Histologic Changes Associated With Placental Separation in Gilts Infected with Porcine Reproductive and Respiratory Syndrome Virus

Experimental Design

Twenty 6-month-old, purebred Landrace gilts from a PRRSV-naive farm were artificially inseminated using semen from Yorkshire boars. Fifteen pregnant gilts were intramuscularly and intranasally inoculated with PRRSV-2 (NLSV 97-7895; 1 × 10 ⁵ 50% tissue culture infective dose total dose) on gestation day 86 ± 0.4 and then euthanized at 2, 5, 8, 12, or 14 dpi (3 gilts per time point), along with one sham-inoculated control at each day postinoculation (5 total). An equivalent volume of minimum essential medium was similarly administered to all sham gilts on gestation day 86 ± 0.5. All gilts were vaccinated prebreeding against porcine parvovirus, porcine circovirus type 2, Leptospira interrogans (multivalent) and Erysipelothrix rhusiopathiae using commercial products. The study protocol adhered to principles established by the Canadian Council on Animal Care and was reviewed and approved by the Animal Research Ethics Board at the University of Saskatchewan (permit no. 20160023).

Gross Pathology

The gravid reproductive tract was removed intact and placed in a specially designed trough in a linear manner. The uterus was cut from left oviduct to right oviduct along the mesometrium (broad ligament) starting at the tip of each horn. Fetuses were numbered sequentially according to their position within each horn starting with “L1” and “R1,” closest to the ovary on the left and right sides, respectively. Sampling of the uterus with adherent fetal placenta (chorioallantois) was performed by cutting uterine wall (full thickness) into rectangular samples centered on the umbilical stump (∼20 cm long side) of each fetus. External fetal preservation status was assessed on the basis of previously published criteria ¹¹ as follows: viable (VIA; normal, white to purple skin with visible hair and regular umbilical cord), meconium stained (MEC; skin covered with inspissated, brownish amniotic fluid and regular umbilical cord with edema), and decomposed (DEC; <50% of skin discolored, no blood in the umbilical cord or from axilla).

Histopathology

From the sample of MFI obtained at the necropsy, rectangular 3 × 1 cm, full-thickness samples adjacent to the umbilical stump of each fetus were fixed in 10% neutral buffered formalin solution and paraffin embedded for histopathologic evaluation. Histopathology of the MFI was assessed by a pathologist blinded to inoculation group, fetal preservation status, time points of infection, and PRRSV ribonucleic acid (RNA) concentration in the endometrium and fetal placenta. Severity of inflammatory lesions in endometrium and fetal placenta was separately assessed on the basis of a previously published grading scheme. ¹⁸ Briefly, the severity of the inflammatory cell infiltrate in endometrium (Figs. S1–S4) and chorioallantois (Figs. S5–S8) was graded as follows: normal = score 0 (no inflammatory cell infiltrate in the tissue), minimal = score 1 (<10% of tissue section contains inflammatory cell infiltrate), mild = score 2 (10%–25% of tissue section contains inflammatory cell infiltrate), moderate = score 3 (25%–50% of tissue section contains inflammatory cell infiltrate), and severe = score 4 (>50% of tissue section contains inflammatory cell infiltrate). The distribution of vasculitis, a PRRSV-specific lesion in MFI, was evaluated separately in the endometrium and fetal placenta on the basis of a previously published scoring scheme ¹⁸ : normal = score 0 (no blood vessels inflamed in the tissue section), focal = score 1 (<30% of inflamed blood vessels in the tissue section), multifocal = score 2 (30%–70% of inflamed blood vessels in the tissue), or diffuse = score 3 (>70% of inflamed blood vessels in the tissue sections inflamed) (Figs. S9–S12). Finally, separation of placenta from the uterus was assessed. To eliminate potential artificial placental separations that may have occurred during tissue processing, only placental detachments associated with inflammation in the endometrium and fetal placenta, and/or the presence of acute cellular swelling (degeneration) of trophoblastic epithelium, and/or the presence of necrotic debris was used for evaluation because they require time to develop and therefore were considered to have occurred antemortem. The severity of these detachments was evaluated using a novel grading scheme that reflected the percentage of maternal-fetal interdigitation area affected by this pathologic process. Therefore, the severity of separation was scored as follows: normal = score 0 (no detachment of placenta along the interdigitation area), mild = score 1 (<25% of interdigitation area affected by detachment of placenta), moderate = score 2 (25%–50% of interdigitation area affected by detachment of placenta), and severe = score 3 (>50% of interdigitation area affected by detachment of placenta) (Figs. S13–S16).

