Maternal immune activation and dietary soy isoflavone supplementation influence pig immune function but not muscle fiber formation

Abstract

The goals of this study were to determine the impact of maternal PRRSV infection on offspring muscle and immune development and the potential of dietary soy isoflavones to mitigate those effects. Thirteen first-parity gilts (“gilts”) were randomly allotted into one of three treatments: not infected and fed a diet devoid of isoflavones (CON), infected with porcine reproductive and respiratory syndrome virus (PRRSV) and fed the control diet (POS) or that supplemented with 1,500 mg/kg soy-derived isoflavones (ISF). Gilts were inoculated with PRRSV intranasally on gestational day (GD) 70. After farrowing (GD 114 ± 2), 1–2 offspring (“pigs”) closest to the average litter weight were selected either at birth (3 ± 2 d of age) or weaning (21 ± 2 d of age) to determine body, muscle, and organ weights as well as muscle cell number and size. Four weaned pigs of average body weight within each litter were selected for postnatal immune challenge. At PND 52, pigs were injected with 5 µg/kg BW lipopolysaccharide (LPS) intraperitoneally. Serum was collected at 0, 4, and 8 h following LPS administration to analyze tumor necrosis factor alpha (TNF-α). At PND 59, pigs were administered a novel vaccine to elicit an adaptive immune response. At PND 59, 66, and 73, peripheral blood mononuclear cells were isolated and T-cell populations determined by flow cytometry. Both POS and ISF pigs exhibited persistent PRRSV infections throughout the study (PND 1-73). At PND 3, whole body, muscle, and organ weights were not different (P > 0.22) between groups, with the exception of relative liver weight, which was increased (P < 0.05) in POS compared with CON pigs. At PND 21, ISF pigs had reduced (P ≤ 0.05) whole body and muscle weights, but greater (P < 0.05) kidney weight compared with CON, and greater (P < 0.05) relative liver weight compared with CON and POS. Muscle fiber number and size were not different (P > 0.39) between groups at birth or weaning. After LPS administration, TNF-α was greatest in ISF pigs (P < 0.05) at both 0 and 8 h post-challenge. At the peak time-point of 4 h post-challenge, ISF pigs had the greatest concentration of TNF-α and CON pigs had the lowest, with POS pigs being intermediate (P = 0.01). After vaccination, ISF offspring had shifts in T-cell populations indicating an impaired immune response. These data indicate that maternal PRRSV infection may impact offspring organ growth and immune function, particularly when the dam is supplemented with isoflavones.

Maternal infection with porcine reproductive and respiratory syndrome virus altered both innate and adaptive immune responses in offspring, and maternal isoflavone supplementation did not mitigate these effects. Maternal infection did not impact offspring muscle development.

Introduction

Although porcine reproductive and respiratory syndrome virus (PRRSV) is estimated to infect approximately 45% of the U.S. sow herd annually (Neumann et al., 2005), very little is known about the effects of PRRSV infection during gestation on the muscle development, postnatal growth, and immune function of offspring. Infection with PRRSV compromised performance of sows (Neumann et al., 2005; Valdes-Donoso et al., 2018; An et al., 2020), young pigs (Rochell et al., 2015; Smith et al., 2019), and finishing pigs (An et al., 2020; Smith et al., 2020), due in part to the inflammatory nature of the disease. Infection with PRRSV resulted in increased circulating pro-inflammatory cytokines including TNF-α, IL-6, and IL-1β (Thanawongnuwech et al., 2001; Liu et al., 2010; An et al., 2020; Lu et al., 2020). Given the prevalence of PRRSV in sow herds, nearly half of pigs produced in the United States may have been prenatally exposed to PRRSV and its concomitant inflammation.

Chronic maternal inflammation induced by obesity was associated with increased cytokine presence in cord blood and potential activation of fetal immune responses (Yockey and Iwasaki, 2018). Maternal exposure to LPS resulted in increased fetal plasma IL-1β concentrations (Ashdown et al., 2005), which suggests that acute maternal inflammation may impact fetal signaling cascades. Although PRRSV infection is associated with chronic rather than acute inflammation in the sow, fetal thymic gene expression suggests that the fetal immune response is more akin to an acute inflammation cascade (Hong et al., 2016). Further, maternal helminth infection led to maternal imprinting of cytokine profiles (Djuardi et al., 2009) and may promote protective immunophenotypes (Brand et al., 2011). It is important to note that variability between individual animal’s immune responses to pathogens is a critical factor in determining not only the type of environment the fetus is exposed to (immune response at the maternal level) but also the robustness of the fetal response (immune response at the fetal level). For example, PRRSV-infected gilts and their litters displayed varying levels of susceptibility or resistance to viral infection that was attributed to inherent differences between animals at either the gilt or fetal level (Malgarin et al., 2019). It is apparent that viral invasion of fetal tissues is necessary for a robust fetal immune response (Malgarin et al., 2019; van Goor et al., 2020), and that the extent of invasion may be controlled by the heartiness of the sow’s immune system as evidenced by the relationship between placental gene expression and fetal viability (van Goor et al., 2020). However, the influence of maternal immune activation during critical periods of fetal immune development on offspring postnatal innate and adaptive responses is relatively unknown.

