Understanding Biological Factors Associated with Pelvic Organ Prolapse in Late Gestation Sows

Iowa State University Capstones, Theses and Graduate Theses and Dissertations Dissertations

2021

Understanding biological factors associated with pelvic organ prolapse in late gestation sows

Zoe E. Kiefer Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/etd

Recommended Citation Kiefer, Zoe E., "Understanding biological factors associated with pelvic organ prolapse in late gestation sows" (2021). Graduate Theses and Dissertations. 18526. https://lib.dr.iastate.edu/etd/18526

This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Understanding biological factors associated with pelvic organ prolapse in late gestation sows

Zoë Elizabeth Kiefer

A thesis submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Major: Animal Physiology (Reproductive Physiology)

Program of Study Committee: Jason W. Ross, Major Professor Aileen F. Keating Stephan Schmitz-Esser

The student author, whose presentation of the scholarship herein was approved by the program of study committee, is solely responsible for the content of this thesis. The Graduate College will ensure this thesis is globally accessible and will not permit alterations after a degree is conferred.

Iowa State University

Ames, Iowa

2021

DEDICATION

I dedicate this thesis to everyone who has encouraged and supported me throughout this journey. To my family, friends, colleagues and mentors, thank you for helping me become the scientist I am today. To the Iowa State University Department of Animal Science thanks for a great experience and excellent education.

LOYAL ● FOREVER ● TRUE

iii

TABLE OF CONTENTS

Page

LIST OF FIGURES ...... v

LIST OF TABLES ...... vi

NOMENCLATURE ...... vii

ACKNOWLEDGMENTS ...... ix

ABSTRACT ...... x

CHAPTER 1. INTRODUCTION ...... 1 References ...... 2

CHAPTER 2. LITERATURE REVIEW ...... 4 Sow Reproduction ...... 4 Onset of Puberty ...... 4 Estrous Cycle ...... 5 Gilt Development ...... 6 Sow Mortality in the U.S. Swine Industry ...... 7 Reproductive Tract Dysfunction Overview ...... 7 Reproductive Disease in Humans...... 8 Reproductive Disease in Cattle ...... 9 Physiology of Pelvic Organ Prolapse ...... 10 Vaginal Microbiota Overview ...... 11 Vaginal Microbiota of Humans ...... 12 Vaginal Microbiota of Cattle ...... 14 Vaginal Microbiota of Swine ...... 15 Role of the Vaginal Microbiota in the Immune System ...... 16 Immune and Inflammatory Response ...... 17 Steroid Hormone Changes During Late Gestation ...... 20 Steroidogenesis ...... 20 Steroid Hormone Effects on the Immune Response ...... 21 Steroid Hormones and their Potential Influence on the Microbiota ...... 22 Summary ...... 23 References ...... 24

CHAPTER 3. VAGINAL MICROBIOME AND SERUM METABOLITE DIFFERENCES IN LATE GESTATION COMMERCIAL SOWS AT RISK FOR PELVIC ORGAN PROLAPSE .33 Abstract ...... 33 Introduction ...... 34 Results ...... 35 Discussion ...... 38 Materials and Methods ...... 43 iv

Statistical analyses ...... 47 Data availability ...... 49 Acknowledgements ...... 49 Conflict of Interest ...... 50 Authors contributions ...... 50 References ...... 50

CHAPTER 4. VAGINAL MICROBIOTA VARIATION DURING LATE GESTATION COMMERCIAL SOWS A RISK FOR PELVIC ORGAN PROLAPSE ...... 63 Abstract ...... 63 Introduction ...... 64 Materials and Methods ...... 66 Results ...... 71 Discussion ...... 75 Acknowledgements ...... 81 Conflict of Interest ...... 81 Authors contributions ...... 81 References ...... 82

CHAPTER 5. CIRCULATING BIOMARKERS ASSOCIATED WITH PELVIC ORGAN PROLASPE RISK IN LATE GESATION SOWS ...... 98 Abstract ...... 98 Introduction ...... 99 Materials and Methods ...... 101 Results ...... 104 Discussion ...... 106 Conclusion ...... 110 Acknowledgements ...... 111 Disclosures ...... 111 References ...... 111

CHAPTER 6. SUMMARY AND CONCLUSION ...... 119 Future Directions ...... 122 References ...... 125

APPENDEX: ADDITIONAL DATA...... 128 v

LIST OF FIGURES

Page

Figure 3.1. Perineal score in late gestation as an indicator of pelvic organ prolapse (POP) risk...... 54

Figure 3.4. Different OTUs in the vaginal microbiota of sows with different perineal scores. ... 57

Figure 3.5. Serum metabolites in sows differing in perineal score (PS) and POP risk...... 58

Figure 3.6. Differences in serum metabolites between PS1 and PS3 sows...... 59

Figure 3.7. Perineal scoring (PS) methodology as an indicator of pelvic organ prolapse (POP)...... 60

Figure 4.1. Perineal score (PS) changes throughout late gestation...... 86

Figure 4.2. Vaginal microbial community comparisons...... 87

Figure 4.3. Alpha diversity of the vaginal microbiota for sows during late gestation...... 88

Figure 4.4. Abundance of OTUs in the vaginal microbiota of late gestation sows...... 89

Figure 5.1. Perineal score (PS) at gestation week fifteen is an indicator of pelvic organ prolapse (POP) risk...... 118 vi

LIST OF TABLES

Page

Table 3.1. Differences in OTUs between vaginal microbiomes of PS1 and PS3 sows...... 61

Table 3.2. Small molecule metabolites in serum from sows with differing risk of ...... 62

Table 4.1. The 50 most abundant bacteria in vaginal samples from sows...... 90

Table 4.2. Differences in OTUs between vaginal microbiota of PS1 and PS3 sows during gestation week 15...... 92

Table 4.3. Differences in OTUs between vaginal microbiota of sows during gestation week 15 assigned PS3 that subsequently did or did not experience pelvic organ prolapse...... 94

Table 4.4. Differences in OTUs between vaginal microbiota of gestation week 12 PS1 sows and gestation week 15 PS1 sows...... 95

Table 4.5. Differences in OTUs between vaginal microbiota of gestation week 12 PS1 sows and gestation week 15 PS3 sows...... 96

Table 5.1. Complete blood count analysis differences between late gestation sows with differing risk for pelvic organ prolapse (POP)1...... 115

Table 5.2. Differences in circulating steroid hormones between sows differing in risk for pelvic organ prolapse (POP)1 ...... 116

Table 5.3. Serum biomarker analysis between late gestation sows differing in risk for pelvic organ prolapse (POP)...... 117

vii

NOMENCLATURE

E2 17β-estradiol APP Acute phase protein ANOSIM Analysis of similarity AI Artificial insemination BV Bacterial vaginosis CAP Canonical analysis of Principle Coordinates CCA Canonical correlation analysis JNK cJun NH2-terminal kinases CBC Complete blood count CL Corpora lutea CRP C-reactive protein CK Creatine kinase LEfSe Effect Size ELISA Enzyme-linked immunosorbent assay EDTA Ethylenediaminetetraacetic acid FC Fold change GC-MS Gas chromatography with tandem mass spectrometry GnRH Gonadotrophin releasing hormone IAV Influenza A virus ISU Iowa State University LDA Linear Discriminant Analysis LPS Lipopolysaccharide LBP Lipopolysaccharide binding protein log2FC Log2-fold change LH Luteinizing hormone MHC Major histocompatibility complex MPV Mean platelet volume MAPL Mitochondria-associated protein ligase MAPK Mitogen-activated protein kinase MHP Mycoplasma hyopneumoniae NF-ĸB Nuclear Factor Kappa B OTUs Operational taxonomic units PID Pelvic inflammatory disease POP Pelvic organ prolapse PS Perineal score PERMANOVA Permutational analysis of variance PBS Phosphate buffered saline viii

PWSY Pigs weaned per sow per year PED Porcine epidemic disease PRRS Porcine reproductive and respiratory syndrome PCoA Principle coordinate analysis P4 Progesterone STI Sexually transmitted infections TLR Toll-like receptor TNF-α Tumor necrosis factor alpha U.S. United States VVC Vulvo-vaginal candidiasis

ACKNOWLEDGMENTS

I would like to thank my committee chair, Dr. Jason Ross, and my committee members,

Dr. Stephan Schmitz-Esser, and Dr. Aileen Keating, for their guidance and support throughout the course of this research. To my major professor, Dr. Jason Ross, whom I cannot express enough gratitude towards, and will forever be grateful for him giving this Chicago girl a shot in the pig world.

I would also like to thank the past and current members of the Ross lab, as well as the physiology group faculty and graduate students for their assistance on all things research related in addition to being great colleagues outside of work, making my time as a graduate student that much more enjoyable. I would specifically like to thank Dr. Elizabeth Hines, for teaching me that pigs are actually pretty cool and instilling in me a passion for research. I also want to extend my appreciation to the undergraduate students, IPIC team members, and industry partners who contributed to these experiments.

Lastly, I would also like to thank my family and friends for their encouragement, and continuous support throughout my time at Iowa State University.

ABSTRACT

Sow reproductive efficiency is critical to the sustainability of the U.S. swine herd, and throughout the last few decades the U.S. swine industry has made significant improvements in several areas of production efficiency, including reproductive performance. However, over the past decade sow mortality rates have increased, with a disproportionate increase due to pelvic organ prolapse (POP). Most commonly, POP occurs during the peripartum period, and approximately 21% of all sow deaths are attributed to POP. Unfortunately, there is no good biological understanding of the root causes for this anatomical phenomenon. A perineal scoring

(PS) system was developed, based on phenotypic observations associated with POP outcomes that can be correlated with POP risk. Late gestation sows were assigned PS using this newly developed system, and sows were assigned PS1 (low), PS2 (moderate), or PS3 (high) based on

POP risk. This system was then utilized to explore differences in the vaginal microbiota in relation to POP risk (Chapter 3 and Chapter 4) in addition to identification of steroid hormones, markers of immune system activation, and other potential biomarkers of POP risk in circulation

(Chapter 3 and Chapter 5). Vaginal swabs and blood were collected on PS3 sows who were parity matched with PS1 sows, and at gestation week 12 on a subset of sows. Vaginal swabs were utilized for 16S rRNA gene sequencing and differences in the vaginal microbiota between

PS1 and PS3 sows, and PS3 sows that subsequently did or did not experience POP. Additionally, several microbes were observed to be consistent across studies, sharing the same trends in abundance differences between PS. Serum molecular features were evaluated via GC-MS, and differences in abundancies were detected between sows at differing risk for POP. Immune cell populations, biomarkers of inflammation and steroid hormone profiles were also altered between assigned PS1 compared to PS3 sows. Collectively, these data suggest differences exist between xi microbial populations, biomarkers for inflammation, and serum metabolites between sows at differing risk for POP during late gestation.

CHAPTER 1. INTRODUCTION

Pork is one of the most consumed meats in the world, and third most in the United States

(U.S.). As the global population continues to expand, farmers must increase production in order to meet the accompanying demands for protein (Roser, 2014). Satiating this demand for a growing population will require increased and more efficient production of safe and wholesome product produced in a sustainable manner. Thus, producers must continually seek improved production strategies while balancing economics, animal welfare and sustainability. The sow herd is an essential part of the swine industry and is key to efficiently producing adequate piglets to meet the global call for increased pork. In the U.S., a sow produces an average of 3.5 litters in her lifetime (Plain and Lawrence, 2003), and has the potential to wean approximately 16 pigs per litter. During 2020, the average sow weaned 11.9 piglets per litter based on PigChamp’s benchmarking report (PigCHAMP, 2020). Additionally, the target number of pigs weaned per sow per year (PWSY) has increased over the last several decades from 20 to 30 pigs, and has the potential for further growth (Koketsu et al., 2017). Increases in sow herd efficiency has improved overall pork production while decreasing the sow population needed (Plain and Lawrence, 2003).

Sows also represent an expensive production cost, thus decreasing the number of sows needed can decrease inputs while reducing the environmental footprint of pork production. The expense of adding sows to the herd fluctuates due to feed costs, productivity, and markets, but in general, a sow must produce at least three parities for the value of the weaned pigs to offset overall costs and become profitable to the enterprise (Stalder et al., 2003). In 2010 the average parity of a sow in the U.S. was 2.7 (Rix and Ketchem, 2010). Therefore, it remains imperative that producers maximize the average lifetime number of parities produced by ensuring and supporting sow health and longevity. 2

The U.S. swine industry has made significant improvements in several areas of production efficiency, including reproductive performance. However, sow mortality rates have been increasing substantially, with an unproportioned increase due to pelvic organ prolapse

(POP) (Supakorn et al., 2014). A survey conducted in 2018 discovered that approximately 21% of sow mortality was the result of POP (Ross, 2019). Phenotypically, POP is characterized by one or more of the pelvic organs (uterus, vagina, and/or rectum) pressing up against or out of the vagina (Jelovsek et al., 2007). Most often, sows experience POP during the peripartum period, leading up to and shortly after farrowing. Due to the timing of POP in sows, it may negatively affect piglet survival as well, furthering the economic and animal welfare consequences.

Maximizing sow lifetime productivity is necessary to improve the sustainability and profitability of the swine industry. While this industry-wide problem is both an animal welfare concern and economic issue, there is currently a lack of understanding of the biological causes of POP, which has created a significant barrier in the development of effective mitigation strategies.

References

Jelovsek, J. E., C. Maher, and M. D. Barber. 2007. Pelvic organ prolapse. Lancet. 369:1027– 1038. doi:10.1016/S0140-6736(07)60462-0.

Koketsu, Y., S. Tani, and R. Iida. 2017. Factors for improving reproductive performance of sows and herd productivity in commercial breeding herds. Porc. Heal. Manag. 3:1. doi:10.1186/s40813-016-0049-7.

PigCHAMP. 2020. Benchmark 2020 - US. PigCHAMP.

Plain, R. L., and J. D. Lawrence. 2003. Swine production. Vet. Clin. North Am. - Food Anim. Pract. 19:319–337. doi:10.1016/S0749-0720(03)00025-2.

Rix, M., and R. Ketchem. 2010. A Closer Look at Sow Herd Parity Structure. Natl. Hog Farmer.

Roser, M. 2014. Future Population Growth. Our World Data.

Ross, J. W. 2019. Identification of putative factors contributing to pelvic organ prolapse in sows (Grant # 17-224) II. Industry Summary. 3

Stalder, K. J., ; R Curt Lacy, T. L. Cross, and G. E. Conatser. 2003. Financial impact of average parity of culled females in a breed-to-wean swine operation using replacement gilt net present value analysis. J. Swine Heal. Prod. 11:69–74.

Supakorn, C., J. D. Stock, C. Hostetler, and K. J. Stalder. 2014. Prolapse Incidence in Swine Breeding Herds Is a Cause for Concern. Open J. Vet. Med. 7:85–97. doi:10.4236/ojvm.2017.78009. 4

CHAPTER 2. LITERATURE REVIEW

Sow Reproduction

Increasing sow farm productivity through improved reproductive efficiency can help increase pork supply while lowering the environmental footprint. To continue making incremental improvements in sow reproduction there must be a continued effort to maximize number of pigs per litter, optimize piglet birth weight, maximize litters per sow per year, and improve longevity and lifetime productivity. Sows enter the breeding herd as gilts (a female pig that has yet to produce a litter) producing their first litter at approximately one year of age.

Onset of Puberty

In a sow’s reproductive lifetime, several critical events precede inclusion into the sow herd beginning with puberty onset. Puberty is defined as the biological changes associated with sexual maturation and development enabling a female of being capable of reproduction. On average it takes five to eight months for a gilt to reach puberty (Graves et al., 2020), and in commercial settings gilts are typically six to eight months of age when entered into the breeding pool with the objective of producing her first litter within a year of age. Vulva width has been correlated to a gilt’s ability to achieve behavioral estrus by 200 days of age and is an indicator of sows lifetime productivity (Graves et al., 2020; Romoser et al., 2020a; Romoser et al., 2020b).

Achievement of puberty is controlled by steroid hormones and neurotransmitters, which induce the secretion of gonadotrophin releasing hormone (GnRH) (Delemarre-Van De Waal, 2002).

Kisspeptin, a neuropeptide, has been shown to initiate GnRH release by acting directly on the hypothalamic neurons (Seminara and Crowley, 2008). Disruption in kisspeptin signaling has 5 been shown to delay puberty onset due to its regulation of GnRH (De Roux et al., 2003). The secretion of GnRH occurs in a pulsatile fashion and involves multiple feedback mechanisms of inducing and suppressive signals (Navarro et al., 2009; Livadas and Chrousos, 2016). Once experiencing their first estrous cycle, the gilt is considered sexually mature, and being a polyestrous animal, the reproductive cycle repeats approximately every 21 days.

Estrous Cycle

The estrous cycle consists of four phases including estrus (days 0-2), metestrus (days 2-5) diestrus (days 5-16) and proestrus (days 17-21) (Senger, 1997). Of these phases, proestrus and estrus represent the follicular phase when follicle development and ovulation occur, while metestrus and diestrus represent the luteal controlled phases of the cycle. Collectively, phases of the estrous cycle are regulated through positive and/or negative feedback regulation of numerous hormones. The follicular phase (estrus and proestrus) is characterized by the period of time when follicles in the ovary are recruited for growth and ovulation. Follicle recruitment is dependent on

GnRH induced release of luteinizing hormone and follicle stimulating hormone (Foxcroft and

Hunter, 1985). During this phase circulating 17β-estradiol (E2) is high and progesterone (P4) is low, and antral follicles progressing towards ovulation are the predominant structures on the ovary (Senger, 1997). The luteal phase (metestrus and diestrus) represents the period of time when corpora lutea (CL) are the predominant ovarian structures. The CL produce copious quantities of P4 and E2 levels begin falling after ovulation, reaching a nadir during the diestrus phase (Senger, 1997).

During estrus, E2 production from ovulatory follicles achieves levels capable of inducing a positive feedback mechanism resulting in a surge of luteinizing hormone (LH) that ruptures the follicle, releasing the oocyte into the uterine horn (Senger, 1997). Also during estrus, gilts are 6 sexually receptive, which includes behavioral responses such as ‘standing’ and vocalization in response to boar stimulus (Hemsworth, 1985; Johnson, 2007). On average, estrus lasts from 24-

72 hours and oocytes are ovulated at approximately 55-60% of the way through the estrus phase

(Soede et al., 2000). During sexual receptivity, when E2 levels are maximal, is the ideal time for sperm to be present within the reproductive tract to allow for the greatest probability of fertilizing an oocyte. Typically, in commercial settings, a female is mated twice during behavioral estrus, utilizing artificial insemination (AI). Following fertilization in the ampullary- isthmic junction of the oviduct, the embryos begin development and eventual migration into the uterine horns. In order to establish and maintain pregnancy a sow must maintain the viability of the CL, have embryos in both horns, and have adequate circulating P4 (Senger, 1997; Spencer and Bazer, 2002). If pregnancy is established, a sow will undergo gestation for approximately

115 days from the onset of estrus (Senger, 1997).

Gilt Development

Following parturition, sows in most commercial U.S. herds will nurse their piglets until weaned at 18-24 days of age. At this time, piglets are transferred from the sow farm and raised at a secondary location to minimize risk of disease spread. Most weaned pigs from commercial farms are reared until reaching market weight and then harvested. Gilts produced from genetic nucleus and multiplier farms are reared in separate production facilities (i.e., a gilt development unit) until transferred back into sow farms to become part of the breeding herd. A sow’s profitability is dependent on her reproductive success and longevity (Stalder et al., 2003). The earlier a gilt can obtain puberty, the more productive she tends to be through two parities

(Romoser et al., 2020a).

Sow Mortality in the U.S. Swine Industry

Sow mortality continues to increase in the U.S., and is a major concern for animal welfare, worker morale, and is a substantial economic loss for producers. As mentioned, a sow only becomes profitable after her third parity at which point her production costs have been offset by the cumulative value of her weaned pigs (Stalder et al., 2008). A barrier to increasing the average parity sows achieve during their lifetime is premature sow mortality, which has been increasing in the U.S. swine industry during the past decade. Sow mortality has increased from

7.9% to 10.7% between the years 2014 and 2017 (PigChamp; Supakorn et al., 2019), and as of

2018 was 12.7% (Ross, 2019) within U.S. herds. An average of 5% is generally considered an acceptable mortality rate in the U.S. (Stalder et al., 2008). Removals from sow herds encompass an assortment of reasons, including reproductive failures, lameness or locomotive problems, and natural death, making up 35-40%, 11-20% and 3-8% of removals, respectively (Supakorn et al.,

2019). In recent years pelvic organ prolapse (POP), including vaginal, uterine and rectal prolapse, has been observed to cause approximately 21% of all sow mortality, with lameness being an additional major cause of death (Ross, 2019). It is estimated 40-50% of sow deaths occur within the periparturient period, corresponding to the last 10% of gestation and the first few days post-partum (Kuratomi and Anil, 2007; Supakorn et al., 2019).

Reproductive Tract Dysfunction Overview

Dysfunction of the female reproductive tract can be caused by multiple reasons and can be further affected by endocrine, paracrine, and autocrine signals (Roy and Matzuk, 2011).

Additionally, developmental abnormalities, environmental factors, hormonal irregularities, or genetic disorders can all influence reproductive tract function (Roy and Matzuk, 2011).

Clinically, disorders of the reproductive system can manifest as pain and discomfort, abnormal 8 menstruation in women, or infertility, which all are symptoms of a variety of diseases further described below.

Reproductive Disease in Humans

Pelvic inflammatory disease (PID) is estimated to affect one million women in the U.S.

(McCormack, 1994). While recognized by the inflammatory response, PID is associated with microbial infection of the upper genital tract caused by pathogenic microorganisms. The upper genital tract consists of the uterus, fallopian tubes and ovaries. This infection can be asymptomatic or have atypical presentations outside of clinical signs, making it difficult to diagnose. These microorganisms can be from sexually transmitted bacterial infections (STI) and have been linked to chlamydia and gonorrhea (McCormack, 1994; Gradison, 2012). Women suffering from PID have a 20% chance of infertility caused by tubal infertility where a blockage will not allow the sperm to meet the egg (McCormack, 1994), and an 18% chance of experiencing chronic pelvic pain (Gradison, 2012). Diagnosis of PID is typically done via clinical evaluation, however it can be difficult to identify, and antibiotic treatment is usually initiated without confirming the exact cause of infection (McCormack, 1994). Many reproductive diseases can manifest and present similar symptoms as others, making it difficult to correctly diagnose or even determine an appropriate route of treatment. Several studies have connected PID with bacterial vaginosis (BV) due to the association of vaginal microbiota changes (Eschenbach et al., 1975; Spiegel, 1991) although there is still some controversy around the comparison (Ness et al., 2005).

It is well known that the human vaginal microbiota is dominated by lactobacilli, and BV is characterized by the disequilibrium in the vaginal microflora. This imbalance can be caused by changes in the most common microbes present and/or an increase in the number of species 9 present (Russo et al., 2019). Similar to PID, BV is associated with upper reproductive tract infections, but has also been connected to lower reproductive tract (external genitalia, the vagina, and lower part of the cervix) infections. There is a common belief that infection of the lower reproductive tract from BV travels upward and eventually manifests as PID (Spiegel, 1991; Ness et al., 2005). The estimated occurrence of BV varies from 5-70% of women, but is considered to be the most common vaginal infection, with the U.S. having a 30% prevalence in reproductive aged women (Kairys and Garg, 2021). Comparable to PID, women suffering from BV have an increased risk of further infections and STIs. Complications with pregnancy and preterm delivery are also increased with BV, as well as spontaneous abortions, and could also increase the probability of infertility (Han et al., 2019). There is a 30% chance BV resolves on its own, however, typically it is treated with the use of antibiotics, probiotics or bacteriophage therapy

(Javed et al., 2019). Unfortunately, reoccurrence is very high for women with BV, and use of antibiotics as treatment leads to a concern for increased risk of antibiotic resistance (Javed et al.,

2019). Vaginal health can influence immune function and BV can activate enzymes that reduce leukocytes, which fight infection, and increase endotoxins that stimulate cytokine and prostaglandin production (Greenbaum et al., 2019). While uterine dysfunction is more commonly studied in humans, livestock species are also affected, although not nearly as much is known.

Reproductive Disease in Cattle

Uterine microbial disease affects nearly half of post-partum dairy cattle, which can cause infertility, and disrupt uterine and ovarian function (Sheldon et al., 2009). After parturition, the uterus has greater potential to be contaminated with a range of bacteria, and depending on the delicate balance between host and pathological bacteria, can result in the development of a clinical disease (Sheldon et al., 2009). Endometritis is a common uterine disease, and can be 10 diagnosed in women, cattle and swine (Barlund et al., 2008; Leyland et al., 2010; Lorenzen et al.,

2015; Sheldon et al., 2019). There is limited knowledge on reproductive dysfunction treatment and even less relating specifically to swine. One common reproductive issue of growing concern and little knowledge across humans and livestock species is POP.

Physiology of Pelvic Organ Prolapse

Pelvic organ prolapse is characterized by the descent of female pelvic organs resulting in protrusion of the vagina, uterus or rectum out of the body (Jelovsek et al., 2007). The pathophysiology behind POP is not well understood in humans or livestock species. The POP rate in humans is extremely variable and is thought to be 3-6% or potentially even higher (Barber and Maher, 2013). It is estimated that approximately 1% of cattle experience POP (Couri et al.,

2012; Carluccio et al., 2020). For the swine industry, it has been observed that the average annualized POP rate is approximately 2.7% in the U.S., and in some cases can reach 10.3%

(Ross, 2019). Overall, POP incidence increased by 4-fold in 2016 compared to 2008 in the U.S. sow herd (Supakorn et al., 2019). Data also suggests there is a seasonal effect on POP occurrence in U.S. commercial sow farms with greatest POP incidence occurring during the winter months. The type of POP experienced can be categorized into uterine, rectal, or vaginal which were discovered to make up 30.8-45.6%, 32.7-41% and 21.2-28.2% of overall POP incidence, respectively, and observations of combinations of these specific types of POP have also been noted (Supakorn et al., 2014). The current strategy to repair POP is surgery, which has been successful in cattle (Miesner and Anderson, 2008), however rate of success varies in humans (Jelovsek et al., 2007). The most common practice in the swine industry to manage POP is by culling and/or euthanizing the sow. The economic loss from POP alone is significant and has likely increased in the last few years (Supakorn et al., 2014). 11

For perspective, average POP mortality alone equates to a 5,000 head sow farm losing

135 sows annually, which is a substantial economic loss. There is limited information available on causes or prevention strategies for POP in any species, making the development and deploying of strategies to mitigate the underlying cause of POP in the swine herd difficult. As mentioned previously, several reproductive tract disorders and overall reproductive health are linked to dysbiosis or alterations in the vaginal microbiota (Sheldon et al., 2009; Javed et al.,

2019; McCormack, 1994), suggesting the microbiota of sows as a potential link to POP.

Vaginal Microbiota Overview

Microorganisms are harbored in the surfaces and cavities of a host and participate in mutualistic relationships (Rosenberg and Zilber-Rosenberg, 2011). Vaginal microbiota in relationship to reproductive health is an emerging field of research and area of interest, as evidenced by exponential publication growth during the last decade, in large part through the use of high throughput DNA sequencing of 16S rRNA genes. The microbiota has an influence on development, physiology, immunity, and nutrition; but surprisingly the functionality and mechanism by which it protects the reproductive tract is poorly understood (Ma et al., 2012;

Ravel and Brotman, 2016). The vaginal microbiota is thought to play a role in both innate and adaptive immune responses in the female reproductive tract, and is a specific growing area of interest (Ravel and Brotman, 2016). Current research has been focused on evaluating and characterizing human vaginal microbiota while knowledge of livestock microbial communities is currently limited. The vaginal microbiota is extremely dynamic and is influenced by a multitude of factors such as sexual development, coitus, personal hygiene, menses, and hormones (Gajer et al., 2012; DiGiulio et al., 2015; Lewis et al., 2017). Certain microbes may have different functions, and can vary within different host species as well as body sites (Chu et al., 2017). 12

Differences have also been observed among ethnicities (Jacques Ravel et al., 2011), leading to the presumption that differences exist across livestock species and between genetic lines within a species. The relationship between vaginal microbiota and health status is dynamic and complex, and requires further research to have a more comprehensive understanding.

Vaginal Microbiota of Humans

The relationship between vaginal health and/or reproduction with the vaginal microbiota has been primarily investigated in women. As example, it has been demonstrated that the microbiota in pregnant women increases in stability, suggesting an important functional role in lowering uterine infection risk of the mother and prevention of preterm labor (Greenbaum et al.,

2019). Conversely, dysbiosis in the vaginal microbiota has been associated with preterm birth in humans leading to the supposition that an important link between the microbiota and maintenance of pregnancy exists (DiGiulio et al., 2015; Greenbaum et al., 2019). There is controversy over the idea of a ‘core’ healthy microbiota, because of many influential contributing factors, but there is strong evidence that postpartum alterations occur (DiGiulio et al., 2015). Generally, vaginal microbial communities are divided into two groups; one associated with healthy individuals, and the second associated with dysbiosis. It is well-defined that the healthy human vaginal microbiota is dominated by Lactobacilli, which produce lactic acid, subsequently aiding in lowering the vaginal pH as a protective measurement. Therefore

Lactobacilli are often utilized in probiotic treatments for dysbiosis (Senok et al., 2009). Changes in the vaginal microbiota have been documented in relation to several reproductive disorders.

A number of urogenital diseases, such as yeast infections, urinary tract infections, STIs,

PID, and BV have shown to be associated with alterations in vaginal microbiota (McCormack,

1994; DiGiulio et al., 2015; Greenbaum et al., 2019). The Nugent scoring system (Nugent et al., 13

1991) is a widely used tool to diagnose BV and is based on gram-stained vaginal smears. While

Lactobacilli are still observed in BV samples, there are increases in Aerococcus, Prevotella,

Prevotellaceae, in addition to several other microbes (Jacques Ravel et al., 2011). It is believed that Gardnerella creates a biofilm during the onset of a BV infection and allows for other opportunistic bacteria to populate, although further work is required to validate this postulate

(Verstraelen and Swidsinski, 2019). The vaginal pH of a woman suffering from BV is normally higher (> 4.5), which could allow for undesirable microbes to inhabit the lower reproductive tract

(Spiegel, 1991). Similar changes in the microbiota of BV patients have also been observed in women suffering from PID and endometritis (Lewis et al., 2017).