Quantification of PRRSV RNA Concentration

Samples of MFI corresponding to each fetus were placed in weigh boats (endometrium and placenta facing up, myometrium facing down) on ice for up to 45 minutes. After gently washing, they were placed with myometrium side down on absorbent tissue, where the fetal placenta was manually separated from endometrium using sterile forceps. RNA extraction from placental and endometrial samples and quantitative polymerase chain reaction (PCR) targeting PRRSV NVSL 97-7895 was performed as previously described in detail. ¹¹ Briefly, RNA was extracted from 30 mg of tissue using the RNeasy MiniKit (QIAGEN, Toronto, ON, Canada) performed on the Qiacube Robotic workstation, and RNA concentration and quality (A260/A280 ∼ 2) were assessed using 2 μL of each sample with a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA). Primers and probe were designed to target a highly conserved region at the C-terminal end of ORF7 of NVSL-97-7895. Reactions (20 μL) contained 2 μL RNA extract, 10 μL Mastermix (Bio-Rad iTaq Univeral Probes 1-Step Kit; Bio-Rad Laboratories, Mississauga, ON, Canada), 0.5 μL reverse-transcriptase/RNase block enzyme Supermix (iScript Reverse transcriptase; Bio-Rad), 0.5 μL of 10 μM PRRS-2F primer 5′-TAA TGG GCT GGC ATT CCT-3′, 0.5 μL of 10 μM PRRS-1 R primer 5′-ACA CGG TCG CCC TAA TTG-3′, 0.5 μL of 10 μM PRRS-P1 probe 5′-HEX-TGT GGT GAA TGG CAC TGA TTG RCA-BHQ2-3′ (reported initially by Kleiboeker et al ⁸ ), and 6 μL RNAse-free water. Each PCR plate contained 35 samples in duplicate, extraction controls, positive control (known positive tissue homogenate), and a quantitative PCR nontemplate control. A dilution series (10 ⁷ –10¹ copies/μL) of HindIII linearized plasmid, pCR2.1TOPO-NVSL, containing a 446-bp sequence of ORF7 was used as a standard curve, run in triplicate on each plate. The PCR protocol was reverse transcription for 30 minutes at 50°C followed by 10 minutes at 95°C, 40 cycles of denaturation for 30 seconds at 95°C, and annealing/extension for 30 seconds at 59°C. Individual samples were rerun if the Cq standard deviation between the duplicates was >1.5 or if 1 of the 2 duplicates had no Ct. Results were reported as log₁₀ target RNA concentration per milligram of tissue. The limits of quantification were defined by the least and most concentrated standards. Samples were recorded as negative if target RNA was not detected or detected but not quantifiable if target RNA was detected but the concentration was below the least concentrated standard. All PCR plates contained appropriate nontemplate and extraction controls and a positive control sample (serum or tissue as appropriate) to track interplate variation. Across all plates, the PCR reaction efficiency averaged 99.85 ± 4.94%.