Further, little is known regarding the effects of maternal immune activation on offspring muscle development, an important factor in overall growth potential. It is understood that animals which experienced prenatal undernutrition had reduced birth weight due to suppressed secondary muscle fiber formation (Dwyer et al., 1995; Picard et al., 2002; Zhu et al., 2006; Reed et al., 2014). Inflammation can be responsible for induction of sickness behaviors, including anorexia (Johnson and von Borell, 1994; Leininger et al., 2000; Jo et al., 2007; Colpoys et al., 2019), and when induced in gestating animals may result in prenatal undernutrition. However, inflammation does not always cause a meaningful reduction in feed intake (Mani et al., 2013; Labussiére et al., 2015; Merlot et al., 2016). Maternal inflammation alone has been linked with downregulated myocyte differentiation (Du et al., 2010). This suggests that offspring exposed to maternal infection may have reduced muscle fiber development, limiting their ability to accrete lean mass postnatally.

Dietary interventions hold promise to improve immune responses to viral and other infections. When supplemented to pigs with PRRSV, dietary soy isoflavones have been shown to reduce viral load and systemic inflammation (Greiner et al., 2001; Rochell et al., 2015), and improve growth and minimize mortality (Smith et al., 2019; Smith et al. 2020). Therefore, the objective of this study was to analyze the effect of dam immune activation during late gestation on muscle development and postnatal immune responses in offspring pigs. It was hypothesized that maternal PRRSV infection would disrupt both muscle development and the innate and adaptive immune responses in pigs, and that maternal supplementation with dietary soy isoflavones would mitigate these effects.

Materials and Methods

All procedures were reviewed and approved by the Institutional Animal Care and Use Committee (protocol 17279) of the University of Illinois.

Gilt husbandry and experimental design

A total of 24 first-parity gilts (PIC Cambrough; hereafter referred to as “gilts”) were obtained from the PRRSV-naïve University of Illinois Swine Research Center herd. Gilts were not inoculated against PRRSV during their lifetime, and all were confirmed to be PRRSV-negative via qRT-PCR using Z-PRRSV Multiplex reagent (Tetracore, Rockville, MD) at the University of Illinois Veterinary Diagnostic Laboratory. Upon enrollment on study, gilts were divided into three cohorts (n = 8 in each cohort) to accommodate space in the biocontainment facility. Gilts were transported from the University of Illinois Swine Research Center to the Edward R. Madigan Laboratory animal care biocontainment facility on gestational day (GD) 65. Gilts were housed in farrowing crates in disease containment chambers, with 1 gilt residing in each chamber. Each chamber was individually HEPA-filtered and strict biosecurity protocols and traffic patterns were enforced throughout the study period. Gilts were randomly allotted to one of three treatments: not infected and fed a nutritionally adequate control diet formulated with soy protein concentrate to be functionally devoid of isoflavones (CON; n = 6), or infected with PRRSV and fed the control diet (POS; n = 9), or that supplemented with 1,500 mg/kg of a soy-derived isoflavone mixture (ISF; n = 9). A description of the ISF source and full dietary compositions can be found in Smith et al. (2020).

Gilts were provided 3.2 kg of their assigned diet once daily as is standard industry practice. Observations of each animal were recorded twice daily to monitor the longitudinal health status and track clinical signs of illness throughout the study. Upon arrival, gilts were given a 5-d acclimation period to adjust to the facility and dietary treatments prior to induction of the gestational challenge. On gestation day 70, POS and ISF gilts were inoculated intranasally with 2.5 × 10⁴ tissue culture infected dose (TCID)₅₀/mL of suspended live PRRSV (NADC₂₀ strain; courtesy of Dr. Federico Zuckermann, University of Illinois, Urbana, IL), with half of the dose administered in each nostril; CON gilts were administered a sham inoculation (sterile saline) in the same manner.

Blood obtained from jugular venipuncture was collected on 0, 7, 14, and 21 d post-inoculation (DPI) and allowed to clot at room temperature. Samples were then centrifuged at 1,300 × g and 20 °C for 20 min to collect serum. Serum samples were submitted to the University of Illinois Veterinary Clinical Pathology Laboratory to be analyzed for PRRSV by qRT-PCR using Z-PRRSV Multiplex reagent (Tetracore, Rockville, MD). A result of non-detectable (ND) was considered negative. A cycle threshold (Ct) count of 0 < Ct ≤ 38 was considered positive, and Ct > 38 was considered suspect.

Throughout the course of gestation, 7 of the 24 gilts were removed from the study. One ISF gilt was removed from the study due to extreme anorexia. Additionally, 2 POS and 4 ISF gilts, despite inoculation with PRRSV, never tested positive for PRRSV and were removed from the study as per veterinary recommendation. All data from gilts removed from study were excluded from the final analysis. One POS gilt never tested positive for PRRSV and did not exhibit any clinical symptoms, so with veterinary approval was reassigned to the CON treatment. This resulted in 7 CON, 6 POS, and 4 ISF gilts completing the study and farrowing naturally on GD 114.

