IMMUNOSTIMULATING EFFECTS OF A COMMERCIAL FEED SUPPLEMENT IN DAIRY HEIFERS VACCINATED WITH A STAPHYLOCOCCUS AUREUS BACTERIN VALERIE JOANNE EUBANKS

IMMUNOSTIMULATING EFFECTS OF A COMMERCIAL FEED SUPPLEMENT IN DAIRY HEIFERS VACCINATED WITH A STAPHYLOCOCCUS AUREUS BACTERIN by VALERIE JOANNE EUBANKS (Under the direction of Stephen C. Nickerson) ABSTRACT The purpose of this study was to evaluate the effect of a general immunostimulant feed supplement on amplifying dairy heifers immune response to Staphylococcus aureus vaccination. Blood and mammary immune parameters were measured. Serum anti-s.aureus titers were similar between supplement-treated heifers and unsupplemented controls indicating no beneficial effect of the feed supplement. Prevalence of overall intramammary infections (86% of heifers, 58% of quarters) and S. aureus intramammary infections (62.1% of heifers, 25.3% of quarters) was similar between treated heifers and unsupplemented controls. Musculoskeletal growth was also similar between treated and control heifers. However, L-selectin mrna expression, phagocytic activity, and reactive oxygen species production was greater in treated heifers compared to controls, suggesting enhanced immunity in supplemented heifers prior to calving. Furthermore, treated heifers exhibited decreased prevalence of postpartum mastitis, decreased somatic cell counts, and increased milk yield compared to control heifers, although these differences were not statistically significant. INDEX WORDS: Nutritional immunostimulation, dairy heifers, mastitis, Staphylococcus aureus

IMMUNOSTIMULATING EFFECTS OF A COMMERCIAL FEED SUPPLEMENT IN DAIRY HEIFERS VACCINATED WITH A STAPHYLOCOCCUS AUREUS BACTERIN by VALERIE JOANNE EUBANKS BS, Clemson University, 2010 A Thesis Submitted to the Graduate Faculty of The University of Georgia in Partial Fulfillment of the Requirements for the Degree MASTER OF SCIENCE ATHENS, GEORGIA 2012

2012 Valerie Eubanks All Rights Reserved

IMMUNOSTIMULATING EFFECTS OF A COMMERCIAL FEED SUPPLEMENT IN DAIRY HEIFERS VACCINATED WITH A STAPHYLOCOCCUS AUREUS BACTERIN by VALERIE JOANNE EUBANKS Major Professor: Committee: Stephen C. Nickerson David J. Hurley Roy D. Berghaus Electronic Version Approved: Maureen Grasso Dean of the Graduate School The University of Georgia May 2012

DEDICATION I would like to dedicate this thesis to my family and fiancé for all the love and support they have given me over the past couple of years. My parents, Joe and Deborah Eubanks, and grandparents have encouraged and supported every decision I have made. I know that without them, I could not have achieved my goals thus far. My fiancé, Matt Ryman, has shown me such tremendous dedication and love, which have helped me along this path, and I know I will forever have my greatest fan beside me. iv

ACKNOWLEDGEMENTS I would like to acknowledge Dr. Stephen C. Nickerson for his continued guidance and mentorship as my major professor during my time at University of Georgia. I also would like to acknowledge my committee members, Dr. David J. Hurley, and Dr. Roy D. Berghaus, for their willingness to assist me during my tenure here as well as offer suggestions to improve my research and writing skills. I also thank Felicia Kautz for her assistance in the lab and her help with sample collection. v

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS...v LIST OF TABLES...x LIST OF FIGURES...xv CHAPTER 1 LITERATURE REVIEW...1 Introduction...1 Mastitis and its causes...4 Prevalence of mastitis in heifers...6 Mastitis prevention and control techniques...8 Mammary gland immune defenses...9 Influence of infection on SCC...12 Effect of S. aureus on mammary tissues...12 Heifer mastitis prevention and control techniques...14 Vaccination...21 Heifer growth...27 Nutritional immunostimulation...28 Objectives...35 2 MATERIALS AND METHODS...36 Overview...36 Initiation of trial...37 vi

Growth measurements and feeding...39 Blood collection and analysis...43 Determination of L-selectin and IL-8R mrna expression...44 Blood mononuclear cell and neutrophil preparation for phagocytic activity and reactive oxygen species (ROS) production...45 Vaccination and antibody titers...47 Teat swab and mammary section collection and analysis...50 Total leukocyte counts...59 Differential leukocyte counts...59 Determination of new IMI and TCI, chronic IMI and TCI, spontaneous cures, and success of DCT...62 Postpartum data collection...63 3 RESULTS AND DISCUSSION...66 Effects of a general immunostimulant feed supplement on heifer growth as measured by average daily gain, body weight, and hip height...66 Effects of a general immunostimulant feed supplement on the immune system as measured by L-selectin mrna and IL-8R mrna expression...76 Effects of a general immunostimulant feed supplement on vaccine efficacy as measured by anti-s. aureus total IgG titers...80 Effects of a general immunostimulant feed supplement on neutrophil and monocyte phagocytosis and reactive oxygen species production...84 Effects of a general immunostimulant feed supplement on prevalence of mastitis in nulligravid and primigravid heifers...91 vii

Effects of a general immunostimulant feed supplement on colony-forming unit (CFU) score of bacterial isolates and classification of infection...127 Effects of general immunostimulant feed supplement on new IMI rate, TCI rate, chronicity of infection, response to antibiotic therapy, and spontaneous cure rate...129 Effects of a general immunostimulant feed supplement on SCC of mammary secretions...131 Effects of a general immunostimulant feed supplement on differential leukocyte counts in mammary secretions...136 Effects of a general immunostimulant feed supplement on prevalence of clinical mastitis...144 Effects of a general immunostimulant feed supplement on postpartum prevalence of mastitis, SCC, and milk production...148 4 CONCLUSIONS...158 REFERENCES...162 APPENDICES...170 1 Initiation dates and distribution of treated and control heifers within 8 age groups...170 2 Mean comparison of anti-s. aureus total IgG titers among treated and control heifers in group 1...171 3 Mean comparison of anti-s. aureus total IgG titers among treated and control heifers in group 2...172 viii

4 Mean comparison of anti-s. aureus total IgG titers among treated and control heifers in group 3...173 5 Mean comparison of anti-s. aureus total IgG titers among treated and control heifers in group 4...174 6 Mean comparison of anti-s. aureus total IgG titers among treated and control heifers in group 5...174 7 Distribution of new IMI and chronic IMI by total number of quarters and quarters/heifer in mammary secretion samples collected from treated heifers...175 8 Distribution of new IMI and chronic IMI by total number of quarters and quarters/heifer in mammary secretion samples collected from control heifers...175 9 Distribution of new TCI and chronic TCI by total number of quarters and quarters/heifer in teat keratin samples collected from treated heifers...175 10 Distribution of new TCI and chronic TCI by total number of quarters and quarters/heifer in teat keratin samples collected from control heifers...175 ix

LIST OF TABLES Table 1. Abbreviated list of OmniGen-AF components...41 Table 2. Wheat silage nutrient analysis...42 Page Table 3. Heifer grain mix ration (lb/ton)...42 Table 4. Dry cow grain mix ration (lb/ton)...43 Table 5. Multivariable generalized estimating equations (GEE) linear regression model for the prediction of body weight (kg) in Holstein heifers that received a dietary supplement (treated, n = 28) or that served as unsupplemented controls (n = 26)...70 Table 6. Multivariable generalized estimating equations (GEE) linear regression model for the prediction of hip height (cm) in Holstein heifers that received a dietary supplement (treated, n = 28) or that served as unsupplemented controls (n = 26)...74 Table 7. Independent comparisons of percent fluorescence of surface-bound and internalized fluorescein-labeled E. coli and S. aureus by blood neutrophils and monocytes collected from control and treated heifers on days 0, 30, and 60...86 Table 8. Univariate comparisons of any case of IMI diagnosed in mammary secretions from treated and control heifers...93 Table 9. Univariate comparisons of Staphylococcus aureus IMI diagnosed in mammary secretions from treated and control heifers...94 x

Table 10. Univariate comparisons of coagulase-negative staphylococci (CNS) IMI diagnosed in mammary secretions from treated and control heifers...95 Table 11. Univariate comparisons of any case of Staphylococcus chromogenes IMI diagnosed in mammary secretions from treated and control heifers...97 Table 12. Univariate comparisons of any case of Staphylococcus hyicus IMI diagnosed in mammary secretions from treated and control heifers...97 Table 13. Univariate comparisons of Staphylococcus simulans IMI diagnosed in mammary secretions from treated and control heifers...97 Table 14. Univariate comparisons of Streptococcus spp. IMI diagnosed in mammary secretions from treated and control heifers...98 Table 15. Univariate comparisons of any case of TCI diagnosed in treated and control heifers...99 Table 16. Univariate comparisons of Staphylococcus aureus TCI diagnosed in treated and control heifers...100 Table 17. Univariate comparisons of coagulase-negative staphylococci (CNS) TCI diagnosed treated and control heifers...101 Table 18. Univariate comparisons of Staphylococcus chromogenes TCI diagnosed in treated and control heifers...101 Table 19. Univariate comparisons of Staphylococcus hyicus TCI diagnosed in treated and control heifers...102 Table 20. Univariate comparisons of Staphylococcus simulans TCI diagnosed in treated and control heifers...102 xi

Table 21. Univariate comparisons of Streptococcus spp. TCI diagnosed in treated and control heifers...102 Table 22. Multivariable logistic regression results for the prediction of IMI based on quarter mammary secretions from treated and control heifers...108 Table 23. Multivariable logistic regression results for the prediction of Staphylococcus aureus IMI based on quarter mammary secretions from treated and control heifers...109 Table 24. Multivariable logistic regression results for the prediction of coagulase-negative staphylococci (CNS) IMI based on quarter mammary secretions from treated and control heifers...109 Table 25. Multivariable logistic regression results for the prediction of Staphylococcus hyicus IMI based on quarter mammary secretions from treated and control heifers...111 Table 26. Multivariable logistic regression results for the prediction of Staphylococcus simulans IMI based on quarter mammary secretions from treated and control heifers...111 Table 27. Distribution of teat score values among quarters in group 1 from August 2010...113 Table 28. Distribution of teat score values among quarters in group 2 from May 2011...113 Table 29. Distribution of teat score values among quarters in group 3 from May 2011...114 Table 30. Distribution of teat score values among quarters in group 4 from September 2011...114 xii

Table 31. Distribution of teat score values among quarters in group 1 from September 2010...115 Table 32. Distribution of mean teat score values among quarters in groups 2, 3, and 4 in the month succeeding fly treatment...115 Table 33. Comparison of teat score values among quarters of treated and control heifers overall and in 4 different heifer groups...117 Table 34. Multivariable logistic regression results for the prediction of Staphylococcus aureus TCI based on teat canal keratin swabs from treated and control heifers...123 Table 35. Multivariable logistic regression results for the prediction of coagulase-negative staphylococci (CNS) TCI based on teat canal keratin swabs from treated and control heifers...123 Table 36. Multivariable logistic regression results for the prediction of coagulase-positive staphylococci (CPS) TCI based on teat canal keratin swabs from treated and control heifers...124 Table 37. Multivariable logistic regression results for the prediction of Staphylococcus chromogenes TCI based on teat canal keratin swabs from treated and control heifers...125 Table 38. Multivariable logistic regression results for the prediction of Staphylococcus hyicus based on teat canal keratin swabs from treated and control heifers...126 Table 39. Multivariable logistic regression results for the prediction of Staphylococcus simulans TCI based on teat canal keratin swabs from treated and control heifers...126 xiii

Table 40. Distribution of new cases, recurring cases, Staphylococcus aureus cases, and front quarter cases of clinical mastitis among quarters from 4 groups of heifers...147 Table 41. Distribution of new cases, recurring cases, Staphylococcus aureus cases, and front quarter cases of clinical mastitis among heifers from 4 groups...147 Table 42. Prepartum (30 to 60 d) somatic cell counts (SCC) and microbiology of mammary secretions collected from supplement-fed (treated) heifers...149 Table 43. Prepartum (30 to 60 d) somatic cell counts (SCC) and microbiology of mammary secretions collected from unsupplemented (control) heifers...150 Table 44. Postpartum somatic cell counts (SCC) and microbiology of mammary secretions collected from supplemented-fed (treated) heifers...152 Table 45. Postpartum somatic cell counts (SCC) and microbiology of mammary secretions collected from unsupplemented (control) heifers...153 xiv

LIST OF FIGURES Page Figure 1. Head lock stanchion system used to feed heifers individually...38 Figure 2. Hip height measurement taken while in head lock...39 Figure 3. Heart girth measurement taken while in head lock....39 Figure 4. Clipping neck prior to blood collection...44 Figure 5. Sanitization of site with 70% alcohol cotton...44 Figure 6. Inserting needle into jugular vein...44 Figure 7. Visible knot in right rear leg after vaccination...48 Figure 8. No scabs on teats (Score 1)...51 Figure 9. Older scabs on teat (Score 2)...51 Figure 10. Fresh scabs on teat (Score 3)...51 Figure 11. Insertion of calcium alginate swab into the teat canal to collect keratin...52 Figure 12. Thick, glue-like secretion from an uninfected quarter (bottom) and thin, watery secretion from and infected quarter (top)...53 Figure 13. Turbid secretion from an infected quarter (left) and clinical secretion from an infected quarter illustrating the settling of clots and flakes (right)...54 Figure 14. Turbid, white secretion from an infected quarter (left) and amber secretion from an uninfected quarter (right)....54 Figure 15. TSA 5% blood agar plate illustrating a heifer with 3 S. aureus-infected quarters (RF, LF, RR) identified by the β-hemolysis and 1 CNS-infected quarters (LR)...55 xv

Figure 16. Results of coagulase test showing a positive result (solidifying of plasma) exhibited by inverted tube...56 Figure 17. Results of MSA test showing 1 MSA+ sample (LR) and 3 MSA- samples (RF, LF, RR)...56 Figure 18. API Staph identification test showing a positive test for S. hyicus...57 Figure 19. Results of Slidex test showing a positive agglutination test in 6/G denoting an ungroupable Streptococcus spp. An API 20 Strep test was subsequently conducted to confirm Strep. uberis...58 Figure 20. Differential leukocyte stain illustrating, LM=Lymphocyte, PMN=Neutrophil, and MA=Macrophage...60 Figure 21. Differential leukocyte staining illustrating cells from an infected quarter demonstrating a predominance of neutrophils...61 Figure 22. Differential leukocyte stain illustrating cells from an uninfected quarter demonstrating a predominance of macrophages...61 Figure 23. Infusing dry cow product into the right rear quarter of heifer using the partial insertion technique...64 Figure 24. Delvotest illustrating positive results for the RF quarter and negative results for the control, composite, and LF, LR, and RR quarters...65 Figure 25. Average daily gains (ADG) for groups 1 to 5 and overall ADG between treated and control heifers...67 Figure 26. Overall Lowess curves for heifer body weight across all groups for control and treated heifers....71 xvi

Figure 27. Lowess curves for heifer body weight by group for control and treated heifers...71 Figure 28. Lowess curves for heifer body weight, by group, for treated and control heifers regardless of treatment...72 Figure 29. Overall Lowess curves for heifer body weight across all groups for control and treated heifers...75 Figure 30. Lowess curves for heifer height, by group, for control and treated heifers...75 Figure 31. Lowess curves for heifer height, by group, for treated and control heifers regardless of treatment...76 Figure 32. Mean L-selectin mrna expression of blood leukocytes in treated and control heifers by sample time (mo) for heifer groups 1, 2, and 3...77 Figure 33. Mean IL-8R mrna expression of blood leukocytes in treated and control heifers by sample time (mo) for heifer groups 1, 2, and 3...79 Figure 34. Comparison of mean anti-s. aureus total IgG titers, by group, prior to and following initial, booster, and semi-annual vaccination...83 Figure 35. Comparison of mean anti-s. aureus total IgG titers among treated and control heifers, regardless of group, prior to (month 0) and following initial, booster, and semi-annual vaccination...84 Figure 36. Comparison of reactive oxygen species (ROS) production (±SD) by neutrophils collected from group 6 and 7 heifers (treated, n=10 and control, n=10) after in vitro stimulation at d 30...87 xvii

Figure 37. Comparison of reactive oxygen species (ROS) production (±SD) by neutrophils collected from groups 6 and 7 heifers (treated, n=10 and control, n=10) after in vitro stimulation 60 d after beginning trial...89 Figure 38. Comparison of reactive oxygen species (ROS) production (±SD) by neutrophils collected from group 8 heifers (treated, n=4 and control, n=5) after in vitro stimulation at d 0...90 Figure 39. Comparison of reactive oxygen species (ROS) production (±SD) by neutrophils collected from group 8 heifers (treated, n=4 and control, n=5) after in vitro stimulation 30 d after beginning trial...91 Figure 40. Comparison of reactive oxygen species (ROS) production (±SD) by neutrophils collected from group 8 heifers (treated, n=4 and control, n=5) after in vitro stimulation at d 60...90 Figure 41. Overall prevalence (%) of IMI by infection status and individual bacterial isolate...105 Figure 42. Distribution of infection status by treatment group among quarters...106 Figure 43. Distribution of IMI in treated and control heifers by bacterial isolate...106 Figure 44. Presence of horn flies on teats...112 Figure 45. Overall prevalence (%) of TCI by infection status and individual bacterial isolate...120 Figure 46. Distribution of infection status by treatment groups among quarters...120 Figure 47. Distribution of TCI in treated and control heifers by bacterial isolate...121 Figure 48. Percentage of total isolates by colony-forming unit (CFU) score between treated (n = 267 quarter samples) and control (n = 214 quarter samples) heifers...128 xviii

Figure 49. Percentage of total Staphylococcus aureus isolates by colony-forming unit (CFU) score between treated (n = 89 quarter samples) and control (n = 52 quarter samples) heifers...128 Figure 50. Percentage of total isolates, excluding Staphylococcus aureus, by CFU score between treated (n = 74 quarter samples) and control (n = 80 quarter samples) heifers...129 Figure 51. Average number of new IMI and chronic IMI between treatment groups...130 Figure 52. Average number of new TCI and chronic TCI between treatment groups...131 Figure 53. Comparison of somatic cell counts (mean ± SD) of quarter mammary secretions from infected and uninfected quarters regardless of treatment group (n = 332 total samples)...133 Figure 54. Comparison of somatic cell counts (mean ± SD) of quarter mammary secretions among treatment groups based on infection status (n = 332 total samples)...133 Figure 55. Comparison of somatic cell counts (mean ± SD) of quarter mammary secretions infected with Staphylococcus aureus, coagulase-positive staphylococci (CPS), and coagulase-negative staphylococci (CNS) among treatment groups (n = 238 total samples)...135 Figure 56. Comparison of somatic cell counts (mean ± SD) of quarter mammary secretions infected with predominant pathogens (excluding Staphylococcus aureus), by bacterial species, among treatment groups (n = 127 total samples)...135 Figure 57. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in quarter mammary secretions from infected and uninfected quarters, regardless of treatment group (n = 283 total samples)...138 xix

Figure 58. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in heifer mammary secretions of treated (n = 185 quarter samples) and control (n = 98 quarter samples) heifers regardless of infection status...138 Figure 59. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in infected mammary secretions of treated (n = 138 quarter samples) and control (n = 76 quarter samples) heifers...139 Figure 60. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in uninfected mammary secretions of treated (n = 47 quarter samples) and control (n = 22 quarter samples) heifers...139 Figure 61. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in Staphylococcus aureus-infected heifer mammary secretions of treated (n = 75 quarter samples) and control (n = 33 quarter samples) heifers....141 Figure 62. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in coagulase-positive staphylococci (CPS) infected mammary secretions of treated (n = 9 quarter samples) and control (n = 3 quarter samples) heifers...141 Figure 63. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in coagulase-negative staphylococci (CNS) infected mammary secretions of treated (n = 45 quarter samples) and control (n = 32 quarter samples) heifers...142 Figure 64. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in Staphylococcus hyicus-infected heifer mammary secretions of treated (n = 27 quarter samples) and control (n = 14 quarter samples) heifers...142 xx

Figure 65. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in Staphylococcus chromogenes-infected mammary secretions of treated (n = 9 quarter samples) and control (n = 15 quarter samples) heifers...143 Figure 66. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in Streptococcus spp.-infected heifer mammary secretions of treated (n = 4 quarter samples) and control (n = 10 quarter samples) heifers...143 Figure 67. Ropy, discolored secretion being expressed from heifer infected with Staphylococcus aureus...144 Figure 68. Secretion from quarter infected with Staphylococcus aureus...144 Figure 69. Swollen right rear quarter and normal left rear both infected with Staphylococcus aureus...144 Figure 70. Comparison of daily milk yield through 14 d in milk (DIM)...156 xxi

CHAPTER 1 LITERATURE REVIEW Introduction Mastitis, defined as an inflammation of the mammary gland, is a disease that costs the dairy industry 2 billion dollars a year due to milk loss from intramammary infections (IMI), discarded milk, and culling of infected animals (Nickerson and Philpot, 2000). Both clinical and subclinical IMI account for these economic losses. Clinical mastitis can be visually identified through abnormalities in milk, such as clots and flakes, as well as in abnormalities of the infected mammary gland itself, such as swelling, induration, heat, and redness. Subclinical mastitis cannot be visually diagnosed but is identified based on microbial culture, increased somatic cell counts (SCC), and a decrease in milk production. The presence of various mastitis-causing pathogens including staphylococci, streptococci, and coliforms results in damage to the milksecreting cells of the mammary gland, thereby compromising the health and productive capacity of the dairy animal. Once infected, mastitis may remain indefinitely unless treated. Factors at stake if causative pathogens are not identified and eradicated in the first-calf heifer include an elevated SCC throughout the first and succeeding lactations, compromised health of mammary tissue, decreased milk yield over the animal s productive life, and most importantly, reduced longevity of the dairy animal (Piepers et al., 2009). Programs in place to decrease the incidence of mastitis are focused on lactating and dry cows, specifically primiparous and multiparous animals. Nulligravid and bred heifers are not usually considered in mastitis prevention and control programs because they are thought to be 1

free of damaging IMI. However, studies have shown that the prevalence of infection in unbred heifers as well as bred heifers prior to calving ranges between 28.9% and 75% of the heifers tested (reviewed by Fox, 2009). In another study, it was reported that 96.9% of heifers and 74.6% of quarters were found to harbor IMI, quarters infected with Staphylococcus aureus were found to have higher SCC compared to quarters infected with coagulase-negative staphylococci (CNS) or the environmental streptococci (Nickerson et al., 1995). S. aureus is a major mastitis-causing pathogen that has been found to infect dairy heifers (Fox et al., 1995; Nickerson et al., 1995). This pathogen has also been identified as one of the most detrimental to milk-secreting tissues, which are in the process of developing in the growing heifer. Tissue from S. aureus-infected animals was found to exhibit decreased alveolar luminal area and increased interalveolar stromal formation, resulting in a decreased quantity of secretory tissue (Trinidad et al., 1990a). Additionally, leukocyte infiltration into S. aureus-infected tissues was markedly higher than in uninfected tissues. Abscesses were also identified in S. aureusinfected tissues, resulting in scar tissue formation. This scar tissue not only leads to a reduction in the capability to produce milk at an optimum level, but also provides a safe haven for bacteria to avoid antibiotic therapy (Sordillo et al., 1989). Antibiotic therapy is relied upon to decrease the number of mastitis cases through 1) the treatment of existing infections, e.g., lactating cow therapy, and 2) the prevention of new IMI, e.g., dry cow therapy. Antibiotic therapy is heavily used in lactating cows to treat clinical cases of mastitis, and in dry cows to prevent new infections, as well as to cure acute infections. Both lactating and dry cow therapy may also be used during gestation and prior to calving in heifers to cure existing infections, which ideally results in a lower SCC and higher milk yield in first-calf heifers. Research has shown that heifers are much more responsive to antibiotic therapy than 2

lactating cows when treated for cases of S. aureus and CNS mastitis (Trinidad et al., 1990c). For example, heifers treated with nonlactating antimicrobial agents 60 days prior to calving exhibited lower incidences of S. aureus at calving than heifers not receiving therapy (Trinidad et al., 1990d). Although antibiotic use is beneficial in many cases, the goal of disease management is prevention. Vaccines have been studied and used to address this goal. Mastitis vaccines are used to 1) prevent new IMI, 2) reduce the severity of existing IMI, and 3) aid in curing IMI in association with antibiotic therapy. In mature cows, vaccination against S. aureus resulted in elevated immunoglobulin levels and lower infection rates when animals were experimentally challenged; however, a less than optimum degree of protection was achieved (Nickerson et al., 1993). In contrast, in heifers a commercial vaccine directed against S. aureus was found to reduce mastitis by 45 to 60% at the time of calving, and reduce SCC by 50% (Nickerson et al., 1999, 2009). Although vaccination was more successful in heifers than in older animals, a reduction in heifer mastitis approaching 90 to 100% at calving would be more desirable. One option to enhance response to vaccination may be through oral immunostimulation by administration of a feed supplement. Increasing public awareness and displeasure with the use of antibiotics in food animals, as well as the increasing need to freshen heifers with low SCC to meet the proposed legal limit of 400,000 cells/ml are motivating factors for improving mammary health. Thus, nutritional supplementation with immunostimulants has been considered to improve immune function in dairy cattle and decrease incidence of mastitis. For example, antioxidants have been identified as important immunomodulatory nutrients to protect mammary epithelial cell membranes and immune cells (Sordillo et al., 1997). Additionally, commercially-available feed additives 3

containing antioxidants have been found to be effective in increasing concentrations of peripheral blood leukocytes in livestock when challenged with stressors such as: 1) injections of dexamethasone, 2) presence of molds in feed, or 3) the act of parturition (Wang et al., 2004, 2007). Improving immune function through feed supplementation with immunostimulants may not only aid in preventing IMI but also aid in decreasing treatment and culling costs due to mastitis. Other feed additives include general immunostimulants, such as yeast extracts, which improve an animal s ability to ward off diseases. For example, the commercial feed supplement, OmniGen-AF, which contains yeast and B complex vitamins, has been used to enhance the activities of L-selectin and interleukins in sheep when included as a dietary supplement (Wang et al., 2004, 2007, 2009). The purpose of this study was to evaluate the effect of OmniGen-AF on amplifying heifers immune response to S. aureus vaccination, with the goal of calving heifers free of infection and having SCC low enough to meet the proposed legal limit of 400,000/mL. Mastitis and its causes Mastitis (inflammation of the mammary gland) may be caused by chemical irritants, physical trauma, or most frequently, bacterial infection. The pathogens most commonly associated with mastitis include staphylococci, streptococci, and coliforms (Fox, 2009). These various pathogens can be further divided into two groups: 1) environmental pathogens and 2) contagious pathogens. Staphylococcus aureus and Streptococcus agalactiae are identified as contagious pathogens, which are spread during milking if sanitary procedures are not in place and followed appropriately. Environmental pathogens are found in mud, manure, and bedding materials. These pathogens, such as the environmental streptococci and coliforms, which are 4

