in many facets of biology. Chemical analysis techniques must be performed to isolate and characterize the structures of essential nutrients. Biochemical and physiological experiments must be performed to determine nutrient metabolic and signaling pathways responsible for homeostasis in the organism of interest. In addition, 36 research on inborn errors of metabolism or other genetic studies is used to identify gene-nutrient interactions. Examples of this include the human recessive defect galactosemia, in which defective lactose-1-phosphate uridyltransferase (GALT) causes galactose to accumulate in the blood, causing health problems including mental retardation, and phenylketonuria (PKU), a trait that results in accumulation of phenylalanine in the blood that can cause neurological damage (156). One limitation of this approach is that one is unable to sufficiently predict and quantify interactions among dietary components and polymorphic alleles. Microarray technology provides a high-throughput functional genomics approach to begin to provide a greater understanding of the complex and reciprocal interactions within the genome at the molecular level (292). Both genetic and dietary variables can be manipulated to design an effective nutritional functional genomics study to examine gene interactions affecting normal and diseased physiological states. Developing a microarray Advances in robotics and biotechnology in the past decade have made possible the fabrication of microarrays for expression screening of tens-of-thousands of genes in a single experiment (128:184:257:259:287:318). Microarrays are hundreds to thousands of DNA fragments (probes) precisely positioned at a high density on a solid support where they can act as molecular detectors to measure relative mRNA abundance between biological samples (Figure 1). Microarrays vary in the type of solid support used (glass slides or filters), the type of probe 37 used (cDNA or oligonucleotides) and the manner in which the probe is deposited on the array (synthesized in situ or spotted). The first-generation arrays were built by spotting cDNAs amplified from clone libraries onto nylon membranes (129:258). A cDNA microarray is generated by amplifying gene specific sequences using polymerase chain reaction (PCR). This approach requires development of a clone library and thus is limited by the presence or absence of genes within the clone library. A olone library is generated by extracting mRNA from a tissue or organism of interest, converting this mRNA into cDNA, fragmenting the cDNA using restriction enzymes, and then “cloning” the cDNA fragments into plasmid vectors. Once inserted into the plasmid vector the sequence can be amplified by growing it in competent bacteria which are able to take up the plasmid. These competent bacteria also contain DNA coding for antibiotic resistance so that clone containing colonies can be selected in bacterial culture. Antibiotic resistant colonies now each contain different plasmids (and thus gene sequence fragments) and can be collected and sequenced to create a gene library. One limitation of cDNA microarrays is that cDNA clones and PCR amplicons must be tracked, which can lead to misidentification of up to 10-30% of clones (314:318). Nylon membrane arrays are limited in their spot density but development of spotting procedures for modified glass slides increased the maximal spot density and allowed addition of more genes probes to an array. Oligonucleotide arrays are a more recent advance that allows designing of probes to ensure unique probe sequences, optimal melting temperature, and 38 similar hybridization efficiency between probes (Figure 1). Control of these factors standardizes hybridization conditions from gene to gene to improve expression data. This approach is not limited by the genes present in clone libraries and requires only a gene sequence from a database. Although the probes used in oligonucleotide arrays are shorter in length than cDNA probes, they experience less problems with secondary structural formation (probes annealing to themselves), low hybridization efficiency, and non-specific hybridization. Oligonucleotide microarrays can be composed of short oligonucleotides (25 bases) synthesized directly onto a solid support using photolithographic technology (Affymetrix arrays) (Figure 2) (184:189) or constructed from long oligonucleotides (55-130 bases) spotted onto glass slides (39;155;338). Large differences in expression data among genes have been reported across microarray platforms (cDNA, oligonucleotide, and Affymetrix arrays), but this variation may be due to probe sequence and annotation (31:169:332). However, data from one study directly comparing gene expression data from a spotted long (80mer) oligonucleotide array to those on an Affymetrix 25mer chip showed that for the majority of genes (92%), no significant effect of platform on gene expression was detected (171). Here, treatment effects were stronger than platform effects. Only 8% of genes showed divergent results between the two platforms, and this divergence may represent incorrect probe identification on the array, incorrect gene identification, or alternative splicing (171). Medium length (70mer) oligonucleotide probes spotted onto glass slides 39 at high density are an excellent and cost-effective microarray platform increasingly used by researchers in many disciplines (314). Conducting a microarray experiment consists of four steps: 1) generation of the array, 2) sample mRNA isolation, conversion to cDNA, amplification and labeling with fluorescence dye, 3) hybridization of labeled sample to the array and measurement of resulting hybridization, and 4) analysis and interpretation of the data (54:89:94:134) (Figures 1, 2). The first step in developing a microarray is to generate a set of sequence- validated probes, each with a unique sequence that has little cross-hybridization to other probes. The gene set should contain a comprehensive representation of the expressed fraction of the genome or a collection of genes specifically tailored to an experimental system of interest (i.e. a set specific to a research field such as metabolism, or type of tissue such as the mammary gland). Following collection of the genes of interest, microarrays can be fabricated in a number of ways (89:94:134;184). In general there are two main types of microarrays: either the oligonucleotides are synthesized on the area in situ (Affymetrix arrays) or they are printed or deposited onto their solid supports. Microarrays rely on a labeled representation of mRNA generated using reverse transcriptase to make a single-stranded cDNA. Following cDNA synthesis different colored fluorescent labels (Cy3 = red, and Cy5 = green) are attached to the aminoallyl modified cDNA targets in the experiment (control or treated) and hybridized to the array. Labeled Cy3 and Cy5 fluoresce at different wavelengths and are detected using a confocal laser scanner. The signal presented by each 40 fluor indicates relative mRNA abundance. To analyze expression, data spots must be correctly identified and aligned to a grid specifying spot location. Background and signal intensities are calculated based on the expected position, size and shape of each spot. To standardize hybridization across an array, normalization procedures are applied to analysis. Normalization procedures may make use of genes that are known to be unaffected by treatments and remain constant in their expression (housekeeping genes), or the average spot intensity across the entire slide, based on the assumption that the majority of genes are not modified between treatments (239). A local weighted linear regression procedure (lowess) is considered the most robust normalization and is often used for data normalization (58). Fold change cutoffs (i.e. greater than 1.5 fold) are a rudimentary criteria for detecting differentially expressed genes still used by some researchers (58). More recently, linear and mixed models allow analysis of variance and determination of statistically significant genes (i.e. P<0.05) (15:190). In addition, many data visualization techniques have been developed for microarray data analysis including clustering and principle component analysis. These methods provide deductive insight especially in data across multiple treatment levels for time sequences (95). Known information on gene activity and relationship to other genes should be used to analyze microarray data in a biological context (239). Recent approaches include determination of biochemical and cell signaling pathways that are coordinately regulated between experimental conditions. Bayesian approaches also allow integration of previously described biological data. Statistical and bioinformatics applications to 41 microarray data analysis is a quickly progressing field and the interested reader is directed to the many reviews of this topic (15:53:190). A number of different designs may be employed in microarray experiments, including direct comparison, dye-flip, referenw, and loop designs (52:158:326) (Figure 3). Often the type of comparison being made in the study as well as availability of biological replicates and funding will determine the type of design used. Although biological replicates (an increased number of individual animals included for cDNA comparisons) are preferred to technical replicates (repeated sampling of a single or pooled sample over a number of different microarrays), the cost of each sample will often determine the number of samples analyzed. Dye-flip experiments, in which two samples are compared on two different microarrays with the labeling dye reversed for each microarray, are preferred to account for preferential binding of either the Cy3 or Cy5 toward given genes (50). Dye-flips, however, are technical replicates, not biological replicates. Nutritional genomics and animal production diseases Discovery of specific genes responsible for disease phenotype will enable improved prediction of disease risk and development of novel methods to prevent and treat diseases specific to an animal’s genotype (293). Diet has been identified as a causative agent for a number of diseases in dairy cattle including ketosis, hepatic lipidosis, ruminal acidosis, laminitis, and periparturient paresis. Each of these diseases is incompletely understood, and prevention solely through careful formulation of diets specific to the metabolic status of the animal 42 is difficult. The precise molecular mechanisms causing diseased phenotypes are not clear. It is highly likely that some diet-regulated genes, including their normal, common splice-variants (genes with DNA sequence rearrangements), play a part in dictating the onset, incidence, progression and severity of each disease. One major factor determining disease susceptibility lies in individual genetic makeup, as many animals in the group will be phenotypically normal with others failing in health. It is possible, however, that dietary intervention based on knowledge of nutritional requirement, nutritional status and genotype can be used to prevent or mitigate these diseases by improving group diet formulation. Dietary factors influence gene expression Multiple mechanisms exist by which dietary factors can elicit cellular responses in biological systems (Figure 4). Nutrients may directly modify gene expression as ligands for transcription factors or stimulate signal transduction pathways (109;156:191:209). Nutrients may also indirectly modify gene expression via their metabolites or through the modification of other cellular pathways (156). One example of a dietary factor acting as a transcription factor ligand (Figure 4A) is illustrated in fatty acid metabolism, which is known to be regulated by the peroxisome proliferator-activated receptor (PPAR) family of transcription factors. The fatty acids palmitic, linoleic and arachidonic acid and eicosinoids have been shown to be ligands for the nuclear receptors of the PPARs, which act as fatty acid sensors (3:12:105:166). The PPAR lipid sensors heterodimerize with the retinoic acid receptor (RXR), and RXR expression can be regulated by dietary 43 retinol (76). In a similar manner, the sterol regulatory element binding proteins (SREBP) are activated by protease cleavage and are regulated by low oxysterol concentrations and changes in insulin/glucose and polyunsaturated fatty acids (92), and the carbohydrate responsive element binding proteins (ChREBP) are proteins activated in response to high glucose and regulated by phosphorylation events (1 56:308). Dietary chemicals can be metabolically converted to ligands that alter gene expression (Figure 43). One example of this is seen in steroid biosyntheis. Concentrations of steroid hormones derived from cholesterol are regulated by 10 steps in their biosynthetic pathway. The various intermediates produced along this pathway are also available to branch into other metabolic pathways (1) (222:229). Specific combinations of alleles regulating the enzymatic steps in these assorted pathways will regulate the concentration of any given ligand branching off of steroid biosynthesis and thus potentially link nutrient intake and gene expression. Dietary chemicals and nutrients have been linked to effects on numerous established signal transduction pathways (Figure 4C). These include polyphenols (Pl3K/Akt pathway) (3), resveratrol (MAPK pathway) (13), phenethyl isothiocyanate (NF-kappa B pathway) (150), genistein (NF-kappa B and PI3K/Akt pathways) (254), and retinoids (Pl3K/Akt pathway) (219). Nutrition and genome interactions 44 Selection pressures exerted on the bovine species in the processes of gene mutation, purifying selection and random drift as well as human selection have given rise to the primary sequence of the bovine genome as well as the genetic variation that exists today within the species (90). One of the most persistent and variable environmental pressures working to shape an organism’s genome is nutrition, which can contribute greatly to genetic variation (102:292). Nutrients and environmental factors can heavily influence the development and viability of a fetus and a growing portion of literature indicate that gene imprinting is an important factor in the penetrance of deleterious genetic flaws (33:159). The ability of dietary components to affect postnatal growth and gene imprinting shows that nutrition is an important in utero selection pressure that may contribute to fixation of altered phenotypes and genetic defects in populations (33:159). Genomic responses to nutritional insults can work at the level of DNA transcription, mRNA translation, and protein and mRNA stability to maintain homeostasis (156). Maternal nutritional status can alter the epigenetic state of the fetal genome and imprint gene expression levels with lifelong consequences (317). These epigenetic alterations of the fetal genome do not alter the primary DNA sequence, but affect gene expression. This could explain the subtle differences observed in monozygotic twins (82). Methylation of DNA has been shown to be a major factor in regulating gene expression. These methylation patterns are established early in life and remain stable throughout life (139:159). Cells can thus use genomic adaptation to regulate nutrient transport and nutrient status, alter nutrient storage capacity, alter the flux of intermediates through 45 metabolic branch points, and restructure the cellular transcriptome and proteome to initiate cellular differentiation, proliferation or apoptosis. Cattle metabolic disease and nutritional genomics Similar limitations exist for assessing the regulation of individual or multiple genes by dietary factors in cattle. One must first separate cause from effect for each gene to evaluate which genes are responsible for a given phenotype. Next, one must determine if expression patterns for one strain or genotype are particular to that genotype. Based on inbred mouse studies, different breeds of cattle may indeed have unique patterns of gene expression based on diet and genotype (110:156:182). This could also help explain differences in metabolic disease susceptibility seen in different breeds of cattle. Increased susceptibility to copper toxicity and periparturient paresis by the Jersey breed, relative to the Holstein would be an example of this difference (88:195). Directed dietary intervention to treat or prevent diseases can be a great challenge. It is difficult to determine the precise way to confront a disease state when multiple genes and environmental variables interact to cause the phenotype. Identifying the genes with the greatest contribution to the disease state must first be performed and then linked to their regulation by dietary variables. To use the example of hepatic lipidosis, a disorder which can arise in overconditioned lactating dairy cows in the immediate postpartum period, much is known about the etiology of the disease, but we do not have a true understanding of the genes responsible. A bovine specific microarray will 46 facilitate the discovery of genes regulating the diseased condition in dairy cattle and forego the need to use model organism microarrays (i.e. mouse) to study these problems. RESULTS Development of the bovine metabolism (BMET) microarray First generation microarrays employing extensive cDNA libraries have allowed high numbers of both known and unidentified genes to be surveyed. Many of these arrays have no or limited spot replication, impeding observation of within- plate variance and increasing the technical variance of expression observations (29). The Human Genome Project has added significantly to databases such as Entrez Gene, the Kyoto Encyclopedia of Genes and Genomes (KEGG), Gene Ontology (G0), The Institute for Genomics Research (I'IGR), and BioCarta. These publicly available resources, paired with recent price reductions in oligonucleotide synthesis, allow researchers to feasibly design and produce microarrays with gene sets tailored to specific research areas. To build a comprehensive metabolism directed gene database, a directed search to collect metabolic enzymes and signal transduction related genes was performed using internet database resources and nucleotide sequences gathered from United States Department of Agriculture expressed sequence tag projects. Genes were collected from several functional categories: metabolic enzymes, mitogen and mitogen binding proteins, growth factors and cytokines, intracellular signaling proteins, transcriptional control regulatory proteins, 47 apoptotic regulators, and cell-cycle regulatory proteins. A gene list of human DNA sequences pertaining to metabolic genes and genes specific to pathways of interest were extracted from the Swiss-Prot Metabolic Pathway website (2), the Kyoto Encyclopedia of Genes and Genomes (KEGG) (3), and BioCarta (4) websites. The resulting genes from this search were annotated using NCBI Gene (5), Swiss-Prot, and Online Mendelian Inheritance in Man (OMIM) (6), along with their KEGG, BioCarta and Gene Ontology (Z) classifications. The human sequences from metabolism related genes were searched against bovine sequences in the GenBank database using the Basic Local Alignment Search Tool (BLAST) (11). These genes with an NCBI Reference Sequence (RefSeq), complete sequence, or 3’-end sequence were saved for oligonucleotide design. Next, we used keywords corresponding to categories of interest to search the Gene Ontology, UniGene and Swiss-Prat databases to gather human sequences of known genes and collected their bovine sequences in the same manner. This process generated a large list of genes corresponding to a UniGene cluster unique ID number with a specific accession number as an identifier (Figure 5). A cluster is a compilation of sequences of overlapping DNA that represent one gene (8). Only those bovine genes with a RefSeq link, complete sequence, 3’- end sequencing, or strong TIGR (g) cluster match were used to facilitate oligonucleotide probe design. All Perl scripts written to generate this data are available upon request under an open source license. From these databases 4,010 human RefSeq genes were identified. From the nucleotide sequences of these 4,010 genes 2,371 matching bovine sequence 48 homologues with an expectation value (e-value) of 1e10'35 or smaller were found. The expectation value is a measurement of sequence similarity, and values below 1e10“-1e10"15 tend to show strong sequence homology between two sequences being compared. A very low e-value score was chosen to ensure correct gene identification. This gene set included 513 GenBank RefSeq genes, 109 completely sequenced genes, and 901 3’-end United States Department of Agriculture sequenced genes and all genes had a TIGR cluster match. Of the 1207 KEGG metabolism genes extracted, greater than 1,100 are represented in this dataset. Using the Oligopicker program (316) 70mer oligonucleotide probes were successfully designed from 2,349 of the genes and splice variants. Each oligonucleotide was designed within specific parameters to standardize hybridization behavior. These 70mers were then synthesized at the Massachusetts General Hospital Oligonucleotide Synthesis Facility (10) and spotted four times per gene on microarray slides at the Massachusetts General Hospital Microarray Core Facility (11) on Poly-L-lysine coated glass slides. Replicate oligonucleotide spots were printed on each array in a global fashion (spot replicates are distributed across the slide to account for spatial variation in hybridization conditions) to ensure adequate spot replication for downstream data analysis. The BMET array contains 9,582 spots including housekeeping (including GAPDH and Beta-actin) and negative (sequences not found in the bovine genome) control genes (10 Arabidopsis thaliana metabolic genes) (314). 49 As part of our effort to build a strong bovine metabolic genome resource, we are currently developing a web-based analytical tool to help with downstream analysis of the BMET array. Known human metabolic and cell regulatory pathways represented within KEGG and BioCarta will be adapted to this bovine- specific gene set using the GenMAPP tool (g) and a GeneLink resource (Figure 5). This is housed at our websites h_ttg/Iwww.nutri-genomics.org (13) and http://nbfgcmsuedu (g). Spot replication improves within-array quality control and increases the statistical power of accurately detecting small changes in expression at a lower cost than slide replication (29). Reduction in technical error to increase statistical power is especially important for metabolic research, in which changes in gene expression are often subtle and the cost per experimental unit is very high. In addition, our focus on only those genes that are relevant to metabolism increases the efficiency of downstream bioinformatics and data analysis for integration of metabolic gene networks. Because all genes included in this design are annotated with corresponding human homologues, the design can be applied to other species to promote our understanding of comparative metabolism. Our design of a focused oligonucleotide microarray with multiple spots per gene will facilitate research in the study of cattle metabolic disease genomics of cattle and can be easily applied to other species and disciplines. Conclusions 50 The merging of nutritional and genomics approaches to the understanding of interactions between diet and genotype is revealing a system of great complexity. Nutritional genomics studies have given researchers the opportunity to target specific dietary interventions to avoid and treat metabolic diseases in humans and cattle. Mechanisms underlying the effects of dietary nutrients on genome stability, imprinting, expression and signaling may become more clear in the near future through the employment of targeted microarrays and experimental design to answer questions remaining on specific metabolic diseases encountered in modern animal production. The design of a microarray containing a large fraction of the genome with known links to metabolism and signal transduction will increase the quality and utility of gene expression data that can be acquired from microarray technology and lead to advancements in the field of nutritionally related production animal diseases. Acknowledgments We would like to thank our collaborator Randy Baldwin for his help in funding the production of the BMET array, and the staff of the Michigan State University Center for Animal Functional Genomics, especially Peter Saama and Steven Suchyta, for their help in assembling the gene database. 51 Website References 1) ht_tp:/Iwww.genome.jp/dbget-bin/show pgthwgflhsg00100+3156z Kyoto Encyclopedia of Genes and Genomes, Steroid Biosynthesis Pathway. 2) http:/Iisexpgsvorgl: The ExPASy (Expert Protein Analysis System) proteomics server of the Swiss Institute of Bioinformatics. 3) mp:I/www.genome.jplkegg/pathway.html: Kyoto Encyclopedia of Genes and Genomes pathways website 4) Mggpnci.nih.gov/PgthwgvslBingtg; Pathways obtained from BioCarta via the Cancer Genome Anatomy Project. 5) http://www.ncbi.nih.gov/entrez/querv.fcgi?db=gene: Entrez Gene. 6) _t1t_tp:/Iwww.ncg.nlm.nih.gov/entrez/guem.fcgi?db=OMIM: The Online Mendelian Inheritance in Man database, a catalog of human genes and genetic disorders. 