Offspring (hereafter referred to as “pigs”) mortality data were collected on all litters. Total pigs per litter were considered the number of pigs born alive plus the number of stillborn pigs. The percentage of pigs born alive was calculated as the number of live pigs at farrowing divided by the total pigs per litter. Survival during the suckling period was calculated by the number of pigs successfully weaned divided by the number of pigs born alive, excluding pigs that were euthanized at PND 3 for sample collection. One POS gilt had 100% litter mortality upon farrowing and therefore did not contribute to postnatal data collection. Therefore, 7 CON, 5 POS, and 4 ISF litters contributed to postnatal data collection. Of these, 7 CON, 4 POS, and 3 ISF litters contributed to PND 3 and PND 21 dissection data. Litters that were excluded had fewer than 6 live pigs at farrowing, and none were culled to ensure survival of at least 4 pigs to weaning for the postnatal immune challenges.

Offspring husbandry and sample collection

At postnatal day (PND) 3 (± 2 d), one male and one female pig closest to the average litter weight were selected for dissection. In some POS and ISF litters, there were inadequate numbers of pigs to allow for selection of both a male and female. In those cases, only one pig was selected. Pigs were euthanized via sedation with a telazol:ketamine:xylazine (100 mg/mL; Telazol C IIIN, 100 mg/mL; KetaVed C III, and 100 mg/mL; Pivetal Anased, Patterson Veterinary Supply, St. Paul, MN) mixture followed by sodium pentobarbital (390 mg/mL; Euthasol Euthanasia Solution C IIIN, Patterson Veterinary Supply, St. Paul, MN) via intracardiac administration. The left and right semitendinosus, left longissimus dorsi, left psoas major, heart, lungs, liver, and both kidneys were extracted and weighed. Relative weights were determined by the following equation:

$${\displaystyle \begin{array}{ll}{\displaystyle \mathrm{R}\mathrm{e}\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{w}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t},\mathrm{\%}=}& (\mathrm{W}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e},\mathrm{k}\mathrm{g}/\\ & \mathrm{W}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\mathrm{o}\mathrm{f}\mathrm{w}\mathrm{h}\mathrm{o}\mathrm{l}\mathrm{e}\mathrm{p}\mathrm{i}\mathrm{g},\mathrm{k}\mathrm{g})\ast 100.\end{array}}$$

Lung tissue samples were dissected and flash frozen using liquid nitrogen. Samples were submitted for PRRSV testing via qRT-PCR at the University of Illinois Veterinary Diagnostic Laboratory. In the event that a result for a particular sample was not congruent with the littermate collected, biosecurity was appropriately adjusted and samples were resubmitted the following week. If the second submission produced a result different from the first, or if the remaining litter remained consistent in their results, the outlier test was considered a false positive or negative and subsequently removed from analysis. The left semitendinosus muscle was submerged in formalin for at least 48 h. Samples were then stored in 70% ethanol at ambient temperature (approx. 23 °C) until further processing. Sections were cut perpendicular to the muscle fibers, individually placed in plastic cassettes, and submitted to the University of Illinois Veterinary Diagnostic Laboratory for hematoxylin and eosin staining. Slides were visualized using a compound microscope. The entire sample was visualized, and areas representative of the entire samples morphology were chosen for imaging. At least 5 digital images of each sample were obtained. These images were then uploaded into Photoshop version 21 (Adobe, San Jose, CA) and the pixel to μm conversion ratio determined using the ruler tool. Muscle fiber area was measured by tracing the perimeter of an individual fiber using the lasso tool. A cross section was obtained from the middle of the semitendinosus muscle. The area of the semitendinosus muscle was determined by tracing the perimeter of this cross section onto acetate paper, then converting these tracings to a Photoshop image using an Intuos tablet (model no. PTK-840, Wacom, Kazo, Japan). The pixel to μm conversion ratio was determined using the ruler tool, and muscle area was measured using the lasso tool. Estimated total fiber number was determined by the following equation:

$${\displaystyle \begin{array}{ll}{\displaystyle \mathrm{E}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{m}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{f}\mathrm{i}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}=}& (\mathrm{A}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{o}\mathrm{f}\mathrm{s}\mathrm{e}\mathrm{m}\mathrm{i}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{o}\mathrm{s}\mathrm{u}\mathrm{s}\mathrm{m}\mathrm{u}\mathrm{s}\mathrm{c}\mathrm{l}\mathrm{e},\mu {\mathrm{m}}^2/\\ & \mathrm{A}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{o}\mathrm{f}\mathrm{i}\mathrm{m}\mathrm{a}\mathrm{g}\mathrm{e},\mu {\mathrm{m}}^2)\\ & \times \mathrm{A}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{m}\mathrm{a}\mathrm{g}\mathrm{e}.\end{array}}$$

Remaining pigs were processed at PND 3 following standard industry procedures. In brief, needle teeth were removed with nippers and tails were docked using sharpened surgical scissors. Pigs received intramuscular injections of 1 mL iron dextran and 0.10 mL Excede (Zoetis, Parsippany-Troy Hills, NJ). Male pigs were castrated using a scalpel. All pigs were monitored for complications during routine twice daily observations.

