Endometrial Cytology in Mares Bred with Frozen Semen

There is a small resident population of neutrophils present in the endometrium after ovulation. Breeding with one full dose or two half doses of frozen semen does not significantly increase the amount of debris or percent neutrophils in low-volume uterine lavage samples collected at 24 - 96 h after breeding.

1. Introduction

Acute post-breeding uterine inflammation in mares is believed to be caused by the introduction of spermatozoa and bacterial contaminants. Delayed uterine clearance of spermatozoa and contaminates has been identified as the underlying cause of persistent uterine inflammation. Some mares bred with frozen semen develop marked persistent inflammation that has been attributed to the removal of seminal plasma during the process of cryopreservation, allergic type hypersensitivity reactions to components of the freezing extenders such as glycerol and egg yolk, or delayed uterine clearance as is reported in mares susceptible to endometritis [1,2]. Seminal plasma has been shown to have a modulating depressant effect on the endometrial inflammatory response. However, the marked uterine inflammatory cell reaction after breeding with frozen semen is not commonly observed in young virgin mares [3]. Sperm-induced persistent inflammation has been described as contributing to lower fertility in mares bred with frozen semen by altering the uterine environment so that it is incapable of supporting a pregnancy. The embryo leaves its tubal phase on about day 5 after ovulation; therefore, uterine inflammation must be controlled by 96 h after ovulation to maximize survival of the embryo. A number of breed associations have recently approved breeding with frozen semen, resulting in its increased use. It has become increasingly important to determine the underlying cause of aggressive post-breeding endometrial inflammation seen in some mares.

Our objectives were to characterize endometrial cytologic features such as the amount of debris, presence of bacteria, total neutrophil numbers, and percentage of neutrophils in mares selected for good breeding potential that were bred with frozen semen and compare the results with those of mares not bred for up to 96 h after breeding. The non-bred control group was chosen to determine the amount of variability in parameters associated with day of the physiologic cycle not related to breeding. We aimed to identify the cytologic features present after breeding with frozen semen and to determine the relationship between the presence of debris, bacteria, and neutrophils in two different insemination protocols. It was hypothesized that most uterine inflammation would peak within 24 h of breeding and then decrease; additionally, any inflammation that persisted >24 h would be attributable to bacterial causes.

2. Materials and Methods

Semen was collected from a fertile Quarter horse stallion using a Missouri-type artificial vagina. Semen was filtered and extended 1:1 with commercial skim milk extender. The seminal plasma was removed by centrifugation at 400 × g for 12 min. The sperm pellet was aspirated and resuspended in Lactose-EDTA extender at a concentration of 400,000,000 sperm/ml. Extended semen was packaged in 0.5-ml straws at room temperature and passively cooled to 4°C over 2 h. The straws were suspended 5 cm above liquid nitrogen for 10 min and plunged. For this study, insemination doses of 800,000,000 total spermatozoa were selected from frozen ejaculates that had a mean ± SD of 47 ± 8% morphologically normal spermatozoa and 25.5 ± 6% average post-thaw motility. The 0.5-ml straws were thawed in a water bath at 37°C for 45 s and then discharged in a sterile tube. One-half milliliter of pre-warmed Kenney extender [a] was aspirated into the insemination pipette followed by the thawed semen.

The experimental protocol was approved by the University of Saskatchewan’s Institutional Animal Care and Use Committee. Thirty-three mares selected for good breeding and potential fertility ranging in age from 2 - 12 yr (mean age = 5 yr) were used for the study. Good breeding potential was defined as good perineal conformation, young age, no history of reproductive problems, and one or more previous ovulations and breedings with serial monitoring of the reproductive tract using ultrasound and palpation without evidence of subfertility. Data was collected from 37 cycles beginning in May and ending in September. Thirty-three mares were sampled one time, three mares were sample two times, and one mare was sampled three times. Mares were intensively housed in large dry lots and fed hay, barley silage, oats, mineral, and salt with free access to water. The mares were individually identified by brands. A GE Ausonics ultrasound machine and a variable-frequency 6-MHz linear array probe was used for ultrasound examinations.

Ultrasound examinations were performed daily during estrus to determine the day of natural ovulation (day 0) and for 96 h after ovulation. There were three treatment groups: group 1, control non-bred; group 2, single breeding on the day of ovulation with 800,000,000 (full dose) frozen thawed sperm; group 3, two breedings (one breeding at ovulation and one breeding 24 h later with 400,000,000 [dose] frozen thawed sperm per insemination and sample collection time). There were four sample collection times (24, 48, 72, and 96 h after ovulation) with four cycles per collection time for groups 1 (n = 4 cycles × 4 sample times = 16 total cycles) and 2 (n = 4 cycles × 4 sample times = 16 total cycles). There was only one sample time for group 3 (24 h after the last breeding; n = 7 cycles). Different sample sizes were chosen because of the greater variability expected in group 3. Mares were sampled at one assigned collection time per estrus.