7) Mlmggneontologvorgl: The Gene Ontology consortium website. 8) http://www.ncbi.nlm.nih.gov/entrez/guery.fcgi?db=un1<fiiez NCBl’s UniGene link. This system automatically partitions GenBank sequences into a non- redundant set of gene-oriented clusters. 9) Mtg/Winnorg/tigr-scripts/tgifr index.<_:gi?species=cattle: The Institute for Genome Research’s bovine gene index. 10) httpszlldnacore.mMarvarc_l.edujsvnthesisfingexshtml; Massachusetts General Hospital Oligonucleotide Synthesis Facility. 52 11) https:/Idnacore.mgh.harvard.ed_u/micrchrav/index.shtml; Massachusetts General Hospital Microarray Core Facility. 12) h1tp:/Mww.genmapp.org: GenMAPP is a free computer application designed to visualize gene expression data on maps representing biological pathways and groupings of genes. 13) http://www.nutri;genomics.<£gz: Home of the BMET microarray database. 14) http://nbfgmsuedul: Michigan State University Center for Animal Functional Genomics database. 53 FIGURES Microarray Overview: Spotted DNA Arrays 1. Total RNA 2. Aminoallyl 3. Cy3/Cy5 4. Co-hybridization 5. Wash isolation modified cDNA dye coupling to slides and scan Control ...B—a—fi B—«é 4 Figure 1. Spotted DNA Arrays. Plasmid clones are propagated in bacteria, and the cloned inserts are amplified by polymerase chain reaction (PCR) and then purified. These purified PCR products are robotically printed onto solid supports. Alternatively, oligonucleotides can be used instead of PCR products. Synthesized oligonucleotides are printed directly onto the glass support. Fluorescently labeled control and test samples are fluorescently labeled with Cy3 or Cy5 dyes and hybridized simultaneously onto the glass slide array. Following washes, hybridization is detected by phosphorimaging or measurement of the excitation of the two fluors at the appropriate wavelengths by a laser scanner. 54 Microarray Overview: Affymetrix GeneChipTM 1. Total RNA 2. Prepare 3. Hybridize, wash, 4. Scan chips isolation Biotin Labeled stain chips cRNA targets Control fig /Ana|yze Test @fi —"—’ E] Cell 5 Probe Pair Figure 2. Affymetrix Arrays. The GeneChip gene “probes” are synthesized in situ on the chip using photolithographic techniques. Gene probes are made up of 25mer oligonucleotides called perfect match probes (PM), and each is paired with a “mismatch" probe (MM) that contains a base pair substitution to interfere with hybridization between the probe and its “target", present in the sample. Test and control samples are labeled with biotin and hybridized to separate chips. ' The hybridized “target” probes are then stained with streptavidin phycoerythrin and scanned for light emitted at 570 nm, which indicates the sample’s signal. Match and mismatch probe intensities can be used for expression data analysis (PM-MM), or match pairs can be compared between the control and test samples alone (PM only) (270). 55 .‘""'"O Figure 3. Microarray experimental designs for two-color arrays. A) Direct comparison with dye-flip: B) Reference design for three treatments with dye-flip: C) Augmented loop design for four treatments. All Affymetrix GeneChips are scanned individually based on a one color system and all chips can be directly compared to each other. 56 Nutrients Ge ne Ligands for transcription Exp ressio n factor receptors Gene regulation or cell signaling Cellular Response Alteration of signal transduction Figure 4. Nutritional genomics model. Nutrients can act directly as ligands for transcription factor receptors (A): may be metabolized by primary or secondary metabolic pathways to alter concentrations of substrates or intermediates involved in gene regulation or cell signaling (B); or might alter signal transduction pathways and signaling (C). 57 Bovine Sequence Extraction ExPASy Biocarta (at CGAP) Metabolic NiceZyme Cell sr nalIn Pathways i'l Enzyme patharays g m I / /KEGG Homo saplens . NT SEQ NCBI Genbank ——_. Reference Sequence (RefSeq) Sequence Sequence I 3'-end I TIGR I Sequenced I Clusters .' 2,349 Bovlne sequence homologs from NCBI Genbank, USDA and TIGR with an . expectation value < 10‘" Complete [ Partial Figure 5. Metabolic gene database assembly. Using database resources such as ExPASy (g), the Kyoto Encyclopedia of Genes and Genomes (3), BioCarta (4), and NCBI GenBank (_5_) 2320 human genomic sequences were identified. Following a BLAST search against bovine sequences provided by NCBI, the United States Department of Agriculture, and The Institute for Genomic Research (9), 2349 bovine sequence homologues were detected at an expectation value better than 1e10°35. 58 Va- Absl The meta Iran; MICI that The 96"” PCI de. CHAPTER 3 Validation of a metabolism-specific long-oligonucleotide microarray for transcription profiling in cattle Abstract The bovine metabolism array (BMET) contains spots of oligonucleotides from 2,348 cDNAs and ESTs and was designed to measure gene expression in metabolism-related studies. To validate this array, comparisons were made of transcripts from adipose, mammary, and liver tissues of lactating cows. Microarray analysis revealed that all 70-mer oligonucleotides showed hybridization to probes from at least one tissue type. Statistical analysis revealed that 239 genes were differentially expressed in adipose versus liver (P<0.05). The adipose versus mammary comparison revealed 365 differentially expressed genes, and 125 genes were different in liver vs. mammary. Quantitative real-time PCR confirmed expression patterns for 7 of 8 genes tested. Our results demonstrate that the BMET array is informative for studies measuring changes in the bovine metabolic transcriptome for a number of tissue types. Keywords: metabolism, gene expression, microarray, Q-RT-PCR 59 0i 0: links undr incl To sli: frc Iis M; An C0; Introduction Advances in a number of scientific fields have allowed for parallel assessment of expression of thousands of messenger RNAs (mRNA) to be performed in a single experiment. Since their inception in 1995 (258), high-density microarrays have been widely used in genomics research in organisms with sufficient genomic resources. With the growing amount of genomic sequence information and improved annotation methods, “boutique” microarrays, which contain cDNA or oligonucleotide probes tailored to include only genes with specific ontological links, are becoming increasingly popular. In an effort to enhance our understanding of metabolic processes achieved by all the chemical compounds, including metabolic enzymes, a set of cDNA sequences encoding a large fraction of the bovine metabolome as well as interacting signal transduction pathway genes was collected (Chapter 2). From this sequence library, 70-mer oligonucleotide probes were designed, synthesized, and hybridized to derivatized glass slides to produce the bovine metabolism (BMET) microarray. To assess the utility of the BMET array, hybridization tests of glass microarray slides containing the 2,348 oligonucleotides were conducted by using target RNA from three tissues (adipose, mammary, and liver) from three lactating cows (9 tissue samples). MATERIALS AND METHODS Animals. Animals used in this study were three primiparous lactating Holstein cows described by VanderKooi et al. (311). These cows were housed in 60 individual tie stalls and exposed to 24 h/d of light and were fed an ad libitum mixed diet of alfalfa haylage, corn silage and grain containing ~1.7 Mcal of NEng and 18% crude protein with free access to water. Feed was offered at 0330 and 1630 h daily. Cows were milked three times daily at 0545 1400 and 2200 h. The Michigan State University All-University Committee on Animal Use and Care approved experimental procedures (311). Sample collection and tissue processing. Cows were killed by captive bolt and exsanguination 181 d post-calving. Liver and mammary tissue was harvested within 20 min of death (311). A 10-g piece of liver tissue, 10 g of adipose tissue taken from the omasal fat deposit, and several 100-g slices of mammary tissue taken from the left hemigland were removed, immediately placed in liquid nitrogen, and stored at -80°C until future use (22;311). RNA isolation procedure. Total RNA was extracted from frozen mammary, liver, and adipose tissues collected from tissue samples (after homogenization) with Trizol reagent (lnvitrogen Life Technologies Corp., Carlsbad, CA) essentially as recommended by the manufacturer (Invitrogen Life Technologies Corp., Carlsbad, CA). Tissue was kept cold using dry ice and ~200 mg of mammary tissue was weighed and directly added to 3 mL of Trizol reagent in a 15 ml culture tube. Tissue was homogenized using a Polytron for ~30 3. Between homogenization of individual samples, the Polytron bit was rinsed with diethyl 61 PI" lub lerr W31 ll'lll pyrocarbonate- (DEPC, Sigma) treated water, and RNAse Away (Molecular BioProducts, San Diego, CA). Samples were split into three, 1-mL samples and incubated at room temperature for 5 min. Chlorofonn (200 pL) was added to each microcentrifuge tube and samples were manually shaken, incubated for 3 min at room temperature, and centrifuged at 10,500 x g for 15 min at 4°C. The upper phase was transferred to a clean tube. lsopropanol (500 uL) was added to the precipitated RNA. The tube was vortexed, incubated at room temperature for 10 min, and centrifuged at 10,500 x g for 10 min at 4°C. The isopropanol was decanted and the remaining pellet was washed with 75% ethanol, centrifuged at 8500 x g for 5 min at 4°C, decanted, and dried. RNA was then subjected to DNase treatment (RQ1 RNase-free DNase: Promega, Madison, WI) following the manufacturer’s protocol (Promega, Madison, WI). Nuclease free water (52 uL), DNase buffer (10 uL of 10X: Promega), and DNase (1 uL of 2 UluL: Promega) were added to the pellet and then incubated at 37°C for 30 min. A phenol/chlorofon'n separation step was performed by adding RNase-free water (37 uL) and phenol/chloroform (100 uL). The tube was shaken and centrifuged for 2 min at 14,000 x g. The upper phase (~90 pL) was transferred to a fresh tube, sodium acetate (3 M, 9 uL) and ethanol (250 pL) were added to this phase, and the mixture was stored overnight at -20°C. The following day, microcentrifuge tubes were centrifuged at 14,000 x g at 4°C for 20 min. The liquid was decanted and the pellet was washed with ethanol (75%, 500 uL). The tube was centrifuged at 14,000 x g at 4°C for 10 min. The ethanol was decanted 62 and the pellet was dried in the hood for 15 min. The pellet was resuspended in 50 uL of nuclease-free water and incubated at 60°C for 10 min. The quantity of RNA isolated was determined using the NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE), and quality was checked using the using the RNA 6000 Nano LabChip kit and Agilent 2100 BioAnalyzer (Agilent Technologies, Palo Alto, CA). BMET microarray experimental design. The BMET microarray was used to compare gene expression profiles of adipose, mammary, and liver samples. The BMET array was described previously (Chapter 2) (97) and contains 2,348 spots representing genes involved with bovine metabolism and signal transduction. This array includs 10 metabolic housekeeping genes from Arabidopsis thaliana used for spike-in controls, 1320 blank spots (printing buffer only spotted) and 64 empty spots (nothing spotted). Complete annotation of the BMET gene set can be found at httpzllwww.nutri- genomics.org/data.html. The oligonucleotides were synthesized and the BMET microarrays were printed at the Massachusetts General Hospital Microarray Core Facility at Harvard University (h_tt_ps://dnfiore.mgh.harvard.edu_/microafigv/protocol-printing.shtml). Briefly, cDNA oligonucleotides (70mers) were suspended in water to 20 pM and rearrayed into Genetix polystyrene V-bottom plates (cat# X6004). Before printing, 5 pl of 1X Printing Buffer (150 mM sodium phosphate, 0.0005% sarcosyl, (pH 8.5) was added to each well. Plates were then sealed using Corning seals and 63 incubated at 37°C for 30 minutes to help resuspend DNA. Plates were shaken near maximum rotational speed on a flat-bed shaker for 1 min and then centrifuged at 2000 x g for 3 min. Each gene was then spotted four times on a Codelink glass slide (Amersham Pharrnacia Biotech: Piscataway, NJ) (10816 total spots) across the array using the Genomic Solutions Omnigrid 100 printer (Genomic Solutions, Harvard Biosciences, Inc., Holliston, MA). Humidity was kept between 30-40% during the print run using a sensor-activated humidifier and dehumidifier in conjunction. Following a print run, slides were sealed with moist paper towels in a chamber containing supersaturated NaCl overnight. Slides were then washed in Blocking Solution (50 mM 2-aminoethanol, 0.1 M Tris (pH 9.0), 0.1% N-Lauroyl sarcosine) prewarrned to 55°C for 15 min, rinsed twice in deionized water, washed again for 30 min in 55°C Washing Solution (4X SSC, 0.1% N-Iauroyl sarcosine), rinsed in deionized water 2x and centrifuged at 800 x g for 5 minutes to dry. Screening for genes differentially expressed between adipose, mammary, and liver was performed on three animals using a loop design (158) balanced for Cy3 (green) and Cy5 (red) fluorescent dye labeling (Amersham Phannacia Biotech: Piscataway, NJ) with dye-flip (Figure 1) for each of the three tissue types (18 arrays total). Preparation of labeled cDNA for BMET hybridization. For cDNA synthesis, 15 pg of sample RNA was used as a template in reverse transcription reactions (SuperScn'pt III Fluorescent Labeling Kit: lnvitrogen Life Technologies Corp., Carlsbad, CA) in which oligo(dT)15-18 was used as a primer. In the Superscript 64 I|l system, cDNA is prepared with a randomly incorporated amino-modified dUTP. Following first-strand synthesis, cDNAs for each tissue within an animal were differentially labeled using N-hydroxysuccinimide-derivatized Cy3 and Cy5 dyes (Amersham Pharrnacia, Ltd., Piscataway, NJ). Differentially labeled cDNAs were combined and concentrated to 10 pl by using Microcon 30 spin concentrators (Millipore Corp., Bedford, MA). Microarray hybridization was performed after addition of 100 pl of SlideHyb-3 hybridization buffer (Ambion lnc., Alameda, CA) to the concentrated Cy3—Cy5-labeled probe cDNAs (7). The labeled cDNAs were incubated at 70°C for 5 min just prior to 18-hour array hybridization in Corning Microarray Hybridization Chambers (Fisher Scientific) placed in water baths set to 42°C. Following hybridization, microarray slides were washed with 2X SSC saline sodium citrate (0.15 M NaCl plus 0.015 M sodium citrate), 0.1% SDS for 4 min at 42°C, 2X SSC for 4 min at room temperature, 0.2X SSC for 4 min at room temperature, 0.06X SSC for 10 sec at room temperature and once in double-distilled H20. Microarrays were dried by centrifugation in a cushioned 50-ml conical tube and immediately scanned with a GeneTAC LS IV microarray scanner (version 3.01). GeneTAC LS software version 3.3.0 (Genomic Solutions, Inc., Ann Arbor, MI) was used to process microarray images, find spots, integrate robot spotting files with the microarray image, and finally to create reports of raw spot intensities. Total intensity values for each dye channel were converted to comma-delimited data files, exported into Excel spreadsheets, and imported into SAS (255) for data normalization and analysis. 65 Quantitative real-time PCR validation of microarray gene expression changes. The expression differences of certain genes found to be significant by microarray analysis were checked using quantitative real-time reverse transcription-PCR (Q-RT-PCR) using an Applied Biosystems 7000 DNA Sequence detection system (Perkin Elmer Corp., Foster City, CA). Total RNA was extracted from each tissue from each of the three cows, quantified and quality checked as described previously. The PCR system used was the ABI PRISM 7000 Sequence Detection System (Applied Biosystems). RNA was converted into first-strand cDNA by adding 1 pg of total RNA to a 12 pl reaction containing 10 mM oligo(dT)18 primer. Following a 5 min incubation at 70°C, the reaction was chilled on ice and pH adjusted by addition of 4 pl of a buffer supplied by the reverse transcriptase manufacturer (Invitrogen Life Technologies). The final reagent concentrations were 50 mM Tris-HCI, pH 8.3, 75 mM KCI, and 3 mM MgCl2, 1 mM deoxynucleoside triphosphates, and 200 units of Superscript II RNase H reverse transcriptase (Invitrogen Life Technologies). Incubation at 37°C continued for 30 min in the presence of RNase H at 42°C to remove the original RNA templates. Heating at 70°C for 15 min subsequently inactivated RNase H. The reaction tubes were removed from the thermocycler and 0.2 pL of 0.5 M EDTA was added and mixed. Then, 25 pL of water, 5 pL 3M sodium acetate, and 125 pL of ethanol (-20°C) were added to the tube. Tubes were then allowed to precipitate overnight at -20°C. The reaction mixture containing the newly-synthesized cDNA was transferred to a new, clean 66 microcentrifuge tube. These tubes were then centrifuged at 14,000 x g and 4°C for 20 min and the supernatant was decanted. Pellets were washed with 250 pL of ethanol (75%, -20°C) and the tube was centrifuged at 14,000 x g and 4°C for 6 min. The supernatant was decanted and the pellet was allowed to dry for 15 min. The pellet was resuspended in 50 pL of water and incubated at 60°C for 5 min. The cDNA concentration was analyzed using a spectrophotometer (NanoDrop), and was then diluted to a final concentration of 10 ng/pL and stored at -80°C until the Q-RT-PCR reaction was initiated. SYBR Green PCR Master Mix (Perkin Elmer Corp.) and gene-specific primers were used to perform Q-RT-PCR reactions. Primer Express Software (Perkin Elmer Corp.) was used to design all primers, ‘which were then synthesized by a commercial facility (lnvitrogen Life Technologies). Primer sequences for genes analyzed in this report are included in Table 1. The amount of primer used was determined by performing an optimization matrix for each primer using three concentrations of primers: 50:50 nM, 3002300 nM, 900:900 nM. Dissociation curves were similar for all concentrations and the 300:300 nM matrix was chosen, thus 3 pL of primer was used for all experiments. Each gene of interest and the control gene were measured in duplicate. Within each well of a 96-well reaction plate (MicroAmp Optical, Applied Biosystems), 30 ng of sample cDNA (3 pL), 6.5 pL DEPC water, 3 pL of each primer, and 12.5 pL Sybr Green (Applied Biosystems) were 'added. To determine an appropriate reference control gene by which relative mRNA abundances for genes of interest could be measured, a number of potential 67 housekeeping genes were screened based on microarray results. These candidate control genes included B-actin (B-actin), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), hydroxymethyI-bilane synthase (HMBS), hypoxanthine phosphoribosyltransferase 1 (HPRT1), TATA box binding protein (TBP), and succinate dehydrogenase complex, subunit A (SDHA). Q-RT-PCR threshold cycle (CT) values for GAPDH for all tissues were consistent among animals and this gene was deemed a suitable reference control. The 2“” method of Q-RT-PCR analysis was performed as previously described (7:65:188). This method enabled relative gene expression changes across treatments based on quantitative differences in the PCR amplified target reaching a fixed 07 number at a set treatment versus other treatments (Table 6). Gene-specific standard errors were estimated using independent analyses of variance (ANOVA). All analyses were performed using the SAS System (255). Statistical analysis of microarray data. M vs. A scatter plots were constructed for each array to visualize potential dye intensity biases in the microarray data sets. Here, log intensity ratios M = Iog(Cy3/Cy5) = logCy3 - logCy5 were plotted against mean log intensities A = (logCy3 + logCy5)/2 for each array spot as previously described (9, 10). Array specific normalization was performed using a robust local regression technique (Cleveland 1979 #660), LOESS, of SAS (255). Monitoring M-A plots for data from each array before and after LOESS normalization assessed efficiency of LOESS normalization. Normalized data were then back-transforrned prior to further statistical analyses using the 68 following formulas: logCy3” = A + M12 and logCy5’ = A — M/2, where IogCy3’ and logCy5’ are the normalized log intensities. Here, M represents each of the normalized M values with M being the LOESS predicted value for each spot. Statistical analysis of LOESS-adjusted log intensities was performed using a mixed model approach consisting of two steps (9, 11). The first step involved an array-specific spatial variability normalization model: yadj = dye array day*array block(array) dye*block(array) + e All terms are random except dye. The second step involved gene-specific analyses for the estimated residuals obtained from the normalization approach obtained from step one to test the effect of treatment on expression profiles for individual genes. These models were: resid = heifer + treat + treat’heifer + dye + array(heifer) + pin + spot(pin*array*heifer) All terms are random except heifer, dye and pin. These analyses were computed using the MIXED procedure of SAS (255). In the gene-specific analysis, genes were considered significantly different if P<0.05. Spotted cDNA sequences representing genes whose expression profiles were significantly altered by hormonal infusion in this analysis were subjected to BLASTn analysis to reveal identities. The biological functions and ontologies of these genes were determined through an extensive Entrez-Gene (httpzllwww.ncbinlm.nih.fl/entrez/gueryfggi’MMD=Search&DB=gene) and 69 PubMed (httpzllwww.ncbi.nlm.nih.gov/entrez/guem.fcgi?) literature search. This information was used to establish clusters of affected genes within broad functional categories for subsequent expression validation and downstream analysis. RESULTS Identification of genes differentially expressed in mammary and liver of lactating cows. The BMET microarray comparison between mammary and liver tissues revealed significant differential expression (P<0.05) of 125 of 2348 genes measured (5.3%) (Supplemental Table 1). The relative mRNA abundance of 78 of these genes was found to be higher in mammary than liver tissue. Genes significantly different in mRNA abundance between the two tissues were grouped within a number of gene ontologies and pathways using the GenMAPP MAPPFinder 2 program (72:87). This analysis indicated overrepresentation of liver specific expression within a number of pathways including alcohol and carbohydrate metabolism (Table 1). Mammary specific ontology overrepresentation was clustered within intracellular signaling and transcription factor binding ontologies (Table 1). Alpha-lactalbumin was the most significantly underexpressed gene in liver relative to mammary tissue (fold-change = 0.16). Identification of genes differentially expressed in adipose and mammary of lactating cows. The BMET microarray comparison between mammary and adipose tissues revealed significant differential expression (P<0.05) of 365 of 70 2348 genes measured (15.5%) (Supplemental Table 2). The relative mRNA abundance of 170 of these genes was found to be higher in mammary than adipose tissue. Significantly altered genes could be broadly classified within a number of gene ontologies and pathways. GenMAPP MAPPFinder 2 pathway analysis (72:87) indicated significant differential expression of 11 genes in mammary tissue (PRLR, MVD, NSDHL, SEC14L2, SC5DL, SC4MOL, NMNAT2, PISD, SCD, FADS1 and MGST2), and three genes in adipose tissue (CMAS, ELOVL1, and STAR) within the lipid biosynthesis ontology (Table 2). Within the insulin signaling pathway eight genes were more highly expressed in adipose than mammary tissue (SLC2A4, GSK3A, GYS1, MAP3K3, MAPK10, MAP2K1, MAP3K11, and MAP3K12). Genes significantly different between the two tissues and involved with RNA polymerase II transcription factor activity tended to cluster within adipose tissue (ATF2, BTF3, HOXC10, IRF4, SF01, TEF, and SOX9) more than in mammary tissue (FOXA3 and MYOD1). Identification of genes differentially expressed in adipose and liver of lactating cows. The BMET microarray comparison between adipose and liver tissues revealed significant differential expression (P<0.05) of 238 of 2348 genes measured (10.1%) (Supplemental Table 3). The relative mRNA abundance of 91 of these genes was found to be higher in adipose than liver tissue. Genes significantly different between the two tissue types were clustered within a number of specific gene ontologies and pathways relevant to tissue-specific expression in the lactating cow as indicated by GenMAPP pathway analysis 71 (Table 3) (72:87). Three genes involved in carboxy-lyase activity (ODC1, PISD, and SGPL1A) were found to be more abundant in liver tissue than adipose. As with the adipose versus mammary tissue comparison, genes falling within the RNA pol ll transcription factor activity gene ontology tended to be more highly expressed in adipose tissue (ATF2, HOXC10, SREBF1, SREBF2, TEF, THRAP6, GTF2, and IRD1) rather than liver (MYOD). Three genes involved with hepatocyte growth factor (HGF) signaling (MAP2K1, STAT3, RASA1) were found to be more abundant in adipose than liver tissue. The adipocyte-specific fatty acid binding protein FABP4 was