At PND 21, one male and one female pig closest to the average litter weight were euthanized, and samples were extracted and weighed as described above. Lung tissue was again submitted for PRRSV testing as described above. From the remaining litter, a total of four pigs closest to the average weight of the litter (n = 16 CON, n = 16 POS, n = 12 ISF) were removed and transported to separate containment housing within the facility. Pigs were housed by litter and given a standard nursery diet ad libitum. Once-weekly oral fluids were collected from each litter by introducing a 100% cotton rope to the pen for 15 min. Fluids were then manually extracted into tubes and submitted for PRRSV testing via qRT-PCR at the University of Illinois Veterinary Diagnostic Laboratory. In the event that a result for a particular sample was not congruent with expectations or clinical signs, biosecurity was appropriately adjusted and samples were resubmitted the following week. If the second submission produced a result different from the first, or if the remaining litter remained consistent in their results, the outlier test was considered a false positive or negative and subsequently removed from analysis.

Innate immune challenge

Litters from cohorts 1 and 2 (4 CON, 4 POS, and 3 ISF litters) were used for postnatal immune challenge analysis. Due to COVID-related shutdowns, no postnatal immune challenges were possible for the third cohort. At PND 52, pigs were administered 5 µg/kg BW reconstituted lipopolysaccharide (LPS; E. coli strain O111:B4, Sigma Aldrich, St. Louis, MO) via a single intraperitoneal injection. All pigs had ad libitum access to feed prior, during, and after the immune challenge. Blood obtained from the subcutaneous abdominal vein was collected at 0, 4, and 8 h after LPS administration. Samples were allowed to clot at room temperature. Blood was then centrifuged at 1,300 × g and 20 °C for 20 min before serum was collected and frozen at −80 °C pending analysis.

Serum samples were analyzed using the porcine TNF-α enzyme-linked immunosorbent assay (ELISA) kit (ThermoFisher Scientific, Waltham, MA) according to the manufacturer’s protocol. Standards (1,500, 750, 375, 187.5, 93.8, 46.9, 23.4, and 0 pg/mL) were created using the provided stock solution (7,500 pg/mL). These standards were used to calculate the linear standard curve. Plates were read at 450 nm as per the included protocol. Absorbances were then applied to the standard curve equation and multiplied by the dilution factor to calculate TNF-α concentration. Manufacturer calculated CV for inter- and intra-assay precision for this kit were 7.3 and 6.0, respectively.

Adaptive immune challenge

At PND 59, the same pigs utilized in the innate immune challenge were exposed to a single 2-mL dose of a porcine circovirus vaccine as per manufacturer instructions (Circumvent PCV-M G2 Swine Vaccine; ValleyVet, Marysville, KS) to induce adaptive immune responses, as reviewed by Kekarainen et al. (2010). Whole blood was collected at 0, 7, and 14 d post-vaccination (DPV) and used for a T-cell immunophenotyping protocol using flow cytometry as described by Smith et al. (2020). Percentages displayed are the incidence of each variable population out of the 10,000 cells phenotyped per sample and populations were defined as follows: total T cells (CD3+), helper T cells (CD3+CD4+), cytotoxic T cells (CD3+CD8+), and dual positive T cells (CD3+CD4+CD8+). The proportion of undifferentiated T cells at each time-point was calculated using the following equation:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{U}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s},\mathrm{\%}=100(\mathrm{H}\mathrm{e}\mathrm{l}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s},\mathrm{\%}}\\ +\mathrm{C}\mathrm{y}\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{o}\mathrm{x}\mathrm{i}\mathrm{c}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s},\mathrm{\%}+\mathrm{D}\mathrm{u}\mathrm{a}\mathrm{l}\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s},\mathrm{\%}),\end{array}}$$

$${\displaystyle \begin{array}{ll}& \mathrm{U}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s}=\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s}(\mathrm{H}\mathrm{e}\mathrm{l}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s}\\ & +\mathrm{C}\mathrm{y}\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{o}\mathrm{x}\mathrm{i}\mathrm{c}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s}+\mathrm{D}\mathrm{u}\mathrm{a}\mathrm{l}\mathrm{P}\mathrm{o}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{T}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{s}).\end{array}}$$