Statistical Analyses

All statistical analyses were performed using Stata 14 (StataCorp, College Station, TX). Descriptive statistics were used to summarize the chronologic progression of fetal lesions over the 5 termination days. The frequency of fetal lesions in VIA compared with compromised (MEC + DEC) fetuses was assessed using a Fisher exact test. MEC and DEC fetuses were combined into a single group (compromised) because there was an insufficient number in each group. Group differences (PRRSV inoculated vs noninoculated) in the severity of each lesion (endometritis, placentitis, vasculitis in the endometrium and fetal placenta, placental separation) were evaluated using a 2-sided Fisher exact test, first by assessing differences across all 5 termination days together and, if significant, by subsequently looking at each individual day. To validate these results, differences in median lesion severity scores between PRRSV inoculated and noninoculated were assessed at each termination day using a Mann-Whitney U test. To determine if lesion severity progressed over time in the PRRSV-inoculated group, median lesion severity scores were compared across termination day using a Kruskal-Wallis test, followed by a post hoc Dunn test (with Sidak multiple comparison adjustment). Finally, the association between the placental separation severity score in the PRRSV-infected samples and PRRS viral load in the endometrium and fetal placenta, severity of endometritis, placentitis and vasculitis at MFI, termination day, and fetal preservation status was analyzed using single-level proportional odds model. For the purpose of these analyses, the values for PRRSV RNA concentration in the endometrium and placenta and placental separation severity scores were categorized on the basis of the natural break points in data. Endometrial viral load categories were 0 if PRRSV RNA concentration was <3 log₁₀ copies/mg, 1 if ≥3 and ≤5.75, and 2 if >5.75. Placental viral load categories were 0 if PRRSV RNA concentration was <2 log₁₀ copies/mg, 1 if ≥2 and ≤5.5, and 2 if > 5.5. For placental separation, because of relatively small numbers of samples with moderate and severe scores compared with normal and mild severity, the moderate and severe severity categories were combined.

Fetal Histopathology

The prevalence of fetal histologic lesions by termination day are summarized in Table S1. The presence of any fetal lesion was 23 times (95% confidence interval, 8.5–62.5) more likely in compromised (MEC + DEC; 13 of 22 [59.1%]) compared with VIA (15 of 239 [5.9%]) fetuses (P < .001). Pathologic changes in fetal tissues observed were minimal to mild in severity and not specific. In total, 276 fetal brains, lungs, hearts, and livers, 228 thymuses, and 219 mesenteric lymph nodes were examined. The first pathologic change in fetal organs was observed 5 dpi in the thalamus and consisted of mild focal areas of gliosis (samples of diencephalon were systematically collected in this study to assess potential neuronal necrosis, particularly in the hypoxia-sensitive hippocampus). ³¹ Similar areas of mild gliosis were also observed in the cortex of diencephalon 8, 12, and 14 dpi. This change was not observed in negative control fetuses. Also, no neuronal necrosis was observed in the PRRSV-infected or negative control samples. Lesions in heart, previously reported to be consistent fetal lesions of PRRS, ²² were first observed in this study 8 dpi and consisted of small aggregates of lymphocytes between cardiomyocytes (mild focal lymphocytic myocarditis), but there was no dramatic change in the prevalence or severity of this lesion over time. No necrosis of cardiomyocytes was seen in this study, including in the negative control samples. The only pathologic change observed in the lung was the occasional presence of small to moderate amounts of meconium in the airways, which was seen starting from 8 dpi and only in MEC and DEC fetuses. Meconium in the bronchioles was also seen in 1 MEC negative control fetus. One PRRSV-infected fetus exhibited a mild increase in lymphocytolysis in the cortex of the thymus 12 dpi, but no thymic atrophy was observed in the whole study population. Finally, 14 dpi, grossly enlarged mesenteric lymph nodes were observed in 5 fetuses at necropsy that microscopically corresponded to mild to moderate follicular hyperplasia (reactive).