In both BV and PID patients it is common to screen for abnormal increases in Chlamydia trachomatis and Neisseria gonorrhoeae, for which antibiotics are typically an effective treatment

(Gradison, 2012). Unfortunately, there are not many other treatment options outside of antibiotics for these types of disease, and as previously mentioned, reoccurrence is high due to this lack of efficacious treatment strategies. Further insight into the microbiota and function of specific microbial species is needed to aid in developing a better understanding of vaginal microbiota with the potential to improve female reproductive health. Interactions between body sites and the microbiota should be considered and the gut-vagina axis may have an important and opportunistic role in understanding and improving women’s health, which is the basis for the concept of probiotic supplementation as a preventative measure or treatment option. Although human vaginal microbiota research is much further along than livestock, it is still unclear through which mechanisms and effects a dysbiosis can alter reproductive health (Donati et al.,

2010; Lewis et al., 2017).

Vaginal Microbiota of Cattle

In dairy cattle populations, within a week of parturition, approximately 40% of the herd is likely to have some form of uterine disease (Sheldon et al., 2009). While the rumen and microbes responsible for maintaining normal function and health are the main focus of study, the cattle vaginal microbiota is just beginning to be defined. Unlike the human vaginal microbiota, cattle have a more diverse vaginal microbial community (Rodrigues et al., 2015). Cattle suffer from several reproductive tract diseases that have been linked to microbiota dysbiosis (Rodrigues et al., 2015). Until the last decade cows suffering from uterine diseases were only identified if they were exhibiting clinical signs or relied heavily on culture-dependent assays. It is well known that many microbes cannot be cultured, therefore a full characterization of what organisms make up the vaginal microbiota is not feasible with cultivation-dependent approaches only and additional cultivation-independent approaches such as 16S rRNA sequencing are required. One study observed operational taxonomic units (OTUs) corresponding to 15 taxa, when evaluating the vaginal microbiota of cows (Rodrigues et al., 2015). Bacteroides made up 28% of the control, healthy, microbiota, along with Enterobacteriaceae and Victivallis representing the other most abundant microbes. Significant differences in the microbial communities between healthy cows and cows suffering from reproductive disorders have been observed (Rodrigues et al., 2015).

Bacteria of the genera Bacteroides, Mycoplasma, Histophilus, Fusobacterium, and Prevotella, as well as the species Escherichia coli and Streptococcus spp., among others are considered pathogenic and associated with genital tract diseases in cattle (Corbeil, 1996; Pfützner and

Sachse, 1996; LeBlanc, 2008). Changes in the microbiota have also been observed in cows suffering from clinical endometritis and metritis (Galvão et al., 2019).

Studies have observed increases in Trueperella pyogenes, Escherichia coli,

Fusobacterium necrophorum, Prevotella melaninogenica, and Bacteroides spp. found in samples 15 from cows with endometritis (Sheldon et al., 2009; Galvão et al., 2019). In healthy cows

Streptococcus spp., Staphylococcus spp., and Bacillus spp. were observed to be in abundance

(Galvão et al., 2019). Alternatively, cows that developed metritis, had greater abundance of

Bacteroidetes and Fusobacteria and lower abundance of Proteobacteria and Tenericutes

(Sheldon et al., 2009; Galvão et al., 2019). Metritis in cows is often characterized by a loss of heterogeneity and decreased species richness within the microbiota (Galvão et al., 2019). While the vaginal microbiota is being defined in cattle and potential pathogenic microbes have been identified, there is still a lack of knowledge and understanding of swine vaginal microbiota.

Vaginal Microbiota of Swine

There are limited studies evaluating vaginal microbiota in pigs and even less looking at the association with reproductive disorders. Prior studies have evaluated the swine vaginal microbiota by cultivation, which limits the microbes able to be detected in the sample (Bara et al., 1993; Maes et al., 1999). The vaginal microbiota of pre- and post-pubertal minipigs from has begun to be characterized by a Denmark group (Lorenzen et al., 2015), but requires further work to validate findings. The minipig’s core vaginal microbiota was established by five phyla, consisting of Firmicutes, Proteobacteria, Tenericutes, Actinobacteria, and Bacteroidetes

(Lorenzen et al., 2015). The same microbes were observed in sexually mature minipigs, however at different abundances than the prepubertal group (Lorenzen et al., 2015). It is important to note that the mature group was hormonally synchronized which may or may not had an effect on the vaginal microbiota. Interestingly, Firmicutes was still observed to be the most abundant phylum, and there was a low (3.5-6%) abundance of the Lactobacillaceae family (Lorenzen et al., 2015).

This characterization of minipig vaginal microbiota would not be comparable to U.S. late gestation sows presumably due to differences in genetic lines, and pregnancy status. A study 16 conducted in China evaluated the vaginal microbiota in healthy sows and sows suffering from endometritis, but had a small sample size of only four animals in each treatment group (Wang et al., 2017). However, even with the small sample size, there were observable differences in the microbiota. In agreement with other studies conducted, Firmicutes, Proteobacteria, and

Bacteroidetes were observed in the vaginal microbiota in both groups (Lorenzen et al., 2015,

Wang et al., 2017).

Recent studies are starting to evaluate the vaginal microbiota in U.S. commercial gilts.

Five common phyla were observed including Fusobacteria, Proteobacteria, Firmicutes,

Bacteroidetes and Tenericutes (Sanglard et al., 2020a). However, Sanglard et al. (2020a) focused largely on heritability traits and genetics in response to PRRS vaccination, and not on reproductive function or dysbiosis. Another approach evaluated the vaginal microbiota in relation to vaccination and reproductive performance, but the sampling was conducted on non- pregnant gilts (Sanglard et al., 2020b). Vaginal microbiota can shift in relation to life stage and during pregnancy, so it may be presumed that late gestation sows experiencing POP will have different vaginal microbial communities than non-pregnant gilts. To date there is no literature evaluating late gestation sow vaginal microbiota, particularly as it relates to reproductive dysfunction.

Role of the Vaginal Microbiota in the Immune System

It is feasible that even small changes in the microbiota can cause major shifts in a pathogenic interaction. Changes in pH have the ability to alter the microbiota, making the environment more favorable to certain ‘beneficial’ or ‘harmful’ bacteria. It is clear that changes in the vaginal microbiota are multifactorial, and complex within itself. Further research is needed to better understand the interactions between vaginal microbiota and reproductive health. 17

Currently, there is little knowledge about the microbiota and its possible association with POP in any species. The immune system plays a role in controlling the diversity of the microbiota through both innate and adaptive immune responses (Mirmonsef et al., 2011). Identifying types of bacteria present in the reproductive tract is believed to be a critical step in advancing the understanding of how the immune system is affected by the microbiota or how it affects the immune system (Mirmonsef et al., 2011). Host microbiota interaction within the immune system can ultimately play a role in reproductive health.

Immune and Inflammatory Response

The inflammatory response is highly coordinated and mediates resident tissue cells as well as recruited inflammatory cells (Spector and Willoughby, 1963). Common signs of inflammation are heat, pain, redness, swelling and impaired function. These symptoms are the localized effects of chemokines, cytokines, and other mediators of inflammation. Inflammation is heavily involved in reproductive success and plays a role in ovulation, menstruation, implantation and parturition (Romero et al., 2008). During acute inflammatory responses, restoration of tissue to homeostasis must occur efficiently to minimize injury and infection.

However, if this response becomes uncontrolled, it may lead to chronic inflammation and disease such as Crohn’s, endometritis, endometriosis, and others (Barlund et al., 2008; Sartor, 2008;

Leyland et al., 2010; Lorenzen et al., 2015; Sheldon et al., 2019). Infections, tissue injury, and other pathogenic factors can induce inflammation, however, this response depends on the specific nature of the stimulus and body location.

The female lower genital tract relies on the innate and adaptive immune systems to fight pathogens. Interactions between vaginal microbiota and the host epithelial cells can be responsible for reproductive health (Wira et al., 2005a). The innate immune system consists of 18 several layers of protection, with the first line of defense being the epithelial cells lining the reproductive tract acting as a physical barrier to protect against pathogens and particles (Wira et al., 2005b). In addition to providing a physical barrier, epithelial cells also produce mucus in attempt to defend against infectious agents (Franklin and Kutteh, 1999; Wira et al., 2005b).

There is an elaborate equilibrium between the immune system, microbiota, and epithelial cells that is responsible for female reproductive health. This equilibrium is quite delicate, and because of this pregnant animals and women are often more susceptible to microbial products (Romero et al., 2008). Vaginal fluid and mucus produce antimicrobial proteins as a defense against pathogens (Valore et al., 2002). Additionally, epithelial cells express receptors needed to elicit an immune response, furthering their protective effect against pathogens (Mirmonsef et al., 2011).

Microbial metabolites and cytokines can be primary inflammatory stimuli, and chronic diseases typically involve common mediators and pathways of inflammation (Chen et al., 2018).

The inflammatory response can be activated in a variety of ways by several different biological pathways. The nuclear factor kappa B (NF-ĸB) pathway regulates pro-inflammatory cytokine production and cell recruitment, which contribute to the inflammatory response (Wira et al.,

2005a). The mitogen-activated protein kinase (MAPK) pathway has a cellular response to inflammatory cytokines such as tumor necrosis factor alpha (TNFα) (Sabio and Davis, 2014).

Epithelial cells produce TNFα as well as other interleukins to recruit immune cells and facilitate an effective immune response, and plays a prominent role in the inflammatory response (Wira et al., 2005b). Inflammatory stimuli and stress typically activate cJun NH2-terminal kinases (JNK) and p38 , which in turn respond by increasing the expression of inflammatory cytokines (Sabio and Davis, 2014).

Pathogen-associated molecular patterns are microbial structures that can activate an inflammatory response (Gudkov and Komarova, 2016). Receptors involved in immune response can include Toll-like receptors (TLR), and major histocompatibility complex (MHC) molecules, which help recognize and eliminate pathogens. Specifically, TLRs play a critical role in the recognition and elimination of pathogenic microorganisms, by TLRs recognition of the bacterial ligands (Rakoff-Nahoum et al., 2004). Activation of these factors quickly result in cytokines, chemokines, and antimicrobial product production in efforts to mitigate the immune response.

Microbial products such as acetic, butyric, and propionic acids, have also been shown to influence an immune response, but have been studied primarily in the gut (Mirmonsef et al.,

2011). If bacterial lipopolysaccharide (LPS) penetrates the protective barriers it can lead to immune activation (Schromm et al., 2021) and is known to activate mitochondria-associated protein ligases (MAPL) and cytokine production (Solomon et al., 1998). This stimulates TLR4 and induces the release of LPS binding protein (LBP), which is critical for a successful immune response (Lu et al., 2008). As part of the immune response LPS is known to activate monocytes production and/or allocation (Takashiba et al., 1999), and is conventionally used to study inflammation due to its effects generated through TLR4 (Tucureanu et al., 2018). Parturition is often considered a localized inflammatory response and is accompanied by an increase in acute phase reactants and changes in specific leukocytes, such as monocytes (Romero et al., 2008). An exaggerated inflammation response has also been linked to preterm birth (Romero et al., 2008), therefore a delicate balance is required for inflammation around the time of parturition.

Steroid Hormone Changes During Late Gestation

Steroid hormones play a crucial role in women’s health, and there are many changes in hormones during pregnancy and leading up to parturition. In sows, P4 is produced by the CL and is necessary for pregnancy establishment and maintenance for the entire duration of pregnancy

(Senger, 1997). It has been observed that P4 within circulation begins a slight decline approximately five days pre-partum with a rapid decline just before farrowing (Molokwu and

Wagner, 1973; Baldwin and Stabenfeldt, 1975). Levels do not decrease further until after farrowing, but plateau to a low concentration during lactation (Molokwu and Wagner, 1973;

Baldwin and Stabenfeldt, 1975). Estrogen levels are inversely related to those of P4 as a steady increase in E2 has been observed approximately 6 to 11 days prior to farrowing, which is followed by a dramatic drop in concentration during farrowing, then remains low throughout lactation (Molokwu and Wagner, 1973; Baldwin and Stabenfeldt, 1975).

Steroidogenesis

The production of steroid hormones is necessary for endocrine regulation of metabolism, inflammation, immune function, and reproductive function in livestock species. All active steroid hormones begin as cholesterol and derivatives are produced through the activity of various steroidogenic enzymes which can be classified into hydroxysteroid dehydrogenases or cytochrome P450 enzymes (Bremer and Miller, 2014; Moon et al., 2016). During this process, cholesterol is converted to various steroid hormones including glucocorticoids, mineralocorticoids and sex steroids such as progestins, androgens, and estrogens (Bremer and

Miller, 2014). Steroid hormones can be synthesized in the adrenal cortex, gonads (testes and ovaries), brain, placenta, and adipose tissues (Falkenstein et al., 2000). Ovarian steroidogenesis results in production of E2 and P4 which regulate the estrous cycle. This process begins with 21 transport of cholesterol into the mitochondria by steroid acute regulatory protein (StAR) followed by conversion of cholesterol to pregnenolone by CYP11A1, a cytochrome P450 enzyme, which is considered to be a rate limiting enzyme in steroid hormone synthesis (Bremer and Miller, 2014; Moon et al., 2016). An additional enzyme of interest is CYP19A1, or aromatase, which plays a critical role in estrogen biosynthesis. Biosynthesis of P4 by the CL also begins with transport of cholesterol into the mitochondria by StAR and conversion of cholesterol into pregnenolone by CYP11A1 (Christenson and Devoto, 2003). This is followed by conversion of pregnenolone to P4 by 3BHSD (Christenson and Devoto, 2003). During late gestation the brain is responsive to changes in the ovarian and uterine hormone secretion (Gilbert et al., 2001).

Steroid hormone changes are very dynamic during the peripartum period and serve as regulators of multiple processes, including regulation of the immune response.

Steroid Hormone Effects on the Immune Response

Immune responsiveness has been observed to differ between males and females as females have been shown to resist bacterial and viral infections more successfully than males, and have a higher survival rate, presumably because of an increased immune capability (Ahmed et al., 1985). Testosterone, which is detected in much greater concentrations in males, has been observed to decrease immune responses (Grossman, 1984). The endocrine changes during pregnancy are also believed to aid in immunosuppression in order to prevent immunological rejection of the fetus (Ahmed et al., 1985). However, there are conflicting results on the pathophysiology of steroid hormones regulation of the immune response. For example, E2 and

P4 have been shown to both suppress and enhance the immune response depending on the specific physiological states (Wyle and Kent, 1977; Mathur et al., 1978; Clemens et al., 1979;

Lawrence et al., 1980; Ahmed et al., 1985). The effect of steroid hormones on an immune 22 response varies depending on the target tissues and could be a contributing factor to these conflicting results. Due to the altered microenvironment caused by sex hormones, migratory changes and localization have been shown to effect lymphocytes, and changed in hormone levels have been observed to suppress T cells (Ahmed et al., 1985). Specifically, testosterone may decease T cell population through apoptosis (Mcmurray et al., 2001), and T helper lymphocytes main function is to produce cytokines during an inflammatory response (Bouman et al., 2005). It has been suggested that E2, and possibly P4, have the ability to decrease monocytes (Ben‐Hur et al., 1995). However, these hormones have also been observed to induce monocyte release from the bone marrow (Bain and England, 1975). Monocytes play a critical role in the immune response, and since steroid hormones have an effect on monocyte levels, could influence the immune response, furthering the differences observed between genders and reproductive status.

Steroid Hormones and their Potential Influence on the Microbiota

The vaginal microbiota are influenced by the hormonal changes throughout the reproductive stages in a woman’s life, including puberty, menstruation, pregnancy, and menopause and may also influence fertility (Farage and Maibach, 2006). It is now believed that there is a range of microbial profiles that are able to produce a stable vaginal microbiota and maintain good reproductive health in women (Farage et al., 2010). The vaginal microflora can be greatly influenced by estrogens (Spiegel, 1991), and over the course of pregnancy, estrogen steadily increases, mainly in the form of estriol (Pagana and Pagana, 1998). Elevated levels for

E2 have been shown to stimulate the proliferation of vaginal epithelial cells, subsequently increasing Lactobacilli within the microflora (Farage and Maibach, 2006). This is further supported by the observation that during reproductive stages when estrogens are highest there is also greater levels of Lactobacilli present in humans (Larsen and Monif, 2001). The lactic acid 23 produced by these microbes lowers the vaginal pH, creating a more favorable environment for colonization of bacteria. During menses, in humans, the vaginal microbiome is disrupted, providing a vulnerable window, and allowing the potential for other organisms to gain dominance (Farage et al., 2010). Recent work has also begun to evaluate bacteria that are capable of estrogen degradation, which could have an effect on women’s reproductive health (Flores et al., 2012; Yu et al., 2013; Fuhrman et al., 2014; Vieira et al., 2017; Y.-L. Chen et al., 2018; Li et al., 2018). Further research evaluating the microbiota during pregnancy and potential hormonal influences is needed, especially in livestock species.

Summary

In order to continue improving the sustainability of swine production and the efficiency of reproduction, improving reproductive performance while decreasing sow mortality must be a research focus. A substantial barrier to pork production efficiency and sow longevity is sow mortality due to POP. Problematically, there is no known biological basis for the root cause of the increased POP occurrence in the U.S. sow herd. Therefore, developing mitigation strategies can only come after understanding the fundamental and physiological basis of POP.

The overarching objectives of this thesis were to: 1) develop a better understanding of the vaginal microbiota in sows and how it may be associated with POP, and 2) identify potential biomarkers that could serve as indicators to assess a sow’s risk for POP. Accomplishing these objectives would create a strengthened understanding of POP and enable a meaningful step towards the long-term goal of developing POP mitigation strategies for the U.S. sow herd. To do this, a scoring system was developed that effectively categorized sows during late gestation into populations that had a high or low relative risk of POP. This system was then utilized to explore 24 differences in the vaginal microbiota in relation to POP risk (Chapter 3 and Chapter 4) in addition to identification of circulating steroid hormone, markers of immune system activation and other potential biomarkers of POP risk (Chapter 3 and Chapter 5).

References

Ahmed, S. A., W. J. Penhale, and N. Talal. 1985. Sex hormones, immune responses, and autoimmune diseases. Mechanisms of sex hormone action. Am. J. Pathol. 121:531–551.

Bain, B. J., and J. M. England. 1975. Variations in Leucocyte Count During Menstrual Cycle. Br. Med. J. 2:473–475. doi:10.1136/bmj.2.5969.473.

Baldwin, D. M., and G. H. Stabenfeldt. 1975. Endocrine Changes in the Pig During Late Pregnancy, Parturition and Lactation 1.

Bara, M., M. McGowan, D. O’Boyle, and R. Cameron. 1993. A study of the microbial flora of the anterior vagina of normal sows during different stages of the reproductive cycle. Aust. Vet. J. 70:256–259. doi:10.1111/j.1751-0813.1993.tb08043.x.

Barber, M. D., and C. Maher. 2013. Epidemiology and outcome assessment of pelvic organ prolapse. Int. Urogynecol. J. Pelvic Floor Dysfunct. 24:1783–1790. doi:10.1007/s00192- 013-2169-9.

Barlund, C. S., T. D. Carruthers, C. L. Waldner, and C. W. Palmer. 2008. A comparison of diagnostic techniques for postpartum endometritis in dairy cattle. Theriogenology. 69:714–723. doi:10.1016/j.theriogenology.2007.12.005.

Ben-Hur, H., G. Mor, V. Insler, I. Blickstein, Y. Amir-Zaltsman, A. Sharp, A. Globerson, and F. Kohen. 1995. Menopause Is Associated With a Significant Increase in Blood Monocyte Number and a Relative Decrease in the Expression of Estrogen Receptors in Human Peripheral Monocytes. Am. J. Reprod. Immunol. 34:363–369. doi:10.1111/j.1600- 0897.1995.tb00965.x.

Bouman, A., M. Jan Heineman, and M. M. Faas. 2005. Sex hormones and the immune response in humans. Hum. Reprod. Update. 11:411–423. doi:10.1093/humupd/dmi008.

Carluccio, A., I. de Amicis, M. Probo, B. Giangaspero, and M. C. Veronesi. 2020. Prevalence, survival and subsequent fertility of dairy and beef cows with uterine prolapse. Acta Vet. Hung. 68:91–94. doi:10.1556/004.2020.00017.

Chen, L., H. Deng, H. Cui, J. Fang, Z. Zuo, J. Deng, Y. Li, X. Wang, and L. Zhao. 2018. Inflammatory responses and inflammation-associated diseases in organs. Oncotarget. 9:7204–7218. doi:10.18632/oncotarget.23208. 25

Chen, Y.-L., H.-Y. Fu, T.-H. Lee, C.-J. Shih, L. Huang, Y.-S. Wang, W. Ismail, and Y.-R. Chiang. 2018. Estrogen Degraders and Estrogen Degradation Pathway Identified in an Activated Sludge. Appl. Environ. Microbiol. 84:e00001-18. doi:10.1128/AEM.00001-18.

Christenson, L. K., and L. Devoto. 2003. Cholesterol transport and steroidogenesis by the corpus luteum. Reprod. Biol. Endocrinol. 1. doi:10.1186/1477-7827-1-90.

Chu, D. M., J. Ma, A. L. Prince, K. M. Antony, M. D. Seferovic, and K. M. Aagaard. 2017. Maturation of the infant microbiome community structure and function across multiple body sites and in relation to mode of delivery. Nat. Med. 23:314–326. doi:10.1038/nm.4272.

Clemens, L. E., P. K. Siiteri, and D. P. Stites. 1979. Mechanism of Immunosuppression of Progesterone on Maternal Lymphocyte Activation during Pregnancy. J. Immunol. 122.

Corbeil, L. B. 1996. Histophilus somni host–parasite relationships. Anim. Heal. Res. Rev. 8:151–160. doi:10.1017/S1466252307001417.

Couri, B. M., A. T. Lenis, A. Borazjani, M. F. R. Paraiso, and M. S. Damaser. 2012. Animal models of female pelvic organ prolapse: Lessons learned. Expert Rev. Obstet. Gynecol. 7:249–260. doi:10.1586/eog.12.24.

De Roux, N., E. Genin, J. C. Carel, F. Matsuda, J. L. Chaussain, and E. Milgrom. 2003. Hypogonadotropic hypogonadism due to loss of function of the KiSS1-derived peptide receptor GPR54. Proc. Natl. Acad. Sci. U. S. A. 100:10972–10976. doi:10.1073/pnas.1834399100.

Delemarre-Van De Waal, H. A. 2002. Regulation of puberty. Best Pract. Res. Clin. Endocrinol. Metab. 16:1–12. doi:10.1053/beem.2001.0176.

DiGiulio, D. B., B. J. Callahan, P. J. McMurdie, E. K. Costello, D. J. Lyell, A. Robaczewska, C. L. Sun, D. S. A. Goltsman, R. J. Wong, G. Shawa, D. K. Stevenson, S. P. Holmes, and D. A. Relman. 2015. Temporal and spatial variation of the human microbiota during pregnancy. Proc. Natl. Acad. Sci. U. S. A. 112:11060–11065. doi:10.1073/pnas.1502875112.

Donati, L., A. Di Vico, M. Nucci, L. Quagliozzi, T. Spagnuolo, A. Labianca, M. Bracaglia, F. Ianniello, A. Caruso, and G. Paradisi. 2010. Vaginal microbial flora and outcome of pregnancy. Arch. Gynecol. Obstet. 281:589–600. doi:10.1007/s00404-009-1318-3.

Eschenbach, D. A., T. M. Buchanan, H. M. Pollock, P. S. Forsyth, E. R. Alexander, J. Lin, S. Wang, B. B. Wentworth, W. M. McCormack, and K. K. Holmes. 1975. Polymicrobial Etiology of Acute Pelvic Inflammatory Disease. N. Engl. J. Med. 293:166–171. doi:10.1056/nejm197507242930403.

Falkenstein E, Tillmann HC, Christ M, Feuring M, Wehling M. Multiple actions of steroid hormones--a focus on rapid, nongenomic effects. Pharmacol Rev. 2000 Dec;52(4):513- 56. PMID: 11121509. 26

Farage, M. A., K. W. Miller, and J. D. Sobel. 2010. Dynamics of the Vaginal Ecosystem— Hormonal Influences. Infect. Dis. Res. Treat. 3:IDRT.S3903. doi:10.4137/idrt.s3903.

Farage, M., and H. Maibach. 2006. Lifetime changes in the vulva and vagina. Arch. Gynecol. Obstet. 273:195–202. doi:10.1007/s00404-005-0079-x.

Flores, R., J. Shi, B. Fuhrman, X. Xu, T. D. Veenstra, M. H. Gail, P. Gajer, J. Ravel, and J. J. Goedert. 2012. Fecal microbial determinants of fecal and systemic estrogens and estrogen metabolites: A cross-sectional study. J. Transl. Med. 10. doi:10.1186/1479-5876-10-253.

Franklin, R. D., and W. H. Kutteh. 1999. Characterization of immunoglobulins and cytokines in human cervical mucus: Influence of exogenous and endogenous hormones. J. Reprod. Immunol. 42:93–106. doi:10.1016/S0165-0378(98)00086-2.

Foxcroft, G. R., and M. G. Hunter. 1985. Basic physiology of follicular maturationin the pig. Journals Reprod. Fertil. Ltd. 33:1–19. Fuhrman, B. J., H. S. Feigelson, R. Flores, M. H. Gail, X. Xu, J. Ravel, and J. J. Goedert. 2014. Associations of the Fecal Microbiome With Urinary Estrogens and Estrogen Metabolites in Postmenopausal Women. J. Clin. Endocrinol. Metab. 99:4632–4640. doi:10.1210/jc.2014-2222. Gajer, P., R. M. Brotman, G. Bai, J. Sakamoto, U. M. E. Schütte, X. Zhong, S. S. K. Koenig, L. Fu, Z. Ma, X. Zhou, Z. Abdo, L. J. Forney, and J. Ravel. 2012. Temporal dynamics of the human vaginal microbiota. Sci. Transl. Med. 4:132ra52-132ra52. doi:10.1126/scitranslmed.3003605.

Galvão, K. N., R. C. Bicalho, and S. J. Jeon. 2019. Symposium review: The uterine microbiome associated with the development of uterine disease in dairy cows. In: Journal of Dairy Science. Vol. 102. Elsevier Inc. p. 11786–11797.

Gilbert, C. L. 2001. Endocrine regulation of periparturient behaviour in pigs. Reprod. Suppl. 58:263–276.

Gradison, M. 2012. Pelvic Inflammatory Disease. Available from: www.aafp.org/afp.

Graves, K. L., B. R. Mordhorst, E. C. Wright, B. J. Hale, K. J. Stalder, A. F. Keating, and J. W. Ross. 2020. Identification of measures predictive of age of puberty onset in gilts. Transl. Anim. Sci. 4:285–292. doi:10.1093/tas/txz173.

Greenbaum, S., G. Greenbaum, J. Moran-Gilad, and A. Y. Weintruab. 2019. Ecological dynamics of the vaginal microbiome in relation to health and disease. Am. J. Obstet. Gynecol. 220:324–335. doi:10.1016/j.ajog.2018.11.1089.

Grossman, C. J. 1984. Regulation of the Immune System by Sex Steroids*. Endocr. Rev. 5:435– 455. doi:10.1210/edrv-5-3-435.

Gudkov, A. V., and E. A. Komarova. 2016. p53 and the carcinogenicity of chronic inflammation. Cold Spring Harb. Perspect. Med. 6. doi:10.1101/cshperspect.a026161. 27

Han, C., H. Li, L. Han, C. Wang, Y. Yan, W. Qi, A. Fan, Y. Wang, and F. Xue. 2019. Aerobic vaginitis in late pregnancy and outcomes of pregnancy. Eur. J. Clin. Microbiol. Infect. Dis. 38:233–239. doi:10.1007/s10096-018-3416-2.

Hemsworth, P. H. 1985. Sexual Behavior of Gilts. J. Anim. Sci. 61:75–85. doi:10.1093/ansci/61.Supplement_3.75.

Javed, A., F. Parvaiz, and S. Manzoor. 2019. Bacterial Vaginosis: An insight into the prevalence, alternative regimen treatments and it’s associated resistance patterns. Microb. Pathog. 127:21–30. doi:10.1016/j.micpath.2018.11.046.

Jelovsek, J. E., C. Maher, and M. D. Barber. 2007. Pelvic organ prolapse. Lancet. 369:1027– 1038. doi:10.1016/S0140-6736(07)60462-0.

Johnson, C. (Iowa P. I. C. 2007. Heat Detection Critical to Success | National Hog Farmer. Natl. Hog Farmer.

Kairys, N., and M. Garg. 2021. Bacterial Vaginosis. StatPearls Publishing.

Kuratomi, M., and S. S. Anil. 2007. Periparturient Risk Factors for Sow Mortality. Pig Site.

Larsen, B., and G. R. G. Monif. 2001. Understanding the bacterial flora of the female genital tract. Clin. Infect. Dis. an Off. Publ. Infect. Dis. Soc. Am. 32:e69-77. doi:10.1086/318710.

Lawrence, R., J. A. Church, W. Richards, and M. Borzy. 1980. Immunological mechanisms in the maintenance of pregnancy. Ann. Allergy. 44:166–173.

Lawrence, T. 2009. The nuclear factor NF-kappaB pathway in inflammation. Cold Spring Harb. Perspect. Biol. 1. doi:10.1101/cshperspect.a001651.

LeBlanc, S. J. 2008. Postpartum uterine disease and dairy herd reproductive performance: A review. Vet. J. 176:102–114. doi:10.1016/j.tvjl.2007.12.019.

Lewis, F. M. T., K. T. Bernstein, and S. O. Aral. 2017. Vaginal Microbiome and Its Relationship to Behavior, Sexual Health, and Sexually Transmitted Diseases HHS Public Access. Obs. Gynecol. 129:643–654. doi:10.1097/AOG.0000000000001932.

Leyland, N., R. Casper, P. Laberge, S. S. Singh, L. Allen, K. Arendas, N. Leyland, C. Allaire, A. Awadalla, C. Best, E. Contestabile, S. Dunn, M. Heywood, N. Leroux, F. Potestio, D. A. Rittenberg, S. S. Singh, R. Soucy, W. L. Wolfman, and V. Senikas. 2010. Endometriosis: Diagnosis and Management. J. Endometr. 2:107–134. doi:10.1177/228402651000200303.

Li, M., X. Zhao, X. Zhang, D. Wu, and S. Leng. 2018. Biodegradation of 17β-estradiol by bacterial co-culture isolated from manure. Sci. Rep. 8:3787. doi:10.1038/s41598-018- 22169-0. 28

Livadas, S., and G. P. Chrousos. 2016. Control of the onset of puberty. Curr. Opin. Pediatr. 28:551–558. doi:10.1097/MOP.0000000000000386

Lorenzen, E., E. Kudirkiene, N. Gutman, A. B. Grossi, J. S. Agerholm, K. Erneholm, C. Skytte, M. D. Dalgaard, and A. M. Bojesen. 2015. The vaginal microbiome is stable in prepubertal and sexually mature Ellegaard Göttingen Minipigs throughout an estrous cycle. Vet. Res. 46:125. doi:10.1186/s13567-015-0274-0.