Statistical analysis

Infection status data were analyzed separately from innate and adaptive immune activation outcomes. Gilt performance, gilt infection status, offspring mortality, and offspring infection status data were analyzed as a one-way ANOVA with cohort as a random variable in the MIXED procedure of SAS 9.4 (SAS Institute, Inc., Cary, NC). Gilt served as the experimental unit for these analyses. Litter served as the experimental unit for mortality and offspring infection analyses. As lower Ct values can indicate greater severity of infection, but a Ct value of 0 indicates a negative result, only the pairwise P-value between POS and ISF treatments is reported to denote differences in infection severity between PRRSV-positive treatments. Pig dissection data were analyzed as a two-way ANOVA in the MIXED procedure of SAS, with the main effects of treatment (CON, POS, or ISF), sex (barrow or gilt), and their interaction. Pig served as the experimental unit for these analyses. Both main effects of sex and treatment × sex interaction had no statistically significant impact on any outcomes; thus, the statistical model was simplified to only include the fixed effect of treatment. Innate and adaptive immune data were analyzed using a split plot design, with litter serving as the whole plot and individual offspring pig serving as the split plot. Data were then run as a two-way ANOVA in the MIXED procedure of SAS, with the main effects of treatment (CON, POS, or ISF), time after dosing (0, 4, or 8 h post-LPS administration or 0, 7, or 14 d post-vaccination, respectively), and their interaction. Repeated measurements (each pig sampled multiple times) were accounted for using an unstructured and compounded covariance structure for the innate and adapted data, respectively. Least squares means were separated using the probability of difference (PDIFF) option and means were considered significantly different at P ≤ 0.05. For illustrative purposes, area under the curve for innate immune response was calculated via the trapezoidal method, using treatment least squares means and without baseline subtraction.

Results

Infection status

All gilts tested negative for PRRSV at trial initiation (Table 1). CON gilts remained negative throughout the study. All POS and ISF gilts (excluding those removed from study) tested positive on DPI 7, 14, and 21. Infection severity as measured by Ct count was similar between POS and ISF gilts at DPI 7 and 14; however, CT counts for ISF gilts were increased compared to POS on DPI 21 (P = 0.04). Although rectal temperatures differed statistically between groups at 4 non-consecutive days during the infection period (P < 0.05; Figure 1), all recorded temperatures were within a range considered normal for an adult pig (36.0–39.0 °C). Infection status impacted litter mortality, with both POS and ISF gilts having fewer pigs born alive and increased incidence of stillborn and mummy pigs (P ≤ 0.04; Table 2). This ultimately leads to POS and ISF gilts having an average of 5 fewer pigs weaned per litter compared to CON gilts (P = 0.04). POS and ISF gilts also had reduced average weight of pigs born alive compared to CON gilts (P = 0.02). One CON offspring euthanized at PND 21 had a Ct count above 38, considered suspect for this assay. Biosecurity was adjusted to protect the other CON litter from possible infection. Because a littermate euthanized at the same time-point was negative, and the remaining litter tested negative for the remaining study period, this test was considered a false positive and this pig was removed from analysis. Therefore, CON pigs tested negative at each time-point (Table 3). Both POS and ISF pigs tested positive throughout the study period. Although not statistically different, ISF pigs had numerically increased Ct counts compared to POS pigs from DPI 28–63, suggesting a reduced viral load.

Table 1.

Gilt infection status over time as represented by viral Ct counts¹

	Treatment
Item	CON	POS	ISF	SEM	P value2
Gilts3, n	7	6	4
DPI 0	ND	ND	ND	.	.
DPI 7	ND	15.20	14.34	5.17	0.90
DPI 14	ND	26.36	24.78	2.52	0.65
DPI 21	ND	27.05	32.45	1.49	0.04

Ct, ND (non-detectable) considered negative, 0 < Ct≤ 38 considered positive, Ct > 38 considered suspect.

Pairwise P-value between POS and ISF treatments.

Gilt served as the experimental unit.

Abbreviations: CON, control treatment (unsupplemented diet without infection); Ct, cycle threshold; DPI, days post-inoculation; ISF, isoflavone-supplemented treatment (with infection); POS, positive control (unsupplemented diet with infection); SEM, standard error of the mean.

Figure 1.

The effect of PRRSV infection status and dietary supplementation with soy isoflavones on gilt rectal temperature. abLeast squares means within a time point lacking a common superscript letter differ (P < 0.05). Gilt served as the experimental unit. Abbreviations: CON, uninfected and fed a diet devoid of soy isoflavones; POS, infected with PRRSV and fed the control diet; ISF, infected with PRRSV and fed the control diet supplemented with 1,500 ppm soy isoflavones. 60 × 36 mm (300 x 300 DPI).

Table 2.

Effects of maternal PRRSV infection and dietary treatment on offspring mortality

Item	Treatment			SEM	P-value
	CON	POS	ISF
Litter1, n	7	6	4
Total piglets per litter2, n	14	12	12	1.76	0.54
Piglets born alive, n	14	9	10	1.81	0.08
Piglets born alive, %	98.19a	75.05b	77.54b	5.84	0.04
Weight of piglets born alive, kg	1.47a	1.11b	1.18b	0.11	0.02
Stillborn piglets, n	0a	3b	3b	0.66	0.04
Mummy, n	0a	5b	3ab	1.26	0.03
Pigs weaned per litter, n	11a	6b	7b	1.54	0.04
Weight of piglets weaned, kg	6.52	4.82	5.39	0.58	0.09
Survival during suckling period3, %	94.5	76.33	80.77	5.55	0.15

LS means within a row having differing superscripts are statistically different (P < 0.05).