present in the cow s surroundings, typically gain access to the interior of the mammary gland during the intermilking period. Staphylococcus spp. are characteristically separated into coagulase-negative staphylococci (CNS) and coagulase-positive staphylococci (CPS). Staphylococcus chromogenes, Staphylococcus hyicus, Staphylococcus simulans, and Staphylococcus epidermidis are examples of CNS that have been isolated from milk samples and secretions (Pyörälä and Taponen, 2009). The CNS typically cause mild cases of mastitis, which are rarely clinical in nature, whereas CPS infections often manifest as clinical mastitis and cause more tissue damage. The most common CPS is S. aureus; however, other bacteria such as Staphylococcus intermedius and S. hyicus can also produce coagulase. While the effects of CNS mastitis cases are not as severe as CPS infections, especially those caused by S. aureus, increased SCC due to CNS IMI can negatively affect milk production, and ultimately the producer s income. Streptococcus spp., considered as major mastitis-causing pathogens, include Strep. agalactiae, Streptococcus dysgalactiae, and Streptococcus uberis. Strep. uberis, being an environmental pathogen, has been identified as the most prominent Streptococcus pathogen to cause an IMI. Harmon et al. (1994) identified these streptococci as major mastitis-causing pathogens, along with S. aureus and coliforms, all of which result in deleterious compositional changes in milk as well as increases in SCC. Acute clinical infections are frequently associated with coliforms, including Escherichia spp., Klebsiella spp., and Enterobacter spp. E. coli, found in the gastrointestinal (GI) tract of warm-blooded animals, is identified as the most prevalent species of the genus Escherichia. Klebsiella and Enterobacter are isolated from GI tracts, soil, water, and grains. Overall, studies 5

have found that, while only 5% of quarters infected are due to coliforms, 25% of cows may have coliform mastitis, and 85% of these cases can result in clinical mastitis (Hogan and Smith, 2003). Prevalence of mastitis in heifers The prevalence of mastitis in nulligravid and pregnant heifers has been studied extensively in the past 2 decades. Formerly, most dairy animals were analyzed for bacterial infections as adult cows, e.g., shortly after calving, during lactation, and at drying off. More recently, studies have investigated breeding age heifers and monitored them throughout gestation, and subsequently found that greater than 90% of heifers 12 to 24 mo of age were infected. The CNS (such as S. chromogenes and S. hyicus) and S. aureus were found to be the most prevalent mastitis-causing pathogens (Trinidad et al., 1990b). In a 1-yr study that involved 10 Jersey heifers evaluated beginning at 10 to 12 mo of age through calving, Boddie et al. (1987) reported the prevalence of mastitis-causing bacteria on teat skin, in teat canals, and in mammary secretions. At the conclusion of the trial, it was determined that the major pathogens isolated from teat canals were as follows: S. chromogenes (41%), S. hyicus (16.8%), S. aureus (10%), Staphylococcus xylosus (1%), Staphylococcus warneri (0.8%), and Staphylococcus sciuri (0.5%). Similar results were found in mammary secretions with the following predominant bacterial isolates: S. chromogenes (49.5%), S. hyicus (21.3%), S. aureus (13%), Strep. uberis (1.4%), and S. xylosus (0.9%). There was a strong positive correlation (82.2%) between presence of teat canal organisms and those pathogens identified from mammary secretions. These results lend support to the hypothesis that heifers may be infected with mastitis pathogens at a young age, and such infections may persist until calving. Another study based on a larger sample (n = 97) of breeding age and pregnant animals found that bacteria colonized 70.7% of quarters and 93.1% of heifers (Trinidad et al., 1990b). S. 6

aureus was identified in teat canal keratin samples from 12.3% of quarters. S. chromogenes (42.9%), S. hyicus (25.2%), Strep. dysgalactiae (0.6%), other staphylococcal spp. (5.7%), other streptococcal spp. (3.1%), and mixed isolates (5.7%) were also identified from teat canal keratin samples in this study. Likewise, IMI were identified in 74.6% of quarters and 96.9% of heifers. Prevalence of infection among quarters was: S. aureus (14.7%), S. chromogenes (43.1%), S. hyicus (24.3%), Strep. dysgalactiae (0.4%), other staphylococcal spp. (3.6%), other streptococcal spp. (3.3%), and mixed isolates (5.1%). Overall, 29% of heifers and 15.1% of quarters exhibited clinical mastitis such as clots, flakes, and blood. Of the quarters diagnosed with clinical mastitis, S. aureus was isolated from 25% of them (Trinidad et al., 1990b). Fox at al. (1995) conducted a study of 28 dairies in Louisiana, California, Washington, and Vermont in which mammary secretions were collected from animals 8 mo to 38 mo of age. One visit per season was conducted to account for any seasonal and regional differences. Prevalence of heifer mastitis across seasons and regions ranged from 21.2% to 52%. Louisiana has a climate more similar to Georgia compared to other regions, such as Washington, so these results are valuable for comparing heifer mastitis. For example, in Louisiana, spring prevalence of mastitis was 39.7% (which was when IMI was greatest across seasons in Louisiana), whereas in Washington, prevalence of mastitis was 21.2%. Prevalence of S. aureus IMI in Louisiana across seasons ranged from 1.3% to 10.1%, with the highest prevalence occurring during the spring and fall (other regions across seasons ranged from 0% to 3%). In Louisiana, prevalence of CNS IMI across seasons ranged from 23% to 35.6%, with the highest prevalence occurring during fall and winter months. 7

Mastitis prevention and control techniques Several management techniques are used in attempts to prevent and treat mastitis. The National Mastitis Council (NMC) recommends a 10-point mastitis control program that includes: 1) determining udder health goals, 2) ensuring a clean environment, 3) practicing proper milking procedures including the use of teat dips, 4) following correct function and use of milking machine, 5) maintaining high-quality records, 6) timely treatment of clinical mastitis, 7) utilizing dry cow therapy, 8) effective biosecurity measures to reduce pathogen exposure, 9) observing mammary health intently, and 10) review of the mastitis control program from time to time (NMC, 2006). Teat dipping with a germicide, prior to milking, is used to reduce the number of mastitiscausing pathogens on teat skin during the milking process. Preparing a cow s udder for milking by predipping with a germicide will decrease the pathogen load on the surface of the teat skin that is present due to exposure to the environment. Conversely, postdipping is used to prevent the spread of contagious pathogens typically found on teat skin after milking. When comparing good udder preparation and postdipping vs. predipping in conjunction with postdipping, the rate of new infection from mastitis-causing pathogens was reduced by as much as 54% with the latter (Pankey et al., 1987). Additionally, Hogan et al. (1987) determined that development of new Staphylococcus spp. infections were greater in cows that were not dipped (11%) after milking versus those that were postdipped (7.2%). While pre- and post-milking germicidal use are important in reducing new IMI caused by environmental and contagious pathogens, a clean environment is also vital in maintaining a healthy herd to prevent new cases of mastitis, specifically those caused by environmental pathogens. High bacteria counts in the environment, such as in bedding, are directly related to 8

increased risk of infection. The goal of a clean environment is to reduce the pathogens that may come into contact with the cow s teats between milkings. Organic bedding materials, especially sawdust and straw, are associated with higher coliform numbers and Streptococcus spp., respectively, due to the availability of nutrients and a moist environment in which bacteria can survive. Sand is preferred to organic bedding materials because it does not support the growth of pathogens. Sand bedding, however, can become heavily contaminated by manure if not properly managed, and may harbor the same (if not higher) levels of pathogens. A direct relationship has been identified between bacterial counts in bedding and the rates of clinical mastitis cases, indicating that greater exposure leads to greater infection rates (Hogan et al., 1989). Mammary gland immune defenses The mammary gland is susceptible to an IMI when the natural innate and adaptive immune processes fail to neutralize harmful pathogens. The primary defense offered by the mammary gland involves a dual physical barrier, which consists of the keratin plug and teat sphincter. Keratin is composed of antimicrobial fatty acids and proteins that reduce bacterial multiplication and aid in preventing new IMI. The lack of a keratin plug results in increased susceptibility to pathogen exposure and ultimately increased rates of infection (Capuco et al., 1992). The teat sphincter provides tight closure at the teat end to prevent bacterial entry into the canal and further into the mammary gland (Sordillo et al., 1997). Innate immunity allows for the detection of pathogens in the mammary gland and involves epithelial cells and leukocytes. Leukocytes, primarily neutrophils, are considered the second line of defense against invading pathogens after the first line provided by the keratin and teat sphincter is breached. Initial contact between macrophages and mammary epithelial cells with bacterial invaders releases chemoattractants, such as cytokines, complement factors, and 9

prostaglandins, in response to the invading pathogens. These signals elicit the influx of neutrophils into the area of infection (Sordillo et al., 1997; Sordillo and Streicher, 2002). Neutrophils and macrophages are nonspecific, phagocytic cells that ingest and destroy invading pathogens. One difference lies in the ability of macrophages to present antigens to B and T cells under some conditions. Neutrophils are short-lived cells that are effective in eliminating most bacterial species via phagocytosis and intracellular killing. L-selectin (CD62L; cluster of differentiation number 62 ligand) is a primary adhesion molecule on the neutrophil cell surface that allows neutrophils to migrate from the blood across the capillaries into the mammary tissue to reach infected areas (Paape et al., 2002; Sordillo and Streicher, 2002). An increase in circulating L-selectin is indicative of an enhanced immune response (Wang et al. 2009). Once neutrophils reach the site of infection, their primary function is to phagocytize and kill bacteria. During phagocytosis, various processes take place that act to kill bacteria. These include the respiratory burst process, which releases free radicals into infected tissue, as well as the degranulation process, which releases enzymes from neutrophils to degrade bacteria. While the respiratory burst and degranulation processes produce reactive oxygen species (ROS) that are instrumental in preventing bacterial growth and colonization, these same ROS can damage tissues in the mammary gland and lead to the death of leukocytes themselves. These damaged areas of tissue may result in scar tissue formation, which leads to a decrease in or total loss of milk production in affected quarters of cows and heifers (Paape et al., 2003; Zhao and Lacasse, 2008). As stated previously, cytokine production leads to an influx of neutrophils into the gland, and includes products like interleukins (IL), colony-stimulating factors, interferons, and tumor necrosis factor. Interleukins are especially important for the function of neutrophils, and include 10

IL-1, IL-2, and IL-8 (Sordillo et al., 1997). Interleukin-8, produced by leukocytes and epithelial cells, mediates neutrophil migration and activation, and also induces inflammation. Inflammation is vital to the activation of immune processes to fight mastitis-causing pathogens. Due to this direct involvement, IL-8 has been described as a very powerful chemoattractant for leukocytes, neutrophils in particular. Interleukin-8, as well as complement factors, IL-1, and IL- 2 activate neutrophils. Production of these cytokines occurs after adhesion molecules, such as L- selectin, allow neutrophils to bind to endothelial surfaces. Research has also demonstrated the ability of IL-8 to induce the release of lysosomal enzymes through degranulation and enhance phagocytosis after opsonization of bacteria. IL-8 also recruits T lymphocytes, indicating the role of this cytokine in cell-mediated immunity. An increase in circulating IL-8 suggests an enhanced immune response during inflammation and infection due its direct role in chemotaxis (Barber and Yang, 1998; Paape et al., 2002). Importance of L-selectin and IL-8 in the immune response L-selectin functions by interacting weakly with the vascular endothelium. This allows the neutrophil to stop rolling along the interior vessel walls, and to monitor surfaces for signals of the presence of bacteria invading adjacent tissue. After adhesion of the neutrophil to vascular endothelium, integrins are expressed on the endothelium and neutrophils, allowing the leukocytes to bind more firmly. Once firm binding of the neutrophil has occurred, other cytokines, such as IL-8, initiate the shedding of L-selectin from the surface of the neutrophil, which allows more adhesion molecules to be expressed, such as Mac-1, an integrin that secures the neutrophil to the endothelial surface. This process facilitates the ongoing migration of neutrophils into the infected tissue. In addition to mediating L-selectin shedding, IL-8 also plays 11

an important role in production of ROS by leukocytes. The ROS molecules are instrumental in killing of bacteria (Paape et al., 2003; Forsberg, 2004). Influence of infection on SCC The movement of leukocytes into mammary tissue during infection results in an elevated SCC in mammary secretions or milk; therefore, the SCC is used as indicator of infection. Boddie et al. (1987) studied the association of SCC with bacterial species in mammary secretions from bred heifers. During gestation, uninfected quarters exhibited a mean SCC of 3.5 x 10 6 /ml, whereas SCC in quarters infected with S. chromogenes, S. hyicus, and S. aureus were 7.8, 8.5, and 9.2 x 10 6 /ml, respectively. At calving, SCC of uninfected quarters remained relatively low at 1.6 x 10 6 /ml, whereas those quarters infected with staphylococcal species averaged 3.2 x 10 6 /ml. Mean SCC through the first 3 mo of lactation were higher for quarters infected with S. aureus (578 x 10 3 /ml) compared to mean SCC of 168 and 193 x 10 3 /ml for S. chromogenes and S. hyicus-infected quarters, respectively. S. aureus has been consistently found to result in the highest SCC among mammary secretions from infected quarters. Trinidad et al. (1990b) found that S. aureus IMI resulted in an average SCC of 17.3 x 10 6 /ml compared to uninfected quarters of bred heifers, which averaged 5.7 x 10 6 /ml, while S. chromogenes and S. hyicus resulted in SCC of 12.8 x 10 6 and 12.4 x 10 6 /ml, respectively. Effect of S. aureus on mammary tissues To establish an infection in the mammary tissue, bacteria must first enter the mammary gland. Typically, the bacteria enter through the teat orifice and travel into the teat canal, where they are able to multiply and colonize the keratin, eventually migrating farther into the gland. A major goal of S. aureus once in the tissue is to attach to the various proteins and elements of 12

epithelial cell surfaces through corresponding binding proteins. Sites of attachment also were found to correspond to sites of lesions and ulcers, giving bacteria the opportunity to destroy tissues more effectively (Gudding et al., 1984). S. aureus produces various toxins including 4 different hemolysins, leukocidins, nucleases, proteases, lipases, hyaluronidase, collagenase, and coagulase. When released, these toxins severely damage mammary epithelial cells. The hemolysin, α-toxin, is the most widely studied, and has been shown to induce apoptosis and necrosis (Zhao and Lacasse, 2008). Cultured mammary epithelial cells damaged by α-toxin are found to be more amenable to attachment of S. aureus, which is an important mechanism in the pathogen s route to colonization and infection (Cifrian et al., 1994). Trinidad et al. (1990a) examined the histopathology of mammary quarters experimentally infected with S. aureus in unbred heifers ranging from 14 to 26 mo of age. The alveolar luminal area of infected tissues was reduced in size due to the increase of interalveolar connective tissue and scar tissue, which may lead to future decreased milk production in heifers. S. aureusinfected tissues also exhibited higher level of infiltration by lymphocytes and neutrophils than uninfected quarters; the latter typically exhibited only a few lymphocytes. Quarters infected with CNS exhibited a greater leukocyte infiltration into parenchymal tissues than uninfected quarters but not to the extent as observed with S. aureus-infected quarters. Tubercule-like micro- and macroscopic abscesses containing lymphocytes, neutrophils, plasma cells, and multinucleated giant cells were also found in S. aureus-infected quarters. These may lead to scar tissue formation. Quarters infected with CNS were more similar to uninfected quarters than to S. aureus-infected quarters, which exhibited increased stromal areas compared to CNS or uninfected quarters, and were usually associated with a decrease in milk-secreting tissues. The 13

damage caused by infiltration of leukocytes as well as reductions in secretory tissue, in young heifers may negatively affect milk production in the first and in subsequent lactations. Likewise, Sordillo et al. (1989) found increased stromal tissue in quarters infected with S. aureus. In some cases, epithelial tissue was deemed inactive, lacking an accumulation of milk components in the lumen. Furthermore, ultrastructural examination showed that the cells damaged by S. aureus were less differentiated than uninfected quarters, and interalveolar tissue areas were found to have higher concentrations of immune cells, especially antibody-producing B cells. While percentage of epithelial area was similar when comparing infected vs. uninfected tissues, the epithelium in infected tissues was deemed less developed 7 d prior to calving and at calving. Additionally, S. aureus-infected quarters exhibited increased epithelial cell areas composed of unoccupied cytoplasm (less developed cells) along with increased leukocyte infiltration. Necrosis of epithelial tissue was also observed in infected quarters, providing an additional piece of evidence linking infection status and milk production, e.g., an increase in necrotic tissue ultimately results in fewer cells able to produce milk. Similarly, other studies have reported tissue degradation in infected quarters due, not only to the S. aureus pathogen itself, but also to the influx of leukocytes, mainly neutrophils, that further damaged alveolar tissue. While not highly prevalent in mammary tissues, S. aureus has been identified in several vacuoles of neutrophils indicating its ability to survive within the phagocytic leukocytes (Nickerson and Heald, 1982; Harmon and Heald, 1982), which compromises the effectiveness of the immune processes. Heifer mastitis prevention and control techniques Several management tools are used on dairy operations to prevent the spread of mastitis among heifers and subsequently to the lactating and nonlactating herds. While antibiotic therapy 14

has been shown to cure mastitis in young heifers prior to calving, the ideal management technique is prevention, e.g., through vaccination programs and other preventative management protocols. Since heifer calves are the replacements for culled cows, these young animals are extremely valuable to a dairy producer. The initial management tool used for reduction of disease transfer involves separating the calf from the mother quickly, which prevents the calf from introducing pathogens into the teat during suckling. Secondly, calves are housed in individual hutches to prevent cross-suckling between calves, which may spread mastitis-causing pathogens such as S. aureus and Strep. agalactiae (Schalm, 1942). Antibiotic therapy in heifers Antibiotics are used in adult cows during lactation to treat existing infections and/or at dry off to prevent new infections during the nonlactating period. Heifers are assumed to be uninfected due to their nonlactating state; however, they may be chronically infected with mastitis pathogens for up to 2 yr, damaging the developing mammary tissue. Infection rates in breeding age dairy heifers as high as 97% have been observed, with S. aureus infections resulting in significant production losses over the course of the first and subsequent lactations (Boddie et al. 1987; Meaney, 1981; Nickerson et al. 1995). Both nonlactating and lactating cow therapeutic products have been used successfully to control heifer mastitis and are discussed below. Nonlactating cow products Dry cow products administered during gestation have been successful in reducing the prevalence of mastitis in heifers upon calving. Trinidad et al. (1990c) studied primigravid heifers (35 treated, 38 untreated) to determine the effectiveness of a single intramammary infusion of a penicillin and dihydrostreptomycin nonlactating cow product given 60 d prior to 15

expected calving date. Among treated animals, the percentage of infected heifers decreased from 97.1% before treatment to 40% at calving. Similarly, 73.2% of quarters among treated animals were infected before treatment, which was reduced to 34% at calving. Among untreated animals, 100% of heifers and 71.2% of quarters were infected initially, and upon calving, 97.4% of heifers and 77.8% of quarters remained infected. Six of the 38 treated heifers (11 quarters) were infected with S. aureus, which was successfully eliminated from all but 1 quarter, yielding a reduction of 91%. Conversely, 10 of the 35 untreated heifers (18 quarters) were infected with S. aureus initially, and 6 heifers (11 quarters) were still infected at calving, yielding a spontaneous cure rate of 39%. Ultimately, overall prevalence of IMI was reduced by 60% and S. aureus infections were reduced by 91%. Owens et al. (1991, 1994) administered a nonlactating cephalosporin-based dry cow product to heifers 8 to 12 wk prepartum after identifying natural S. aureus infections or after experimentally inducing S. aureus IMI. All (100%) of the experimentally induced infections and 87% of the natural mammary gland infections that were treated were found to have cleared the infection at calving. Conversely, untreated control quarters that remained infected at calving exhibited a cure rate of only 50-56% when given a lactating cow product (cephapirin benzathine) shortly after calving. The average SCC of infected quarters that were subsequently treated and cured decreased from 15 x 10 6 /ml to 4 x 10 6 /ml 1 wk later, and finally to 700 x 10 3 /ml at calving. All untreated infected quarters remained infected at calving, and exhibited a mean SCC of 5 x 10 6 /ml. While studies have established that treating heifers prior to calving with nonlactating products results in a reduction of IMI and SCC, the optimal time to administer treatment is still a consideration. Owens et al. (2001) conducted a 2-yr study to determine the efficacy of 16

administering a nonlactating cow product at various times during gestation using a total of 233 heifers. Initially, quarter samples were taken from each heifer that was confirmed pregnant and sampling continued every 4 wk. A total of 56.5% of quarters were reported to be infected with some type of mastitis pathogen, and 15.4% of quarters were infected with S. aureus. After sampling, animals were infused once either during the first, second, or third trimester of pregnancy with one of the following nonlactating cow products: 1) combination of penicillin and dihydrostreptomycin, 2) cephapirin benzathine, 3) combination of novobicin and penicillin, 4) tilmicosin, or 5) cephalonium (used only during the first trimester). Cure rates associated with these products ranged from 67 to 100%. However, untreated animals exhibited a spontaneous cure rate of only 25%. Results showed no differences in cure rate when the heifers were treated during the first, second, or third trimester of pregnancy. While non-lactating cow antibiotic therapy has been found to be effective in reducing infection rate and SCC at the time of calving, the use of a teat sealant can further protect the young mammary gland from bacterial invasion prior to calving by serving as a physical barrier. For example, as a heifer nears parturition, her teats will eventually dilate due to fluid accumulation in the udder, and an open (leaky) teat is much more likely to allow the entry of bacteria into the gland (Kromker and Friedrich, 2009). Teat sealant infused into the teats of heifers prior to calving has shown to reduce the prevalence of IMI after calving. Parker et al. (2008) used a bismuth subnitrate teat seal in 2,082 quarters and found that compared to quarters that did not receive a teat seal, those treated were less likely to be diagnosed with a new IMI (relative risk [RR] = 0.34) and less likely to experience clinical mastitis (RR = 0.26). 17

Lactating cow products Lactating products have been used 1 to 3 wk prepartum to treat mastitis in bred heifers, and have been found to be successful in reducing prevalence of environmental streptococci and CNS infections after calving. Oliver et al. (1992) studied the efficacy of administering a lactating cow product in reducing the prevalence of mastitis in 115 pregnant Jersey heifers. At the time of infusion (1 wk prepartum), the infection rate among animals was 90%, but after treatment, only 17.6% remained infected at calving. Those left untreated exhibited a high infection rate of 78% at calving. At 10 d postcalving, untreated animals still had an infection rate of 29% compared to a 4% infection rate in treated animals. Unfortunately, 17.4% of colostrums collected from 33 heifers tested positive for antibiotic residues when treated with cloxacillin. However, heifers infused with cloxacillin > 7 d prior to calving tested negative for residues. In addition, 28% of milk samples taken 3 d after parturition from heifers treated with cephapirin tested positive for antibiotic residues. Based on these results, subsequent prepartum treatment was administered 14 d or more prior to calving to minimize antibiotic residues. In a study conducted by Oliver et al. (1997), 67% and 64% of control and treated heifer mammary glands, respectively, were found to be infected 14 d prior to expected calving. After the initial sampling, the treated quarters were infused with cephapirin sodium. The prevalence of infection decreased to 56% and 16% in control and treated heifers, respectively, at 3 d postcalving, and by 30 d postpartum, control animals had a 36% infection rate compared to an infection rate of 8% in treated animals. Treated animals also had a lower infection rate during the remainder of the lactation compared to untreated control heifers (45% vs. 12%). Only 3.1% of milk samples taken 3 d postpartum tested positive for antibiotic residues. 18