Lu, Y. C., W. C. Yeh, and P. S. Ohashi. 2008. LPS/TLR4 signal transduction pathway. Cytokine. 42:145–151. doi:10.1016/j.cyto.2008.01.006.

Ma, B., L. J. Forney, and J. Ravel. 2012. Vaginal Microbiome: Rethinking Health and Disease. Annu. Rev. Microbiol. 66:371–389. doi:10.1146/annurev-micro-092611-150157.

Maes, D., M. Verdonck, and A. De Kruif. 1999. Vaginal Microecology and Vulval Discharge in Swine. Old Herborn University seminar monograph. Vol. 12.

Mathur, S., R. S. Mathur, and H. Dowda. 1978. Sex steroid hormones and antibodies to Candida albicans. Clin. Res. 26:79–87.

McCormack, W. M. 1994. Pelvic Inflammatory Disease. N. Engl. J. Med. 330:115–119. doi:10.1056/NEJM199401133300207.

Mcmurray, R. W., S. Suwannaroj, K. Ndebele, and J. K. Jenkins. 2001. Differential effects of sex steroids on T and B cells: Modulation of cell cycle phase distribution, apoptosis and bcl-2 protein levels. Pathobiology. 69:44–58. doi:10.1159/000048757.

Miesner, M. D., and D. E. Anderson. 2008. Management of Uterine and Vaginal Prolapse in the Bovine. Vet. Clin. North Am. - Food Anim. Pract. 24:409–419. doi:10.1016/j.cvfa.2008.02.008.

Mirmonsef, P., D. Gilbert, M. R. Zariffard, B. R. Hamaker, A. Kaur, A. L. Landay, and G. T. Spear. 2011. The Effects of Commensal Bacteria on Innate Immune Responses in the Female Genital Tract. Am. J. Reprod. Immunol. 65:190–195. doi:10.1111/j.1600- 0897.2010.00943.x.

Molokwu, E. C., and W. X. Wagner. 1973. Endocrine physiology of the puerperal sow. J. Anim. Sci. 36:1158–1163. doi:10.2527/jas1973.3661158x.

Moon, J. Y., M. H. Choi, and J. Kim. 2016. Metabolic profiling of cholesterol and sex steroid hormones to monitor urological diseases. Endocr. Relat. Cancer. 23:R455–R467. doi:10.1530/ERC-16-0285.

Navarro, V. M., M. L. Gottsch, C. Chavkin, H. Okamura, D. K. Clifton, and R. A. Steiner. 2009. Regulation of gonadotropin-releasing hormone secretion by kisspeptin/dynorphin/neurokinin B neurons in the arcuate nucleus of the mouse. J. Neurosci. 29:11859–11866. doi:10.1523/JNEUROSCI.1569-09.2009. 29

Ness, R. B., K. E. Kip, S. L. Hillier, D. E. Soper, C. A. Stamm, R. L. Sweet, P. Rice, and H. E. Richter. 2005. A Cluster Analysis of Bacterial Vaginosis–associated Microflora and Pelvic Inflammatory Disease. Am. J. Epidemiol. 162:585–590. doi:10.1093/aje/kwi243.

Nugent, R. P., M. A. Krohn, and S. L. Hillier. 1991. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J. Clin. Microbiol. 29.

Pagana, K. D., and T. J. Pagana. 1998. Mosby’s Manual of Diagnostic and Laboratory Tests.

Pfützner, H., and K. Sachse. 1996. Mycoplasma bovis as an agent of mastitis, pneumonia, arthritis and genital disorders in cattle. OIE Rev. Sci. Tech. 15:1477–1494. doi:10.20506/rst.15.4.987.

PigChamp. Benchmarking Summaries | Swine Management Software | PigCHAMP.com.

Rakoff-Nahoum, S., J. Paglino, F. Eslami-Varzaneh, S. Edberg, and R. Medzhitov. 2004. Recognition of commensal microflora by toll-like receptors is required for intestinal homeostasis. Cell. 118:229–241. doi:10.1016/j.cell.2004.07.002.

Ravel, J., and R. M. Brotman. 2016. Translating the vaginal microbiome: Gaps and challenges. Genome Med. 8:1–3. doi:10.1186/s13073-016-0291-2.

Ravel, J., P. Gajer, Z. Abdo, G. M. Schneider, S. S. K. Koenig, S. L. McCulle, S. Karlebach, R. Gorle, J. Russell, C. O. Tacket, R. M. Brotman, C. C. Davis, K. Ault, L. Peralta, and L. J. Forney. 2011. Vaginal microbiome of reproductive-age women. Proc. Natl. Acad. Sci. U. S. A. 108:4680–4687. doi:10.1073/pnas.1002611107.

Rodrigues, N. F., J. Kästle, T. J. D. Coutinho, A. T. Amorim, G. B. Campos, V. M. Santos, L. M. Marques, J. Timenetsky, and S. T. de Farias. 2015. Qualitative analysis of the vaginal microbiota of healthy cattle and cattle with genital-tract disease. Genet. Mol. Res. 14:6518–6528. doi:10.4238/2015.June.12.4.

Romero, R., F. Gotsch, B. Pineles, and J. P. Kusanovic. 2008. Inflammation in Pregnancy: Its Roles in Reproductive Physiology, Obstetrical Complications, and Fetal Injury. Nutr. Rev. 65:S194–S202. doi:10.1111/j.1753-4887.2007.tb00362.x.

Romoser, M. R., T. Gall, L. H. Baumgard, A. F. Keating, K. J. Stalder, and J. W. Ross. 2020a. Evaluating the efficacy of visual assessment of gilt vulva size prior to puberty on subsequent reproductive performance. Appl. Anim. Sci. 36:694–700. doi:10.15232/aas.2020-02034.

Romoser, M. R., B. J. Hale, J. T. Seibert, T. Gall, C. J. Rademacher, K. J. Stalder, L. H. Baumgard, A. F. Keating, and J. W. Ross. 2020b. Methods for reproductive tract scoring as a tool for improving sow productivity12. Transl. Anim. Sci. 4:275–284. doi:10.1093/tas/txz160. 30

Rosenberg, E., and I. Zilber-Rosenberg. 2011. Symbiosis and development: The hologenome concept. Birth Defects Res. Part C Embryo Today Rev. 93:56–66. doi:10.1002/bdrc.20196.

Ross, J. W. 2019. Identification of putative factors contributing to pelvic organ prolapse in sows (Grant # 17-224) II. Industry Summary.

Roy, A., and M. M. Matzuk. 2011. Reproductive tract function and dysfunction in women. Nat. Rev. Endocrinol. 7:517–525. doi:10.1038/nrendo.2011.79.

Russo, R., E. Karadja, and F. De Seta. 2019. Evidence-based mixture containing Lactobacillus strains and lactoferrin to prevent recurrent bacterial vaginosis: A double blind, placebo controlled, randomised clinical trial. Benef. Microbes. 10:19–26. doi:10.3920/BM2018.0075.

Sabio, G., and R. J. Davis. 2014. TNF and MAP kinase signalling pathways. Semin. Immunol. 26:237–245. doi:10.1016/j.smim.2014.02.009.

Sanglard, L. P., S. Schmitz-Esser, K. A. Gray, D. C. L. Linhares, C. J. Yeoman, J. C. M. Dekkers, M. C. Niederwerder, and N. V. L. Serão. 2020a. Investigating the relationship between vaginal microbiota and host genetics and their impact on immune response and farrowing traits in commercial gilts. J. Anim. Breed. Genet. 137:84–102. doi:10.1111/jbg.12456.

Sanglard, L. P., S. Schmitz-Esser, K. A. Gray, D. C. L. Linhares, C. J. Yeoman, J. C. M. Dekkers, M. C. Niederwerder, and N. V. L. Serão. 2020b. Vaginal microbiota diverges in sows with low and high reproductive performance after porcine reproductive and respiratory syndrome vaccination. Sci. Rep. 10. doi:10.1038/s41598-020-59955-8.

Sartor, R. B. 2008. Microbial Influences in Inflammatory Bowel Diseases. Gastroenterology. 134:577–594. doi:10.1053/j.gastro.2007.11.059.

Schromm, B., U. Seydel Mareile Mueller, K. Brandenburg, and R. Dedrick. 2021. LPS-Binding Protein (LPS)-Induced Cell Activation: A Role for Phospholipids Inhibit Lipopolysaccharide. J Immunol Ref. 174:1091–1096. doi:10.4049/jimmunol.174.2.1091.

Senger, P. L. 1997. Pathways to pregnancy and parturition. Pathways to pregnancy parturition.

Senok, A. C., H. Verstraelen, M. Temmerman, and G. A. Botta. 2009. Probiotics for the treatment of bacterial vaginosis. Cochrane Database Syst. Rev. doi:10.1002/14651858.CD006289.pub2.

Seminara, S. B., and W. F. Crowley. 2008. Kisspeptin and GPR54: Discovery of a novel pathway in reproduction. J. Neuroendocrinol. 20:727–731. doi:10.1111/j.1365- 2826.2008.01731. 31

Sheldon, I. M., J. G. Cronin, and J. J. Bromfield. 2019. Tolerance and Innate Immunity Shape the Development of Postpartum Uterine Disease and the Impact of Endometritis in Dairy Cattle. Annu. Rev. Anim. Biosci. 7:361–384. doi:10.1146/annurev-animal-020518- 115227.

Sheldon, I. M., J. Cronin, L. Goetze, G. Donofrio, and H.-J. Schuberth. 2009. Defining Postpartum Uterine Disease and the Mechanisms of Infection and Immunity in the Female Reproductive Tract in Cattle1. Biol. Reprod. 81:1025–1032. doi:10.1095/biolreprod.109.077370.

Skinner, D. C., N. P. Evans, B. Delaleu, R. L. Goodman, P. Bouchard, and A. Caraty. 1998. The negative feedback actions of progesterone on gonadotropin-releasing hormone secretion are transduced by the classical progesterone receptor. Proc. Natl. Acad. Sci. U. S. A. 95:10978–10983. doi:10.1073/pnas.95.18.10978.

Soede, N. M., A. K. Nissen, and B. Kemp. 2000. Timing of insemination relative to ovulation in pigs: Effects on sex ratio of offspring. Theriogenology. 53:1003–1011. doi:10.1016/S0093-691X(00)00246-6.

Solomon, K. R., E. A. Kurt-Jones, R. A. Saladino, A. M. Stack, I. F. Dunn, M. Ferretti, D. Golenbock, G. R. Fleisher, and R. W. Finberg. 1998. Heterotrimeric G proteins physically associated with the lipopolysaccharide receptor CD14 modulate both in vivo and in vitro responses to lipopolysaccharide. J. Clin. Invest. 102:2019–2027. doi:10.1172/JCI4317.

Spector, W. G., and D. A. Willoughby. 1963. The inflammatory response. Bacteriol. Rev. 27:117–154. doi:10.1128/mmbr.27.2.117-154.1963

Spencer, T. E., and F. W. Bazer. 2002. Biology of Progesterone Action During Pregnancy Recognition and Maintenance of Pregnancy. Front. Biosci. 7:1879–1898.

Spiegel, C. A. 1991. Bacterial vaginosis. Clin. Microbiol. Rev. 4:485–502. doi:10.1128/CMR.4.4.485.

Stalder, K., L. Karriker, and A. Johnson. 2008. Understanding sow longevity and mortality. Animal Science Conference Proceedings and Presentations. 22.

Supakorn, C. 2019. A review of aetiology and risk factors affecting sow mortality. CAB Rev. Perspect. Agric. Vet. Sci. Nutr. Nat. Resour. 14. doi:10.1079/PAVSNNR201914026..

Supakorn, C., J. D. Stock, C. Hostetler, and K. J. Stalder. 2014a. Prolapse Incidence in Swine Breeding Herds Is a Cause for Concern. Open J. Vet. Med. 7:85–97. doi:10.4236/ojvm.2017.78009. 32

Takashiba, S., T. E. Van Dyke, S. Amar, Y. Murayama, A. W. Soskolne, and L. Shapira. 1999. Differentiation of monocytes to macrophages primes cells for lipopolysaccharide stimulation via accumulation of cytoplasmic nuclear factor κB. Infect. Immun. 67:5573– 5578. doi:10.1128/iai.67.11.5573-5578.1999.

Tucureanu, M. M., D. Rebleanu, C. A. Constantinescu, M. Deleanu, G. Voicu, E. Butoi, M. Calin, and I. Manduteanu. 2018. Lipopolysaccharide-induced inflammation in monocytes/macrophages is blocked by liposomal delivery of Gi-protein inhibitor. Int. J. Nanomedicine. 13:63–76. doi:10.2147/IJN.S150918.

Valore, E. V., C. H. Park, S. L. Igreti, and T. Ganz. 2002. Antimicrobial components of vaginal fluid. Am. J. Obstet. Gynecol. 187:561–568. doi:10.1067/mob.2002.125280.

Verstraelen, H., and A. Swidsinski. 2019. The biofilm in bacterial vaginosis: Implications for epidemiology, diagnosis and treatment: 2018 update. Curr. Opin. Infect. Dis. 32:38–42. doi:10.1097/QCO.0000000000000516.

Vieira, A. T., P. M. Castelo, D. A. Ribeiro, and C. M. Ferreira. 2017. Influence of Oral and Gut Microbiota in the Health of Menopausal Women. Front. Microbiol. 8:1884. doi:10.3389/fmicb.2017.01884.

Wang, J., C. Li, L. T. Nesengani, Y. Gong, S. Zhang, and W. Lu. 2017. Characterization of vaginal microbiota of endometritis and healthy sows using high-throughput pyrosequencing of 16S rRNA gene. Microb. Pathog. 111:325–330. doi:10.1016/J.MICPATH.2017.08.030.

Wira, C. R., J. V. Fahey, C. L. Sentman, P. A. Pioli, and L. Shen. 2005a. Innate and adaptive immunity in female genital tract: Cellular responses and interactions. Immunol. Rev. 206:306–335. doi:10.1111/j.0105-2896.2005.00287.x.

Wira, C. R., K. S. Grant-Tschudy, and M. A. Crane-Godreau. 2005b. Epithelial cells in the female reproductive tract: A central role as sentinels of immune protection. Am. J. Reprod. Immunol. 53:65–76. doi:10.1111/j.1600-0897.2004.00248.x.

Wyle, F. A., and J. R. Kent. 1977. Immunosuppression by sex steroid hormones. The effect upon PHA- and PPD-stimulated lymphocytes. Clin. Exp. Immunol. 27:407–15.

Yu, C. P., R. A. Deeb, and K. H. Chu. 2013. Microbial degradation of steroidal estrogens. Chemosphere. 91:1225–1235. doi:10.1016/j.chemosphere.2013.01.112.

CHAPTER 3. VAGINAL MICROBIOME AND SERUM METABOLITE DIFFERENCES IN LATE GESTATION COMMERCIAL SOWS AT RISK FOR PELVIC ORGAN PROLAPSE

Zoë E. Kiefer1, Lucas R. Koester2,3, Lucas Showman4, Jamie M. Studer1, Amanda L.

Chipman5, Aileen F. Keating1, Stephan Schmitz-Esser 1,3, Jason W. Ross1,5

1 Department of Animal Science, Iowa State University, Ames, Iowa, United States

2 Department of Veterinary Microbiology and Preventive Medicine, Iowa State University

3 Interdepartmental Microbiology Graduate Program, Iowa State University

4 W. M. Keck Metabolomics Research Laboratory, Iowa State University

5 Iowa Pork Industry Center, Ames, Iowa

Modified from a manuscript published in 2021 by Scientific Reports

doi: 10.1038/s41598-021-85367-3

Abstract

Sow mortality attributable to pelvic organ prolapse (POP) has increased in the U.S. swine industry and continues to worsen. Two main objectives of this study were, 1) to develop a perineal scoring system that can be correlated with POP risk, and 2) identify POP risk-associated biological factors. To assess POP risk during late gestation, sows (n = 213) were scored using a newly developed perineal scoring (PS) system. Sows scored as PS1 (low), PS2 (moderate), or

PS3 (high) based on POP risk. Subsequently, 1.5, 0.8, and 23.1% of sows scored PS1, PS2, or

PS3, respectively, experienced POP. To identify biomarkers, serum and vaginal swabs were 34 collected from late gestation sows differing in PS. Using GC-MS, 82 serum metabolite differences between PS1 and PS3 animals (P < 0.05) were identified. Vaginal swabs were utilized for 16S rRNA gene sequencing and differences in vaginal microbiomes between PS1 and PS3 animals were detected on a community level (P < 0.01) along with differences in abundances of 89 operational taxonomic units (P < 0.05). Collectively, these data demonstrate that sows with greater POP risk have differential serum metabolites and vaginal microflora.

Additionally, an initial and novel characterization of the sow vaginal microbiome was determined.

Introduction

Sow reproductive performance across the U.S. swine industry has increased in the past decade culminating in some farms achieving thirty pigs weaned per sow per year. Despite marked improvements in sow key performance indicators (farrowing rate, litter size, pigs weaned), there has also been a substantial increase in sow mortality due to pelvic organ prolapse

(POP) during late gestation and early lactation (Supakorn et al., 2014a). In a survey of the U.S. swine industry POP was determined to contribute to approximately 21% of sow deaths annually.

Pelvic organ prolapse is an anatomical disorder characterized by one or more pelvic organs

(uterus, rectum and/or vagina) pressing up against or out of the vagina (Jelovsek et al., 2007). In sows, POP occurs predominantly within a few days of farrowing and can result in loss of both the sow and her offspring. The incidence of sow POP continues to increase, yet there is a lack of mitigation strategies since the biological causative underpinnings are ill-defined (Supakorn et al.,

2014a).

The surfaces and cavities of all mammalian species exposed to the environment are host to microorganisms (J. Ravel et al., 2011) and can have substantial influence on animal health and 35 well-being. Recent discoveries in swine have revealed the contribution of the microbiota to gut function (ND et al., 2019; Maltecca et al., 2020). Information on the vaginal microbiome in swine is limited and, in some cases, focused on genetic lines used in commercial production

(Lorenzen et al., 2015; Wang et al., 2017; Sanglard et al., 2020a; Sanglard et al., 2020b).

Findings in other species support the notion that changes in the microbiota of the reproductive tract can compromise reproductive function; as alterations to and increased diversity in the vaginal microbiome affect susceptibility to gynecologic infections (Green et al., 2015).

Serum contains biomarkers including lipids, amino acids, peptides, nucleic acids, organic acids, vitamins, thiols and carbohydrates. These are important in biological systems and have the ability to assist further understanding of disease phenotypes (Arakaki et al., 2008). Non-targeted metabolomics is a global unbiased analysis of small-molecule metabolites present in a given biological sample (Naz et al., 2014). The objective of this study was to develop a perineal scoring system to evaluate risk of POP of sows during late gestation and additionally assess differences in vaginal microbial populations and molecular features associated with POP risk to serve as potential biomarkers to better understand biological alterations associated with POP.

The hypothesis that both the vaginal microbiome and serum small-molecule repertoires would differ between sows with differing risk of POP was tested in late gestation sows.

Results

Differences in perineal score is associated with differing risk of pelvic organ prolapse

Of the 213 sows assigned a perineal score (PS) during late gestation, 68, 119 and 26 were assigned PS1, PS2 and PS3, respectively. Retrospectively, there was no difference (P = 0.81) in the number of sows that experience POP between PS1 (1.5%) and PS2 (0.8%) while sows that 36 scored a PS3 (23.1%) had greater (P < 0.01) incidence of POP compared to PS1 or PS2 scored sows (Figure 3.1). Additionally, parity had no effect on PS (P = 0.54).

16S rRNA gene amplicon sequencing identified association between vaginal microbiota and perineal score

Of 6,437 operational taxonomic units (OTUs) obtained from 42 samples (23 PS1 and 19

PS3), A total of 1,711 OTUs remained after quality control and removal of OTUs representing less than ten sequences. The average number of sequences per sample was 13,353 with a standard deviation of 8,422. The majority of reads were bacterial (98.27%), and 1.73% were archaeal. Alpha diversity estimators revealed no significant differences between samples regarding species richness, community evenness and diversity (Figure A.3.1). Out of the 1,711

OTUs 26 phyla were identified. Firmicutes, Bacteroidetes, Proteobacteria, Actinobacteria, and

Fusobacteria were the most abundant, representing 49.6%, 24.0%, 12.6%, 4.0%, and 3.3% of all reads, respectively (Figure 2A). There were 571 genera identified of which the most abundant included Pasteurellaceae unclassified (7.6%), Veillonella (6.2%), Clostridium cluster I (4.5%),

Bacteroides (3.4%), and Prevotellaceae unclassified (3.3%) (Figure 2B). The 50 most abundant vaginal tract OTUs are reported in Table A.3.1. The most abundant OTU was identified as

Veillonella, and accounted for 6.2% of all the reads. OTUs 1, 2, 3, 4, and 5 were classified as

Pasteurellaceae unclassified, Fusobacterium, Prevotellaceae UCG-001, and

Phascolarctobacterium, respectively, and accounted for 5.2%, 2.6%, 2.4%, 2.2%, and 2.1%, respectively.

When comparing the microbiota of PS1 and PS3 sows, there were differences on a whole community level using ANOSIM (P < 0.01) and some separation using canonical 37 correspondence analysis (Figure 3.3A). However, principal coordinate analyses revealed no distinct clustering of PS1 and PS3 samples (Figure 3.3B). Using LEfSe, abundance differences

(P < 0.05) in 89 total OTUs were observed. Of these, 26 OTUs were more abundant in PS1 and

68 were more abundant in PS3. Out of the 89 significantly different OTUs, 24 were within the

100 most abundant OTUs (Figure 3.4, Table 3.1). Of those, 12 OTUs were more abundant in

PS1, including OTUs 1, 3, 7, 22, 31, 37, 49, 52, 54, 55, 57, and 71. There were also 12 OTUs that were more abundant in PS3, including OTUs 4, 14, 28, 38, 43, 46, 51, 58, 70, 73, 84, and 97.

A higher abundance of Prevotellaceae, and Treponema within several OTUs of each were noted in PS3 animals compared to PS1. In addition, a Streptococcus dysgalactiae OTU was more abundant (P < 0.01) in PS3 sows.

Serum small molecule metabolites differ with POP risk

Following GC-MS, a total of 960 molecular features were detected in the 44 serum samples used (25 PS1 and 19 PS3). There were 82 differently abundant metabolites between treatment groups (P < 0.05), and four were more abundant in PS1 animals (Figure 3.5). Overall metabolites between PS1 and PS3 animals were determined and an expected overlap was noted, but distinct differences were also observed (Figure A.3.2). Grouping of PS based on serum metabolite abundances was also determined (Figure 3.6). Out of the 960 molecular features detected, 93 of them were identifiable. When evaluating just the 93 identified metabolites 16 differed (P < 0.05) between PS1 and PS3 (Table 3.2). Metabolites such as L-methionine, L- alanine, 2-aminobutanoic acid, lactic acid, D-glucose and several others were more abundant (P

< 0.05) in PS3 compared to PS1 scored sows. All 16 of the identified differential metabolites were more abundant in PS3 animals. One of particular interest is D-Fructose which was 5.3-fold 38 more abundant in PS3 sows and had a 78.5% accuracy as a potential POP risk biomarker. There were several other potential biomarkers identified all above 74% area under the curve (Table

3.2). Nine of these potential biomarkers met the biomarker criteria, however, eight are currently not identifiable.

Discussion

Despite dramatic improvements in reproductive performance in the U.S. swine industry,

POP still remains an issue. Of all causes of sow mortality, 21% is due to POP, although the biological explanation remains unknown. This is considerably higher than other livestock species, such as the bovine, which experiences less than 1% (Carluccio et al., 2020). Critical for the development of mitigation strategies is the better understanding of the biological events that precede POP. To this end, this study tested the hypothesis that a PS system could have utility in defining POP risk and further, that differential vaginal microbiota and serum metabolites would be associated with altered POP risk in sows. Our data demonstrates that sows with variable risk for POP can be identified during late gestation through phenotypic evaluation of the perineal region. This evaluation system of the perineal region of sows is significant for the swine industry as identification of animals with elevated risk for POP has proven difficult. While the PS was effective in assessing POP risk, the biological factors contributing to the phenotypic variation observed remain unknown.

Previous studies investigating the sow microbiome have not been conducted in commercial settings and/or were not conducted with pregnant sows, which could influence the environmental microbiota presence and potentially affect the vaginal microbiome (Lorenzen et al., 2015; Wang et al., 2017; Sanglard et al., 2020a). Other studies have attempted to evaluate the 39 vaginal microbiota in swine via in vitro culture (Bara et al., 1993). While providing some novel information, cultivation approaches may not accurately capture the in vivo representation of the microbiota in swine as not all microorganisms grow effectively in a culture system. This study is our initial characterization of the vaginal microbiome of commercial sows during gestation. The data demonstrates a diverse microbial community within the vagina of commercial sows while also identifying differences in the vaginal microbiome between sows with differing risk for POP.

When comparing overall microbial community structure between sows with low and high risk of POP there were differences on a whole community level. It is well established that the most abundant microbe in the human vaginal microbiome is Lactobacillus (Ma et al., 2012). In this study, Lactobacillus presence was observed, but at a much lower average abundance compared to human vaginal microbiomes. In agreement with this, other published pig vaginal microbiota studies also determined a relatively low overall abundance of Lactobacillus

(Lorenzen et al., 2015; Wang et al., 2017; Sanglard et al., 2020b; Sanglard et al., 2020a). A higher abundance of Streptococcus dysgalactiae was noted in PS3 sows. Streptococcus dysgalactiae is a major pathogen in humans and animals and has also been associated with equine reproductive system infections (Pinho et al., 2016; Baracco, 2019). Further, similarities in the genera of species present were consistent with previous studies on the swine vaginal microbiome. Sanglard et al. reported Fusobacterium presence in non-pregnant gilts at a relative abundance of 11.4%. In the current study Fusobacterium was observed in late gestation sows at a lower relative abundance (2.6%) and was greater in sows with low risk for POP compared to sows with high risk. A higher abundance of Fusobacterium in the sow vaginal microbiota with endometritis has been determined (Wang et al., 2017). The most abundant microbe present in the vaginal microbiome of sows in this study was Veillonella, with a relative abundance of 6.2% 40 which also differed in relative abundance between sows of different POP risk with a higher abundance of Veillonella in PS1 sows. Previous studies have reported that Veillonella is present in the vaginal microbiome of non-pregnant gilts at a relative abundance of 2.0% (Sanglard et al.,

2020a) and was also noted in non-pregnant sows (Wang et al., 2017). Further data is needed to confirm that changes observed are associated with differences in PS.

Differences between this study and that performed in non-pregnant females are congruent with observations of shifts in microbial populations during pregnancy and parturition in humans

(Aagaard et al., 2012). An increase in Treponema OTUs in the PS3 sows was noted in this study though it is unknown if the higher abundance of Treponema OTUs in PS3 sows is associated with a higher risk for POP. It is noteworthy to mention that in humans Treponema pallidum is the causative agent of syphilis (Danforth’s Obstetrics and Gynecology - Google Books).

Furthermore, an increased abundance of Treponema on the vaginal microbiota has previously been associated with genital disease and reproductive disorders in cattle (Rodrigues et al., 2015).

Interestingly, Treponema has been linked to swine dysentery (Hughes et al., 1975), a severe infectious disease that is often characterized by inflammation in the large intestine. While functional association with POP needs to be verified in future studies, the results presented herein highlight candidate phylotypes for future attention.

Interaction between the microbiome and host has become an increasing area of interest.

The vaginal microbiome, in particular, is being investigated in humans in relation to reproductive disease, and this work has begun to transfer into livestock. Further, compromised human reproductive health, including pelvic inflammatory disease (PID), has been linked to changes in the vaginal microbiome (Green et al., 2015) and a biological link between pelvic infections and a disequilibrium of the vaginal flora is thought to exist (Sharma et al., 2014). However, PID is not 41 the only disease linked to a disruption of the vaginal microbiome. Bacterial vaginosis (BV) in humans, although resulting from different pathogenic infections, has also been linked to altered vaginal microbiome (Hedges et al., 2006; Nardis et al., 2013). It has been reported that microorganisms related to BV have been linked to an elevated risk of acquiring PID (Ness et al.,

2005). Vaginal microbiota consistent with BV or other changes in the vaginal microbiome are associated with an increased risk for viral sexually transmitted diseases and subsequent ascending infections (Nardis et al., 2013; Green et al., 2015). As a result, BV contributes to endometritis and cervical inflammation which includes immune cell infiltration and localized erythema. Additionally, sexually transmitted infections have been linked to alterations in the vaginal microbiota (Sheldon et al., 2009; Nardis et al., 2013). Humans suffering from BV are reported to have substantial increases in Prevotella (Ceccarani et al., 2019), which was also elevated in PS3 sows. Differences in relative abundance of Prevotella and Fusobacterium in the vaginal microbiome between PS1 and PS3 sows was also discovered. Both of these microbial genera have been reported to be associated with infections caused by gram negative bacteria

(Garrett and Onderdonk), relevant since these are associated with BV and PID. Identifying differences in the vaginal microbiome associated with phenotypes of differential POP risk creates a starting point to mechanistically explore the contribution of vaginal dysbiosis to POP in swine.

In addition to characterization of the vaginal microbiome in sows with variable risk for

POP, changes in components in circulation may also provide understanding of physiological contributions to POP risk. Serum contains a plethora of information about various small molecules suggesting the serum metabolome may have value as a predictive phenotype, particularly if coupled with microbial changes. Changes observed in the metabolome of a 42 diseased individual may serve as primary indicators and in some cases are already used in clinical practice (Zhang et al., 2012). Alterations to the serum metabolome between sows with high and low risk for POP were observed in this study, known and potentially novel unidentified

POP risk associated molecules. Similar to observations of the vaginal microbiome, major shifts in the overall serum metabolome were not observed, which was anticipated as only a subset of metabolites would be expected to be associated with such a specific phenotype in otherwise consistent biological samples. In this study, 26 metabolites were discovered which differed between low and high POP risk animals, 21 of which were increased in high risk of POP sows.