Litter served as the experimental unit.

Total piglets per litter = no. piglets born alive + no. of stillborns.

Survival during suckling period, % = (no. of piglets successfully weaned/ total piglets per litter excluding piglets euthanized at PND 3 for sample collection) × 100.

Abbreviations: CON, control treatment (unsupplemented diet without infection); POS, positive control (unsupplemented diet with infection); ISF, isoflavone-supplemented treatment (with infection); SEM, standard error of the mean.

Table 3.

Piglet infection status over time as represented by viral Ct counts¹

Item	Treatment			SEM	P-value2
	CON	POS	ISF
Litters3, n	7	5	4
PND 3	ND	8.12	10.87	7.26	0.78
PND 21	ND	34.66	21.59	26.91	0.79
PND 28	ND	11.00	29.40	7.79	0.12
PND 35	ND	15.13	32.09	6.19	0.07
PND 42	ND	16.03	34.12	9.41	0.13
PND 49	ND	16.53	33.55	9.55	0.15
PND 56	ND	17.62	34.50	10.21	0.18
PND 63	ND	16.87	33.34	6.93	0.11

Ct, ND (non-detectable) considered negative, 0 < Ct≤ 38 considered positive, Ct > 38 considered suspect.

Pairwise P-value between POS and ISF treatments.

Litter served as the experimental unit.

Abbreviations: CON, control treatment (unsupplemented diet without infection); ISF, isoflavone-supplemented treatment (with infection); POS, positive control (unsupplemented diet with infection); PND, postnatal day; SEM, standard error of the mean.

Growth performance and muscle outcomes

Up to DPI 21, feed intake was reduced (P < 0.05) by over 40% in POS and ISF gilts compared with CON gilts (Figure 2). During the post-infection period (i.e., after DPI 21), gilt daily feed intake was not different (P > 0.05) between treatment groups. At PND 3, body, muscle, and organ weights did not differ (P > 0.05; Table 4) between groups with the exception of POS pigs having a greater (P = 0.04) relative liver weight compared with CON pigs. At PND 21, body weight was increased approximately 36% in CON pigs compared to ISF pigs (P = 0.02; Table 5). As a function of this increase in body weight, longissimus dorsi, semitendinosus, and psoas major weights were increased by approximately 61%, 60%, and 47%, respectively, in CON pigs compared to ISF pigs (P < 0.05). CON pigs also had larger (P = 0.04) kidneys compared to POS and ISF pigs. ISF pigs had a larger (P < 0.01) relative liver weight compared to both CON and POS pigs. At both PND 3 (Figure 3) and 21 (Figure 4), there was no statistical difference (P > 0.05) in muscle fiber number or area between treatments.

Figure 2.

The effect of PRRSV infection status and dietary supplementation with soy isoflavones on gilt feed disappearance. abLeast squares means within a time point lacking a common superscript letter differ (P < 0.05). Gilt served as the experimental unit. Abbreviations: CON, uninfected and fed a diet devoid of soy isoflavones; POS, infected with PRRSV and fed the control diet; ISF, infected with PRRSV and fed the control diet supplemented with 1,500 ppm soy isoflavones. 57 x 36 mm (300 x 300 DPI).

Table 4.

The effect of maternal PRRSV infection and dietary supplementation with isoflavone on piglet muscle and organ weights at PND 3

Item	Treatment			SEM	P-value
	CON	POS	ISF
Pigs1, n	14	6	6
Body wt, kg	1.81	1.65	1.99	0.15	0.27
Longissimus dorsi, g	14.72	13.19	16.75	2.24	0.51
% body wt	0.80	0.80	0.82	0.08	0.98
Semitendinosus, g	5.22	4.20	5.34	0.59	0.30
% body wt	0.29	0.25	0.27	0.03	0.64
Psoas major, g	3.43	3.09	4.17	0.46	0.22
% body wt	0.19	0.19	0.21	0.02	0.54
Heart, g	15.63	14.28	16.96	1.32	0.40
% body wt	0.86	0.87	0.86	0.05	0.98
Lungs, g	36.73	38.14	42.09	5.46	0.69
% body wt	1.99	2.34	2.14	0.25	0.50
Liver, g	69.78	76.01	77.70	7.54	0.59
% body wt	3.81a	4.64b	3.91ab	0.25	0.04
Kidney, g	16.13	16.92	18.76	1.44	0.30
% body wt	0.90	1.03	0.94	0.05	0.13

Pig served as the experimental unit.

LS means within a row having differing superscripts are statistically different (P < 0.05).

Abbreviations: CON, control treatment (unsupplemented diet without infection); ISF, isoflavone-supplemented treatment (with infection); PND, postnatal day; POS, positive control (unsupplemented diet with infection); SEM, standard error of the mean.

Table 5.