Histopathology of the MFI

Results of histopathologic lesions at the MFI are summarized in Table S2. The earliest lesion in the endometrium of PRRSV-infected gilts was focal vasculitis observed 2 dpi affecting medium-size venules in the endometrium (Fig. 1). Focal vasculitis (<30% blood vessels affected) was characterized by accumulation of small numbers of lymphocytes and histiocytes under the endothelial lining without associated damage. By 8 dpi, more than half of the PRRSV-infected samples exhibited multifocal (30%–70% blood vessels) or diffuse (>70% blood vessels) endometrial vasculitis. On 12 dpi, all PRRSV-infected samples exhibited varying degrees of vasculitis in the endometrium, with the majority of samples having a diffuse pattern of distribution (Fig. 2). The severity of vasculitis increased across days, progressively affecting medium-size arterioles and occasionally larger veins, but the predominant inflammatory cell types of vasculitis, lymphocytes, and histiocytes remained the same. The endothelial cell lining remained intact in endometrial tissue samples on all days postinoculation. No vasculitis was observed in PRRSV-infected fetal placental samples or in the endometrium or placenta of noninfected control samples.

OPEN IN VIEWER

Figures 1–6. Maternal-fetal interface of pregnant gilts infected by porcine reproductive and respiratory syndrome virus 2 (PRRSV-2), pig, uterus and placenta. Figure 1. Uterus and endometrium, 2 days postinoculation. Focal vasculitis in a medium-sized venule (arrow and inset at higher magnification). Hematoxylin and eosin (HE). Figure 2. Uterus and endometrium, 12 days postinoculation. Diffuse vasculitis (arrow and inset). HE. Figure 3. Chorioallantois, 14 days postinoculation. Placental macrophages within the chorioallantoic stroma. HE. Figure 4. Uterus with placenta. Mild placental separation characterized by focal area of placental detachment (arrow). HE. Figure 5. Uterus with placenta. Moderate placental separation characterized by locally extensive area of placental detachment (arrow and inset). HE. Figure 6. Uterus with placenta. Severe placental separation characterized by diffuse area of placental detachment (arrow and inset). HE.

In PRRSV-infected animals, severe endometritis was observed for the first time 8 dpi, and by 14 dpi, more than half of samples were affected. However, placentitis in PRRSV-infected samples differed from endometritis in its severity (mostly milder, rarely moderate or severe) and by type of inflammatory cells involved. Severe placentitis was seen in only a single sample at 12 dpi. Although endometritis was characterized by large numbers of lymphocytes, small numbers of histiocytes, and rare plasma cells, placentitis was composed exclusively of varying numbers of plump vacuolated macrophages residing along the placental mesenchyme in row-like formation in the area between MFI and placental vasculature (Fig. 3). PRRSV-negative control animals had either normal endometrium or minimal endometritis, except for single samples at 2, 12, and 14 dpi that scored mild, mild, and moderate, respectively. As for placentitis in control animals, all samples were normal or minimal, except for a single 14-dpi sample that scored mild.

Placental separation was observed only microscopically (not grossly) and characterized mainly by moderate to severe inflammation in the endometrium and fetal placenta, marked acute vacuolar degeneration of detached trophoblastic epithelium, mild to moderate focal areas of necrosis of uterine epithelium, and varying amounts of necrotic debris in the detachment gaps. Mild separation was commonly observed in noninfected control gilts. Moderate and severe placental separation were each observed in 1 noninfected fetus, both VIA. The only noninoculated (control) MEC fetus showed no evidence of placental separation. In the PRRSV-infected MFI samples, frequent areas of mild (Fig. 4) and moderate (Fig. 5) microscopic placental separation were both seen from 2 to 8 dpi. Severe placental separation (Fig. 6) was observed 12 and 14 dpi only.

Group Differences in Microscopic Lesions at the MFI

Potential differences in lesion severity in MFI samples from PRRSV-inoculated and noninoculated gilts were assessed using 2 methods. In the first assessment method using contingency tables, Fisher exact tests revealed that the endometrial vasculitis, endometritis, and placentitis severity scores differed significantly in PRRSV-inoculated compared with negative control samples (P < .001 for all) when all days were considered together. Further analyses of group differences in these lesion scores at each termination day revealed that endometritis severity (Fig. 7) differed on all 5 termination days, whereas placentitis severity (Fig. 8) and vasculitis severity (Fig. 9) differed at 5 to 14 dpi but not 2 dpi, indicating that endometrial vasculitis and placentitis require more time to develop. Although placental separation severity scores (Fig. 10) also differed between PRRSV-inoculated and noninoculated samples when all days were considered together (P = .037), the group differences were driven mainly by a significant difference in severity at 5 dpi.