found to be very highly expressed in adipose tissue relative to liver (3.42 fold difference). Two other genes involved in fatty acid binding (FABP2, ANXA2) were also more abundant in adipose than liver fissue. Q-RT-PCR confirms IGF-l induced expression changes in 7 of 8 genes significantly altered by microarray analysis. Significant differential expression of six of seven genes between liver and mammary tissue, five of six genes between adipose and liver, and five of six genes between adipose and mammary tissue as indicated by BMET microarray analysis were confirmed by Q-RT-PCR (Table 5). For all genes tested, magnitude of fold-change was much larger by Q- RT-PCR analysis than by the BMET microarray. For example mRNA coding for suppressor of cytokine signaling 3 (SOCS3) was confirmed as more abundant in mammary than liver tissue. Microarray analysis generated a ratio of 0.85 for SOCS3, indicating a slight increase in mRNA abundance in mammary tissue 72 relative to liver. However, Q-RT-PCR analysis revealed a much larger magnitude in relative abundance with SOCS3 upregulated 54-fold in mammary relative to liver. Similarly, the cytokine inducible SH2-containing protein (CISH) was confirmed as overexpressed in mammary relative to liver tissue. CISH was 9-fold more abundant in mammary than liver by Q-RT-PCR, though microarray analysis predicted a 0.65 ratio of mammary to liver. In the adipose versus mammary comparison, interferon alpha 2 (IFNA2) gene expression was predicted to be 1.55-fold higher in adipose versus mammary tissue (P=0.058). This fold-change difference was magnified to 75-fold higher in adipose than mammary by Q-RT- PCR analysis. Only one gene, Absent, small, or homeotic-Iike 2 (ASH2L) was invalidated by Q-RT-PCR analysis. All three tissue mRNA abundance comparisons for this gene were significant in the opposite direction indicated by BMET microarray analysis. In all, 14 of 17 Q-RT-PCR tissue comparison mRNA abundance comparisons were in accordance with the predicted directionality of microarray relative abundance (Table 5). DISCUSSION Tissue specificity of expression One critical test for the value of the BMET microarray lies in its ability to detect differences in expression of genes known to be differentially expressed in different tissues. Overall, the BMET performed well at achieving this goal. For example, genes involved in coagulation, such as plasminogen and fibrinogen, were both significantly more abundant in liver than adipose and mammary, as 73 expected (337). Also expressed in a liver-specific manner included a number of genes involved with alcohol metabolism, derivation of energy from oxidation of organic compounds, carbohydrate metabolism and carboxy-lyase activity (Tables 1 and 3). All of these gene ontologies are consistent with processes known to be highly regulated within the liver. The liver specific isofonns of phosphofructokinase (PFKL) and fatty acid binding protein 1 (FABP1) were significantly overexpressed in liver compared to adipose tissue. Genes expressed in a greater abundance in adipose tissue were identified and seemed to fit well with adipose-tissue related physiological processes. Specifically, the significant differences in expression of in FABP5 in adipose vs. liver, FABP4 in both adipose vs. liver and mammary comparisons stand out. The signaling activity of adipose tissue through the mitogen activated protein kinase (MAPK) cascade seems to be quite high based on the frequency of genes highly expressed within this ontology. The growth factor pathways for insulin and hematopoietic growth factor (HGF) and a number of intracellular signaling components were also highly represented in the significantly expressed adipose gene group (Tables 2 and 3). Mammary tissue gene expression in the lactating dairy cow is highly dynamic. As might be expected, the most significantly overexpressed genes in mammary compared to adipose tissue were alpha-Iactalbumin and beta-Iactoglobulin. These genes code for two proteins produced in great abundance in mammary tissue during lactation. Similarly, 11 genes involved with lipid biosynthesis were found to be highly expressed in mammary relative to adipose tissue, including a 74 number of genes known to be regulated during milk fat production such as stearoyl coA desaturase (SOD), and fatty acid desaturase 1 (FADS1) (20;244:280). A number of genes involved with intracellular signaling were also found to be predominantly expressed in mammary tissue (Table 1). Of these, CISH and SOCS3, two genes which are known to inhibit prolactin (PRL) receptor-mediated activation of STAT5 in the mammary gland, were confirmed as overrepresented in mammary versus liver tissue (302). It is possible that these genes are expressed to a greater degree late in lactation (181 days in milk) to promote mammary gland involution by increasing apoptosis and inhibiting proliferation of mammary epithelial cells (174:295:302). In conclusion, the use of microarrays to help understand complex gene expression patterns in various tissue types in cattle is becoming increasingly helpful in understanding phenotype and genotype interactions. The BMET microarray should provide a useful platform for helping to understand bovine disease states, and nutritional and hormonal regulation of gene expression to improve the production and quality of food in the future. 75 FIGURE Figure 1: Tissue comparison microarray loop design. Each of three tissues for three animals with flip dye labeling was compared using the BMET microarray (18 arrays total). 76 XXXXX EuEEuE ><>< ><><><>< ><><>< X .83.. £529.. .599 P 89:8 9982888,... 2.5“: e emmco.mcm=_>moo:u. 59.888 .382 295.. 29m :92. m camcmmoefmo 98.08. ..(uquw. 880.39. 99.39.90 .8820 3.2.4.9.. .598 F emmSE 29829385. 5.1. 28 2%,. m 82:0me nucaanoo 0.590 .0 cow—«2.8 >3 cosgteu 3.25 .9839 m .885 .cetoqwcgoo 8829:5609 m 2E9 5E8 93.0w .6. 6828029229 8m=mEom_.cwm..o:w Emma: 22.... $92. m 8282258 98.8.... 3.5.... F 890385883”. 59.88:. 82.22.83“. 8.2.5. m 5.6.89.9. ..o .2938 can .8389. 965 ...nmoow. «-9532» 25.3.6 .0 380595 .mmdz. m 8.926 o.m_>cm:2m .6389 8.33 2.92.62 ...:m8. 539.. 8.59835 «3.89.. 8:28 8.5 9:28:25 2838 93.9.2» 5.2.89.5 320303.353 280 9n».a:< unis—.00 can»... baEEufi .e> .3... .v 03a... ww..m<._. 77 ><><>< 3585a! XX XX 50>... ><><>< c2353 :3 9585. 3.505 350:uo..0u.x0 >uo.8:0.>u35am mo..-E-o 3 882.; . 033.. :. :o.mm2nxm .93... 3:35:93 «83.2.. I x 393:. m 82288.. 22g: $051. w 89388.. 9.9m... .mazaz. n 5662532 .180. $mcmmo§nmo 4.8-5226 :Iomz. mx__$mcomeu>fio 229.». 22058.. 6.92 3.394.. . mam... 6.0.8.2698 250.853.3552 .254. .3525 < wE>~cmoo-_>omoxo-m 3.68532. m $9828.38 4. mE>~cmoo-_>.mo< 229. 2% $92. m ommcmmoazmu 22.82 619.. .> 3%. m 32086on 6:82 280 358.6. $8229. $2323 $820 6104.. .> 8%. o magmasgcmu .282 5.1. 38 35. m 82298: 78 XXXX XXXX ><><><><><><>< 025:3 EaEEu! €59. v. .0589 .98. £265 .8399“. .oxaé. o 895. 299.. 8658502.: :33. @5085 P >m> 50.. F .28. £265 .8329”. gums... . $9.... $2.. 299... 8.3.8 comes... coucfixmcfio £35 .0 9.033.501 .3..ng N 8906:97w 9.25936 38890.5. .59.“: 89:58... 28 as. 99.8 68. .mmm.:.mmmu-m-9.m3 89.38% we cozmmcom 6320. 382...? Bow o_c_Em.5oc_bmom-z mamcamocaocoe 9.53.6 «.8553... an... 323953.553 0.30 «325.2 925.50 2.3: baEEaE .u> 0333‘ .N can» 79 xxx x XXX ><><><>< xx 3252 EuEEa! x .vo.mm~m 2...... $9886»: 3.8.3. 3.58 3803020....5 .253. F «88 F 882 8888...... 8.888.. 8882.5 68mm. 988 5.28 8.8...-Sw ”coma £205 nougoommvgm 35m. 882 88.88.. 8.88 22 m 882 2...-va ...vzmoo. 2...-.. 88882 2.828 8&8 E8 3 ......x. 5 882 2.12.88. cmvmz. .88. F 2.89m ... 88:8 2.3 3.3323 Em: 392. .o .203 520:2 .8052: = .2. (21 Eat :ozntomcat. 30.3:038358a 230 .85. m 882 88882. 8.88m .3080... = .2. (zm Eat cozatumcaz. 80 ><><><><>< ><><><><><><>< x x 3252 EuEEaE ><><><><><>< 8.8%.. 0.88 .5888 8.88-<.m 8m. :88... 88888-...m .m<..&>._. .N< mmma..o...amo...a .mEowow>.. m 33.638383 2 :22. P 88.939.308.382 .2002. r 5.82888 .0580. .8632 .N..u< .5922. F 8.8.88.8 8.888 85.0.. mEEmm m .208. .893: 2.628%: .mem. m 8..-? 8.8. 8.55.28 88. .Em Gm... .208. oEQCnEm o...ao..o.>:k .Em. Em 7883888888 .3"... v .208. bo.m_:mo. 5.9.9... 8.9.0.... o5 .88 09.8: .25. m 8.8. 88.88.. 8.8m .32. m 88. 8.8.8.8.. 8.828... 8.2.8 .28. 8.8.888: __ .2. <><><><>< >< mama: E :o_mm2nxo 3:9; 22:35:?» «335:. I x Amxon: m mmmc2>xo£mo 2923650232.”. @5849 N4. game .cmtoamcm: mmofimfimdoa mm 2.83 ..mEmo 320w 619v m muncmooazmn 9.28.. 8x1; 32:98... S .5”: Ammmahmcmzamooéjéca.m-N 352359 F 3903553803“. 530% F $a.m_mcm.=>w83.o 529a 2&5: mmmEEofl 29.38%. 38:22 Em=35mE $98: Az>d @2080: 838:0 86.9 .93 2.8.3 EosmmEm> 7mg; 5.5% v 5589 .903 5:65 «$305: EmzonaamE 398... 3033035353 250 88%: m 8.88.. 8388 .93 £59.39 0:339:30 255$ 6558 2 37809 $22 $2835.05 NEESE NF 82: $8: 38: 529a Bfizsmémmogs. : Fxmaszv 3 $2: $2: $22 5205 85269535 335$ F 32: 82: Essa 8%.sz 58:: 658:: m: 82: .86st coszocamofiAtéEmEb 3.0533430 33:3 cause. 2.323 530.5 82 XXX ><><><>< 38.3 XXX XXX 50>... 33:8 ES: :xo€momu mctmmiuzmcamosa :89 P 8959.88 8.5.50 5333 wag--3350 5.3 m 82:6me 2mm: www.mEom_ 92393 $820 $10.: m $mcmmo§cmo $89 «320023 ...—8:333 3:82 68%: 8% 52% 96:5 <5 CPS—2v F www.mzmcfigoamEEi Comov mamboficmEScmfioé $5:ch Cx.m_a< E335 anabfimcmbtfii @523 ~< 5x25. .335: 1 53.5 8.3.3 28 Eu“. aamfiv N £205 9.6% 28 ran. 9.28 an... awe—fico‘aaifian 950 £3.22 c.2250 2.3:. 33.. .m> 3093‘ .n 2am... 83 XXX XX 3253 ><><><><>< ><><><><><>< X .52.. Eggs; .22 En“. 95955 5.3»... amok: mmm£c>m m Eocmamumoi COmov mmmEEcmbSochod 95an A3>odv P 92582 .mom3m3m .~o_m:zw$ mnfim >38 5ch ace. Em> 6 5.6086 Rink: .. 33.658539; 850$ m 892825 (8-32 Exagv m 38:. £29.. omfizsgmmogs Ammoowv m 9:3ch mciogo 6 53935 :xmaie F 32; 32:. 599.. 8.9.? 83.5. 6§a<§ 9 32: 522a nemzsmgmogs. Agagv m 82: $2: 822 5990. 8626953.: €539 v .BEmE .cotoamcmz 3836 0055.8: N 2E5 5E3 920m 60:): F newscasts gamma: Somfiuhwv F mEEmEooéfiEou 60am: _ __ 563 cosatomcm: .9200 Ami/Ext 0 £205 oofifiowwm 8339 9.950: 22?: Sub .908 253950 0:30:95» ...mme N .203 coautomcmb 96:3 «580.0 382.69 .905 Emmmwv F .209 86:5 53% rename .9me 86x05 20 3.. 8on ants N .203 83529. mc=m>8< £335 .203 539.835.: = .2. (zm >mo_2:9>§>5ua 2.00 6§m<§ 2 9.2: 52.5 86282033 8x835 m 32: $2: $2: 529... Emzsmémmogs. 84 ><>xo€momo mc_.mm_>u..mcamozm $3.5. 5906 30.9309 938 0.5902905 35% . 2.1.32 .mo_m.vm:m No.35“: mEom 3.8 Ease ace. 30> .o cozmmcem Earn... 2% :92. m 8mcmmo§cou 29.68.. .2849 m .889: .cotoamcm..oo 38:35:68. m 2E5. .mEmo 220m 213.6. v .853 .3535: 38:5 83:35. N 2.6.3 .053 mSEm $0.6. _ $29620 2&0. mmm..mEom. mfifimoca 9.83.0 $9.83. m 522.. m..._TommEmoccmE-mc_ocmzcm-co_.mu9mou mm axasz. o 82: c655 8.9.8-5852 .82. N 5.8. 83.8%.. 959.2 awn—hm. 323:8 2:22.205 .3353 :2. E mfifimEEE. @5335 83.3808 223.2530 328:0.»«353 6.xa<§ o. 38:. 529a B.m>_.om-cmmo._2 2.00 5232. . $2... $2: 5205 862.8 Smog: 255.. . E92,. 9552.8 $85. 3.3.8 5%... 3.. 9?. .3352. 55.9.9» xm<2 A 250$ P .8868 2:5. 520-828.: $82.53 «60.33 85 ><><><><>< 309.5 ><><>< xxxx .83.. XXX moaémé B .852; . can»: 5 :ofinaaxm 5:2: ascooEcofi «8355 I x Snowy P mag. 2232a 8.88an 52% 0333988 mc:0m.>u=mcamoca :89 ' 89>on88 82:50 8>QDcm=m=>o>2>9m ,ODOV c2309 €88.95 038m 505 ON 3926 99>:sz :99 333.620 :28 x. $9255 2:850 228 _> 323:5 289.8 33on 3.23 338953955}. 980 220; 2% ¢o .83.. buEEa! .m> ones—Z 33... .u> emoa5< N x NQ>>< N 62.9.... 8628... u & moaémd .Qm 8.6.50.2... :0 E8536 8: n N Homoafim E 09093 so: 338:. u > God v n: moaémd .3 8.8%.. u x Ammoowv m-mc__mcm_m 9.3.030 B 53935 828va CE 522.. 9.6% £228 85 5.2% m 2% 599.25 cam/E. v 592.. 8.9.5 28 Em”. £98 :68... 9.59835 2989.. 9.295 319: 9:76.852 8 .__m_..m .58an 22 .283 __-< 50.9898... 235. 5532587226. 230 .8331 moan—CG coated—:3 oaamfi .m 23¢... 88 CHAPTER 4 IGF-I and leptin hormonal infusions alter the cell proliferation and transcriptional profile of the prepubertal bovine mammary gland ABSTRACT In dairy cattle high-energy diets before puberty promoting body growth rates above 1 kgld impair mammary development. Rapid growth regimens enable earlier calving but also lead to reduced milk production. Biological mechanisms to explain this result are not well understood, but insulin-like growth factor-I (IGF-I) and leptin likely play a role. Adipocytes produce the protein leptin, and local leptin concentrations in the mammary gland increase in response to increased fat deposition. IGF-l stimulates and leptin inhibits proliferation of bovine mammary epithelial cells in vitro and in vivo. This is in contrast to human and mouse studies, which show a proliferatory effect of leptin on mammary epithelial cells. Our objective was to elucidate the effects of IGF-I and leptin infusion on cell cycling in mammary epithelial cells, identify key genes controlling the interaction of these two hormones in bovine mammary tissue, and determine if Ieptin alters lGF-l-induced expression of genes involved in the major IGF-l signaling pathways. A functional genomics approach was used to analyze parenchyma tissue collected from prepubertal Holstein heifers after 7 d of intramammary hormone infusions. With this model Silva (2002) previously showed that IGF-I stimulated proliferation of epithelial cells 60% and leptin 89 blocked this mitogenic effect of IGF-I (278). Each mammary quarter of six heifers received infusions via the streak canal of one of four treatments: IGF-I, Ieptin, IGF-l plus Ieptin, or saline plus BSA control. Intramammary infusion of IGF-I increased the percentage of BrdU-labeled mammary epithelial cells in the S-phase of the cell cycle by 52%. Intramammary infusion of Ieptin decreased BrdU-labeling 48% in IGF-l-treated quarters, and 19% in saline-treated quarters. No differences were observed in the number of cells within the cell cycle, as indicated by the Ki-67 labeling index. A significant increase in apoptosis, as marked by caspase-3 immunostaining, was seen in IGF-I and leptin treated quarters relative to saline control. One possibility is that increased concentrations of IGF-I may promote the proliferation of mammary epithelial cells by increasing the progression of cells into the S-phase of the cell cycle in vivo in mammary epithelial cells in prepubertal heifers. Conversely, increasing the Ieptin concentration above normal may reduce epithelial cell proliferation by limiting the progression of cells into the S-phase of the cell cycle and stalling them within the G1 phase of the cell cycle. We hypothesized that changes in cell proliferation should be accompanied by alteration of the profile of transcribed genes involved in cell proliferation and survival. Gene expression profiles of total parenchyma mRNA from each quarter of the six animals were examined using bovine-specific cDNA and oligonucleotide microarrays. Numerous genes mediating intracellular signaling were significantly altered (P<0.05) by the treatments, all with small fold changes (< 1.5-fold up- or down-regulation). IGF-l induced changes within the JAK/STAT pathway, the Ras/Raf pathway, the PI3K pathway, and the Bad/Bcl 9O pathway. Similarly, Ieptin altered expression within the JAK/STAT pathway, the Ras/Raf pathway, and the PI3K pathway including increased expression of the suppressor of cytokine signaling protein SOCSS. These data suggest that Ieptin may counter-regulate IGF-l action through IGF and cytokine receptor families’ downstream signaling pathways. Overall, this study has demonstrated that IGF-I and leptin intramammary infusions in prepubertal heifers can serve as a useful model for study of hormonal regulation of cell cycle progression and gene transcription in the developing gland. Introduction Mammary growth and development are precisely controlled processes that are governed by the complex interplay of circulating hormones and locally acting factors. Classic studies in the laboratory and in domestic species established that normal development of the mammary gland is contingent upon the interactions of a wide range of circulating hormones secreted from the ovary and hypothalamus (69;194;214). The importance of tissue-derived factors in promoting mammogenesis has more recently been investigated (9;68;237;327). Mammary epithelial cells form the foundation for future lactation potential, and numerous studies have indicated that high planes of nutrition have a negative impact on this cell population in ruminants (38;43;117;265;267;324). High energy diets increase the serum concentrations of both IGF-I and leptin (28:241). IGF-l has a well-established role in promoting mammary epithelial cell proliferation and survival, mediated by the IGF-l receptor (IGF-IR), which is able to activate 91 several distinct signaling pathways (124). Leptin is a hormone secreted by fat cells (112;113;336) that mediates central hormonal and metabolic responses to the nutritional state of mammalian species. Leptin has been shown to be synthesized in mammary tissue (285) and can also be found in the circulation, suggesting that either locally or extra-mammary-produced Ieptin may influence the milieu surrounding mammary epithelial cells (278). Recently our lab has shown that exogenous bovine and ovine Ieptin inhibit the IGF-l stimulated proliferation of bovine mammary cells in vitro in the MAC-T cell line (276) and in vivo in a mammary infusion study (278). It is believed that Ieptin acts in the mammary gland by signaling through the long form of the Ieptin receptor (Ob- Rb), which is expressed by mammary epithelial cells (Silva, VandeHaar, et al. 2002 #560). Interaction between members of the IGF family and interleukin-6 family of class I cytokine receptors has been demonstrated within the central nervous system (47;48;260;261), skeletal muscle (303), adipose tissue (246), ovary (4:288:333), liver (46), and mammary epithelial cells (284). In accordance with these results, it is here proposed that Ieptin, a hormone produced by adipocytes, could potentially mediate the inhibition of lGF-l-stimulated mammary parenchymal development through crosstalk between the IGF-I and leptin receptors. Evidence supporting distinct interactions between IGF-I and leptin on cell proliferation both in the mammary gland and other tissues supports the hypothesis that these factors may act synergistically to alter the profile of 92 transcribed genes in the prepubertal bovine mammary gland to affect mammogenesis. It is believed that intramammary infusions of the hormones IGF-I and leptin alter the profile of transcribed genes mediating crosstalk between the IGF-I and leptin and cell proliferation of mammary epithelial cells. To test this hypothesis the expression profiles of genes likely to control mammogenesis from quarters infused with IGF-I and or Ieptin hormones in prepubertal heifers was measured using a functional genomics approach. The first objective of this study was to extend the analysis of Silva (2002) on the effects of in vivo hormonal infusion on mammary epithelial cell proliferation (278). This was done by measuring the total number of cells committed to the cell cycle and apoptosis using immunohistochemical analysis of Ki-67 and Caspase-3, respectively. The second objective of this study was to examine global expression of transcribed genes using cDNA and oligonucleotide microarray platforms. Thirdly, quantitative RT-PCR was used to analyze the mRNA expression profile of major components of the IGF-I and leptin signaling pathways likely to mediate crosstalk between the two hormones. MATERIALS AND METHODS Animals. Six prepubertal Holstein heifers weighing 209110 kg were purchased and housed in individual stalls at the Beef Cattle Research Center of Michigan State University. Heifers were fed once a day a corn silage-based diet at restricted intake formulated to achieve a body weight gain of 700 gld, according 93 to standard recommendations (216293). This growth rate was selected to promote normal mammary development (Sejrsen, Purup, et al. 2000 #5150). Heifers had free access to water. The Michigan State University All-University Committee on Animal Use and Care approved experimental procedures. Infusion procedure. Intramammary infusions of hormones were performed as described in Silva (2002). Six heifers were used to test the effects of intramammary infusion of leptin on proliferation of mammary epithelial cells. Heifers were adapted to diet and environment for 10 d before the start of infusions. The front or rear halves of the udders were treated as separate strips and randomly assigned to receive intramammary infusions of either IGF-l or saline (278). To balance for the effects of strip, two heifers were infused at each time, one with IGF-l infusions in the front half, and the other with lGF-l infusions in the rear half. The left and right side of each strip were assigned randomly to receive intramammary infusions of either recombinant ovine Ieptin (oLeptin) or saline over a 7 d infusion period. Treatments were then defined as IGF-I (IGF), IGF-l + oLeptin (IGF+LEP), saline (SAL) and saline + oLeptin (LEP) (Silva 2002 #5810). Thus, each treatment was applied to a total of six quarters, one in each heifer, and the effect of oLeptin always was compared against its contralateral quarter, which was not infused with Ieptin (278). Lyophilized recombinant human IGF-l (GroPep Pty Ltd, North Adelaide, Australia), which is identical to bovine IGF-l, was dissolved in 10 mM HCI, neutralized and diluted in sterile, physiological saline at a concentration of 1 pg 94 lGF-l/ml. The saline contained 1 mg of bovine serum albumin/ml (lnvitrogen, Carlsbad, CA). Ten micrograms of IGF-I were infused daily using a 12-ml syringe with an intramammary infusion cannula (278). Recombinant ovine Ieptin was purchased from Dr. Arieh Gertler of the Hebrew University, Jerusalem, Israel, and was produced as described by Gertler et al. (120). Lyophilized Ieptin was dissolved in saline at a concentration of 10 pg oLeptin/ml. Leptin-treated quarters were infused daily with 100 pg of oLeptin. Solutions were infused at 0800 h each day (278). On the last day of treatment, additional infusions were given at 2000 h to maximize the effect of the infused hormones on BrdU incorporation the following morning. Heifers were killed 14 h after the last infusion. To avoid any microbial contamination, all materials used during infusion were sterile. The teats were cleaned with an iodine solution and the teat ends scrubbed with ethanol before infusion (278). The solutions infused were tested for the presence of endotoxins using a commercial kit with sensitivity to detect 0.006 ng of endotoxin/ml (LAL, BioWhittaker, Walkersville, MD) (278). Sample collection and tissue processing. Mammary parenchymal tissue was harvested from the heifers as described previously (278). At slaughter, mammary glands were removed within 5 min of death and samples from three regions of the mammary parenchyma (proximal, intermediate, and distal to the teat) were collected and fixed in neutral buffered formalin for 18 h (Sigma). At this time mammary samples (~5 g each) from each of the three regions of the mammary parenchyma were collected for 95 RNA extraction, sealed individually and snap frozen in liquid nitrogen for storage at -80°C until use (278). BrdU immunohistochemistry Bromodeoxyuride-incorporation immunohistochemistry was performed by Silva (2002). Total epithelial cells and BrdU-labeled epithelial cells were quantified in three microscopic fields from each of three slide sections (Figure 1A). Two independent evaluators blinded to the identity of the samples enumerated cells, and an average of 2935 epithelial cells were counted in each quarter to calculate the percentage of cells labeled with BrdU (278). Ki-67 immunohistochemistry lmmunohistochemical analysis of Ki-67 labeling was performed by Matti Kiupel, Michigan State University Department of Pathology and Diagnostic Investigation, as described by Kiupel et al. (162). Room temperature paraffin- embedded tissue samples were incubated in 10 mM HCl/citrate buffer (pH 6.0) and heated in a microwave oven (600 W) for 25 minutes. Slides were transferred to 100% ethanol and immersed in methanol containing 1% hydrogen peroxide for 30 min to block endogenous peroxidase activity. Sections were then washed in distilled water and PBS (pH 7.6) containing normal equine serum for 15 min to block nonspecific antibody binding. A mouse anti-Ki-67 antibody was used for immunostaining (Immunotech S.A., Marseille, France) at a dilution of 1:200 in 10% normal equine serum for 60 minutes. Antibody binding was visualized with 96 the Vector ABC Elite kit (Vector PK4000) with 3,3’-diaminobenzidine substrate (Vector SK4100). Final washing was performed in distilled, deionized water. Sections were then counterstained with Mayer’s hematoxylin, then dehydrated, cleared, and mounted in DPX (BDH). Equine lymphoid tissue was used as a positive control (162). Total epithelial cells and Ki-67-labeled epithelial cells were quantified in three microscopic fields from each of three slide sections (Figure 1B). Two independent evaluators blinded to the identity of the samples determined the percentage of Ki-67 labeling, counting an average of 1710 epithelial cells per quarter. Caspase-3 immunohistochemistry Cell death analysis was performed by Matti Kiupel essentially as described by Kiupel et al. (161). Tissue sections were deparaffinized and incubated in target retrieval solution (Dako Corporation, Santa Barbara, CA) for 15 min at 95-990. lmmunostaining was performed