On the day of ovulation, each mare was randomly assigned a group, and if they were in group 1 or 2, they were also randomly assigned a sampling time (group 3 mares only had one sampling time). Mares assigned treatment two or three were bred using aseptic transcervical artificial insemination techniques. A low-volume uterine lavage was performed using 60 ml of phosphate buffered saline (PBS). The mare was prepared as for artificial insemination (AI), and an AI pipette was introduced through the cervix into the uterus. Sixty milliliters of PBS was infused, and the uterus was massaged per rectum to obtain a representative sample of the uterine environment. Fluid was aspirated back through the AI pipette to obtain the sample. Approximately 10% of the fluid was recovered. Recovered fluid was measured and placed in a 10-ml centrifuge tube and centrifuged at 400 × g for 10 min. The cell pellet was resuspended with 0.5 ml of PBS. Cells were counted using a hemocytometer, and the number of cells per milliliter of flushed solution was recorded and adjusted to represent the cells in the 60-ml sample volume. Differential cell counts (300 cells) were performed on slides stained with a rapid Wright Giemsa stain [b], and the percentage of neutrophils was multiplied by the total number of cells to determine the number of neutrophils per 60 milliliters infusion of PBS. Debris and bacterial scores were also evaluated on each slide. The amount of debris was given a score from 1 to 4 (1, <25%; 2, <50%; 3, <75%; 4, >75%) on the proportionate amount of debris that covered a high power (1000 times) microscopic field. Bacteria was given a score from 1 to 5 based on the number of bacteria per high power (1000 times) field (1, no bacteria per 30 fields; 2, one bacterium per 30 fields; 3, one bacterium per 10 fields; 4, 2 - 10 bacteria per field; 5, 11 - 50 bacteria per field).

The equality of variances in the data was tested using the Shapiro Wilk test. A significance level of p < 0.05 was used for all statistical tests. A Kruskal Wallis non-parametric analysis of variance (ANOVA) was used to evaluate differences between and within groups. Comparison of mean rank was used for determining differences between groups. Spearman Rank correlation was used to determine relationships between endometrial cytologic variables, time, and treatment group. Data was analyzed using a statistical computer software program [c].

3. Results

Shapiro Wilk tests showed that the variances of the data were not equal; therefore, the p values are from the non-parametric Kruskal Wallis one-way ANOVA. There were no within treatment group differences of sample time post-ovulation for group 1 on bacterial score (p = 0.3860), debris (p = 0.7961), total neutrophils (p = 0.5274), or percent neutrophils (p = 0.3266). There were no within treatment group differences of sample time post-ovulation for group 2 on bacterial score (p = 0.4769), debris (p = 0.1465), total neutrophils (p = 0.8552), and percent neutrophils (p = 0.6231). The pooled descriptive data from all time points in groups 1 and 2 are shown in Table 1. Pooled descriptive data from the 24-h post-ovulation time-point (groups 1 and 2) or after the last breeding (group 3) are shown in Table 2. There were no significant differences between the three groups in endometrial cytology parameters in treatment groups at the 24-h sample time: bacterial scores (p = 0.4707), debris (p = 0.1866), total (p = 0.3805), and percent neutrophils (p = 0.9446). Effects of treatment in the pooled data on endometrial cytology parameters including bacterial scores, debris, total neutrophils, and percent neutrophils in groups 1 - 3 are shown in Table 3. Table 4 shows the correlation coefficients and p values between endometrial cytologic parameters, time, and treatment group.

4. Discussion

Persistent post-breeding inflammation is a concern of practitioners, because it is associated with lower fertility. It has also been reported as a cause of decreased pregnancy rates in mares bred with frozen semen. Presently, practitioners do not have a normal range of endometrial cytologic parameters that allow them to evaluate mares post-breeding and determine if intervention is necessary. There is a lack of studies that systematically evaluate endometrial cytology after breeding. Endometrial debris has not previously been evaluated as a means of determining post-breeding inflammation, and we described this parameter in mares after breeding. The main goal of this research was to describe endometrial cytologic parameters in mares selected for good breeding potential at various time intervals after breeding with frozen semen. Additionally, post-breeding inflammation was evaluated using two different insemination protocols in the time frame when practitioners are frequently evaluating mares for post-breeding inflammation.

There are contradictory reports of the effect of two breedings on fertility. Metcalf [4] reported that a previous breeding with a full dose of frozen semen did not interfere with the fertility of a second breeding with frozen semen within 12 h. However, Troedsson et al. [5] reported different pregnancy rates in mares that were first treated with extender or dead sperm and then bred without seminal plasma (1 of 22 mares) compared with those bred with seminal plasma only (17 of 22 mares). This study suggested that insemination into an inflamed uterine environment without the addition of seminal plasma resulted in low fertility. Seminal plasma has been associated with a reduction in neutrophil binding to sperm in vitro [6], and it is removed in the process of cryopreservation. Large field trials have reported that pregnancy rates were the same in mares bred one time in the periovulatory period with a full dose of semen and mares bred two times 24 h apart with a half dose of semen. This indicates that the post-breeding inflammation was not detrimental to fertility under these conditions [7].