Oliver et al. (2004) subsequently evaluated the effectiveness of a penicillin-novobicin and a pirlimycin lactating cow product administered 2 wk prepartum in reducing mastitis during early lactation in bred Holstein heifers. At the start of the trial, heifers were sampled 14 d prior to calving, and 73% of heifers were found to be infected. After treatment with penicillinnovobicin or pirlimycin, 24% and 41%, respectively, of heifers were infected at calving, while in the control group, 74% remained infected. The predominant pathogens infecting these heifers were CNS (44%) and S. aureus (30%). A replica of this same study was conducted using Jersey heifers as well, and 96% of these animals were found to be infected 14 days prior to calving. However, after treatment with penicillin-novobicin or pirlmycin 14 d prepartum, prevalence decreased to 25% and 13%, respectively, at time of calving. Conversely, 44% of control heifers were found to be infected at calving. Pathogens identified in this group were mainly CNS (61%), followed by Streptococcus spp. (19%), and S. aureus (8%). A subsequent study reported the efficacy of cephapirin on infection rate at calving when administered to heifers 10 to 21 d prepartum (Borm et al., 2006). Control quarters exhibited a cure rate of 31.7% compared to a cure rate of 59.5% for heifers treated with the lactating cow product. These studies show that lactating cow products administered 7 to 21 d prepartum have been effective for CNS- and Streptococcus-infected quarters. However, quarters infected with S. aureus may be chronically infected, and it is likely that lactating cow treatment would be of questionable value. Developing mammary glands that have experienced a S. aureus infection may have reduced productive capacity due to the negative effects of the pathogen on mammary tissue development. Treatment of S. aureus-infected quarters prior to parturition with 19

nonlactating cow products should be administered earlier in gestation to: 1) cure existing infections, 2) reduce chronic inflammation, and 3) allow mammary tissues to undergo normal development throughout the remainder of gestation (Nickerson, 2009). Fly control The presence of horn flies (Haematobia irritans) may also influence the infection status of heifers on a dairy operation. Horn flies insert their proboscis into the teat end capillaries to draw blood, and in the process inject S. aureus into the skin, resulting in small abscesses and formation of scabs (Owens et al., 1998). The spread of S. aureus by horn flies among dairy animals was initially established when reports showed that farms using some form of fly control had lower rates of mastitis in heifers compared with herds using no fly control. Also, the animals from the herds not using a fly control method had more scabs and lesions on their teats, which were associated with S. aureus colonization (Nickerson et al., 1995). With the advent of DNA technology, animal strains of pathogens and suspected vectors were studied to determine what relationship, if any, existed. Owens et al. (1998) identified a strain of S. aureus colonizing horn flies, which was identical to the strain identified on teat ends of heifers that had been exposed to these same flies. Teat ends exposed to flies were covered in scabs, which contained the identified strain of S. aureus. Once it was established that the horn fly was a vector in the transmission of mastitiscausing bacteria, management practices were evaluated to reduce flies and lower the prevalence of IMI. Insecticide-impregnated tags placed on the tail switch in close proximity to the udder during the spring and summer months were successful in reducing horn fly populations by 60% as well as the incidence of mastitis during the first 2 mo after placement (Nickerson et al., 1998). In heifers with tail tags, mastitis incidence increased from 8.6 to 15% (1.7-fold increase) during 20

spring and summer months, while in controls, incidence increased from 17.1 to 52.4% (3.1-fold increase). However, after 2 mo, tags fell off and replacing them was impractical from a management standpoint. In a subsequent trial, daily dietary supplementation of an insect growth regulator helped to suppress fly populations but not enough to prevent new cases of mastitis in dairy heifers (Owens et al., 2000). Lastly, the use of an insecticidal pour-on every 2 wk for 6 wk followed by placement of insecticidal ear tags reduced fly populations and decreased the incidence of new S. aureus in heifers by 83% in a 6-mo efficacy trial (Owens et al., 2002). Vaccination Currently, vaccination is not a major mastitis management tool in heifers, but is heavily used in the prevention of other calfhood diseases, such as clostridial infections, Brucellosis, respiratory infections, and enteric viruses. Vaccines directed at controlling mastitis are used more frequently in lactating cows. Presently, commercial vaccines are available to decrease the prevalence of mastitis caused by coliforms and S. aureus. Coliform vaccines Several vaccines are available for immunization against coliforms, and have been found to increase antibody titers and increase the recovery of the cow (Tomita et al., 2000). Coliform vaccines are based on gram-negative core lipopolysaccharide (antigen), which lacks the O-side chains that protect the lipopolysaccharide component. This characteristic is important because the immunoglobulins produced by the animal are specific to the exposed lipopolysaccharides, which are common to all gram-negative bacteria. E. coli is the predominant organism that is targeted with the majority of the commercially available vaccines; however, there have been some reports that indicate protection against Klebsiella, Pseudomonas, Serratia, and Proteus as well. Studies have shown that vaccinated animals exhibit a 46.7% decrease in the number of 21

clinical cases of mastitis caused by coliforms compared to unvaccinated animals (Hogan et al., 1992). S. aureus vaccines S. aureus is a contagious pathogen, and its primary reservoirs are teat and udder skin and mammary tissues of infected cattle. This pathogen is spread from cow to cow during the milking process, and once established in the mammary gland, infections are difficult to cure with antibiotic therapy. Vaccines using whole-killed bacterins, live attenuated bacterins, and vaccines containing various virulence factors have been used to reduce the rates of S. aureus mastitis, as well as reduce the severity of infections. Vaccine trials have shown that the rate of spontaneous recovery from mastitis is higher in vaccinated cows than control animals. Additionally, the number of clinical mastitis cases is decreased when animals are vaccinated versus those that are not vaccinated (Pankey et al., 1985). Watson (1984) reported that after a live, attenuated S. aureus vaccine containing pseudocapsule was administered to heifers late in gestation, serum antistaphylococcal IgG 1 and IgG 2 levels were higher compared to control animals. The influx of IgG 2 into the mammary gland indicated an influx of circulating neutrophils given the fact that IgG 2 is cytophilically bound to this leukocyte. An influx of neutrophils and IgG 2 (directed against the pseudocapsule) to the site of inflammation would allow increased opsonization and phagocytosis of S. aureus. In a study conducted by Pankey et al. (1985), 30 Jersey cows in their first lactation were immunized with either a protein-a based S. aureus vaccine (n = 10) or a commercial staphylococcal bacterin (Somatostaph, Anchor Laboratories, St. Joseph, MO, n = 10); 10 animals served as unvaccinated controls. Within each treatment group, 5 animals were immunized 9 wk prior to parturition and the other 5 were immunized 2 to 4 wk prior to 22

parturition. After parturition, all cows were challenged by teat end immersion in a suspension of S. aureus once a day for 5 d a week over 6 mo, and new IMI were monitored. Results showed that the spontaneous cure rates were higher in protein-a and bacterin-vaccinated groups (87 and 75%, respectively) when compared to control animals (40%). Vaccinated cows also exhibited lower rates of clinical mastitis compared to the control group. Vaccinated animals (protein A and bacterin) had lower SCC (316 x 10 3 and 290 x 10 3 /ml, respectively) than control animals (772 x 10 3 /ml) In a trial conducted by Sears et al. (1990), 20 Holstein heifers were used 2 mo prepartum to determine the efficacy of a S. aureus vaccine against experimental challenge. Vaccinated animals (n=20) received injections 4 and 2 wk prior to expected calving. In the control group (n=5), 75% of the quarters were found to be infected after calving; however, one quarter spontaneously cured within 7 d resulting in a 64% chronic IMI rate due to the challenge. Conversely, only 33% of the vaccinated heifers experienced an IMI after challenge, with 9 quarters spontaneously curing, resulting in a 12% chronic S. aureus IMI. Nickerson et al. (1993) studied the effect of a S. aureus bacterin (developed by Watson, 1984), on serum antibody, new infection rate, and mammary development given either intramuscularly or in the area of the supramammary lymph node (SMLN). Nonpregnant cows were vaccinated 1 d after dry off and 6 wk later. The animals were experimentally challenged intracisternally with S. aureus 4 wk after the booster was administered. Vaccinated animals exhibited a 4.7-fold increase in serum antibody titers over unvaccinated animals and pretreatment levels. IgG 1 and IgG 2 titers remained higher in vaccinated animals compared to control animals. Eleven out of 12 control quarters (91.7%) became infected compared to 10 out of 21 quarters in vaccinated cows (47.6%). Approximately 36% (4 out of 11) of quarters 23

vaccinated intramuscularly became infected compared to 60% (6 out of 10) of quarters in cows vaccinated in the SMLN. Ultimately, vaccination reduced new IMI by 48.1% compared to controls. Nordhaug et al. (1994) used 108 heifers to determine effectiveness of a vaccine containing whole, inactivated bacteria with α and β toxins when given in the area of the SMLN. Heifers received an injection 8 wk before calving and another 2 wk before calving. Results showed that 6% of control animals had to be treated for clinical mastitis caused by S. aureus during lactation whereas none of vaccinated heifers had to be treated for clinical symptoms due to S. aureus. Subclinical infections caused by S. aureus were higher in control animals compared to vaccinated animals (14% vs. 8.6%, respectively). Vaccinated animals showed a significant response to vaccination in total serum IgG, IgG 1, and IgG 2 levels against the pseudocapsule and α-toxin, which were higher than control animals and remained higher throughout lactation. Another study using encapsulated antigens and toxic products produced by S. aureus reported reductions in IMI caused by S. aureus (Giraudo et al., 1997). A vaccine made from inactivated, encapsulated S. aureus, S. aureus exopolysaccharides, and inactivated, unencapsulated S. aureus and Streptococcus spp. strains was used in 30 bred heifers. Group 1 received vaccinations at 8 and 4 wk before expected calving; group 2 received vaccinations 1 and 5 wk after calving, and group 3 received a placebo 8 and 4 wk before expected calving. Groups 1 and 2 heifer quarters exhibited a 2.5% incidence of clinical mastitis caused by S. aureus, whereas 12.5% of control quarters exhibited clinical mastitis. In total (subclinical, clinical, and latent infections), groups 1 and 2 exhibited a quarter IMI rate of 32.5% and 27.5%, respectively, whereas control quarters exhibited an infection rate of 75%. 24

One commercial vaccine (Lysigin, Boehringer Ingelheim Vetmedica Inc., St. Joseph MO, USA) in particular, was formulated to target specific capsulated serotypes that have been identified as most often infecting dairy animals. Lysigin is a lysed culture of polyvalent antigens of 5 S. aureus serotypes combined with an aluminum hydroxide adjuvant. The five serotypes include one serotype 5, two serotypes 8, and two serotypes 336. Nickerson et al. (1999) tested Lysigin using the dairy heifer as a model. The vaccine was administered to 35 Jersey heifers at 6 mo of age in the semimembranosus muscle (5 ml), a booster was given 14 d later, and again every 6 mo; 35 heifers served as controls. Results demonstrated a reduction in new IMI during gestation of 44.8% and a reduction in new IMI at calving of 44.7%. Antistaphylococcal antibody titers remained higher in vaccinated animals compared to controls and were significantly higher 2 mo after the initial vaccination and 2 mo after the booster vaccination. In a follow-up study (Nickerson et al., 2009), infection rate and SCC in Holstein heifers receiving the same S. aureus bacterin (Lysigin ) were evaluated using the vaccination schedule from the previous study (Nickerson et al., 1999). At calving, a reduction of 60.9% was reported in animals that were vaccinated compared to unvaccinated controls. Somatic cell counts in milk samples collected during the first week of lactation irrespective of infection status were 45% lower in vaccinates compared with controls (287,317 vs. 522,345/mL). The SCC from uninfected heifers for vaccinates and controls were 66,095 and 132,754/mL, respectively; a 50.2% reduction; and for infected heifers, SCC were 441,764 and 892,176/mL, respectively; a 50.5% reduction. An examination of the 305-d lactation milk yield for the first lactation of vaccinated and unvaccinated heifers demonstrated an 8.6% increase in yield in vaccinates vs. controls (11,217 vs. 10,332 kg, respectively) or a difference of 886 kg. Actual 305-day weights (in 25

kilograms) of both fat and protein were higher in vaccinates than controls (fat: 408 vs. 339 kg, respectively; protein: 330 vs. 315 kg, respectively). An examination of the number of days in milk for the first lactation demonstrated that vaccinates persisted 13 days longer than unvaccinated controls (349 vs. 336 days). In addition, the average first lactation SCC was 11,000 cells/ml lower in vaccinates compared with controls (49,000 vs. 60,000/mL). Authors concluded that vaccinating dairy heifers with Lysigin reduced the number of new S. aureus IMI at time of calving by 60.9%, lowered the SCC by 50%, and decreased the culling rate by approximately one-third. In addition, overall milk yield, production of fat and protein, and number of days in milk were greater in vaccinated heifers compared with controls. Luby and Middleton (2005) evaluated Lysigin in 12 lactating cows with chronic S. aureus infections. The animals were assigned to; 1) a group that received extended pirlimycin therapy or 2) pirlimycin and Lysigin. During the trial, cattle were vaccinated 14 d and 2 d before they received antibiotic therapy, and again 7 d following antibiotic therapy, or received antibiotic therapy only. A total of 50% of quarters were cured in the pirlimycin treatment only group compared to 66% of quarters in the pirlimycin plus Lysigin group; however, these differences were not statistically significant. Overall, it did not appear to be advantageous to combine vaccination with antibiotic therapy in cows suffering from chronic S. aureus infections. Lysigin was also the focus of a follow-up study conducted by Middleton et al. (2006) to determine the efficacy of this commercial vaccine compared to 2 experimental bacterins. Forty-seven heifers received the following treatments: 1) an experimental bacterin containing 3 S. aureus isolates (n=11); 2) an experimental bacterin containing 5 S. aureus isolates (n=11); 3) Lysigin (n=14); and 4) unvaccinated (n=11). Vaccines were subcutaneously administered 42 and 14 d before expected calving. There were no significant differences in spontaneous cure 26

rates, clinical mastitis prevention rates, or SCC within the vaccinated groups and between the vaccinated and the control groups. However, Lysigin -vaccinated animals did have a significantly lower mean duration of clinical mastitis than the unvaccinated animals; 4 d compared to 10 d. Also, while not statistically significant, vaccinated heifers did exhibit decreased SCC, increased milk yield, and a decrease in severity of clinical mastitis cases (Middleton et al., 2006). In a subsequent trial, Luby et al. (2007) evaluated serum and mammary secretion samples from groups 3 and 4 from the above study (Middleton et al., 2006). Enzymelinked immunosorbent assays (ELISA) demonstrated a significant increase in total anti-s. aureus IgG titers in serum of vaccinated animals, as well as an increase in serum IgG 1 and IgG 2 concentrations. Heifer growth Monitoring dairy heifer growth is necessary to determining if animals are developing normally to reach the appropriate height and weight for optimal future milk production. A healthy immune system should result in a properly growing animal because less energy is expended fighting calfhood diseases. In this context, using a feed additive to enhance the immune system may in fact enhance the growth performance of heifers. To analyze growth, standard growth charts are typically used as references that provide weight and height values for east comparison. Typically, weight measurements taken using tapes, are performed around the heart girth, and are generally accurate within 5 to 7% of the actual body weight based on scales. In terms of height measurements, a height stick positioned at the hip is commonly used; however, wither height is often used as an alternative. These values can then be compared to average values obtained from a large data set. 27

From 1991 through 1992, the National Dairy Heifer Evaluation Project (a USDAsponsored study) documented various management practices on growth and development of young calves and replacement heifers for the Holstein population (Heinrichs and Losinger, 1998; Heinrichs et al., 1994). Results showed that dairy heifers in the Southeast grew poorest. They were behind the Midwest, West, and Northeast in growth in that study. For example, in the Southeast, expected weights on days 180, 208, 365, and 545 were 170, 190, 294, and 400 kg, respectively, compared to body weights of heifers in the Midwest for these same days (184, 206, 322, and 442 kg, respectively) (Heinrichs and Losinger, 1998). Using this USDA study along with data from surrounding Pennsylvania farms, the Penn State calf and heifer growth charts were formulated, which have been instrumental in comparing heifer growth in research studies to average animals in commercial dairy operations. Average daily gains (ADG) are also useful in monitoring the performance of replacement dairy heifers. Using past research and field survey data, upper and lower values for ADG under ideal management has been established (Hoffman, 1997). On average, the upper value for ADG for Holstein heifers was 1.84 lb/d and the lower value was 1.68 lb/d. Nutritional immunostimulation Optimal growth and control of livestock diseases, such as mastitis, through nutritional supplementation is being encouraged by the US government as well as consumers due to the general disapproval of antibiotic use in food animals. With an increasing consumer awareness of food sources and animal welfare, producers are now more focused on management techniques to address livestock health and product quality without relying on drugs or antibiotics. Discovering ways to boost mammary gland immunity in conjunction with vaccination programs to prevent mastitis is an emerging area of interest. Healthy cows have a balanced diet consisting primarily 28

of energy and protein to meet their maintenance and lactation requirements. Other nutrients administered for general health and wellbeing include vitamins, macrominerals, and trace minerals. While heifers receive a balanced diet for maintenance and growth that focuses on sufficient levels of protein and energy, trace vitamins and minerals are reported to be equally important to optimize immune function and udder health. The most frequently studied trace minerals and vitamins that are found to have immunomodulatory effects on the mammary gland include vitamin E, selenium, vitamin A, β-carotene, vitamin D, copper, and zinc (Heinrichs et al., 2009; Sordillo et al., 1997). Vitamin E, primarily found in green grasses, has been identified as a fundamental component of antioxidant processes that are involved in the reduction of free radicals released during phagocytosis and killing of bacteria by leukocytes. The active component of vitamin E has been identified as α-tocopherol, and for an animal to receive the proper amount, it must receive the proper amount of vitamin E. The antioxidant, vitamin E, serves as an integral part of cell membranes that prevents the adverse effects of lipid peroxidation of membrane-bound phospholipids caused by free radicals. Neutrophil cell membranes contain these membranebound phospholipids, which makes them especially susceptible to peroxidation by free radicals. Hogan et al. (1990b) used 21 multiparous Holsteins to determine the efficacy of supplementing a lactating cow s diet with vitamin E, selenium, or both and found that neutrophils from vitamin E-supplemented cows exhibited increased intracellular killing. Vitamin E appears to function by increasing neutrophil viability, thereby allowing for more effective killing of bacteria (Hogan et al., 1993). Selenium, found primarily in soils, is an integral component of the enzyme glutathione peroxidase (GSHpx), which converts hydrogen peroxide (produced by the conversion of 29

superoxide to hydrogen peroxide) to water (Zhao and Lacasse, 2008). Hydrogen peroxide is a reactive oxygen species (ROS) produced during the phagocytosis and killing of bacteria by leukocytes, which is deleterious to cell membranes. The conversion of ROS into less threatening substances via neutralization is beneficial to the health of the animal because ROS will damage healthy cells if not neutralized. Hogan et al. (1990b) found that in cows supplemented with Se, GSHpx activity was higher compared to unsupplemented animals. In addition, Se supplementation led to increased intracellular killing of S. aureus compared to unsupplemented animals. Similarly, Smith et al. (1984, 1997) found that dietary supplementation with both Se and vitamin E and resulted in reduced rates and severity of clinical mastitis in dairy cows; however, in another study the effects were not additive (Hogan et al., 1990b). It was also documented that the speed of immune response, intracellular killing of bacteria, and longevity of neutrophils were heightened in heifers receiving Se supplementation (Sordillo et al., 1997). Vitamin A and β-carotene are actively involved at the cell membrane level to ensure functional and healthy epithelial cells by aiding in the growth of the epithelium in addition to acting as antioxidants. While vitamin E acts on the peroxidation process once the process has started and attempts to terminate it, the main function of β carotene or vitamin A, is to prevent events leading to the initiation of peroxidation and generation of ROS (Weiss, 2002). β-carotene specifically acts as a neutralizer of ROS, which causes oxidative stress during parturition and bouts of clinical mastitis (Sordillo et al., 1997). Studies have shown that there is an increased rate of new IMI when animals are fed diets that are deficient in vitamin A and β-carotene (Dahlquist and Chew, 1985). Copper is also recognized as being an instrumental nutrient in immunoregulation. It is a primary component in the enzyme superoxide dismutase, which works to neutralize ROS by 30

converting superoxide to oxygen and hydrogen peroxide. GSHpx then reduces free hydrogen peroxide thus eliminating this harmful ROS. Jones and Suttle (1981) reported a decrease in microbicidal activity of peripheral leukocytes due to a copper deficiency in cattle, which may have been due to a decrease in superoxide dismutase, leading to leukocyte damage. Research has also found that copper is an integral component of ceruplasmin, an acute phase protein, which is believed to function as a modulator of lysosomal enzymes released during inflammation and assist in iron absorption and transport. Cows suffering from mastitis reportedly have higher levels of circulating ceruplasmin compared to noninfected cows as reviewed by Nickerson (2011). Various studies involving dietary supplementation with copper have reported decreased SCC, increased neutrophil bactericidal activity, and a reduction in number of mastitis cases (Heinrichs et al., 2009). Zinc is vital to mammary health due to its role in protecting the integrity of the teat canal, specifically in the formation of the keratin plug. Like copper, zinc is an integral component of superoxide dismutase, which aids in neutralizing ROS by converting superoxide into oxygen and hydrogen peroxide (later reduced by GSHpx). Furthermore, deficiencies in dietary zinc have been found to lead to thymus atrophy, affecting various other lymphoid tissues in mice as reviewed by Hogan et al. (1990a). Cell mediated immunity and antibody response may also be affected by a zinc deficiency due to a decrease in T and B lymphocytes in humans; however, the function of zinc in bovine immunology is not well documented (Sherman, 1992). Research does not support a conclusive relationship with zinc and the health of the bovine mammary system; however, the role of zinc in superoxide dismutase indicates its importance and possible immunomodulatory function. 31

Vitamin D has also been cited as an important, yet indirect, immune regulator. While it is understood that a deficiency in vitamin D affects calcium metabolism, which ultimately may have a secondary effect on the immune system, it is not widely reported that vitamin D has a direct relationship to immune function. E. coli vaccine trials using the active form of vitamin D as an adjuvant with a vaccine have shown that antibody titers in milk and blood serum are greater compared to using the vaccine alone (Reinhardt et al., 1999). Additionally, infected mammary tissues and monocytes have an increased expression of the enzyme that catalyzes vitamin D to its active form (Nelson et al., 2010), indicating a possible role for vitamin D in the mammary immune system. Commercial feed additives incorporating an array of the above supplements to improve animal health have not been intensely researched in the past; however, new efforts are focused on formulating products to enhance the immune system against livestock diseases, such as mastitis. Wang et al. (2004) analyzed the effects of a feed additive, OmniGen-AF, containing various B-vitamins, trace nutrients, yeast products, and fungal organisms on improving immunity in wethers and ewe lambs. This trial showed that when sheep were challenged with the immunosuppressor, dexamethasone, L-selectin levels were restored in blood samples from animals supplemented with OmniGen-AF compared with unsupplemented animals. L-selectin, expressed by neutrophils on their cell surface, interacts weakly with the capillary endothelium, which allows neutrophils to roll along the interior vessel walls and to monitor surfaces for the presence of signals, such as cytokines, signaling an infection. In this same study, Wang et al. (2004) challenged sheep with a mold commonly found in forages, which has been found to suppress the immune function. When the animals were fed OmniGen-AF after exposure to the mold, L-selectin levels were restored and remained elevated compared to unsupplemented 32

control animals similar to results after administration of dexamethasone. Levels of IL-1β, a cytokine important in the inflammatory response and adaptive immunity, also increased after supplementation. Rowson et al. (2010) infused 4 different mastitis-causing pathogens (Strep. uberis, E. coli, S. aureus, and Klebsiella pneumonia) into the mammary glands of mice. Mice that received OmniGen-AF 2 wk prior to infusion and during lactation exhibited significantly reduced concentrations of mammary bacterial DNA concentrations (> 99% reduction in S. uberis, 71% reduction in E. coli, and 94% reduction in S. aureus), indicating a positive effect of the feed supplement on the ability of the mammary gland to resist infection. While the mechanism by which this supplement enhances these immune mediators is unknown, it is evident that there is a positive relationship between this feed additive and immune mediator levels in the blood of experimental animals. Researchers hypothesized that the yeast and fungal cell wall components interact with receptors in the gastrointestinal tract that initiate an innate immune response, priming cells for future exposure to antigens (Harris et al., 2006; Heine and Ulmer, 2005). In a subsequent trial, Wang et al. (2007) evaluated the effect of OmniGen-AF on innate immune markers in ewes and wethers after the administration of the immunosuppressor, dexamethasone alone or dexamethasone in combination with a pathogen (mold). Both immunosuppressors decreased expression levels of L-selectin and IL-1β, which were restored with the addition of OmniGen-AF. These levels, once reversed, were higher than those found in unsupplemented control animals. IL-1β was affected to a greater extent, e.g., decreased expression, when dexamethasone and the mold were administered together rather than dexamethasone alone. This is important, because IL-1β assists neutrophils in migrating from the 33

endothelium into the tissue by allowing neutrophils to adhere to the endothelial walls (Janeway et al., 2005). More recently, Wang et al. (2009) studied 8 periparturient Jersey cows under natural stress (parturition) to determine if dietary supplementation with OmniGen-AF expressed the same positive effects on immune function. Animals were fed a control or treatment (OmniGen- AF -supplemented) dry cow ration daily 1 mo prior to parturition, and after calving, blood samples were taken to measure IL-4R and interleukin converting enzyme (ICE) activities. Researchers hypothesized that the mechanism by which IL-1β activity was increased was through the increased activation of ICE. Results also showed an increase in the average concentration of blood leukocytes in supplemented animals compared to controls (6,950 vs. 10,300/mm 3 ), and neutrophil numbers tended to increase (2,654 vs. 4,200/mm 3 ). These changes in regulator and cytokine gene expression indicate a positive association between the feed supplement and the immune system of the animals used in this study (Wang et al., 2009). 34

Objectives The overall purpose of this study was to determine the effects of OmniGen-AF on dairy heifer s immune system prior to breeding, throughout gestation and through calving. Specific objectives were established and evaluated the following parameters: Monitor any differences in growth between treated and control animals. Measure blood leukocyte RNA expressions of the immune markers L-selectin and IL-8R. Determine the prevalence of teat canal and intramammary infections prior to and during gestation among treated and control animals. Measure SCC, CFU count, and differential leukocyte counts of mammary secretions in treated and control animals. Determine differences in blood leukocyte phagocytic ability and ROS production in treated vs. control animals. Monitor IMI, SCC, antibiotic residues, and any differences in milk yield postcalving between treated and control animals. 35

CHAPTER 2 MATERIALS AND METHODS Overview A total of 83 Holstein heifers from the UGA Teaching Dairy were studied over a 22-mo period (April 2010 to February 2012) to determine the effect of OmniGen-AF (Prince Agri Products, Inc., Quincy IL, USA) on immune function. Due to the difference in ages, groups began the trial at varying times. Heifer numbers and initiation dates are listed in Appendix 1. For this reason, all groups could not be analyzed for each parameter, and all animals were not enrolled for the full 22 mo. Prepartum parameters measured included musculoskeletal growth, serum antibody titers against S. aureus, prevalence of teat canal infections, prevalence of intramammary infections, mammary secretion differential counts, mammary secretion SCC, mrna expression of L-selectin and interleukin-8 receptor (IL-8R) on blood leukocytes, phagocytic (surface-binding and internalization) activity of blood neutrophils and monocytes, and reactive oxygen species (ROS) production by blood neutrophils. Postpartum parameters measured included prevalence of IMI at 3 and 10 d postpartum, milk SCC at 3 and 10 d postpartum, antibiotic residues, and milk yield through 14 days in milk (DIM). At the start of the trial, groups of 9 to 12 heifers were allotted to 1- to 2- acre pastures; groups began the trial at an average age of 6.5 mo. Once acclimated to the stanchion head-lock system constructed in each pasture, blood samples were collected from the jugular vein to obtain pre-trial sera for baseline antibody titers and mrna expression levels of L-selectin and IL-8R. Blood collection continued monthly throughout the trial until an average of 14 d prior to calving. Pretrial hip heights and body weights were also recorded and continued monthly until an average 36

of 14 d prior to calving. After collection of blood samples, a small aliquot of each sample was added to premeasured vials of TRIzol (Life Technologies, Carlsbad CA, USA), which were sent to OmniGen Research LLC to measure mrna leukocyte expression of L-selectin and IL-8R. Serum titers against S. aureus were also measured in all blood samples throughout the course of the study. For a subset of heifers (groups 6, 7, and 8), phagocytic activity and ROS production were measured. After the initial pre-trial bleeding and growth measurements, heifers began receiving the daily supplement (OmniGen-AF ) or control (no OmniGen-AF ) diets. One month after trial initiation, heifers were vaccinated with 5 cc of Lysigin (Boehringer Ingelheim Vetmedica Inc., St. Joseph MO, USA). A 5-cc booster of vaccine was administered 14 d later, and heifers were boosted semi-annually until calving. At an average of 15 mo of age, teat canal and mammary secretion samples were collected when possible, and plated for microbiological analysis and identification. Mammary secretions were also used to determine differential leukocyte counts and SCC. At approximately 4 to 6 wk prior to expected calving, all quarters were treated with Spectramast DC (Pfizer Inc., New York, NY, USA) regardless of infection status. Samples were not collected from heifers once they had been treated so as not to remove any antibiotic from the mammary gland. After calving, 3- and 10-d quarter milk samples were analyzed for SCC, and bacterial growth. Antibiotic residues, if any, were also measured in 3-d milk samples. Milk weights were also collected and analyzed to determine differences between treatment groups after 14 d in milk. Initiation of trial Groups of heifers, ranging in size from 9 to 12 heifers per group, were started on the trial once they reached an average age of 6.5 mo. Heifer numbers, corresponding group numbers, and 37

initiation dates are outlined in Appendix 1. Heifers were allowed an adjustment period before receiving treatment (feed supplement) and before sampling was performed. Each group of heifers was placed in a pasture containing a stanchion head-lock system (Figure 1) to enable animals to be secured during bleedings, samplings, growth measurements, breeding, pregnancy exams, and vaccinations. Upon being placed in the pasture, heifers were designated as either receiving the treatment diet or the control diet. Treated heifers were designated as even numbered heifers and control heifers were designated as odd numbered heifers based on ear tag identification numbers given to them as calves based on birth order. Each group contained a 50:50 ratio of treated and control heifers to allow all heifers to experience the same environmental conditions. During feeding, heifers were monitored to ensure that little to no cross-feeding occurred between treated and control animals. Figure 1. Head lock stanchion system used to feed heifers individually. 38

Growth measurements and feeding Pre-trial heights and body weights were recorded while heifers were secured in stanchions. Heights were taken (in centimeters) at the hip using a height stick (Figure 2). Weights were taken (in kilograms) around the heart girth using a weigh tape (Figure 3). The determination of heights and weights continued once a month for the remainder of the study. Figure 2. Hip height measurement taken while in head lock. Figure 3. Heart girth measurement taken while in head lock. 39