using the Dako Autostainer (Dako Corporation, Santa Barbara, CA) by incubation with a polyclonal anti- mouse anti-activated caspase-3 antibody (R&D Systems, Minneapolis, MN) at a dilution of 1:100 at -4C overnight, followed by a 1:500 dilution of a biotinylated goat anti-rabbit secondary antibody (Dako Corporation, Santa Barbara, CA). A peroxidase labeled streptavidin-biotin complex (Dako Corporation, Santa Barbara, CA) was used for antibody binding localization, after which antibody binding was visualized with a 3,3’-diaminobenzidine as a chromogen substrate. After a final wash in automation buffer, sections were counterstained with 97 Lemer’s hematoxylin. Sections incubated with isotype control antibodies were used as negative controls (161). Total epithelial cells and caspase-3-labeled epithelial cells were quantified in three microscopic fields from each of three slide sections (Figure 10). Two independent evaluators blinded to the identity of the samples determined the percentage of caspase-3 labeling, with an average of 3136 epithelial cells counted per quarter. Infusion study statistical analyses. Data from the IGF-I and leptin infusion study were analyzed by least squares ANOVA using the following model, where IGF is saline or IGF-l infusion, Ieptin is saline or Ieptin infusion and region is the parenchymal regions sampled (proximal, intermediate and distal to the teat), as described by Silva (2002). The model was: Y= p-I- heifer+ IGF+LEP+ lGFxLEP+e The MIXED procedure of SAS (2000) was utilized with the data log- transformed to remove heterogeneity of variance (255). Data are presented as back-transformed. Estimates of variance were calculated as confidence intervals from the transformed data. BrdU-labeling among parenchymal regions was compared by Bonferroni t-test adjusted for 3 comparisons (217). RNA isolation procedure. Total RNA was extracted from frozen mammary tissues from intermediate tissue samples (after homogenization) with Trizol reagent (Invitrogen Life Technologies Corp., Carlsbad, CA) essentially as recommended by the manufacturer (lnvitrogen Life Technologies Corp., 98 Carlsbad, CA). Tissue was kept cold using dry ice and ~200 mg of mammary tissue was weighed and directly added to 3 mL of Trizol reagent in a 15 ml culture tube. Tissue was homogenized using a Polytron for ~30 3. Between homogenization of individual samples, the Polytron bit was rinsed with diethyl pyrocarbonate (DEPC, Sigma) treated water, and RNAse Away (Molecular BioProducts, San Diego, CA). Samples were aliquotted into three 1.5-mL centrifuge tubes and incubated at room temperature for 5 min. Chloroform (200 pL) was added to each tube and samples were manually shaken, incubated for 3 min at room temperature, and centrifuged at 10,500 x g for 15 min at 4°C. The upper phase was transferred to a clean tube. lsopropanol (500 pL) was added to precipitate RNA. The tube was vortexed, incubated at room temperature for 10 min, and centrifuged at 10,500 x g for 10 min at 4°C. The isopropanol was decanted and the remaining pellet was washed with 75% ethanol, centrifuged at 8500 x g for 5 min at 4°C, decanted, and dried. RNA was then subjected to DNase treatment (RQ1 RNase-free DNase; Promega, Madison, WI) following the manufacturer’s protocol (Promega, Madison, WI). Nuclease free water (52 pL), DNase buffer (10 pL of 10X; Promega), and DNase (1 pL of 2U/pL; Promega) were added to the pellet and then incubated at 37°C for 30 min. A phenol/chloroform separation step was performed by adding RNase-free water ' (37 pL) and phenol/chloroform (100 pL). The tube was shaken and centrifuged for 2 min at 14,000 x g. The upper phase (~90 uL) was transferred to a fresh tube, sodium acetate (3 M, 9 uL) and ethanol (250 pL) were added to this phase, and the mixture was stored overnight at -20°C. The following day, 99 microcentrifuge tubes were centrifuged at 14,000 x g at 4°C for 20 min. The liquid was decanted and the pellet was washed with ethanol (75%, 500 pL). The tube was centrifuged at 14,000 x g at 4°C for 10 min. The ethanol was decanted and the pellet was dried under air for 15 min. The pellet was resuspended in 50 pL of nuclease-free water and incubated at 60°C for 10 min. RNA used in the BOTL4 microarray study was quantified using spectrophotometry at 260- and 280-nm and quality was checked by 288 and 188 rRNA band visualization following gel electrophoresis and ethidium bromide staining. The quantity of RNA isolated for the BMET microarray study was determined using the NanoDrop ND- 1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE) and quality was checked with a RNA 6000 Nano LabChip kit and Agilent 2100 BioAnalyzer (Agilent Technologies, Palo Alto, CA). BOTL4 microarray experimental design. The BOTL4 microarray used is an expanded version of those previously described (7;67;98;328). The BOTL4 microarray contains 4800 total spots consisting of 709 bovine EST clone inserts developed from a normalized total leukocyte (BOTL) cDNA library (328) and an additional 627 amplicons representing additional genes including cytokines, growth factor associated proteins, receptors and intracellular signaling components (67). A list of known genes represented on the BOTL4 micorarray and their sequences can be viewed at Mowhiteansmsu.edu/public php/gd- bfovine-immgnolm Amplicons were suspended in 50% DMSO and spotted in triplicate on glass slides in a 4 x 12 pattern of 48 patches with 100 spots per 100 patch. This array included 144 spots representing GAPDH (three spots in each of 48 patches), 75 spots representing B-actin, 78 spots representing ribosomal protein L-19 (RPL-19), 48 synthetic Lambda Q gene spots (1 per patch), 303 negative control spots (DMSO only), and 144 blank spots (nothing spotted). Screening for leptin-, IGF-I- and IGF-l + Ieptin-induced mammary parenchymal gene expression changes was performed on all six animals using a loop design (157:158) balanced for Cy3 (green) and Cy5 (red) fluorescent dye labeling (Amersham Phannacia Biotech; Piscataway, NJ) for each of the four treatments (24 arrays total). RNA pools of 20 pg per sample were used for cDNA synthesis and the two dye coupling reactions; the pools were separated into two aliquots of 10 pg each for reverse transcription and dye-labeling. Within a four comparison loop, the first comparison made was between SAL vs. IGF, the second comparison was between IGF and IGF+LEP, the third comparison was between IGF+LEP and LEP, and the fourth comparison was between LEP and SAL. Three loops were run in the fonlvard direction and three in reverse (i.e. in array no. 1 three animals had Cy3 dye-labeling of the IGF-I treatment and three had Cy5). Comparisons included within the loop were designed to test the major comparisons of interest with regard to the BrdU incorporation results, in which the most significant results were found between the SAL and IGF (array no. 1) and IGF and IGF+LEP (array no. 4) treated quarters. BMET microarray experimental design. The BMET microarray was used to compare gene expression profiles of SAL, IGF and IGF+LEP infused quarters. 101 The BMET array was described previously (Chapter 2) (97) and contains 2,348 spots representing genes involved with bovine metabolism and signal transduction. This array included 10 metabolic housekeeping genes from Arabidopsis thaliana used for spike-in controls, 1320 blank spots (printing buffer only spotted) and 64 empty spots (nothing spotted). Complete annotation of the BMET gene set can be found at h_ttp:/Iwww.nutri- genomics.org/data.html. The BMET microarrays were printed at the Massachusetts General Hospital Microarray Core Facility at Harvard University (httpszlldnacore.mgh.harvard.edulmicroarray/protocol-printing§html) as described in Chapter 3 of this thesis (Materials and Methods). Two microarray comparisons, SAL versus IGF, and IGF versus lGF+LEP, were screened directly for all six animals with the design balanced for Cy3 (green) and Cy5 (red) fluorescent dye labeling (Amersham Pharrnacia Biotech; Piscataway, NJ) (12 arrays total). Total RNA (15 ug) from each quarter of the three treatments of interest was used for cDNA synthesis and dye coupling reactions. In the first comparison (SAL vs. IGF), SAL cDNA was labeled with the Cy3 dye (Amersham Biosciences, Piscataway, NJ) for three heifers (20, 28, and 29) and IGF-l was labeled with Cy5 (Amersham Biosciences, Piscataway, NJ); SAL was labeled with the Cy5 dye in the other three heifers (35, 36 and 88) with IGF-I samples labeled with the Cy3 dye. This reciprocal dye design was repeated for the second comparison (IGF vs. IGF+LEP). 102 Preparation of labeled cDNA for BOTL4 hybridization. Gene expression profiles from SAL, IGF, LEP, and IGF+LEP treated quarters of all six heifers were evaluated on cDNA microarrays (24 microarrays total). Synthetic lambda Q gene containing an engineered poly-A tail was spiked into each cDNA synthesis reaction (650 pg) to provide a control for cDNA synthesis and dye labeling efficiency. For cDNA synthesis 8 pg of sample RNA was converted to cDNA using the Atlas Powerscript fluorescent labeling kit (BD Biosciences, Alameda, CA). First strand cDNA synthesis, RNase-H cleanup of template RNA, and sample purification were performed according to procedures outlined by Aho et al. (7). Resulting cDNAs were coupled to Cy3 and Cy5 dyes using the Atlas Powerscript fluorescent labeling kit (BD Biosciences, Alameda, CA) using the manufacturer’s protocol. Labeled cDNAs were purified to remove unincorporated dyes, combined, and concentrated according to procedures outlined previously (7). The labeled cDNAs were incubated at 70°C for 5 min just prior to an 18-hour array hybridization using a GeneTAC Hybridization Station (Genomic Solutions, Ann Arbor, MI) With a step-down hybridization protocol as previously described (64:66). Following hybridization, microarray slides were washed twice in low stringency buffer and once in high stringency buffer in the Hybstation unit (Genomic Solutions, Inc., Ann Arbor, MI). Finally, microarrays were rinsed once in 2X SSC (saline sodium citrate, 0.15 M NaCl plus 0.015 M sodium citrate) and once in double-distilled H20. Microarrays were dried by centrifugation in a cushioned 50-ml conical tube and immediately scanned with a GeneTAC LS IV 103 microarray scanner (version 3.01). GeneTAC L8 software version 3.3.0 (Genomic Solutions, Inc., Ann Arbor, MI) was used to process microarray images, find spots, integrate robot spotting files with the microarray image, and to create reports of raw spot intensities. Total intensity values for each dye channel were converted to comma-separated value data files and exported into Excel spreadsheets and loaded into SAS (255) for data normalization and analysis. Preparation of labeled cDNA for BMET hybridization. Gene expression profiles from SAL, IGF, and IGF+LEP treated quarters of all six heifers were evaluated on oligonucleotide microarrays (12 microarrays total). For cDNA synthesis 15 pg of sample RNA was used as a template in reverse transcription reactions (SuperScn'pt III Fluorescent Labeling Kit; lnvitrogen Life Technologies Corp., Carlsbad, CA) in which random hexamers were used as a primer. In the Superscript Ill system, cDNA is prepared with a randomly incorporated amino- modified dUTP. Following first-strand synthesis, cDNAs for each tissue within an animal were differentially labeled using N-hydroxysuccinimide-derivatized Cy3 and Cy5 dyes (Amersham Pharrnacia, Ltd., Piscataway, NJ). Differentially labeled were combined and concentrated to 10 pl by using Microcon 30 spin concentrators (Millipore Corp., Bedford, MA). Microarray hybridization was performed after addition of 100 pl of SlideHyb-3 (Ambion Inc., Alameda, CA) to the concentrated Cy3—Cy5-labeled probe cDNAs (7). The labeled cDNAs were incubated at 70°C for 5 min just prior to 18-hour array hybridization in Corning Microarray Hybridization Chambers (Fisher Scientific) placed in water baths set 104 to 42°C. Following hybridization, microarray slides were washed with 2xSSC saline sodium citrate (0.15 M NaCl plus 0.015 M sodium citrate), 0.1% SDS for 4 min at 42 C, 2X SSC for 4 min at room temperature, 0.2XSSC for 4 min at room temperature, 0.06X SSC for 10 sec at room temperature and once in double- distilled H20. Microarrays were dried by centrifugation in a cushioned 50-ml conical tube and immediately scanned with a GeneTAC LS IV microarray scanner (version 3.01). Genepix Pro software version 6.0 (Molecular Devices Corporation, Sunnyvale, CA) was used to process microarray images, find spots, integrate robot spotting files with the microarray image, and to create reports of raw spot intensities. Total intensity values for each dye channel were converted to comma-separated value data files, exported into Excel spreadsheets, and loaded into SAS (255) for data normalization and analysis. Statistical analysis of microarray data. Microarray M vs. A scatterplots to detect and account for potential dye intensity biases analysis was performed essentially as described in Chapter 3 of this thesis. Statistical analysis of LOESS-adjusted log intensities was performed using a mixed model approach consisting of two steps (9, 11). For the BOTL4 microarray, the first step involved an array-specific spatial variability normalization model: Iog(ygijk/mn) = p + Geneg + Treatment,- + Dye] + Animalk + SIide(Arra y) kloop + Patch(SIide)k/m + Egijklmn 105 Where ijjklmn represents each observed fluorescent intensity signal; p is an overall mean value; T. is the main effect of treatment i (SAL, IGF-l, IGF+LEP, LEP); D; is the main effect of dye j; A. is the main effect of animal k; S(A)m is the effect of slide k within array I; P(S)klm is the effect of patch m within slide k within loop I; and egg-km... is a stochastic error (assumed to be normally distributed with mean 0 and variance oz). The second step involved gene-specific analyses for the estimated residuals (eukm) obtained from the normalization approach obtained from step one to test the effect of treatment on expression profiles for individual genes. These models were: égijklmn = mg + Tgi + ng + T'Agik + Agk + S(A)gkl + Sp(8)gkln + egijklmn Where all the effects have the same definitions as for the normalization model, except that P(S)gk.n represents the random effect of each spot, and 9 index is added to account for effects now being gene specific. The error terms ejjklmn were assumed to have independent normal distributions with gene-specific variances 092. Similarly, for the BMET microarray analysis data were analyzed first with the following model, where all terms are random except dye: yadj = dye + array + day*array + block(array) + dye*block(array) + e Next, residuals calculated from step one were analyzed in a second step according to the following model: resid = heifer + treat + treat*heifer + dye spot(heifer) + e 106 Where heifer and array are confounded, and treat was tested by treat*heifer. These analyses were computed using the MIXED procedure of SAS (255). In the gene-specific analysis, genes were considered significantly different if P<0.05. Quantitative RT-PCR validation of microarray gene expression changes. The expression differences of certain genes found to be significant by microarray analysis were checked using quantitative real-time reverse transcription-PCR (Q- RT-PCR) using an Applied Biosystems 7000 DNA Sequence detection system (Perkin Elmer Corp., Foster City, CA). Total RNA was extracted from each quarter from each of the six heifers, quantified and checked for quality as described previously. The PCR system used was the ABI PRISM 7000 Sequence Detection System (Applied Biosystems). RNA was converted into first-strand cDNA by adding 1 pg of total RNA to a 12-pl reaction containing 10 mM oligo(dT)13 primer. Following a 5-min incubation at 70°C, the reaction was chilled on ice and pH adjusted by addition of 4 pl of a buffer supplied by the reverse transcriptase manufacturer (final reagent concentrations were 50 mM Tris-HCI, pH 8.3, 75 mM KCI, and 3 mM MgClz), 1 mM deoxynucleoside triphosphates, and 200 units of Superscript II RNase H reverse transcriptase (Invitrogen Life Technologies). Incubation at 37°C continued for 30 min in the presence of RNase-H to remove the original RNA templates. Heating at 70°C for 15 min subsequently inactivated RNase-H. The 107 reaction tubes were removed from the thermocycler and 0.2 pL of 0.5 M EDTA was added and mixed. Then, 25 pL of water, 5 pL of 3M sodium acetate, and 125 pL of ethanol (-20°C) were added to the tube. Tubes were then allowed to precipitate overnight at -20°C. The reaction mixture containing the newly- synthesized cDNA was transferred to a new, clean microcentrifuge tube. These tubes were then centrifuged at 14,000 x g and 4°C for 20 min and the supernatant was decanted. Pellets were washed with 250 pL of ethanol (75%, - 20°C) and the tube was centrifuged at 14,000 x g and 4°C for 6 min. The supernatant was decanted and the pellet was allowed to dry for 15 min. The pellet was resuspended in 50 pL of water and incubated at 60°C for 5 min. The cDNA concentration was analyzed using a spectrophotometer (NanoDrop), and was then diluted to a final concentration of 10 ng/pL and stored at -80°C until the Q-RT-PCR reaction was initiated. SYBR Green PCR Master Mix (Perkin Elmer Corp.) and gene-specific primers were used to perform Q-RT—PCR reactions. Primer Express Software (Perkin Elmer Corp.) was used to design all primers, which were then synthesized by a commercial facility (lnvitrogen Life Technologies). Primer sequences for genes analyzed in this report are included in Table 1. The amount of primer used was determined by performing an optimization matrix for each primer using three concentrations of primers: 50:50 nM, 300:300 nM, 900:900 nM. Dissociation curves were similar for all concentrations and the 3002300 nM matrix was chosen, thus 3 pL of primer was used for all experiments. Each gene of interest and the control gene were measured in duplicate. Within 108 each well of a 96-well reaction plate (MicroAmp Optical, Applied Biosystems), 30 ng of sample cDNA (3 pL), 6.5 pL DEPC water, 3 pL primer, and 12.5 pL Sybr Green (Applied Biosystems) were added. To determine an appropriate reference control gene by which relative mRNA abundances for genes of interest could be measured, a number of potential housekeeping genes were screened. These candidate control genes included [3- actin (B-actin), glyceraIdehyde-3-phosphate dehydrogenase (GAPDH), hydroxymethyl-bilane synthase (HMBS), hypoxanthine phosphoribosyltransferase 1 (HPRT1), TATA box binding protein (T BP), and succinate dehydrogenase complex, subunit A (SDHA) (Table 6). The variation in gene expression across all treatments for these six genes was tested using the methods of Vandesompele et al. (312). The expression of GAPDH and HMBS were not significantly different from 1 nor each other across all treatments (SAL, IGF, LEP, IGF+LEP) and were therefore suitable reference controls. The 2”" method of Q-RT—PCR analysis was performed as previously described (7:65:188). In this study analysis GAPDH was used instead of HMBS to serve as the control gene to calculate initial ACT values, as it has previously been shown to be an appropriate control gene in experiments with mammary tissue (284;286). This method allows for the determination of relative gene expression changes across treatments based on quantitative differences in the PCR amplified target reaching a fixed threshold cycle (CT) number for a specific 2-AACT treatment versus other treatments (Table 7). In our analysis, the CT for SAL was the calibrator used to determine relative expression changes for IGF, ' 109 LEP, and IGF+LEP treated quarters for each GAPDH-normalized test gene. Gene-specific standard errors were estimated using independent analyses of variance (ANOVA), which included the effects of quarter and treatment. All analyses were performed using the SAS System (255). RESULTS Intramammary infusion of oLeptin decreased BrdU-labeling of mammary epithelial cells. As reported by Silva (2002), in both saline- and lGF-l-treated quarters, infusion of leptin decreased the percentage of cells in the S-phase of the cell cycle (Figure 2B) (278). Leptin decreased (P < 0.001) BrdU-labeling by 48% in lGF-l-treated quarters (4.1 vs. 7.9 %) and decreased (P = 0.01) BrdU-labeling only 19% (5.0 vs. 6.2 %) in saline-treated quarters (Figure 2). In contrast to the significant increase in BrdU labeling in lGF-l-treated quarters relative to SAL reported in a previous study (275), no differences were detected between IGF and SAL in this study. Cell cycle commitment Is unaffected by treatment. The number of cells committed to the cell cycle was unaffected by hormonal infusion as measured by the percentage of cells labeled by the nuclear cell proliferation antigen Ki-67 (Table 1). The percentage of Ki-67 positive mammary epithelial cells ranged from 8.5-11.7%, but did not differ significantly across SAL, IGF, LEP, and IGF+LEP treatments (Figure 3). 110 The percentage of proliferating cells occupying the S-phase of the cell cycle was significantly reduced by Ieptin (Table 1). This was calculated by dividing the percentage of BrdU-labeled cells by the percentage of Ki-67-labeled cells. Both the main effect of Ieptin ((LEP + IGF+LEP) — (SAL + IGF)) (P=0.015) and the IGF versus IGF+LEP contrast (P=0.02) were highly significant (Figure 5). Apoptosis in prepubertal mammary epithelial cells is low. Using caspase-3 immunohistochemical staining as a marker of cell apoptosis (Figure 4), the pattern of programmed cell death in each quarter was measured. The overall percentage of apoptotic cells was very low (less than 0.25%). Although apoptosis was significantly increased in IGF, IGF+LEP and LEP treated quarters relative to SAL (Figure 4). The low numerical percentages of caspase-3 immunostaining in each quarter do not indicate a large biological significance due to any treatment. BOTL4 microarray SAL vs. IGF differentially expressed genes. IGF-l infusion significantly increased the expression of 31 genes and decreased 34 genes relative to SAL treatment on the BOTL4 cDNA microarray (P<0.05) (Supplemental Table 1). Of these 65 genes altered by IGF-I treatment, 15 were without known function according to BLASTn searches and were ignored for subsequent analysis. Gene ontology analysis revealed that IGF-l treatment induced changes in 38 genes that could be divided into four broad ontologies (Table 2). These included genes involved with extracellular receptor mediated 111 signal transduction (n=18), cell proliferation and apoptosis (n=7), transcriptional regulation (n=6), and cell structure and metabolism (n=8). BOTL4 microarray IGF-I vs. IGF-I + Ieptin differentially expressed genes. The comparison between IGF-I and IGF+LEP treated quarters revealed expression changes with 65 transcripts (P<0.05) (Supplemental Table 1). Eighteen genes were without known function, 34 were downregulated and 31 upregulated by IGF+LEP treatment relative to IGF. Genes of interest with regard to IGF and leptin signaling were involved with extracellular receptor mediated signal transduction (n=17), cell proliferation and apoptosis (n=9), and cell structure and metabolism (n=2) (I' able 3). BOTL4 microarray SAL vs. LEP differentially expressed genes. Leptin infusion altered the expression of 92 mRNAs relative to SAL treatment (P<0.05) (Supplemental Table 1). Of these, 25 had no known function according to BLASTn searches, 32 were downregulated and 60 upregulated. As with the IGF treatment results, genes of particular interest could be broadly categorized into four major categories (Table 4). Included were genes implicated in extracellular receptor mediated signal transduction (n=18), cell proliferation and apoptosis (n=7), transcriptional regulation (n=9), and cell structure and metabolism (n=16). BMET microarray SAL vs. IGF differentially expressed genes. Sufficient data for statistical analysis was not found for 25 of 2,348 bovine genes in the IGF vs. 112 SAL BMET microarray comparison. IGF infusion significantly altered the expression of 135 (5.8%) genes compared to SAL control on the BMET array (P<0.05) (Supplemental Table 2). This included the upregulation of 102 genes by IGF-I and downregulation of 33 mRNAs relative to control treatment (Supplemental Table 2). Gene ontology analysis also revealed that IGF treatment induced changes in 39 genes of interest that could be divided into four broad categories related to IGF and leptin action in the mammary gland (Table 5). These included genes involved with receptor-mediated signal transduction (n=19), cell proliferation and apoptosis (n=10), transcriptional regulation (n=6), and cell structure and metabolism (n=4). Significantly altered genes grouped within a number of pathways and gene ontologies according to GenMAPP analysis (72;87). These included the insulin signaling pathway, which included upregulation of the mRNAs encoding FOXO1A, IRS1, RAD, RAF1, PKCalpha, MAP3K4, MAPK13, MAP2K5, and downregulation of TSC2, EGR1 (Figure 9). Seven genes found within the TGF- beta signaling pathway were significantly upregulated by IGF-l treatment including TGFBR1, TGFBR3, FST, MADH2, STAT1, and EGF (Figure 8). Genes falling within the pathway regulating G1 to S phase cell cycle control were also significantly overrepresented in the analysis of control vs. IGF-l treatments. These included upregulation of mRNAs coding for MYC, CCNE1 (P=0.06), and E2F4 (P=0.07), and downregulation of CCND2, PCNA (P=0.07), and CREB3L1 (Figure 10). 