OPEN IN VIEWER

Figures 7–10. Median severity and distribution scores for histologic lesions across different postinoculation time points in porcine reproductive and respiratory syndrome virus 2 (PRRSV-2)–infected and noninfected control maternal-fetal interface (MFI) samples. Figure 7. Endometritis severity. Figure 8. Placentitis severity. Figure 9. Distribution of endometrial vasculitis. Figure 10. Placental separation severity. Error bars represent interquartile ranges. Superscripts within each figure indicate statistically significant changes in severity score across the 5 termination days (2, 5, 8, 12, and 14 days postinoculation [dpi]) for the PRRSV-inoculated compared with control group.

The second method of assessment involved the comparison of median lesion scores by group (Mann-Whitney U test), initially across all termination days, then individually by day. Differences in median rank lesion severity scores between PRRSV-inoculated and noninoculated samples were identical to those using the contingency method.

Progression of Microscopic Lesions at the MFI Over Time

These analyses included samples from PRRSV-inoculated gilts only and were based on the hypothesis that lesion severity at the MFI would increase over time, potentially leading to more severe placental separation. Spatiotemporal changes in the severity of vasculitis, endometritis, placentitis, and placental separation are shown in heat maps in which darker colors represent increased severity (Figs. S17–S20). On the basis of a Dunn test, the median severity ranks for endometritis, placentitis, and endometrial vasculitis increased over time (P < .001 for all) (Figs. 7–9). Surprisingly, placental separation (Fig. 10) did not differ across day of termination, when all PRRSV-inoculated fetuses were considered (P = .85), or when only considering the viable fetuses (P = .56).

Factors Associated With Placental Separation

Having found wide spatiotemporal variation in placental separation scores (Fig. S20) but no progression in severity over time in the PRRSV-inoculated group, together with overall group differences in severity mainly driven by a single day (5 dpi), a proportional odds model was used to better understand factors potentially associated with placental separation severity. Although the study population was naturally clustered by gilt, a 2-level hierarchical model was not required. Of 7 factors unconditionally evaluated, only placental inflammation severity, vasculitis distribution, and fetal preservation were unconditionally associated with placental separation severity (P < .20) and thus included in the full model. Importantly, termination day, endometritis severity, and viral load in endometrium and placenta were not associated with placental separation score. After backward stepwise elimination, placentitis severity remained the most significant factor associated with placental separation (P < .004), whereas fetal preservation score only trended toward significance (P = .10). Post hoc pairwise analysis revealed that placental separation severity was associated only with fetal death; the probability of a DEC fetus having moderate or severe placental separation was .51 ± .16, compared with .06 ± .04 for MEC and .09 ± .05 for VIA fetuses (Fig. 11). The relationship of placental separation severity to placentitis is shown in Fig. 12. The probability of severe or moderate placental separation was .26 ± .13 if placentitis severity was mild to severe, compared with less than .08 if placentitis was minimal or none.

OPEN IN VIEWER

Figures 11 and 12. Factors associated with the probability of placental separation severity score. The color of the bar represents the placental separation severity: none or absent (light gray), mild (darker gray), moderate and severe (black). The y axis indicates the probability of fetuses having a given placental severity score. Figure 11. Relationship of fetal preservation status to placental separation. The probability of moderate or severe placental separation (black bars) in viable (VIA) and meconium-stained (MEC) fetuses is less than 10%, whereas it is more than 50% in decomposed (DEC) fetuses. Figure 12. Relationship of placental inflammation score to placental separation. The probability of moderate or severe placental separation (black bars) in fetuses with no or minimal placental inflammation is less than 10%, whereas it is 26% in fetuses with moderate or severe placental inflammation. Together these bar plots indicate that moderate and severe placental separation is associated with fetal death (decomposition) and placental inflammation but is not associated with meconium staining. Error bars represent standard deviation.