Heifer body weights and hip heights were compared between treated and control animals for groups 1 to 5 through October 2011 using population-averaged, generalized estimating equations linear regression models. Heights and weights were modeled using linear regression with robust standard errors and an exchangeable working correlation structure to account for repeated measurements with the same animals. Lowess curves were produced to visualize growth curves by age, group, and treatment. Age was centered at 14 mo (average age of the 5 groups analyzed) to avoid collinearity with the quadratic terms. A two-sided alternative hypothesis was assumed. P<0.05 was considered statistically significant. Analyses were performed with Stata, version 11.1 (StataCorp, L.P., College Station, TX). Two-way ANOVA was used to evaluate the effects of treatment (treated vs. control) and age (groups 1 to 5 from April 2010 to October 2011) on average daily gain (ADG). A two-sided alternative hypothesis was assumed and P < 0.05 was considered statistically significant. Analyses were performed using Stata version 11.1 (StataCorp, L.P., College Station, TX) OmniGen-AF contained a mixture of active dried Saccharomyces cerevisiae, dried Trichoderma longibrachiatum fermentation product, niacin, vitamin B12, riboflavin-5- phosphate, d-calcium pantothenate, choline chloride, biotin, thiamine monohydrate, pyridoxine hydrochloride, menodione dimethylpyrimidinol bisulfate, folic acid, calcium aluminosilicate, sodium aluminosilicate, diatomaceous earth, calcium carbonate, rice hulls, and mineral oil (abbreviated supplement formulation shown in Table 1). Treated heifers received OmniGen- AF at a rate of 4 g/100 lb of body weight/day. Based on this rate, the supplement contained 10% OmniGen-AF, 10% molasses to allow for binding, and 80% heifer grain mix. Control supplement was heifer grain mix only. Treatment or control diet was fed directly on the concrete slab in front of the heifers. 40

Table 1. Abbreviated list of OmniGen-AF components. OmniGen-AF label ingredients 1 Crude protein 2.5% Vitamin B-12 90 µ/lb Riboflavin 6.5 mg/lb d-panthothenic acid 69 mg/lb Thiamine 34 mg/lb Niacin 225 mg/lb Vitamin B-6 20 mg/lb Folic acid 3 mg/lb Choline 35 mg/lb Biotin 1 mg/lb 1 Complete formulation is proprietary; also includes Saccharomyces cerevisiae cell wall and Trichoderma longibrachiatum extract. The average body weight that was recorded each month determined the amount of OmnigGen-AF /grain mixture that groups would receive. Mean group weights were converted to pounds to determine the amount of treatment or control supplement to administer. This average corresponded to a designated amount of OmniGen-AF /grain mix that was fed once a day. For example, at the beginning of the trial, heifers received 0.5 lb/heifer/d of either the supplement/grain mix or 0.5 lb/heifer/d of grain as the control diet. For every 100-lb increase in group mean body weight, the amount of treatment diet increased by 0.1 lb/heifer/d. In addition to the supplement, heifers were given 11 kg/heifer/d of wheat silage or sorghum silage (sample silage ration composition shown in Table 2) depending on the season, and approximately 2.5 kg/heifer/d of grain mix (Table 3). Water and Bermudagrass hay were available to the heifers ad libitum. At approximately 2 to 3 wk prepartum, silage was topdressed with approximately 0.8 kg/heifer/d of dietary cation anion diet (DCAD) mix, 2.7 kg/heifer/d of dry cow grain (Table 4), and 4 oz/heifer/d of limestone. 41

Table 2. Wheat silage nutrient analysis. NIR analysis of small grain wheat silage Dry Matter Crude Protein 12.4% Crude Fiber 26.7% Neutral Detergent Fiber (NDF) 57.5% Acid Detergent Fiber (ADF) 35% Lignin 4.6% Total Digestible Nutrients 56.2% Phosphorous 0.24% Potassium 1.34% Calcium 0.5% Magnesium 0.21% Manganese 98 parts per million (PPM) Iron 306 PPM Aluminum 482 PPM Copper 7 PPM Zinc 33 PPM Sodium 404 PPM Table 3. Heifer grain mix ration (lb/ton). Pounds/ton Corn 905 Corn gluten feed 750.6 Soy hulls 275.3 Cotton seed meal 42 6.6 Limestone 36.8 Dicalcium phosphate 12.1 Salt 10 Magnesium oxide 5 Bovatec 1 0.4 Trace mineral pack 1 Vitamin A, D, and E 1 1 Pfizer, New York, NY. 42

Table 4. Dry cow grain mix ration (lb/ton). Pounds/ton Rolled corn 1740 Soybean meal 222 Dicalcium phosphate 17.1 Clarifly 1 9.5 Salt 8.4 Trace mineral pack 4.4 Vitamin A, D, and E 4.4 Zinpro Performance Minerals 2 3.4 Limestone 0.9 1 Akey, Lexisburg OH, USA. 2 Zinpro Corporation, Eden Prairie MN, USA. Blood collection and analysis All blood samples were taken via jugular venipuncture. While secured in the head stanchion, heifers were restrained using a halter and, if necessary, excess hair was removed using an electric clipper (Figure 4). The area of the jugular vein was sanitized with a cotton pad soaked in 70% isopropyl alcohol (Figure 5). A 4.5-mL BD vaccutainer (Becton, Dickenson and Company, Franklin Lakes NJ, USA) blood tube was pre-labeled with the heifer number and date. The blood collection tube contained buffered sodium citrate. A plastic sleeve was used to house the 18-G x 1-inch vacuette needle (Greiner Bio-One GmBH, Kremsmünster, Austria) during blood collection (Figure 6). After collection, blood samples were placed in a styrofoam cooler containing ice packs. Upon returning to the lab, 50 µl of blood was removed from the 4.5-mL sample tubes using a sterile pipette and mixed with 0.5 ml of TRIzol (Invitrogen Corp., Carlsbad CA, USA) in a 1.2-mL cryogenic vial. TRIzol is a mixture of phenol, guanidine isothiocyanate, red dye, and other components, which was used to isolate total RNA (at a later date) using manufacturer s guidelines. After adding the blood to the reagent, the sample was homogenized and frozen at - 20 C until shipment. All frozen blood samples were shipped to the OmniGen Research, LLC lab 43

in Corvallis, Oregon. The remaining blood samples were processed by centrifugation (10 min at 450 x g in a centrifuge with a swinging bucket rotor), the plasma collected and frozen at -80 C until utilized for determination of anti-s. aureus total IgG titers by ELISA. Figure 4. Clipping neck prior Figure 5. Sanitization of site Figure 6. Inserting needle to blood collection. with 70% alcohol cotton. into jugular vein. Determination L-selectin and IL-8R mrna expression Upon receipt, samples were allowed to come to room temperature and incubate for 5 min. An aliquot of 0.2 ml of chloroform was added to the vial containing the blood/trizol suspension (an additional 0.5 ml of TRIzol was added), and the mixture was shaken for 15 sec followed by incubation at room temperature for 2 to 3 min. Next, the samples were centrifuged at 12,000 x g at 4 C for 15 min, the aqueous phase was transferred into a new tube, and RNA was precipitated by mixing with 0.5 ml of isopropyl alcohol. This solution was incubated at room temperature for 10 min then centrifuged at 12,000 x g at 4 C for 10 min, and the pellet was washed with 1 ml 75% ethanol. Next, the sample was vortexed then centrifuged at 7,500 x g at 4 C for 5 min. A Bio-Rad icycler (Bio-Rad Laboratories, Hercules, CA, USA) was used to conduct a three-color polymerase chain reaction (TaqMan RT-PCR) for L-selectin, IL-8R, RPL- 44

19, plus their standards. Standards were cdna sequences for each gene of interest. L-selectin and IL-8R mrna expressions were normalized to the expression of RPL-19. Absolute quantification using the standard curves method was employed to determine the unknown amount of L-selectin and IL-8R relative to known standard curves generated using the cdna standards. Blood samples from groups 1 to 3 for 15 total mo on trial for each group were used in the analysis of L-selectin and IL-8R mrna expression. Least squares means of L-selectin and IL-8R mrna expression of treated and control heifers from groups 1 to 3 at each sample time point were compared; analyses were performed with SAS (SAS Institute Inc., Cary, NC). A two-sided alternative hypothesis was assumed; P < 0.05 was considered statistically significant. Blood mononuclear cell and neutrophil preparation for phagocytic activity and reactive oxygen species (ROS) production A subset of heifers (n=29) from groups 6, 7, and 8 was used for the analysis of phagocytic (surface-binding and internalization) ability of peripheral blood neutrophils and monocytes and for ROS production by neutrophils. A 60-cc syringe containing 1.5 ml of 100 mm EDTA fitted with a 15-G 1¾-in needle was used to obtain peripheral blood as follows: 1) the animal s head was restrained using a halter; 2) the area over the jugular vein was sanitized with a cotton ball soaked in 70% isopropyl alcohol; 3) the needle was inserted into the jugular vein; 4) 60 ml of blood was collected; and 5) the blood was transported on ice to the lab. Processing began by taking 4 aliquots of 15 ml of blood/heifer, which were placed into 4 separate centrifuge tubes. Blood was centrifuged at 2,600 rpm for 20 min and the buffy coat was collected (plus about 2 mm of red blood cells (RBC)). All aliquots from each heifer were pooled into 1 50-mL tube and then diluted to 40 ml using Ca 2+ and Mg 2+ free PBS followed by centrifugation at 2,200 rpm for 10 min to wash cells. 45

To isolate monocytes and neutrophils from the buffy coat for use in the phagocytosis and ROS assays, the buffy coat was diluted in 40 ml of PBS and layered over 10 ml lymphocyte separation medium (LSM) with a density of 1.077 gm/ml in a 50 ml tube. Next, the tube was centrifuged at 2,200 rpm for 30 min then the PBS layer was removed. The monocytes were removed from the top of the LSM using a 5-mL pipette, and placed into a 15-mL tube, and then the LSM was removed to expose the pellet. The tube was then filled with PBS and cells washed by centrifugation at 2,000 rpm for 5 min. The PBS was then removed and RBC were lysed by suspending the pellet into 10 ml of sterile water, vortexing vigorously for 40 sec, adding 10 ml double strength PBS, and filling each tube to 50 ml with PBS. Following this step, the tubes were centrifuged at 2,000 rpm for 5 min. Both pellets of isolated monocytes and neutrophils were washed and cells suspended in 10 ml of PBS. A viable cell count was performed using Trypan blue. Phagocytic activity Phagocytic activity including surface-binding and internalization of bacteria was measured according to the techniques described in Hart et al. (2011) and Wiggins et al. (2011). Differences in mean percentage fluorescence of surface-bound and internalized bacteria among neutrophils and monocytes were compared between treated and control heifers from groups 6, 7, and 8 using two-sample t-tests. Normality and homogeneity of variance were evaluated using the Kolmogorov-Smirnov test and the F-test, respectively. Analyses were performed with InStat (Graphpad Software, La Jolla, CA). A two-sided alternative hypothesis was assumed; P < 0.05 was considered statistically significant. 46

Reactive oxygen species production Reactive oxygen species production (hydrogen peroxide, hydroxyl radical, hypochlorous acid, nitric oxide, peroxide radical, peroxynitrite anion, singlet oxygen, and superoxide anion) by neutrophils was measured according to techniques described in Wiggins et al. (2011). Stimulants used in this trial included Salmonella typhimurium lipopolysaccharide (LPS) at 1 µg/ml, S. aureus peptidoglycan (PGN) at 1 µg/ml, staphylococcal enterotoxin B (SEB), S. aureus whole cell antigen at 1:100, S. aureus whole cell antigen at 1:200, S. aureus whole cell antigen at 1:400, S. aureus whole cell antigen at 1:800, and phorbol myristate (PMA) acetate at 10-7 to measure total ROS capacity. Total ROS production by neutrophils alone (without stimulation) was also measured and reported. Differences in mean ROS production as measured by arbitrary fluorescence units (AFU) among neutrophils were compared between treated and control heifers from groups 6 and 7, and group 8 separately using two-sample t-tests. Normality and homogeneity of variance were evaluated using the Kolmogorov-Smirnov test and the F-test, respectively. Analyses were performed with InStat (Graphpad Software, La Jolla, CA). A twosided alternative hypothesis was assumed; P < 0.05 was considered statistically significant. Vaccination and antibody titers One month after trial initiation, the heifers were vaccinated with 5 cc Lysigin (Boehringer Ingelheim Vetmedica Inc., St. Joseph MO, USA) in the right semimembranosus muscle of the leg. At 14 d after the initial vaccination, a booster injection of 5 cc was administered in the left semimembranosus muscle of the leg. Lysigin was composed of a lysed culture of polyvalent antigens containing phage types I, II, III, IV, and other groups of S. aureus combined with an aluminum hydroxide adjuvant. The serotypes included one serotype 5, two serotypes 8, and two serotypes 336. These serotypes were found to be associated with S. aureus 47

mastitis in dairy animals. Post-vaccination sites were monitored for 2 wk to ensure that no adverse reaction occurred. Any injection site reactions were measured, in centimeters, and recorded as well as any abscess formation (Figure 7). Revaccinations were administered every 6 mo following the initial vaccination (alternating left and right sides) until time of calving. Figure 7. Visible injection site reaction in right rear leg after vaccination. Jugular bleedings and analysis continued once a month for the remainder of the study. To determine antibody titers in heifers, a heat-killed culture of S. aureus (positive control) was prepared prior to blood sample collection and serum antibody titer determination. A bacterial culture of S. aureus from a chronically infected cow was used in the preparation of the positive control. S. aureus cultures were prepared in a brain heart infusion broth (BHI, Difco, Detroit, MI) by incubating bacteria for 48 h at 37 C. After reaching a high turbidity with the dense mass forming a pellet, samples were centrifuged at 3,000 rpm for 15 min at 4 C. Cells were pooled in 50 ml of 1xPBS and suspended by vortexing. This suspension was split into 2 50-mL sterile 48

tubes and centrifuged at 3,000 rpm for 15 min at 4 C. Pellets were then pooled in 10 ml of 1x PBS and suspended by vortexing. Bacterial cells were placed into 1 50-mL tube, which was placed in a 60 C water bath for 1 h to heat kill the bacteria. The bacteria were washed in 40 ml of 1xPBS, and this suspension was centrifuged for 15 min at 3000 rpm at 4 C. The resulting pellet was suspended in 1xPBS and sonicated at 3 cycles at 80% power using a microprobe tip in an ice bath. One-mL aliquots of the sonicate were divided and placed into a -80 C freezer for future use. The antigen was diluted in coating buffer at a ratio of 1:5000, and then 100 μl of the diluted antigen was placed into each well of a 96-well ELISA plate and incubated for 24 h at 4 C. After incubation, plates were washed 3 times with 200 μl of wash buffer per well. Next, 200 μl of blocking buffer was added to each well, and after a 1-h incubation period at room temperature, the plate was again washed with 200 μl/well of wash buffer 3 times. Heifer samples were added to the first column of ELISA plates, and 2-fold dilutions were made across the plate leaving the end column free of sample. The last column contained 4 wells of 100 μl of 1:10 diluted fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA, USA), and 4 wells of 100 μl of 1:160 diluted serum from the positive control (chronic S. aureus-infected cow). Plates were incubated for 1 hr at room temperature then washed 3x with 200 μl wash buffer. One hundred μl of rabbit anti-bovine IgG (h+1) peroxidase conjugate (diluted at 1:1,000) was added to each well and incubated for 30 min. Plates were washed 4 x with 200 μl/ well. Hydrogen peroxide (30%) was added to ABTS solution (20μL/11 ml), and 100 μl of this solution was added to each well and incubated for approximately 30 min in a dark area. Using ELISA software, the optical density (OD) was read at 405 nm. The highest endpoint was compared relative to 2x the OD of the mean FBS OD. Endpoint titers were identified and recorded for 49

groups 1 to 5 through December 2011. Jugular bleedings and antibody titer analysis continued once a month for the remainder of the study. Teat swab and mammary secretion collection and analysis Upon reaching an average age of 15 mo, quarter teat canal keratin samples and mammary secretions were collected from eligible heifers. Teat scores were recorded at this time to denote the health status of teat skin at each sampling period. A score of 1 corresponded to healthy skin with no scabs on the teat (Figure 8); a score of 2 corresponded to older scabs that were darker and red-brown in color (Figure 9); and a score of 3 corresponded to fresh scabs caused by bloodsucking horn flies that were bloody and dark red in color (Figure 10). Teat scores for groups 1 to 4 over a period of 2 months/group are presented in the discussion and results section. One of the months (per group) represents the higher teat end scores at the beginning of fly season prior to fly treatment administration, and the second mo (per group) represents teat scores 30 d after administration of fly treatment (Ultra Boss, Merck Animal Health, Summit, NJ). The additional months are not presented because administration of fly treatment kept fly populations low and teat ends healthy (mean scores of 1). Furthermore, during winter, fly treatment was not needed as fly populations remained low and teats remained healthy (mean scores of 1). 50

Figure 8. No scabs on teats Figure 9. Older scabs on teat Figure 10. Fresh scabs on teats (Score 1). (Score 2). (Score 3). After scoring and prior to sampling, any scabs at the teat orifice were removed to decrease possible contamination as well as to access the teat canal using a cotton ball soaked in 70% isopropyl alcohol. A calcium-alginate fiber-tipped sterile swab was then inserted into the sanitized teat canal orifice to a depth of 2 to 3 mm and rotated 360 to collect a keratin sample (Figure 11). The swab was placed in a sterile plastic tube containing 0.5 ml of physiological saline for storage. 51

Figure 11. Insertion of calcium alginate swab into the teat canal to collect keratin. Following teat swab collection, mammary secretions were also collected if volume was sufficient. Teat ends were sanitized again, and any secretions present were expressed from each quarter into a sterile 12-mL tube. Any signs of clinical mastitis, e.g., swelling, redness, or induration were recorded for each quarter. After sample collection, teats were dipped in a 1% iodine solution (FS-103 X, IBA, Inc., Millbury, MA) to sanitize teats and prevent the transmittance of bacteria into the teat canal. Gloves were worn during the entire sampling process to minimize contamination. Upon returning to the lab, mammary secretions were visually inspected for clots, flakes, and other signs of a clinical infection. Differences in color and consistency were also noted (Figures 12, 13, and 14). Teat swab solutions and mammary secretions were mixed by vortexing and plated on trypticase soy agar (TSA) with 5% blood plates using sterile, flamed 10-µL loops. Each loop was flamed between platings to ensure that all samples were handled aseptically. Plates were incubated for 48 hr at 37 C after which time, they were visually inspected for 52

presence of colonial growth and hemolysis (Figure 15). By presumptive identification based on colony color, morphology, and presence of hemolysis, a determination could be made regarding whether other pathogens were a Staphylococcus spp., Streptococcus spp., or coliform If needed, a catalase test was performed to determine whether the bacterial colony was a staphylococcus (catalase positive) or a streptococcus (catalase negative). A catalase test was considered positive if hydrogen peroxide was readily decomposed to water and oxygen as evidenced by bubbling of colonies after hydrogen peroxide was introduced to the colony of interest. Figure 12. Thick, glue-like secretion from an uninfected quarter (bottom) and thin, watery secretion from an infected quarter (top). 53

Figure 13. Turbid secretion from an infected quarter (left) and clinical secretion from an infected quarter illustrating the settling of clots and flakes (right). Figure 14. Turbid, white secretion from an infected quarter (left) and amber secretion from an uninfected quarter (right). 54

LF RF LR RR Figure 15. TSA 5% blood agar plate illustrating a heifer with 3 S. aureus-infected quarters (RF, LF, RR) identified by the β-hemolysis and 1 CNS-infected quarters (LR). Colonies that were creamy white to grayish-white or yellow, smooth, circular, catalase negative, and exhibited β-hemolysis were identified as S. aureus. Atypical colonies identified as staphylococci, but not S. aureus (because no β hemolysis was present) were tested using the coagulase test (Remel, Lenexa, KS) and mannitol salt agar test (MSA). A coagulase test was positive if the 0.5 ml plasma reagent inoculated with 3 to 5 typical colonies of interest coagulated within 24 h of incubation (Figure 16). A positive test verified that the bacteria contained the coagulase enzyme, allowing the plasma to clot within the test tube. A MSA test was positive if the bacteria colonized the agar surface and if the plate turned a bright yellow, indicating that the pathogen was able to grow in a high saline environment and fermented mannitol to produce the color change (Figure 17). A coagulase-positive, MSA-positive result was categorized as S. aureus given its growth requirements and characteristics despite the lack of β-hemolysis. These atypical S. aureus samples (nonhemolytic, but coagulase and MSA positive) 55

were described as colonies that were creamy to gray in color, displayed a small, white raised dot in the center of the circular colony, may exhibit a narrow zone of α-hemolysis, and displayed a greenish hue around the culture. Figure 16. Results of coagulase test showing a positive result (solidifying of plasma) exhibited by inverted tube. LF RF LR RR Figure 17. Results of MSA test showing 1 MSA+ sample (LR) and 3 MSA- samples (RF, LF, RR). 56

Samples that were identified as S. aureus using these tests were documented and distinct characteristics noted. Samples that were not identified as S. aureus (coagulase positive/msa negative or coagulase negative/msa negative) were tested using the API Staph test system (biomérieux, Marcy l'etoile, France). This system utilized biochemical and fermentation tests, results of which were entered in the associated database following which an identification was provided based on the tests completed (Figure 18). If an identification was not provided using this system, the bacteria were classified as coagulase-negative staphylococci (CNS) or coagulase-positive staphylococci (CPS) based on previous tests. In the presentation of results and discussion of findings, CPS will refer to any isolates that were nonhemolytic coagulasepositive, MSA-negative. Figure 18. API Staph identification test showing a positive test for S. hyicus. Samples that were identified as streptococci based on the catalase reaction were further tested using the Slidex Strepto Plus test (biomérieux, Marcy l'etoile, France). This test utilized 3 to 5 typical colonies of interest, which were then added to 0.4 ml of extraction enzyme (increases antigen yield) and vortexed. After a 10-min incubation period at 37 C and another vortex, 15 μl of the suspension was placed on 6 circles of a test kit corresponding to 6 different group antigens (Groups A to G, excluding E). Latex particles were mixed with each of the 15- μl aliquots of suspension. A positive test was evident after 2 min if agglutination of the latex 57

was visible (Figure 19). If there was agglutination in an ungroupable antigen category or no agglutination at all, an API 20 Strep test (biomérieux, Marcy l'etoile, France) was conducted. This test was conducted in a similar fashion to the staphylococcal test (API Staph test); however, the biochemical and fermentation tests were different. Any sample that could not be identified based on presumptive analysis or biochemical testing was tested by completing a Gram stain. Based on microscopic morphology, the isolate was classified and/or identified as a gram-positive or gram-negative coccus or bacillus. Figure 19. Results of Slidex test showing a positive agglutination test in 6/G denoting an ungroupable Streptococcus spp. An API 20 Strep test was subsequently conducted to confirm Strep. uberis. The prevalences of mastitis among quarters of treated and control heifers in groups 1 to 3 through October of 2011 were compared using population-averaged, generalized estimating equations logistic regression models. Robust standard errors and an exchangeable working correlation structure were used to account for the clustering of quarters within cows. A two- 58

sided alternative hypothesis was assumed. Analyses were performed with Stata, version 11.1 (StataCorp, L.P., College Station, TX); P < 0.05 was considered statistically significant. Differences in the cow-level prevalence of mastitis among treated and control heifers in groups 1 to 4 through October 2011 were compared using Fisher s exact test. A two-sided alternative hypothesis was assumed. Analyses were performed with Stata, version 11.1 (StataCorp, L.P., College Station, TX); P < 0.05 was considered statistically significant. Colony-forming unit (CFU) scores were also recorded for isolates from secretions of each quarter based on the following scale: a CFU score of 0 was assigned to uninfected quarters; a CFU score of 1 was identified as less than 20 CFU; a CFU score of 2 was identified as 20 to 200 CFU; and a CFU score of 3 was identified as greater than 200 CFU. Average CFU score for each bacterial species was determined. Additionally, average CFU score for quarters from treated animals was compared to average CFU score for quarters from control animals in groups 1 to 3 through October 2011. Total leukocyte counts The SCC of mammary secretions were determined using the DeLaval Direct Cell Counter (DeLaval, Tumba, Sweden) if volume allowed. After several months of sampling, a maximum SCC of approximately 6,000 x 10 3 cells/ml was identified. Secretions that were clinical in nature were retrospectively assigned a SCC of 7,000 x 10 3 cells/ml because the actual cell count could not be determined due to clots and flakes in the samples. Mean SCC was calculated for groups 1 to 3 through October 2011. Differential leukocyte counts Mammary secretions were used to prepare differential smears for enumeration of neutrophils, macrophages, and lymphocytes. For the preparation of the differential smear, 50 µl 59

of 7.5% bovine serum albumin (BSA) and 25 µl of secretion sample were added to a cytospin well. After being secured in a metal holder with a clean microscope slide, the prepared secretion sample was placed in a Cytospin 2 centrifuge (Shandon, Pittsburgh PA, USA) and operated for 2 min at 1200 rpm. After the slide was removed and air-dried, the smear was stained using the Wright stain method (Wright, 1902). Once dry, the sample was examined at 1000x under an oil immersion lens, and percentages of lymphocytes, macrophages, and neutrophils were recorded (Figure 20). A total of 100 cells/slide were counted to determine the population distribution. Differences in leukocyte populations between infected (Figure 21) and uninfected quarters (Figure 22) were examined as well as treatment differences. Mean differential leukocyte counts were analyzed for groups 1 to 3 through October 2011; heifers in these groups were developed sufficiently to produce an appropriate volume of mammary secretion for study. PMN MA LM Figure 20. Differential leukocyte stain illustrating, LM=Lymphocyte, PMN=Neutrophil, and MA=Macrophage. 60

PMN Figure 21. Differential leukocyte staining illustrating cells from an infected quarter demonstrating a predominance of neutrophils. MA Figure 22. Differential leukocyte stain illustrating cells from an uninfected quarter demonstrating a predominance of macrophages. 61

Determination of new IMI and TCI, chronic IMI and TCI, spontaneous cures, and success of DCT The numbers of new IMI and teat canal infections (TCI), chronic IMI and TCI, spontaneous cures, and success of antibiotic treatment for each quarter from which teat keratin samples and mammary secretions were collected were determined as follows: 1) A new IMI (or TCI) was identified as an IMI (or TCI) that was not present at the first sampling, but was present in at least 2 out the 3 consecutive subsequent samplings (including the first sampling); 2) a chronic infection was identified as an IMI that was present in at least 3 out of 4 consecutive samplings; 3) a spontaneous cure was identified as a quarter that cultured negative for at least 2 consecutive samplings after having established a new or chronic IMI; and 4) a successful prepartum treatment was identified as a quarter that was treated with SpectramastDC and cured for at least 2 consecutive samplings during the trial (efficacy of treatment administered 30 to 60 days prior to calving was determined using 3-d milk culture results). Totals were determined for treatment and control groups for teat keratin samples and for mammary secretion samples, and then averaged on a per heifer basis to determine the differences in treatment and control. For example, out of 11 heifers receiving the feed supplement eligible for analysis in this category, 26 new IMI were diagnosed in mammary secretions over the course of the study period for an average of 2.5 quarters infected per heifer. Similarly, out of 14 unsupplemented control heifers eligible for analysis in this category, 31 new IMI were diagnosed in mammary secretions over the course of the study period for an average of 2.6 quarters infected per heifer. Groups 1 to 3 from August 2010 through October 2011 were used in the analysis of these parameters. 62