113 BMET microarray IGF-I vs. IGF+LEP differentially expressed genes. The BMET microarray comparison between IGF and IGF+LEP treated quarters revealed significant differential expression (P<0.05) of 182 of 2323 genes measured (7.8%) (Supplemental Table 3). The IGF+LEP treatment significantly increased the expression of 108 and decreased 73 mRNAs relative to IGF-l treatment. Significantly altered genes could be broadly classified within a number of gene ontologies relevant to IGF-I- and Ieptin-related phenotype within the mammary gland (Table 6). These included receptor-mediated signal transduction genes (n=28), genes involved in cell proliferation and apoptosis (n=5), regulators of transcription (n=3), and cell structure and metabolism (n=2). GenMAPP pathway analysis (72:87) indicated overrepresentation of significantly altered genes within the insulin signaling pathway, including the Ieptin-induced upregulation of GSK3A, CAP1, FRAP1, GYS1, MAP3K11 (P=0.06), and SHIP2, and downregulation of P|3KA (#006), SRF, MAP2K4, and PFKL (Figure 12). The TGF-beta signaling pathway was also significantly overrepresented in IGF-l plus Ieptin versus IGF-l treatment comparison when a genes tending to be altered by treatment (P<0.1) were also included in the GenMAPP analysis, including upregulation of ENG, MADH3 (P<0.05), TNF, CTNNB1, MADH7, and MADH2, and downregulation of NFKB1 (P<0.05) (Figure 11). Q-RT-PCR confirms IGF-l induced expression changes in 4 of 21 genes significantly altered by microarray analysis. As shown in Figure 6, lGF-l-induced expression changes indicated by BOTL4 and BMET microarray 114 analysis were confirmed by Q-RT-PCR for four tested genes (Table 8). F K506 binding protein 12-rapamycin associated protein 1 (FRAP1) was confirmed as upregulated 56% with Q-RT-PCR analysis, closely mirroring its predicted change by BOTL4 microarray analysis. Although the BOTL4 microarray analyses predicted a 26% upregulation of the apoptosis-related cysteine peptidase caspase-8 (CASP8) by IGF-l treatment relative to SAL infusion, a 110% upregulation was observed with Q-RT-PCR. However, although microarray analysis predicted that this gene was additionally downregulated by IGF+LEP relative to IGF-I treatment, no significant differences were detected. 14-3-3- protein, eta isoform (YWHAH) similarly was confirmed for only one of two comparisons deemed significant by microarray analysis, showing a downregulation of 55% by IGF-I treatment relative to control. For a number of genes, a reversal of the direction of fold change was seen using Q-RT-PCR compared to that of microarrays. Although microarrays indicated an upregulation of NF-KB p65 by 17%, Q-RT—PCR showed a 47% increase in expression in the IGF+LEP treatment compared to IGF, but no significant difference in the SAL vs. IGF contrast. Similar disparities in comparing microarray and Q-RT—PCR results for genes found to be significantly altered by treatment for both platforms were seen with brain derived neurotrophic factor (BDNF) (upregulated 2.86 in IGF relative to SAL); mitogen activated protein kinase kinase kinase 1 (MAP3K1) (upregulated in IGF relative to SAL); NF-kappa B p65 (RELA) (upregulated 1.40 in IGF+LEP relative to IGF), and phosphatidylinositol glycan, class F (PIGF) (upregulated 1.52 in IGF relative to IGF+LEP). In three cases, a false-negative 115 gene was reported based on microarray analysis. While the BMET microarrays predicted a non-significant increase in SOCS3 mRNA of 1.13 in IGF+LEP relative to IGF (p=0.21), this gene was found to be upregulated 3.5-fold as measured by Q-RT-PCR. In addition, F RAP1, though found to be significantly upregulated by IGF-l relative to control for both the BOTL4 microarray and Q-RT-PCR, was not significantly upregulated (p=0.16) by 1.07 by the BMET microarray. YWHAH was also not predicted to be significantly different in the SAL vs. IGF comparison by the BMET array (0.98 fold, p=0.83) though this change was detected by the BOTL4 microarray and Q-RT—PCR. Expression changes for eight genes expected to be differentially expressed by BOTL4 microarrays were not significantly different by any treatment comparison when checked by Q-RT-PCR (OB-R, v-Jun, TLE1, IL8RB, SAP18, MAD2L1, FAK2, and BCAR1) (Table 2-4). Similarly, the BMET genes ALG12, CYCD2, IF NA2, MYC, PRKD2, PTCH, and SST were not found to be significantly changed in any treatment contrast. Q-RT-PCR analysis reveals potential IGF-I and Ieptin signaling pathway interaction. Seven genes believed to be potentially involved with IGF-I and leptin signaling in the mammary gland, signal transducer and activator of transcription 3 (STAT3), STAT5A, STAT5B, cytokine inducible SH2-containing protein (CISH), suppressor of cytokine signaling 1 (SOCS1), SOCS3 and 80085, were measured by Q-RT-PCR. No significant differences in any treatment contrasts were detected for any of these genes except SOCS3 Expression of 80083 was shown to be significantly downregulated in IGF and in 116 LEP quarters relative to IGF+LEP (Figure 6). Leptin upregulated SOCS3 mRNA expression 34-fold in IGF-treated quarters (P<0.05), and IGF-I increased Ieptin- SOCS3 expression 24-fold in Ieptin treated quarters (P<0.05). DISCUSSION Silva (2002) found that an elevated concentration of leptin in the mammary gland in vivo can decelerate IGF-l stimulated epithelial cell proliferation (278). The reduction in the number of cells in the S-phase of the cell cycle was consistent with our previous work showing that Ieptin inhibits proliferation of the bovine mammary epithelial cell line MAC-T (276). The current study is the first to provide a possible intracellular mechanism by which Ieptin might inhibit proliferation of bovine mammary epithelial cells at a supraphysiological dose. Studies examining the effects of Ieptin on human breast cell lines have shown stimulation of cell proliferation by leptin in three studies (85;145;173), but inhibition at elevated doses (17). It is possible that this is a species-specific difference between humans and cattle. A number of lines of evidence suggest that an interaction between the hormones IGF-I and leptin likely plays a role in the nutritional inhibition of mammogenesis in prepubertal ruminants. A definitive molecular mechanism linking high energy diets, which are known to increase circulating but not local mammary IGF-l concentrations, and mammary parenchymal development is still unknown (269;320;321). Paradoxically, increased serum IGF-l levels are correlated with high planes of nutrition but do not promote mammogenesis even 117 though IGF-l is a potent stimulator of mammary epithelial cells in culture (269). Proliferation of primary bovine mammary epithelial cells is maximized with ~25 ng of lGF-l/ml, a concentration much lower than that found in serum (231 ;319;320). Although the presence of IGF binding proteins in the extracellular milieu can reduce the amount of free IGF-I available to bind its receptor, the results shown in this and our previous study (275) indicate that the bovine mammary gland is responsive to an increase in IGF-l above the normal physiological concentration found in heifers fed for moderate growth. A similar effect was seen in a study using pregnant beef heifers in which the intramammary infusion of IGF-I for 10 d increased the mass of mammary parenchyma (59). In line with this local effect of IGF-I on mammogenesis, transgenic mice overexpressing mammary IGF-l showed stimulated development of mammary alveolar buds (319). Yet the fact that high energy intake in ruminants decreases mammary growth while increasing body growth suggests that factors other than IGF-l must be involved in the control of mammary epithelial cell proliferation. Our lab has recently proposed that Ieptin mediates the nutritional impairment of mammary development in ruminants based on a number of lines of evidence (28;225;276;277). Serum leptin concentrations are elevated in high energy diets (28) and decreased mammogenesis in prepubertal heifers is associated with increased body fat mass (277). In a study by McFadden and Cockrell (1993), the presence of adipose tissue in mammary explant cultures inhibited proliferation of bovine mammary epithelial cells (199). The signaling form of the Ieptin receptor, Ob-Rb, is expressed in the epithelial cells of bovine 118 mammary tissue (276). Leptin inhibits proliferation in vitro of the bovine mammary epithelial cell line MAC-T (276). And finally, Ob-Rb is a member of the interleukin-6 (IL-6) receptor family, and may share similar signaling capabilities with the lL-6 receptor in the mammary gland (19:225). Interleukin-6 has been shown to inhibit bovine mammary epithelial cell proliferation (225). In our study, Ieptin decreased the fraction of cells in the S-phase of the cell cycle to a greater extent in the lGF-l-treated quarters than in saline-treated quarters. This inhibition of cell proliferation agrees with our in vitro experiments in which Ieptin decreased DNA synthesis of MAC-T cells when they were stimulated with IGF-l or fetal bovine serum (FBS), but not when they were grown in serum free basal media (2762??) Leptin was likely able to reduce the percentage of cells in the S-phase of the cell cycle in saline-treated quarters in vivo by inhibiting the mitogenic effects of endogenous IGF-I or other growth factors present in the intact mammary gland. In culture, doses of leptin from 16 to 1,000 nglml decreased lGF-l- or FBS- stimulated DNA synthesis 10 to 30% (276). The dose of leptin infused in this experiment was likely supraphysiological. As reported by Silva (2002), leptin concentrations in the saline and IGF-l infused quarters were 6 ng Ieptin/g of tissue, while Ieptin-infused quarters reached 170 nglg of parenchymal tissue 14 h after the final Ieptin infusion (278). The mass of mammary parenchyma in a 200 kg heifer is ~100 g / quarter (unpublished data); therefore, control quarters contained ~600 ng of leptin and this was increased temporarily to 100 #9. or 1000 nglg, in Ieptin-treated quarters (Silva 2002 #5810). 119 Thus, the half-life of infused Ieptin was 5 to 6 h, and the Ieptin concentration would have been ~50 nglg of tissue at 24-h after infusion, which is at least 8-fold above normal during the entire 7-d treatment period (278). Mammary Ieptin concentrations elicited by either high energy intake or obesity are likely much lower, and whether these smaller increases in mammary Ieptin also would decrease proliferation of mammary epithelial cells remains to be determined. Although a number of studies in humans and mice have described a proliferative response of mammary gland to Ieptin expression, the physiological state of the tissue used may be significantly different from the bovine mammary tissue in our studies. Our conclusion that Ieptin inhibits bovine mammogenesis is based on intramammary infusions of leptin in physiologically normal heifers, whereas the reports showing that Ieptin stimulates mammogenesis in rodents are based on mice with genetic deficiencies of Ieptin or the Ieptin receptor. Three studies showed an increase in cell proliferation in human tumor cells in response to Ieptin added in the range of 50 ng to 16 ug in two different cell lines (85;173;226), and a fourth study reported no change in proliferation when 100 ng of leptin were added to HTB26 tumor cells (224). However, in normal mammary cell lines the effect of leptin on mammary epithelial cell proliferation is conflicting. Hu and colleagues (2002) reported an increase in mammary epithelial cell proliferation at when 100 ng of Ieptin was added to human HBL100 cells (145), but Baratta et al, (2003) showed that mammary epithelial cell growth was inhibited by high concentrations of leptin (1.5 and 15 uM) in the mouse HC11 cell line (17). This report is consistent with our laboratory’s study showing a decrease 120 In bm 25 ng rode blah and isu ha: leg in bovine mammary epithelial cell proliferation in the MAC-T cell in response to 25 ng of Ieptin (276:277). Another possible explanation for the apparent discrepancy of ruminant, rodent and human results is the Ieptin dose. It is possible that Ieptin exhibits a biphasic effect on mammary cell proliferation, with low doses of leptin stimulating, and high doses inhibiting mammogenesis. Such a biphasic effect of Ieptin has been reported previously, for granulosa cells (251) and human mammary epithelial cells (17). At such high leptin concentrations, it is possible that Ieptin is eliciting its action by binding to a cognate receptor within the interleukin-6 family of cytokines. The biological actions of leptin are elicited by the binding of leptin to the Ieptin receptor (Ob-Rb). Ob-Rb is the long isoform form of the Ieptin receptor which contains a transmembrane intracellular tail that is required for endocrine function and signaling (32;34;62;100;116:175;176;208;296;297;334). Alternative splicing of the primary Ob-R transcript can result in the production of multiple isoforms (Ob-Ra to -e) that contain a common extracellular domain but lack transmembrane signaling capacity (100;175;176). Only the activation of Ob-Rb leads to activation of STAT3 (26). The binding of leptin and IGF-l to their respective receptors induces a cascade of intracellular signaling events that regulate key cellular functions. Leptin is a member of the gp130 family of membrane bound receptors, a group that includes oncostatin M, IL-6, lL-11, LIF, and ciliary neurotrophic factor (146). These cytokines can signal through a common signaling subunit, glycoprotein 121 130, to produce a variety of responses, including growth arrest of mammary epithelial cells (146). Insulin-like growth factor-I is a key regulator of cellular proliferation, survival and differentiation in numerous cell types, and the activation of the IGF-l receptor (IGF-IR) is now known to generate signals within at least four distinct signal transduction pathways (124). These pathways are not necessarily specific to IGF-l, but can be activated by a number of growth factor receptors to allow for significant signaling complexity (124). Increasingly, studies are revealing that insulin and leptin signal transduction pathways show significant amounts of crosstalk across multiple tissue types (21;46-48;168;220;221). Intracellular signaling for insulin and IGF are similar, and signaling interactions between IGF-I and Ieptin in the developing mammary gland seem likely. Our analysis of cell cycle regulation by leptin and IGF-l indicate that Ieptin may inhibit mammary epithelial cell proliferation by altering cell cycle kinetics. lmmunohistochemical labeling of the proliferating cell nuclear antigen Ki-67 indicated that the number of cells committed to the cell cycle was unaffected by treatment. The percentage of Ki-67 labeled cells ranged from 8.5-11.7%, which strongly agrees with work by Capuco et al. (42) examining cycling in prepubertal mammary epithelial cells. Less than 0.25% of mammary epithelial cells across all treatments in our study were apoptotic as measured by caspase-3 labeling. This value is in line with the fraction of cells found to be apoptotic in the prepubertal mammary glands of heifers as measured by terminal deoxynucleotidyltransferase-mediated dUTP nick-end labeling (TUNEL) (212). A smaller fraction of cells in the Ieptin-treated quarters occupied the S-phase of the 122 cell cycle, indicated by the relative labeling between the percentage of cells labeled by Ki-67 and BrdU. A number of studies have shown that IGF-l stimulates cyclin-dependent kinases and cyclins during the late G1 phase to advance cells into the S-phase of the cell cycle and accelerate proliferation (115227). In addition, although the increase in BrdU labeling by IGF-l was not significant in this study, we have previously shown that IGF-I intramammary infusion does increase the fraction of cells in the S-phase of the cell cycle in a study designed to specifically test the IGF vs. SAL comparison with higher statistical power (275). The results presented in this study extend this information and indicate that Ieptin may inhibit the recruitment of mammary epithelial cells into the S-phase of the cell cycle and leave them stranded in the G1-phase, thereby slowing proliferation. Such an effect of leptin on decreased cell-cycling could be mediated through the inhibition of the STAT3 and ERK signaling pathways, known to control proliferation in mammary epithelial cells (205;233;272). Although no studies have shown a direct inhibitory effect of Ieptin on cell cycle progression, it is possible that the interference with IGF-l-stimulated proliferation could be mediated through a member of the suppressor of cytokine signaling (SOCS) family. Links between the interleukin-6 class I cytokine receptors, including Ieptin, as well as the IGF family, and the SOCS families have been well documented (19). Inhibition of the JAK/STAT signaling pathway by the SOCS proteins can arise from several distinct molecular interactions. SOCS proteins are able to bind directly to the SH2 domain of the tyrosine phosphorylated cytokine and IGF-I 123 receptors, or bind directly to the janus kinase (JAK) tyrosine kinases, which also interact with the lL-6 and IGF receptor families (84;96:213;218;310:329). Activation of these cytokines can lead to the activation of STAT3. Upon its phosphorylation and activation, STAT3 translocates to the nucleus and mediates gene transcription, including that of 80083, which is able to attenuate Ieptin receptor signaling in a feedback manner by interfering with further JAK2 activation (16:19;25;26:44:179:289:322;339). SOCS genes vary in expression level across tissues, particularly SOCS1 and SOCS3. This expression is often not constitutive and can be rapidly regulated by various cytokines as well as GH, prolactin and leptin (315). The promoters of many SOCS genes contain STAT- responsive DNA elements, consistent with their expression dependence on the JAK/STAT signaling pathway (213:290). However, JAK/STAT-independent regulation of SOCS expression has also been described for SOCS3 in neutrophils stimulated by lL-10 (49). The SOCS3 protein has been shown to bind . to JAKs, but with lower affinity than SOCS1 and with a weaker inhibition of catalytic activity (127). SOCS3 may, however, be more effective at interacting with activated cytokine receptors. Hansen et al., (1999) showed that SOCS3 associates with the tyrosine-activated GH receptor, thereby inhibiting GH- activated STAT5 signaling. Here the mode of SOCS action is not by prevention of the docking of STAT5 to its JAK, but the result is still the inhibition of JAK2 activation (127). In this study, it was shown that SOCS3 mRNA is upregulated 34-fold in quarters infused with IGF-l plus Ieptin relative to IGF-l alone. A Ieptin-induced 124 increase in SOCS3 protein level when the IGF-l receptor is activated, assuming protein changes mirror those of mRNA, could possibly explain the inhibition of proliferation in mammary epithelial cells (Figure 7). A SOCS-mediated feedback mechanism may exist that works to decrease mammary epithelial cell sensitivity to the IGF-I mitogenic effect and thereby inhibits proliferation during periods of high energy feeding or increased adiposity in the ruminant. Decreased signaling through the IGF-l receptor would reduce progression to the S-phase of the cell cycle and slow proliferation. Future experiments measuring SOCSS protein levels and activation of critical components of the IGF-I and leptin signaling pathways, including STAT3, insulin receptor substrate 1 (IRS1), STAT5A and B, and JAK2 will be necessary to elucidate such a mechanism. Microarray gene expression study The experimental design put forth in this study is well-suited for studying the individual and counter-regulatory effects of the hormones IGF-I and leptin on mammary parenchyma gene expression. The within-animal comparisons employed by infusion of different hormonal treatments to adjacent quarters within the same animal contribute to increased power in testing the interactions of these hormones, as animal variability is removed. Alteration of expression profiles of genes regulating proliferation and apoptosis, intracellular signaling, and cell morphogenesis and metabolism by IGF-I and leptin were of particular interest in this study given the BrdU incorporation results. Microarray results for each treatment comparison revealed 125 numerous candidates for mediation of the cell proliferation response by IGF-I and its abolishment by Ieptin. However, many of these changes in gene expression were invalidated by Q-RT-PCR. To date, real-time PCR has been performed on 21 genes of interest deemed significant by analysis with the BOTL4 and BMET microarrays (Table 8). Of these 21 genes, four were found to be in agreement with microarray results with regard to up- and down-regulation direction within the appropriate treatment contrast. An additional four genes found to be significant by microarray analysis were also found to be significant by Q-RT-PCR but were either reversed in their fold-change directionality or were significant for a different treatment contrast (Figure 6). At this time, it seems that the use of the BOTL4 and BMET microarrays were not well-suited for detecting differences in relative gene expression in the tissues of study. The FK506 binding protein 12-rapamycin associated protein (FRAP1) gene was upregulated 56% in IGF-I treated tissues relative to saline infused control. Similarly, caspase-8 mRNA was increased by IGF-l by over 2-fold. The 14-3-3 protein eta isoform (YWHAH) gene was significantly downregulated in IGF-l treated tissues relative to saline control. Although BOTL4 microarray analysis predicted very small fold changes due to the treatments (all <1 .5-fold at P<0.05), Q-RT—PCR showed greater relative differences for each of these genes. This fold-change magnification is likely correlated with the inherent muted dynamic range of microarrays in measuring relative expression changes. The F RAP1 gene encodes the protein belonging to a family of phosphatidylinositol kinase-related kinases. It is also commonly referred to as 126 mTOR (mammalian target of rapamycin), and has numerous links to the IGF-l signaling pathways Pl3K/akt, raf/ERK, and MEK/MAPK in various cell types including mammary epithelial cells (71). Studies in neuroblastoma cells have shown that maintenance of lGF-I-induced proliferation requires mTOR function along the MAPK pathway (204). The well established role for F RAP1 in IGF-l signaling supports a possible role in regulation of mammary epithelial cell proliferation as observed in this infusion study. The 14-3-3 genes code for a family of proteins that have previously been shown to bind to and regulate other proteins, including the IGF-IR, to modulate neurodevelopment, cell-division, signal transduction and gene transcription (Wong, Likhodi, et al. 2005 #5720). These proteins are phosphoserine (pSer)- binding proteins that have been shown to interact directly with oncogenic products and components of mitogenic and apoptotic signaling pathways (114). The 14-3-3 eta isoform has most commonly been linked with brain functions, and one study showed that it was downregulated by IGF-I in Purkinje cells (335). The downregulation of 14—3—3 eta observed in this experiment may contribute to the increased mammary epithelial cell proliferation observed by interacting with cyclins and cyclin dependent kinases to modify cell cycling. Caspase-8 is known to be involved with antiproliferative and apoptotic signaling in mammary epithelial cells (252). However, although a significant 2- fold increase in caspase-8 mRNA expression was observed in IGF-l tissues relative to control, no only minor differences in apoptosis were observed in our tissues based on caspase-3 immunostaining (Figure 4). It seems unlikely that 127 caspase-8 regulation provides a proliferative stimulus in the mammary gland, and this change in gene expression can not be explained at the moment. Quantitative-RT-PCR on microarray genes of interest revealed significant differences in expression of mitogen activated protein kinase kinase kinase 1 (MAP3K1), brain derived neurotrophic factor (BDNF), NF-kappa B p65 (RELA), and phosphatidylinositol glycan, class F (PIGF) in this mammary gland infusion model. Each of these genes was significantly altered in at least one treatment contrast not predicted by either BOTL4 or BMET microarray analysis. MAP3K1 is a mitogen activated protein kinase (MAPK) that is activated in normal mouse mammary and human breast cancer cells in response to prolactin (73). Our results seem to follow with this study and the upregulation of MAP3K1 in IGF-l quarters relative to control could indicate amplified signal transduction through the MAPK pathway to enhance mitogenesis in mammary epithelial cells. BDNF is a member of the nerve growth factor family and is necessary for survival of neurons in the brain. This growth factor has been shown to interact with IGF-l in the brain to mediate neuronal plasticity by amplifying IGF-l signaling through the extracellular signal-regulated kinase (ERK) pathway (152). Although BDNF mRNA has not been previously found in the mammary tissue, its upregulation in IGF-l treated quarters relative to control could indicate some mitogenic function for this growth factor in the gland. NF-KB activity in normal mammogenesis in humans and mice has been previously reported in numerous studies (36:55:160). These studies implicate this transcription factor in a role in cell survival, protecting cells that have 128 received an apoptotic stimulus. Immunological functions of high NF-KB levels have been reported in bovine mammary glands infected with mastitis (35). A study by Lappas et al., (2005) showed that the NF-KB transcription pathway is important to regulation of lL-6 in adipose tissue (170). The upregulation of NF-KB p65 mRNA in IGF-l plus Ieptin treated quarters relative to IGF-l could possibly indicate a role for this gene in the negative regulation of leptin secretion. PIGF is a protein involved in glycosylphosphatidylinositol (GPl)-anchor biosynthesis and modification (273). Blood cells use the GPI anchor to attach proteins to the cell surface. Currently, no literature exists linking PIGF to either IGF-l, Ieptin or the mammary gland, and the biological relevance of changes in PIGF gene expression in our study cannot be explained at this time. The inconsistency between microarray and Q-RT—PCR results revealed in this study are somewhat puzzling. Two major reasons behind the weak correlation between the two techniques are apparent. First of all, the experimental design used in this study to measure changes in gene expression in response to the hormones IGF-I and leptin could also be improved for the application of a functional genomics approach. Samples were harvested at day 7 of this infusion trial; this time point may be unsuitable for functional genomics analysis, as mRNA expression changes could have occurred early in response to the hormonal infusions and reached a lower, steady state by day 7. Tissue harvested at different time points, especially interspersed within the first 24 hr of infusion may likely be better suited for mRNA analysis. 