The effect of maternal PRRSV infection and dietary supplementation with isoflavones on pig muscle and organ weights at PND 21

Item	Treatment1			SEM	P-value
	CON	POS	ISF
Pigs1, n	14	6	6
Body wt, kg	6.88a	5.69ab	5.06b	0.68	0.02
Longissimus dorsi, g	100.24a	81.70ab	62.14b	12.78	0.01
% body wt	1.46	1.34	1.18	0.14	0.10
Semitendinosus, g	26.13a	21.87ab	16.38b	3.93	0.04
% body wt	0.38	0.38	0.32	0.05	0.32
Psoas major, g	16.21a	12.04ab	11.01b	2.41	0.05
% body wt	0.24	0.22	0.21	0.03	0.66
Heart, g	46.74	44.09	35.54	4.91	0.06
% body wt	0.68	0.76	0.74	0.07	0.39
Lungs, g	104.08	95.98	108.40	15.07	0.81
% body wt	1.50	1.67	2.34	0.38	0.06
Liver, g	184.74	177.76	178.40	17.06	0.88
% body wt	2.69a	3.13a	3.67b	0.20	< 0.01
Kidney, g	45.10a	43.18ab	35.16b	4.12	0.04
% body wt	0.65	0.78	0.72	0.07	0.24

Pig served as the experimental unit.

LS means within a row having differing superscripts are statistically different (P < 0.05).

Abbreviations: CON, control treatment (unsupplemented diet without infection); ISF, isoflavone-supplemented treatment (with infection); PND, postnatal day; POS, positive control (unsupplemented diet with infection); SEM, standard error of the mean.

Figure 3.

The effect of maternal PRRSV infection and dietary supplementation on offspring semitendinosus muscle fiber size and number at PND 3. (A) The effect of maternal PRRSV infection and dietary supplementation on offspring muscle fiber size. (B) The effect of maternal PRRSV infection and dietary supplementation on estimated offspring muscle fiber number. (C) Example of CON sample with H&E staining. (D) Example of POS sample with H&E staining. (E) Example of ISF sample with H&E staining. Pig served as the experimental unit. Abbreviations: CON, uninfected and fed a diet devoid of soy isoflavones; POS, infected with PRRSV and fed the control diet; ISF, infected with PRRSV and fed the control diet supplemented with 1,500 ppm soy isoflavones. 65 x 43 mm (300 x 300 DPI).

Figure 4.

The effect of maternal PRRSV infection and dietary supplementation on offspring semitendinosus muscle fiber size and number at PND 21. (A) The effect of maternal PRRSV infection and dietary supplementation on offspring muscle fiber size. (B) The effect of maternal PRRSV infection and dietary supplementation on estimated offspring muscle fiber number. (C) Example of CON sample with H&E staining. (D) Example of POS sample with H&E staining. (E) Example of ISF sample with H&E staining. Pig served as the experimental unit. Abbreviations: CON, uninfected and fed a diet devoid of soy isoflavones; POS, infected with PRRSV and fed the control diet; ISF, infected with PRRSV and fed the control diet supplemented with 1,500 ppm soy isoflavones. 63 x 43 mm (300 x 300 DPI).

Innate immune challenge

Main and interactive effects (P ≤ 0.05) of time and treatment were evident following the postnatal innate immune challenge (Figure 5). Prior to LPS administration (0 h post-injection), ISF pigs had a greater (P = 0.05) TNF-α concentration compared to CON pigs, with POS pigs being intermediate. In ISF pigs, TNF- α concentration was nearly four times greater than CON and in POS pigs, it was more than doubled compared with CON. For all treatments, TNF-α was increased (P < 0.01) at 4 h post-injection compared with the pre-challenge time-point (0 h post-injection). At 8 h post-injection, TNF-α was reduced (P < 0.01) compared to 4 h post-injection but increased compared to 0 h post-injection for all treatments. At 4 h post-injection, ISF pigs again had a higher (P = 0.05) TNF-α concentration (1,880 pg/mL) compared to CON pigs (945 pg/mL), with POS pigs being intermediate (1,355 pg/mL). At 8 h post-injection, CON and POS pigs (388 and 511 pg/mL, respectively) were not different, but both were lower (P = 0.05) than ISF pigs (890 pg/mL). The area under the curve was largest for ISF pigs at 9,710 TNF-α (pg/mL × h post injection), whereas POS had an area of 6,698 TNF-α (pg/mL × h post injection) and CON an area of 4,660 TNF-α (pg/mL × h post-injection). From 0- to 4-h post-injection, TNF-α increased by 9-, 10.6-, and 18-fold for ISF, POS, and CON pigs, respectively. From 4- to 8-h post-injection, ISF, POS, and CON pigs had decreases of 2-, 2.7-, and 2.5-fold, respectively. When comparing 0- and 8-h post-injection, ISF and POS pigs were increased by 4-fold, whereas CON pigs were increased by 7.5-fold.

Figure 5.