Postpartum data collection A total of 16 heifers (7 treated, 9 control) from groups 1 to 3 have calved and were used in postpartum analysis. Tables are presented in results and discussion with heifer and group numbers. When animals were 30 to 60 d from calving, final teat canal keratin and mammary secretions sample were collected, then each quarter was infused with 5 cc of Spectramast DC (Figure 23). From this point on, the heifer was not sampled to ensure that the antibiotic would remain in the mammary gland. Once the heifer calved, 3- and 10-d postpartum milk samples were collected. The 3-d quarter milk samples were tested for antibiotic residues using a Delvotest (DSM Food Specialties USA, Inc., Parsippany NJ, USA). A positive or negative result was identified by the color change after a 3-h incubation (Figure 24). A positive reaction was represented by a blue color, whereas yellow was indicative of a negative result (no residue present). The 3- and 10-d milk samples were analyzed to determine SCC using the same technique as used for the heifer mammary secretions. All SCC values were log transformed, and differences in mean log SCC at 3 d and 10 d postpartum were compared between treated and control heifers using two-sample t-tests. Analyses were performed with Stata, version 11.1 (StataCorp, L.P., College Station, TX). A two-sided alternative hypothesis was assumed; P < 0.05 was considered statistically significant. The 3- and 10-d samples were also plated for microbiological growth and analysis on 5% blood TSA plates. If any bacteria were present on a plate, a coagulase test, MSA test, and/or Gram stain were conducted to identify the pathogen. Differences in the cow-level prevalence of mastitis among treated and control heifers were compared using Fisher s exact test. Analyses 63

were performed with Stata, version 11.1 (StataCorp, L.P., College Station, TX). A two-sided alternative hypothesis was assumed; P < 0.05 was considered statistically significant. Daily milk weights through 14 d in milk (DIM) were collected for each heifer after calving using the ALPRO (DeLaval, Tumba, Sweden) system installed at the UGA Teaching Dairy. For each heifer, all milk weights through 14 DIM were summed to determine total milk production/heifer. Differences in mean milk yield were compared between treated and control heifers using two-sample t-tests. Analyses were performed with Stata, version 11.1 (StataCorp, L.P., College Station, TX). A two-sided alternative hypothesis was assumed; P < 0.05 was considered statistically significant. Figure 23. Infusing dry cow product into the right rear quarter of heifer using the partial insertion technique. 64

Figure 24. Delvotest illustrating positive results for the RF quarter and negative results for the control, composite, and LF, LR, and RR quarters. 65

CHAPTER 3 RESULTS AND DISCUSSION As outlined in the Materials and Methods section, due to the initiation of animal groups into the trial at differing time periods because of minimum age requirements, different groups were used in various sections of this chapter. Groups used in the analysis of various parameters are described prior to presentation of results and discussion of findings. Effects of a general immunostimulant feed supplement on heifer growth as measured by average daily gain, body weight, and hip height Average daily gain For analysis of heifer growth, groups 1 through 5 (April 2010 through October 2011) were used due to availability of growth data through a minimum of 1 yr of age (refer to Appendix 1 to observe group initiation dates). Overall average daily gain (ADG) across all 5 groups (n = 54) was 1.81 lb/d (±0.29 lb) for treated heifers (n = 28) and 1.73 lb/d (±0.27 lb) for control heifers (n = 26) (Figure 25). Among groups, treated heifers tended to have a higher ADG compared to control heifers, with the exception of group 3. The ADG for treated heifers in group 3 was 1.70 lb/d compared to an ADG of 1.78 lb/d for control heifers. In general, the heifers that were enrolled in this study grew at a similar rate. However, there was no significant difference between treated and control animals (P = 0.291). Mean ADG for the 5 groups as well as the overall ADG (Figure 25) generally fell within the range of 1.68 to 1.84 lb/d described by Hoffman (1997) in a review of growth rates of Holstein replacement heifers. Heifers from groups 1 and 2, however, had lower average ADG than the other groups. The heifers in groups 1 and 2 were born in the summer months, August and September, which may have affected their 66

Average daily gain (lb/d) ADG during prepubertal and postpubertal growth periods. Additionally, these animals were older at the time of analysis so their growth had most likely peaked, whereas groups 3, 4, and 5 were younger and still growing. O Brien et al. (2009) found that heat stress-inducing temperatures resulted in decreased dry matter intake (12%). While not significantly different, O Brien also found that heat-stressed calves had a lower ADG than calves kept in a thermoneutral climate. At approximately 1 yr (±17 d) of age, the average weight for the 54 Holstein heifers in the present trial was 328.9 kg (± 30.6 kg), which is greater than the average weight of 294 kg for year-old Holstein heifers in the Southeast (Heinrichs and Losinger, 1998). Thus, ADG and average body weight for yearlings in the present study were similar to those weights observed in general for Holstein heifers. Figure 25. Mean average daily gains (ADG) ± SD for groups 1 to 5 1 and overall ADG between treated and control heifers. Treated Control 2.5 2 1.5 1 0.5 2.08 1.97 1.88 1.95 1.81 1.78 1.73 1.70 1.63 1.70 1.61 1.47 0 Group 1 Group 2 Group 3 Group 4 Group 5 Overall 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n = 12) treated (n = 6) and control (n = 6); group 3 (n = 12) treated (n = 5) and control (n = 7); group 4 (n = 9) treated (n = 4) and control (n = 5); and group 5 (n = 12) treated (n = 7) and control (n = 5). 67

Body weight Results for the multivariable generalized estimating equations linear regression model for the prediction of body weight over time are shown in Table 5. Age was centered at 14 mo of age to avoid collinearity. Treated heifers weighed 3.57 kg less than the control heifers at 14 mo of age; however, this difference was not significant (P = 0.613). The test to determine differences in heifer weights among groups 1 to 5 yielded a P-value of < 0.001, indicating that there was a weight difference among the 5 groups at 14 mo of age. Referent to group 1, group 5 was the lightest group (88.6 kg lighter than group 1), whereas group 1 was the heaviest group at 14 mo of age. After accounting for growth over time and interaction terms (Age, Age 2, Treatment X Age, Treatment X Age 2, Group X Age, and Group X Age 2 ), the joint test of Treatment X Age and Treatment X Age 2 yielded a P-value of 0.063, which coincidentally was the same P-value as Treatment X Age 2. This indicates that the change in weight over time was not significantly different between treated and control heifers, although, the P-value approaches significance as evidenced by the diverging growth curves between treated and control heifers at 20 to 25 mo of age (Figure 26). However, groups 1 and 2 are the only groups (out of the 5 analyzed) that have data beyond 20 mo. Based on Lowess curves for each age group (Figure 27), it is probable that group 2 weight values at 20 to 25 mo contributed to the previous P-value (0.063). The 5 groups, regardless of treatment, exhibited significantly different growth curves as indicated by the joint test of Group X Age and Group X Age 2 that yielded P-value < 0.001. Figure 28 shows the different growth curves among the 5 groups. While treated heifers had a greater overall mean ADG of 1.81 lb/d compared to overall mean ADG of 1.73 lb/d (Figure 25), multivariable generalized estimating equations linear regression showed that, while not significant, treated heifers weighed 3.57 kg less than control 68

heifers (Table 5). The reason for this relationship is that control heifers began the trial heavier (171.1 kg) compared to treated heifers (165.6 kg). The differences in beginning weights were even more exaggerated in groups 2, 3, and 4 in which control heifers began the trial 8 to 16 kg heavier than treated heifers. 69

Table 5. Multivariable generalized estimating equations (GEE) linear regression model for the prediction of body weight (kg) in Holstein heifers that received a dietary supplement (treated, n = 28) or that served as unsupplemented controls (n = 26). Variable Treatment Treated Control Coefficient (Robust SE) -3.57 (7.07) Referent 95% Confidence interval -17.4, 10.3 P-value 0.613 Group 1 2 3 4 5 Referent -38.9 (6.84) -50.1 (9.21) -59.2 (10.7) -88.6 (11.6) -52.3, -25.5-68.1, -32.0-80.2, -38.1-111, -65.9 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 * Age 28.6 (0.646) 27.3, 29.9 < 0.001 ( * Age) 2-1.21 (0.057) -1.32, -1.10 < 0.001 Treatment X * Age 0.848 (0.710) -0.544, 2.24 0.233 Treatment X ( * Age) 2 0.135 (0.072) -0.007, 0.277 0.063 Group X * Age 1 2 3 4 5 Referent -3.56 (0.863) -6.72 (1.15) -4.49 (1.28) -12.5 (3.33) -5.25, -1.87-8.97, -4.46-7.01, -1.98-19.1, -6.02 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 Group X ( * Age) 2 1 2 3 4 5 Referent 0.190 (0.073) 0.371 (0.130) 0.424 (0.169) -0.200 (0.464) 0.048, 0.332 0.116, 0.625 0.093, 0.754-1.11, 0.709 0.002 0.009 0.004 0.012 0.666 Constant 429 (5.86) 418, 441 < 0.001 * Age was centered (i.e., age in months 14) to avoid collinearity with the quadratic term. 70

Weight (kg) Weight (kg) Figure 26. Overall Lowess curves for heifer body weight across all groups for control and treated heifers. 600 Lowess Curves for Heifer Body Weight 500 400 300 200 100 Control Treated 5 10 15 20 25 Age (months) Figure 27. Lowess curves for heifer body weight by group for control and treated heifers. 600 Lowess Curves for Heifer Body Weight 1 2 3 400 200 0 5 10 15 20 25 600 4 5 400 200 0 5 10 15 20 25 5 10 15 20 25 Age (months) Control Treated 71

Weight (kg) Figure 28. Lowess curves for heifer body weight, by group, for all heifers regardless of treatment. 600 Lowess Curves for Heifer Body Weight 500 400 300 200 100 5 10 15 20 25 Age (months) Group 1 Group 2 Group 3 Group 4 Group 5 Hip height The results for the multivariable generalized estimating equations linear regression model for the prediction of hip height are shown in Table 6. Treated heifers were 0.739 cm shorter than the control heifers at 14 mo of age; however, this was not significant (P = 0.341). The test to determine differences among heifer hip heights among groups 1 to 5 yielded a P-value of 0.001, indicating a difference in hip heights among the 5 groups at 14 mo of age. All groups were shorter than group 1 at 14 mo except for group 2, which was 0.368 cm taller. After accounting for growth over time and interaction terms (Age, Age 2, Treatment X Age, Treatment X Age 2, Group X Age, and Group X Age 2 ), the joint test of Treatment X Age and Treatment X Age 2 yielded a P-value of 0.622. This indicated that the change in height over time was not significantly different between treated and control heifers (Figure 29). Among the 5 groups, regardless of treatment, heifers grew significantly different over time as indicated by the joint 72

test Group X Age and Group X Age 2 that yielded a P-value of < 0.001. Graphically, the difference in groups can be observed (Figure 30). Figure 31 shows the growth curves supporting the test determining that the groups grew differently in height over time. Overall, heifers began the trial at an average hip height of 112.1 cm at 6.5 mo of age. The average hip height of heifers at the beginning of the present trial was greater than the average hip height (101 cm) for Holstein heifers at 6 mo of age found by Brown et al. (2001) after measuring 35 Holstein heifers, and was most likely due to the average age of the present study s heifers being slightly greater than 6 mo. Brown et al. (2001) found that at 14 mo of age, the Holstein heifer was an average of 125 cm, whereas the heifers in the present study were an average of 137 cm. Each animal in the present trial was not measured at exactly 14 mo of age, which may be resulting in the numerical differences between the present trial and those of Brown et al. (2001). Other factors that may be leading to the observed differences in height include differences in genetics of heifers and nutritional programs. 73

Table 6. Multivariable generalized estimating equations (GEE) linear regression model for the prediction of hip height (cm) in Holstein heifers that received a dietary supplement (treated, n = 28) or that served as unsupplemented controls (n = 26). Variable Treatment Treated Control Coefficient (Robust SE) -0.739 (0.766) Referent 95% Confidence interval -2.23, 0.773 P-value 0.341 Group 1 2 3 4 5 Referent -3.00 (0.885) 0.368 (0.949) -.393 (0.995) -3.79 (1.40) -4.71, -1.27-1.49, -2.23-2.34, 1.56-6.52, -1.05 < 0.001 0.001 0.698 0.693 0.007 * Age 2.22 (0.054) 2.12, 2.33 < 0.001 ( * Age) 2-0.100 (0.005) -0.110, -0.091 < 0.001 Treatment X * Age -0.017 (0.058) -0.130, 0.097 0.770 Treatment X ( * Age) 2-0.003 (0.006) -0.015, 0.010 0.663 Group X * Age 1 2 3 4 5 Referent -0.070 (0.061) 0.125 (0.083) 0.153 (0.106) -0.201 (0.419) -0.189, 0.049-0.038, 0.289-0.055, 0.362-1.021, -0.619 0.037 0.248 0.133 0.149 0.631 Group X ( * Age) 2 1 2 3 4 5 Referent 0.008 (0.005) -0.046 (0.009) -0.082 (0.017) -0.064 (0.050) -0.001, 0.017-0.064, -0.028-0.116, -0.049-0.163, 0.035 < 0.001 0.099 < 0.001 < 0.001 0.204 Constant 138 (0.618) 136, 139 < 0.001 * Age was centered (i.e., age in months 14) to avoid collinearity with the quadratic term. 74

Height (cm) Height (cm) Figure 29. Overall Lowess curves for heifer hip height across all groups for control and treated heifers. 150 Lowess Curves for Heifer Height 140 130 120 110 100 Control Treated 5 10 15 20 25 Age (months) Figure 30. Lowess curves for heifer hip height, by group, for control and treated heifers. 160 Lowess Curves for Heifer Height 1 2 3 140 120 100 5 10 15 20 25 160 4 5 140 120 100 5 10 15 20 25 5 10 15 20 25 Age (months) Control Treated 75

Height (cm) Figure 31. Lowess curves for heifer hip height, by group, for all heifers regardless of treatment. 150 Lowess Curves for Heifer Height 140 130 120 110 100 5 10 15 20 25 Age (months) Group 1 Group 2 Group 3 Group 4 Group 5 Effects of a general immunostimulant feed supplement on the immune system as measured by L-selectin mrna and IL-8R mrna expression L-selectin mrna expression For analysis of L-selectin mrna and IL-8R mrna expression in blood leukocytes, groups 1 through 3 were chosen as representative samples. Sample time 0 represents pre-trial (prior to feed supplementation) and sample times 1 to 15 represent each month s collection after beginning the feed supplement or control diet. For the majority of the sample time periods shown in Figure 32, treated heifers exhibited increased overall L-selectin mrna expression compared to control heifers (P = 0.001). Prior to the diet being administered (sample time 0), blood leukocytes from treated heifers exhibited a mean L-selectin mrna expression of 65 fg/pg compared to blood leukocytes from control heifers having an expression of 79 fg/pg. Thirty days after trial initiation (sample time 1), L-selectin mrna expression for treated heifers increased to 76

L-selectin conc./ RPL-19 conc. (fg/pg) 116 fg/pg, whereas expression for the control heifers decreased to 28 fg/pg. Throughout the trial, similar trends for the mrna expression of L-selectin on blood leukocytes were observed for treated and control heifers. By 15 mo, L-selectin mrna expression was still greater in treated heifers (91 fg/pg) than control heifers (77 fg/pg). At each sample time point, with the exception of time 0, L-selectin mrna expression was numerically greater in treated heifers indicating a positive advantage in L-selectin mrna expression on blood leukocytes when supplemented with OmniGen-AF. Although, the overall effect of the feed supplement was significant (P = 0.001) when analyzing least square means, only the following individual sample times represent periods when L-selectin mrna expression in treated heifers was significantly greater than L-selectin mrna expression in control heifers: 7 mo (P = 0.043), 12 mo (P = 0.006), and 14 mo (P = 0.046). Figure 32. Mean L-selectin mrna expression (± SE) of blood leukocytes in treated and control heifers by sample time (mo) for heifer groups 1, 2, and 3 1. 450 400 350 300 250 200 150 100 50 0 Treated 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n=12) treated (n = 6) and control (n =6); and group 3 (n=12) treated (n = 5) and control (n = 7). *Least squares mean L-selectin mrna expression was greater in blood leukocytes from treated heifers vs. control heifers (P = 0.043, P = 0.006, and P = 0.046). * Control 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sample time (mo) * * 77

IL-8R mrna expression Similar results were found when analyzing IL-8R mrna expression. Treated heifers tended to have increased expression of IL-8R mrna compared to control heifers (Figures 33). Prior to starting the diet, treated heifers exhibited an IL-8R mrna expression of 1.7 pg/pg, whereas leukocytes from control heifers exhibited an IL-8R expression of 2.5 pg/pg. Thirty days after trial initiation (sample time 1), control heifers still had a higher IL-8R expression at 2.6 pg/pg, which was numerically similar to the pre-trial expression. Treated heifers, on the other hand, increased to 2.2 pg/pg. By 15 mo, treated heifers had a greater IL-8R mrna expression (1.9 pg/pg) than control heifers (1.6 pg/pg). The overall effect of the feed supplement was not significant (P = 0.106) when analyzing least square means; however, the sample taken at 4 mo (P = 0.030) was significant for greater IL-8R mrna expression in treated heifers compared to IL- 8R mrna expression in control heifers. 78

IL-8R conc./ RPL-19 conc. (pg/pg) Figure 33. Mean IL-8R mrna expression (± SE) of blood leukocytes in treated and control heifers by sample time (mo) for heifer groups 1, 2, and 3 1. 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 * Treated Control 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sample time (mo) 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n=12) treated (n = 6) and control (n =6); and group 3 (n=12) treated (n = 5) and control (n = 7). *Least squares mean L-selectin mrna expression was greater in blood leukocytes from treated heifers vs. control heifers (P = 0.030). These numerical differences in L-selectin and IL-8 R mrna expression suggest that the treatment diet was being provided in sufficient amounts to result in immunological marker expression changes. The observed increases in L-selectin mrna expression in whole blood leukocytes is similar to the increase in L-selectin mrna expression in neutrophils observed in previous experiments by Wang et al. (2007). L-selectin, expressed by neutrophils on their cell surfaces, interacts weakly with the vascular endothelium, which allows the neutrophil to roll along the interior vessel walls and to monitor surfaces for the presence of signals, such as cytokines, which indicate a bacterial invasion at the site of local inflammation. Once adhesion has occurred, other cytokines, such as IL-8, initiate the shedding of L-selectin from the surface of the neutrophil, which allows more adhesive molecules to be expressed. IL-8 also plays an important role in production of ROS, which is instrumental in killing of bacteria (Paape et al., 2003; Forsberg, 2004). 79

While the mechanism by which the feed supplement enhances L-selectin and IL-8R mrna expression in dairy heifers is unknown, it is hypothesized that yeast and fungal cell wall components in the supplement contain molecules that interact with the innate immune system to prime leukocytes for antimicrobial processes, specifically neutrophils (Harris et al., 2006; Heine and Ulmer, 2005). Alternatively, the interaction of supplement components (yeast and fungal cell walls) in the gut may initiate the innate immune response leading to a cascade of events associated with enhanced adhesion molecules (L-selectin) as well as increased proinflammatory cytokine receptors (IL-8R). Additionally, increases in cytokine receptors, e.g. IL-8R, may indicate that the leukocytes are more primed and receptive to signaling once a pathogen is recognized. Significant increases in L-selectin and IL-8R mrna expression enhanced by nutritional stimulation signify a direct relationship between nutrition and the innate immune response. Effects of a general immunostimulant feed supplement on vaccine efficacy as measured by anti-s. aureus total IgG titers Groups 1 through 5 were used to determine anti-s. aureus total IgG titers at time of initial vaccination, and monthly through the trial to include at least 1 semi-annual vaccination. Mean anti-s. aureus IgG titers among the groups, regardless of treatment, are shown in Figure 34 (Appendices 2, 3, 4, 5, and 6). Overall, Group 1 did not exhibit a positive response to the initial vaccination or to the booster. Groups 2 and 3 exhibited a small increase in anti-s. aureus total IgG titers after initial vaccination, e.g., 1 mo later, with group 3 displaying a greater response. Groups 4 and 5 did not exhibit a response 1 mo following the vaccination. Titers did not increase in most groups until the spring of 2011. During this time, horn fly populations increased dramatically in the pastures where the heifers were located, and the 80

probable influence of the fly population on increasing anti-s. aureus titers is represented by the dramatic increase in titers beginning in April of 2011 and continuing through August of 2011. This increase was most likely a result of natural vaccination initiated by horn flies transmitting S. aureus. For example, Owens et al. (1998) observed that after S. aureus-infected horn flies were allowed to feed off of heifer teat ends, abscess formation occurred due to the fly bites. Horn flies inserted their proboscis into teat end capillaries to draw blood, and in the process injected S. aureus into the skin, resulting in small abscesses and formation of scabs (Owens et al. 1998). Analysis of scab homogenates from these abscesses revealed that the same strain of S. aureus from the horn flies was isolated from both specimens (scabs and horn flies). Additionally, Nickerson et al. (1995) found that teats with abscesses and lesions caused by the horn flies exhibited an IMI prevalence of 70% compared to an IMI prevalence of 40% in heifers whose teats were not exposed to the flies. The association between horn flies and S. aureus IMI, abscesses on teat ends, and S. aureus transmission is supported by these studies (Nickerson et al., 1995; Owens et al., 1998) as well as the present study. It is hypothesized that the horn fly bites and subsequent abscess formation act as a natural immunization against S. aureus. The ineffective response to the artificial vaccination (Lysigin ) may have been that the serotype used in the vaccine was not being recognized when ELISA assays were conducted because the whole cell S. aureus antigen prepared in the lab was of a different serotype. Among the 5 groups, treated and control animals exhibited similar total IgG titers against S. aureus. Following the initial vaccination and 14-d booster, no trends were observed between treated and control animals in groups 1 to 5. Overall examination of treated vs. control animals displayed similar results (Figure 35). Actual titers for groups 1 to 5 (treated and control heifers) can be found in Appendices 2-6. Titers were not enhanced in supplement-fed animals. The 81

reason for increases in antibody titers observed between mo 10 and mo 14 is attributed to the increases in the fly population as observed in Figure 34. While the heifers (regardless of treatment) did not respond effectively to the initial vaccination, they did appear to respond, however minimally, to the semi-annual vaccination at 6 mo, as titers doubled 1 mo later (7 mo), if only transiently. Regardless of treatment group, flies appeared to affect anti-s. aureus total IgG titers more than the vaccine itself as evidenced by the increase in titers observed beginning in mo 10 (May 2011). Nickerson et al. (1993) observed a 4.7-fold increase in S. aureusvaccinated heifers compared to controls and pretreatment titers, whereas in the present study, the greatest increase, in relation to vaccination, was a transient 2-fold increase in titers (7 mo) after the 1 st semi-annual vaccination at 6 mo. An increase in titers at 13 mo following the 2 nd booster administered at 12 mo was not observed, which was similar to the lack of response noted after the initial vaccination (mo 0 and 1). 82

Mean anti-s. aureus IgG titers Figure 34. Comparison of mean anti-s. aureus total IgG titers, by group 1, prior to and following 2 initial 3, booster 4, and semi-annual vaccination 5. 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 GROUP 1 GROUP 2 1 1 GROUP 5 1 Group 1 (n = 9); group 2 (n = 12); group 3 (n = 12); group 4 (n = 9); and group 5 (n = 12). 2 Colored arrows correspond to 30 d after initial vaccination and 14 d after booster injection. 3 Initial vaccination dates: Group 1, 06/2010; group 2, 07/2010; group 3, 08/2010; group 4, 11/2010; and group 5, 05/2011. 4 Booster vaccinations were administered 14 d following initial vaccination. 5 Semi-annual vaccinations for groups 1 to 5 were administered 6 mo after initial vaccination and the succeeding vaccination. Sample date 83

Mean anti-s. aureus IgG titers Figure 35. Comparison of mean anti-s. aureus total IgG titers between treated and control heifers 1, regardless of group, prior to (month 0) and following initial 2, booster 3, and semi-annual vaccination 4. 5000 4500 4000 3500 3000 2500 2000 1500 1000 500 0 1 Treated (n = 26) and control (n = 28). 2 Initial vaccination administered at month 0. 3 Booster administered 14 d after initial vaccination between months 0 and 1. 4 Semi-annual vaccination administered at mo 6 and mo 12. Arrows mark these booster vaccinations. Treated Control 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after initial vaccination Effects of a general immunostimulant feed supplement on neutrophil and monocyte phagocytosis and reactive oxygen species production Phagocytic activity As a subset, groups 6, 7, and 8 (group 6 (n = 10) treated (n = 5) and control (n = 5); group 7 (n = 10) treated (n = 5) and control (n = 5); and group 8 (n = 9) treated (n = 4) and control (n = 5) were used in the analysis of phagocytic activity and ROS production to determine treatment differences 1) before supplementation (d 0), 30 d after the start of supplementation (d 30), and 60 d after the start of supplementation (d 60). On d 0 (prior to feed supplementation), there were no significant differences in the overall phagocytic activities (surface-bound and internalized bacteria) of blood neutrophils and monocytes collected from control and treated heifers 84

regardless of bacterial species (E. coli and S. aureus) used in culture (Table 7). Similar phagocytic activities between groups were expected as the dietary supplement had not yet been administered to the animals. On d 30, there was a significant increase in surface-bound (21.5% vs. 37.8%) and internalized E. coli (18% vs. 28.7%) in neutrophils (but not monocytes) collected from treated heifers compared to neutrophils collected from control heifers. Additionally, there was a tendency (P < 0.10) for neutrophils from treated heifers to have increased surface-binding (18.9% vs. 22.8%) and internalization (18.0% vs. 22.6%) of S. aureus on d 30 compared to neutrophils from control heifers on d 30. There were no significant differences observed in the phagocytic ability of monocytes at d 30. On d 60, there was a tendency for neutrophils from treated heifers to exhibit increased surface-binding (20.6% vs. 26.8%) and internalization 16.6% vs. 20.3%) of E. coli compared to neutrophils from control heifers. Surface-binding ability of S. aureus at d 60 tended to increase (P < 0.1) in neutrophils from treated heifers compared to controls (14.5% vs. 17.1%) but not internalized S. aureus. There were no significant differences observed in the phagocytic activity of monocytes at d 60. 85