129 Secondly, it is possible that the changes in gene expression within the mammary parenchyma in our study are quite small and difficult to detect using the microarrays because of the inherent limitations of sensitivity of the platform. Prepubertal bovine mammary parenchymal tissue consists of approximately 10- 20% epithelial cells, 40—50% stromal tissue and 30—40% fat cells at the initiation of puberty (265:266). Although approximately 30% of the cells in our parenchymal tissue appear to be epithelial (unpublished observations), Ki-67- labeling shows that only 8-12% of these epithelial cells are proliferating. It is therefore possible that the gene expression profile of our target cell group is highly diluted by stromal and fat cell mRNA. Following this logic, the large fold change revealed for SOCS3 may indicate that it is regulated in stromal cells rather than epithelial cells. However, the fact that a 3.4-fold change was detected for SOCS3 gene mRNA that was undetected by microarray analysis is puzzling. Microarray sensitivity or analysis must surely be improved here to reveal such obvious differences in expression. The use of false discovery rate (FDR) analysis may help to explain the failure of validation of microarray results. Although FDR analysis was not applied to the BOTL4 microarrays, it was used to determine the probability of validation for genes measured in the BMET microarray. In the IGF-I vs. IGF+LEP treatment comparison only 15 genes were found to have an FDR value below P=0.3. This indicates a 30% probability of validation for each of these genes. Expression of the most highly ranked gene (MYC) and 15th gene (IF NA2) were measured by Q-RT—PCR and validation failed. Further Q-RT-PCR analysis of the remaining 13 genes could lead to validation, though all predicted 130 fold-changes for these genes were less than 16% up- or down-regulated. Similarly, only three genes in the control versus IGF-l microarray comparison had FDR values below P=0.41. The problem of false-negative generation by microarray analysis was also revealed in this study. Based on a comparison of the two microarray platforms used with Q-RT-PCR, three genes were found to be false negatives in BMET microarray analysis. Though the BMET microarrays predicted a non-significant increase in SOCS3 mRNA of 13% in IGF+LEP relative to IGF (P=0.21), this gene was found to be significantly upregulated by Q-RT-PCR measurements. FRAP1 was found to be significantly upregulated by IGF relative to SAL for both the BOTL4 microarray and Q-RT—PCR, but was insignificant in BMET analysis. Similarly, YWHAH was also not detected as differentially expressed in the SAL vs. IGF comparison by the BMET array (0.98 fold, p=0.83), but a significant BOTL4 microarray change was confirmed by Q-RT-PCR. Therefore, the reasons behind the inversion of fold change by real-time PCR relative to microarray predicted changes, differences in significant contrasts, and false negative prediction are hereto unclear and will require further analysis of the microarray platforms used in this study. In conclusion, increased concentrations of IGF-I promote the proliferation of mammary epithelial cells by increasing the progression of cells into the S- phase of the cell cycle in vivo in mammary epithelial cells in prepubertal heifers. Conversely, increasing the Ieptin concentration above normal reduces proliferation by limiting the progression of cells into the S-phase of the cell cycle 131 and stalls them within the G1 phase of the cell cycle. These effects on cell proliferation are likely regulated by changes in gene expression induced by IGF-I and leptin within the mammary parenchymal tissue. The SOCS3 could potentially be a primary Ieptin-induced inhibitor of IGF-I signaling and cell proliferation to mediate this effect. Overall, this study has demonstrated that IGF-I and leptin intramammary infusions in prepubertal heifers can serve as a useful model for the study of hormonal regulation of cell cycle progression and gene transcription in the developing gland. 132 FIGURES A) Figure 1A. BrdU labeling. 133 Figure 1B. KI-67 labeling. 134 C) Figure 1C. Caspase-3 labeling. Figure 1. Representative mammary parenchymal sections immunostained for bromodeoxyuridine (BrdU), proliferating cell nuclear antigen (Ki-67) and caspase- Panel A) B = BrdU-positive cells (darker), E=epithelial cells, S=stromal cells, and A=adipocytes. BrdU-labeling was calculated by dividing the number of BrdU- positive epithelial cells by the total number of epithelial cells (278). Panel B) K = Ki-67—positive cells (darker), E = epithelial cells, S = stromal cells, and A = adipocytes. Ki-67-Iabeling percentage equals the number of Ki-67- positive epithelial cells divided by the total number of epithelial cells. Panel C) C = Caspase-3-positive cells (darker), E = epithelial cells, S = stromal cells, and A = adipocytes. Caspase-3-labeling percentage equals the number of caspase-3-positive epithelial cells divided by the total number of epithelial cells. 135 109 9-1 IGF-I + Lenin Latin Figure 2. Effects of intramammary infusion of ovine Ieptin and IGF-l on bromodeoxyuridine- (BrdU) labeling of mammary epithelial cells in prepubertal heifers. Individual mammary glands of six heifers were infused for 7 d with either saline or IGF-l (10 pg/d). Contralateral quarters were infused with either saline+leptin (100 pg Ieptin/d) or IGF-l + Ieptin (100 pg Ieptin/d). BrdU was infused intravenously about 2 to 3 h before slaughter. Leptin decreased (P<0.02) mammary epithelial cell proliferation in both saline- and lGF-I-treated quarters. Results are back-transfonned. Estimates of variance were calculated as confidence intervals from the transformed data for comparisons of leptin effect within IGF-l treatments (Data taken from Silva (278)). 136 Ad—h-I ON&°, L % labeled epithelial cells ONIhGQ Oatrd IGF-l IGF-1+Lqfin Latin Figure 3. Effects of intramammary infusion of ovine leptin and IGF-l on proliferating cell nuclear antigen (Ki-67) labeling in prepubertal heifers. The number of cells committed to the cell cycle was measured by the percentage of proliferating cell nuclear antigen (Ki-67) labeling. No significant differences in the number of cycling cells were found among the four treatments. 137 % labeled gpithelial gells S 'a.‘ 8 hi 8 ,O ‘3‘ O 1 IGF-I + Latin Latin Figure 4. Caspase-3 immunostaining indicates apoptosis in the mammary gland. lmmunohistochemical staining of caspase-3 was performed on sections taken from all treated quarters. Two contrasts were significantly different (P<0.05): LEP vs. SAL, IGF vs. LEP and one contrast showed a tendency toward a difference (P<0.07): IGF-LEP vs. SAL. Although statistical differences exist, there is very little evidence of apoptosis plays a large role in shaping the mammary gland, as numerically all caspase-3 percentages across treatments were very small (<0.25%). 138 165 r 14- 12« I 10* IKi—67 new ICm-a % labeled epithelial cells Oortrol IGF-l IGF-I +Leptin Leptin Figure 5. Effects of ovine leptin and IGF-l on the percentage of cells committed to the S-phase of the cell cycle. Leptin treatment significantly reduced the number of proliferating cells occupying the S-phase of the cell cycle. The fraction of proliferating cells within the S-phase of the cell cycle was determined by the dividing the percentage of BrdU-labeled cells by the percentage of Ki-67-labeled cells. The main effect of leptin ((LEP + IGF+LEP) — (SAL + IGF)) (P=0.015) and the IGF versus IGF+LEP contrast (P=0.02) were both significant. 139 A) 1.5 . 0 5 1. 1 .935 go". 10 o B) ‘1 0 5 0. .6 0 0 1 7. 8:26 so“. 0.0< C) 5 2 0 Z 5 0 8:28 2o". 5 Q 0 a 140 20. D) s. o. .m a 0.0. 20- .. 1 m 5 09.2.0 v.0"— 0. 1 1 E) 4 o 6. 4. .. O B O _ . "v l 2 L 1 2 . 0°C“: U—O Z 852.0 3o". F) 141 G) H5 Fold Change H) 2.5- 20- 1.0 ~ 0.5 _ i 0.0 ‘ I r r C I I. L Figure 6. Genes found to be significantly altered by treatment in mammary tissue from six prepubertal bovine heifers. Data were derived from Q-RT-PCR assay of RNA from six cows and expressed as ratios of expression relative to SAL infused control. Total RNA isolated from treated quarters was converted to first-strand cDNA and subjected to Q-RT-PCR as described previously and in Materials and Methods. Data presented are the mean +/- standard error of the mean relative expression levels of each gene within group. calculated as 2““, as described, with GAPDH as a control gene and the mean ACt value for each gene in the control samples as the calibrator. Fold Fold Change a: 142 A) IGF-l treatment significantly decreased the expression of 14-3-3 protein, eta isoform, 55% relative to SAL control. B) Brain derived neurotophic factor (BDNF) was significantly upregulated 2.86- fold by in IGF-l treated quarters and 3.25-fold by Ieptin relative to SAL control. C) Caspase-8 mRNA was significantly upregulated 2-fold in IGF relative to SAL quarters. D) IGF-I treatment significantly increased the expression of FRAP1 56% relative to SAL control. E) MAP3K1 mRNA expression was upregulated 46% by IGF-l treatment relative to SAL. F) NF-kB p65 (RELA) expression was upregulated 34% in IGF-LEP treated quarters relative to IGF, and 47% relative to SAL. G) For the phosphatidyl inositol glycan, type F (PIGF) gene IGF quarters were significantly different than both SAL (52%) and IGF-LEP (46%). LEP tended to be different from SAL (P<0.08) (37%). Heifer 88 was an outlier for this gene and was deleted from analysis. H) Suppressor of cytokine signaling-3 (SOCS3) mRNA was significantly upregulated 3.4-fold in IGF-LEP quarters relative to IGF. 143 IGF-l Leptin Receptor Receptor DOOOOOOOOOOO.......OOOOOOOOU ......OOOOOOO......OOOOOOOOI P|P2 ~‘1‘PIP3 PDK1 JAK2 I... I=.Pllfci _RS -_AK2 ' _li..' g Figure 7. A schematic of IGF-I and Ieptin intracellular signal transduction. Binding of IGF-I to its receptor induces intrinsic tyrosine kinase activation and receptor autophosphorylation. The insulin receptor substrate (IRS) molecules are then recruited to the receptor, are activated by tyrosine phosphorylation, and are then able to activate downstream targets such as the regulatory subunit (p85) of phosphatidylinositol 3-kinase (PI3K), which also contains a catalytic (p110) subunit. Phosphatidylinositol triphosphate is activated by the p110 subunit of PI3K to signal downstream to the phosphor-inositide dependent kinase (PDK1), protein kinase B (PKB/Akt) and protein kinase C (PKC). Leptin binding to its receptor induces activation of the janus kinase (JAK) proteins, dimerization of its receptor, and JAK-mediated phosphorylation of the intracellular portion of the Ieptin receptor. JAK2 activation leads to recruitment of the signal transducers and activators of transcription-3 (STAT3) protein, which binds to tyrosine phosphorylated SH2 proteins on the receptor and is activated by phosphorylation. Phosphorylated STAT3 translocates to the nucleus and induces the transcription of target genes such as the suppressor of cytokine signaling-3 (SOCSS) protein. SOCS3 has the ability to interfere with signaling along both of these pathways. It binds directly to the SH2 docking sites on the IGF-l receptor (83) as well as IRS substrates (86). SOCS3 can directly mediate feedback inhibition of the Ieptin receptor by preventing the docking of STAT3 (25) and STAT5 genes (135:302), or by directly interfering with JAK2 activation (330). Thus, I hypothesize that the effects of IGF-I and Ieptin in the prepubertal mammary gland may result from crosstalk mediated by this common signaling pathway and SOCS3 (Adapted from Niswender and Schwartz (221)). 144 TGF Beta Signaling Pathway llgnnds: ITGBS G“ W a TGFB' 01;" _ - NI~~_M11.” . ELM-1 MN mm. m +—- a on" may“ m._g_g_u E7/ l— E Mi,*-W 0.0m L and Tm 1!th coder-c1 W I I No cm ‘ Nd found a . I o m r— m lano- , R ”T “I.“ epmsou o 5541 , -m m @ ~ MmIE Co-amads El" ‘5“ ‘ “’t’ Comm, 1 0 me u ‘ l l : .34: I“ " 3:“ I —@ ozm—p ou , _— Audion Nunt Gal ‘ a" - 0.13:7 E-md‘: Nunligpostlau acil SKIL . Lasurrodflod: 311/02 5 °-‘“ edited by Naman Salomoms Figure 8. Genes differentially expressed in the SAL vs. IGF treatment comparison are significantly overrepresented in the TGF-beta signaling pathway. GenMAPP analysis indicated that seven genes significant at P<0.1 were altered in the SAL vs. IGF comparison within the TGF-beta signaling pathway. Gene significance is indicated by the number next to the gene on the pathway map. Unaffected genes are indicated in light gray and genes absent from the BMET microarray are indicated by white boxes. 145 Author. Diabetes Genome f-nalom. PleOd Imestlgalors Human 'flSUlin Signaling Gone 0mg”. umoyumdm tee and ueagan learam mammal": 9a Emu:lold,n,eo@1':5Iln namrd ecu E’EE‘P" Datuat “3‘ W“: “3’40”: limo. Tandem, CvlanatyleLL“ Color Is: Mattresses on on. ooooooooooooooo Myu“. ""P .. 5W" “.331... 5W” It!!! ................ I'nwl'" ROCOP‘flIA‘i'I'G“ R'COPW19.5.2.1!12:22:12.3: Glucose Transport 1 r Insulin Receptor Substrates PI Minus RasJMPK Signaling PIK3CA 0qu mm um :22; "m 2:3 "m GR82 om ms um PIK3C8 PIK3R2 IR53 sncs \ GRBIO Rm com PIK3CD P111383 "154 G88“ PIK3CG PlKJRl ”9 °°" PIK3C2A GA81 was $051 RACI W" PIK3C26 05m 5052 ems RACZ PIKJCZI \ HEW? KIM!“ PDKIAkt Signaling ' MAPK1 oIosI MP3K1 PDK1 o “21 SGK Tor Signaling HAPK) MAPJKZ AKTI osm SGK2 \ NARI“ om HAWK) 013s: AKT2 SGKL IMPKB can “PSI“ oms 55"“ “m _I_ IIAPKI ems sums 55"” ”"2 IIAPKI IIAPJKG J / \ ““53 W“ IMPKS out: MAPJKI 1 IIAPK10 oIIm IMP3KB sass Vesicular Trafficking Transcriptional Regulators & Immediate Early Genes 0 ms MAPKII MAPJKS non CAP roxou mu / \ “m” °"“ mpmo non cm osass roxom WK" ”“3 “”3"" “m STXBPI ch mu 0 ms p70 56 Klnasss Translation Regulation “PK“ ‘ ms :MAP3K121sm2 5mm CBLB 03m [RpssKaI o.ssIElF4EBP1IPIlASl 0150! “m“ “m ””3“” XBLC _ 2; com Mom I. EIFIE ”PM W" "”3"“ STX8P3 can “m ros W, IAPZKII om: mam arm STXBP‘ cm W" JUN “m MAPZKI osss HAP4K2 VAMPZ TC 10 IAPZKS oom HANK} ST)“ MAPIKG 05m MAPJKA Mon . SNAP23 MAPZKI HAWKS SNAPZS KIFM HAWKS KIFSB * ARFI 0.6m NYOIC . MetabolIc 9,901.50” Nodulators of Insulin Action ARF6 PFKII o u” p90 Rslt Klnasss LAR I an PKC alpha oossr 45160 PFKL o m was RHOJ ' PTPIII PKC baIa osm UPEI'HSL 0614! PCT 01211 PKC delta 0‘“ RA“ GYSI ours I om: APS PTEN PKC III 0.6046 EH01 cvsz SHIPZ otzos PKC Iota um P" sum PKC mm EH02 cvc suxzs INPP‘A PKC ma om: SOC 51 0.14! AMPK A1 SOCS3 0.113s AHPK AZ IKK 8 a mi XBP‘I ossst Figure 9: Genes differentially expressed in the SAL vs. IGF treatment comparison are significantly overrepresented in the insulin signaling pathway. Nine genes within the insulin signaling pathway were different in this treatment comparison at P<0.1 according to GenMAPP analysis. Upregulated genes are indicated in red, and downregulated genes are indicated in blue. Unaffected genes are light gray on this map. White boxes indicate genes not present on the BMET microarray. 146 39.8.0? ..mEm 05 :o «:39: Co: 353 98:35 35:3 SE; :5... .EoEfiob E uouootmcs mocom 9m $533 .35 En: .:o=m_=mo.a: 8286:. no. ucm ..__m:m md<_2:ow .3 Cdvn: .__o>:_ 856 52.5 .3353 6:50 20.6 .60 w 9 PO 9: :_ noucowo.no:o>o 3:592:05 9m :omtmano Eons—we.“ “.9 .m> .___m::o.ot_u 3:00 .3 2:9". a E tag :2. o a a ...—H..- E a I II g g 32.. 3.83... no 38 «Ngu — r [-58— E a a 888:. ism a 8.... .381: $5 is .6 1...: E 9”. 38g E E _. . . . . _ a: _. I555 E ¢8~.o.=l.zo.n “.0. 05 :_ :08th 3:3 9: :_ 823059.90 9“ >955“... m:__m:m_w Soniwh m5 .0 £359: 9: 85 £85. m_m>_m:m n_n_ “.9 $5 :_ 83990 252925. 8:...5 .3 952". a accrue—aw 5:32 .3 3:3 , main ”bog: 33 ... .51 % _ _ ...ua.:m3moa@at:z ”2w .....EE .2 L _ ......zé. .. £30800 £93950”. 4 28.... ) ......iflaa @ I. E \ Alisa Big 8.3 gulf?! a 836E L a. §§< Iii... E .la.a§ooeo~..la.u: E lawn}. E 3230 E “ODE-O: >6350n. 956.55 500 ".0... 148 a: 55:593.: .555 9: :o «:39: .0: «:5: 98:5 35:3 933 :5 59:59. .3 :95th 9.5: :5 35:3 55 En: .:o_5_:mw.a: 352:5 :9 :5 .....Q 9 359:8 amid? :_ 55.30955: 149 ...:ku Sign an. ......an a ...-mm- and... IE. Ini a IEHHMH- a a E E ...-Hm- Inl HH- a I'd—“MN“ 1345 a a I33... 1:5 HH- i.nmmm alm- §.§ .....nHu EINI ...: Emma .numfi HH- ..... Hun-unn- i.fiuu .....nnmu azlfil .55 IE 38:21:83 .8313 Ella: so}; al.-NH I a Ham a E a Eidfiuu . E HRH—sun ....IHI E Sing ”a —HHNH_ - a! an..a E Ehuamflflmfl i.§_=— ;E. .Hm- Imurmm 319233.523 i535!!!» ...-...: 3.!HI . _ 3E E .....BI; 3.36 / \ :8 IE 0 at. . .— > 5 $3 - 1 y 58.8.2... .555 m5 :0 239: .0: «0:3 032:... 35:3 223 :cm .2353: 3 83:2: 8.59 2: mega» >55 Em... .cozmfimoaa 33:5 :9 :5 £0. 2 :Qmano $3-5. 5 5338.58 8:865 SE .358: 82.358 2258.. .25 c. :92: 2285ch em 3353 mc__mcm_m 5.35 :5 553, 95.3 «meow .0 598:: a 332:5 «529.: c.2250 5:353 95:ch 5.35 9: £5? :mucommamtgo 228553 2: comtmano “cm—Emma nM.—£0. .m> “.0. m5 E 83998 bfizcmfiE: $50 .9. 2:2“. 151 TABLES Table 1. Ki-67 and BrdU labeling comparison Ki-67 Labeling Treatment Labeling °/o S.E. Pr > ltl SAL 8.7336 2.9049 * IGF 8.4869 2.9049 * IGF+LEP 11.6778 2.9049 * LEP 9.7680 2.9049 * BrdU Labeling (From Silva, 2002) Treatment Labeling. % §._§. Pr > it] SAL 6.8915 0.9737 * IGF 9.1526 0.9737 * IGF+LEP 4.9540 0.9737 * LEP 5.3680 0.9737 * Ki-67 and BrdU Difference Treatment Jifference S_._§. Pr > In SAL 1.8421 2.7042 0.51 lGF -0.6658 2.7042 0.81 IGF+LEP 6.7238 2.7042 ** LEP 4.3999 2.7042 0.12 * = Treatment labeling percentage is different than 0 (P<0.01) ** = Difference between Ki-67 and BrdU is significant (P<0.05) 152 0000.300”. 00.300000... 0000.300”. 0.00.02? 0.00.00.00-02 .900“. 000000000; 52:80 0888......» 00000 00.00.0000. 0000000000: 02.0.0590 ..00 .0 0000.300”. 02.0002 02.030000... .0090 02.0.0590 :00 22.0 :00 .o 8.8380 00000 0.000000300080520 =00 . 0000.00.90 =00 05.0005 0.0.0.90 0350.. 05.0090 0.0.0.0 5.0000”. 00.0300 0.0.05-0 0.003004. .0 0000300. 0350... 05.0005 0.0.0.0 .20000m 00.0300 0.0.0.90 9208.0 .2880 9...-..3 b.>_.0< 0000? 00.000.055.00 0.0.0.0 0000:0000; .0005 T:0.:0..0ES.0._. 88:88: 08.0 0.2“. >0350n. 05.0090 0.0.0.0 5.0000”. 00.0300 0.0.0.80 c8885» 2.0.0 2.0.0.. .288... 830 22:80 02.082 8.2.80 0.23.0.2... 0000:0000... .0090 0-00.? 0000.303. 000000000... 00002.00 030.00 x032 5.2.00 0000... 00.00.... 0.0.0.0 0005. 0.0.0.0 00.000.509.000 00000 0000:0000... .0090 030.005.002.000”. 0030000 5.0 5.0 000.0 50.0 5.0 5.0 «0.0 00.0 5.0 5.0 5.0 5.0 00.0 00.0 No.0 5.0 5.0 5.0 5.0 5.0 50.0 v1 2.. 3 .2 .28. 02052.... 2828: .08. 3 0000000. 9.0.2 00008.0 00.0.0.6 000.0004. a: .500. 2258 08 .28. 8.8.882 020.2 8.. .080. m .8828 02.728852 :0 .85. m .852: 2.0.2 05 a0: .3200. F 02.2280 8828.25 .880 .820 0: £222. 3.0. 00: .023; m 2...» .2880 0.223222 2.. .0220. 022.258 0.2.8 8.22;... 2.20 522.. 0 0: .9200. at 82.0. 2.08.5880 0. a 8.2.8 2.8 .2000. .288. 2.2052900 ta ..-... 72232.2... :0 . 80200. =. 2280 m: .22 .30.. .55.... 222.. 8.2880 28582-09... a: 5.00. .. .288. .02.. 0.0. 8.88:0 003 .000. .288. 5.8.. 8.. .3030. v 80.2.82“. 0.20.88 .08 «3 2.8 8.22.32 0 0320 Re .92.... N 8.2.... 8.828 .08.. N: .20... _ 82.2 222. 8828.3. 0000.5 0.0“. 00.00 0000 ..ob0oo 8 0.50.0. 16.. 00000 080.0005: 0.03000 0.000.020 Show >0 030800 «00.5003 .30 0:0..0> :0. >0 020:0 00000 a 28» 153 20000000005 028...... .80 880 0.2.3.82 08 2.8.0.0 :00 00:000000... .5390 ..00 .0 02.05000 00.009000 02.000009: 00000 00.00.0000 0000000009... 8.8.380 22.0 .80 0000000 00:0000 0.2.0090. .00..< 0.020 ..00 .0. 00:05.0 0.3.0090. 00000 0.00.000< 000 0000.00.20 =00 ..._>:0< 00000. 000090000000 0.0.90 0.2.2.. 89.0. 22.82.. 00:0:0009... .000.0 00:99:90 ..00 00:0:0009: .000.0 0500.0 000000000000 00:0:0009: .0000 5.00000 000030 ..00 00000 0000:0009: .000.0 030.000.00.030”. 0030030 00.0 00.0 No.0 50.0 00.0 5.0 No.0 00.0 00.0 5.0 00.0 5.0 v1 ... 280 00040-0 .5 8.02.9... a .0 2.2.... 00090-0 .3 8.8._m> . 20.0000 3 0300.0. .00.. 00000 080.000.0300 0.2.2.200 802.00 0.2.38.2 ..8 888: 0.0... 90:30.00 0000:00 0.2.8200 88.60 90:30.05. 000 000009... 0.0... 90:30.05. 00003.0 00000 0.020030! 000 0.30.50 ..00 00.0 5.0 00.0 00.0 No.0 5.0 00... .zoo. .880 0.... ....808. 2.82.0 5.0 30-00.900.200 <20... .9080. 00.900.090.000. 08.0 80.20. 2 02.02.28 8.280890 00.0 000.0 ..0.0000. 000000.. 0.00.0: 00.0 2.09.. 709 .0 8000.003 00000. 8.0 2.20.00. 5 .230... 8.2.... 080802008 ..8... 00.00.22. 3.0 0.22.. 982.888.... 0.... .92. m 05:80. 8.2.... 22.83. 8.0 00.0.00 .3... 0.212;. 0.2.0 8 0.220 0.0.! 00.0 .90000. .0.0>:00 0000000090 003000029... 000000 0.0.... 2000 0000 3.. 8.2.2 3.. .0000. 82208.. 0.0.. 2.8820 0: 8.2.2 0: 80.2.2. 0 89.290208... ......22 00... 00930000 <00._>900.0 0...: 80.00. 0 00000000. 0000... 00000090200 0.0250 154 5:835.» .29... ..8 mo... 2... .38. 2 2:25 2282.. :25 2.880 xza 3 o .....o 65:25.. 32.. 522.. «2...... 0.2.. 5.88%: .25 8:98 3... a8... 3%... m 8.: .288”. 95.3.9... 3.2.3. 39.2 ac.c8...283? ......0m .3 388% 2253.. ....o. 25.2, .2731; 3.8... 8.30 ”n 035... 155 050.0“. 508.0 .252... 38....0828._m.02 52.2.50! 2.0 0.58.50 ..00 2.0.8.300”. 5.03.0 2.0 530.0 ..00 5.8.02.0... =00 .m0..< 0.2.0 ..00 5.83:. 0.0202? 5.8.300”. 0.00823 2.28:2: 0.0082? 850 0.82.8... .2... 8083...... =8 E>=o< 0025. 508.... 00.0 v0.0 30 00.0 00.0 00.0 90.0 ..00.0 00.0 00.0 50.0 n00 00.0 00.0 a 2.2. «025-0 .... 3.8.2:. a 8.9.. $5.222“. Ava—2.... v 0000.90.02.20: .0 5.5.5.. 0:00.... 50.0 508... 8000.330 8...... 0.3.9.00 89.2. 50.9.. 5.2.5»... .382 .5530 < 59.2.. 0.0.5.220. £08... 05.00008... 05.00 as... a $230. 2% E28. 9.22 32.9. 2.0.20 0.9.8... 00.0 #00 33. 50.9. 8.0.082 £80 25.200. 22.0.0 2002.... £00.00 156 0055.505. 05 530.0 =00 2.2.2.. 82.0. 5.2.20.2 ...... .888: 0...: 8.8.2 ..00 0.02.005... .2888.» 0.0.. 2...... 8.80 52.830: 0.... 2:830 =00 0.00.52 5000.00.60 5505.005... .05“. 505.005.... 0055.502 05 52.0.0 =00 b.>..0< 5.00“. 505.005.... 3.2.0... 5.0.200”. 5.59.0.0 5525.005... 5.5.200”. 505.005.... 55.500 525.005.... 0050 5.5.000”. 55000005.... 5.5.0520 :00 ..o 5.5.300”. 055002 5.5.05.0... :00 8.8.2.20 :00 .o 8.8.382 02.082 2.8.85 0.2.: 00:00 0.0302? 05 500.050... =00 5002005.... 55.0 202.500 0555.0 50.0... 5.00000 00.0300 50.0.0-0 5.63005... 55.0 0000.5 5m 0000000 0555.0 5.2.0005... 5.83005; 55.0 0025.. 5.0000”. 02.00 b.>..0< 0005. 055055550 50.0... 00050. 052.05-02-02 .0 5.52.04. 500505; 55.0 5.0000”. 225”. 5.00000 00 KG 0050 5003005.... .000.0 005.005.5530”. cozocau 00.0 N00 00.0 F00 5.0 50.0 00.0 00.0 00.0 00.0 30 F000 00.0 F00 _.0.0 _.0.0 00.0 3.0 00.0 00.0 No.0 V0.0 5.0 ..0.0 vi a: 8.2.2.8.. 0883-852 8 22.52.2205 3.. 5.2. ...oo 2...... m 08:58.. 0... .0000. 08.288 ...... 8.2.88... 0: .2500. .858... 88.8 .xx 8... . 2...... 5.. 52.20. ....38 88. ....E. .N 5.3.8 005 55005.. 0.00 05.5 5.50.. 005. .0050 50.0-0050 3.. 2.8. F .28.. 8......85. .mm o... .22.. 2.885 22.. .57.. 2.. $09.5. 522.. 8.2.... <20 «<5 0... 05.88.. 72:. [0.59 8.. 880. m .8828 2...-..888... 5.. 82.0. 830 89.8 .882 .... .85. m .850... 5.5. 05 ...: 2000. 8.8. . .8.-.....m.u..m..-.._..m ... 8.0.0.. 8.. ...._<. 7.28. 02825.... 5.8.... 5.. 3852 2.. .88.... cm 2.8.. 05.6.8.0 3.. 5.00. .. .288. ....o... 0-0. 2.52.2.0 2.. 5525.. F 2.22.. 8.82.8 0805 ....m ow... 32... . o8... 88.. mm. 62.200. ....38 20.. .288. 02.88 2-5.2885 3.. 802.0. ... 8.80 0... .55... no: 30...... 50.0.0 50.5.2.0 05.503005... 005.8 0.0". 0.5: 0:00 005.30 3 2.00.0. :50... 00:00 005.3005: 0.0205 >555? 0.....00 5 030300 «5.500.: ..(0 0:0.0> mm.— 3 00.8.0 00:00 ...0 030... 157 5.5.828 amass”. 3... 3.... 9.5.... 55.888 53.8 No... ...... 37.2.2. 