Small molecules within circulation may be useful to help differentiate between diseased and non-diseased status (Nicholson and Lindon, 2008). Circulating glucose levels have been reported to change during BV (Ceccarani et al., 2019). In our study of the serum metabolome, changes in glucose and its derivatives were identified between PS1 and PS3 animals. Vulvo- vaginal candidiasis (VVC) in humans has similar effects on the metabolome as BV (Ceccarani et al., 2019). Changes in lactic acid as well as changes in amino acids were determined in this study, suggesting molecular feature differences could be valuable in assessing reproductive dysfunction in sows. Increased abundance of 2-aminobutanoic acid, a derivative of butyric acid, which has been shown to correlate with inflammation (Leonel and Alvarez-Leite, 2012; Grilli et al., 2016), was noted in PS3 sows compared to PS1 sows. Changes in the serum metabolome associated with reproductive dysfunction along with changes in the vaginal microbiome have been reported (Aagaard et al., 2012; Ceccarani et al., 2019) consistent with the findings reported herein.

Collectively, this study established a phenotypic scoring system to identify sows with differential POP risk. This approach enabled identification of biological contributors to POP risk 43 which could serve as reliable POP biomarkers. Development of the scoring system will have utility for both producers and researchers. In addition, several bacterial and metabolomic candidates of interest have been discovered and the vaginal microbiome of pregnant sows has been established. The putative markers identified in this work will require determination of causality. Further, a clear biological relationship between the vaginal microbiome and serum metabolite differences is a future research avenue and addition of POP risk adds a further layer of both complexity and translatability.

Materials and Methods

Animals

All experiments in this study were approved by the Iowa State University Institutional

Animal Care and Use Committee and all methods were conducted in accordance with relevant guidelines and regulations. Two hundred and thirteen pregnant females (gestation days 105-117) from two commercial sow farms were used. All animals were individually housed in the commercial sow farms.

Perineal scoring system

To categorize sows with differing risk of POP, a perineal scoring system was developed to assign sows into low (PS1), moderate (PS2) or high (PS3) POP risk (Figure 3.7). To assign the score, the perineal region was visually evaluated for swelling, redness and protrusion. If a sow lacked swelling, redness and protrusion they were considered low risk for POP and assigned as

PS1. Sows with some characteristics such as moderate swelling, redness and protrusion of the perineal area were considered moderate risk for POP and assigned as PS2. Sows demonstrating 44 all of the characteristics of severe swelling, redness and protrusion of the perineal area were assigned to the PS3 category and considered high risk for POP. Importantly, a sow scoring a PS3 would be considered abnormal for any stage of gestation (Figure 3.7). To minimize variation, animals were scored at one time point during the farm visit by the same individual, while the sow was lying down, and performed on animals between days 105 and 117 of gestation. The variation in gestation stage evaluation was the result of the timing of the visit to the farm and the emphasis was placed on sows within two weeks of expected farrowing date. Subsequently, the average parity was 3.04 ± 0.21 and the average gestation day was 110.7 ± 0.5, neither of which were significant between PS (P > 0.3). Animals chosen for sampling were parity matched. Animals were then monitored for subsequent POP occurrence during and following parturition.

Sample collection

Vaginal swabs for microbiota DNA extraction and blood for serum collection were collected from sows classified as PS1 (n = 29) or PS3 (n = 23). Vaginal swabs were collected by aseptically inserting a 7-inch histology brush (2199, Puritan Medical Products) into the vagina and brushing the vaginal orifice for approximately 15 sec. Swabs were removed and immediately placed in sterile 1x phosphate buffered saline (PBS), frozen on dry ice, and stored at -80⁰C until pelleted and used for DNA extraction. Blood samples were collected by jugular venipuncture into vacutainer tubes, which were kept at room temperature to allow appropriate clotting before separation of serum. All serum samples were processed by centrifugation at 20⁰C for 15 min at

2,000 × g followed by aspiration of serum for storage at -80⁰C until further analysis.

DNA extraction

Tubes containing vaginal microbiota swabs in PBS were thawed, vortexed for 5 minutes to detach cells, and centrifuged at 5,000 × g at room temperature for 15 minutes to form a pellet.

Supernatant was discarded and pellets were resuspended in 1 mL of sterile PBS and DNA was extracted with a Qiagen DNeasy Powerlyzer Powersoil kit (Qiagen, Germantown, MD) per manufacturer’s protocol. Briefly, mechanical cell lysis was performed using a Fisher Scientific

Beadmill 24. The concentration of the isolated DNA was measured using a Qubit 3.0

Fluorometer and diluted to 25 ng/µL for each sample.

16S rRNA gene sequencing

16S rRNA gene sequences were amplified from the vaginal swab samples of individual animals. PCR amplification of the V4 region of bacterial 16S rRNA genes was performed with the conserved primers [515F (5′-GTGYCAGCMGCCGCGGTAA-3′;(Parada et al., 2016)) and

(806R (5′-GGACTACNVGGGTWTCTAAT-3′; (Apprill et al., 2015)) as previously described

(Kozich et al., 2013)] with sequence tags (bar codes) and sequencing primers incorporated into each PCR primer. PCR mixtures contained 200 µM (each) deoxyribonucleotide triphosphate, 2.0

µM (each) primer, 2.0 U Ampligold Taq polymerase (Applied Biosystems, Foster City, CA), 2.5 mM MgCl2, 50 ng template DNA, Ampligold Taq buffer (Applied Biosystems), and sterile water to 50 µl. PCRs were performed in a PTC-225 thermal cycler (BioRad DNA Engine Dyad

Peltier Thermal Cycler) with the following protocol: 3 minutes at 95°C; 21 cycles of 1 minute at

95°C, 30 seconds at 56°C, and 45 seconds at 72°C; and a final elongation step for 3 minutes at

72°C. All PCR products were purified with a QIAquick 96 PCR Purification Kit (Qiagen,

Hilden, Germany) according to the manufacturer’s instructions. PCR bar-coded amplicons were 46 mixed at equal molar ratios and used for Illumina MiSeq paired-end sequencing with 250 bp read length and cluster generation with 10% PhiX control DNA on an Illumina MiSeq platform.

Sequence analysis

Sequence analysis was performed with Mothur V1.40.4 following the Mothur MiSeq

Standard Operating Procedure (Kozich et al., 2013). Barcode sequences, primers and low-quality sequences were trimmed using a minimum average quality score of 35, with a sliding window size of 50 bp. Chimeric sequences were removed with the “Chimera.uchime” command. For alignment and taxonomic classification of OTUs, the SILVA SSU NR reference database (V132) provided by the Mothur website was used. Sequences were clustered into OTUs with a cutoff of

99% 16S rRNA gene similarity (=0.01 distance) and were ordered from most to least abundant.

Representative sequences for the 50 most abundant OTUs were classified using NCBI BLAST, to improve classification accuracy.

Gas Chromatography Mass Spectrometry (GC-MS)

Serum samples from the identical sows used for microbiota analysis were analyzed by non-targeted GC-MS at the ISU W.M. Keck Metabolomics Research Laboratory. Sample preparation was conducted using a modified version of the methanolic extraction and sample preparation methods established by A et al. (Jiye et al., 2005). Serum (300 µL) from each sow was spiked with internal standards (10 µg of nonadecanoic acid and 10 µg ribitol (Sigma-Aldrich

CO., St. Louis, MO)) and extracted twice using 0.8mL cold methanol (Fisher Scientific,

Waltham, MA), extracts were pooled and dried (0.8 mL of each extract) overnight using a

SpeedVac system in preparation for gas chromatography with tandem mass spectrometry (GC- 47

MS) analysis. The dried extracts were subjected to methoximation with methoxyamine hydrochloride (50 µL of 20 mg/mL in pyridine (Sigma-Aldrich CO., St. Louis, MO)) at 30°C for

90 min. Samples were silylated with 70 µL BSTFA/1% TCMS (Sigma-Aldrich CO., St. Louis,

MO) at 42°C for 30 minutes, and then subjected to GC-MS analysis on an Agilent 7890C gas chromatograph in tandem with a 5975C MSD. The GC oven program began at 80°C and was ramped at 5°C/minute to 320°C which was held for 6 minutes. The mass range was set from 40–

800 m/z. The separation column was an HP5MSI (30 m long, 0.250 mm ID, 0.25 μm film thickness). The mass spectrometer operated under standard conditions with a 230°C ion source.

Identification and quantification were conducted using AMDIS with a manually curated retention indexed GC-MS library with additional identification performed using the NIST17 and

Wiley 11 GC-MS spectral library.

Statistical analyses

The Mixed procedure in SAS was used to evaluate the effect of PS on the occurrence of

POP by evaluating differences of least square means followed by comparisons between different

PS using the probability of differences function.

Microbial sequence analysis

To compare alpha diversity between experimental groups, reads were randomly subsampled to accommodate the sample with the lowest number of reads across data sets (3,000 sequences). Measurements of Chao species richness, Shannon Diversity, and Simpson evenness were taken to compare community structures between experimental groups. The means of the experimental group alpha diversity measures were compared using a pooled t-test assuming equal variance.

To compare overall microbial community structure, samples were given a Bray-Curtis dissimilarity value and means were then compared using the analysis of similarity (ANOSIM) package provided by Mothur. Bray-Curtis was selected as the dissimilarity coefficient because of its ability to compare closely-related samples (Bray and Curtis, 1957).

All plotting was completed using ggplot2, v2_3.1.1 graphing package(R: The R Project for Statistical Computing) in R 3.6.0. Overall variation in bacterial communities was visualized using principle coordinate analysis (PCoA). Canonical correlation analysis (CCA) was used to visualize the variation strictly due to PS. This information was generated with the Phyloseq

(v1.28.0 (McMurdie and Holmes, 2013)) and Vegan (v2.5-5, (Oksanen et al., 2010)) packages using the shared and taxonomy file generated in Mothur. Bray-Curtis dissimilarity measures were used to generate distances between samples for the PCoA and CCA plots.

Differences in individual OTUs were compared using Linear Discriminant Analysis

(LDA) Effect Size (LEfSe, (Segata et al., 2011)), identifying OTUs that most likely explain the greatest between-group variation. LEfSe performs a Kruskall-Wallis test to analyze all OTUs, broadly selecting OTUs that show the most variation between sample types. The retained features then undergo a pairwise Wilcoxon test, removing any OTUs that do not differ in 49 ranking. In the last step, a linear discriminant analysis model is built from the retained OTUs to determine the effect sizes for each feature. P-values of < 0.05 were considered significant.

Metabolite analysis

MetaboAnalyst (Chong et al., 2019), an R based statistical package was utilized to provide statistical evaluation of the GC-MS metabolite data. Important features were visualized using MetaboAnalyist’s univariate and multivariate analysis methods. For two-group data, fold change (FC) analysis, t-tests, and volcano plots, which is a combination of the first two methods, were produced. Hierarchical Clustering was created using the hclust function in package stat, and was displayed as a heatmap. Distances were calculated using Euclidean measures and clustering using the ward.D algorithm.

Data availability

The 16S rRNA gene sequences have been submitted to the NCBI Sequence Read Archive SRA and are available under the BioProject ID PRJNA623913.

Acknowledgements

This project was supported by the National Pork Board and the Foundation for Food and

Agriculture Research grant #18-147.

Conflict of Interest

Any opinion, findings, conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the funding agency. No conflicts of interest, financial or otherwise are declared by the authors.

Authors contributions

ZEK designed experiments, completed experiments, performed analysis and wrote the draft of the manuscript. LRK and LS contributed to execution of experiments, data analysis and edited the manuscript. JMS and ALC contributed to experimental design, execution of experiments, and edited the manuscript. AFK contributed to experimental design and edited the manuscript. SSE contributed to experimental design, data analysis and edited manuscript. JWR contributed to experimental design, project funding, provided project oversight and edited the manuscript.

References

Supakorn C, Stock JD, Hostetler C, Stalder KJ. Prolapse incidence in swine breeding herds is a cause for concern. Open J. Vet. Med. 2014;7:85–97.

Jelovsek JE, Maher C, Barber MD. Pelvic organ prolapse. Lancet. 2007;369:1027–1038. doi: 10.1016/S0140-6736(07)60462-0. - DOI - PubMed

Ravel J, et al. Vaginal microbiome of reproductive-age women. Proc. Natl. Acad. Sci. USA. 2011;108:4680–4687. doi: 10.1073/pnas.1002611107. - DOI - PMC - PubMed

Aluthuge ND, et al. Board invited review: The pig microbiota and the potential for harnessing the power of the microbiome to improve growth and health. J. Anim. Sci. 2019;97:3741– 3757. doi: 10.1093/jas/skz208. - DOI - PMC - PubMed

Maltecca C, Bergamaschi M, Tiezzi F. The interaction between microbiome and pig efficiency: A review. J. Anim. Breed. Genet. 2020;137:4–13. doi: 10.1111/jbg.12443. - DOI - PubMed 51

Wang J, et al. Characterization of vaginal microbiota of endometritis and healthy sows using high-throughput pyrosequencing of 16S rRNA gene. Microb. Pathog. 2017;111:325–330. doi: 10.1016/j.micpath.2017.08.030. - DOI - PubMed

Lorenzen E, et al. The vaginal microbiome is stable in prepubertal and sexually mature Ellegaard Göttingen Minipigs throughout an estrous cycle. Vet. Res. 2015;46:125. doi: 10.1186/s13567-015-0274-0. - DOI - PMC - PubMed

Sanglard LP, et al. Investigating the relationship between vaginal microbiota and host genetics and their impact on immune response and farrowing traits in commercial gilts. J. Anim. Breed. Genet. 2020;137:84–102. doi: 10.1111/jbg.12456. - DOI - PubMed

Sanglard LP, et al. Vaginal microbiota diverges in sows with low and high reproductive performance after porcine reproductive and respiratory syndrome vaccination. Sci. Rep. 2020;10:1–11. doi: 10.1038/s41598-020-59955-8. - DOI - PMC - PubMed

Green KA, Zarek SM, Catherino WH. Gynecologic health and disease in relation to the microbiome of the female reproductive tract. Fertil. Steril. 2015;104:1351–1357. doi: 10.1016/j.fertnstert.2015.10.010. - DOI - PubMed

Arakaki AK, Skolnick J, McDonald JF. Marker metabolites can be therapeutic targets as well. Nature. 2008;456:443. doi: 10.1038/456443c. - DOI - PubMed

Naz S, Vallejo M, García A, Barbas C. Method validation strategies involved in non-targeted metabolomics. J. Chromatogr. A. 2014;1353:99–105. doi: 10.1016/j.chroma.2014.04.071. - DOI - PubMed

Carluccio A, de Amicis I, Probo M, Giangaspero B, Veronesi MC. Prevalence, survival and subsequent fertility of dairy and beef cows with uterine prolapse. Acta Vet. Hung. 2020;68:91–94. doi: 10.1556/004.2020.00017. - DOI - PubMed

Bara M, Mcgowan M, O’boyle D, Cameron R. A study of the microbial flora of the anterior vagina of normal sows during different stages of the reproductive cycle. Aust. Vet. J. 2008;70:256–259. doi: 10.1111/j.1751-0813.1993.tb08043.x. - DOI - PubMed

Ma B, Forney LJ, Ravel J. Vaginal microbiome: Rethinking Health and disease. Annu. Rev. Microbiol. 2012;66:371–389. doi: 10.1146/annurev-micro-092611-150157. - DOI - PMC - PubMed

Baracco GJ. Infections Caused By group C and G Streptococcus (Streptococcus dysgalactiae subsp. equisimilis and Others): Epidemiological and clinical aspects. Microbiol. Spectr. 2019;7:275–283. doi: 10.1128/microbiolspec.GPP3-0016-2018. - DOI - PubMed

Pinho MD, et al. Beta-hemolytic Streptococcus dysgalactiae strains isolated from horses are a genetically distinct population within the Streptococcus dysgalactiae taxon. Sci. Rep. 2016;6:1–9. doi: 10.1038/srep31736. - DOI - PMC - PubMed 52

Aagaard K, et al. A Metagenomic approach to characterization of the vaginal microbiome signature in pregnancy. PLoS ONE. 2012;7:e36466. doi: 10.1371/journal.pone.0036466. - DOI - PMC - PubMed

Danforth’s Obstetrics and Gynecology: Google Books. https://books.google.com/books?hl=en&lr=&id=v4krPhqFG8sC&oi=fnd&pg=PR11&....

Rodrigues NF, et al. Qualitative analysis of the vaginal microbiota of healthy cattle and cattle with genital-tract disease. Genet. Mol. Res. 2015;14:6518–6528. doi: 10.4238/2015.June.12.4. - DOI - PubMed

Hughes R, Olander HJ, Williams CB. Swine dysentery: Pathogenicity of Treponema hyodysenteriae. Am. J. Vet. Res. 1975;36:971–977. - PubMed

Sharma H, Tal R, Clark N, Segars J. Microbiota and pelvic inflammatory disease. Semin. Reprod. Med. 2014;32:043–049. doi: 10.1055/s-0033-1361822. - DOI - PMC - PubMed

Nardis C, Mosca L, Mastromarino P. Vaginal microbiota and viral infections Vaginal microbiota and viral sexually transmitted diseases. Ig. 2013;25:443–456. - PubMed

Hedges SR, Barrientes F, Desmond RA, Schwebke JR. Local and systemic cytokine levels in relation to changes in vaginal flora. J. Infect. Dis. 2006;193:556–562. doi: 10.1086/499824. - DOI - PubMed

Ness RB, et al. A cluster analysis of bacterial vaginosis–associated microflora and pelvic inflammatory disease. Am. J. Epidemiol. 2005;162:585–590. doi: 10.1093/aje/kwi243. - DOI - PubMed

Sheldon IM, Cronin J, Goetze L, Donofrio G, Schuberth H-J. Defining postpartum uterine disease and the mechanisms of infection and immunity in the female reproductive tract in cattle1. Biol. Reprod. 2009;81:1025–1032. doi: 10.1095/biolreprod.109.077370. - DOI - PMC - PubMed

Ceccarani C, et al. Diversity of vaginal microbiome and metabolome during genital infections. Sci. Rep. 2019;9:1–12. doi: 10.1038/s41598-019-50410-x. - DOI - PMC - PubMed

Garrett, W. S. & Onderdonk, A. B. Bacteroides, Prevotella, Porphyromonas, and Fusobacterium Species (and Other Medically Important Anaerobic Gram-Negative Bacilli).

Zhang A, Sun H, Wang X. Serum metabolomics as a novel diagnostic approach for disease: A systematic review. Anal. Bioanal. Chem. 2012;404:1239–1245. doi: 10.1007/s00216- 012-6117-1. - DOI - PubMed

Nicholson JK, Lindon JC. Systems biology: Metabonomics. Nature. 2008;455:1054–1056. doi: 10.1038/4551054a. - DOI - PubMed 53

Grilli E, Tugnoli B, Foerster CJ, Piva A. Butyrate modulates inflammatory cytokines and tight junctions components along the gut of weaned pigs. J. Anim. Sci. 2016;94:433–436. doi: 10.2527/jas.2015-9787. - DOI

Leonel AJ, Alvarez-Leite JI. Butyrate: Implications for intestinal function. Curr. Opin. Clin. Nutr. Metab. Care. 2012;15:474–479. doi: 10.1097/MCO.0b013e32835665fa. - DOI - PubMed

Parada AE, Needham DM, Fuhrman JA. Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environ. Microbiol. 2016;18:1403–1414. doi: 10.1111/1462-2920.13023. - DOI - PubMed

Apprill A, McNally S, Parsons R, Weber L. Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquat. Microb. Ecol. 2015;75:129–137. doi: 10.3354/ame01753. - DOI

Kozich JJ, Westcott SL, Baxter NT, Highlander SK, Schloss PD. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the miseq illumina sequencing platform. Appl. Environ. Microbiol. 2013;79:5112–5120. doi: 10.1128/AEM.01043-13. - DOI - PMC - PubMed

Jiye A, et al. Extraction and GC/MS analysis of the human blood plasma metabolome. Anal. Chem. 2005;77:8086–8094. doi: 10.1021/ac051211v. - DOI - PubMed

Bray JR, Curtis JT. An ordination of the upland forest communities of Southern Wisconsin. Ecol. Monogr. 1957;27:325–349. doi: 10.2307/1942268. - DOI

R: The R Project for Statistical Computing. https://www.r-project.org/. (Accessed 7 May 2020)

McMurdie PJ, Holmes S. phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE. 2013;8:e61217. doi: 10.1371/journal.pone.0061217. - DOI - PMC - PubMed

Oksanen, J. et al. Vegan: Community Ecology Package. R package version (2010).

Segata N, et al. Metagenomic biomarker discovery and explanation. Genome Biol. 2011;12:R60. doi: 10.1186/gb-2011-12-6-r60. - DOI - PMC - PubMed

Chong J, Wishart DS, Xia J. Using MetaboAnalyst 4.0 for comprehensive and integrative metabolomics data analysis. Curr. Protoc. Bioinform. 2019;68:e86. doi: 10.1002/cpbi.86. - DOI – PubMed

Tables and Figures

Perineal Score as an Indicator of POP Risk

25% 23.1%

20%

15%

10% Percent Prolapsed Percent

5% 1.5% 0.8% 0% PSPS1 1 PS2PS 2 PSPS3 3

Figure 3.1. Perineal score in late gestation as an indicator of pelvic organ prolapse (POP) risk. Perineal scoring was conducted on sows (n = 213) during late gestation (days 105-117). Sows assigned a PS1 (n = 68), PS2 (n = 119) and PS3 (n = 26) were assumed low, medium or high risk, respectively, for POP. As predicted, 23.1% of PS3 scored sows subsequently died due to POP while 1.5 and 0.8% of sows scoring PS1 and PS2, died respectively. These data demonstrate the ability to distinguish differential risk of POP for animals during late gestation.

A B

Figure 2. Most abundant microbial phyla (A) and genera (B) in vaginal swabs from sows with low and high risk for prolapse. (A) Vaginal swabs were collected from sows with low (PS1) and high (PS3) risk for prolapse and microbial detection and abundance were determined through 16s rRNA sequencing and analysis. Spirochaetes abundance between PS1 and PS3 sows was significantly different (P < 0.05). (B) Stacked bar charts comparing the ten most abundant vaginal microbial genera. Relative abundance of vaginal swab microbial communities on a genus level represent the mean across each perineal score. Animals assumed low risk for POP are considered PS1 and high risk animals for POP are considered PS3.

A B

Figure 3. Vaginal swab community comparisons. (A) Canonical correspondence analysis showing maximum variation between samples in community due to perineal score (PS). Data for low risk animals (PS1) are depicted in light grey and high-risk animals (PS3) are in black. All points represent Bray Curtis dissimilarity measures for each sample. (B) Principal Coordinates Analysis showing differences in beta diversity of vaginal microbial communities from sows with different perineal scores (PS). All points represent Bray Curtis dissimilarity measures for each sample. Ellipses were based on a multivariate normal distribution of data points. There were statistical differences in overall microbial communities were found due to perineal score ANOSIM (P < 0.01). 57

Figure 3.2. Different OTUs in the vaginal microbiota of sows with different perineal scores. Significantly different (P < 0.05) OTUs were identified with LEfSe. Only significant OTUs within the 100 most abundant are shown. Error bars represent the standard error of the mean. Out of all significantly different OTUs, 24 were considered significantly different among the 100 most abundant OTUs. Out of these 24 OTUs, 12 were more abundant in PS1 animals and 12 were more abundant in PS3 animals. See Table 3.1 for more details.

3.0 3.0

2.5

2.0

1.5

log10(p)

1.0

0.5

0.0 0.0

- -5 0 5 10 15 10 log2(FC) Figure 3.3. Serum metabolites in sows differing in perineal score (PS) and POP risk. This plot represents important features and demonstrates those which are different between sows with a PS3 (high risk for POP) compared to PS1 (low risk for POP) score. The pink circles represent features above the threshold (2-fold change and P < 0.05). PS1 averages were set as the base level and PS3 values of individual metabolites are represented by points. The pink dots represent the individual metabolites that are considered significantly different between PS. Pink dots in the upper righthand are metabolites that are more abundant, on average, in PS3 animals, and pink dots in the upper left hand less abundant, on average, in PS3 animals. Note both fold changes and P-values are log transformed. The further the position from the (0,0), the greater level of statistical significance for that the feature.

. 59

Figure 3.4. Differences in serum metabolites between PS1 and PS3 sows. This heat map displaying the 75 molecular features with the highest statistical significance (lowest p- values), illustrates relative abundance of serum metabolites (represented by color) between low POP risk. (PS1) and high POP risk (PS3) sows. PS3 animals (black) cluster to the left of the dendrogram separate from PS1 animals (grey).

C A B

Figure 3.5. Perineal scoring (PS) methodology as an indicator of pelvic organ prolapse (POP). The perineal region was visually evaluated for swelling, redness and protrusion to assign scores. (A) If a sow lacked swelling, redness and protrusion they were considered low risk for POP and assigned as PS1. (B) Sows with some characteristics such as moderate swelling, redness and protrusion of the perineal area were considered moderate risk for POP and assigned as PS2. (C) Sows demonstrating all of the characteristics of severe swelling, redness and protrusion of the perineal area were assigned to the PS3 category and considered high risk for POP. 61

Table 3.1. Differences in OTUs between vaginal microbiomes of PS1 and PS3 sows. OTU Taxonomy (Silva v132) NCBI BLAST Classification LDA2 PS3 P-value OTU 1 Veillonella Veillonella caviae PV1 4.08 1 0.03 OTU 3 Fusobacterium Fusobacterium gastrosuis CDW1 4.13 1 < 0.01 OTU 4 Prevotellaceae UCG-001 Duncaniella sp. TLL-A3 4.08 3 < 0.01 OTU 7 Parvimonas Parvimonas sp. KA00067 3.70 1 0.02 OTU 14 Bacteroidales unclassified Muribaculum sp. S4 3.82 3 0.04 OTU 22 Porphyromonas Porphyromonas somerae KA00683 3.27 1 < 0.01 OTU 28 Prevotellaceae NK3B31 group Prevotellaceae bacterium 3.63 3 0.03 OTU 31 Anaerococcus Anaerococcus tetradius 3.74 1 < 0.01 OTU 37 Streptococcus Streptococcus hyovaginalis TRG26 2.84 1 0.04 OTU 38 Treponema 2 Treponema bryantii, 3.47 3 0.02 OTU 43 Ruminococcus 1 Bacterium MA2007 3.36 3 0.02 OTU 46 Prevotellaceae UCG-001 Muribaculum sp. S4 3.23 3 < 0.01 OTU 49 Porphyromonas Porphyromonas katsikii JF5581 3.76 1 < 0.01 OTU 51 Lachnospiraceae XPB1014 group Absiella sp. strain 1XD42-72 2.99 3 0.05 OTU 52 Streptococcus Streptococcus gallolyticus 96L8 3.14 1 0.02 OTU 54 Ezakiella Clostridiales bacterium S9 PR-1 3.58 1 < 0.01 OTU 55 Gallicola Peptoniphilaceae bacterium SIT14 3.32 1 < 0.01 OTU 57 Porphyromonas Porphyromonadaceae bacterium FC4 3.29 1 < 0.01 OTU 58 Prevotellaceae unclassified Prevotella copri DSM 108494 3.45 3 0.03 OTU 70 Treponema 2 Candidatus Treponema suis 3.31 3 0.04 OTU 71 Murdochiella Levyella sp. Marseille-P3170 2.20 1 < 0.01 OTU 73 Treponema 2 Treponema porcinum 14V28 3.27 3 0.01 OTU 84 Streptococcus Streptococcus dysgalactiae PK 2.86 3 < 0.01 OTU 97 Lachnospiraceae unclassified Kineothrix alysoides KNHs209 2.92 3 < 0.01 1 Individual microbes were assigned in ordered of abundance and classified into operational taxonomic units (OTUs). 2 Linear discriminate analysis (LDA) was used for comparison between perineal scores. 3 Sows were assigned a perineal score (PS) based on their relative risk of experiencing a pelvic organ prolapse (POP). Sows assigned PS1 were presumed low risk for POP while sows assigned PS3 were presumed high risk for POP.

Table 3.2. Small molecule metabolites in serum from sows with differing risk of 1 Perineal Score 2 Metabolite PS1 PS3 Fold Change P-value D-Fructose 1.48E-11 5.80E-10 5.29 < 0.01 Pseudo uridine 3.93E-12 1.28E-10 5.02 < 0.01 L-Methionine 2.92E-11 7.92E-10 4.76 0.01 L-Alanine 2.51E-10 7.36E-09 4.87 0.02 2-Aminobutanoic acid 2.30E-12 8.18E-11 5.16 0.02 Ethanolamine 2.98E-12 1.04E-10 5.12 0.02 L-Glutamic acid 1.49E-11 5.03E-10 5.07 0.03 L-Proline 1.80E-10 5.12E-09 4.83 0.04 Glycine 3.19E-10 9.25E-09 4.86 0.04 L-Tryptophan 2.46E-11 6.53E-10 4.73 0.04 Myo-Inositol 2.47E-10 6.19E-09 4.65 0.04 Lactic Acid 1.75E-09 4.17E-08 4.57 0.04 Hippuric acid 9.31E-12 7.94E-11 3.09 0.04 Oleic Acid 7.71E-11 2.25E-09 4.87 0.05 D-Glucose 7.94E-10 1.97E-08 4.64 0.05 L-Asparagine 8.45E-12 2.47E-10 4.87 0.05 1 Sows were assigned a perineal score based on their relative risk of experiencing a pelvic organ prolapse (POP). Sows assigned PS1 were presumed low risk for POP while sows assigned PS3 were presumed high risk for POP. 2 Metabolite list only includes those that were identifiable and statistically significant (P < 0.05) between perineal scores. Values reported as moles per mg. 63

CHAPTER 4. VAGINAL MICROBIOTA VARIATION DURING LATE GESTATION COMMERCIAL SOWS A RISK FOR PELVIC ORGAN PROLAPSE

Zoë E. Kiefer1, Lucas R. Koester2,3, Jamie M. Studer1, Amanda L. Chipman4, Christine

Mainquist-Whigham5, Aileen F. Keating1, Stephan Schmitz-Esser 1,3, Jason W. Ross*,1,4

1 Department of Animal Science, Iowa State University, Ames, Iowa, United States

2 Department of Veterinary Microbiology and Preventive Medicine, Iowa State University

3 Interdepartmental Microbiology Graduate Program, Iowa State University

4 Iowa Pork Industry Center, Ames, Iowa

5 Pillen Family Farms, Columbus, NE

Grant Support: This project was supported in part by the National Pork Board and Foundation for Food and Agriculture Research.