Table 7. Independent comparisons of percent fluorescence of surface-bound and internalized fluorescein-labeled E. coli and S. aureus by blood neutrophils and monocytes collected from control and treated heifers on days 0, 30, and 60. Day 0 Day 30 Day 60 Control Treated Control Treated Control Treated E. coli Neutrophils Surface-bound 21.1 23.8 21.5 a 37.8 b 20.6 c 26.8 d Internalized 15.7 16.3 18.0 a 28.7 b 16.6 c 20.3 d Monocytes Surface-bound 19.4 17.2 21.6 20.6 13.8 15.0 Internalized 17.7 14.2 18.8 17.0 11.9 13.0 S. aureus Neutrophils Surface-bound 19.7 17.8 18.9 c 22.8 d 14.5 c 17.1 d Internalized 18.4 16.4 18.0 c 22.6 d 12.5 13.6 Monocytes Surface-bound 11.2 9.0 13.1 11.6 8.7 8.5 Internalized 10.9 9.2 12.9 11.8 8.0 8.1 a,b Means within rows and comparisons with different subscripts differ (P < 0.05). c,d Means within rows and comparisons with different subscripts differ (P < 0.10). ROS production Groups 6 and 7 were analyzed separately from group 8 for ROS (hydrogen peroxide, hydroxyl radical, hypochlorous acid, nitric oxide, peroxide radical, peroxynitrite anion, singlet oxygen, and superoxide anion) production. Groups 6 and 7 blood samples were not collected and analyzed for ROS production on d 0. There were no significant differences in ROS production by neutrophils collected from treated and control heifers at d 30 (P > 0.05) although numerically, ROS production was higher in neutrophils stimulated by all whole cell S. aureus dilutions (100, 200, 400, and 800) of treated vs. control heifers (Figure 36). At d 60, ROS production by leukocytes from treated animals was numerically greater after each type of stimulation compared to leukocytes from controls, and there was a significant difference in ROS production by neutrophils collected from treated heifers (177 AFU) after PMA stimulation compared to neutrophils from control heifers (149 AFU) (P < 0.05) (Figure 37). Similarly, 86

ROS production (AFU) Wiggins et al. (2011) identified an increase in ROS production following stimulation with PMA, whereas there were no differences after LPS and PGN stimulation. Likewise, Heyneman et al. (1990) found greater ROS production after neutrophils were stimulated with PMA. The ROS production stimulated by PMA indicates the greatest capacity of the neutrophil to produce ROS by initiating activation of neutrophil oxidase (Li et al. 2001). The results of the present trial show that neutrophils from treated animals may be primed and more readily available to begin producing ROS for bacterial killing. Additionally, these data indicate that it may take up to 60 d for a difference to be detected in leukocyte function. At 60 d, ROS production was numerically higher in neutrophils stimulated by all S. aureus dilutions of treated vs. control heifers. Figure 36. Comparison of reactive oxygen species (ROS) production (±SD) by neutrophils collected from group 6 and 7 heifers (treated, n=10 and control, n=10) after in vitro stimulation 1 at d 30. 250 Treated Control 161 200 163 150 101 97 88 100 50 52 55 55 56 54 53 51 52 78 71 69 63 62 0 CELLS LPS PGN SEB SA 100 SA 200 SA 400 SA 800 PMA Stimulant 1 Cells = no stimulation, LPS = lipopolysaccharide, PGN = peptidoglycan, SEB = staphylococcal enterotoxin B, SA 100 = whole cell S. aureus antigen diluted to 1:100, SA 200 = whole cell S. aureus antigen diluted to 1:200, SA 400 = whole cell S. aureus antigen diluted to 1:400, SA 800 = whole cell S. aureus antigen diluted to 1:800, PMA = phorbol myristate acetate. 87

ROS production (AFU) Figure 37. Comparison 1 of reactive oxygen species (ROS) production (±SD) by neutrophils collected from groups 6 and 7 heifers (treated, n=10 and control, n=10) after in vitro stimulation 2 60 d after beginning trial. 300 Treated Control 250 200 177 149 150 100 50 47 44 48 45 47 45 47 45 81 87 75 67 63 54 51 47 0 CELLS LPS PGN SEB SA 100 SA 200 SA 400 SA 800 PMA Stimulant 1 Means within stimulants with different superscripts differ (P < 0.05). 2 Cells = no stimulation, LPS = lipopolysaccharide, PGN = peptidoglycan, SEB = staphylococcal enterotoxin B, SA 100 = whole cell S. aureus antigen diluted to 1:100, SA 200 = whole cell S. aureus antigen diluted to 1:200, SA 400 = whole cell S. aureus antigen diluted to 1:400, SA 800 = whole cell S. aureus antigen diluted to 1:800, PMA = phorbol myristate acetate. Prior to beginning the feed supplement (d 0), there were no significant differences in ROS production by neutrophils collected from group 8 treated and control heifers when exposed to the various stimulants (P > 0.05) (Figure 38). This indicates that there was no advantage for either treatment group prior to beginning the feed supplement. In fact, neutrophils from control heifers exhibited numerically greater ROS production compared to neutrophils from treated heifers. Thirty d after beginning the trial (day of vaccination), blood neutrophils from group 8 treated heifers stimulated with whole S. aureus antigen at 1:100 and PMA exhibited significantly (P < 0.05) higher in vitro ROS production (82 AFU and 142 AFU) compared to ROS production by neutrophils from control heifers (72 and 94 AFU) (Figure 39). Both Wiggins et al. (2011) and Heyneman et al. (1990) found increased ROS production after PMA stimulation. Sixty d 88

ROS production (AFU) after beginning trial (30 d after vaccination), there were no significant differences in ROS production by neutrophils among treated and control heifers, and no trends were observed (Figure 40). The significant differences between treated and control heifers observed at d 30, but not d 60 indicate that the heifers innate immune system may have adjusted to the feed supplement, resulting in no added effect beyond the 30-d period. The difference observed in results among groups 6, 7, and 8 may be due to month of sampling, age of heifers, and health status of animals, or maybe random variation. Figure 38. Comparison of reactive oxygen species (ROS) production (±SD) by neutrophils collected from group 8 heifers (treated, n=4 and control, n=5) after in vitro stimulation 1 at d 0. 120 Treated Control 85 100 80 74 60 40 45 41 42 42 43 43 42 43 54 51 51 50 45 47 44 45 20 0 CELLS LPS PGN SEB SA 100 SA 200 SA 400 SA 800 PMA Stimulant 1 Cells = no stimulation, LPS = lipopolysaccharide, PGN = peptidoglycan, SEB = staphylococcal enterotoxin B, SA 100 = whole cell S. aureus antigen diluted to 1:100, SA 200 = whole cell S. aureus antigen diluted to 1:200, SA 400 = whole cell S. aureus antigen diluted to 1:400, SA 800 = whole cell S. aureus antigen diluted to 1:800, PMA = phorbol myristate acetate. 89

ROS production (AFU) Figure 39. Comparison 1 of reactive oxygen species (ROS) production by neutrophils collected from group 8 heifers (treated, n=4 and control, n=5) after in vitro stimulation 2 30 d after beginning trial. 250 Treated Control 200 142 a 150 94 b 100 83 a 72 b 65 70 59 50 46 41 44 41 44 42 43 41 49 55 52 0 CELLS LPS PGN SEB SA 100 SA 200 SA 400 SA 800 PMA Stimulant 1 Means within stimulants with different superscripts differ (P < 0.05). 2 Cells = no stimulation, LPS = lipopolysaccharide, PGN = peptidoglycan, SEB = staphylococcal enterotoxin B, SA 100 = whole cell S. aureus antigen diluted to 1:100, SA 200 = whole cell S. aureus antigen diluted to 1:200, SA 400 = whole cell S. aureus antigen diluted to 1:400, SA 800 = whole cell S. aureus antigen diluted to 1:800, PMA = phorbol myristate acetate. 90

ROS production (AFU) Figure 40. Comparison of reactive oxygen species (ROS) production by neutrophils collected from group 8 heifers (treated, n=4 and control, n=5) after in vitro stimulation 1 at d 60. 180 Treated Control 160 140 120 100 80 60 40 20 0 21 18 17 17 31 28 35 21 70 64 1 Cells = no stimulation, LPS = lipopolysaccharide, PGN = peptidoglycan, SEB = staphylococcal enterotoxin B, SA 100 = whole cell S. aureus antigen diluted to 1:100, SA 200 = whole cell S. aureus antigen diluted to 1:200, SA 400 = whole cell S. aureus antigen diluted to 1:400, SA 800 = whole cell S. aureus antigen diluted to 1:800, PMA = phorbol myristate acetate. 46 51 30 32 15 11 110 109 CELLS LPS PGN SEB SA 100 SA 200 SA 400 SA 800 PMA Stimulant Effects of a general immunostimulant feed supplement on prevalence of mastitis in nulligravid and primigravid heifers Prevalence of IMI in heifers Groups 1 to 3 were used in the analysis of IMI prevalence among treated and control heifers due to the completeness of data and availability of secretion samples. Ability to collect secretions varied between age groups (due to the development of the mammary gland) and infection status (infected vs. uninfected). Over the course of the trial, one or more quarters of a mammary gland was diagnosed with an IMI in 86% of heifers (groups 1 to 3). Similar infection rates have been identified in past studies. For example, Oliver et al. (1992) found a 90% prevalence of IMI among heifers prior to calving. Trinidad et al. (1990b) reported that 96.9% of heifers exhibited an IMI. Other studies 91

have reported infections rates as high as 100% (Boddie et al. 1987; Trinidad et al., 1990c; Nickerson et al., 1995), whereas some studies such as Owens et al. (1994) identified a prevalence of only 50.7% among heifers. In terms of treatment group in the present trial, treated heifers exhibited a higher infection rate (92.9%) compared to control heifers (80%), although this difference was not significant (Table 8). Among the 3 groups that were analyzed, there were no significant differences in prevalence of infection (80-90%) (P = 1.00) (Table 8). The overall prevalence of IMI caused by S. aureus across treatments was 62.1%. While treated heifers had a higher prevalence (71.4%) compared to control heifers (53.3%), the difference was not significant (Table 9). Among the 3 groups, there was a difference in the prevalence of mastitis caused by S. aureus. Group 1 had the highest prevalence of mastitis caused by S. aureus at 88.9% compared to groups 2 (30%) and 3 (70%). Trinidad et al. (1990b) found a lower prevalence among heifers of S. aureus IMI (37.1%) compared to the overall 62.1% prevalence in the present trial. Fox et al. (1995) found a greater prevalence of S. aureus IMI in the south (Louisiana) compared to other regions of the country (California, Vermont, and Washington) at various times of the year. For example, in the Spring and Fall, prevalence of S. aureus IMI among quarters in the south was 10.1% and 7.9%, whereas the other areas of the country did not exhibit prevalences greater than 3.0% during either of these seasons. The differences observed in geographical regions is most likely caused by the warm, humid climate, which is favorable to fly larvae growth and survival as well as an extended horn fly season, allowing for prolonged exposure of udders and teats to enhanced bacterial loads introduced by the horn flies. The increased prevalence S. aureus IMI in the present trial (62.1%) compared to 92

Trinidad et al. (1990b) (37.1%) may be due to difference in regional location, weather, fly populations, breed of heifers (Holstein vs. Jersey), and age of heifers at time of sampling. The high prevalence of S. aureus IMI in the present trial is of concern. S. aureus IMI prior to and during pregnancy may lead to 10% less milk production during the first few months of lactation if not treated with nonlactating dry cow product (Trinidad et al., 1990c). Owens et al. (1991) found that S. aureus-infected quarters produced 11% less milk compared to quarters that were cured with antibiotic therapy. Nickerson et al. (1999) studied vaccinated (against S. aureus) and unvaccinated heifers throughout the first lactation and found that vaccinated heifers produced 8.6% more milk during the first lactation. Although it was hypothesized that vaccination against S. aureus in conjunction with administration of the feed supplement would lead to a lower prevalence of S. aureus in supplement-treated heifers compared to controls, such a beneficial effect was not observed. The role that horn flies play in the transmittance and infection of heifer mammary glands by S. aureus may be greater than previously noted. Table 8. Univariate comparisons of any case of IMI diagnosed in mammary secretions from treated and control heifers. Variable Sample size Number with any IMI P-value (%) Treatment Treated Control 14 15 13 (92.9) 12 (80.0) 0.598 Group 1 2 3 9 10 10 8 (88.9) 9 (90.0) 8 (80.0) 1.00 93

Table 9. Univariate comparisons of Staphylococcus aureus IMI diagnosed in mammary secretions from treated and control heifers. Variable Sample size Number with S. P-value aureus IMI (%) Treatment Treated Control 14 15 10 (71.4) 8 (53.3) 0.450 Group 1 2 3 9 10 10 8 (88.9) 3 (30.0) 7 (70.0) 0.029 Overall prevalence of CNS among heifers, regardless of treatment group was 62.1% (Table 10). Treated animals exhibited a greater prevalence of mastitis caused by CNS (71.4%) compared to control heifers (53.3%); however, this difference was not significant (P = 0.450). In addition, there was no difference in prevalence of mastitis caused by CNS among the 3 groups (P = 0.249). Fox et al. (1995) observed prevalences of CNS IMI that ranged from 23% to 35.6% among heifers, which was lower than the prevalence of 62.1% that was identified in the present study. While other studies did not report CNS IMI among heifers, CNS IMI was the most prevalent of all bacterial species among quarters, indicating that aside from S. aureus, CNS pathogens are important in the analysis of mastitis-causing pathogens. For example, Trinidad et al. (1990b) found that 67.4% of bacterial isolates identified were CNS species. Oliver et al. (1992) found a similar trend, with CNS IMI observed as the most prevalent among quarters (44.3%). Nickerson et al. (1992) and Oliver et al. (1992) reported CNS IMI prevalences among quarters of 27.9% and 39%, respectively. 94

Table 10. Univariate comparisons of coagulase-negative staphylococci (CNS) IMI diagnosed in mammary secretions from treated and control heifers. Variable Sample size Number with CNS P-value IMI (%) Treatment Treated Control 14 15 10 (71.4) 8 53.3) 0.450 Group 1 2 3 9 10 10 7 (77.8) 7 (70.0) 4 (40.0) 0.249 S. chromogenes, S. hyicus, S. simulans, and Streptococcus spp. were identified as being the most prevalent mastitis-causing pathogens (after S. aureus) isolated from mammary secretions. Overall, prevalence of IMI caused by S. chromogenes across treatments was 24%. Treated heifers exhibited a higher prevalence (35.7%) compared to control heifers (13.3%), but there was no significant difference (P = 0.385) (Table 11). There was, however, a significant difference in the prevalence of S. chromogenes mastitis among the 3 groups (P = 0.025). Group 2 exhibited the greatest prevalence (50.0%) compared to both groups 1 and 3 (22.2% and 0.00%, respectively). It appears that colonization of S. aureus in the mammary gland prevents colonization of other pathogens. Groups 1 (88.9%) and 3 (70%) exhibited the greatest prevalence of S. aureus in contrast to group 2 (30%). The reverse can be said for S. chromogenes, a CNS species. For example, Group 2 exhibited the greatest prevalence of S. chromogenes IMI (50%) and the lowest prevalence of S. aureus IMI, whereas groups 1 (22.2%) and 3 (0%) demonstrated the lowest prevalences of S. chromogenes IMI. This relationship was identified in a study conducted by Nickerson and Boddie (1994), which found that uninfected quarters prior to challenge with S. aureus had a prevalence 3 times greater than quarters infected with CNS. The results from the 95

past and current studies indicate that quarters infected with Staphylococcus spp., other than S. aureus, may be somewhat protected against development of S. aureus IMI. Overall prevalence of mastitis caused by S. hyicus was 45%, and was not different between treated (42.9%) and control heifers (46.7%) (Table 12). Differences among groups were significant (P = 0.050), with group 1 having the greatest prevalence of 77.8% and groups 2 and 3 having the lowest (40% and 20%, respectively). Analysis of IMI caused by S. simulans showed that the overall prevalence was 21%. While treated heifers had a greater prevalence (28.6%) than control heifers (13.3%), there was no significant difference (P = 0.390) (Table 13). Additionally, there were no differences among groups 1, 2, or 3 (22.2%, 10%, and 30%, respectively) (P = 0.642). The prevalence of mastitis caused by Streptococcus spp. was 14%. There was no significant difference in prevalence between treatment or age groups (treatment, P = 0.598 and age, P = 0.094) (Table 14). Prevalence of S. chromogenes, S. hyicus, S. simulans, and Streptococcus spp. IMI among heifers is not widely reported; however, prevalence among quarters in one study was observed to be 43.1%, 24.3%, 0.7%, and 3.6%, respectively (Trinidad et al., 1990b). Nickerson et al. (1995) also found that quarter samples from heifers revealed a 35%, 17%, and 9% prevalence of S. chromogenes, S. hyicus, and Streptococcus spp. IMI. Comparisons between the present and previous trials will be made in later sections when discussing IMI and TCI prevalence among quarters. 96

Table 11. Univariate comparisons of any case of Staphylococcus chromogenes IMI diagnosed in mammary secretions from treated and control heifers. Variable Sample size Number with S. P-value chromogenes IMI (%) Treatment Treated Control 14 15 5 (35.7) 2 (13.3) 0.385 Group 1 2 3 9 10 10 2 (22.2) 5 (50.0) 0 (0.00) 0.025 Table 12. Univariate comparisons of any case of Staphylococcus hyicus IMI diagnosed in mammary secretions from treated and control heifers. Variable Sample size Number with S. hyicus P-value IMI (%) Treatment Treated Control 14 15 6 (42.9) 7 (46.7) 1.00 Group 1 2 3 9 10 10 7 (77.8) 4 (40.0) 2 (20.0) 0.050 Table 13. Univariate comparisons of Staphylococcus simulans IMI diagnosed in mammary secretions from treated and control heifers. Variable Sample size Number with S. P-value simulans IMI (%) Treatment Treated Control 14 15 4 (28.6) 2 (13.3) 0.390 Group 1 2 3 9 10 10 2 (22.2) 1 (10.0) 3 (30.0) 0.642 97

Table 14. Univariate comparisons of Streptococcus spp. IMI diagnosed in mammary secretions from treated and control heifers. Variable Sample size Number with Strep. P-value IMI (%) Treatment Treated Control 14 15 1 (7.14) 3 (20.0) 0.598 Group 1 2 3 9 10 10 3 (33.3) 1 (10.0) 0 (0.00) 0.094 Prevalence of TCI in heifers Groups 1 to 4 were used in the analysis of teat canal infections because in all age groups, teat canals were of sufficient diameter to accommodate collection of mastitis-causing bacteria, which were colonizing keratin. A total of 83% of heifers (across treatments) experienced a teat canal infection (TCI) caused by any bacterial species during the course of the study, with control animals exhibiting a greater prevalence (87%) compared to treated heifers (77.8%); however, there was no significant difference (P = 0.679) (Table 15). Overall prevalence of TCI was similar to overall prevalence of IMI in heifers. This indicates a direct relationship between infection in the teat canal and infection in the mammary gland. This relationship was also recognized by Boddie et al. (1987). In the present trial, there was a significant difference (P = 0.028) in prevalence of TCI among age groups. Groups 2 exhibited the greatest prevalence of TCI (100%), whereas group 4 exhibited the least (50.0%). Group 4 was the youngest group, so these data support the idea that as heifers mature, there is a greater chance that they will come into contact with bacterial pathogens that will subsequently develop into a TCI. Trinidad et al. (1990b) observed a similar overall TCI prevalence of 93.1% among heifers as the overall prevalence found in the present trial (83%). 98

Overall prevalence of TCI among heifers caused by S. aureus was 61%; however, there was no difference between treatments (61.1% vs. 60.9%) (P = 1.00) (Table 16). This percentage was similar to the prevalence of IMI found among heifers discussed previously (62.1%) (Table 8). There was a significant difference (P = 0.008) among age groups. Group 1 exhibited the greatest prevalence (88.9%) followed by group 3 (83.3%). Prevalences in groups 2 and 4 were 41.7% and 25%, respectively. Differences in group prevalences may be due to population of flies within pastures, as similar infection rates were seen in reference to S. aureus IMI among heifers in groups 1 (88.9%) and 3 (70%). Trinidad et al. (1990b) reported a prevalence of S. aureus TCI among heifers of 31%, which was lower than the prevalence of S. aureus TCI (61%) exhibited by heifers in this study. Factors attributing to these differences are similar to the reasons offered in the previous section, and may be age of heifers at time of sampling, fly populations, regional location, weather, and breed of heifer. Table 15. Univariate comparisons of any case of TCI diagnosed in treated and control heifers. Variable Sample size Number with any TCI P-value (%) Treatment Treated Control 18 23 14 (77.8) 20 (87.0) 0.679 Group 1 2 3 4 9 12 12 8 8 (88.9) 12 (100.0) 10 (83.3) 4 (50.0) 0.028 99

Table 16. Univariate comparisons of Staphylococcus aureus TCI diagnosed in treated and control heifers. Variable Sample size Number with S. P-value aureus TCI(%) Treatment Treated Control 18 23 11 (61.1) 14 (60.9) 1.00 Group 1 2 3 4 9 12 12 8 8 (88.9) 5 (41.7) 10 (83.3) 2 (25.0) 0.008 Overall prevalence of TCI caused by CNS was 59% of heifers. Prevalence of TCI in treated (55.6%) vs. control (60.9%) heifers was similar (P = 0.760) as shown in Table 17. Furthermore, there was no difference among age groups (P = 0.129). Overall prevalence of TCI in treated and control heifers caused by S. chromogenes was 34% of heifers. Prevalence was similar between treatments (33.3% vs. 34.8%) (P = 1.00) (Table 18), and there was no difference among age groups (group 1: 44.4%, group 2: 50%, group 3: 8.33%, and group 4: 37.5%) (P = 0.130). Overall prevalence of TCI caused by S. hyicus was 39% of heifers. Similar to other tests, there was no difference in prevalence between treatments (33.3% vs. 43.5%) (P = 0.540) (Table 19). There was, however, a significant difference among age groups (P = 0.012). Group 1 had the greatest prevalence (77.8%), compared to group 3 having the lowest prevalence (8.33%). Groups 2 and 4 had prevalences of 41.7% and 37.5%, respectively. Analysis of TCI caused by S. simulans and Streptococcus spp. yielded similar results (Tables 20 and 21). Overall prevalences of TCI caused by S. simulans and Streptococcus spp. were approximately 29% and 7%, respectively. In regard to TCI caused by S. simulans, control heifers had a slightly higher prevalence at 30.4% compared to treated heifers at 27.8%, but there was no significant difference (P = 1.00). There was also no difference among age groups (group 1: 33.3%, group 2: 16.7%, 100

group 3: 50%, and group 4: 12.5%) (P = 0.253). For TCI caused Streptococcus spp., there were no differences between treatment groups (0% vs. 13%) (P=1.00) or among age groups (group 1: 22.2%, group 2: 8.33%, group 3: 0%, and group 4: 0%) (P = 0.249) While a few studies report TCI among quarters, rather than among heifers, comparisons may be made in trends based on the predominant species identified. Trinidad et al. (1990b) observed predominant bacterial isolates among quarters, such as S. chromogenes (42.9%), S. hyicus (25.2), S. simulans (1.25%), and Streptococcus spp. (3.7%), that were similar to the predominant species isolated from teat canal keratin samples in the present study. Table 17. Univariate comparisons of coagulase-negative staphylococci (CNS) TCI diagnosed treated and control heifers. Variable Sample Size Number with CNS P-value (%) Treatment Treated Control 18 23 10 (55.6) 14 (60.9) 0.760 Group 1 2 3 4 9 12 12 8 7 (77.8) 9 (75.0) 5 (41.7) 3 (37.5) 0.129 Table 18. Univariate comparisons of Staphylococcus chromogenes TCI diagnosed in treated and control heifers. Variable Sample size Number with S. P-value chromogenes TCI (%) Treatment Treated Control 18 23 6 (33.3) 8 (34.8) 1.00 Group 1 2 3 4 9 12 12 8 4 (44.4) 6 (50.0) 1 (8.33) 3 (37.5) 0.130 101

Table 19. Univariate comparisons of Staphylococcus hyicus TCI diagnosed in treated and control heifers. Variable Sample size Number with S. hyicus P-value TCI (%) Treatment Treated Control 18 23 6 (33.3) 10 (43.5) 0.540 Group 1 2 3 4 9 12 12 8 7 (77.8) 5 (41.7) 1 (8.33) 3 (37.5) 0.012 Table 20. Univariate comparisons of Staphylococcus simulans TCI diagnosed in treated and control heifers. Variable Sample Size Number with S. P-value simulans (%) Treatment Treated Control 18 23 5 (27.8) 7 (30.4) 1.00 Group 1 2 3 4 9 12 12 8 3 (33.3) 2 (16.7) 6 (50.0) 1 (12.5) 0.253 Table 21. Univariate comparisons of Streptococcus spp. TCI diagnosed in treated and control heifers. Variable Sample Size Number with Strep. P-value (%) Treatment Treated Control 18 23 0 (0.00) 3 (13.0) 1.00 Group 1 2 3 4 9 12 12 8 2 (22.2) 1 (8.33) 0 (0.00) 0 (0.00) 0.249 102

Prevalence of IMI by quarter Overall prevalence of IMI among quarters sampled from August 2010 through October 2011 was 58% (427 infected quarters out of 736 quarters sampled) (Figure 41). Conversely, Boddie et al. (1987) found an overall prevalence of IMI of 86.1%. Likewise, Trinidad et al. (1990b) observed an overall prevalence of IMI of 74.6%. The heifers in the previously mentioned trials were Jersey heifers whereas the heifers in the present study are Holstein heifers, which tend to exhibit a lower prevalence of IMI (Nickerson et al., 1995). Also, heifers in the present study were enrolled at a younger age than those from the studies by Boddie et al. (1987) and Trinidad et al. (1990b). Younger heifers are inherently less likely to be infected than older animals, which have had more time to be exposed to various pathogens and stressors. Similar to this trial, Owens et al. (2001) found an overall IMI prevalence of 56.5% among quarters, and he sampled heifers around the same age as that of the heifers from the present trial. In the present trial, S. aureus was the most prevalent pathogen among quarters with a prevalence of 25.3% (186 S. aureus-infected quarters out of 736 quarters sampled) (Figure 41). Contrary to the results found in the present trial, Boddie et al. (1987) found that only 13% of quarters were infected with S. aureus. Likewise, Trinidad et al. (1990b) observed 19.9% of quarters infected with S. aureus, and Owens et al. (2001) observed 15.4% of quarters infected with S. aureus. The higher prevalence of quarter infections with S. aureus IMI identified in the present trial may be due to seasonal differences and management techniques, e.g., fly control to reduce horn fly populations. In the present trial, S. hyicus was identified in 11.1% of quarters (81 S. hyicus-infected quarters out of 736 quarters sampled), S. chromogenes in 8.7% of quarters (64 S. chromogenesinfected quarters out of 736 quarters sampled), S. simulans in 4.5% of quarters (33 S. simulans- 103

infected quarters out of 736 quarters sampled), and S. intermedius in 1.6% of quarters (12 S. intermedius-infected quarters out of 736 quarters sampled) (Figure 41). Boddie et al. (1987) found greater prevalences of S. chromogenes IMI (49.5%) and S. hyicus (21.3%) IMI among quarters compared to the present trial. Trinidad et al. (1990b) also found increased prevalences of S. chromogenes IMI (43.1%) and S. hyicus IMI (24.3%) among quarters compared to the present trial. Given that Boddie et al. (1987) and Trinidad et al. (1990b) identified greater prevalences of IMI (86.1% and 76.4%, respectively) than the present trial (58%) and lower prevalences of S. aureus (13% and 19.9%, respectively) than the present trial (25.3%), it follows that the prevalence of CNS species would be greater in the two previous trials in comparison to the present trial. Findings from past studies indicate that quarters infected with Staphylococcus spp., other than S. aureus, (e.g., CNS) may be somewhat protected against development of S. aureus IMI. Since animals in the past studies (Boddie et al., 1987 and Trinidad et al., 1990b) had a greater prevalence of CNS, they may have been protected against S. aureus IMI. However, in the present trial, the effect that fly populations have on pathogen transmission appeared to play a greater role in S. aureus IMI development rather than in development of CNS infections. The CNS (overall quarter prevalence of 22.7%, 167 out of 736 quarters sampled) identified included S. chromogenes, S. hyicus, S. simulans, S. xylosus, S. warneri, S. auricularis, and S. caprae. The CPS (overall prevalence of 5.7%, 42 out of 736 quarters sampled) identified included S. intermedius, S. hyicus, S. sciuri, and S. simulans. The streptococci (overall prevalence of 3.1%, 23 out of 736 quarters sampled) identified include Strep. uberis and Strep. dysgalactiae. Various other microorganisms (3.7%, 27 out of 736 quarters sampled) were identified including gram-positive cocci, Micrococcus spp., Yeast, and Serratia spp. 104