9 82.29.2902 5.5.2 @528 50.9.. ...... .....o .25... 592.. 3.758.820 £85535 3.80.2.2 9...:n. Food 3... 9.0a _. 9.... $353.35.. 0353.39.05. 05.3... 58:32.85 .8... . k... .28. 5.88 8:8 55.33.... .2... 2:82“ ..8 5.53%.. 8.0.82...» 3... 8... .352. 522. 3.58.3.8... 5.8.8.8 8.3883 6... an»... .996. 9 3.5.85.8. 3.58335... 55.88%: ......gm «.3-an 55.538”. 8.5.8.5: 5... ...... 50.9.. 8.59-952 5 9:5 «9.00 50.5353. 350.3539... 5.5.9.9.... ..8 3... 8... 89:2. 8.2.2. o 599.. 6.55.3... 58.82 .8..< 20.6 ..8 No... 3... 5558 5.538.. 58.82 No... 8... 805:. 529.. 8.583... 7.988. “.25 06:00 «.3322 2:. 50:88:05... .30 8880 95.96% 5.2.88.5 no... 8... 2.6.8. «5.6-9.0 ....>..o< 32.x m5..8.....o5.mm 50.9“. 3... a8... ..xmasz. ca 599.. 95.9.8958... 53800 $99385. 95:09:35.3 5.905 vod mm... 9...... (N www.mcamofi £90.... $99.30.... 958.: 5905 no... mod n...o< 38.582922 3... m... a 3.2.2. 2 82998.53: 5.6.... 158 53:94. «3535:. vod 5.898 58.60 3.0 8.0 and a mam: 595-0 3 8529:. a A :wOmv F¢m8£mcgo§méé 28.3 :9ng magic. 159 0.52 9325 .0893. 3:22.529... 532.029. 5 3295 39.: mc_com.:=o< 5:033:92 _mcm_m vouflnmsdszo 33:8 9.2235 8996 9m_>cmum £99m 0 5:895 .0689 mc_xo>o.oc_xo$o 9.59.50 9.385291 9583 9:966 55:89.5 “.9239 533880 90389 95966 c. cozficomoa new 9398 Susie.— oocoo 53032.9.— _aco_m 3.0.3:...2903. c0325... coo woo ro.o Nod vo.o _.o.o moo woo woo vo.o woo woo woo woo voo hood woo woo woo mooo mood coo woo woo vt 2.. Exé . .282... .32 8.. 69.8: 8.3828 882.882.. 82... 522.. 2.. 68.2.. m .88... 2.8.. .28. 28.8.. .25; 2.. $3 > 8.88282 2.. some 82.... 522.. 8.8: 208 .83... ..oo 2.. fzooo : 2.0.6 8.. 5-..: 2.58.2... Re Emcee... 888. .28. 5:58 8283223“. 8.. 830% m .558 8:8 88528 8.2 856V 5 8.: 88.96 292:8... 8.. .830: _ 8.: .288. 23-5» coca—.0 20". «En: 2.00 cob—.3 3 0559 ...—G: nocoo u8a3o9n: «732:. 85.3.9... hwim as $8083. 2.253.... ..(m «:29, 3.0. 3 v98.“ «930 u... 282 160 £3558... Rom 0:25 ”5:335 95 w.w>.oo>.0 Em._onm.mo Rom 5.8m 520.3555 856 E2355! 0:: 83026 .50 55:85.. .0 5.5.53. 55:85.. .0 5.5.59". 350 5.5.35”. 55.5528... £3.55 :8 .o 55:69 .8500 90.6 =8 9.. .0 5.5.52 9.582 5.8522 .m:m.m m.mo.aoam-..:< 00:00 £352? .5: 5388...»... .50 £39m ..mo 5 5.9 5.2.89.5 555?. 5589 5.58 c.9990 2588 5.555 5.2.085. 2200 955. E32. 5205 3550955.? 8.9.52. 3:00 5.5522... .555 55.55.5530”. :ozocau 3.0 No.0 mod mod No.0 No.0 3.0 mod No.0 wood mod vQ Bo :Ozm. 2% .. 398m 8: .5. 89685.. 520 85:99 mu_5m8__eu§o mod 85:. o 3389 5&3 329.5: 8.0 .3 E5. x: 522: 59... EN Ed .393. 33.2: 2.2.0.5 um.m_9-m.mo.aoa< Rd Ezaz. 8.. 5998586 5.8.82 mad Guwmuzb m 5: 5.: .6589 55m. $3.8: 583 5... 22.25. 8 seoaoocogesg no... 852. m 8389 358 0.5552 8: .855. um 3926 99.55 8: :55. F 83232.. 558% 58 35:0 v.0“. 0:5: 500 53:3 8 3.5.2 3.0.. 353 385.5953“... 5.5.55.2: v5 £35555... .5.mm¢5xo 0:5 $5.50”. 5.99.. 5.203 :5 $55: 5 080.2 5 5.5.5. 00:00 83.380! ..5 23035 .50 5.5558. .mcmfi ucm 55:85:. :. 529:. 599555.... 595 .5: $53.50 .. .oa 03:.00 v8.0 0... :. 55:00:02 .0008 00.0.00... 83:80 2.00.8 88.... 0.88.0 .88... .....8. 8.. 83200 2.0-2.0.. 88.... 8.88.0 2.0.2.0 5.02.00 8.5200 08.058 5.0000. 0......0... ...<...w.v.<.. 20:9... 0:..0c08 05.0.5 2208800820 .2008. :.0.0.0 0 20:2... .20.:00. 5:80.55 ..00 .200. 2.30.0 00.0.008 0:..0:0.0 000:... 0:82.... 02.8.2... 82.80 2.0.2.... 08.058 9.02 2...... 00208. $03500 0:..0:0.0 v.03). .:0800.0>00 0.0.800 5.0.008 8.5500 05.058 8.5 00:00 55:005.... .055 020.005.5508”. :000::0 3.0 3.0 No.0 3.0 3.0 30.0 00.0 3.0 3.0 3.0 «0.0 00.0 3.0 00.0 3.0 00.0 00.0 00.0 3.0 No.0 3.0 3.0 3.0 v1 0... QNF.F 0... .... 0... 0... 0... 0... 0... .0. 00.. 00.. 00.. «0.. 0.0.. 00.. 00.. 09.2.0 .200 50.0.00. 2.00 5.0000. 00.02.00 02.80.55 080008.00 ...000. 00.020.00.05... ... .28. 820808.. 000 ..00:. . .28. 020808.. 0.8.00: ..080. . 8.8208285... 02.3.0 ......0. . 08.... 9....200 .0500. 09.00 ...220 8.2.... 0.0 0.. 00.0 8.0.8 ..00 .2022. 050005 00.0.0. .0..> 080.082.60.38 022$ .0>s.. 0:00.080: 0:00005 .0..> 080.082.60.38 022$ 8.-.... 9-0.0.802... .0000. .2008. 5.0.80.2...0 .0000. ..::..:0 0:20.00 .0 0002000005 :.0.0.0 .02.... N .00808 .2.80..00:0 .200. 0.00.00: .08: ... .000. 000:... .2008. 00.050.50.20 0 ..0<0.... 0. 0.220 8.2.5 80... .000. 80.82.80 .0000. 0 .28. 2.8.88 .800 5...... e .2008. 9...-..8 .03.. 20.. ..-c...:0..2c_ .m 2004.. .2008. .-m 7020.0 0.0.05.4. .00... .28. 28.... 2.8.80: .0 _ wz - .= .m> _ mz - _ .m> o wz - _ .m> o .. Nvd .= .m> _ . weed .__ .9 _ .. «we . .m> o .. El .__ .m> . wz - . .m> U - oz ... .m> _ .. 9.; _ .m> o - - .__ .m> o 2. 8+ _ .m> o - 02 .= .m> _ 2. SN _ .m> o wz - .= .....5 _ . 8N _ .m> o 02 - ... .m> _ .... ovd _ .m> 0 03.25". 00:20.20". gag—coo £2.35 «0.75.6 00.0 050 0N0 00.0 rad NWO 00.0 9.0 #00 9.0 9.0 3.0 00.0 o:_a>.n_ once—.0 Son. 38:50 009 50.0 3.9 50.0 92—. 00.0 «0.0 00.0 vu... DZ 5.... 506 00.0 0?... 00.0 00.0 ... .m>_ ....m: .90 _.w>0 .=.m>_ .__ .m>_ _ .m> 0 ..= .m>_ _ .m> O .__ .m>_ _ .w> 0 4. .m>_ _ .m> O .= .m>_ _ .m> 0 02 02 .= .m> _ . .m> 0 #523: 5.2m 00.0 . 03225 moans—-0 «=23 Eva—3.85 0N._. 05.0 mwé RN... 0:. 00.0 on... 5.0 omé 05.0 0Né 50.0 4. .9: .= .9... _ .m> 0 ..m>0 02 DZ 4. .w>_ _ .m> 0 .= .m>_ _ .m> 0 .= .m>. _ .m> 0 .= .m>_ _ .m> 0 .= .m> _ _ .m> 0 ..= .m> _ _ .m> 0 2:9:— oucugo Son. $82.00 5523: 35m Fwd. mam; N¥_ mz - .. .m> _ mz - do: wz - do: m2 - ._ .m> _ wz - .. .w>_ mz .. _ .m> o mz - ._ .w> _ wz - _ .m> o wz - .. .m>. wz - _ .m> 0 e255. 00:20.20“. «nab-30 £2.35 mo..-E.o it «33>... 09.2.0 So“. «9550 00.0 N...P 00.0 006 N... N00 00.0 ..0... .553 ca 2 9 use. mazes“ 52. .858 8 8%: » 585:9... 82 n mz 85.59% 32 u oz .. .m>. ... .m> _ 4. .m>_ .. .m). .. .m> _ .. .m>_ _ .m> 0 .. .m>. . .m> 0 4..w>. ..m>0 >atn20=z ...mEm t 00.0 .. 00.0 e 0.? 3.25. $55 Eon. .2530 .9355 Shoo oz oz oz _ .9 0 oz 4. .2: _ .9 o o. .2: _ .2, o ....m>. ..w>0 ..0.0 vn. .... 00.0vl .. hmm 10h“... Nevin. 0>2 N; chomp ($blast); print "\nWhat format is the BLAST output in? (l=xml 2=html)\n"; my $format = ; chomp (Sformat); if ($format == ) I print "\nParsing BLAST output ...... \n"; @parseout = ParseblastXML ($blast); } 206 if (Sformat == ) { print "\nParsing BLAST output ...... \n"; @parseout = ParseblastHTML ($blast); I if (Sformat > 2) I print "PLEASE ENTER A VALID SELECTION\n"; die; } if (Sformat =~ /\D/g) I print "PLEASE ENTER A VALID SELECTION\n"; die; } $/ = "\n": print "\nWould you like to print out a PASTA file of sequences that did not return a hit?\n"; print "Another BLAST could be peformed with this set.\n"; print "Perhaps with less stringent conditions.\n\n"; print "If you answer yes, you will need the original PASTA file used for input\n"; print "(1=yes 2=no)\n"; my $nohit = ; chomp ($nohit); if ($nohit == l){ NoHits(); } if ($nohit == 2) { print "You entered no, continuing ...... \n"; I if ($nohit > 2) { print "PLEASE ENTER A VALID SELECTION\n"; die; I if ($nohit =~ /\D/g) I print "PLEASE ENTER A VALID SELECTION\n"; die; } ############Ask if you want to get missing ones from internet print "There are often sequences for whose locuslink IDs could not be extracted from loc2ref file\n"; print "Would you like to extract these IDs from the internet? (requires open connection and LWP::Simple module)\n"; print "\n(1=yes 2=no)\n"; $/ = "\n"; my $ID = ; chomp ($ID); if ($ID > 2) I print "PLEASE ENTER A VALID SELECTION\n"; die; I if ($ID =~ /\D/gl I print "PLEASE ENTER A VALID SELECTION\n"; die; I s/ = "\n"; 207 print "\nEnter the desired output filename\n"; print "Ouput file will be saved in :\n$dir\n"; my Soutfile = ; chomp (Soutfile); print "Would you like to create an html file output in addition to the tab delimited one?\n"; print "(1=yes 2=no)\n"; my $htm = ; chomp ($htm); my $html; if ($htm == l){ print "\nEnter the desired name of the HTML output file\n"; print "It will be saved in $dir\n"; Shtml = ; chomp ($html); } if ($htm > 2) { print "PLEASE ENTER A VALID SELECTION\n"; die; I if ($htm =~ /\D/g) I print "PLEASE ENTER A VALID SELECTION\n"; die; } print "Running the script with the desired options, (script can take 30min to several hours to run)\n"; print "Retreiving Locuslink IDs, please wait, this can take a while....\n"; @locusinput = GetlocusIDs (@parseout); #################Find ones that are missing my %hits; foreach (@parseout){ my (Sname, $gi, $acc, Sdef, $score, $evalue) = split "\t"; $hits{$name} = $evalue; } my %linksfound; foreach (@locusinput){ my (Sclone, $locusID, $evalue) = split "\t"; $linksfound{$clone} = $evalue; } my @missing = (I; foreach (keys %hits) { push(@missing, $_) unless exists $linksfound{$_}; } print "\nDone\n"; if ($ID == )I print "Extracting locuslink IDs from the internet...\n"; my @missinginput = AlternateGetlocusIDs(@parseout); @locusinput = (@locusinput, @missinginput); 208 print "Web extraction complete\n"; I if ($ID == 2) I print "You did not want to extract missing IDs from the web, continuing ...... \n"; } ############################################### print "\nParsing LocusLink Database file....\n"; my @locusoutputl = Locuslinkparse (@locusinput); 4* LocuslinkparseZ (@locusoutputl); print "\nDone writing to tab delimited file\n"; 5/ = "\n".- a if ($htm == l){ csv2HTML($outfile); print "\nDone writing to html file\n"; I if ($htm == ZII print "\a\a\a\nScript complete, check output file\n"; I unlink "locustemp.txt"; unlink "nohits.txt"; print "\a\a\a\nScript complete, check output files\n"; sub ParseblastXML { my $blast = shift @_; my ($queryname, $gi, $aCC, $def, $score, $evalue, $nohits, @parse); $/ =''; open (NOHITS, ">nohits.txt"); print "$blast\n"; open (BLAST, "<$blast") ll die ("Error Opening $blast \n"); while () { ($queryname, $gi, $aCC, $def, $score, $evalue, $nohits) = II N . I if ($_ =~ /(.+?)<\/BlastOutput_query-def>/g) { Squeryname = $1; if ($queryname =~ /\#/g) I Squeryname =~ s/\#/\t/g; I else{ Squeryname = "$queryname\t"; *=8t=fl==fl==fl==fl= I 209 if ($_ =~ /gi\l(.+?)\lref\l(.+?)\l<\/Hit_id>/g){ Sgi = $1; $acc = $2; if (S =~ /(.+?)<\/Hit_def>/g) { $def = $1; if ($def =~ /gi\|(.+?)\|.+?\|(.+?)\|.+?/g) I 591 = $1; Sacc = $2; if ($_ =~ /(.+?)<\/Hsp_bit-score>/g) I $score = $1; if ($_ =~ /(.+?)<\/Hsp_evalue>/g) I $evalue = $1; I if ($_ =~ /(.+?)<\/Iteration+_message>/g) { $nohits = $1; I if ($nohits =~ /No Hits Found/ig) { print NOHITS "$queryname\n"; ($queryname, $gi, Sacc, $def, $score, $evalue, $nohits) = ""; next; I else { push @parse, "$queryname\thi\t$acc\t$def\t$score\t$evalue\n"; ($queryname, $gi, $aCC, $def, $score, $evalue, $nohits) = ""; next; I I close NOHITS; close BLAST; return @parse; I sub NoHits { open NOHITS, "; chomp (Soriginal); open (PASTA, "; chomp ($nohits_out); open (FASTA_OUT, ">$nohits_out") ll die ("Error Opening $original my (@clone_with_hits, @clones); S/ = a)"; ; while () { my (Sname, Ssequence) = split "\n"; #my ($clone, STC) = split "\t", $name; push @clone_with_hits, "$name\t$sequence"; I s/ = "\n"; while () I $_ =~ s/>//g; $_ =~ s/\n//g; $ =~ s/\r//g; pEsh @clones, "$_"; I foreach my $clone_name (@clones) { foreach (@clonefiwith_hits) { my ($clone, $sequence) = split "\t"; if ($clone =~ /$clone_name/i) { print FASTA_OUT ">$clone\n$sequence\n"; I else { next; I I I print "FASTA file complete\n"; close FASTA_OUT; close FASTA; close NOHITS; sub AlternateGetlocusIDs { use LWP::Simple qw($ua get); $ua->agent("Mozilla"); # hide the fact that this is libwww-perl my @missinginput; my $locusID; S/ ._.. "\n"; foreach (@_) { chomp; my (Shame, $gi, $aCC, $def, $score, $evalue) = split "\t"; foreach (@missing){ if ($_ eq $name){ 211 my $url = "http\:\/\/www\.ncbi\.nlm\.nih\.gov\/entrez\/viewer\.fcgi\?val\=$gi\&db \=Nucleotide\&dopt\=GenBank"; my $ncbi_text = get($url); if ($ncbi_text =~ m/LocusID:(.+?)<\/a>/g) I q $locusID = $1; push @missinginput, "$name\t$locusID\t$evalue\n"; print "."; next; else{ next; I I return @missinginput; sub GetlocusIDs { open (LOCZREF, ") { my (Sid, $accession, $number, $status, Sprotacc, $num) = split "\t"; $accession =~ s/\..+//g; push @loc2ref, "$accession\t$id"; next; I foreach (@_) I chomp; my ($name, $gi, $acc, $def, $score, $evalue) = split "\t"; $acc =~ s/\..+//g; foreach (@loc2ref) { my ($accession, Sid) = split "\t"; if ($accession =~ /$acc/ig and $evalue =~ /\d/g) { print "."; push @locusinput2, "$name\t$id\t$evalue\n"; next; I else { next; I next; } close LOCZREF; return @locusinput2; 212 sub Locuslinkparse { S/ = "\n"; open (LOCUSPARSE, ">>locustemp.txt"); my (@locuslinkIDs, @locuslinkparse); ">n; I while () { my (Sid, $refseq, $symbol, Sname, $summary, Somim, @DBKEGG, @locuslinkparse); if ($_ =~ /LOCUSID: (.+?)\n/g) { if I @DBKEGG = m/(DB_DESCR: KEGG pathway.+?DB_LINK: ($_ ($_ ($_ ($_ ($_ ($_ (3 Sid = $1; Sid =~ s/\r//g; I =~ /NM: (.+?)\n/gl I $refseq = $1; $refseq =~ s/\r//g; =~ /OFFICIAL_SYMBOL: (.+?)\n/g) I $symbol = $1; $symbol =~ s/\r//g; =~ /OFFICIAL_GENE_NAME: (.+?)\n/g) I Sname = $1; Sname =~ s/\r//g; =~ /PREFERRED_SYMBOL: (.+?)\n/g) { $symbol = $1; $symbol =~ s/\r//g; =~ /PREFERRED_GENE_NAME: (.+?)\n/g) I Sname = $1; $name =~ s/\r//g: =~ /SUMMARY: (.+?)\n/g) I $summary = $1; $summary =~ s/\r//g; =~ /OMIM: (.+?)\n/g) I Somim = $1; Somim =~ s/\r//g; foreach (@DBKEGG) { 213 "$locusID\t$clone\t$evalue"; #(LL3_O40211.txt .+?)\n/sg; $ =~ s/DB_DESCR: / /g; $ =~ s/DB_LINK: /<>/g; $ =~ s/\r//g; S =~ s/\n//g; I @G0 = m/(GO: .+?)\n/sg; foreach (@GO) { S =~ s/GO: / /g; $_ =~ s/\r//g; $_ =~ s/\n//g; I push @locuslinkparse, "$id\t$symbol\t$name\t$refseq\t$omim\t$summary\t@DBKEGG\t@GO"; #print it out in tab delimited format print LOCUSPARSE "$id\t$symbol\t$name\t$refseq\t$omim\t$summary\t@DBKEGG\t@GO\n"; (Sid, $symbol, Sname, $refseq, Somim, $summary, @DBKEGG, (960) = "n; I close LOCUSLINK; close LOCUSPARSE; return @locuslinkIDs; I sub Locuslinkparse2{ my @GOslimbio = ("behavior GO:OOO7610", "biological_process unknown GO:OOOOOO4", "cell communication GO:OOO7154", "cell recognition GO:0008037", "cell-cell signaling GO:OOO7267", ”host- pathogen interaction GO:OO30383", "response to endogenous stimulus GO:0009719", "response to external stimulus GO:OOO9605", "response to abiotic stimulus GO:OOO9628", "response to biotic stimulus GO:OOO9607", "signal transduction GO:OOO7165", "cell growth and or maintenance GO:0008151", "cell cycle GO:OOO7049", "cell growth GO:0016049", "cell organization and biogenesis GO:0016043", "cytoplasm organization and biogenesis GO:OOO7028", "organelle organization and biogenesis GO:OOO6996", "mitochondrion organization and biogenesis GO:OOO7005", "cytoskeleton organization and biogenesis GO:OOO7010", "cell proliferation GO:OOO8283", "chemi- mechanical coupling GO:OOO6943", "cell homeostasis GO:0019725", "metabolism GO:0008152", "amino acid and derivative metabolism GO:OOO6519", "biosynthesis GO:OOO9058", "carbohydrate metabolism GO:0005975", "catabolism GO:OOO9056", "coenzymes and prosthetic group metabolism GO:0006731", ”electron transport GO:OOO6118", "energy pathways GO:OOO6091", "lipid metabolism GO:0006629", "nucleobase nucleoside nucleotide and nucleic acid metabolism GO:OOO6139", "DNA metabolism GO:OOO6259", "transcription GO:OOO6350", "protein metabolism GO:0019538", "protein biosynthesis GO:0006412", "protein biosynthesis GO:OOO6416", "protein biosynthesis GO:0006453", "protein modification GO:0006464", "secondary metabolism GO:0019748", "response to stress GO:0006950", "transport GO:OOO6810", "ion transport GO:OOO6811", "protein transport GO:0015031", "death GO:0016265", "cell death GO:0008219", "development GO:OOO727S", "cell differentiation GO:0030154", "embryonic development GO:OOO9790", "morphogenesis GO:OOO9653", "regulation of gene expression epigenetic 214 GO:OO40029", "growth GO:OO40007", "reproduction GO:OOOOOOB", "physiological processes GO:OOO7582", "viral life cycle GO:0016032"); my @GOslimcell = ("cell GO:0005623", "intracellular GO:0005622", "chromosome GO:OOOS694", "cilium GO:OOOS929", "cytoplasm GO:OOOS737", "cytoplasmic chromosome GO:0000229", "cytoplasmic vesicle GO:0016023", "cytoskeleton GO:0005856", "cytosol GO:0005829", "endoplasmic reticulum GO:0005783", "endosome GO:0005768", "Golgi apparatus GO:0005794", "lipid particle GO:OOOSBll", "microtubule organizing center GO:0005815", "mitochondrion GO:0005739", "peroxisome GO:OOOS777", "plastid GO:OOO9536", "ribosome GO:0005840", "vacuole GO:OOOS773", "lysosome GO:0005764", "nucleus GO:0005634", "nuclear chromosome GO:0000228", "nuclear membrane GO:0005635", "nucleolus GO:0005730", "nucleoplasm GO:OOOS654", "plasma membrane GO:0005886", "thylakoid GO:OOO9579", "cellular_component unknown GO:0008372", "external encapsulating structure GO:OO30312", "cell envelope GO:0030313", "cell wall GO:0005618", "extracellular GO:0005576", "extracellular matrix GO:0005578", "extracellular space GO:OOOS615", "unlocalized GO:0005941"); my @GOslimmolec = ("antioxidant activity GO:0016209", "apoptosis regulator activity GO:0016329", "binding GO:0005488", "calcium ion binding GO:0005509", "carbohydrate binding GO:0030246", "lipid binding GO:0008289", "nucleic acid binding GO:0003676", "DNA binding GO:0003677", "chromatin binding GO:OOO3682", "transcription factor activity GO:0003700", "transcription factor activity GO:0000130", "nuclease activity GO:OOO4518", "RNA binding GO:0003723", "translation factor activity nucleic acid binding GO:0008135", "nucleotide binding GO:OOOOl66", "oxygen binding GO:OOl9825", "protein binding GO:0005515", "cytoskeletal protein binding GO:0008092", "actin binding GO:0003779", "cell adhesion molecule activity GO:0005194", "chaperone activity GO:0003754", "chaperone activity GO:0003757", "chaperone activity GO:0003758", "chaperone activity GO:0003760", "chaperone activity GO:OOO3761", "chaperone regulator activity GO:OO30188", "defense and immunity protein activity GO:0003793", "catalytic activity GO:0003824", "kinase GO:0016301", "protein kinase activity GO:0004672", "hydrolase GO:0016787", "peptidase activity GO:0008233", "protein phosphatase activity GO:OOO4721", "transferase GO:0016740", "enzyme regulator activity GO:0030234", "molecular_function unknown GO:0005554", "motor activity GO:0003774", "protein stabilization activity GO:0017028", "protein tagging activity GO:0008638", "signal transducer activity GO:OOO4871", "receptor binding GO:OOOSlOZ", "receptor activity GO:OOO4872", "nutrient reservoir activity GO:0045735", "structural molecule activity GO:0005198", "transcription regulator activity GO:0030528", "translation regulator GO:OO45182", "transporter activity GO:0005215", "electron transporter activity GO:0005489", "ion channel activity GO:0005216", "neurotransmitter transporter activity GO:0005326", "triplet codon-amino acid adaptor activity GO:0030533"); $/ = "\n"; open (OUT, ">$outfile") || die ("Error Opening Soutfile \n$!\n"); 215 print "Finding matches in LocusLink, compiling GeneLink file, this may take quite a while (about 25 mins, but could be hours) and use a lot of CPU resources....\n"; open LOCUSPARSE, ") { my ($id, $symbol, $name, $refseq, $omim, $summary, SDBKEGG, $60) = split "\t"; my ($acc, $gi, $type) = split (/\l/, $refseq); my ($GObio, $GOcell, $GOmolec); foreach (@GOslimbio)I my ($name, $GOnum) = split "\t"; if (SGO =~ /$GOnum/g)I $GObio = "$name\|$GOnum"; I foreach (@GOslimcell){ my ($name, $GOnum) = split "\t"; if ($GO =~ /$GOnum/g){ $GOcell = "$name\|$GOnum"; I foreach (@GOslimmolec){ my ($name, $GOnum) = split "\t"; if ($GO =~ /$GOnum/g)( $GOmolec = "$name\l$GOnum"; I I foreach (@_)I my ($llID, $clone, $evalue) = split "\t"; if ($llID == $id) { #print OUT "$clone\t$TC\t$evalue\t$id\t$name\t$symbol\thi\t$acc\t$omim\t$summary\ t$DBKEGG\t$GObio\t$GOcell\tSGOmolec\n"; print "."; push @output, "$clone\t$evalue\t$id\t$name\t$symbol\tSsummary\t$omim\t$DBKEGG\t$GObio \tSGOcell\t$GOmolec"; next; I I next; I #make sure there are only unique entries in the output my @uniq; my %seen = (I; foreach my $item (@output) I push(@uniq, Sitem) unless $seen{$item}++; I foreach (@uniq){ print OUT "$_\n"; I 216 s/ ___ "\n"; open (NOHITS, ") { chomp; #my ($clonename, $TCnumber) print OUT "$_\n"; split "\t"; I close NOHITS; close OUT; I sub csv2HTML { my $csv = shift @_; my $blank = ""; my @out; $/ = "\n"; open (HTML, ">$html") || die ("Error Opening Shtml \n"); print HTML "\n\n"; print HTML "\n\n\n"; print HTML "\n"; print HTML "\n"; print HTML "\n"; print HTML "\n"; print HTML "\n"; print HTML "\n"; ########re-ordering edit print HTML "\n"; print HTML "\n"; print HTML "\n"; ################## ##print HTML "\n"; ##print HTML "\n"; ##print HTML "\n"; ##print HTML "\n"; ################### print HTML "\n"; print HTML "\n"; print HTML "\n"; print HTML "\n"; open (CSV, "){ ##my ($clone, $TC, $evalue, $id, $name, $symbol, $gi, $acc, $omim, $summary, $DBKEGG, $GObio, $GOcell, $GOmolec) = split "\t"; my ($clone, $evalue, $id, $name, $symbol, $summary, $omim, SDBKEGG, $GObio, $GOcell, $GOmolec) = split "\t"; my ($pubmed, $pubmedsearch); if (Sid != $blank) { $pubmed = "Pubmed Links<\/A>"; 217 $pubmedsearch = "Pubmed Search<\/A>"; I #Sacc = "$acc<\/A>"; $id = " $id<\/A>"; if ($omim != $blank) { $omim = "$omim<\/A>"; I my ($GOnamebio, $GOnumbio) = split /\l/, $GObio; my ($GOnamecell, $GOnumcell) = split /\|/, $GOcell; my ($GOnamemolec, $GOnummolec) = split /\l/. $GOmolec; if ($GObio =~ /\D+/g) I $GObio = "$GObio<\/A>"; I if ($GOcell =~ /\D+/g) I $GOcell = "$GOcell<\/A>"; } if ($GOmolec =~ /\D+/g) { $GOmolec = "$GOmolec<\/A>"; I my @KEGG; my @pathways = split / /, $DBKEGG; foreach (@pathways)( my (SKEGGname, $KEGGurl) = split "<>"; push @KEGG, "$KEGGname<\/A>"; ##my $line = "$clone\t$TC\t$evalue\t$id\t$symbol\t$name\t$acc\t$omim\t$pubmed $pubmedsearch\t$summary\t@KEGG\t$GObio\t$GOcell\tSGOmolec\n"; my $line = "$clone\t$evalue\t$id\t$name\t$symbol\t$summary\t$omim\t$pubmed $pubmedsearch\t@KEGG\t$GObio\t$GOcell\t$GOmolec\n"; print HTML "\n"; I print HTML "
Sequence nameE-valueLocusLink IDOfficial Gene NameOfficial SymbolSummary of Gene FunctionOMIMBibliographyAccessionOMIMBibliographySummary of Gene FunctionKEGG PathwaysGO Biological ProcessGO Cellular ComponentGO Molecular Function
"; $line =~ s/\t/<\/TD>/g; print HTML "$line"; print HTML "