Within groups 1 and 2, there was no time effect on bacterial score, debris, total neutrophils, or percent neutrophils, suggesting that sampling a mare between 24 and 96 h after breeding will provide similar results. This is the time frame when many veterinarians are evaluating mares for persistent inflammation. The pooled data showed low levels of neutrophils in non-bred and bred mares. In group 1, the median (first, second, and third quartiles) percent neutrophils was 2 (range, 1 - 6) with a median bacterial score of 2 (range, 1 - 2). This means that there was one bacterium per 30 high-powered (1000×) microscope fields. This finding differs from the findings of Kotilainen et al. [8]; that study found that the control mares had no neutrophils when the low-volume uterine lavage technique was used. However, their findings were in agreement with Nikolakopoulos and Watson [9], who also found a low percentage of neutrophils in control mares.

Our data show that there is a low percentage of neutrophils in the uterus, which could be attributed to a constant resident population that does not vary within the first 96 h of ovulation. The physiological inflammation described in several studies [2,8,9] was not reflected in the median values observed in samples from mares bred with frozen semen at 24 - 96 h. Group 2 had median percent neutrophils of 5 (range, 3 - 10) with a bacterial median of 3 (range, 2 - 3). The median percent neutrophils for group 3 was 3 (range, 1 - 28), and the bacterial median was 2 (range, 1 - 3). The neutrophil percentages in the pooled data were similar across treatments. This suggests that mild inflammation (i.e., median percent neutrophils of 5) is common in mares selected for good breeding potential after breeding. There were a few lavage samples in each group with neutrophil counts and percents above the median. These were often associated with higher bacterial scores, which suggests that iatrogenic contamination resulted in a greater degree of inflammation. We would interpret this data to indicate that a mare with >5% neutrophils within 96 h of ovulation or breeding with frozen semen in a low-volume uterine lavage should have further consideration for possible treatment.

We compared endometrial cytologic parameters 24 h after last breeding in groups 2 and 3, the time frame when peak inflammation was described, or 24 h after ovulation in group 1. There were no differences in bacterial score, debris, total neutrophils, or percent neutrophils in groups 2 and 3. These parameters were not different in this sample set, suggesting that inflammation peaked earlier or was minimal when these breeding protocols were used.

In the comparison of the treatments in the pooled data, there was a significant effect of treatment on bacterial score (p = 0.0001) and total neutrophils (p = 0.0201). Comparison of bacteria by mean rank showed that group 1 ranks were less than the ranks of groups 2 and 3, but groups 2 and 3 ranks were similar. Mares bred one or two times had similar bacterial populations but higher bacterial scores than control samples, suggesting that the transcervical insemination and/or the inseminate introduced these microorganisms. Total neutrophils were also different between the groups; group 2 was different from group 1, but group 3 was not different from groups 1 or 2. Larger numbers of mares may be required to determine if this is a real difference or caused by chance. The amount of debris in the mare was not increased by breeding with frozen semen and is present in low levels in mares selected for potential fertility.

Troedsson et al. [2] suggested that persistent sperm-induced inflammation may be a more important cause of infertility in susceptible mares than infectious endometritis. Our research suggests, in agreement with Maloufi et al. [10], that physiological inflammation is negligible by 24 h after breeding. Maloufi et al. [10] concluded that elevated neutrophil counts at 96 h were associated with bacterial infection and persistent inflammation. However, elevated numbers were not found after the frozen semen or extender challenge was injected into susceptible mares. Cytologic parameters measured 24 - 96 h after ovulation and compared with normal values will facilitate timely treatment of mares with post-breeding endometritis.

The Spearman rank correlations showed significant correlations between bacteria scores and total neutrophil number (p = 0.0524) and bacteria scores and treatment (p < 0.001). This was expected, because bacteria are known to stimulate neutrophil chemotaxis and influx. Fewer bacteria were expected in the non-bred mares, which we believe explained the relationship with treatment. Debris was correlated to total neutrophil number (p = 0.005) and percent neutrophils (p = 0.0015), suggesting that it is primarily a residue left behind from the degeneration of neutrophils. Total neutrophil number was predictably related to percent neutrophils (p = 0.002) and treatment group (p = 0.0176). We anticipated higher neutrophils counts in the bred mares because of the introduction of sperm and bacterial contaminants.

In conclusion, there is a small resident population of neutrophils in the endometrium after ovulation. Breeding with one full dose or two half doses of frozen semen does not significantly increase the amount of debris or the percent neutrophils in low-volume uterine lavage samples collected from 24 to 96 h post-breeding. Bacterial scores are higher in mares bred with frozen semen compared with non-bred mares. Further research will be required to determine if this finding is related to contamination during breeding.

We thank the Alberta Agriculture Research Institute and the Equine Health Research Fund of the Western College of Veterinary Medicine for financial support.

Footnotes

1. EZ-Mixin BF extender, Animal Reproduction Systems, Chino, CA 91710.
2. Diff Quik, Fischer Diagnostics, Middletown, VA 22645-0307.
3. Student Edition of Statistix version 7.0 for Windows 2.0, Tallahassee, FL 32317.