Modified from a manuscript to be submitted to Biology of Reproduction

Abstract

Sow mortality in the U.S. swine industry has increased the last decade and is a cause of concern for pork producers. Approximately 21% of all sow deaths are attributed to pelvic organ prolapse (POP) during late gestation and lactation. In order to evaluate the biological basis to

POP, sows were assigned a perineal score (PS) throughout late gestation (weeks 12-15), and categorized into PS1 (low), PS2 (moderate), or PS3 (high), based on phenotypic observations 64 associated with putative POP outcomes. Recent evidence suggests microbial differences in the vagina may be related to POP risk. The objective of this study was to further evaluate the sow vaginal microbiota during late gestation to identify microbial population changes that may be associated with POP. Our hypothesis is that the vaginal microbiota would differ between sows with variable risk for POP during late gestation and that the changes in the microbiota during the last 3 weeks of gestation would be different for sows with differing risk levels for POP. Of the

2,864 individual sows scored on gestation week 15, 1.0%, 2.7%, and 23.4% of PS1, PS2, or PS3 sows, respectively, subsequently experienced POP. Vaginal swabs were utilized for 16S rRNA gene sequencing and differences (P < 0.05) in vaginal microbiota between PS1 and PS3 sows were detected on a community level along with differences (Q < 0.05) in 53 operational taxonomic units (OTU). Further, differences (Q < 0.05) in two OTUs were observed in PS3 sows

(high risk) that did or did not subsequently experience POP. Additional differences (Q < 0.05) were observed when comparing sows with low risk for POP at gestation week 12 and subsequently low or high risk at gestation week 15 suggesting that the vaginal microbial population of sows shifts during late gestation differently as a sow increases risk for POP.

Collectively, these data demonstrate that sows with greater POP risk have unique vaginal microflora and a further characterization of the sow vaginal microbiota was determined.

Introduction

The swine industry has made significant improvements in several areas of production efficiency, including reproductive performance, throughout the past decade. Unfortunately, during the same time sow mortality has increased with a substantial proportion of the amplified mortality being due to pelvic organ prolapse (POP). During a 2018 industry survey it was 65 discovered that approximately 21% of sow deaths in the U.S. could be attributed to POP (Ross,

2019). Sows most commonly experience POP during the peripartum period, which is defined as the period leading up to and shortly after farrowing (Supakorn et al., 2014). Commonly referred to as an anatomical disorder, POP is characterized as a condition where one or more of the pelvic organs presses up against or out of the vagina (Jelovsek et al., 2007). While this industry-wide problem is both an animal welfare concern and economic issue, there is currently a lack of understanding of the biological causes of POP preventing the development of effective mitigation strategies.

Microorganisms harbored on surfaces and in cavities of a host typically participate in a symbiotic relationship that can influence host health. Dysbiosis, or changes in the microbiota, particularly within the reproductive tract have been linked to reproductive dysfunction, and alterations may affect susceptibility to gynecological infections (Green et al., 2015). In humans, the vaginal microbiota can be influenced by sexual development, coitus, personal hygiene, menses, and hormones. During pregnancy, the human vaginal microbiota has been observed to increase in stability, suggesting a protective effect against infections during pregnancy (Romero et al., 2014; Greenbaum et al., 2019).

In humans, pelvic inflammatory disease (PID) and bacterial vaginosis (BV) are microbial infections of the reproductive tract that are linked to vaginal microbiota changes (Eschenbach et al., 1975; Spiegel, 1991). Symptoms of PID and BV include reproductive dysfunction such as discomfort, inflammation, and infertility, among others. The vaginal microbiota has also been studied in cattle and linked to reproductive disease (LeBlanc, 2008; Sheldon et al., 2009).

Unfortunately, as of now, there is limited knowledge on the vaginal microbiota of sows in general, and current studies have not probed discovery in specific relation to late gestation 66 commercial sows (Wang et al., 2017; He et al., 2020; Sanglard et al., 2020a; Sanglard et al.,

2020b). However, a previous study from our group demonstrated sows with different risk for

POP also had notable changes in the vaginal microbiota (Kiefer et al., 2021). In order to better understand changes in the vaginal microbiota and its association with POP risk, a need to further examine the vaginal microbiota across production systems, genetic lines, and geographical areas exists. Therefore, additional research is warranted to analyze if and how the vaginal microbiota changes during late gestation in sows preceding POP compared to those that do not experience

POP. Thus, the objective of this study was to determine sow vaginal microbiota changes in late gestation in relation to POP. To accomplish this objective, the current study tested the hypothesis that the vaginal microbiota would differ between sows with variable risk of POP during late gestation and that shifts in the microbial populations during the last 3 weeks of gestation would be different for sows with acquiring different risk levels of POP prior to parturition.

Materials and Methods

Animals

All experiments involving animals were approved by the Iowa State University (ISU)

Institutional Animal Care and Use Committee. This work was conducted on two commercial sow farms (designated Farm A and Farm B). The farms were part of the same production system with alike genetics, feed, and housing type, and were within one mile from each other. Additionally, farms had a similar health status being porcine reproductive and respiratory syndrome naïve,

Mycoplasma hyopneumoniae stable, porcine epidemic disease naïve, and influenza A virus stable.

Perineal scoring system

Using an already established perineal scoring system to assess risk of POP in late gestation sows (Kiefer et al., 2021), 3,035 pregnant sows (gestation weeks 12-15), from two different farms, were categorized during late gestation into three perineal score (PS) classes varying from presumed low to high risk for POP. Sows were scored, one time each week, from gestation week 12 through 15 (based on the week of breeding). Sows were scored only while lying, precluding some sows from being scored each week. Additionally, in a few instances, a small number of sows were removed from the herd prior to farrowing. As before (Kiefer et al.,

2021), sows presumed low risk for POP were assigned PS1, moderate risk assigned PS2, and high risk were designated PS3.

Sample collection

At both farms, vaginal swabs for microbiota DNA extraction were collected during gestation week 15 from sows classified as PS3 (wk15PS3; n = 121) or PS1 (wk15PS1; n = 101), which were matched for parity. In Farm B only, vaginal swabs were collected from a group of sows assigned PS1 at gestation week 12 (wk12PS1; n = 39). Vaginal swabs were collected by aseptically inserting a 7-inch histology brush (2199, Puritan Medical Products) into the vagina and brushing the vaginal orifice for approximately 15 seconds. Swabs were removed and immediately placed in sterile 1x phosphate buffered saline (PBS), kept on wet ice, vortexed for 5 minutes to detach cells, and centrifuged at 5,000 × g at room temperature for 15 minutes to form a pellet, supernatant was discarded, and the pellet was stored at -80⁰C until they were used for

DNA extraction.

DNA extraction

Tubes containing vaginal microbiota pellets were thawed, and DNA was extracted with a

Qiagen DNeasy Powerlyzer Powersoil kit (Qiagen, Germantown, MD) per manufacturer’s protocol. Mechanical cell lysis was performed using a Fisher Scientific Beadmill 24. The concentration of the isolated DNA was measured using a Qubit® 3.0 Fluorometer.

16S rRNA gene sequencing

16S rRNA gene sequences were amplified from the vaginal swab samples of individual sows. Sample DNA was diluted in sterile water to a concentration between 25 - 75 ng/µL then sent to the ISU DNA facility for sequencing using the Illumina MiSeq platform (Illumina, San

Diego, CA, USA). Briefly, the genomic DNA from each sample was amplified using Platinum™

Taq DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA) with one replicate per sample using universal 16S rRNA gene bacterial primers [515F (5′-

GTGYCAGCMGCCGCGGTAA-3′; 26) and 806R (5′-GGACTACNVGGGTWTCTAAT-3′;

27)] amplifying the variable region V4, as previously described (Kozich et al., 2013). All samples underwent PCR with an initial denaturation step at 94°C for 3 minutes, followed by 45 seconds of denaturing at 94°C, 20 seconds of annealing at 50°C, and 90 seconds of extension at

72°C. This was repeated for 35 total PCR cycles and finished with a 10 minute extension at

72°C. All the PCR products were then purified with the QIAquick 96 PCR Purification Kit

(Qiagen, Hilden, Germany) according to the manufacturer’s instructions. PCR bar-coded amplicons were mixed at equal molar ratios and used for Illumina MiSeq paired-end sequencing with 250 bp read length and cluster generation with 10% PhiX control DNA.

Sequence analysis

Sequence analysis was performed with Mothur V1.43.0 following the Mothur MiSeq

Standard Operating Procedure (Kozich et al., 2013). Barcode sequences, primers and low-quality sequences were trimmed using a minimum average quality score of 35, with a sliding window size of 50 bp. Chimeric sequences were removed with the “Chimera.vsearch” command. For alignment and taxonomic classification of operational taxonomic units (OTUs), the SILVA SSU

NR reference database (V138) provided by the Mothur website was used. Sequences were clustered into OTUs with a cutoff of 99% 16S rRNA gene similarity (=0.01 distance) and were ordered from most to least abundant. Representative sequences for the 50 most abundant OTUs were classified using NCBI BLAST, to improve classification accuracy. The OTU counts table and taxonomy assignments were then imported into Phyloseq (v1.34.0 (McMurdie and Holmes,

2013)) and Vegan (v2.5-5 (Oksanen et al., 2010)) for microbial sequence analysis.

Statistical analysis

Microbial sequence analysis

To compare alpha diversity between experimental groups, reads were randomly subsampled to accommodate the sample with the lowest number of reads across data sets (9,000 sequences). Measurements of Chao species richness, Shannon Diversity, and Simpson evenness were taken to compare community structures between experimental groups. The means of the experimental group alpha diversity measures were compared using a pooled t-test assuming equal variance.

To compare overall microbial community structure, samples were given a Bray-Curtis dissimilarity value and means were then compared using the permutational analysis of variance 70

(PERMANOVA) package built in to Phyloseq. Bray-Curtis dissimilarity was selected because of its ability to compare closely-related samples (Bray and Curtis, 1957).

All plotting was completed using the ggplot2, v2_3.1.1 graphing package (R: The R

Project for Statistical Computing) in R 4.0.3. Canonical analysis of Principle Coordinates (CAP) was used to visualize the variation capture by PS, Farm and the interaction between the two.

The absolute abundances of the 100 most abundant OTUs among samples were analyzed using a negative binomial distribution in GLIMMIX procedure of SAS (Version 9.4, SAS Inst.,

Cary, NC), and they were offset by the total library count for a given sample. Using the

MULTITEST procedure of SAS, all P-values were corrected for false discovery rates. The

PROC MIXED procedure of SAS was used to analyze diversity indices. Using Fisher’s Least

Significant Difference test, least square means were separated, and treatment differences were considered significant if P or Q values were ≤ 0.05. For the top 100 OTUs with a Q value of ≤

0.05, the Log2-fold change (log2FC) between treatment groups were calculated using R, and plotted using ggplot2. Three analysis were completed; (1) wk15 PS1 sows were compared to wk15PS3, (2) wk15PS3 sows that did experience POP compared to wk15PS3 sows that did not experience POP, and (3) gestation wk12PS1 sows compared to gestation wk15PS1 and wk15PS3 sows separately.

Effect of perineal score on subsequent pelvic organ prolapse

Effects of PS, Farm, and the interaction between PS and Farm, were assessed in SAS 9.4

(Cary, NC) utilizing a PROC MIXED analysis. Data are considered significant if P ≤ 0.05 and a tendency for biological meaning if 0.05 < P < 0.10.

Results

Changes in perineal score throughout late gestation

On farm observations demonstrated the earliest time point a sow with a PS3 designation were observable was approximately gestation week 12. In total, 1,563, 2,046, 2,492, and 2,864 sows during gestation weeks 12, 13, 14, and 15, respectively, were scored across both farms.

Sows assigned PS1 from gestation week 12 to 15 decreased in incidence from 85.3% to 54.8%, while PS3 assigned sows increased from 0.6% to 4.5% during the same time (Figure 4.1). Parity was influenced (P < 0.01) by PS with sows assigned PS1 having an average parity of 1.9 ± 0.05 and PS3 average parity being 3.3 ± 0.18.

Changes in perineal score during late gestation are associated with differing risk of POP

Of the 2,864 sows scored throughout late gestation, during gestation week 15; 1,570,

1,166 and 128 were assigned PS1, PS2 and PS3, respectively. As expected, a difference (P <

0.01) in POP rates was observed between PS1, PS2, and PS3 assigned sows, with 23.4% of PS3 sows subsequently experiencing POP while 1.0% and 2.7% of PS1 and PS2 sows, respectively, subsequently experienced POP (Chapter 5). With respect to Farm influence on POP prevalence,

1.6% of sows scored during week 15 at Farm A experienced POP compared to 3.7% of the sows at Farm B having the same outcome.

16S rRNA gene amplicon sequencing identified vaginal microbiota of late gestation sows

A total of 116,112 OTUs were obtained from 261 samples, however, 7,216 OTUs remained after quality control and removal of OTUs represented by less than ten sequences.

Average sequencing depth per sample was 21,291 sequences with a standard deviation of 5,906 72 sequences. Bacterial reads made up 99.6% of the reads while 0.34% were archaeal. Twenty- seven and 512 unique phyla and genera, respectively, were represented across the 7,216 OTUs used in this dataset. The 50 most abundant vaginal tract OTUs are reported in Table 4.1.

Differences in vaginal microbiota exist between sows at varying risk of POP during week 15 of gestation

When evaluating the average microbial community structure on whole-community level at gestation week 15, differences (P ≤ 0.05) were detected using PERMANOVA between PS,

Farm and the interaction of PS and Farm (Figure 4.2A). Alpha diversity estimators revealed no significant differences between samples regarding species richness, community evenness, and diversity (Figure 4.3A). When evaluating the 100 most abundant OTUs, significant differences for 51, 37, and 3 OTU’s between PS, Farm, and the interaction of PS and Farm, respectively, were observed. Of the OTUs that differed due to PS, 18 were more abundant in PS1 sows and 33 were more abundant in PS3 sows (Table 4.2, Figure 4.4). A few OTUs of interest include increases (Q ≤ 0.01) in Clostridium (OTU 3, 7, 18, 39, and 50), Streptococcus (OTU 4, and 32), and Treponema (OTU 47) in the PS3 sows. Staphylococcus agnetis (OTU 57) had the highest log2 fold change (log2FC) in microbes that were more abundant in PS3 sows. Farm also had an effect (Q ≤ 0.05) on the microbiota with 14 OTUs more abundant on Farm B and 23 more abundant on Farm A (data not shown). When evaluating the interaction of PS and Farm, OTU

25, 39, and 85 were different (Q ≤ 0.03).

For sows with increased risk of POP, those experiencing POP have differences in vaginal microbiota compared to those that did not

When evaluating the vaginal microbiota of PS3 sows at gestation week 15 that did and did not subsequently experience POP, there were observable differences (P < 0.01) on a whole microbial community based on POP outcome and Farm, but no interaction was detected (Figure

4.2B). Alpha diversity estimators revealed no significant differences between samples regarding species richness, community evenness and diversity (Figure 4.3B). When evaluating the 100 most abundant OTUs, only two OTUs (affiliated to the genera Actinobacillus and Veillonella) were observed to be increased (Q ≤ 0.01) in sows not experiencing POP compared to those that did (Table 4.3, Figure 4.4).

The vaginal microbiota community differs between sows at low and high risk for POP during gestation week 15 compared to sows at low risk for POP at gestation week 12

In efforts to understand the vaginal microbiota change over time, a subset of sows from

Farm B were sampled at 12 and 15 weeks. We evaluated wk12PS1 communities compared to wk15PS1 and wk15PS3 sows within Farm B only. There were observable differences (P < 0.01) between wk12PS1 and wk15PS1 as well as wk12PS1 and wk15PS3 comparisons based on CAP

(Figure 4.2C). Alpha diversity estimators revealed differences (P ≤ 0.03) between samples regarding species richness, community evenness and diversity (Figure 4.3C). Shannon and

Simpson revealed wk12PS1 tended to be different (P < 0.06) from wk15PS1 sows and was different (P = 0.01) from wk15PS3 sows.

When evaluating the top 100 OTUs, a total of 20 OTUs were different (P < 0.05, Q <

0.05) between wk12PS1 and wk15PS1 sows, of which 10 were more abundant in wk12PS1 and 74

10 more abundant in wk15PS1 (Table 4.4, Figure 4.4C). Streptococcus (OTU 4 and 32),

Porphyromonas (OTU 23 and 51), Staphylococcus (OTU 24), Gallicola (OTU 25 and 98),

Ezakiella (OTU 33), Peptococcus (OTU 45), and Peptoniphilus (OTU 45) were observed to be higher in abundance in wk12PS1 sows compared to wk15PS1. Conversely, Porphyromonas

(OTU 28 and 84), Staphylococcus (OTU 57), Anaerococcus (OTU 61 and 70), Enterococcus

(OTU 63), Campylobacter (OTU 76), Lachnospiraceae (OTU 78), Trueperella (OTU 89), and

Corynebacterium (OTU 94) were all higher in abundance in wk15PS1 sows.

A total of 40 OTUs were different (P < 0.05, Q < 0.05) between wk12PS1 and wk15PS3 sows, when evaluating the most abundant 100 OTUs. Of these, 23 OTUs were more abundant in wk12PS1 sows and 17 OTUs more abundant in wk15PS3 (Table 4.5, Figure 4.4D).

Streptococcus (OTU 4, 13, 32, and 48), Clostridium (OTU 5), Romboutsia (OTU 6),

Corynebacterium (OTU 10 and 53), Veillonella (OTU 12), Finegoldia (OTU 15),

Methanobrevibacter (OTU 22), Porphyromonas (OTU 23 and 51), Staphylococcus (OTU 24),

Gallicola (OTU 25), Escherichia (OTU 29), Ezakiella (OTU 33), Peptococcus (OTU 40),

Peptoniphilus (OTU 41 and 45), and Anaerococcus (OTU 54, 61, and 70) were more abundant in wk12PS1 compared to wk15PS3 scored sows. By comparison the vaginal microbiome of wk15PS3 sows had a higher abundance of Kurthia (OTU 20), Porphyromonas (OTU 28 and 84),

Anaerococcus (OTU 34), Corynebacterium (OTU 38, 74, and 94), Staphylococcus (OTU 57),

Enterococcus (OTU 63), Facklamia (OTU 72), Campylobacter OTU 76 and 100),

Lachnospiraceae (OTU 78), Peptostreptococcus (OTU 81), Trueperella (OTU 89),

Peptoniphilus (OTU 81), and Gallicola (OTU 98) compared to wk12PS1.

Discussion

The U.S. swine industry has experienced marked improvements in production within the last decade, however, an increasing mortality rate, of which 21% is due to POP, is an animal welfare and productivity concern. The POP prevalence is higher than reported for other livestock species and the biological underpinnings of why this may be occurring in commercial sow farms remain unclear. Developing successful mitigation strategies requires a better understanding of the biological events preceding POP. This study tested the hypothesis that the vaginal microbiota differs between sows with variable risk for POP during late gestation and that microbial population shifts during the last three weeks of gestation are altered for sows acquiring differing risk levels for POP as sows completed gestation. Understanding the vaginal microbiota and its relation to animal health and reproduction is an emerging field in swine physiology (Wang et al.,

2017; Sanglard et al., 2020a; Sanglard et al., 2020b; Kiefer et al., 2021).

In general, the limited knowledge of the vaginal microbiota of late gestation sows in commercial production systems makes it difficult to describe what is considered a normal and/or healthy microbial population; thus, one objective of the current study was to evaluate if findings from previous work could be validated (Kiefer et al., 2021). This study differs from Kiefer et al.

(2021) in that the two studies were conducted in different years, in different geographical locations, and with genetic lines of sows, management, and health statuses. None-the-less, similar observations were detected in assessing the microbial differences in sows differing in PS across the two studies thereby validating previous results (Kiefer et al., 2021). Comparing the 50 most abundant OTUs in both studies, 16 had identical NCBI blast classifications and included

Actinobacillus porcinus, Duncaniella, Porphyromonas somerae, Treponema bryantii, and

Veillonella caviae; all of which expressed similar changes in abundance in relation to POP risk. 76

Identifying OTUs with similar representation across both studies may provide novel description of the core vaginal microbiota of swine and more specifically the relationship between the vaginal microbiota with pregnancy and POP risk. While these studies expand the current characterization of the sow vaginal microbiome, a greater understanding is needed to strengthen associations with specific reproductive outcomes in sows.

When evaluating specific OTUs within the microbial community structure between sows at differing risk for POP, there were distinct differences. A number of Peptoniphilus OTUs were more abundant in PS1 scored sows. Although the function of Peptoniphilus in the vaginal microbiota is currently unknown, a higher abundance of Peptoniphilus was observed in the vaginal microbiota of thermoneutral pregnant sows compared to those experiencing heat stress

(He et al., 2020) demonstrating its prior detection in sows and differential abundance in response to environmental changes.

A higher abundance of several Porphyromonas OTUs including Porphyromonas somerae was noted in PS1 sows in this study confirming previous work (Kiefer et al., 2021) suggesting the plausibility of some Porphyromonas members could have a beneficial influence on the vaginal microbiota in late gestation sows. However, increases in Porphyromonas somerae can have a pathogenic effect in humans (Summanen et al., 2005), and have been linked to uterine disease in cattle (Jeon and Galvão, 2018). It is known that microbial populations and abundance can be altered depending on the host species and could also function differently within different host species as well as across different host tissues within a species (Chu et al., 2017).

Porphyromonas is an organism of interest because of its higher abundance in sows at low risk for

POP; however, having been described as pathogenic in other species indicates further 77 investigation into its function within the swine reproductive tract during late gestation is warranted.

Several Anaerococcus OTUs were also observed to be greater in the PS1 sows compared to those at high risk for POP. Interestingly, Anaerococcus was more abundant in the vagina of thermoneutral pregnant sows compared to sows exposed to heat stress suggesting thermal stress can alter specific vagina microbiota (He et al., 2020). Anaerococcus has also been identified in swine at other stages of production as well, being increased in the fecal microbiota of weaned piglets that tended to be heavier than their counterparts with lower Anaerococcus abundance

(Han et al., 2017). Observations of the Anaerococcus differences within the vaginal microbiota in sows with low risk for POP are conflicting with studies done in humans in relation to PID and

BV, emphasizing the need for better defining this microbe in sows and its function with respect to reproductive health.

While 20 OTUs were observed to be increased in low risk sows (PS1), 33 OTUs were in greater abundance in high-risk sows (PS3). Among those, several OTUs of the genus

Corynebacterium were found to be higher in PS3 sows. The function of Corynebacterium in the pig vaginal microbiota is, however, currently unexplored and warrants future investigation. In addition, several OTUs of Clostridium cluster I were detected to be more abundant in PS3 scored sows. Interestingly, while both the Corynebacterium and Clostridium cluster I OTUs were detected in in high abundance previously, they were not significantly different with respect to

POP risk (Kiefer et al., 2021). However, Clostridium cluster I has been demonstrated to be more abundant in the vaginal microbiota of sows with endometritis (Wang et al., 2017), suggesting a possible negative role of Clostridium cluster I with pig reproductive health. 78

Increases in Duncaniella, Streptococcus dysgalactiae , and Treponema bryantii were observed in sows assigned PS3 in the current study, which is consistent with the initial characterization of sows with elevated risk for POP (Kiefer et al., 2021). Duncaniella is a genus within the Prevotellaceae family and Prevotellaceae have been detected to be increased in the fecal microbiota of women suffering from gestational diabetes mellitus when blood sugar is increased (Fugmann et al., 2015). Interestingly, glucose and its derivatives are increased in circulation of sows with elevated risk of POP (Kiefer et al., 2021) indicating a possible biological relationship that could be further explored.

Mucin acts as a barrier to pathogens in the reproductive tract, and its degradation could lead to inflammation, potentially due to lipopolysaccharide (LPS) (Fugmann et al., 2015). Gram negative bacteria are known to produce LPS, which is a glycolipid surface molecule on most

Gram negative bacteria, and a well-known immune system stimulant (Bertani and Ruiz, 2018).

Interestingly, Prevotella is a species that can degrade mucin and thereby increase cellular permeability (Brown et al., 2011). Elevated Prevotella in the sow reproductive tract could theoretically increase permeability in the reproductive tract as well, which would be expected to elicit some level of an immunological response. In support of this posit, LPS binding protein

(LBP), a marker of inflammation, is increased in sows at high risk of POP (Chapter 5).

Furthermore, in the vaginal microbiota of women suffering from BV, a higher abundance in

Prevotella has also been noted strengthening the connection of this particular species of bacteria to compromised reproductive tract health (Ceccarani et al., 2019).

Streptococcus dysgalactiae is an additional microbe of particular interest, particularly as it can possess virulence factors, and is associated with several diseases in humans and animals

(Takahashi et al., 2011; Jensen and Kilian, 2012; Loubinoux et al., 2013; Rantala, 2014). 79

Virulence factors are products of bacteria that aid in eluding host defenses (Cross, 2008)

Streptococcus dysgalactiae is considered a pathogen in humans when found in the female genital tract and is associated with reproductive dysfunction in equids (Pinho et al., 2016; Baracco,

2019). Observations in this study, consistent with our previous work, have found Streptococcus dysgalactiae to be increased in the vaginal microbiota of late gestation sows at high risk of POP

(Kiefer et al., 2021) making it a noteworthy target for further investigation to decipher its potential role in affecting POP risk.

Like Streptococcus dysgalactiae, Treponema has also been linked to mammalian reproductive disorders (Rodrigues et al., 2015), as well as diseases that cause inflammation in the swine intestinal tract (Hughes et al., 1975). These findings demonstrating increased Treponema in the vaginal microbiota in PS3 scored sows is consistent with previous work (Kiefer et al.,

2021), and provides potential explanation of accompanied increases in biomarkers of inflammation (i.e. LPS binding protein) in PS3 sows (Chapter 5).

The OTU with the greatest log2FC increase in PS3 sows in this study was

Staphylococcus agnetis. Staphylococcus agnetis is an emerging pathogen in poultry and may have an effect on collagen and fibronectin (Al-Rubaye et al., 2015; Szafraniec et al., 2020).

Collagen plays a role in the structure of the muscles and tissue in the female pelvic floor and reproductive tract (Dhital et al., 2016). Thus, bacteria potentially capable of interfering with the host animal’s connective tissue may be a possible mechanism through which Staphylococcus agnetis could be associated with POP. This, however, will require verification in future studies focused on the mechanistic actions of this microbe.

Only Actinobacillus porcinus and Veillonella caviae were differentially abundant between sows at high risk for POP that subsequently did or did not experience POP. Both 80 microbes were greater in PS3 scored sows that did not subsequently experience POP compared to those that did. Veillonella has been detected in pigs previously and was observed in higher abundance in healthy pig fecal microbiota compared to those with intestinal diseases (Kraatz and

Tarast, 2008; Liu et al., 2015). In this study, wk15PS1 sows had greater Veillonella caviae presence compared to wk12PS1 and was also greater in wk12PS1 compared to wk15PS3 sows.

These observations collectively indicate that progression from week 12 of gestation to week 15 is accompanied increased Veillonella caviae in sows that remain low risk for POP, although this progression does not occur if a sow’s risk for POP increases during this same time period. Based on these observations and the consistency with prior work (Kiefer et al., 2021), Veillonella represents a potential beneficial microorganism warranting further exploration to determine potential roles in regulation of the microbiota as it relates to POP risk in sows.

The vaginal microbiota is known to change throughout gestation in humans (Greenbaum et al., 2019) and this may also happen in sows. In this study, the sample size was small for individual sows assigned PS3 during week 15 after previously being assigned PS1 during week

12, however, the results of this work suggest a temporal assessment of sows during gestation may be beneficial to better understand normal changes to vaginal microbial populations and shifts as reproductive disorders emerge.

Collectively this study validates the phenotypic PS system to identify sows at higher risk for POP. Bacterial candidates of interest were identified consistent with prior work and that may be associated with POP, in addition to providing further characterization of the vaginal microbiota of pregnant sows. These data aid in the understanding of the biological association leading up to POP in the U.S. commercial swine herd. Further, additional research that is 81 mechanistic by design is needed to demonstrate POP risk causality of specific microbes to continue moving this research area forward.

Acknowledgements

This project was supported by the National Pork Board and the Foundation for Food and

Agriculture Research grant #18-147.

Conflict of Interest

Authors contributions

ZEK designed experiments, completed experiments, performed analysis and wrote the draft of the manuscript. LRK contributed to execution of experiments, data analysis and edited the manuscript. JMS and ALC contributed to experimental design, execution of experiments, and edited the manuscript. CMW contributed to execution of experiment, and edited the manuscript.

AFK contributed to experimental design and edited the manuscript. SSE contributed to experimental design, data analysis and edited manuscript. JWR contributed to experimental design, project funding, provided project oversight and edited the manuscript.

References

Al-Rubaye, A. A. K., M. B. Couger, S. Ojha, J. F. Pummill, J. A. Koon, R. F. Wideman, and D. D. Rhoads. 2015. Genome Analysis of Staphylococcus agnetis, an Agent of Lameness in Broiler Chickens. PLoS One. 10:e0143336. doi:10.1371/journal.pone.0143336.

Baracco, G. J. 2019. Infections Caused by Group C and G Streptococcus (Streptococcus dysgalactiae subsp. equisimilis and Others): Epidemiological and Clinical Aspects. Microbiol. Spectr. 7. doi:10.1128/microbiolspec.gpp3-0016-2018.

Bertani, B., and N. Ruiz. 2018. Function and Biogenesis of Lipopolysaccharides. EcoSal Plus. 8. doi:10.1128/ecosalplus.esp-0001-2018.

Bray, J. R., and J. T. Curtis. 1957. An Ordination of the Upland Forest Communities of Southern Wisconsin. Ecol. Monogr. 27:325–349. doi:10.2307/1942268.

Brown, C. T., A. G. Davis-Richardson, A. Giongo, K. A. Gano, D. B. Crabb, N. Mukherjee, G. Casella, J. C. Drew, J. Ilonen, M. Knip, H. Hyöty, R. Veijola, T. Simell, O. Simell, J. Neu, C. H. Wasserfall, D. Schatz, M. A. Atkinson, and E. W. Triplett. 2011. Gut Microbiome Metagenomics Analysis Suggests a Functional Model for the Development of Autoimmunity for Type 1 Diabetes. PLoS One. 6:e25792. doi:10.1371/journal.pone.0025792.

Ceccarani, C., C. Foschi, C. Parolin, A. D’Antuono, V. Gaspari, C. Consolandi, L. Laghi, T. Camboni, B. Vitali, M. Severgnini, and A. Marangoni. 2019. Diversity of vaginal microbiome and metabolome during genital infections. Sci. Rep. 9:1–12. doi:10.1038/s41598-019-50410-x.

Cross, A. S. 2008. What is a virulence factor? Crit. Care. 12:196. doi:10.1186/cc7127.

Dhital, B., F. Gul-E-Noor, K. T. Downing, S. Hirsch, and G. S. Boutis. 2016. Pregnancy-Induced Dynamical and Structural Changes of Reproductive Tract Collagen. Biophys. J. 111:57– 68. doi:10.1016/j.bpj.2016.05.049.