\n"; .218 print HTML "\n"; close CSV; close HTML; I sub ParseblastHTMLI print "Not included in genelink.pl yet, please try to do your BLAST output in XML\n"; die; I 219 05.0 05.0 05.0 V50 V50 V50 05.0 N50 N50 05.0 05.0 05.0 000.0 000.0 000.0 000.0 500.0 000.0 000.0 000.0 000.0 50.0 e:.u> ll :0 00.0 00.0 0...0 00.0 50.0 no.0 00.0 0 ...0 00.0 ...00 00.0 00.0 00.0 00.0 5.0 .....0 No.0 no.0 00.0 00.0 00.0 . thm 09.2.0 coatanEoO ..0.0 0.... 0w... mué 0w.w 0...? 0...? 00.0 mm... 00.0 000 0P... «00 NN... Nwé 0N... RN... 0.... Eur N...w mm... 00.0 Eon. .....U. m> ..(m ”.0. 0> ..<0 “.0. 0> ..<0 “.0. 0> ...(m “.0. 0> ..<0 “.0. m) ..(m ”.0. 0> ..(m “.0. 0> ... ..<0 ...0. 0> 4.40 ”.0. 0> 4% ...0. m> .. 4(0 “.0. 0> ..(w “.0. 0> ...(m “.0. 0> ..(m “.0. m> ..<0 “.0. 0> .. ...(m “.0. 0> ..<0 “.0. 0> ..(m “.0. 0> ..Nc0om_ 00053. 0005090200 2923 9.00 .2080. 5.00.. 8 .e .38. 8535.... 5898: :36:ch «:90 5.058;... ..8 moo 0.29:8 0cm .0 0:0 v 398 0:00 .On..2.... 5.0.9682... 0090300526 0550.80.59.80 >>>> o. ..m__E_0 3.000 565:: £0.20 £0.60 2 cm._E_m 5.08 x. .288. 08:. 0-0. 056520 5.0009 5.02.00 comosEmma 003.8059: 88 86.9-9.2 E. 9:20 :26:ch 35.2. 2%.. 5.00 050.2 0000.90 050.08 00202? 00053. 5.09.00 .000. 92.... 2 565.5 eczoom 000 .900. 500.805.. 00003-02 20.00. _ 000:0. £0.90 0902.00.50 Snow. m .8550 05.0.0300th 2 8.3828 8888353 0:20 2:0: 0:00 00:00 5.3203. Show ESE—5.0 3. 03¢... ranges—0.00:0 . E0EEmm -2“... (sz 5.0009 mEEmméoBteE. 081.800.8350 8882? 00.52 99.2 9888:0302 84550100002 NomrfiooSSBo «5:82.235... 81188235.... «omoo mwixuooooafiom. «858885350 380810082 m> ..(m “.0. 0> ..(m 00. 0> ..(m “.0. 0> .. ..(m 00. 0> ..(m 00. 0> .. ..(m 00. 0> ..(m 00. 0> .. ..<0 00. 0> .. ..(m ...0. 0> .(m “.0. 0> .(m “.9 0> ..(w ...0. 0> ..(w 00. 0> .(m 00. 0> ..(m “.0. 0> .. ..(w “.0. 0> ..(m “.0. 0) ..(w 00. 0) ..(w ”.0. 0> 40.0 050 .300. 020.000.0020 =00 5505.:0 $050. . 00:50.00. :000..00...:0 00:00 .000.0 0.000.me T... 02.000200 2.2.8 0:032... 2...... .5290 0. :.0.0.0 050:... 02.00.00: 05:0..0 5.05.5 0:00 00.0.8000-:0..0E:.0..:. :0 1:52 00. 2.00.000 0002... 02_~ 0. .....e... 2.00; 5505.5. .052... 0.. 50.0.0820 5505:: .8200. ... 2.80 00.2.2 0:.E0.00000:0.. 5505.5 ..0:0N0.2.2 0.0.30 .230 . :90 .82. 0 00.300. 000:... 0.0.3000 .0..0:0:00...2 ...20 0.0 50.0... 0.0.! 00.05.0000 (00-39005 .520. 0.8 » .... 80:080. :00..:0 05050.00 ..00 0:080:00 5.65.:0 m .00:0>000 .0200. 50.0.0 00.0.8000 50050000000... 3.0 .02. 5.5.5. 000:... .:00:0000.:._0>0 30:383....50 80.883350 $200. <20... .P 00:30.00. c000..00-..:0 .00:00 .000.0 00830000002 .... 0805008350 K500X00000IF 04200 00000000 00002 1000060000 ..0.._.00 000:. 38 .350 .052... 0.. 50.900020 80:383....50 80:. 88 .350 00.2.2 0559300000... 3E000.-0mv_0 500v00m0I06002 221 wood wood vood wood wood ”8.0 nocd wood wood wood wood omod owed 03.0 03.0 53.0 25.0 03.0 mvod mvod mvod 35.0 N36 N36 506 95.0 000.0 mmod mod mod mod nod 00.0 3.0 50.0 mod nod cod no.0 9.0 mod :0 9.0 mod 9.... 0.20 mod mod mod aod nod No.0 3.0 8.0 3.0 mod 0.... mad 3% 2% NNé 55.0 .8. mm... K... 9.... 3.0 0:. 3... no.0 0.2.. m... mad and mo... 0...? mo... 9.... ..F... N... mod o...w 96 3.? am.— m> ._4._4 ..(m am... w> ..(m mm. m> .. ._ .._ ..(w “.0. m> ..(m “.0. m> 4 ._ ..(m “.0. m> ..(w “.0. m> ..(m ...0. m> ._ 4(0 “.0. m> ..(m “.0. m> ..(w “.0. m> ...(m ...0. m> 46 m0. m> ..(w ”.0. m> .73 :3: £905 .85....» 3.750335... 5.65:: .3655 565...: :39.ch .28. £508 CK... .2535: Ron oEEm .mzam: mew. SEE £80685 c2655 .8 Km. oo 30 59...... .833: 20.93 050: :30:ch , 9.0m __ 09¢ wwwcmmofiéwu 2230305.... 3.35 6%.... m 8.: .238. 3:32.95 ESE. .28. ....88 cBchca :30:ch 9.2m 3.38. 958.9. 99...... .5385. ....82m 565...: .325. 8.55 .mEmw. 9.2 92:8 8.....88. .23 :26:ch .85. m .859: .....Em. 0.8 :26...ch ..-mm. ..o .8333... came... 59. :26:ch 8.3.8 8% 3-50. 5589 8.359.900 60mm. 3323...... 2.... 5.3885 9.0805 2.605....me 9.6.0530 80:. 38 6.5m BmJFSoSfiOm New... 88 6.:0m NomuNEooEfiOm 3<|88235m (2%.. .28. 5.88 NEISSESPS o _. Ilo ..ooo ..o.:.0m moOIN w coo S.E.Om moxuoSooSfiom. 20:882....5m .93.... m 3... .238. 95%....9... E22 > ..<0 am.— 0> ..(m am.— 0> .. 0% am.— 0> 05 0m... 0) ..(w 0w.— 0> ..(m 0m... 0> ..(m 0w._ 0> ..(m am.— 0> ..(m am. 0> ..(m am... 0> ..(m 0w. 0) ..<0 am.— 0> .70 0M.— 0> 0% 0w; 0> ..(w am.— 0> ..(w 0m.. 0.. ..<0 0m; 0.. ...(w am.— 0> ..(w am.— 0> .20 am.— 00 ...(w 0w; 0) ..(w am; 0> ..(w .025... 00$. 80.20. N 50.0.0 5.0:0000905 0:00 9000... .288. =8.» 20.. > 30200. 58:0 0.00 5.0000. 0.0.0000 25.0.0005 .02. ...28. 0205502... 20.8.2 CS. .00000:0.. 0.00 0550 .0000: 00.0. .0500 50.0-000.0 55055: $552. . 0:00 :0..000.0:0.. 050005... 0000.. 0.0005... 00.0.00000 000005. 000.0 5.0.0005»... .0. 000 0.0.0500 0:0 .0 0:0 0 0:98 0:00 .002... 5.0600550... 550550 .0205. .2580 8.0. ....E. .N c.0200 .8200. ... 8.80 00<> 5.0.00. 0000 ...< 0.0000) .300 <-<. 5950 0.500005... _ 000.0 .2 55055... ..000. 8.8. . 8.0.8.0220... ... 020.00 .3852. 25.-.8022. 0000.. 0:52.00 .0020 5.055. «0.402 00.0.2032 5 0320 ...-00.. _. .900. 0.30.0 0575.005 .000m. 0 .00:0::0 05.-50000:0.... E00550 ... 0882...... 02888.0...» 0.0(0 80.0. 22°... 2...-.....o.0..0..0 ..<.m..00. 50.0090 :000..00 .XX 00... F 0:0.< m. . 5.30000 .3500 £0.20. N 50.0.0 0005000905 0:00 22:88:52.8 <20... 60.200. 538 0.00 0.0000. 0.0.0000 ...-5.0.0005 <20... .02. 7.0.00. 50.055005 00.0-2.0 00.00. 000 now 0.08.3500 .o.:o.8o.o...00 0.00.88.00.00 05:00:88.8 0.0:. .8235... 820.08.00.00 80:. .80....500 .o0:~.o8.3.-00 <22... .58 <.<. 80...... 2.82:2... _ 08.0 s. 80:88 ...-:00 00.00.808.350 «5:58.350 8858:0082 ...-00.. __ .900. 0.30.0 0575.005 .0. (205 05.5 80:. 80.3.00 80:088.:050 out-0022:0052 083088.850 3<:. .80 ......00 223 .000 000.0 000.0 000.0 0N0.0 5N0.0 5N0.0 500.0 000.0 0N0.0 0N0.0 00.0.0 vw0.0 000.0 NN0.0 NN0.0 .N0.0 000.0 0N0.0 0.0.0 0.0.0 0.0.0 0.0.0 0.0.0 0.0.0 5.0.0 00.0 0.0 00.0 50.0 0.0 ...0 00.0 00.0 00.0 00.0 ...0 00.0 0.0 00.0 V. .0 50.0 50.0 00.0 50.0 00.0 00.0 00.0 00.0 00.0 ...0 00.0 v... 00.0 m... N... 00.0 00.0 5... v... 0... 0.... 0m. 00.0 00.0 0... 05.0 0.... v... 0... v... .... 0... 00.0 00.0 8.0 .0.0 0... 0m. 0.. ..(w am. 0> ..(m 0w. 0) ..(m cm. 0> ...(w 0m..0>.. ..<0 am. 0> ..<0 am. 0> .5 0w. 0> .5 am... 0> .5 0m. 0> ..(m am. 0> ...(0 am. 0> .. ..(m am. 0> ..(m am; 0> 0% am; 0> .5 am. 0> .5 am. 0> :20 0m. 05 .50. am.— 0> .(fl 0m. 0> ..(w 0m. 0) ... ..(m 50055.. 0.280. 0050.000 00.0.0. .0. 2020000 0025.00 .2030. .200. 0.30.0 .05.00.0m 550550 .0228. 50008.0. 55.02. 0.20.3. 0085-80... 8 0.2022222... .0<000.0. ...220 8.080000 .00... .20000. .000.0:00000 0500.... 50.0.0 .-00. .0 50020000 000:... .09. .00005. ...220 8.2.... <20 0<. 550550 550550 . -..0: 0:20 ..20... 0:0000:0 0.0.0 :0...0 .52... . 00.8... 02.2 00. .00<0... 50.0.0 00.0.00000 ...0.0000. 02... 0082.2. 72:. 5.0-2.. 005 000.0 55:50 c5550 c3055: .02... .28 2...... 0 002.028: 55055: 5.2.0.... 55500 .000.0. 0 .00... .2080. 002020008 8.02... 5220 ...08 20 2220 0.0.... ...220 2-0.2.2 .0000. 800.208.. 0.0.. 0.8820 00.050... 00.00.05.000 0>..0.0:0000 85882350 80:. .80 .350 80H282350 :0 882350 2<:882350 0808....<:00002 20:808350 83.88350 80:208350 . 2:088 .350 .52... <20... .. 0.02.... 082 00. .00<0... 50.0.0 00.0.00000 ...0.0000. 02 0. 80:. 88 350 005 000.0 80:282350 20:882350 20:882350 0805082350 20:282350 80:088350 080.808.0050 080500880500 085.580.0050 2.220 2.0.2.... 80:208350 224 mvod 536 56.0 mvod 3.0 «3.6 mvod 35.0 35.0 03.0 mmod mmod mmod wmod hmod smod smod omod owed omod owed mmod mmod 33.6 vmod vmod 00.0 50.0 mod mod 8.9 mod no.0 .96 mod mod mod 96 mod 96 no.0 9.0 mod mod mod Eno nod .....0 mod n v.0 Nod mod mo._. of? E? and owé 3% mo... mad and Q? moé or... had mwd N_..F 8.0 ore 9.9 and 3.0 cad mad moé pod «...? 8.0 mm.— m> .75 am.— m> ..(m am; m> ._ ..(m cm.— 9.2% am.— w>._ ..(m am; m> ..(m aw4w>4444444 .73 am; m> ..(w am; m> ..(w mm.— w> ..(m am.— w> ..(m am4 m> ..(w aw; m> 45 mm; m) ..(m cm.- m> ..(m mm.— m> ..(m am4 m> ..(m ...mcom Ee _ €3.85me 2. 25:8 :26:ch :26:ch 80:5 86.8.. o 5295 5.35% 50.82 :38ch $68 5 .288. 985 0.8 2.6520 :26...ch :xmagv 32: $acc. $acc. £92m umfiéomémmogs :30:ch :30:ch suaam=om mmmnzb cam 32.-me a 3389 x 28 22:2 oEuNSoSSBm «mooxmoooSSBm moonfiooSSBm mamxvoooSSBm mvaXmococleJPOm 5mm thwm Owumz m P .5388 :38 :xngé 32: mmmci $acc. £22m umfi>=omémm§8 mounfiooosfig F Gum SooSfiOm $86885:an $3 .25 momumoooSSSm m8 no. 0310. m> “.0. amino. m> ”.0. amino. m> “5. amino. w> "a. $33.0. m> “.0. amino. m> ".0. 033.0. m> .00. 03.10. m> “.0. 053.6. 9 “.0. 033.0. m> “.0. amino. m> ”.0. “mile. m> ".0. 033.0. m> no. $53.0. m> “.6. 033.0. m> “.0. am... m> ..(w aw.— m> ..(m am: .4. .20 am; m> ..(m aw.— m> ..(m 5.65:: 5.65:: 68.20 2% 828 0:22 mmmoooa 052m? 020.00% $2000 .28. 9:855»: 38mg can 565:: .2039 :29: 59:0: ocflnEmEmcg 550 555?: $13.: mm ..._ ...0. aw._+.._mu_ m> “.0. aw._+u_0_ w> ”.0. nM.—3.0. m) ".0. awnTuO. m> ".0. aw4+u0_ m> ...0. amid? w> “.0. amélmu. m> “.0. aw4+u0_ m> “.0. Qw._+.._0_ m> “.0. amfimu. m> n.0— nM.—.35. m> n.0— mm..+u0. m> ”.0. twink! m> “.0. aw._+n_0. m> ...0. am._+..._0_ w> ”.0. .5435. 9 “.0. amid—G. m> uO. amid—0. A859 83m $95 5282 3.399: .933 mcEiEoo .29 33 5x2; memommwmuowumz :25:ch a: 59:3 5.30.3585. .EmEoo «Ea» m:.... 33 mans 0 mvo: mguooumuooumz 390:8..-«356 2056 00m x _. ..mXNoooo 93.—bu Aflwza “.0. awfimu. m> “.0. nM.—...mmu. m> “.0. nM.—+5 m> “.0. mw._+..._0_ w> ”.0. mm.._+n.0_ m> n.0— await"! m> “.0. aw._+u_0. m> “.0. amid—0. m> ...0. awfimv. M5 “.0. am._+.._0_ m> n.0— am._+u_0_ m> n.0— amid—0. m> “.0. amidG. m> m0. awfimu. w> “.0. nM....imv. m) ....0. nM.—.350. m> ”.0. aw..+...0. w> ...0. czomé :26:ch is; F 82:. 8:8 505 £98.. 83983 .083 93.980 395. .9325 002:: 325 8652 3an8.50 0885.5 «2:953 885.855 a 22386.2”: m8.“ 2% 3. $-29 m8 2%58 793 $96: 8582 =9; 28 > :25: 8.953900 (sz 539a 3.703203 3:3 5.33 < :EEE :26:ch c2653: :25:ch 3258 80 £88 “.0. nM.-3.0. m> “.0. am._+u_0. m> n.0— mm..+u0. m> “.0. awfimv. m> “.0. amid—0. m> m0. mm._+n_0_ m> m0. amid—0. m> “.0. awélmu. m> “.0. aw._+n_0_ a) “.0. mwfimu. m) “.0. aw..+....0. m> n.0— nM.—.10. m> ”.0. 35.28.80 N 8885.8 .9. 2. .8 522: .883”. “29035833 £9825 58138238.: 599: 8:: 88.: 83E 2 8000.080: 28.: E 8596 0.58:5 .35 3.2: 8209 850.88 93 82:. 5.29: 883 8.2.. 95.2.. :26:ch 6.2.838: :26:ch :30:ch 8%. F «85:. :88 5.83: m 82 8389 «55885 «83882358 $88>>__E& .258 223 8..8:6:.m:32 3.6 6.6 o... Ewan. m. 8228.888... 89.88 888.3932 3.... 6... .... Bum... . .8585 3888.. 2.8562. m 2.5288 .8888 852.582 886:6:5:32 2.... 6.6 8.. .220 . 88828 2.2.8 85.225 «88:22 9.9m .0220... 0. .2905 059.3 3:022... mEEwam 88:63:32 3... 6.6 .... 8.2.0 8288. 282822258288. 2526.... 882.20 .88:6:.m:32 3.6 8.6 v. .. ..Nmm<> 2.: 8285.8 ._Dn_ szm max 7.... .F 0.0.0.0050 6000.00.03 00000020300 :92 m .80 00.000 00010.06 v 08.0800030000000 000000200 000.0 68.0.0900 < 0530000 $00.00... 2 .3505 ......8. v 3208.208 00200.0 3. E£0.0.0 000009002000. 300.00.... v< 380.000....0: 000.5320 .000.0.8 .0. N 3.4.500 N 0050.0200000090 0.0000000... 00.000.02.200 20.00.00 .03 0.0.0.00000000 5.00.000. 0.5800006? 3.00.00? 5.9.09.5 m .0200... 5.5000000 .0009: .900. 0.00.000 .005. .0328. w $0.0.000.=>£0E>x0.0>0 00000 F .000.0E ..mmx E08? 00000000.. 0.00.053 3.00.0 so... 305.00.600.30. v 2.0.0. .0050 05.00 00050.? 0.0805620920. N ....m. < .080. 00000020 :0..0..00:0.. P 000.050.... 2000000023 ...00000000000 _. 08.0.0200 (zm .200. 00000020 P .28. 80.528. .0 $0.0 .5058 00.. 0080.000035082 00.830005030020800 $0000.00. .....Ew. m 00000090500 00.5020 . 0000.000300600 03 522. 0582.. 50885. N 9.38 N 08:... 053.920.... 00000020300 4. 0053000 3.0.0.60. : 0.0.6 0.00.0030... .0020. .9 000.000 503::22 anammwm oomwowmm hwgmmxo 8820.0 385322 3330.0 0922422 855580 0808.. «88.00 «+3335 883% «.89: :2 8880.0 0350.0 2.0.5.0 .8830 0892422 ..mwmmwfizz 03292 «$32422 8830.0 865.0 5.590 F 38523 832.00 «83...:22 ..mvovt:zz 231 Eummuvouanaws ammummwomamnwz 8on S .m aw: www.mnrouanaw: okanpcuannm: «98:38:82 8.5495182 mmmmmIFonanams. moamuanauams. oceansuauawz moommuvonauaws. 22mu.ou.mu:w2 .85135182 whmumuanauam: o8~mu.o...mu:ws_ mmormn.ou.mu:ms_ Ermmuvcuauaw: Rowmnsnauam: omwmmnsnannwz wfimmuanauamz mammasuauamz SmNQIFonaaamz mmtmnsuauam: Swamnwonanaw: maimmwouauam: .33 515 am: 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 00.0 00.0 00.0 no.0 8.0 00.0 00.0 00.0 00.0 00.0 no.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 hwé 00.0 0...? 9% 00.0 00.0 00... moé 3.0 05.0 N00 50 00.0 :6 #00 9.6 NNH 9.6 3.? N—L. 3.... 05 V00 26 of? 50.0 (mmmvon. Ewan. 0.0.20 w>00< pad. 041 FOO< em EZN mmhz man—O ..mzmxo Nmaz palm 0010— u0m> anima— ”5‘02 >n_z .th0 5.2.5 0002 Ema 141—“.0 m0<¥ma mam/‘0 9.2m ..oawom. .23. 526.0 09532065 F 33289... 3.050QO .26 c-2269 $9929.35832869.53 -z wt.Em.3_>moo:_m_>m2om_mm.c>c_59:9.Ewan -2.asEmmoUm_m_.vo-mfi_m-_28m-zdoa Ems. F «3.0.6 «53:08 F :99. 2....3qnésoi 82.843 «8.. 5005833 0328 .w 33:58 8 TNT... 9.0.3 «2 £90... .mmcc 2% 889 29.... 6328.89... 8388. 8.2.9.9256 225.. 3.825 F 023328 33 6.5.8:. ..ofiwom. 0.92.65 3 .2300. 0230a o_.m.:_..mco_..mv m 3926 25.62.90 .238. 2.33 2.92.6: Z 5205 52.83.. «8.50 A+0 v 3323033 .o.«$202.38..2-3.8.3882: wmfloficflgégmmooagbwom -z-~. 733 520532900. 79.23 3.8ch... > 2538.39. _ .= .203 cozatomcmz .8950 3288.9 50.9. 238.2%... :28 .F 9:38 m .BEmE .Emzato. o 2.239... @338 325...? 3836 9...: asim mama. ”.8388 32009253 Eon 99. Sago 35:89 .0 Eves—too m: 5829. 520 8655 §§§25u .2538. 28.8% 9.78.... new 898 98 .3598 ..5288.n_2< 30 ..EU owmohgo «waivuzz .828 2.55.0 «$829.2 9.5.50 3ng «.mfivtuzz 3805.0 «93.322 Nomiflzz 38th 0839.0 ..oavtnzz sanctum 03055.0 v.30 va>< 0x. F 5000 vomhwfiwm gsohhxo «8821.22 828% «392422 88..me 232 gmusuaunms. 3 Fumnwonfinams. «Eumusuauaws. «38:58 8.... 83m 515:8: Evnflsnauam: 83m Suauams. Efimuwouanams. massaging Emmmuananamz Summasuauamz owthouanaw: 85 3:58:82 Nemmuanauawz $551385? guanwouauaws. Swans...m|:ws Rammusuauam: ”$318.56; motmuvonauams. nmRmJonEnamz aummnvouamamz Q88 3 3. :ms. gausgunm: monmmnroufiuaw: tmsusnanaws. «88:65:35 oaausnauams. 3d 3.0 3.0 3d Yvd 3.6 3.0 3.0 3.6 3.0 36 3.0 3.0 3.0 3.0 3.0 $6 36 3.0 3.0 3.0 3.0 3.0 36 Yvd :6 3.0 3.0 3.0 3.0 3.0 cod rod cod 3.0 3.0 3.0 vod vod vod v0.0 vod cod 3.0 3.0 3.0 cod 3.0 3.0 3.0 vod 3.0 we... nod mad mod NN... NNé 9.... 9.6 86 mo... 3.... ON... PN... N...‘ mod and 94 0.? cad 92? 2... av... 8.0 0.2.. .....w NF... ...... VN... ,8? 9% N96 VNé ....v. 92... EXS. ”.qu pamumvh wwm. 5000 5.... Fwaa .225 .23 ms. .omzN . o5 Fzmm :5. ~89. 68 m“. mEzw 0mm m<~0..w 503m mXO..< 592 me0 $16 8.25.. 2.3005 35> mEoEmm 95.0. v cactmxoaN->EmI 5.-., «.962 8.8. 85.83969. 529.. 9.2.5 .25 <5: ... $em§.8 393......“ 2.898 3.3323 F5. 332.8. 038.0323... 2.28 53.92. $358.36.. 92508. 9.553 an... m 8.6% 58.8.... <20. 83.5.8 F c322...» 3529.... __ 5.8. 5.6.380 m r 9.5m .5205 0. 50.05 @5059 05.00.03: 95:39 .cman.o>oo 8.0.6.8 0.5.5.5 w .359: ..Eofi? t. ..mtoqmcg Eon 055m 0.5.50. \. 2.5m. .oEmo 05.8 F 38055.... < 95.5.80 3.23 mmmcmgxongwm 0.9.3.585 F 9... m 32.. «5358 P 3533. 3:098 339......» o.._o «Equaaoa fiwuwmomuz unwavmxo mwvomvmo .3898an ~39..sz mmwowm>< ammommmo w .ommm ..vn.< 85.86 8398 59.86 5.3.3522 t 5mm» 322 388“... «.mmmmtnzz 35.8.0 285322 68sz 385.0 tkmmtuzz $888 ... wmvowv>< wmuvov>>< «govtuzz 3256 28:33.: 233.0 5385 233 2000:5493... 0800:5800: 0800103000.... 330:5..01002 0003:5860... 2000:5850: mmmmmlwolfilnw! 0090MBH0H002 0003 S .0 00.2 0.000u00u.mu00_2 0020:0398: Svmmusnanawz «36:58:00.... 0030:8800: 305100.050: {30:58:00.2 000815.550: 20033000.... 028151.030: 00000:.95120: 00081030100... 0005:0300? 0020:5800: 850.001.0002 monpmuwomanamz .0000 5 .0 00.2 2000:5860... 3.0 3.0 3‘... 3.0 3.0 :6 :6 0.0.0 3.0 3.0 vvd 3.0 3.0 30.0 3.0 3.0 36 3.0 3.0 3.0 3.0 3.0 3.0 3.0 36 3.0 30.0 mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod vod cod vod cod 006 36 mo... and h v ... mm... mm... 8.9 an; 2... 3.0 we... no... Nr... .1... mmé 9... 00.? on... 9? 0N6 C... no... mwd de 96 ..w... mNé m... mhmOS OmOm 0a.... <0“. 5.045.. ..(ONO..m $300... mhzu. Namzm 00m Nn...m 9.0.x (In—m Nm._.0< 5.... (man—n. m>oo< n....zn. ooxmn. 9...... men..o< Inc 95102... N0 05.00.800.00 600000005000. 00.00. 0.305 0500.00 N ... 000005.000 <21 00.00. 00.00050 N 00000. 00.000.500.000 05.000.500.030. .00. 50.0.0900 .< 003:0 .x0.0500 00000020200 0.050000 N 00.: 00.0000. .. 50008600 F 00000.0x00 000.0 ...0.0000. 00.02.00 020005.90 050006.00 0 000.26 0.030000 0000.. 5.000000 00.000200 0.50.00 .00.0>..00-<20 .0000... 0.0.90 0 00.0000. 0....-.0. F 000000. 5.0000500 300009060005 0.00 0550000. 000.08.050.00 0550000 0 000505 .....5000000 000000. .200. 0.00.000 .052 N .00 505 00000000.. 0.00 0.0... R 2.50. .0500 05.00 N 05.0090 050.050 .0 0000000000 v 0....-0000. 00.0.0000 00.0.0200. 00.00 0.0.. 0.000 000. 00> .0 02.0520 v 3.5.00 0x...000.0.0005>50§m.0550m 80009.0 mvmomwxo muowo Em wdovmmonz Sgwmwm gamma 5.8020050 0.0802422 30008.0 0.890;... 00380.0 0305.0 00:02:22 0.000502 00502:; 0008000.. 0050.00 3.0.00.0 0:00.000 02.00.02 0050082 0.0088422 P .0003: mmmvvmxo 0.802.302 000.300 Nwmooo>< 234 0.3043000: 00... 05~0H.0H.0H00.2 3.0 0.05 S .0 00.2 3.... mod mod mod 3.. 3.0 mm... m<0._.. m>uz Exam 250.5 NmmmOAw mi: 220. 10...... .95 ww>mv x. 53.0... ID2 .0000 30030 0502 0006 .0000000005 .N 000.5000 350.0000 ..0... N 0.0.0.0 00.050 000 $0.05 3000 000.00 .00..000200 000.0 ..0.0000. 5.0000...>. > 000.0 5.00.5 0.00 .> 5.00. :0..0..00:0.. 00.0.... .. 8.02.00 <2”. 5.0000. 0.0.0.0 ... .3505 00.0.00 t. 00000000.. 0.00 00.50 0.05.00. . 2.50. .0500 0.0.00 N0 .3505 .00 2.50. .0500 0.0.00 000.0550 0.030.520 N 000.0 .00.0..0.5 0.0.08.0. 00.029. 0000.0. .000.0. . 00.0500 00000000 0N0 00.... 0.00.00.00.500 0.0... 3.0005. . 0000.0..0 0000020 . .0 000000200 2.50. N 000.0.000...>00020 00: 000000900000 < 05>~0000 0.035.050.3990? .< 05500093000 ... 000.0 .00020 60005300200000 5.000005. . 00.0500 500050.. 0.33 00.0 . 0000.. 000.0.000...w 00550.20 .000.0.000...>000005.o. . .000.0 ..000... 00.0500 N. 05.0...00020 005.700.00.008 .0>.. .0005..0.00..0000000 . 000.0.000...w 00550.20 .05000.0.5 3.. 0.0.0... .0588... 0000902.... .0030. 00.0500 00000000 .0..> 500.050.30.005 005$ No 0005.. 50.0.0 0.30.3.2 .03....0 0.00.5422 0.0.80.0 «0000000 0.900.322 0.. {00.0 5.00.8.0 .08....uzz 000.....0 {0000.20 0.00.0.0 00089.0 3.000.380 mvaomxo 05800.00 ...0 3.9.0 00 .3320 or 529.0 mmhmmvmo ..o 9009.0 ...mmvm $0... mwmvEO 0. BIG 0m._+_.n_0_ .2. “.0. "00000 000.020.... ...mim 5005005 "n 030... 2.80052000m 236 .00.0u......0u002 N0000|.....0u002 .0. 3.43000... ~.0~0.....|.0|002 0008:5500: 00....1...n.0....0.... 0800......0u002 . . 00:80:00.2 0005.31.05? 0.051.010-002 5000:8050... 8000......0n00... 0008:8800... 0.000u......0:00_2 000......0:.0..00.2 0008:5000... 0.30:...n0n002 ~0000:.....0..00: v0..0:...|.0u00s. 0300:5050... 3000:2393... 0800......01002 80.01.010-00... .0..0I......0..002 300013030... ~0000..on.0n002 30.0....u.0..002 bod ...md hmd hm... hm... hm... hmd hm... hmd hmd hmd no.0 um... um... no... hm... hm... hmd ......o find hmd hm... nmd ..md 5...... had um... Nod Nod No.0 No.0 N90 N06 Nod Nod No.0 Nod Nod Nod No.0 No.0 No.0 No.0 3.... ..od 3... 5.0 3.0 3.0 5.0 5.0 56 ..od .0... mm... 8.0 mm... vu... cad or... m... no... m... 3... 8.0 3.0 cm... 9.... mad or... mo... no... mod and NP... mwd mm... 5.... 5N. omd m... Fam< m_mv.....z 0.000 0080. ....) 20>... 20>: 3.3... S: 1044. 350(5— SRNOOE N900... 0.01 020.0 ...0... 00...... ..x02... Pam/‘10 m10< N000 010 m 5.200 .01.. ..2.x< ._ wIm< Nwom ..00000. 0005.8 0550. . 50.0.0 0505.. 05.5.50 0.0.. 02.0.05 0..00.0 5 .0000000 0000 05.000200 .00.. 0000.. .0 .900. 00.0.... .. 003. 500.50....000 000002.00 0.0.0.020 0588.0 .3069. 00...... .0 000.080.8580... 000000.03... .00 0000.00.80.09... 0:00.. .0..000... 02.00000» N .00505 2.50. 000.0000. 0.03.8...006056..0 5030. 002.00 050.00.00.00... 00000000 00.0.0. .0..> 000.050.300.035 0.53 v ......000. 0....-.6. ..< .3505 2.50. .. 00000090500 00500.0 0 0005.. 0005.. 50.0.0 00.02.00-502? 3.000.... 0.0.0... 000058.... N 0.00 ..o.00. 530.0 05.50.0005 00.00. .0..000 .< 5.0.00000... .200. 530.0 0300.000: 000000203000 00.55.300.350 ... 000.0 00020 65005350000000 ..00 E0500 00005.0 0200000030.. v .200. :0..0..00:0.. 05.02.00 . 65.03000. 00000035 050.. . 50.0.0 0505.. 0.00 0.05.0. .5200 .000.0 000... ..>. 350.005.053.000 00me 50.0.0 0505.. 0...0. N0 0.030 00.0.20 ..00 . €0.00. :0..0..00:0.. 0505.. 5050.0 0000000. 50.00 00.0. .900. 0000000. 50.00 02.000 .. 0005.. 50000 00.0000. 5.3902005 -0050. . 0.0.0... 0000050800.... 2.08.. so. . 5.8 0.500005. 05.15.0059. .0 ...050 .5030. E00 N 0535.0 0500.500 5.050000 m. 53.9.0 vvonmv>>< 00.00.35 0.08.0.0 0000.50 0.0.0024? 0. 3.0.0 90000... 0.003.322 ~.00~v..|s.2 00000.06 00000.00 0 ...: .00 .000... .00. N. ...0... .422 0000.00 .0.....0>< 03020.0 Ndwo ..m FIZZ ...Nmmmooz< 308000080 cacao—hm .. 0.39.2 0mm 5.9.0 0.2.0.00 0020.00 ..000...|.22 237 8.8:58:32 «88:58:32 808:58:32 «85:58:32 5:0:58:32 . 5 5:58:32 805:58:32 508:5:.m:32 055:58:32 085:58:32 880:58:32 508:58:32 8.5:58:32 805:58:32 .88:58:32 085:5:.0:32 85:58:32 858:58:32 088:58:32 88:58:32 808:58:32 088:5:.m:32 558:58:32 8 8:58:32 808:5:.0:32 hmd nmd hmd 50.0 hmd ~06 50.0 50.0 50.0 50.0 50.0 50.0 50.0 50.0 50.0 50.0 #00 no.0 50.0 50.0 00.0 50.0 50.0 50.0 no.0 mod mod mod 00.0 00.0 00.0 mod mod mod mod No.0 «0.0 No.0 No.0 No.0 No.0 No.0 No.0 No.0 No.0 No.0 mod No.0 No.0 «0.0 had .....w 8.0 0.; 8.0 006 2.... 5.... 00.0 mm... 00.0 or.—. 5.... N00 00.0 00.0 «N... 90.0 N...F 005 N06 50.0 00.0 50% no.0 (am—20 Kuhn. .0012 ...0: 51:0 mewou. Fur—.2 wa> 3&2: Exam... ....nEZ. 0.: 1000 92.—0m «(“502 5.73 Nm 8000050005805. 3 0000.20.02.20... 5.000. c 00.2.. 000.0. 0.80.0 .00.. 5 0570000000005 2050000200 .0..000. AN 0.00 00.0025. 0 50.00025 000.00 3500.00.60. ... 0.0.0.0 0.0.0.3 0550.0 0502.0 00000.0 .0. P 00.00.00 050.00 09.0? .0 0.0.0.5003 05.0 F $555.03 0005020200 Io. 00050.0 F 05.0 50.0.0 05050 .000.0.0 02000000. 0520 N .0 000.0 0550.50 500.00 5.00.50.05.00 0.000 0200. .000.00.0>00 000 05.0 0.200.000.0000. 0000.. 00000000 00..., 0.0.0030. 900.055.00.020: 000.050.0000 00.000000 2.000002055030000 ..0... 5 20.000025 05.0000 00:00. .0 00.005 20.... < 822.0 0.058505060 .208. 00550.08 $150000 5 .000me Nhnvwmw>< 8885 0.80.582 0800282 0888:22 0888.2 0.80522 0.88::22 8088.0 808000 0.080222 038.022 82.50 0.88::22 0.88222 05580.0 E8050 0508.0 .5256 88800 885.0 m wmovmxo .25:th 238 «36:65:85. mmwmmusuumuaws. Smumurouauam: gamupouauam: mwmumusuanam: moowmnsnfiuaw: moonmuauaunmz mmmNmIEJmnaws. mavwmusuauamz «85:35:82 «otmusuauaws. mtmmuauauamz mmvuausnanaws. mmamusuauam: ~68n8u512m2 Bagusuauams. Summusnauams. Rzmnponauams. omzmnsnanams. atmusnaunws. 30315155? 8833982 Emmmusnanams. opmmmnronanam: Semiramis? unmansnanamz msg#95132 Emmmususnnms no.0 nmd hmd bmd hmd hmd 500 had hmd nmd hmd find .36 hmd hmd de hmd hmd hmd hmd nmd nod nod find .36 find bod hmd 3.0 3.0 vod cod vo.o vo.o 3.0 3.0 3.0 3.0 mod mod mod no.0 mod no.0 mod nod mod no.0 mod mod mod mod mad nod mod mod 8.0 mo._. and Ed cad med 9.... no; 3.0 56 m2. mmd 9.... mad 3.0 m2. wad 39 «ad mad 2.... NF... Saw woé :6 mad 2.... mad m8< mum: ww<4< s. 2.5ch «moo ch E 2N NOOm Nwéw mud. ammo must. nmmomI n_z @803}! Km: men=D< 000< wU 5&0 00>: <30. 2. 5203 5333 5333 m 9:53:00 .038 6823 25:0 m 889. E2262 c9325 F 83%» ..mgou .mfic_.§o.oc_Em t 9:”: 9.30.. :30 cu mEoooEon m 8339 £88 98 053E050 v: £995 59.: can EmEou “EN 55 EmEQm mzmmoamcg omen catacofige .52: N ommEEmSB 3.5; N .28. £22.35. fianchSEégooScva N .905 mczflaEzm @200 m .208 35832238.. ...z... \. mmflmEofléEu 292» new -93 m 63:393ch Bo.2w.m-§mn.>x9u>c custozamofi 3.80.03: 8:35 memo 8833. F .98. 53:35: E858: F .0339 589593 mcEfiEoo EmEou $8.60 new 05 6.5093 c .3508 6280.28“ 0230.03: 2.58 5E8 .mcucofiogé 8 £59— 5E8 923 v mm£>xo£wowu mscoEEESmocovm v 3:039 5266 2.60 5238.9 N 9:383:60 2:8.mfimsom N 023253 ggmcgsmossau .3 88 222333.83 P 03%....on .< @5538 .2 2.58 SE 2.958% F $58 82.5% 3. .28. angers sewage JAEQE 632552 385322 8833. 53930 «69.86 955.0 «838.0 38mg mwumwwmo 33.56 3856 «5321.22 $39.8 tmmkxo 8333.0 hmv —. cmEamm 28382 F9832 «$52422 tooovtnzz 38.0.2152 38:35.2 3856 human—553mm $385 ....omvtnzz Nfiofivuzz v88? :2 8856 239 505:5:.m:82 055:5:.m:82 ”53:58:82 «85:58:82 «88:58:82 808:5:5:82 308:5:.m:82 808:5:5:82 .5~m:5:.m:82 0058:5405? 205:5:5302 058:5:au82 053:5:.m:82 2.05:5:5:82 858:5:.m:82 058:5:.m:82 855:5:.m:82 035:5:.0:82 5.....:5:5:82 «98:58:82 «005:5:5302 805:5:.m:82 508:5:.m:82 585:5:.m:82 ..88:5:.m:82 850:5:5:82 ~00 5:5:.m:82 5005:8882 . 5 5:58:82 hmd hm... km... 3.0 sad sad hmd no.0 nmd nod hmd hmd hmd de wmd hmd nmd sad umd hmd nmd No.0 nmd find 500 sad um... km... 3.0 mod mod mod mod mod mod mod mod mod v0.0 vod 3.0 3.0 3.0 cod #06 3.6 3.0 3.... wed vod cod 3.0 3.0 3.... rod vod we... we... mwd no... 3.... so... :9 3.0 3x? and mad and ..md 9... no... 3.? 00.? 3.... Saw Npé 8.0 5.0 0.2.. 3.0 2.... mad .....v hmd hoe omd and ...S.< ..me wz_mn_DDZ C 50.0.0 50.0.8 0m0>00.0 05029 000.0.0c0.._>...0505E0 F 0005.. 8.0000. 00.300.50.90 0 F 50.0.0 05050 55050.00 < 80500.0 05.000 5.0.350: . 50.2.. 8.0.033 cEERNK. 50.9.. 052... 89.. F 50.2.. 9.02.00 $005 :00 0 Q... .8888. 080.0. 80.88. 00.5020 No 5.96 P 50.9.. 5.00.05... .22 20:20:02.5. N 8050.53.00 0.923.800.5005 3.0000. N .200. 539m 05.5.85 5.0000. ..m 70.5.0 6.900060 m 5005050: 3.2885. 0 00.059. 0.0 .2000 09.8... .920 .< 50050000. . 020... .m c0m05ma00 0000». F 50.0... u0.0.00000.000.0>0 0.0.2.000 ...(0 C 000.0 600.3 .0..005...0..0..000..0 5.080.058 0.050005005009035 .0 000.020 w 0000.0:c0x0 ..000. 0.5.0 00.... 05.000200 N .09....03000. 000.0589... 0500.... 50.0.... .. 00....” 5558 x0558 ...0.0.m0.500 .038 05.02.00 0.5.0 .0 .900. .0062. 0500.000: N 3.00.0.5 0005.. 3.0000. 00.900.50.80 0 09.00 00.00. 00.0.0800 ....mb 50.9.. 8.25 .6. <2» ... 000.0556. <20 0....» or 9.2.5.3.... 0 $30. .200. 09.9.9.0 05.00.02. 05:0am $000.09. 000.5,. .P 50.0590: 0:05:05... 000500.030“. Io< wNmNmbmm NNwRoum $03222 0095.0 3802:22 240 5»5:5:8:82 85:58:82 «55:58:82 «85:58:82 5005:5840? 085:5:8:82 0005:8882 585:5:.m:82 85:58:82 mmomm:5:.m:82 535:5:8:82 055:5:.m:82 055:5:.m:82 055:5:.m:82 805:5:.m:82 0505:5853 88:58:82 085:5:5:82 5:5:5:.m:82 00 55:58:82 505:5:.m:82 585:5:.m:82 5005:8882 855:5:.m:82 505:5:.m:82 555:5:8:82 «85:58:82 555:5:8:82 55:58:82 hmd hmd de sad had sad sad hmd smd hmd hmd wmd had had had 36 hmd hmd nmd hmd hmd wmd no.0 smd .56 bad hmd no.0 smd mod mod mod mod mod mod mod mod mod 3.0 3.0 3.0 3.0 3.0 3.0 3.0 cod 36 cod 3.0 3.0 cod 3.0 3.0 3.0 cod wed v0.0 vod and no... 3;. 50H :# wad Ear mmd de mwd Pad 95 no.9 :6 86 3.6 3.? Nwé 8.0 Ed 9% vmd 9; mad :6 had 50.? cad and 0.52 550 .2546 (002 Pan—(mu. wa<02£0Eo5E0 _. 00050. 8.0000. 00.080.5088 0 F 508.0 @5050 58050.00 < 80800.0 05.000 0.8.050: P 8.8.. 8.0.088 c§§§ 8.9.. 559.3 88.. F 50.90 8002.8 8005 :00 0 0x: A8383. 0000.0. 0008.80. 00.5020 8 50.6 r 50.9.. 8:00.826 52 58.88.8058 N 0005588900 0.0>:.>0_oc0oz0moc0 8.0000. N .200. 5308 0575.85 8.0000. ..m 70:50 6.900080 m 500.880: @2888. 5 50659. 0.8 7.5050 2059.. .920 2 000050003 _ 000.0 .m 000050000 800$ F 508.0 008.800.000.08 0.030000 .93 0 000.0 60020 8500535880030 5.0820080 0.050005005000002. .0 000.050 _. 0000.02.98 5000. 05.0 00.5 _. 05.000200 N 00388000. 0000080020 0500.». 508.0 P 057m 5808 x2088 P80089500 8:90 05.02.00 0050 .v 8.00. 80.02. 0502000... N 880.05 00050. .2080. 00.080.50.90 0 0950 .880. 00.0.0800 50.5 50.90 550:... x8 (.2: ... 80.0% <20 005 S c2305.; 0 Emmy. 880. 098208 05808:: 05:08 #000801 $802 09.5 .F 599020.. 3:05:53 308592500 :22 0.00:2:22 0.2.5::22 02 E8 3052:22 5850.0 «8830 9850.0 o~55>< 00058.0 80550.0 0.009222 05 5 E8 38300 0.009222 08850 835.0 8850.0 P 5588 5.855222 2550.0 0.2.3::22 5.0053 wwwmohxo ...th@ wm>< ..Nommbmm Nwwkoum 58:32: 0095.0 5.803.322 240 0005:38802 038:38502 .020:.o:.m::02 0030:58802 0008:58502 0300:3880... 806:58::02 808:...:.m::02 8.5:...8502 9 50:38:82 0030:58802 0020:88802 305:8:8:a02 .0000:.c:.m::02 8000:38302 .005J08::02 «88:38:82 0900:58302 0~0~0:.o:.m::02 8000:38502 t0.0:.o:.m::02 3000:.o:.0::02 0800:8880: «88:58:52 B.8:.o:.m::02 um... 2:90 nmd hmd hmd hmd had .36 3.0 2nd find 206 2b.... sad sad hmd hmd hmd nmd No.0 no.0 de hmd hmd nmd mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod mod no.0 mod mod mod mod mod mod mod 9.? mm... 9.... ON... NNé omd wad 5.0 vwd 92? 8.0 so... omd 0N... moé of? mad NNé Nwd N... or; 9% mad mud :é E...0<.mw<>> 000.0...c20 <21.-_2c0cao.02.. 022 650 $.20 .000....0....9...0 600030000000 .0530... 0002.00000..0.0=0.0E 0:05.00... «.522 N 027E030. 00.089. 0000005 .0..> 0.00.00.00.02... 028$ $0.02 .02... P 0.8.0 ... .00c0..c0 0:00 00.500260 .29. 00000. .0 .200. .0202: 90.0 . 000.8826 m P... 0.0.. .F 50.30002. .300 9...-.. 9.00.. 9.2.. 0.0. 2.6205 5.3 . 08.02.8209... 2.E$§_0.20.8rz-..o: Ewe... F 000...:20 05.00.230.309... m 5.56 m. 000...... 00.0.30. c0..0.2.0..000..0-.>..0...00.2. 2..0....0000-.0=0 mmomhmxo 9.08.0 «0.00282 800002. 03.33... 63.30 .8092 .0380 380::22 cummmem 0.003282 0.0.9.222 0.9.0:.22 3.80530 00830.0 2:002:22 00003.0 wonmnnxo momenta 0025.0 0832:22 5.80230 3.0.38.0 00005.0 0028.0 241 Smmmnslanamz Bo ”33:38:35. 3.0 mod mod mwd 3.0 Xmmww _‘._m>:m $95. «.83 cozmmcgm 2838.30 28 .B v 9:523. var—$90 2355.0 242