Across treatments, 42% of quarters (309 quarters out of 736) were uninfected. Out of 390 quarters of treated animals sampled, 39.0% (152 quarters) were uninfected compared to 45.4% of quarters sampled (157 out of 346) from control animals (Figure 42). Across treatments, 58% of quarters (427 out of 736) were infected. Among quarters of treated animals, 61% of quarters sampled were infected (238 out of 390) compared to 54.6% of quarters sampled (189 out of 346) from control animals. These data indicate a numerically higher prevalence of infection among quarters from treated heifers. Quarters from treated animals exhibited a higher prevalence of S. aureus IMI (28.2%, 110 out of 390) compared to quarters from control heifers (22%, 76 out of 346) (Figure 43). A similar trend was observed in quarter infections caused by CNS (23.6%, 92 out of 390 vs. 21.7%, 75 out of 346) and CPS (5.4%, 21 out of 390 vs. 6.1%, 21 out of 346). Quarters of controls generally had a greater prevalence of infection caused by streptococci (4.9%, 17 out of 346 vs. 1.5%, 6 out of 390) than quarters of treated animals. Figure 41. Overall prevalence (%) of IMI by infection status and individual bacterial isolate. 11.1 8.7 4.5 1.6 25.3 3.7 3.1 *Other includes unidentified CNS, S. auricularis, Serratia spp., S. xylosus, S. sciuri, yeast, grampositive cocci, Micrococcus spp., S. warneri, and S. caprae. **Streptococci include Strep. uberis and Strep. dysgalactiae. Overall prevalence of infection = 58% 42 Uninfected S. aureus S. hyicus S. chromogenes S. simulans S. intermedius Other microoorganisms* Streptococci** 105

% of quarters Percentage of quarters (%) Figure 42. Distribution of infection status by treatment group among quarters. 60 Treated Control 61 54.6 50 45.4 40 39 30 20 10 0 Uninfected Infected Figure 43. Distribution of IMI in treated and control heifers by bacterial isolate. 35 Treated Control 30 28.2 25 20 23.6 22 21.7 15 10 5 0 5.4 6.1 4.9 1.5 2.3 S. aureus CNS CPS Streptococci Other* Infection status *Species include gram-positive cocci, Micrococcus spp., yeast, and Serratia spp. Multivariable logistic regression analyzing the effects of treatment, group, and quarter location on prevalence of mastitis on a quarter basis is shown in Table 22. The odds of IMI caused by any pathogen were twice as high (OR = 2.0) for treated heifers compared to control 106

heifers; however, this difference was not significant (P = 0.278). Odds of infection among quarters from heifers in groups 2 and 3 were significantly different than quarters in group 1 (referent). The odds of IMI in quarters from group 2 heifers were 95% lower (OR = 0.05, P = 0.002), than quarters from group 1 heifers. The odds of IMI in quarters from group 3 heifers were 92% lower (OR = 0.08, P = 0.020) than group 1 heifers. Interestingly, the odds of IMI in front quarters were 5.1 times higher (OR = 5.1) than the rear quarters (P = 0.002). The cause of this difference is attributed to the role that the horn flies play in transmission of infection, specifically S. aureus infection. This relationship is discussed in more detail later in this section. Multivariable logistic regression for the prediction of S. aureus IMI on a quarter basis revealed that the odds of S. aureus IMI was 2.1 times higher (OR = 2.1) than controls; however, this difference was not significant (P = 0.190) (Table 23). The odds of S. aureus IMI in quarters from group 2 heifers were 89% lower (OR = 0.11) than quarters from group 1 heifers (P = 0.004). Nickerson and Boddie (1994) found that quarters infected with Staphylococcus spp. (e.g., CNS) other than S. aureus, may be shielded against development of S. aureus IMI. The odds of S. aureus IMI in quarters from group 3 heifers (OR = 0.55) was not different from quarters of group 1 heifers (P = 0.337). Similar to overall prevalence, the odds of S. aureus IMI diagnosed in front quarters were 3.9 times higher (OR = 3.9) than rear quarters (P = 0.005). The cause of this difference is attributed to the role that the horn flies play in transmission S. aureus IMI. Horn flies are attracted to the navel area of the heifers, which is in close proximity to the front teats. Also, the tail switch is effective at repelling flies from biting the rear teats. This relationship is discussed in greater detail later in this section. Table 24 displays the multivariable logistic results for the analysis of CNS mastitis by quarter. No difference was found between quarters for treated and control heifers (OR = 1.8, P = 107

0.204). The odds of CNS IMI in quarters from group 2 heifers were 77% lower (OR = 0.33, P = 0.029) compared to quarters from group 1 heifers. Likewise, the odds of CNS IMI in quarters from group 3 animals were 81% lower than quarters from group 1 heifers. Nickerson and Boddie (1994) found that prevalence of new S. aureus IMI was 3 times greater in quarters that were previously uninfected compared to quarters infected with CNS, illustrating the potential protective nature of CNS infections against S. aureus IMI. Table 22. Multivariable logistic regression results for the prediction of IMI based on quarter mammary secretions from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control 0.690 (0.626) Referent 2.0 (0.58, 6.7) 0.278 Group 1 2 3 Referent -3.06 (1.01) -2.48 (1.06) 0.05 (0.01, 0.34) 0.08 (0.01, 0.68) 0.002 0.020 Quarter location Rear Front Referent 1.62 (0.515) 5.1 (1.8, 14) 0.002 Constant 1.62 (0.778) NA 0.037 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n = 12) treated (n = 6) and control (n = 6); and group 3 (n = 12) treated (n = 5) and control (n =7). 108

Table 23. Multivariable logistic regression results for the prediction of Staphylococcus aureus IMI based on quarter mammary secretions from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control 0.736 (0.561) Referent 2.1 (0.70, 6.3) 0.190 Group 1 2 3 Referent -2.18 (0.753) -0.596 (0.621) 0.11 (0.03, 0.49) 0.55 (0.16, 1.9) 0.004 0.337 Quarter location Rear Front Referent 1.36(0.489) 3.9 (1.5, 10) 0.005 Constant -1.17 (0.582) NA 0.044 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n = 12) treated (n = 6) and control (n = 6); and group 3 (n = 12) treated (n = 5) and control (n =7). Table 24. Multivariable logistic regression results for the prediction of coagulase-negative staphylococci (CNS) IMI based on quarter mammary secretions from treated and control heifers 1. Variable Treatment Treated Control Coefficient (Semirobust SE) 0.590 (0.464) Referent OR (95% CI) P-value 1.8 (0.73, 4.5) 0.204 Group 1 2 3 Referent -1.10 (0.505) -1.67 (0.616) 0.33 (0.12, 0.89) 0.19 (0.06, 0.63) 0.029 0.007 Quarter location Rear Front Referent -3.11e -15 (0.540) 1.0 (0.35, 2.9) 1.00 Constant -0.605 (0.599) NA 0.312 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n = 12) treated (n = 6) and control (n = 6); and group 3 (n = 12) treated (n = 5) and control (n =7). 109

The multivariable model for the prediction of an IMI caused by S. chromogenes would not converge because no quarters in group 3 were infected with S. chromogenes over the course of the trial. Additionally, the multivariable model for IMI caused by streptococci would not converge because there were no quarters in group 3 infected with streptococci during the course of the trial. Multivariable logistic regression analyzing IMI caused by S. hyicus yielded no difference in the odds of infection between treatments (OR = 1.0, P = 0.928) (Table 25). The odds of S. hyicus IMI in quarters from groups 2 and 3 were 72% (OR = 0.28) and 89% (OR = 0.11), respectively, lower (P = 0.041 and 0.003, respectively) than group 1. Group 1 was the oldest group, so those heifers had more time to be exposed to various pathogens. There was no difference in the odds between front and rear quarters (OR = 1.2, P = 0.812). For the prediction of IMI caused by S. simulans, there was no difference in the odds between quarters from treated and control heifers (OR = 2.9, P = 0.252), quarters of groups 2 and 3 (OR = 0.42, P = 0.499 and OR = 1.9, P = 0.478, respectively) compared to group 1, or front vs. rear quarters (OR = 0.14, P = 0.122) when multivariable logistic regression was performed (Table 26). 110

Table 25. Multivariable logistic regression results for the prediction of Staphylococcus hyicus IMI based on quarter mammary secretions from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control 0.048 (0.534) Referent 1.0 (0.37, 3.0) 0.928 Group 1 2 3 Referent -1.26 (0.614) -2.25 (0.757) 0.28 (0.09, 0.95) 0.11 (0.02, 0.46) 0.041 0.003 Quarter location Rear Front Referent 0.140 (0.589) 1.2 (0.36, 3.6) 0.812 Constant -0.785 (0.559) NA 0.160 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n = 12) treated (n = 6) and control (n = 6); and group 3 (n = 12) treated (n = 5) and control (n =7). Table 26. Multivariable logistic regression results for the prediction of Staphylococcus simulans IMI based on quarter mammary secretions from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control 1.08 (0.944) Referent 2.9 (0.46, 19) 0.252 Group 1 2 3 Referent -0.879 (1.30) 0.622 (0.877) 0.42 (0.03, 5.3) 1.9 (0.33, 10) 0.499 0.478 Quarter location Rear Front Referent -1.94 (1.26) 0.14 (0.01, 1.7) 0.122 Constant -2.85 (0.701) NA 0.000 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n = 12) treated (n = 6) and control (n = 6); and group 3 (n = 12) treated (n = 5) and control (n =7). 111

Influence of flies on teat health The differences exhibited between front and rear quarters for overall IMI and S. aureus IMI were most likely influenced by horn flies, and teat scores were subsequently analyzed to observe any effects of fly populations on teat end health. Horn flies were observed to congregate around the navel area, migrating back to the front teats of the mammary gland. Rear While rear teats were affected by the horn flies as well, front teats consistently exhibited greater numbers (Figure 44). Regardless of treatment group, results demonstrated Front decreasing teat health (based on scores) at the time that horn fly populations drastically Figure 44. Presence of horn flies on teats. increased during the spring and summer months. For example, in August of 2010, teat health score collection began on group 1, and at this time, teat scores were initially recorded and horn fly populations visually noted (but not quantified). Up until this time, fly control treatment had not been applied. A score of 1 corresponded to healthy skin with no scabs on the teat; a score of 2 corresponded to older scabs that were darker and red-brown in color; and a score of 3 corresponded to fresh, new scabs caused by blood-sucking horn flies that were bloody and red in color. In August 2010, front quarters of group 1 heifers were observed to have more scabs, new or old, compared to rear quarters (Table 27). For example, right front teats and left front teats exhibited higher mean teat scores (2.6 and 2.4) compared to mean scores of left rear teats and right rear teats (1.7 and 1.6). Similar trends were found in groups 2, 3, and 4 (Tables 28, 29, and 112

30). Mean right front and left front teat scores for group 2 prior to fly treatment were 2.4 and 2.5, whereas the left rear and right rear teats had mean of 1.5 and 1.4 (Table 28). Mean right front (2.5) and left front (2.5) scores for group 3 heifers were numerically greater than both the mean left rear and right rear (both scores of 1) (Table 29). Mean right front (2.3) and left front (2.3) scores for group 4 heifers were numerically greater than both the mean left rear and right rear (both scores of 1.3) (Table 30). Table 27. Distribution of teat score values among quarters in group 1 from August 2010. Heifer ID Right front Left front Left rear Right rear 9026 2 2 2 2 9027 3 3 1 2 9028 3 3 1 1 9029 2 2 1 1 9030 3 3 2 2 9031 2 2 1 1 9033 3 2 2 1 9034 3 3 3 3 9035 2 2 2 1 Mean score 2.6 2.4 1.7 1.6 Table 28. Distribution of teat score values among quarters in group 2 from May 2011. Heifer ID Right front Left front Left rear Right rear 9036 2 2 2 2 9037 3 3 1 1 9038 3 3 3 3 9039 3 3 3 3 9040 2 2 1 1 9041 1 2 2 1 9042 2 2 1 1 9043 3 3 1 1 9044 2 2 1 1 9045 2 2 1 1 9046 3 3 1 1 9047 3 3 1 1 Mean score 2.4 2.5 1.5 1.4 113

Table 29. Distribution of teat score values among quarters in group 3 from May 2011. Heifer ID RF LF LR RR 1 2 2 1 1 2-1 - - - 3 3 3 1 1 4 3 3 1 1 5 2 2 1 1 6 3 3 1 1 9 3 3 1 1 10 3 3 1 1 11 2 2 1 1 13 - - - - 14 3 3 1 1 17 1 1 1 1 Mean score 2.5 2.5 1 1 1 No score taken; heifer would not lock up. Table 30. Distribution of teat score values among quarters in group 4 from September 2011. Heifer ID RF LF LR RR 15 2 2 2 2 18 2 2 1 1 19 2 2 1 1 20 3 3 1 1 21 2 2 1 1 22-1 - - - 25 2 2 2 2 26 2 2 1 1 27 3 3 1 1 Mean score 2.3 2.3 1.3 1.3 1 No score taken; heifer would not lock up. Once teat scores were assigned, pour-on fly treatment (Ultra Boss, Merck Animal Health, Summit, NJ) was administered to all heifers every 14-21 d to minimize the fly population and allow teats to heal and remain free of scabs. Within 2-4 wk of fly treatment, horn fly populations had dropped drastically and teats had become healthier, exhibiting lower scores. For example, in September of 2010 (1 mo after administering fly control), average scores for all quarters from group 1 animals were lower (RF 1.8, LF 1.7, LR 1, and RR 1) (Table 31) than the scores assigned in the previous month (RF 2.6, LF 2.4, LR 1.7, and RR 1.6) (compare tables 27 114

and 31). Mean teat scores for groups 2 to 4 in the month after fly treatment are displayed in Table 32. Group 3 left rear and right rear mean teat scores from 1 mo after treatment (1.6 and 1.7) (Table 32) are slightly higher than the previous month s mean teat scores prior to fly treatment (1 and 1) (Table 29) because heifers were due to be treated the following day with fly treatment. Mean scores for the right front and left front are, however, higher prior to fly treatment (2.5 and 2.5) (Table 29) compared to 1 mo after fly treatment (1.8 and 1.9) (Table 32). This reduction in teat scores illustrates the importance of fly control in replacement heifer management as described by Nickerson et al. (1995). Table 31. Distribution of teat score values among quarters in group 1 from September 2010*. Heifer ID Right front Left front Left rear Right rear 9026 1 1 1 1 9027 3 3 1 1 9028 3 3 1 1 9029 2 1 1 1 9030 1 1 1 1 9031 1 1 1 1 9033 1 2 1 1 9034 3 2 1 1 9035 1 1 1 1 Mean score 1.8 1.7 1 1 *Compare to pre-fly treatment table 27. Table 32. Distribution of mean teat score values among quarters in groups 2, 3, and 4 in the month succeeding fly treatment*. Group Right front Left front Left rear Right rear 2 1.4 1.5 1.2 1 3 1.8 1.9 1.6 1.7 4 1.1 1.3 1 1 *Compare to pre-fly treatment Tables 28, 29, and 30. Investigation of overall teat score values among quarters of heifers in the 4 different groups showed that front quarters exhibited more severe teat scores than rear quarters, regardless of treatment group (Table 33). Teat score data, in conjunction with the previous data showing 115

that front quarters are more likely be diagnosed with an IMI, more specifically S. aureus IMI, support the documented findings of other researchers (Nickerson et al. 1995; Owens et al. 1998) that flies play a role in the transmission and establishment of S. aureus IMI. Although there were numerical differences in scores between quarters from treated (RF 2.6, LF 2.6, LR 1.4, and RR 1.4) and control (RF 2.3, LF 2.3, LR 1.3, and RR 1.3) animals, overall standard deviations were quite high and suggest no significant differences. To account for confounding factors, the 4 groups were examined individually (Table 33), and while sample sizes were small, results confirm the association between front quarters and severity of scabs compared to rear quarters, regardless of treatment group. Treated animals had mean scores that were generally higher than control animals. For example, in group 1 treated heifers RF and LF mean teat scores were 2.8 for both; however, RF and LF mean teat scores from group 1 control heifers were 2.4 and 2.2, respectively. Mean teat scores for group 1 treated animals LR and RR teats were 2 for both, whereas LR and RR mean teat scores for group 1 control animals were 1.4 and 1.2, respectively. 116

Table 33. Comparison of teat score values among quarters of treated and control heifers overall and in 4 different heifer groups. RF LF LR RR Variable n Mean SD Mean SD Mean SD Mean SD Overall a Treated Control 17 22 2.6 2.3 0.51 0.65 2.6 2.3 0.51 0.57 1.4 1.3 0.71 0.57 1.4 1.2 0.71 0.53 Group 1 b Treated Control 4 5 2.8 2.4 0.5 0.49 2.8 2.2 0.5 0.4 2 1.4 0.82 0.49 2 1.2 0.82 0.4 Group 2 c Treated Control 6 6 2.3 2.5 0.52 0.84 2.3 2.7 0.52 0.52 1.5 1.5 0.84 0.84 1.5 1.3 0.84 0.82 Group 3 d Treated Control 4 6 3 2.2 0 0.75 3 2.2 0 0.75 1 1 0 0 1 1 0 0 Group 4 e Treated Control 3 5 2.3 2.2 0.58 0.45 2.3 2.2 0.58 0.45 1 1.4 0 0.55 1 1.4 0 0.55 a Overall mean teat scores from 4 groups during August 2010, May 2011, and September 2011. b Mean teat scores from August 2010. c Mean teat scores from May 2011. d Mean teat scores from May 2011. e Mean teat scores from September 2011. Prevalence of TCI by quarter Overall prevalence of TCI was 47.6% of quarters (790 infected teat canals out of 1,507 teat canals sampled) (Figure 45). Boddie et al. (1987) and Trinidad et al. (1990b) found greater prevalences of TCI (70.1% and 70.7% of quarters, respectively) compared to the present trial. Heifers in the present trial were initially sampled at a younger age than the heifers in the trials by Boddie et al. (1987) and Trinidad et al. (1990b). Younger heifers are inherently less likely to be infected than older animals that have had more time to be exposed to various pathogens and stressors. 117

In the present trial, S. aureus was the most prevalent pathogen with a prevalence of 20.3% among quarters (306 S. aureus-infected teat canals out of 1,507 teat canals sampled) (Figure 45). Boddie et al. (1987) found a much lower prevalence of S. aureus TCI (10%) as did Trinidad et al. (1990b) (12.3%). The higher prevalence of S. aureus IMI identified in the present trial may be due to seasonal differences (as evidenced by Fox et al. 1995), and horn fly populations. In the present trial, S. hyicus was identified in 8.8% of quarters (132 S. hyicus-infected teat canals out of 1,507 teat canals sampled), S. chromogenes in 6.0% of quarters (90 S. chromogenes-infected teat canals out of 1,507 teat canals sampled), S. simulans in 4.4% of quarters (66 S. simulans-infected teat canals out of 1,507 teat canals sampled), and S. intermedius in 2.2% of quarters (33 S. intermedius-infected teat canals out of 1,507 teat canals sampled) (Figure 45). Boddie et al. (1987) found greater prevalences of S. chromogenes (41%) TCI and S. hyicus (16.8%) compared to the present trial. Trinidad et al. (1990b) also identified greater prevalences of S. chromogenes (42.9%) TCI and S. hyicus (25.2%) TCI compared to the present trial. Overall, greater prevalences of TCI were identified by Boddie et al. (1987) and Trinidad et al. (1990b) with lower prevalences of S. aureus. Fly populations may have influenced the prevalence of S. aureus IMI in the present trial. Additionally, Nickerson and Boddie (1994), found that quarters infected with Staphylococcus spp., other than S. aureus (e.g., CNS), may be somewhat protected against development of S. aureus IMI. This may explain the lower prevalence of S. aureus IMI and greater prevalence of CNS IMI found in Trinidad et al. (1990b) and Boddie et al. (1987) compared to the present study. The CNS (overall prevalence of 15%; 232 out of 1,507 teat canals sampled) identified included S. chromogenes, S. hyicus, S. simulans, S. sciuri, S warneri, S. auricularis, S. hominis, 118

and S. caprae. The CPS (overall prevalence of 9.4%; 141 out of 1,507 teat canals sampled) identified included S. intermedius, S. hyicus, S. sciuri, S. carnosus, and S. simulans. Streptococci (overall prevalence of 1.6%, 24 out of 1,507 teat canals sampled) identified include Strep. uberis and Strep. dysgalactiae. Various other microorganisms were identified (4.3%, 64 out of 1,507 teat canals sampled) including gram-positive cocci, Micrococcus spp., Kocuria spp., E. coli, and Serratia spp. Across treatments, 52.4% of quarters were uninfected. Out of 716 teat canals of treated animals sampled, 51.4% (368 teat canals) were uninfected compared to 53.4% of teat canals sampled (422 out of 792) from controls (Figure 42). Across treatments, 47.6% of quarters were infected. Among quarters of treated animals, 48.6% of teat canals sampled were infected (348 out of 716) compared to 46.6% of teat canals sampled (369 out of 791) from controls. These data indicate a numerically higher prevalence of infection among quarters from treated heifers. Teat canals from treated animals exhibited a similar prevalence of S. aureus IMI (20.9%, 150 out of 716) compared to teat canals from control heifers (19.7%, 156 out of 791) (Figure 43). Teat canal infections caused by CNS (16.8%, 120 out of 716 vs. 14.2%, 112 out of 791) and those caused by CPS (8.5%, 61 out of 716 vs. 10.1%, 80 out of 791) showed contrasting relationships; however, numerically, percentages were not largely different. Quarters of control animals generally had a greater prevalence of infection caused by streptococci (2.7%, 21 out of 790 vs. 0.4%, 3 out of 716) compared to quarters from treated animals. 119

% of quarters Figure 45. Overall prevalence (%) of TCI by infection status and individual bacterial isolate. 1.6 2.2 4.3 4.4 6 8.8 52.4 20.3 Overall prevalence of infection = 47.6% Uninfected S. aureus S. hyicus S. chromogenes S. simulans S. intermedius Other* Streptococci * Includes S. xylosus, Serratia spp., S. sciuri, S. auricularis, Kocuria spp., S. caprae, grampositive cocci, E. coli, S. hominis, S. warneri, S. carnosus, Micrococcus spp., and CNS and CPS not identified by API Staph test. Figure 46. Distribution of TCI infection status by treatment groups among quarters. 55 53 51 51.4 53.4 Treated Control 49 48.6 47 46.6 45 43 41 Uninfected Infected 120

% of quarters Figure 47. Distribution of TCI in treated and control heifers by bacterial isolate. 25 20 15 20.9 19.7 16.8 14.2 Treated Control 10 8.5 10.1 5 0 2.7 1.3 0.4 S. aureus CNS CPS Streptococci Other* Infection status *Other includes Kocuria spp., E. coli, gram-positive cocci, Serratia spp., Micrococcus spp. The multivariable model analyzing prevalence of TCI among quarters would not converge because every quarter of the treated animals in group 1 had a teat canal infection at some time during study. Due to this reason, a table for prevalence of TCI caused by any pathogen is not displayed. Multivariable logistic regression for the prediction of S. aureus TCI revealed that the odds of S. aureus TCI was 1.5 times higher (OR = 1.5) in treated heifers compared to controls; however, this difference was not significant (P = 0.436) (Table 34). Regardless of treatment, the odds of S. aureus TCI in quarters from group 2 heifers were 75% lower (OR = 0.25) than quarters from group 1 heifers (P = 0.027). Additionally, the odds of S. aureus TCI in quarters from group 4 heifers were 85% lower (OR = 0.15) than group 1 (P = 0.044). There was no difference in the odds of S. aureus TCI between groups 1 and 3 (OR = 0.96, P = 0.942). The odds of S. aureus TCI in front teats were 4.5 times (OR = 4.5) higher to than rear teats (P < 0.001). The odds of S. aureus IMI in front quarters (vs. rear quarters) are higher because this location is in closer proximity to the navel region of the heifers, which is 121

where flies tend to congregate. Additionally, rear teats are protected by the tail switch that effectively repels flies. Table 35 displays the multivariable logistic results for the analysis of TCI caused by CNS. No difference was found between treated and control quarters (OR = 0.86, P = 0.717). The odds of teat canals from group 3 heifers to be diagnosed with CNS TCI were 84% lower (OR = 0.16, P = 0.005) compared to quarters from group 1 heifers. There was no difference in the odds of front teats compared to rear teats (OR = 0.32, P = 0.736) (Table 33). When examining the odds of teats from treated heifers being lower than teats from control quarters to be infected with CPS, there was no significant difference (OR = 0.87, P = 0.789) (Table 36). There were group differences, however. The odds of teats from groups 2 (OR = 0.18, P = 0.008), 3 (OR = 0.13, P = 0.007), and 4 (OR = 0.14, P = 0.024) being diagnosed with CPS TCI were lower than teats from group 1. 122

Table 34. Multivariable logistic regression results for the prediction of Staphylococcus aureus TCI based on teat canal keratin swabs from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control 0.387 (0.497) Referent 1.5 (0.56, 3.0) 0.436 Group 1 2 3 4 Referent -1.40 (0.632) -0.042 (0.576) -1.92 (0.954) 0.25 (0.07, 0.85) 0.96 (0.31, 3.0) 0.15 (0.02, 0.95) 0.027 0.942 0.044 Quarter location Rear Front Referent 1.51 (0.366) 4.5 (2.2, 9.3) < 0.001 Constant -1.19 (0.521) NA 0.023 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n =12) treated (n = 6) and control (n = 6); group 3 (n = 12) treated (n = 5) and control (n = 7); and group 4 (n = 9) treated (n = 4) and control (n = 5). Table 35. Multivariable logistic regression results for the prediction of coagulase-negative staphylococci (CNS) TCI based on teat canal keratin swabs from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control -0.153 (0.423) Referent 0.86 (0.37, 2.0) 0.717 Group 1 2 3 4 Referent -0.752 (0.492) -1.83 (0.656) -1.14 (0.682) 0.47 (0.18, 1.2) 0.16 (0.04, 0.58) 0.32 (0.08, 1.2) 0.126 0.005 0.094 Quarter location Rear Front Referent 0.148 (0.438) 1.2 (0.49, 2.7) 0.736 Constant -0.344 (0.486) NA 0.478 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n =12) treated (n = 6) and control (n = 6); group 3 (n = 12) treated (n = 5) and control (n = 7); and group 4 (n = 9) treated (n = 4) and control (n = 5). 123