Fugmann, M., M. Breier, M. Rottenkolber, F. Banning, U. Ferrari, V. Sacco, H. Grallert, K. G. Parhofer, J. Seissler, T. Clavel, and A. Lechner. 2015. The stool microbiota of insulin resistant women with recent gestational diabetes, a high risk group for type 2 diabetes. Sci. Rep. 5. doi:10.1038/srep13212. 83

Green, K. A., S. M. Zarek, and W. H. Catherino. 2015. Gynecologic health and disease in relation to the microbiome of the female reproductive tract. Fertil. Steril. 104:1351–1357. doi:10.1016/j.fertnstert.2015.10.010.

Han, G. G., J. Y. Lee, G. D. Jin, J. Park, Y. H. Choi, B. J. Chae, E. B. Kim, and Y. J. Choi. 2017. Evaluating the association between body weight and the intestinal microbiota of weaned piglets via 16S rRNA sequencing. Appl. Microbiol. Biotechnol. 101:5903–5911. doi:10.1007/s00253-017-8304-7.

He, J., W. Zheng, C. Tao, H. Guo, Y. Xue, R. Zhao, and W. Yao. 2020. Heat stress during late gestation disrupts maternal microbial transmission with altered offspring’s gut microbial colonization and serum metabolites in a pig model. Environ. Pollut. 266:115111. doi:10.1016/j.envpol.2020.115111.

Hughes, R., H. J. Olander, and C. B. Williams. 1975. Swine dysentery: pathogenicity of Treponema hyodysenteriae. Am. J. Vet. Res. 36:971–7.

Jelovsek, J. E., C. Maher, and M. D. Barber. 2007. Pelvic organ prolapse. Lancet. 369:1027– 1038. doi:10.1016/S0140-6736(07)60462-0.

Jensen, A., and M. Kilian. 2012. Delineation of Streptococcus dysgalactiae, its subspecies, and its clinical and phylogenetic relationship to Streptococcus pyogenes. J. Clin. Microbiol. 50:113–126. doi:10.1128/JCM.05900-11.

Jeon, S. J., and K. N. Galvão. 2018. An Advanced Understanding of Uterine Microbial Ecology Associated with Metritis in Dairy Cows. Genomics Inform. 16:e21. doi:10.5808/gi.2018.16.4.e21.

Kiefer, Z. E., L. R. Koester, L. Showman, J. M. Studer, A. L. Chipman, A. F. Keating, S. Schmitz-Esser, and J. W. Ross. 2021. Vaginal microbiome and serum metabolite differences in late gestation commercial sows at risk for pelvic organ prolapse. Sci. Rep. 11:6189. doi:10.1038/s41598-021-85367-3.

Kozich, J. J., S. L. Westcott, N. T. Baxter, S. K. Highlander, and P. D. Schloss. 2013. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the miseq illumina sequencing platform. Appl. Environ. Microbiol. 79:5112–5120. doi:10.1128/AEM.01043-13.

Kraatz, M., and D. Tarast. 2008. Veillonella magna sp. nov., isolated from the jejunal mucosa of a healthy pig, and emended description of Veillonella ratti. Int. J. Syst. Evol. Microbiol. 58:2755–2761. doi:10.1099/ijs.0.2008/001032-0.

LeBlanc, S. J. 2008. Postpartum uterine disease and dairy herd reproductive performance: A review. Vet. J. 176:102–114. doi:10.1016/j.tvjl.2007.12.019. 84

Liu, B., X. Yin, H. Yu, Y. Feng, X. Ying, J. Gong, and C. L. Gyles. 2015. Alteration of the microbiota and virulence gene expression in E. Coli O157:H7 in pig ligated intestine with and without AE lesions. PLoS One. 10. doi:10.1371/journal.pone.0130272.

Loubinoux, J., C. Plainvert, G. Collobert, G. Touak, A. Bouvet, and C. Poyart. 2013. Adult invasive and noninvasive infections due to Streptococcus dysgalactiae subsp. equisimilis in France from 2006 to 2010. J. Clin. Microbiol. 51:2724–2727. doi:10.1128/JCM.01262- 13.

McMurdie, P. J., and S. Holmes. 2013. phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. M. Watson, editor. PLoS One. 8:e61217. doi:10.1371/journal.pone.0061217.

Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, B. O’Hara, G. Simpson, P. Solymos, H. Stevens, and H. Wagner. 2010. Vegan: Community Ecology Package. In: R package version. Vol. 1.

Pinho, M. D., E. Erol, B. Ribeiro-Goncalves, C. I. Mendes, J. A. Carriço, S. C. Matos, S. Preziuso, A. Luebke-Becker, L. H. Wieler, J. Melo-Cristino, and M. Ramirez. 2016. Beta-hemolytic Streptococcus dysgalactiae strains isolated from horses are a genetically distinct population within the Streptococcus dysgalactiae taxon. Sci. Rep. 6. doi:10.1038/srep31736.

R: The R Project for Statistical Computing.

Rantala, S. 2014. Streptococcus dysgalactiae subsp. equisimilis bacteremia: An emerging infection. Eur. J. Clin. Microbiol. Infect. Dis. 33:1303–1310. doi:10.1007/s10096-014- 2092-0.

Romero, R., S. S. Hassan, P. Gajer, A. L. Tarca, D. W. Fadrosh, L. Nikita, M. Galuppi, R. F. Lamont, P. Chaemsaithong, J. Miranda, T. Chaiworapongsa, and J. Ravel. 2014. The composition and stability of the vaginal microbiota of normal pregnant women is different from that of non-pregnant women. Microbiome. 2:1–19. doi:10.1186/2049- 2618-2-4.

Ross, J. W. 2019. Identification of putative factors contributing to pelvic organ prolapse in sows (Grant # 17-224) II. Industry Summary.

Spiegel, C. A. 1991. Bacterial vaginosis. Clin. Microbiol. Rev. 4:485–502. doi:10.1128/CMR.4.4.485.

Summanen, P. H., B. Durmaz, M. L. Väisänen, C. Liu, D. Molitoris, E. Eerola, I. M. Helander, and S. M. Finegold. 2005. Porphyromonas somerae sp. nov., a pathogen isolated from humans and distinct from Porphyromonas levii. J. Clin. Microbiol. 43:4455–4459. doi:10.1128/JCM.43.9.4455-4459.2005.

Supakorn, C., J. D. Stock, C. Hostetler, and K. J. Stalder. 2014. Prolapse Incidence in Swine Breeding Herds Is a Cause for Concern. Open J. Vet. Med. 7:85–97. doi:10.4236/ojvm.2017.78009.

Szafraniec, G. M., P. Szeleszczuk, and B. Dolka. 2020. A review of current knowledge on staphylococcus agnetis in poultry. Animals. 10:1–19. doi:10.3390/ani10081421.

Takahashi, T., K. Ubukata, and H. Watanabe. 2011. Invasive infection caused by Streptococcus dysgalactiae subsp. equisimilis: Characteristics of strains and clinical features. J. Infect. Chemother. 17:1–10. doi:10.1007/s10156-010-0084-2.

Tables and Figures

100% PS1 90% 85.3% 81.2% PS2 80% PS3 69.5% 70% 60% 54.8% 50% 40.7% 40% 29.5% 30% 18.3% 20% 14.1%

Percent of total sows sows scoredoftotalPercent 10% 4.5% 0.6% 0.5% 1.0% 0% 12 13 14 15 Gestation Week

Figure 4.1. Perineal score (PS) changes throughout late gestation. Perineal scoring was conducted on sows during late gestation. Sows assigned a PS1, PS2 and PS3 were presumed low, medium and high risk, respectively, for pelvic organ prolapse. A total of 1,563, 2,046, 2,492, and 2,864 sows were scored weekly during gestation weeks 12, 13, 14, and 15, respectively. Percentage of sows assigned PS2 or PS3 increased from gestation week 12 to 15 while the percentage of sows assigned PS1 decreased.

Figure 4.2. Vaginal microbial community comparisons. Canonical analysis of principal coordinates (CAP) showing the maximum variation of beta-diversity between vaginal microbiota communities. (A) CAP from sows with assumed low (PS1, n = 101) or high (PS3, n= 121) risk for pelvic organ prolapse (POP) during gestation week 15 (day 108-115) from two separate farms. Statistical differences (P < 0.05) were detected in overall microbial communities between PS, Farm and the interaction of PS and Farm at gestation week 15. (B) CAP from a subset of sows that subsequently experienced POP that also were scored PS3 (Yes n = 28, No n = 93). Statistical differences (P < 0.05) were detected in overall microbial communities between POP outcome, and Farm. (C) CAP using sows from Farm B only, at gestation week 12 scored PS1 (wk12PS1, n = 39), gestation week 15 PS1 (wk15PS1, n = 61), and gestation week 15 PS3 (wk15PS3, n = 78). Statistical differences (P < 0.05) were detected in overall microbial communities between gestation week, score and the interaction of week and score. All sample points were based on Bray-Curtis dissimilarities of the overall composition of microbial communities.

Figure 4.3. Alpha diversity of the vaginal microbiota for sows during late gestation. The above graphics compare alpha diversity measurements species evenness (Simpson), richness (TSO, Chao1) and diversity (Shannon) across the different variables within this study. (A) Alpha diversity measurements for sow vaginal microbiota at low (PS1 (grey points, n = 101)) or high (PS3 (black points, n= 121)) risk for pelvic organ prolapse (POP) at gestation week 15 (day 108-115). (B) Alpha diversity measurements for sows that subsequently experienced POP that also were scored PS3 (Yes (black points, n = 28), No (grey points, n = 93)). (C) Alpha diversity measurements for sows, only from farm B, at gestation week 12 scored PS1 (wk12PS1; light blue points, n = 39), gestation week 15 PS1 (wk15PS1; grey points, n = 61), and gestation week 15 PS3 (wk15PS3; black points, n = 78). Shannon and Simpson diversity were observed to be significantly different (P ≤ 0.03) with differences between week 12 sows and both week 15 PS1 and PS3 sows.

Figure 4.4. Abundance of OTUs in the vaginal microbiota of late gestation sows. Bar color denotes which treatment group the OTU is more abundant in. Comparisons are shown for: (A) OTUs different between sows with assumed low (PS1 (gray bars; n = 101)) or high (PS3 (black bars; n = 121)) risk for pelvic organ prolapse (POP) at gestation week 15 (day 108-115) from two separate farms. (B) OTUs from a subset of sows that were scored PS3 that did (Yes; n = 28) or did not (No (black bars; n = 93)) subsequently experience POP. (C) OTUs from farm B, at gestation week 12 scored PS1 (wk12PS1 (blue bars; n = 39)) compared gestation week 15 PS1 sows (wk15PS1 (grey bars; n = 61)). (D) OTUs from farm B, at gestation week 12 scored PS1 (wk12PS1 (blue bars; n = 39)) and week 15 PS3 sows (wk15PS3 (black bars; n = 78)). Only log2 fold changes of OTUs with significantly different abundance between comparisons are shown.

Table 4.1. The 50 most abundant bacteria in vaginal samples from sows NCBI BLAST Relative abundance 2 OTU1 (%) Phylum Taxonomy (Silva v138)3 Classification Similarity (%) Accession no. OTU 1 6.62 Proteobacteria Actinobacillus Actinobacillus porcinus strain 35NTS 99.6 FJ437063.1 OTU 2 5.83 Firmicutes Turicibacter Turicibacter sanguinis strain MOL361 99.6 CP053187.1 OTU 3 4.90 Firmicutes Clostridium_sensu_stricto_1 Clostridium cellulovorans strain 22rA 98.4 KF528156.1 Streptococcus dysgalactiae subsp. equisimilis OTU 4 4.74 Firmicutes Streptococcus strain TPCH-A88 99.6 CP053074.1 OTU 5 4.38 Firmicutes Clostridium_sensu_stricto_1 Clostridium moniliforme strain 2055 99.2 NR_104892.1 OTU 6 4.27 Firmicutes Romboutsia Romboutsia timonensis strain DR1 97.6 NR_144740.1 OTU 7 3.08 Firmicutes Clostridium_sensu_stricto_1 Clostridium chauvoei strain SBP 98.8 CP027286.1 OTU 8 2.84 Firmicutes Anaerococcus Anaerococcus prevotii DSM 20548 99.6 CP001708.1 Terrisporobacter petrolearius strain OTU 9 2.41 Firmicutes Terrisporobacter LAM0A37 99.2 NR_137408.1 OTU 10 2.14 Actinobacteriota Corynebacterium Corynebacterium xerosis strain GS 100 CP032788.1

90 OTU 11 1.41 Fusobacteriota Fusobacterium Fusobacterium gastrosuis strain CDWK1 99.6 NR_146837.2

OTU 12 1.27 Firmicutes Veillonella Veillonella caviae strain PV1 99.2 NR_025762.1 OTU 13 1.17 Firmicutes Streptococcus Streptococcus suis strain SS-CLA1926B 99.6 MT256098.1 OTU 14 1.14 Firmicutes Terrisporobacter Terrisporobacter glycolicus strain RD-1 98.8 MN733184.1 OTU 15 1.05 Firmicutes Finegoldia Finegoldia magna strain FDAARGOS 99.2 CP054000.1 OTU 16 1.04 Firmicutes Nosocomiicoccus Staphylococcaceae bacterium NML 99-ST-011 100 AY841364.1 OTU 17 1.04 Firmicutes Parvimonas Parvimonas micra strain KCOM 2339 98.4 MT982357.1 OTU 18 1.00 Firmicutes Clostridium_sensu_stricto_1 Clostridium butyricum strain 5467 100 MT510294.1 OTU 19 0.88 Firmicutes Peptostreptococcus Peptostreptococcus sp. DSM 106284 100 MN537513.1 OTU 20 0.87 Firmicutes Kurthia Kurthia gibsonii strain EMB4 99.6 KY048434.1 OTU 21 0.81 Bacteroidota Bacteroides Bacteroides massiliensis V081 96.4 LC515611.1 OTU 22 0.80 Euryarchaeota Methanobrevibacter Methanobrevibacter sp. YE315 100 CP010834.1 OTU 23 0.78 Bacteroidota Porphyromonas Porphyromonas levii DSM 23370 90.5 NR_113089.1 OTU 24 0.77 Firmicutes Staphylococcus Staphylococcus simulans strain D14 99.6 MT568571.1 OTU 25 0.75 Firmicutes Gallicola Peptoniphilaceae bacterium SIT14 97.6 LN870299.1 OTU 26 0.74 Actinobacteriota Corynebacterium Corynebacteirum maris strain Coryn-1 99.6 NR_121700.2 OTU 27 0.71 Firmicutes Anaerococcus Anaerococcus prevotii DSM 20548 98.4 CP001708.1 OTU 28 0.67 Bacteroidota Porphyromonas Porphyromonas somerae strain KA00683 99.2 KP192301.1

Table 4.1 Continued OTU 29 0.65 Proteobacteria Escherichia-Shigella Escherichia fergusonii strain SPK 99.6 MW624513.1 OTU 30 0.59 Firmicutes Facklamia Facklamia tabacinasalis strain GF112B 99.6 NR_026482.1 OTU 31 0.59 Proteobacteria Actinobacillus Actinobacillus rossii strain JF2167 99.6 AY465365.1 OTU 32 0.58 Firmicutes Streptococcus Streptococcus pasteurianus strain 2323 99.6 MT604782.1 OTU 33 0.57 Firmicutes Ezakiella Bacteroides coagulans strain EUH 581-73 94.5 NR_104900.1 OTU 34 0.55 Firmicutes Anaerococcus Anaerococcus sp. strain AGMB00486 99.6 MT568623.1 OTU 35 0.49 Actinobacteriota Corynebacterium Corynebacterium phoceense strain JZ R-177 99.6 MH119724.1 OTU 36 0.47 Firmicutes Lactobacillus Lactobacillus amylovorus strain 5081 99.6 MT459395.1 OTU 37 0.47 Bacteroidota Porphyromonas Porphyromonas endodontalis 96.1 LT680662.1 Corynebacterium stationis strain VRD1 333 OTU 38 0.46 Actinobacteriota Corynebacterium N7 99.6 MN840614.1 OTU 39 0.46 Firmicutes Clostridium_sensu_stricto_1 Clostridium perfringens strain 3116 99.6 MT613499.1 OTU 40 0.45 Firmicutes Peptococcus Peptococcus simiae strain M108 99.6 NR_153710.1 OTU 41 0.45 Firmicutes Peptoniphilus Peptoniphilus sp. strain AGMB00490 99.6 MT396160.1 OTU 42 0.40 Actinobacteriota Corynebacterium Corynebacterium amycolatum strain 1MR 99.6 MT423433.1

91 OTU 43 0.40 Bacteroidota Prevotellaceae_UCG-001 Duncaniella freteri strain TLL-A3 84.6 NR_170509.1

Jeotgalicoccus schoeneichii strain 140805- OTU 44 0.38 Firmicutes Staphylococcaceae_unclassified STR-02 99.6 NR_151981.1 OTU 45 0.38 Firmicutes Peptoniphilus Peptoniphilus sp. 1804121828 99.6 MK945758.1 OTU 46 0.37 Actinobacteriota Corynebacterium Corynebacterium pollutisoli strain VDS11 99.6 NR_151947.1 OTU 47 0.37 Spirochaetota Treponema Treponema bryantii 99.6 AB849328.1 OTU 48 0.37 Firmicutes Streptococcus Streptococcus hyovaginalis strain TRG26 99.6 MH329638.1 OTU 49 0.35 Firmicutes Peptoniphilus Peptoniphilus harei strain DCW_SL_25A 99.6 MK424030.1 OTU 50 0.35 Firmicutes Clostridium_sensu_stricto_1 Clostridium cellulovorans strain 22rA 99.6 KF528156.1 1Individual microbes were assigned in order of abundance and classified into operational taxonomic units (OTUs). 2 Relative abundance of the specific OTU in the vaginal swabs collected in this study. 3 Taxonomy was assigned using SILVA SSU NR reference database (V138).

Table 4.2. Differences in OTUs between vaginal microbiota of PS1 and PS3 sows during gestation week 15. NCBI BLAST OTU1 Taxonomy (silva v138)2 Classification Similarity (%) PS3 Log2FC4 Q-value OTU 2 Turicibacter Turicibacter sanguinis strain MOL361 99.6 PS3 0.32 < 0.01 OTU 3 Clostridium_sensu_stricto_1 Clostridium cellulovorans strain 22rA 98.4 PS3 0.31 < 0.01 Streptococcus dysgalactiae subsp. equisimilis strain OTU 4 Streptococcus 99.6 PS3 0.34 < 0.01 TPCH-A88 OTU 5 Clostridium_sensu_stricto_1 Clostridium moniliforme strain 2055 99.2 PS3 0.36 < 0.01 OTU 6 Romboutsia Romboutsia timonensis strain DR1 97.6 PS3 0.32 < 0.01 OTU 7 Clostridium_sensu_stricto_1 Clostridium chauvoei strain SBP 98.8 PS3 0.31 < 0.01 OTU 9 Terrisporobacter Terrisporobacter petrolearius strain LAM0A37 99.2 PS3 0.30 < 0.01 OTU 10 Corynebacterium Corynebacterium xerosis strain GS 100 PS3 0.72 < 0.01 OTU 14 Terrisporobacter Terrisporobacter glycolicus strain RD-1 98.8 PS3 0.32 < 0.01 OTU 16 Nosocomiicoccus Staphylococcaceae bacterium NML 99-ST-011 100 PS3 0.42 < 0.01 OTU 18 Clostridium_sensu_stricto_1 Clostridium butyricum strain 5467 100 PS3 0.34 < 0.01 92

OTU 20 Kurthia Kurthia gibsonii strain EMB4 99.6 PS3 0.35 0.01 OTU 22 Methanobrevibacter Methanobrevibacter sp. YE315 100 PS3 0.38 < 0.01 OTU 23 Porphyromonas Porphyromonas levii DSM 23370 90.5 PS1 1.21 < 0.01 OTU 26 Corynebacterium Corynebacteirum maris strain Coryn-1 99.6 PS3 0.57 < 0.01 OTU 27 Anaerococcus Anaerococcus prevotii DSM 20548 98.4 PS3 0.57 < 0.01 OTU 28 Porphyromonas Porphyromonas somerae strain KA00683 99.2 PS3 2.07 < 0.01 OTU 30 Facklamia Facklamia tabacinasalis strain GF112B 99.6 PS3 0.32 0.02 OTU 32 Streptococcus Streptococcus pasteurianus strain 2323 99.6 PS3 0.36 < 0.01 OTU 33 Ezakiella Bacteroides coagulans strain EUH 581-73 94.5 PS1 2.10 < 0.01 OTU 34 Anaerococcus Anaerococcus sp. strain AGMB00486 99.6 PS1 1.09 < 0.01 OTU 35 Corynebacterium Corynebacterium phoceense strain JZ R-177 99.6 PS3 0.49 0.02 OTU 40 Peptococcus Peptococcus simiae strain M108 99.6 PS1 1.66 < 0.01 OTU 41 Peptoniphilus Peptoniphilus sp. strain AGMB00490 99.6 PS1 1.89 < 0.01 OTU 43 Prevotellaceae_UCG-001 Duncaniella freteri strain TLL-A3 84.6 PS3 0.45 0.01 OTU 44 Staphylococcaceae_unclassified Jeotgalicoccus schoeneichii strain 140805-STR-02 99.6 PS3 0.42 0.01 OTU 45 Peptoniphilus Peptoniphilus sp. 1804121828 99.6 PS1 0.71 < 0.01 OTU 46 Corynebacterium Corynebacterium pollutisoli strain VDS11 99.6 PS3 0.58 0.01

Table 4.2 Continued OTU 47 Treponema Treponema bryantii 99.6 PS3 0.61 < 0.01 OTU 50 Clostridium_sensu_stricto_1 Clostridium cellulovorans strain 22rA 99.6 PS3 0.34 < 0.01 OTU 51 Porphyromonas Porphyromonas canoris strain JCM 16132 92.5 PS1 2.36 < 0.01 OTU 53 Corynebacterium Corynebacterium callunae strain AS67 99.2 PS3 0.49 0.02 OTU 54 Anaerococcus Anaerococcus sp. Marseille-P3915 98.4 PS1 2.45 < 0.01 OTU 55 Corynebacterium Corynebacterium urealyticum strain 2431 99.6 PS3 0.72 < 0.01 OTU 57 Staphylococcus Staphylococcus agnetis strain PR5962A 99.6 PS3 2.81 < 0.01 OTU 60 Atopostipes Atopostipes sp. strain ZH16 99.6 PS3 0.40 0.01 OTU 61 Anaerococcus Anaerococcus nagyae strain ENR0686 94.8 PS1 1.21 < 0.01 OTU 62 Lachnospiraceae_UCG-007 Lachnotalea glycerini strain DLD10 96.8 PS3 0.33 0.01 OTU 66 Bifidobacterium Bifidobacterium pseudolongum A1 100 PS3 0.49 < 0.01 OTU 70 Anaerococcus Anaerococcus sp. Marseille-P3915 99.6 PS1 2.13 < 0.01 OTU 71 Lactobacillus Lactobacillus vaginalis strain 17465 99.6 PS3 0.41 < 0.01 OTU 72 Facklamia Facklamia hominis strain DNF00119 98.4 PS1 1.19 < 0.01

OTU 73 Firmicutes_unclassified Clostridium oryzae strain KC3 90.1 PS3 0.47 < 0.01 93

Corynebacterium glucuronolyticum strain V17 1.78 OTU 74 Corynebacterium 2011556 99.6 PS1 < 0.01 OTU 76 Campylobacter Campylobacter ureolyticus strain LMG 6451 96.5 PS1 2.46 0.01 OTU 81 Peptostreptococcus Peptostreptococcus anaerobius strain WH7 91.3 PS1 1.97 < 0.01 OTU 84 Porphyromonas Porphyromonas asaccharolytica strain HA3347-27 99.6 PS1 1.50 0.02 OTU 89 Trueperella Trueperella pyogenes strain TN2 99.6 PS1 1.07 0.01 OTU 96 Peptoniphilus Peptoniphilus olsenii strain WAL 12922 97.2 PS1 1.62 < 0.01 OTU 97 Facklamia Facklamia hominis strain DNF00119 98.4 PS3 0.46 0.04 OTU 98 Gallicola Gallicola sp. RM-6 92.5 PS1 2.75 < 0.01 1Individual microbes were assigned in ordered of abundance and classified into operational taxonomic units (OTUs). 2 Taxonomy was assigned using SILVA SSU NR reference database (V138). 3Sows were assigned a perineal score (PS) based on their relative risk of experiencing a pelvic organ prolapse (POP). Sows assigned PS1 were presumed low risk for POP and sows assigned PS3 were presumed high risk for POP. Specific OTUs are more abundant in sows with the PS indicated. 4Log2 Fold change.

Table 4.3. Differences in OTUs between vaginal microbiota of sows during gestation week 15 assigned PS3 that subsequently did or did not experience pelvic organ prolapse. NCBI BLAST POP OTU1 Taxonomy (silva v138)2 Classification Similarity (%) Outcome3 Log2FC4 Q-value OTU 1 Actinobacillus Actinobacillus porcinus strain 35NTS 99.6 No 0.88 < 0.01 OTU 12 Veillonella Veillonella caviae strain PV1 99.2 No 1.22 0.01 1Individual microbes were assigned in ordered of abundance and classified into operational taxonomic units (OTUs). 2Taxonomy was assigned using SILVA SSU NR reference database (V138). 3POP outcome refers to whether a perineal score 3 sows at high risk for pelvic organ prolapse (POP) did (Yes) or did not (No) subsequently experience POP. Specific OTUs are more abundant in outcome designated. 4Log2 fold change.

Table 4.4. Differences in OTUs between vaginal microbiota of gestation week 12 PS1 sows and gestation week 15 PS1 sows.

NCBI BLAST wk12PS1 vs OTU1 Taxonomy (silva v138)2 Classification Similarity (%) wk15PS13 Log2FC4 P-value Q- value Streptococcus dysgalactiae subsp. equisimilis OTU 4 Streptococcus strain TPCH-A88 99.6 wk12PS1 0.77 < 0.01 < 0.01 OTU 23 Porphyromonas Porphyromonas levii DSM 23370 90.5 wk12PS1 0.83 0.02 < 0.01 OTU 24 Staphylococcus Staphylococcus simulans strain D14 99.6 wk12PS1 1.21 0.03 0.05 OTU 25 Gallicola Peptoniphilaceae bacterium SIT14 97.6 wk12PS1 0.50 0.02 < 0.01 OTU 28 Porphyromonas Porphyromonas somerae strain KA00683 99.2 wk15PS1 1.00 < 0.01 < 0.01 OTU 32 Streptococcus Streptococcus pasteurianus strain 2323 99.6 wk12PS1 0.47 < 0.01 0.01 OTU 33 Ezakiella Bacteroides coagulans strain EUH 581-73 94.5 wk12PS1 1.13 0.02 < 0.01 OTU 40 Peptococcus Peptococcus simiae strain M108 99.6 wk12PS1 0.89 < 0.01 < 0.01 OTU 45 Peptoniphilus Peptoniphilus sp. 1804121828 99.6 wk12PS1 0.49 < 0.01 < 0.01 OTU 51 Porphyromonas Porphyromonas canoris strain JCM 16132 92.5 wk12PS1 1.37 < 0.01 < 0.01

OTU 57 Staphylococcus Staphylococcus agnetis strain PR5962A 99.6 wk15PS1 3.48 0.01 < 0.01 OTU 61 Anaerococcus Anaerococcus nagyae strain ENR0686 94.8 wk15PS1 0.76 0.01 < 0.01 OTU 63 Enterococcus Enterococcus faecium strain FFNL3053 99.6 wk15PS1 1.18 < 0.01 < 0.01 OTU 70 Anaerococcus Anaerococcus sp. Marseille-P3915 99.6 wk15PS1 0.74 0.02 < 0.01 OTU 76 Campylobacter Campylobacter ureolyticus strain LMG 6451 96.5 wk15PS1 2.27 < 0.01 < 0.01 OTU 78 Lachnospiraceae_XPB1014_group Lachnospiraceae bacterium CA63 94.1 wk15PS1 0.51 < 0.01 0.02 Porphyromonas asaccharolytica strain OTU 84 Porphyromonas HA3347-27 99.6 wk15PS1 1.23 0.03 < 0.01 OTU 89 Trueperella Trueperella pyogenes strain TN2 99.6 wk15PS1 0.92 0.02 < 0.01 OTU 94 Corynebacterium Corynebacterium kutscheri strain NCTC3655 99.6 wk15PS1 0.40 0.03 0.03 OTU 98 Gallicola Gallicola sp. RM-6 92.5 wk12PS1 1.57 0.01 < 0.01 1Individual microbes were assigned in ordered of abundance and classified into operational taxonomic units (OTUs). 2Taxonomy was assigned using SILVA SSU NR reference database (V138). 3Sows were assigned a perineal score (PS) based on their relative risk of experiencing a pelvic organ prolapse (POP). Sows assigned PS1 were presumed low risk for POP and compared between gestation week 12 (wk12PS1) and 15 (wk15PS1). Specific OTUs are more abundant in sows with the week and PS indicated. 4Log2 fold change.

Table 4.5. Differences in OTUs between vaginal microbiota of gestation week 12 PS1 sows and gestation week 15 PS3 sows.

NCBI BLAST wk12PS1 vs OTU1 Taxonomy (silva v138)2 Classification Similarity (%) wk15PS33 Log2FC4 P-value Q-value Streptococcus dysgalactiae subsp. OTU 4 Streptococcus equisimilis strain TPCH-A88 99.6 wk12PS1 0.96 < 0.01 < 0.01 OTU 5 Clostridium_sensu_stricto_1 Clostridium moniliforme strain 2055 99.2 wk12PS1 0.34 < 0.01 0.03 OTU 6 Romboutsia Romboutsia timonensis strain DR1 97.6 wk12PS1 0.31 0.01 0.04 OTU 10 Corynebacterium Corynebacterium xerosis strain GS 100 wk12PS1 0.33 0.05 0.02 OTU 12 Veillonella Veillonella caviae strain PV1 99.2 wk12PS1 0.41 0.02 0.01 OTU 13 Streptococcus Streptococcus suis strain SS-CLA1926B 99.6 wk12PS1 0.60 < 0.01 0.03 OTU 15 Finegoldia Finegoldia magna strain FDAARGOS 99.2 wk12PS1 0.41 0.02 0.01 OTU 20 Kurthia Kurthia gibsonii strain EMB4 99.6 wk15PS3 0.46 < 0.01 0.03 OTU 22 Methanobrevibacter Methanobrevibacter sp. YE315 100 wk12PS1 0.39 < 0.01 0.03 OTU 23 Porphyromonas Porphyromonas levii DSM 23370 90.5 wk12PS1 2.84 < 0.01 < 0.01

OTU 24 Staphylococcus Staphylococcus simulans strain D14 99.6 wk12PS1 1.61 0.01 0.05 OTU 25 Gallicola Peptoniphilaceae bacterium SIT14 97.6 wk12PS1 1.01 < 0.01 < 0.01 OTU 28 Porphyromonas Porphyromonas somerae strain KA00683 99.2 wk15PS3 2.96 < 0.01 < 0.01 OTU 29 Escherichia-Shigella Escherichia fergusonii strain SPK 99.6 wk12PS1 0.72 < 0.01 < 0.01 OTU 32 Streptococcus Streptococcus pasteurianus strain 2323 99.6 wk12PS1 0.54 < 0.01 0.01 OTU 33 Ezakiella Bacteroides coagulans strain EUH 581-73 94.5 wk12PS1 3.77 < 0.01 < 0.01 OTU 34 Anaerococcus Anaerococcus sp. strain AGMB00486 99.6 wk15PS3 1.74 < 0.01 < 0.01 Corynebacterium stationis strain VRD1 333 OTU 38 Corynebacterium N7 99.6 wk15PS3 0.36 0.01 0.01 OTU 40 Peptococcus Peptococcus simiae strain M108 99.6 wk12PS1 2.65 < 0.01 < 0.01 OTU 41 Peptoniphilus Peptoniphilus sp. strain AGMB00490 99.6 wk12PS1 3.93 < 0.01 < 0.01 OTU 45 Peptoniphilus Peptoniphilus sp. 1804121828 99.6 wk12PS1 1.14 < 0.01 < 0.01 OTU 48 Streptococcus Streptococcus hyovaginalis strain TRG26 99.6 wk12PS1 0.68 0.02 < 0.01 OTU 51 Porphyromonas Porphyromonas canoris strain JCM 16132 92.5 wk12PS1 4.29 < 0.01 < 0.01 OTU 53 Corynebacterium Corynebacterium callunae strain AS67 99.2 wk12PS1 0.44 0.01 0.01 OTU 54 Anaerococcus Anaerococcus sp. Marseille-P3915 98.4 wk12PS1 3.86 < 0.01 < 0.01 OTU 57 Staphylococcus Staphylococcus agnetis strain PR5962A 99.6 wk15PS3 5.87 < 0.01 < 0.01 OTU 61 Anaerococcus Anaerococcus nagyae strain ENR0686 94.8 wk12PS1 2.29 < 0.01 < 0.01

Table 4.5 Continued OTU 63 Enterococcus Enterococcus faecium strain FFNL3053 99.6 wk15PS3 1.01 < 0.01 < 0.01 OTU 70 Anaerococcus Anaerococcus sp. Marseille-P3915 99.6 wk12PS1 3.09 < 0.01 < 0.01 OTU 72 Facklamia Facklamia hominis strain DNF00119 98.4 wk15PS3 1.91 < 0.01 < 0.01 Corynebacterium glucuronolyticum strain OTU 74 Corynebacterium V17 2011556 99.6 wk15PS3 2.82 < 0.01 < 0.01 OTU 76 Campylobacter Campylobacter ureolyticus strain LMG 6451 96.5 wk15PS3 5.30 < 0.01 < 0.01 OTU 78 Lachnospiraceae_XPB1014_group Lachnospiraceae bacterium CA63 94.1 wk15PS3 0.51 0.01 0.02 OTU 81 Peptostreptococcus Peptostreptococcus anaerobius strain WH7 91.3 wk15PS3 2.89 < 0.01 < 0.01 Porphyromonas asaccharolytica strain OTU 84 Porphyromonas HA3347-27 99.6 wk15PS3 2.79 < 0.01 < 0.01 OTU 89 Trueperella Trueperella pyogenes strain TN2 99.6 wk15PS3 2.22 < 0.01 < 0.01 Corynebacterium kutscheri strain OTU 94 Corynebacterium NCTC3655 99.6 wk15PS3 0.57 < 0.01 0.03 OTU 96 Peptoniphilus Peptoniphilus olsenii strain WAL 12922 97.2 wk15PS3 3.07 < 0.01 < 0.01 OTU 98 Gallicola Gallicola sp. RM-6 92.5 wk15PS3 4.84 < 0.01 < 0.01 Campylobacter corcagiensis strain LMG 97 OTU 100 Campylobacter 27932 99.6 wk15PS3 5.26 < 0.01 < 0.01 1Individual microbes were assigned in ordered of abundance and classified into operational taxonomic units (OTUs). 2Taxonomy was assigned using SILVA SSU NR reference database (V138). 3Sows were assigned a perineal score (PS) based on their relative risk of experiencing a pelvic organ prolapse (POP). Sows assigned PS1 were presumed low risk for POP and sows assigned PS3 were presumed high risk of POP. OTUs were compared between low risk sows at gestation week 12 (wk12PS1) to high risk sows at gestation week15 (wk15PS3). Specific OTUs are more abundant in sows with the week and PS indicated in the column. 4Log2 fold change.

CHAPTER 5. CIRCULATING BIOMARKERS ASSOCIATED WITH PELVIC

ORGAN PROLASPE RISK IN LATE GESATION SOWS

Zoë E. Kiefer*, Jamie M. Studer*, Amanda L. Chipman†, Malavika K. Adur*, Christine

Mainquist-Whigham‡, Nicholas K. Gabler*, Aileen F. Keating*, Jason W. Ross*,†,

*Department of Animal Science, Iowa State University, Ames, IA, 50011, United States

‡Pillen Family Farms, Columbus NE 685601, United States

†Iowa Pork Industry Center, Ames IA 50011, United States

1This project was supported by the National Pork Board and the Foundation for Food and

Agriculture Research.

Modified from a manuscript submitted to Journal of Animal Science

Abstract

Sow mortality, as the result of pelvic organ prolapse (POP), has been increasing the last decade in the U.S. swine industry. The objective of this study was to identify potential biological markers associated with risk of POP in sows. We hypothesized that sows differing in perineal score (PS) from PS1-PS3 (PS1 - a presumed low POP risk; PS2 - a presumed moderate POP risk; and PS3 - a presumed high POP risk) would differ in circulatory biomarkers of inflammation and hormonal profiles. On gestation week 15, 2,864 individual sows were assigned a PS, and 99 subsequently, 1.0, 2.7, and 23.4% of PS1, PS2, or PS3 sows, respectively, experienced POP.

During PS assignment at week 15 of gestation, blood samples were collected from sows of two farms of similar genetics, feed sources, and health status. Whole blood was subjected to complete blood count (CBC) analysis (n = 212) and steroid hormones were measured in serum from a subset (n = 110) of animals. Lipopolysaccharide binding protein (LBP), tumor necrosis factor alpha (TNF-α), haptoglobin, C-reactive protein (CRP), and creatine kinase (CK) levels were also evaluated. Complete blood count analysis revealed decreased (P ≤ 0.03) mean platelet values (4.2%), lymphocytes (7.8%), and monocytes (6.8%) in PS3 compared to PS1 sows.

Increased (P ≤ 0.05) abundance of androstenedione (10.7%), androsterone (18.1%), estrone

(21.2%), 17β-estradiol (23.3%), and testosterone (14.8%) was observed in PS3 compared to PS1 sows. Additionally, a 21.3% increase (P = 0.06) in LBP in PS3 compared to PS1 sows was observed. Collectively, these data indicate differences in circulating biomarkers of inflammation and hormone profiles exist between groups of sows differing in POP risk.

Keywords: biomarkers, pelvic organ prolapse, sow, swine

Introduction

Sow survivability and longevity is reduced as the result of pelvic organ prolapse (POP), which incidence rates have been consistently increasing in the U.S. swine industry over the past decade (Supakorn et al., 2014a). Our group previously conducted a large, U.S. industry-wide survey demonstrating that 21% of all sow mortality is due to POP. The phenotype of POP is characterized by one or more pelvic organs (uterus, rectum and/or vagina) shifting from their normal location within the pelvic area and descending from the body cavity (Jelovsek et al., 100

2007). Affected sows typically experience POP around the time of parturition, which almost invariably results in the euthanasia of sows and often in the loss of offspring, furthering the economic losses and animal welfare concerns of this biological phenomenon. Since the biological causes of POP in swine are not well-understood, a significant barrier exists in the development of strategies to mitigate POP.

The perineal area of a sow during late gestation can phenotypically be red and swollen in comparison to other stages of gestation, as parturition looms. It has been shown that sows at greater relative POP risk have an abnormal amount of swelling, indicating a potential inflammatory response (Kiefer et al., 2021). Characterization of a late gestation perineal swelling phenotype has associated edema in the perineal area with POP and a scoring system can distinguish sows with low and high risk of POP (Kiefer et al., 2021).

With respect to immune modulating factors, the vaginal environment is neither constant or well defined in most species (Farage et al., 2010) and the immune cell repertoire in the lower genital tract differs between species and during different phases of the reproductive cycle

(Jacques Ravel et al., 2011; Chu et al., 2017). This may be partly influenced by commensal bacteria colonization of the lower genital tract that can affect immune responses (Mirmonsef et al., 2011). The epithelial cells of the female tract express receptors for factors involved in immune responses, such as toll-like receptors (TLR) and major histocompatibility complex

(MHC) molecules, which help identify, process and initiate cellular and humoral immune responses (Mirmonsef et al., 2011). These epithelial cells can, upon activation, produce a variety of cytokines and chemokines that aid in recruiting and activating cells of the immune system in the female reproductive tract (Franklin and Kutteh, 1999).

Understanding how differences in the vaginal microflora relate to alterations in steroid 101 hormones and/or immune system modulation through interactions with cells mediating immunological responses remains an important area for discovery, and differential vaginal microflora populations have been discovered in sows with low and high POP risk (Kiefer et al.,

2021). This is of significance, as one function of the unique mucosal immune system within the female reproductive tract is to identify, and tolerate or eliminate pathological microbes (Franklin and Kutteh, 1999). Further, fluctuations of immune responses and inflammation in the reproductive tract can occur through differing mechanisms, and are regulated by the sex steroid hormones 17β-estradiol (E2) and progesterone (P4) (Vegeto et al., 1999; Rakasz and Lynch,

2002; García-Gómez et al., 2013; García-Gómez et al., 2020). In addition to immune cell migration to the reproductive tract, there is a transudation of immunoglobulins from the blood through the extracellular matrix (García-Gómez et al., 2020).

Therefore, the objectives of this study were to evaluate differences in inflammatory markers and steroid hormone concentrations in sows with high and low POP risk. To accomplish this, we hypothesized that high risk POP sows in late gestation would have increased serum biomarkers associated with inflammation and/or immune modulation, in addition to alterations in steroid hormone profiles compared to low risk sows.

Materials and Methods

Animals

All experiments involving animals were approved by the Iowa State University

Institutional Animal Care and Use Committee. All animals were individually housed in commercial sow farms located throughout the Midwest U.S.

102

Perineal scoring system

Sows were of commercial genetics and from two different farms approximately one mile from each other, populated from the same gilt multiplier, fed identically formulated diets from the same feed mill, and housed and managed similarly. Additionally, farms had the same health status of porcine reproductive and respiratory syndrome (PRRS) naïve, Mycoplasma hyopneumoniae (MHP) stable, porcine epidemic disease (PED) naïve, influenza A virus (IAV) stable. Utilizing an established perineal scoring (PS) system, 2,864 pregnant sows (gestation days 107-116), from two different farms, were categorized during late gestation into three PS categories of presumed low to high POP risk (Kiefer et al., 2021). To minimize variation, all animals were scored as PS1, PS2 or PS3, by the same two individuals, while the sow was laying down. It is important to note that a PS3 sow is considered abnormal for any stage of gestation.

Following scoring, sows were monitored for subsequent POP incidence.

Sample Collection

On a subset of sows, at the time of PS assignment, biological samples were collected from sows classified as PS3 (n = 127) along with a parity matched PS1 sow (n = 103). Blood samples were collected by jugular venipuncture into multiple vacutainer tubes (Becton,

Dickinson and Company, Franklin Lakes, NJ). Blood (3 mL) collected into EDTA tubes was used for complete blood count (CBC) analysis. Blood samples (10 mL) were also collected and allowed to coagulate at room temperature before separation of serum by centrifugation at 20 ⁰C for 15 minutes at 2,000 × g followed by aspiration and storage at -80 ⁰C until further analysis.

103

Complete blood count analysis

Whole blood collected into EDTA tubes was used for CBC analysis. Samples were submitted from PS1 (n = 90) and PS3 (n = 122) animals to the Iowa State University Clinical

Pathology laboratory (Ames, IA) for CBC with automated differential analysis.

Steroid hormone analysis

Serum samples from PS3 scored sows (n = 60) and parity matched PS1 scored sows (n =

50) were quantified using a AbsoluteIDQ®Stero17 Kit from Biocrates (Innsbruck, Austria) from a 17-panel quantification of endogenous steroids (Koal et al., 2012). This panel included deoxycorticosterone, deoxycortisol, hydroxyprogesterone, aldosterone, androstenedione, androsterone, corticosterone, cortisol, cortisone, estrone, β-estradiol, progesterone, and testosterone. Sample preparation procedure was based on solid phase extraction technique in a

96-well plate format, which is necessary for cleaning and concentrating of the target steroid hormones. The HPLC-MS/MS analysis was conducted in multiple reaction monitoring mode using a Waters Xevo™ TQ-S-micro (Waters, Vienna, Austria). For quantification of all 17 steroid analytes, 7-point external calibration curves and 13 stable isotope-labeled internal standards were used as previously described (Koal et al., 2012).

Analysis of markers of inflammation

Serum from PS1 and PS3 sows was used to quantify cytokine and acute phase protein

(APP) markers associated with inflammation via ELISA. Lipopolysaccharide binding protein

(LBP) ELISA (Cat. # HK503, HycultBiotech, Plymouth Meeting, PA) was used to measure LBP.

The detection range for the assay was 1.6 to 100 ng/mL and the inter-assay CV was 12.21%. 104

Porcine tumor necrosis factor alpha (TNF-α) quantikine ELISA (Cat. # PTA00, R&D Systems,

Minneapolis, MN) was used to measure TNF-α. The detection range for this assay was 23.4 to

1,500 pg/mL and the inter-assay CV was 11.02%. Porcine haptoglobin ELISA (Cat. # 41-

HAPPO-E01, Alpco Ltd., Salem, NH). The detection range for this assay was 25 to 400 ng/mL and the inter-assay CV was 16.4%. Porcine C-reactive protein (CRP) ELISA (Cat. # 41-CRPPO-

E01, Alpco Ltd., Salem, NH) was used to measure CRP. The detection range for this assay was

6.25 to 200 ng/mL and the inter-assay CV was 6.18%. All ELISA assays were conducted in accordance with the manufacturer’s protocols. Serum aliquots from a subset of sows, based on serum availability, for those scoring PS1 (n = 47) and PS3 (n = 55) were also submitted to the

Iowa State University Clinical Pathology laboratory (Ames, IA) for quantification of creatine kinase (CK) using a dry chemistry method via a VITROS 5000.

Statistical analysis

Main effects of PS, Farm, and the interaction between PS and Farm were evaluated in

SAS 9.4 (Cary, NC) utilizing the PROC MIXED procedure. Data are represented as least square means and considered significant if P ≤ 0.05 and a tendency for biological meaning if 0.05 < P <

0.10. Comparisons between individual effects are presented as differences of least squared means.

Results

Differences in perineal score is associated with differing risk of POP

Of the 2,864 sows scored during late gestation, 1,570, 1,166 and 128 were assigned PS1,

PS2 and PS3, respectively. A difference (P < 0.01) in POP rate was observed between PS1, PS2, 105 and PS3, with 1.0% of PS1, 2.7% of PS2, and 23.4% of PS3 scored sows, subsequently experiencing POP (Figure 5.1). Distribution of PS1, PS2, and PS3 at Farm A was 55.5%, 41.3%, and 3.2% of the population, respectively. Similarly, distribution of PS at Farm B was 54.1%,

40.1% and 5.7% of sows scored as PS1, PS2 and PS3, respectively. Additionally, Farm A had a

1.6% POP rate while Farm B had a 3.7% POP rate during the duration of this study. For all of the sows used in the study, average parity was 2.2 ± 0.04 and average gestation day was 112.3 ±

0.03 at the time of PS assignment. Sow parity differed (P < 0.01) between PS1 and PS3 scored sows with an average parity of 1.9 and 3.3 ± 0.18, respectively.

Complete blood count parameters differ in sows with high and low POP risk

During CBC analysis, a total of 16 blood parameters were evaluated and reported in

Table 5.1. Farm A had increased (P ≤ 0.04) red blood cells, hematocrit, eosinophils and basophils by 4.0%, 2.6%, 10.5%, and 15.0%, respectively, when compared to Farm B. Farm B had an increase of 1.9% in mean corpuscular hemoglobin and 0.6% increase in mean corpuscular hemoglobin concentration (P ≤ 0.05) compared to Farm A. Hemoglobin and neutrophils tended to be increased in Farm A (P ≤ 0.08) by 2.1% and 8.2%, respectively. There was an effect of PS on mean platelet volume (MPV), lymphocytes and monocytes (P ≤ 0.03) with an increase of

4.2%, 7.8%, and 6.8%, respectively, in PS1 compared to PS3 sows. There tended to be a 10.5% increase (P = 0.09) in eosinophils in PS1 compared to PS3 sows. Farm A had a 25.6% increase in eosinophils (P < 0.01) in PS1 compared to PS3 sows and a tendency for increased basophils

(19.2%; P = 0.07) in PS1 compared to PS3.

106

Steroid hormones in serum are differ between sows with high and low POP risk

Serum steroid hormone concentrations are reported in Table 5.2. Etiocholanolone and dihydrotestosterone were undetectable in all samples. Dehydroepiandrosterone-sulfate was detected in only three sows from Farm B and was therefore not analyzed. Of the remaining 14 hormones, Farm A had 15.5%, 14.3%, and 19.6% greater levels of androstenedione, estrone and

β-estradiol in serum, respectively (P ≤ 0.03; Table 5.2). There was a tendency for Farm A to have 19.2 and 29.7% lower serum levels of deoxycorticosterone and aldosterone, respectively (P

≤ 0.08). Serum androstenedione, androsterone, estrone, β-estradiol, and testosterone levels were

10.7%, 18.1%, 21.2%, 23.3%, and 14.8% higher (P ≤ 0.05; Table 5.2), respectively, in PS3 compared to PS1 sows. Deoxycortisol and aldosterone tended (P ≤ 0.09) to have a 27.3% and

27.2% increase, respectively, in PS3 compared to PS1 sows. No significant Farm by PS interactions were detected for any of the hormones evaluated (Table 5.2).

Acute phase protein and cytokine profiles in sows identified as high risk for POP

Circulating concentrations of functional LBP tended to be 21.3% higher (P = 0.06) in

PS3 compared to PS1 sows (Table 5.3). When comparing PS3 to PS1 sows, no difference in

TNF-α concentrations (56.5 pg/mL ± 7.1) were observed between PS, Farm, or an interaction between PS and Farm (P ≥ 0.34; Table 5.3). Serum haptoglobin concentrations (928.3 ± 138.4) were also not different between PS, Farm, or an interaction between PS and Farm (P ≥ 0.11).

Serum CRP concentrations (31.4 µg/mL ± 5.4) tended (P = 0.07) to be different between Farms, but were not observed to be different between PS or the interaction between PS and Farm. Farm

A had 58.0% higher (P < 0.01) serum CK concentrations compared to Farm B, however, there was no observable difference (P ≥ 0.27) between PS or an interaction of PS and Farm on CK concentration (Table 5.3). 107

Discussion

Although the biological explanation for POP remains unknown it is critical that mitigation strategies be developed which are predicated on understanding the biological causes contributing to POP since the incidence of POP currently accounts for approximately 21% of sow mortality. Sows during late gestation were selected based upon their putative risk for POP using an established scoring strategy (Kiefer et al., 2021) and evaluated for blood markers commonly associated with inflammation and immune system function.

An important component for this study was that the PS system employed was effective in assessing relative risk of POP, as sows scored PS3 had a markedly higher prolapse rate compared to those assigned PS1, recapitulating our previous findings (Kiefer et al., 2021). The observation that sows with an elevated risk of POP also experience decreases in circulating lymphocytes and monocytes could be explained by the reallocation of immune cells to target tissues, which has been observed previously (Kratofil et al., 2017). Monocyte recruitment plays a critical role in the host’s defense by promoting inflammation or facilitating tissue repair (Kratofil et al., 2017). An additional observation from the CBC analysis was that MPV was decreased in PS3 compared to

PS1. Previous studies have associated changes in MPV with inflammatory diseases, such as inflammatory bowel disease, in humans and swine (Järemo and Sandberg-Gertzen, 1996;

Kapsoritakis et al., 2001). In humans a negative correlation between MPV and Crohn’s disease is also reported (Shah et al., 1989). The observations with alterations in lymphocyte abundance and

MPV between sows with differing POP risk have also been reported to have an effect on cytokines and their receptors (Kapsoritakis et al., 2001; D’Ambrosio et al., 2002). The drastic difference in phenotype along with observable differences in immune cell populations between 108 high and low risk sows led us to investigate an inflammatory response in PS3 sows. These finding suggest that sows at greater risk for POP are experiencing an inflammatory response.

After penetration of protective barriers, bacterial endotoxin can lead to immune activation

(Schromm et al., 2021), and phagocytes, such as monocytes, have a role in the pathophysiology of inflammation and associated cytokine production (Heine et al., 1999). The observations in high risk sows having elevated LBP may help explain, at least in part, the significant swelling and presumed local inflammation observed in PS3 sows. When tissues are injured by bacteria, trauma, toxins, heat or various other reasons, cells release chemicals that cause swelling as part of the inflammatory response (Mirmonsef et al., 2011). Tumor necrosis factor α is an inflammatory mediator rapidly secreted in response to LPS (Takashiba et al., 1999) and increases during the initial phases of an inflammatory response (Petersen et al., 2004). Given that the timing of sample collection in this study was presumably well past initial exposure to potential elicitors of inflammation, this could explain why there was no difference in TNF-α between sows with differing in POP risk. These results support that a bacterial infection may be associated with the inflammatory response observed in PS3 sows.

Creatine kinase is an enzyme found in skeletal muscle and other tissues, and is a biomarker found in blood following muscle and tissue damage or muscle usage, such as exercise

(Brancaccio et al., 2007; Brancaccio et al., 2010). Serum CK in this study was evaluated with respect to POP risk to determine if CK was a potential marker of the presumed tissue and muscle damage experienced by sows at high risk of POP. Interestingly, CK was not different between sows at low and high risk for POP, but was significantly different between farms. Further, and contrary to our expectations, the farm with the greater level of POP, during the project period and historically, had lower serum CK on average. Why CK could be different between farms is 109 puzzling given the similarity of the genetics of the sows, facilities and nutrition. Further investigation would be required to better understand if the differences observed between farms can be attributed to anything specific and/or if these differences have biological meaning and application.

Haptoglobin is also a modulator of the inflammatory response due to LPS exposure and regulates monocyte activation (Arredouani et al., 2005). Samples were collected at a single time point, and it is difficult to pinpoint the onset of an inflammatory reaction, since the temporal pattern of the immune response cannot be evaluated. Similarly, CRP is a rapid reaction APP and serum levels have previously been observed to increase in response to inflammation (Gabay and

Kushner, 1999; Petersen et al., 2004). The acute-phase response to inflammation can include fever and hormonal changes (Petersen et al., 2004).

Specifically, the production of pro-inflammatory cytokines that LBP binds to can stimulate the production of APP, and activate the hypothalamic-pituitary-adrenal (HPA) axis

(Steel and Whitehead, 1994; Ulevitch and Tobias, 1995; Nordgreen et al., 2018) and steroid hormone levels have been observed to be altered in response to inflammation (García-Gómez et al., 2013). With a PS3 sow phenotypic observations reveal significant local inflammation, however it is not known if the alteration in steroid hormones are a cause or effect of POP. An increase in both LBP and steroid hormones were observed in the PS3 compared to PS1 sows.

These data support that there is a potential connection between the increased LPB levels in PS3 sows, increased circulating hormones, and the inflammatory response observed in PS3 sows.

Immune responses, as well as changes in the microbiome have been associated with fluctuations in steroid hormones (García-Gómez et al., 2013). Interestingly, distinct differences in the vaginal microbiome are present in sows at high risk for POP (Kiefer et al., 2021) and 110 increased steroid hormones in PS3 sows were observed in this study. The communication between sex steroids and the host microbiome has the potential to determine the outcome of an infection (Hughes and Sperandio, 2008) and P4, E2, and testosterone can alter the immune response against bacterial infection (Vegeto et al., 1999; García-Gómez et al., 2013). In general, estrogens upregulate proinflammatory cytokines while testosterone has been observed to decrease this response (Ahmed et al., 1985), although the mechanism through which this is accomplished is not established. Another potential beneficial role of E2 is that it has been demonstrated to protect immune cells against apoptosis during in vitro culture, specifically TNF-

α induced apoptosis (Sorachi et al., 1993; Vegeto et al., 1999). Modulation of the immune system by steroid hormones may control pro- and anti-inflammatory cytokine expression and regulate the activity of immune cells, specifically lymphocytes (García-Gómez et al., 2013), which were increased in high risk for POP sows. In addition, E2 promotes a pro-inflammatory response and is increased in humans with endometriosis, a disease characterized by excessive inflammation and increased CYP19A1 activity (García-Gómez et al., 2020) and E2 was increased in PS3 compared to PS1 sows. The production of androsterone, estrone and E2 are all catalyzed by CYP19A1 and were increased in PS3 compared to PS1 sows, which could indicate that CYP19A1 protein level or function could be associated with POP risk. E2 levels have been observed to increase in humans who have experienced POP (Bai et al., 2002).

Conclusion

This study further validated a phenotypic PS system to effectively categorize sows by putative risk for POP. Additionally, serum biomarkers associated with inflammation and/or immune modulation, in addition to differences in steroid hormone profiles, exist between sows 111 with different relative risk of POP. Collectively, discovery of these differences in circulating biomarkers of inflammation and steroid hormones is an important step in contributing to the understanding of the biological basis of POP in the U.S. swine herd.

Acknowledgements

The authors would like to acknowledge the support from the farm staff who supported the efforts of this study.

Disclosures

References

Ahmed, S. A., W. J. Penhale, and N. Talal. 1985. Sex hormones, immune responses, and autoimmune diseases. Mechanisms of sex hormone action. Am. J. Pathol. 121:531–551.

Arredouani, M. S., A. Kasran, J. A. Vanoirbeek, F. G. Berger, H. Baumann, and J. L. Ceuppens. 2005. Haptoglobin dampens endotoxin-induced inflammatory effects both in vitro and in vivo. Immunology. 114:263–271. doi:10.1111/j.1365-2567.2004.02071.x.

Bai, S. W., B. H. Jung, B. C. Chung, S. U. Kim, J. Y. Kim, K. H. Rha, S. K. Kim, and K. H. Park. 2002. Steroid hormone metabolism in women with pelvic organ prolapse - PubMed. J. Reprod. Med. 2002 Apr;47(4):303-8. PMID: 12012882.

Brancaccio, P., G. Lippi, and N. Maffulli. 2010. Biochemical markers of muscular damage. Clin. Chem. Lab. Med. 48:757–767. doi:10.1515/CCLM.2010.179.

Brancaccio, P., N. Maffulli, and F. M. Limongelli. 2007. Creatine kinase monitoring in sport medicine. Br. Med. Bull. 81–82:209–230. doi:10.1093/bmb/ldm014. 112

D’Ambrosio, D., C. Albanesi, R. Lang, G. Girolomoni, F. Sinigaglia, and C. Laudanna. 2002. Quantitative Differences in Chemokine Receptor Engagement Generate Diversity in Integrin-Dependent Lymphocyte Adhesion. J. Immunol. 169:2303–2312. doi:10.4049/jimmunol.169.5.2303.

Farage, M. A., K. W. Miller, and J. D. Sobel. 2010. Dynamics of the Vaginal Ecosystem— Hormonal Influences. Infect. Dis. Res. Treat. 3:IDRT.S3903. doi:10.4137/idrt.s3903.

Gabay, C., and I. Kushner. 1999. Acute-Phase Proteins and Other Systemic Responses to Inflammation. N. Engl. J. Med. 340:448–454. doi:10.1056/nejm199902113400607.

García-Gómez, E., B. González-Pedrajo, and I. Camacho-Arroyo. 2013. Role of sex steroid hormones in bacterial-host interactions. Biomed Res. Int. 2013. doi:10.1155/2013/928290.

García-Gómez, E., E. R. Vázquez-Martínez, C. Reyes-Mayoral, O. P. Cruz-Orozco, I. Camacho- Arroyo, and M. Cerbón. 2020. Regulation of Inflammation Pathways and Inflammasome by Sex Steroid Hormones in Endometriosis. Front. Endocrinol. (Lausanne). 10:935. doi:10.3389/fendo.2019.00935.

Heine, H., R. L. Delude, B. G. Monks, T. Espevik, and D. T. Golenbock. 1999. Bacterial lipopolysaccharide induces expression of the stress response genes hop and H411. J. Biol. Chem. 274:21049–21055. doi:10.1074/jbc.274.30.21049.

Hughes, D. T., and V. Sperandio. 2008. Inter-kingdom signalling: Communication between bacteria and their hosts. Nat. Rev. Microbiol. 6:111–120. doi:10.1038/nrmicro1836.

Järemo, P., and H. Sandberg-Gertzen. 1996. Platelet density and size in inflammatory bowel disease. Thromb. Haemost. 75:560–561. doi:10.1055/s-0038-1650321.

Jelovsek, J. E., C. Maher, and M. D. Barber. 2007. Pelvic organ prolapse. Lancet. 369:1027– 1038. doi:10.1016/S0140-6736(07)60462-0.

Kapsoritakis, A. N., M. I. Koukourakis, A. Sfiridaki, S. P. Potamianos, M. G. Kosmadaki, I. E. Koutroubakis, and E. A. Kouroumalis. 2001. Mean platelet volume: A useful marker of inflammatory bowel disease activity. Am. J. Gastroenterol. 96:776–781. doi:10.1016/S0002-9270(00)02414-X. 113

Koal, T., D. Schmiederer, H. Pham-Tuan, C. Röhring, and M. Rauh. 2012. Standardized LC- MS/MS based steroid hormone profile-analysis. J. Steroid Biochem. Mol. Biol. 129:129– 138. doi:10.1016/j.jsbmb.2011.12.001.

Kratofil, R. M., P. Kubes, and J. F. Deniset. 2017. Monocyte Conversion During Inflammation and Injury. Arterioscler. Thromb. Vasc. Biol. 37:35–42. doi:10.1161/ATVBAHA.116.308198.