Table 36. Multivariable logistic regression results for the prediction of coagulase-positive staphylococci (CPS) TCI based on teat canal keratin swabs from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control -0.142 (0.529) Referent 0.87 (0.31, 2.4) 0.789 Group 1 2 3 4 Referent -1.72 (0.648) -2.03 (0.752) -2.00 (0.883) 0.18 (0.05, 0.64) 0.13 (0.03, 0.57) 0.14 (0.02, 0.77) 0.008 0.007 0.024 Quarter location Rear Front Referent 1.08 (0.497) 3.0 (1.1, 7.8) 0.029 Constant -0.966 (0.552) NA 0.080 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n =12) treated (n = 6) and control (n = 6); group 3 (n = 12) treated (n = 5) and control (n = 7); and group 4 (n = 9) treated (n = 4) and control (n = 5). Multivariable logistic regressions studying the prediction of TCI caused by S. chromogenes demonstrated there was no difference between quarters from treated and control heifers (OR = 0.88, P = 0.827) (Table 37). Regardless of treatment, the odds of quarters to be diagnosed with S. chromogenes TCI from group 3 heifers were 91% lower than quarters from group 1 heifers (OR = 0.09, P = 0.032). Quarters from groups 2 and 4 were not significantly different from quarters of group 1 heifers (OR = 0.73, P = 0.648 and OR = 0.58, P = 0.495, respectively) There was no difference between front and rear quarters (OR = 1.9, P = 0.231). There was no difference in the odds of teats being infected with S. hyicus between treated or control heifers (OR = 0.75, Table 38). The odds of being diagnosed with S. hyicus TCI in groups 3 and 4 were significantly lower than teats from group 1 (OR = 0.04, P = 0.002 and OR = 0.17, P = 0.005, respectively). There was no difference in odds between teats from groups 1 and 124

2 (OR = 0.35, P = 0.062). Additionally, there was no difference in odds between front and rear teats (OR = 1.7, P = 0.314). In regards to the odds of teats being infected with S. simulans, there was no significant difference in odds between teats from treated and control heifers (OR = 1.3, P = 0.780) (Table 39). There were no group differences (OR = 0.34, P = 0.242; OR = 1.6, P = 0.488; and OR = 0.26, P = 0.232) identified compared to group 1, as well as no differences found between front and rear quarters (OR = 1.2, P = 0.811). The multivariable model analyzing TCI caused by streptococci would not converge because there were not any quarters in group 3 infected with streptococci during the study. Table 37. Multivariable logistic regression results for the prediction of Staphylococcus chromogenes TCI based on teat canal keratin swabs from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control -0.124 (0.565) Referent 0.88 (0.29, 2.7) 0.827 Group 1 2 3 4 Referent -0.314 (0.688) -2.41 (1.12) -0.537 (0.788) 0.73 (0.19, 2.8) 0.09 (0.01, 0.81) 0.58 (0.12, 2.7) 0.648 0.032 0.495 Quarter location Rear Front Referent 0.634 (0.529) 1.9 (0.67, 5.3) 0.231 Constant -1.73 (0.753) NA 0.022 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n =12) treated (n = 6) and control (n = 6); group 3 (n = 12) treated (n = 5) and control (n = 7); and group 4 (n = 9) treated (n = 4) and control (n = 5). 125

Table 38. Multivariable logistic regression results for the prediction of Staphylococcus hyicus TCI based on teat canal keratin swabs from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control -0.293 Referent 0.75 (0.29, 1.9) 0.548 Group 1 2 3 4 Referent -1.04 (0.558) -3.32 (1.10) -1.74 (0.617) 0.35 (0.12, 1.1) 0.04 (0.004, 0.31) 0.17 (0.05, 0.59) 0.062 0.002 0.005 Quarter location Rear Front Referent 0.544 (0.540) 1.7 (0.60, 5.0) 0.314 Constant -0.725 (0.596) NA 0.224 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n =12) treated (n = 6) and control (n = 6); group 3 (n = 12) treated (n = 5) and control (n = 7); and group 4 (n = 9) treated (n = 4) and control (n = 5). Table 39. Multivariable logistic regression results for the prediction of Staphylococcus simulans TCI based on teat canal keratin swabs from treated and control heifers 1. Variable Coefficient (Semirobust SE) OR (95% CI) P-value Treatment Treated Control 0.260 (0.575) Referent 1.3 (0.38, 3.6) 0.780 Group 1 2 3 4 Referent -1.07 (0.912) 0.475 (0.686) -1.35 (1.13) 0.34 (0.06, 2.1) 1.6 (0.42, 6.2) 0.26 (0.03, 2.4) 0.242 0.488 0.232 Quarter location Rear Front Referent 0.153 (0.638) 1.2 (0.33, 4.1) 0.811 Constant -2.23 (0.539) NA < 0.001 1 Group 1 (n = 9) treated (n = 4) and control (n = 5); group 2 (n =12) treated (n = 6) and control (n = 6); group 3 (n = 12) treated (n = 5) and control (n = 7); and group 4 (n = 9) treated (n = 4) and control (n = 5). 126

Effects of a general immunostimulant feed supplement on colony-forming unit (CFU) score of bacterial isolates and classification of infection Groups 1 through 4 were used in the analysis of CFU score. Based on the analysis of bacterial growth on blood agar plates from mammary secretions, CFU scores were assigned to each isolate where 0 = no growth, 1 = 0 to 20 CFU, 2 = 20 to 200 CFU, and 3 = CFU were too numerous to count. Across all bacterial species, quarters from control heifers tended to result in more 0 scores than quarters from treated heifers, indicating a lower prevalence of infection (39 vs. 31% of quarters) (Figure 48). Although there was no significant difference in prevalence among treated and control heifers as described previously, control heifers did exhibit a numerically lower prevalence of infection than treated heifers. Isolates from quarters of treated heifers tended to have higher CFU scores of 1 and 3 (5 and 46%, respectively) than isolates from quarters of control heifers (3 and 40%), and as discussed, prevalence of mastitis was higher in treated heifers than control heifers. Percentage of quarters between treated and control heifers with a CFU score of 2 were numerically the same (18%). Figure 49 presents the CFU scores for S. aureus, the most predominant mastitis-causing pathogen identified. S. aureus has also been documented as being one of the most detrimental pathogens to developing mammary tissue (Sordillo et al., 1989). Interestingly, the majority of S. aureus isolates, regardless of treatment group, were associated with a score of 3, indicating an association between severity of infection and CFU score, e.g., a score of 3 suggests that the bacteria have the ability to actively multiply and are not being controlled by the immune system. Of the S. aureus isolates quantified, treated animals were associated with lower CFU scores (69.7%) compared to controls (78.8%). For all other bacterial isolates (grouped together) the majority of species exhibited a CFU score of 3 (67.6% treated and 43% control) (Figure 50). 127

% of total S. aureus isolates % of quarters Figure 48. Percentage of total isolates by colony-forming unit (CFU) score 1 between treated (n = 267 quarter samples) and control (n = 214 quarter samples) heifers. 50 45 40 35 30 25 20 15 10 5 0 31 39 1 Scores were assigned as follows: 0 = No growth on plate, 1= 0 to 20 CFU, 2 = 21 to 200 CFU, or 3 = too numerous to count (TNTC). 5 Treated 3 Control 0 1 2 3 CFU score 18 18 46 40 Figure 49. Percentage of total Staphylococcus aureus isolates by colony-forming unit (CFU) score 1 between treated (n = 89 quarter samples) and control (n = 52 quarter samples) heifers. 80 70 60 50 40 30 20 10 0 1 Scores were assigned as follows: 0 = No growth on plate, 1= 0 to 20 CFU, 2 = 21 to 200 CFU, or 3 = too numerous to count (TNTC). 9 1.9 Treated 21.3 Control 19.2 67.7 1 2 3 CFU score 78.8 128

% of total isolates Figure 50. Percentage of total isolates 1, excluding Staphylococcus aureus, by CFU score 2 between treated (n = 74 quarter samples) and control (n = 80 quarter samples) heifers. 80 70 60 50 40 30 20 10 0 6.8 7.5 1 Isolates included the following bacterial species: S. chromogenes, S. hyicus, S. simulans, S. xylosus, S. intermedius, S. auricularis, Serratia spp., Streptococcus spp., gram-positive cocci, Yeast, Micrococcus spp. 2 Scores were assigned as follows: 0 = No growth on plate, 1= 0 to 20 CFU, 2 = 21 to 200 CFU, or 3 = too numerous to count (TNTC). Treated 25.7 Control 38.8 67.6 1 2 3 CFU Score 53.8 Effects of general immunostimulant feed supplement on new IMI rate, TCI rate, chronicity of infection, response to antibiotic therapy, and spontaneous cure rate Heifers from groups 1 to 4 from August 2010 through October 2011 were used in the analysis of these parameters. New IMI rate of quarters becoming infected over the trial period was similar among treated and control heifers (Figure 51). In both treatment groups, an average of approximately 2.5 quarters/heifer exhibited a new infection determined by taking the number of cases (among quarters) then dividing that value by the number of heifers used in the analysis. Additionally, both treatment groups exhibited a similar number of quarters/heifer with a chronic infection (3.2 vs. 3.1), which included existing infections present at the first sampling. Thus, treated animals did not exhibit any advantages by receiving the supplement compared to unsupplemented controls based on new or chronic IMI. Cure rate of quarters from treated heifers 129

Number of quarters treated with dry cow antibiotic treatment (DCT) was 100% (29 quarters treated, 29 quarters cured) compared to a cure rate of 88.9% among controls (14 quarters treated, 13 quarters cured). There were no spontaneous cures in either treatment group; all IMI established during the trial remained unless controlled by dry cow antibiotic therapy. Analysis of TCI (Figure 52) yielded slightly different results than those from IMI. Number of new TCI was greater in quarters of treated heifers vs. controls (3.2 vs. 2.7). Likewise, number of chronic TCI was greater in quarters of treated heifers vs. controls (3.6 vs. 3.3). Quarters from treated animals exhibited fewer cases of spontaneous cures (4 cases) compared to 7 cases among control animals. While numerical differences were found in regards to these categories of infection, overall prevalence of TCI was not significantly different between treatment groups as evidenced previously. Total numbers for all events can be found in Appendices 7, 8, 9, and 10. Figure 51. Average number of new IMI and chronic IMI between treatment groups. 4 3.5 3 2.5 2 1.5 1 0.5 Treated Control 3.2 3.1 2.5 2.6 0 New IMI Chronic IMI 130

Number of quarters Figure 52. Average number of new TCI and chronic TCI, between treatment groups. 4 Treated Control 3.5 3 2.5 2 1.5 1 0.5 3.6 3.2 3.3 2.7 0 New TCI Chronic TCI Effects of a general immunostimulant feed supplement on SCC of mammary secretions Regardless of treatment, secretions from infected quarters across bacterial species exhibited an overall average SCC of 3,251 x 10 3 /ml, whereas secretions from uninfected quarters exhibited an average SCC of 2,044 x 10 3 /ml (Figure 53). Boddie et al. (1987) found similar trends in SCC after studying 10 unbred heifers, and reported that the SCC of infected quarters, depending on the causal pathogen, ranged from 7.8 to 9.2 x 10 6 /ml, whereas uninfected quarters exhibited a SCC of 3.5 x 10 6 /ml. Trinidad et al. (1990b) also reported similar trends when analyzing SCC of uninfected (5.7 x 10 6 /ml) and infected quarters (12.4 x 10 6 /ml to 17.3 x 10 6 /ml). The lower overall SCC found in the present compared to Boddie et al. (1987) and Trinidad et al. (1990b) may be explained due to the technique used to determine SCC. This trial used the DeLaval Direct Cell Counter (DeLaval, Tumba, Sweden); whereas the other trials, used a Fossomatic electronic cell counter (ALSN Foss, Hillerød, Denmark). Using a hemocytometer to determine total leukocyte counts, Jackson et al. (2011) reported a SCC of 8,600 x 10 3 /ml prior to challenging quarters with Strep. uberis, which was higher than SCC of uninfected 131

quarters in the present trial (2,044 x 10 3 /ml). Challenged quarters peaked at 24,300 x 10 3 /ml, whereas unchallenged (uninfected) quarters were 2,800 x 10 3 /ml. Both infected SCC and uninfected SCC after challenge in the study by Jackson et al. (2011), were greater than uninfected and infected SCC of mammary secretions from heifers in the present study because total leukocyte counts were determined using a hemocytometer, which most likely explains the differences. Additionally, Jackson et al. (2011) experimentally challenged quarters, whereas the present trial analyzed SCC from naturally infected quarters. When comparing SCC of treated and control heifers, secretions collected from infected quarters of treated heifers exhibited a mean SCC of 3,304 x 10 3 /ml and secretions from infected quarters of control heifers exhibited a SCC of 3,220 x 10 3 /ml (Figure 54), illustrating no effect of treatment. Similar trends were observed when SCC in secretions from uninfected quarters of treated and control heifers were measured demonstrating 2,200 x 10 3 /ml vs. 1,925 x 10 3 /ml, respectively (Figure 56). It was contended that supplementation would decrease prevalence of new IMI, and hence, lower the SCC during this time. Conversely, if SCC was higher in treated animals, it could be argued that supplementation enhanced the extravasation of leukocytes into the mammary gland, thus quarters were more protected and leukocytes were readily available to defend against bacterial infection. However, no difference between treatments was observed. In contrast to the present study, Wang et al. (2009) observed an increase in the total blood leukocyte counts in cows supplemented with OmniGen-AF compared to unsupplemented controls (10,300 vs. 6,950/mm 3 ), which could provide a pool for extravasation into the mammary gland for increasing SCC. 132

SCC x 10 3 /ml SCC x 10 3 /ml Figure 53. Comparison of somatic cell counts (mean ± SD) of quarter mammary secretions from infected and uninfected quarters regardless of treatment group (n = 332 total samples). Infected Uninfected 6000 5000 3,251 4000 2,044 3000 2000 1000 0 Figure 54. Comparison of somatic cell counts (mean ± SD) of quarter mammary secretions among treatment groups based on infection status (n = 332 total samples). 6000 5000 4000 3,304 3,220 Treated Control 2,200 1,925 3000 2000 1000 0 Infected Infection status Uninfected The mean SCC in secretions from S. aureus-infected quarters of treated heifers was 3,513 x 10 3 /ml compared to a mean SCC of 3,190 x 10 3 /ml of control heifers (Figure 55). It could be argued that the influx of leukocytes into the mammary gland would provide greater protection against bacterial invasion; however, the process of diapedesis and ROS release destroys milk- 133

producing cells, potentially decreasing future milk production as discussed in a review by Akers and Nickerson (2011). For those quarters infected with CPS and CNS, treated heifers had a slightly lower SCC (2,684 x 10 3 /ml and 3,043 x 10 3 /ml, respectively) compared to control heifers (2,932 x 10 3 /ml and 3,133 x 10 3 /ml, respectively) (Figure 55). Boddie et al. (1987) and Trinidad et al. (1990b) found greater SCC in S. aureus-infected quarters (9.2 x 10 6 /ml and 17.3 x 10 6 /ml) than was found in the present trial most likely to due to the method of SCC determination (electronic cell counter vs. direct cell counter). Analysis of other individual bacterial isolates yielded similar results (Figure 56). Quarter secretions from treated heifers infected with S. chromogenes, S. hyicus, and Streptococcus spp. exhibited somewhat lower mean SCC (2,755, 3,214, and 2,795 x 10 3 /ml, respectively) than secretions infected with those same pathogens collected from quarters of control heifers (3,170, 3,321, and 3,536 x 10 3 /ml, respectively). Quarter secretions from treated heifers infected with S. simulans; however, exhibited a higher mean SCC (3,010 x 10 3 /ml) compared to secretions infected with S. simulans collected from control heifers (2,541 x 10 3 /ml). Boddie et al. (1987) and Trinidad et al. (1990b) observed greater SCC in S. chromogenes-infected quarters (7.8 x 10 6 /ml and 12.8 x 10 6 /ml, respectively) and S. hyicus-infected quarters (8.5 x 10 6 /ml and 12.4 x 10 6 /ml, respectively) compared to those found in the present trial. This difference was most likely to due to the method of SCC determination used in the studies. 134

Mean SCC x 10 3 /ml Mean SCC x 10 3 /ml Figure 55. Comparison of somatic cell counts (mean ± SD) of quarter mammary secretions infected with Staphylococcus aureus, coagulase-positive staphylococci (CPS), and coagulasenegative staphylococci (CNS) among treatment groups (n = 238 total samples). 6000 5000 3,513 3,190 Treated Control 2,932 2,684 3,043 3,133 4000 3000 2000 1000 0 S. aureus CPS CNS Bacterial isolate Figure 56. Comparison of somatic cell counts (mean ± SD) of quarter mammary secretions infected with predominant pathogens (excluding Staphylococcus aureus), by bacterial species, among treatment groups (n = 127 total samples). 7000 6000 5000 2,755 3,170 Treated Control 3,214 3,321 2,541 3,010 2,795 3,536 4000 3000 2000 1000 0 S. chromogenes S. hyicus S. simulans Streptococcus spp. Bacterial isolate 135

Effects of a general immunostimulant feed supplement on differential leukocyte counts in mammary secretions An analysis of differential leukocyte counts showed that secretions collected from infected quarters exhibited similar percentages of lymphocytes in infected and uninfected quarters (13.4% and 13.9 %, respectively) (Figure 57). Conversely, secretions from uninfected quarters displayed predominately macrophages compared to secretions from infected quarters (70.9% vs. 45.9%, respectively). Secretions collected from infected quarters exhibited more neutrophils than secretions from uninfected quarters (40.7% vs. 15.2%, respectively). Regardless of infection status, macrophages were the predominant leukocyte type found in secretions (45.9 to 70.9%), which was similar to other studies. For example, Nickerson et al. (1995) found that infected and uninfected quarters exhibited 45.5% and 52.9% macrophages, respectively, and that neutrophils accounted for 20.6% and 16.5% (infected and uninfected, respectively) and lymphocytes accounted for 33.9% and 30.6% (infected and uninfected, respectively). Jackson et al. (2011) found that uninfected quarters, prior to Strep. uberis challenge, exhibited 15.3% lymphocytes, 81% macrophages, and 3.7% neutrophils. The differential counts reported by Jackson et al. (2011) are similar to those found in the present trial (13.9%, 70.9%, and 15.2%) as macrophages were the most predominant cell type in both studies. In Jackson et al. (2011), infected quarters exhibited 3% lymphocytes, 22% macrophages, and 75% neutrophils. Infected quarters from the present trial did not exhibit the increased neutrophils identified by Jackson et al. (2011). The heifers in the previous study were experimentally challenged, whereas the heifers in the present trial were naturally infected, some chronically, which may explain the differences in differential leukocyte counts of infected quarters. 136

Regardless of infection status, secretions from quarters of treated and control animals exhibited similar overall differential leukocyte counts (Figure 58). In secretions from treated heifers, lymphocytes accounted for 13.2% of leukocytes, macrophages 52.6%, and neutrophils 34.2%. In secretions from control heifers, lymphocytes accounted for 13.8% of leukocytes, 51% of macrophages, and 35.2% of neutrophils. These same trends were identified in the analysis of secretions collected from infected quarters of treated (lymphocytes 13.0%, macrophages 45.6%, and neutrophils 41.4%) and control heifers (lymphocytes 13.9%, macrophages 46.7%, and neutrophils 39.4%), which exhibited similar overall differential leukocyte counts between treatment groups (Figure 59). Secretions from uninfected quarters (Figure 60) of treated and control quarters exhibited similar percentages of lymphocytes (14.0 and 13.7%, respectively), whereas secretions from treated heifers exhibited more macrophages (73.2%) and fewer neutrophils (12.8%) than those from control heifers (65.8% and 20.5%, respectively). If neutrophils were lower in treated animals, this may have indicated less ROS production, thereby decreased destruction of alveoli. Additionally, if lymphocytes were greater, specific (adaptive) immunity may be able to assist more efficiently in mammary defense. Conversely, if both neutrophils and lymphocytes were greater in treated animals, it could be argued that the mammary gland was better able to challenge any invading pathogens and sustain specific defenses against those pathogens. However, no treatment differences were observed. In contrast to the present study, Wang et al. (2009) observed an increase in the circulating neutrophils and lymphocytes in cows supplemented with OmniGen-AF compared to unsupplemented controls (4,200 vs. 2,456/mm 3 and 5,960 vs. 4,000/mm 3, respectively), which could provide a pool for extravasation into the mammary gland for increasing neutrophils and lymphocytes. 137

% leukocytes % leukocytes Figure 57. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in quarter mammary secretions from infected and uninfected quarters, regardless of treatment group (n = 283 total samples). 100 90 80 70 60 50 40 30 20 10 0 Infected Uninfected 70.9 45.9 40.7 15.2 13.4 13.9 Lymphocyte Macrophage Neutrophil Type of leukocyte Figure 58. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in heifer mammary secretions of treated (n = 185) and control (n = 98) heifers regardless of infection status. 100 90 80 70 60 50 40 30 20 10 0 Treated Control 52.6 51 34.2 35.2 13.2 13.8 Lymphocyte Macrophage Neutrophil Type of leukocyte 138

% leukocytes % leukocytes Figure 59. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in infected mammary secretions of treated (n = 138) and control (n = 76). 100 90 80 70 60 50 40 30 20 10 0 Treated Control 45.6 46.7 41.4 39.4 13.0 13.9 Lymphocyte Macrophage Neutrophil Type of leukocyte Figure 60. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in uninfected mammary secretions of treated (n = 47) and control (n = 22) heifers. 100 90 80 70 60 50 40 30 20 10 0 Treated Control 73.2 65.8 20.5 12.8 14.0 13.7 Lymphocyte Macrophage Neutrophil Type of leukocyte Only very small numerical differences in the differential counts of secretions from S. aureus-infected quarters of treated and control animals were noted (Figure 61). Secretions collected from quarters of treated heifers exhibited 11.7% lymphocytes, 44.3% macrophages, and 44.0% neutrophils. Secretions from control heifers exhibited 14.1% lymphocytes, 42.9% 139

macrophages, and 43%) neutrophils, which exhibited similar overall differential leukocyte counts between treatment groups. Differential counts of secretions of quarters infected with CPS showed that secretions from quarters of treated animals and control animals exhibited similar percentage of lymphocytes (12.3 vs. 12.6%) (Figure 62). Secretions from treated animals tended to exhibit more macrophages (52.6%) and fewer neutrophils (35.1%) than secretions from quarters of control heifers (57.7 and 29.7%, respectively). While similar trends were identified regarding the differential counts of secretions from quarters infected with CNS, the values were numerically closer than those previously discussed (Figure 63). Secretions from treated heifers exhibited 14.4% lymphocytes, 45.8% macrophages, and 39.9% neutrophils. Similarly, secretions from control heifers exhibited 14.5% lymphocytes, 46.7% macrophages, and 38.8% neutrophils. Quarters infected with S. hyicus showed similar trends (Figure 64). Secretions from treated heifers exhibited 12.7% lymphocytes, 45.2% macrophages, and 42.1% neutrophils. Secretions from control heifers exhibited 15.0% lymphocytes, 43.3% macrophages, and 41.7% neutrophils. Differential counts of secretions collected from quarters infected with S. chromogenes are displayed in Figure 65. Secretions from quarters of treated animals exhibited fewer lymphocytes (9.8%), fewer macrophages (29.9%), and greater neutrophils (60.3%) than secretions from control heifers (14.9%, 44.8%, and 40.3%, respectively). While the numerical differences were not as great in differential counts collected from quarters infected with streptococci, similar trends were identified (Figure 66). Secretions from treated animals tended to display slightly more lymphocytes than secretions from control heifers (14.5 vs. 10.9%). Conversely, secretions from control heifers exhibited more macrophages (57.8%) compared to secretions from treated 140

% of leukocytes % of leukocytes heifers (43.0%). Secretions from treated heifers exhibited more neutrophils (42.5%) than secretions from control animals (31.3%). Figure 61. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in Staphylococcus aureus-infected heifer mammary secretions of treated (n = 75) and control heifers (n = 33). 100 90 80 70 60 50 40 30 20 10 0 Treated Control 44.3 44 42.9 43.0 11.7 14.1 Lymphocyte Macrophage Neutrophil Type of leukocyte Figure 62. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in coagulasepositive staphylococci (CPS) infected mammary secretions of treated (n = 9) and control (n = 3) heifers. 100 90 80 70 60 50 40 30 20 10 0 Treated Control 52.6 57.7 35.1 29.7 12.3 12.6 Lymphocyte Macrophage Neutrophil Type of leukocyte 141

% of leukocytes % of leukocytes Figure 63. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in coagulasenegative staphylococci (CNS) infected mammary secretions of treated (n = 45) and control (n = 32) heifers. 100 90 80 70 60 50 40 30 20 10 0 Treated Control 45.8 46.7 39.9 38.8 14.4 14.5 Lymphocyte Macrophage Neutrophil Type of leukocyte Figure 64. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in Staphylococcus hyicus-infected heifer mammary secretions of treated (n = 27) and control (n = 14) heifers. 100 90 80 70 60 50 40 30 20 10 0 Treated Control 45.2 43.3 42.1 41.7 12.7 15 Lymphocyte Macrophage Neutrophil Type of leukocyte 142

% of leukocytes % of leukocytes Figure 65. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in Staphylococcus chromogenes-infected mammary secretions of treated (n = 9) and control (n = 15) heifers. 100 90 80 70 60 50 40 30 20 10 0 Treated Control 60.3 44.8 40.3 29.9 14.9 9.8 Lymphocyte Macrophage Neutrophil Type of leukocyte Figure 66. Percentage (mean ± SD) of lymphocytes, macrophages, and neutrophils in Streptococcus spp.-infected heifer mammary secretions of treated (n = 4) and control (n = 10) heifers. 100 90 80 70 60 50 40 30 20 10 0 Treated Control 57.8 43 42.5 31.3 14.5 10.9 Lymphocyte Macrophage Neutrophil Type of leukocyte 143

Effects of a general immunostimulant feed supplement on prevalence of clinical mastitis Presence of clinical symptoms in mammary secretions was assessed at the time of sampling as well as by visual observation of secretions in test tubes prior to determination of SCC. Clinical signs of infection included ropy, clotted, and bloody secretions (Figure 67 and Figure 68). Most quarters appeared normal and healthy, whereas very few exhibited swollen quarters (Figure 69). A SCC was assigned to clinical secretions retrospectively after a maximum threshold approaching 7,000 x 10 3 /ml was determined. All clinical secretions were assigned this SCC for data analysis. Percentage of new clinical cases, recurring clinical cases, clinical cases caused by S. aureus, and clinical cases occurring in the front quarters (right front or left front) are shown in Tables 40 (among Figure 68. Secretion from quarter infected with Staphylococcus aureus. heifers) and 41(among quarters). Out of the 4 groups LR RR Figure 67. Ropy, discolored secretion being expressed from heifer infected with Staphylococcus aureus. Figure 69. Swollen right rear quarter and normal left rear both infected with Staphylococcus aureus. 144