Nordgreen, J., C. Munsterhjelm, F. Aae, A. Popova, P. Boysen, B. Ranheim, M. Heinonen, J. Raszplewicz, P. Piepponen, A. Lervik, A. Valros, and A. M. Janczak. 2018. The effect of lipopolysaccharide (LPS) on inflammatory markers in blood and brain and on behavior in individually-housed pigs. Physiol. Behav. 195:98–111. doi:10.1016/j.physbeh.2018.07.013.

Petersen, H. H., J. P. Nielsen, and P. M. H. Heegaard. 2004. Application of acute phase protein measurements in veterinary clinical chemistry. Vet. Res. 35:163–187. doi:10.1051/vetres:2004002.

Rakasz, E., and R. G. Lynch. 2002. Female sex hormones as regulatory factors in the vaginal immune compartment. Int. Rev. Immunol. 21:497–513. doi:10.1080/08830180215016.

Shah, A., G. Morgan, J. D. R. Rose, R. Fifield, and J. Rhodes. 1989. Platelet number and size in relation to serum orosomucoid concentration in Crohn’s disease. Med. Lab. Sci. 46:79– 80.

Sorachi, K. ichi, S. Kumagai, M. Sugita, J. Yodoi, and H. Imura. 1993. Enhancing effect of 17β- estradiol on human NK cell activity. Immunol. Lett. 36:31–35. doi:10.1016/0165- 2478(93)90065-A. 114

Steel, D. M., and A. S. Whitehead. 1994. The major acute phase reactants: C-reactive protein, serum amyloid P component and serum amyloid A protein. Immunol. Today. 15:81–88. doi:10.1016/0167-5699(94)90138-4.

Supakorn, C., J. D. Stock, C. Hostetler, and K. J. Stalder. 2014. Prolapse Incidence in Swine Breeding Herds Is a Cause for Concern. Open J. Vet. Med. 7:85–97. doi:10.4236/ojvm.2017.78009.

Ulevitch, R. J., and P. S. Tobias. 1995. Receptor-Dependent Mechanisms of Cell Stimulation by Bacterial Endotoxin. Annu. Rev. Immunol. 13:437–457. doi:10.1146/annurev.iy.13.040195.002253.

Vegeto, E., G. Pollio, C. Pellicciari, and A. Maggi. 1999. Estrogen and progesterone induction of survival of monoblastoid cells undergoing TNF‐α‐induced apoptosis. FASEB J. 13:793– 803. doi:10.1096/fasebj.13.8.793.

Tables and Figures

Table 5.1. Complete blood count analysis differences between late gestation sows with differing risk for pelvic organ prolapse (POP)1.

Farm A Farm B P-value Parameter (units) PS1 PS3 PS1 PS3 SEM Farm PS Farm*PS 3 White blood cells, 10 /µL 13.63 12.92 13.04 12.74 0.41 0.30 0.19 0.57 6 Red blood cells, 10 /µL 5.45 5.26 5.16 5.03 0.10 0.01 0.18 0.82 Hemoglobin, gm/dL 10.9 10.9 10.7 10.6 0.1 0.07 0.31 0.71 Hematocrit, % 33.0 32.9 32.4 31.8 0.5 0.02 0.29 0.64 Mean corpuscular volume, fl 62.1 62.7 63.1 63.5 0.6 0.11 0.32 0.82

Mean corpuscular hemoglobin, pg 20.5 20.7 20.9 21.1 0.2 0.02 0.30 1.00 115

Mean corpuscular hemoglobin concentration,gm/dL 33.0 33.0 33.2 33.2 0.1 0.05 0.81 0.59 Red cell distribution width, % 16.5 16.2 16.5 16.7 0.2 0.13 0.85 0.12 3 Platelets, 10 /µL 194 196 178 213 25 0.91 0.29 0.45 Mean platelet volume, fl 10.2 10.0 10.5 9.9 0.2 0.88 0.02 0.31 3 Neutrophils, 10 /µL 6.70 6.57 6.00 6.18 0.34 0.08 0.83 0.60 3 Lymphocytes, 10 /µL 5.30 5.05 5.47 4.94 0.21 0.96 0.02 0.45 3 Monocytes, 10 /µL 0.54 0.47 0.50 0.49 0.02 0.58 0.03 0.19 3 Eosinophils, 10 /µL 0.96 0.71 0.95 1.00 0.69 0.02 0.09 0.02 3 Basophils, 10 /µL 0.05 0.04 0.04 0.04 < 0.01 0.04 0.26 0.09 3 Absolute large unstained cells, 10 /µl 0.09 0.08 0.08 0.09 < 0.01 0.86 0.90 0.39 1Sows were assigned a perineal score (PS) based on their relative risk of experiencing POP. Sows assigned PS1 were considered low risk, while sows assigned PS3 were considered high risk for POP.

Table 5.2. Differences in circulating steroid hormones between sows differing in risk for pelvic organ prolapse (POP)1 Farm A Farm B P-value Steroid Hormone PS1 PS3 PS1 PS3 SEM Farm PS Farm*PS Deoxycorticosterone, nM 0.23 0.27 0.28 0.34 0.04 0.08 0.10 0.73 Deoxycortisol, nM 0.39 0.56 0.45 0.60 0.12 0.59 0.09 0.91 Hydroxyprogesterone, nM 0.112 0.13 0.119 0.106 0.011 0.35 0.81 0.11 Aldosterone, nM 0.20 0.38 0.38 0.46 0.08 0.06 0.07 0.45 Androstenedione, nM 0.107 0.138 0.102 0.106 0.009 0.01 0.02 0.06 Androsterone, nM 0.06 0.073 0.056 0.068 0.006 0.39 0.01 0.95 Corticosterone, nM 0.60 0.58 0.68 0.72 0.11 0.23 0.77 0.75 Cortisol, nM 34 35 40 39 5 0.26 0.98 0.94 Cortisone, nM 10.6 12.6 11.5 11.4 0.7 0.82 0.13 0.08 Estrone, nM 8.8 12.7 8.4 10.0 0.8 0.03 < 0.01 0.10 116

β-Estradiol, nM 0.69 0.99 0.60 0.75 0.06 < 0.01 < 0.01 0.18 Progesterone, nM 26 23 24 24 1 0.53 0.15 0.45 Testosterone, nM 0.041 0.056 0.048 0.052 0.006 0.84 0.05 0.27 1Sows were assigned a perineal score (PS) based on their relative risk of experiencing POP. Sows assigned PS1 were considered low risk, while sows assigned PS3 were considered high risk for POP.

Table 5.3. Serum biomarker analysis between late gestation sows differing in risk for pelvic organ prolapse (POP). Farm A Farm B P-value Parameter (units) PS1 PS3 PS1 PS3 SEM Farm PS Farm*PS Lipopolysaccharide binding protein (ng/mL) 6,726 8,614 5,793 7,400 1,131.2 0.26 0.06 0.88 Tumor Necrosis Factor alpha (pg/mL) 52.2 54.6 47.0 48.2 7.1 0.34 0.78 0.92 C-reactive protein (µg/mL) 33.44 35.93 23.87 28.65 5.40 0.07 0.35 0.80 Haptoglobin (µg/mL) 889.2 1036.3 768.3 770.9 138.4 0.11 0.64 0.55 Creatine Kinase (pg/nm) 1,629 1,046 540 562 318 < 0.01 0.31 0.27 1Sows were assigned a perineal score (PS) based on their presumed risk of experiencing POP. Sows assigned PS1 were considered low risk, while sows assigned PS3 were considered high risk for POP.

117

118

25% 23.4%

20%

15%

10% Percent ProlapsedPercent 5% 2.7% 1.0% 0% PS1 PS2 PS3

Figure 5.1. Perineal score (PS) at gestation week fifteen is an indicator of pelvic organ prolapse (POP) risk. Perineal scoring was conducted on sows (n = 2865) during late gestation (days 107-116). Sows assigned a PS1 (n = 1570), PS2 (n = 1166) and PS3 (n = 128) were considered low, medium and high risk, respectively, for POP. There was a difference in POP rates between PS1, PS2, and PS3 sows, with 23.4% of PS3 sows experiencing POP while 1.0% and 2.7% of PS1 and PS2 sows, respectively, experienced POP (P < 0.01). These data demonstrate a method to distinguish differential risk of POP for late gestation sows.

119

CHAPTER 6. SUMMARY AND CONCLUSION

Satisfying the demand for protein for a growing population will require increased production of safe and wholesome pork products. In order to meet these demands an increase in sow efficiency and production is necessary. The U.S. swine industry has made notable improvements in various areas of production, including reproductive performance, within the last decade. Unfortunately, at the same time, sow mortality is still increasing, causing a welfare concern while lowering profits (Supakorn, 2019). A 2018 survey revealed approximately 21% of sow deaths can be attributed to POP during the peripartum period (Ross, 2019). While the phenotype of POP is understood, the biological underpinnings in pigs remains unclear. In order to develop successful mitigation strategies a better understanding of the biological events preceding POP is urgently needed. The objectives of this thesis were to: 1) develop a better understanding of the vaginal microbiota in sows and how it may be associated with POP and 2) identify potential biomarkers that could serve as indicators to assess a sow’s risk for POP. In accomplishing these objectives, it has strengthened the understanding on POP and allowed for steps to be taken towards the long-term goal of development of mitigation strategies for POP within the U.S. sow herd.

An important component of these studies was the PS system that was developed, be effective in assessing risk of POP, as well as understandable enough to be implemented commercially. This is important for future work when evaluating mitigation strategy effectiveness or targeting treatments to high risk sows. In the current work, over 23% of PS3 assigned sows subsequently experienced POP, which was a significantly higher rate than those assigned a PS1 or PS2. These findings validate the PS system and prove it is effective in identifying high risk sows prior to experiencing POP. Additionally, utilization of this system

120 allowed for the exploration for changes in the vaginal microbiota in relation to POP risk (Chapter

3 and Chapter 4) as well as the identification of steroid hormone, markers of immune system activation and other potential biomarkers of POP risk in circulation (Chapter 3 and Chapter 5).

Vaginal microbiota is an emerging field of research, and the relationship to animal health and reproduction is a specific area of focus. Previous work has linked vaginal microbiota dysbiosis to reproductive dysfunction in both humans and cattle (Rodrigues et al., 2015; Lewis et al., 2017). The relationship between sow vaginal microbiota and health status is dynamic and complex. While current investigations into the reproductive microbiota focuses on the characterization, there is limited knowledge of the microbial community in late gestation sows.

In this thesis, novel discoveries were made into the characterization of the vaginal microbiota of late gestation sows (Chapter 3), along with further validation of these results (Chapter 4). This allowed for the identification of OTUs with similar representation across both studies, providing an initial description of the core vaginal microbiota in late gestation commercial sows.

Additionally, specific microbes of interest were acknowledged to be associated with POP risk in sows (Chapter 3 and 4). Specific OTUs of interest that were similar across both studies were identified, and categorized into potentially beneficial (Anaerococcus, Porphyromonas, and

Veillonella) or pathogenic (Prevotella, Streptococcus dysgalactiae, and Treponema) organisms in relation to POP risk, however further investigation is required. In order to further understand changes in the vaginal microbiota and its association with POP risk, a need to examine the vaginal microbiota across additional production systems, genetic lines, and geographical areas exists. In addition to the evaluation of vaginal microbiota, biomarkers to indicate POP risk in sows were assessed.

121

The phenotype for a sow at high risk for POP (PS3) is abnormal, at any stage of gestation. Being able to identify biomarkers associated with inflammation aid in the understanding of the biological events preceding POP. Serum molecular features (Chapter 3) along with inflammatory markers, immune cells, and steroid hormones (Chapter 5) were assessed in relation to POP risk in late gestation sows. Changes observed in the metabolome of diseased individuals may serve as primary indicators and is used in clinical practice (Zhang et al., 2012). Dissimilarities in the metabolome have the potential to help identify sows at risk during earlier stages of gestation. The current study (Chapter 3) identified 16 metabolites that were different in sows differing in PS, interestingly, all of which were increased in high risk

(PS3) sows. Increases in biomarkers commonly associated with inflammation (LBP), as well as decreases in immune cell populations (MPV, lymphocytes, and monocytes) were observed in sows at high risk for POP compared to those at low risk (Chapter 5). Interestingly, immune responses, and changes in the microbiome have been associated with fluctuations in steroid hormones (García-Gómez et al., 2013). The current study (Chapter 5) observed steroid hormone profiles to be altered between sows differing in PS, with hormone levels increase in PS3 sows, as well as distinct differences in the vaginal microbiome in sows at high risk for POP compared to their counterparts (Chapter 3 and 4).

Collectively, data from this thesis aids in the understanding of the biological associations preceding POP in the U.S. commercial swine herd. Because the biological associations with POP in commercial sows in an emerging field, further research is required to determine causality.

With novel areas of research additional biological questions are always generated, requiring future work to be pursued.

122

Future Directions

Microbiota

This thesis has provided the initial characterization of the vaginal microbiota of late gestation sows, however further work is required to capitalize on this information and move this area of research forward to accomplish our long-term goals of improving sow health and reproduction. Observations of the changes within the vaginal microbiota in sows with low risk for POP are not in completed alignment with similar work done in humans in relation to reproductive disorders. This further validates a need to define the vaginal microbiota in sows as well as function of these microbes within the swine reproductive tract. One way to evaluate the understanding of microbiota observed in our study is to evaluate the function of certain microbes by metagenome shotgun sequencing, which could provide specific information about the function of the microbes present. Shotgun metagenomics provides information on the microbial genomes present in a sample, which is then used to profile taxonomic composition and functional potential of the microbial communities (Quince et al., 2017). Mechanistic research is needed to demonstrate causality of POP risk by specific microbes in order to move this research area forward.

A more diverse sample size, for example, from additional production systems, genetic lines, and geographical areas, may be useful in the further characterization of a core vaginal microbiota for sows. Increased sample size could also validate if specific microbes identified within this study (Chapter 3 and 4) are beneficial or pathogenic. Presently, studies have only been conducted on commercial farms with high POP incident rates. Alternatively evaluating the microbiota of sows on farms with low POP incident rates could provide valuable data about the vaginal microbial community as it relates to sow health. The interaction between the microbiota and different body sites should also be considered.

123

The gut-vagina axis may have an important role with the potential to improve reproductive health. In order to better understand if and what the interaction would be, the fecal microbiota of sows at varying risk for POP should be evaluated to determine the relationship between the fecal and vaginal microbiomes. Evaluating the fecal microbiota of sows would provide information on if a feeding strategy could be a feasible mitigation strategy.

Mitigation strategies are another avenue worth pursuing. When attempting to alter the microbiota, anti- or probiotics may be considered. Given that the use of antibiotics in the swine industry is under scrutiny, it may be beneficial to focus on mitigation strategies utilizing probiotics. Benefits of probiotic use have been proven including inducing changes in the fecal microbiota (Hemarajata and Versalovic, 2013), and are a treatment option for women suffering from BV (Javed et.al.,2019). However, further research is necessary to determine if similar changes could be induced in the vaginal microbiota of sows, and whether these changes could have clinical benefits. Administration of the mitigation strategy is also something to consider, and there are benefits to both an oral administration or an intravaginal injection. However, in order to test the efficacy of a mitigation strategy, better understanding of the biological preceding

POP is necessary.

Biological samples

In order to better our understanding of what precedes POP, analysis of reproductive tissues is necessary. Immune cell populations are decreased in sows at higher risk for POP

(Chapter 5), and our working hypothesis is immune cells are migrating and have a localization effect to the reproductive tract. Immunohistochemistry staining of tissue samples would be able

124 to detect if there is a localization of immune cells that were observed to be decreased in circulation of PS3 sows.

Increases in steroid hormones were observed in high risk (PS3) sows (Chapter 5) and many of them were catalyzed by CYP19A1. The enzyme CYP19A1 could be measured in target tissues to compare levels within the reproductive tract between low and high risk for POP sows.

In order to evaluate CYP19A1 abundance, it is necessary to analyze tissue samples. Specific tissue of interest would be the ovaries, endometrial tissue, as well as the ligaments and other structures that hold the reproductive tract in place within the body cavity, and potentially the placenta.

In humans, damage to the levator ani muscle, as well as lengthening and decreased resilience of the uterosacral ligament are associated with an increased risk of POP (Reay Jones et al., 2003; De Lancey, 2016). Since humans and swine have anatomical similarities it would be interesting if the same observations could be made. Alternatively, extracellular matrix (ECM) components of the lower reproductive tract in females may play a critical role in preventing POP

(Alarab et al., 2014).

Collagenase activity

Collagen is an ECM protein which determines soft tissue strength whereas elastin allows the reproductive tract tissues to stretch and return to their original shape/structure. Collagen plays a role in the structure of the muscles and tissue in the female pelvic floor and reproductive tract

(Dhital et al., 2016). Women who are affected by genetic diseases which involve mutations in connective tissue, like those found in the reproductive tract, have an increased incidence of POP

(Carley and Schaffer, 2000). This leads to question the collagen activity within sows in relation

125 to POP. Decreases in total collagen have been observed in women suffering from POP (Jackson et al., 1996). Alternatively, E2 has been observed to increase collagen gene transcription (Clark et al., 2005), acting via the estrogen receptors in the nuclei of connective tissues and smooth muscle (Rechberger et al., 1993; Franz et al., 1996; Lang et al., 2003). Collagenase is an enzyme that breaks down collagen, and microbes have the potential to produce collagenases (Duarte et al., 2016). Therefore, bacteria may have the capability of interfering with the host animal’s connective tissue and could be a possible mechanism associated with POP. Further research would be required to evaluate collagen levels in sows at risk for POP and should be pursued.

References

Alarab, M., H. Kufaishi, S. Lye, H. Drutz, and O. Shynlova. 2014. Expression of extracellular matrix-remodeling proteins is altered in vaginal tissue of premenopausal women with severe pelvic organ prolapse. Reprod. Sci. 21:704–715. doi:10.1177/1933719113512529.

Carley, M. E., and J. Schaffer. 2000. Urinary incontinence and pelvic organ prolapse in women with Marfan or Ehlers-Danlos syndrome. Am. J. Obstet. Gynecol. 182:1021–1023. doi:10.1067/mob.2000.105410.

Clark, A. L., O. D. Slayden, K. Hettrich, and R. M. Brenner. 2005. Estrogen increases collagen I and III mRNA expression in the pelvic support tissues of the rhesus macaque. Am. J. Obstet. Gynecol. 192:1523–1529. doi:10.1016/j.ajog.2004.11.042.

Duarte, A. S., A. Correia, and A. C. Esteves. 2016. Bacterial collagenases - A review. Crit. Rev. Microbiol. 42:106–126. doi:10.3109/1040841X.2014.904270.

Franz, H. B. G., D. Wendler, and G. Oettling. 1996. Immunohistochemical assessment of steroid hormone receptors in tissues of the anal canal: Implications for anal incontinence? Acta Obstet. Gynecol. Scand. 75:892–895. doi:10.3109/00016349609055023.

García-Gómez, E., B. González-Pedrajo, and I. Camacho-Arroyo. 2013. Role of sex steroid hormones in bacterial-host interactions. Biomed Res. Int. 2013. doi:10.1155/2013/928290.

126

Hemarajata, P., and J. Versalovic. 2013. Effects of probiotics on gut microbiota: Mechanisms of intestinal immunomodulation and neuromodulation. Therap. Adv. Gastroenterol. 6:39– 51. doi:10.1177/1756283X12459294.

Jackson, S. R., N. C. Avery, J. F. Tarlton, S. D. Eckford, P. Abrams, and A. J. Bailey. 1996. Changes in metabolism of collagen in genitourinary prolapse. Lancet. 347:1658–1661. doi:10.1016/S0140-6736(96)91489-0.

De Lancey, J. O. L. 2016. What’s new in the functional anatomy of pelvic organ prolapse? Curr. Opin. Obstet. Gynecol. 28:420–429. doi:10.1097/GCO.0000000000000312.

Lang, J. H., L. Zhu, Z. J. Sun, and J. Chen. 2003. Estrogen levels and estrogen receptors in patients with stress urinary incontinence and pelvic organ prolapse. Int. J. Gynecol. Obstet. 80:35–39. doi:10.1016/S0020-7292(02)00232-1.

Quince, C., A. W. Walker, J. T. Simpson, N. J. Loman, and N. Segata. 2017. Shotgun metagenomics, from sampling to analysis. Nat. Biotechnol. 35:833–844. doi:10.1038/nbt.3935.

Reay Jones, N. H. J., J. C. Healy, L. J. King, S. Saini, S. Shousha, and T. G. Allen-Mersh. 2003. Pelvic connective tissue resilience decreases with vaginal delivery, menopause and uterine prolapse. Br. J. Surg. 90:466–472. doi:10.1002/bjs.4065.

Rechberger, T., H. Donica, W. Baranowski, and J. Jakowicki. 1993. Female urinary stress incontinence in terms of connective tissue biochemistry. Eur. J. Obstet. Gynecol. Reprod. Biol. 49:187–191. doi:10.1016/0028-2243(93)90268-H.

Ross, J. W. 2019. Identification of putative factors contributing to pelvic organ prolapse in sows (Grant # 17-224) II. Industry Summary.

Supakorn, C. 2019. A review of aetiology and risk factors affecting sow mortality. CAB Rev. Perspect. Agric. Vet. Sci. Nutr. Nat. Resour. 14. doi:10.1079/PAVSNNR201914026.

127

Zhang, A., H. Sun, and X. Wang. 2012. Serum metabolomics as a novel diagnostic approach for disease: A systematic review. Anal. Bioanal. Chem. 404:1239–1245. doi:10.1007/s00216-012-6117-1.

128

APPENDEX: ADDITIONAL DATA

Figure A.3.1. Alpha diversity of vaginal microbiome for sows with low and high risk for prolapse. The samples show species richness and diversity. Chao species richness is an estimate of the true species richness based on the sample size and Shannon and Simpson analysis evaluate diversity.

129

PS 6 1 3

0 Component(3.4%) 2

- - 0 2 4 6 Figure A.3.2. Serum metaboliteComponent differences 1 (8.4%) overview. Sparse Partial Least Squares Discriminant Analysis (SPLS-DA) was produced by the MetaboAnalyst program using the swine prolapse serum data. Grey dots represent sows with low risk for POP (perineal score 1 (PS1)), and animals with high risk (PS3) for POP are represented by the black dots. This figure compares serum metabolites between PS1 and PS3 on an individual animal basis. These data suggest there is some overlap between the metabolites within PS, which is to be expected, but there are also differences. This figure was made using the following Xia, J., Wishart, D. Web-based inference of biological patterns, functions and pathways from metabolomic data using MetaboAnalyst. Nat Protoc 6, 743–760 (2011). https://doi.org/10.1038/nprot.2011.319

Table A.3.1. Fifty most abundant OTUs1 within swine late gestation vaginal communities.

NCBI BLAST Relative abundance OTU (%) Phylum Taxonomy (Silva v132) Classification Similarity (%) Accession no. OTU 1 6.19 Firmicutes Veillonella Veillonella caviae PV1 99.6 NR_025762.1 OTU 2 5.23 Proteobacteria Pasteurellaceae_unclassified Actinobacillus porcinus 35NTS 100 FJ437063.1 OTU 3 2.61 Fusobacteria Fusobacterium Fusobacterium gastrosuis CDW1 100 NR_146837.2 OTU 4 2.44 Bacteroidetes Prevotellaceae_UCG-001 Duncaniella sp. TLL-A3 84.7 MK521456.1 OTU 5 2.18 Firmicutes Phascolarctobacterium Phascolarctobacterium succinatutens YIT 99.6 NR_112902.1 OTU 6 2.10 Firmicutes Clostridium_sensu_stricto_1 Clostridium sp. CL-2 100 KF528156.1 OTU 7 2.02 Firmicutes Parvimonas Parvimonas sp. KA00067 99.6 KP192297.1 OTU 8 1.97 Firmicutes Turicibacter Turicibacter sp. H121 100 CP013476.1 OTU 9 1.72 Bacteroidetes Bacteroides Bacteroides massiliensis DCW_SL_46 96.1 MK424043.1 OTU 10 1.54 Proteobacteria Pasteurellaceae_unclassified Pasteurella mairii strain CCUG 27189 100 NR_042886.1

OTU 11 1.38 Proteobacteria Escherichia-Shigella Escherichia coli 99.2 AJ567617.1 130 OTU 12 1.23 Firmicutes Romboutsia Romboutsia sp. ilealis 100 LN555523.1 OTU 13 1.12 Euryarchaeota Methanobrevibacter Methanobrevibacter sp. N58C 100 LN610763.1 OTU 14 1.06 Bacteroidetes Bacteroidales_unclassified Muribaculum sp. S4 85.0 MK287698.1 OTU 15 1.03 Firmicutes Terrisporobacter Bacterium WH2-11 100 JQ269302.1 OTU 16 0.97 Proteobacteria Pasteurellaceae_unclassified Pasteurella aerogenes ZWR1 99.6 MN006266.1 OTU 17 0.96 Bacteroidetes Prevotellaceae_unclassified Prevotella timonensis SSA5 93.7 MH714863.1 OTU 18 0.81 Firmicutes Clostridium_sensu_stricto_1 Clostridium moniliforme HYN0057 99.6 KY079341.1 OTU 19 0.79 Firmicutes Peptostreptococcus Peptostreptococcus sp. DSM 106284 99.6 MN537513.1 OTU 20 0.77 Firmicutes Lactobacillus Lactobacillus sp. strain DSM 107288 100 MN537532.1 OTU 21 0.72 Firmicutes Kurthia Kurthia gibsonii EMB4 100 KY048434.1 OTU 22 0.72 Bacteroidetes Porphyromonas Porphyromonas somerae KA00683 99.6 KP192301.1 OTU 23 0.70 Bacteroidetes Prevotellaceae_NK3B31_group Prevotellaceae bacterium 96.1 LC333724.1

Table A.3.1 Continued OTU 24 0.69 Actinobacteria Corynebacterium_1 Corynebacterium xerosis FDAARGOS_674 100 CP046322.1 OTU 25 0.66 Firmicutes Streptococcus Streptococcus suis SS248 100 KR819490.1 OTU 26 0.66 Firmicutes Jeotgalibaca Jeotgalibaca dankookensis HM 99.6 KU714711.1 OTU 27 0.63 Bacteroidetes Prevotella_1 Prevotellaceae bacterium strain AGP2-01-07-05 99.2 MH699323.1 OTU 28 0.63 Bacteroidetes Prevotellaceae_NK3B31_group Prevotellaceae bacterium 98.4 LC333719.1 OTU 29 0.62 Fusobacteria Fusobacterium Fusobacterium necrophorum FDAARGOS_565 100 CP033837.1 OTU 30 0.60 Firmicutes Clostridiales_unclassified Unidentified Eubacterium 91.7 AJ229202.1 OTU 31 0.57 Firmicutes Anaerococcus Anaerococcus tetradius 99.6 LC036320.1 OTU 32 0.55 Bacteroidetes Prevotellaceae_unclassified Prevotella sp. NAIP4D 96.1 JQ797590.1 OTU 33 0.55 Bacteroidetes p-251-o5_ge Parabacteroides distasonis WYJ14_D4 85.2 MN081648.1 OTU 34 0.54 Firmicutes Clostridium_sensu_stricto_1 Clostridium butyricum CFSA3989 100 CP033249.1 OTU 35 0.53 Firmicutes Clostridium_sensu_stricto_1 Clostridium celatum 99.6 AB971795.1 OTU 36 0.52 Proteobacteria Acinetobacter Acinetobacter defluvii WCHA30 99.2 CP029397.2 OTU 37 0.49 Firmicutes Streptococcus Streptococcus hyovaginalis TRG26 100 MH329638.1

OTU 38 0.48 Spirochaetes Treponema_2 Treponema bryantii, 99.6 AB849328.1 131 OTU 39 0.46 Bacteroidetes Bacteroides Prevotella sp. cp02.11 95.3 AY827858.1 OTU 40 0.45 Firmicutes Phascolarctobacterium Phascolarctobacterium succinatutens 99.2 AB490812.1 OTU 41 0.45 Bacteroidetes Prevotellaceae_NK3B31_group Prevotellaceae bacterium 98.8 LC333719.1 OTU 42 0.44 Proteobacteria Acinetobacter Acinetobacter sp. MSRC7 100 MH447439.2 OTU 43 0.43 Firmicutes Ruminococcus_1 Bacterium MA2007 99.6 KF698129.1 OTU 44 0.43 Bacteroidetes Porphyromonas Porphyromonas endodontalis 96.4 LT680662.1 OTU 45 0.43 Firmicutes Kurthia Kurthia gibsonii SAU_AFB01 99.6 MN658386.1 OTU 46 0.42 Bacteroidetes Prevotellaceae_UCG-001 Muribaculum sp. S4 86.1 MK287698.1 OTU 47 0.42 Bacteroidetes Bacteroides Bacteroides fragilis LPB0329 100 MN629228.1 OTU 48 0.42 Bacteroidetes Prevotella_9 Prevotellaceae bacterium AGP1-12-14-09 100 MH699319.1 OTU 49 0.41 Bacteroidetes Porphyromonas Porphyromonas katsikii JF5581 91.3 KM360064.1 OTU 50 0.41 Firmicutes Terrisporobacter Terrisporobacter sp. CCK3R4-PYG-107 99.2 KR364793.1 1Individual microbes were assigned in order of abundance and classified into operational taxonomic units (OTUs)

132

Table A.3.2. Small molecule metabolites in serum from sows with differing risk of POP.

Perineal Score1 2 Metabolite PS1 PS3 Fold Change P-value RI=1242.7, 11.7229 min, 35889 1.47E-13 7.78E-12 5.72 < 0.01 RI=1693.9, 17.2652 min, 20024 8.99E-13 3.54E-11 5.30 < 0.01 D-Fructose 1.48E-11 5.80E-10 5.29 < 0.01 RI=2105.9, 20.8495 min, 17937 1.04E-13 3.21E-11 8.27 < 0.01 RI=1208.7, 11.2167 min, 40555 4.75E-14 3.51E-12 6.21 < 0.01 RI=1915.0, 19.2855 min, 17937 2.26E-12 1.31E-10 5.86 < 0.01 RI=1180.6, 10.7977 min, 36473 2.04E-12 7.66E-11 5.23 < 0.01 RI=2109.7, 20.8792 min, 17937 1.49E-13 3.33E-11 7.81 < 0.01 RI=1758.2, 17.8880 min, 40555 4.25E-14 3.85E-12 6.50 < 0.01 1 Sows were assigned a perineal score based on their relative risk of experiencing a pelvic organ prolapse (POP). Sows assigned PS1 were presumed low risk for POP while sows assigned PS3 were presumed high risk for POP. 2 Metabolite list only includes those that were identifiable and statistically significant (P < 0.01) between perineal scores. Values reported as moles per mg.