The effect of maternal PRRSV infection and dietary supplementation on offspring innate immune response. a–fLeast squares means lacking a common superscript letter differ (P < 0.05). Data were analyzed as a split plot design, with litter serving as whole plot and piglet serving as the split plot. Abbreviations: CON, uninfected and fed a diet devoid of soy isoflavones; POS, infected with PRRSV and fed the control diet; ISF, infected with PRRSV and fed the control diet supplemented with 1,500 ppm soy isoflavones. 52 × 34 mm (300 × 300 DPI).

Adaptive immune challenge

There was no significant effect of treatment on any of the T cell populations measured (P ≥ 0.15). There was a significant effect of time on total T cell percentage (P = 0.03), undifferentiated T cell percentage (P < 0.01), helper T cell percentage (P < 0.01), dual positive T cell percentage (P = 0.01), and helper:cytotoxic T cells (P < 0.01). CON pigs had the highest (P = 0.03) percentage of total T cells at DPV 14 compared to DPV 7, with DPV 0 being intermediate. POS pigs also had the highest (P = 0.03) percentage of total T cells at DPV 14, but had the lowest percentage at DPV 0, with DPV 7 being intermediate. CON pigs had the highest (P < 0.01) percentage of undifferentiated T cells at DPV 14 compared to DPV 7, with DPV 0 being intermediate. Both POS and ISF pigs had the highest (P < 0.01) percentage of undifferentiated T cells at DPV 14 compared to both DPV 0 and 7, which did not differ. CON pigs had the lowest (P < 0.01) percentage of helper T cells at DPV 14 compared to DPV 7, which was significantly lower (P < 0.01) than DPV 0. POS and ISF pigs also had the lowest (P < 0.01) percentage of helper T cells at DPV 14 compared to DPV 0 and 7, which did not differ. CON pigs had the highest (P = 0.01) percentage of dual positive T cells at DPV 7 compared to DPV 0, with DPV 14 being intermediate. The percentage of dual positive T cells in POS pigs did not differ across days. ISF pigs had the lowest (P = 0.01) percentage of dual positive T cells at DPV 14 compared to DPV 0 and 7, which did not differ. CON pigs had a greater (P < 0.01) helper:cytotoxic T cells at DPV 0 compared to DPV 7, which was significantly greater (P < 0.01) compared to DPV 14. POS pigs had the lowest (P < 0.01) helper:cytotoxic T cells at DPV 14 compared to both DPV 0 and 7, which did not differ. Helper:cytotoxic T cells did not differ in ISF pigs across days.

There was a significant treatment × time effect on helper T cell percentage (P = 0.02) and dual positive T cells (P = 0.02). As discussed above, percentage of helper T cells differed across days within each treatment; however, CON pigs significantly decreased across all 3 d, whereas POS and ISF pigs only significantly decreased at DPV 14. At DPV 0, ISF pigs had a greater (P = 0.02) percentage of dual positive T cells compared to CON pigs, with POS pigs being intermediate. Further, CON pigs had the lowest (P = 0.02) percentage of T cells at DPV 0 compared to DPV 7, with DPV 14 being intermediate. Meanwhile, ISF pigs had the lowest (P = 0.02) percentage of dual positive T cells at DPV 14 compared to DPV 0 and 7, which did not differ. Dual positive T cells did not differ in POS pigs across days.

Discussion

Gilts successfully infected with PRRSV exhibited expected clinical signs of illness, and dietary isoflavones provided little protection against the live infection. Anorexia is a clinical sign of PRRSV (Done et al., 1996; Zimmerman et al., 1997; Schweer et al., 2017; Helm et al., 2020; Smith et al., 2020) and was observed in the present study in infected gilts. Although fever has been reported as an outcome of PRRSV infection in growing pigs (Greiner et al., 2001; Li et al., 2017; Smith et al., 2020), similar to the present study, others investigating sows infected with PRRSV report no consistent elevation in body temperature (Prieto et al., 1996; Kranker et al., 1998; Karniychuk et al., 2011), suggesting that fever may not be a universally expressed symptom. As isoflavone supplementation in healthy sows has been reported to have minimal effect on reproductive performance (Rehfeldt et al., 2007; Farmer et al., 2013; Hu et al., 2015; Farmer et al., 2016), isoflavones were only provided to PRRSV-infected sows in an effort to discern their reported antiviral properties (Andres et al., 2009; Rochell et al., 2015; Smith et al., 2018; Smith et al., 2020). Supplementation with isoflavones did not alter feed disappearance or body temperature. However, gilts supplemented with isoflavones had an increased viral Ct count at DPI 21, indicating decreased circulating viral load. The isoflavone concentration used in the current study (1,500 mg/kg) largely exceeds that which is observed in traditional swine diets (Payne et al., 2001; Kuhn et al., 2004; Smith and Dilger, 2018; Zhuo et al., 2020; Wu et al., 2021), implying that excess isoflavone supplementation may be useful in reducing viral load during active PRRSV infection. Increased litter mortality has been consistently reported as a consequence of PRRSV infection (Kranker et al., 1998; Zimmerman et al., 1997; Holtkamp et al., 2013) and was also observed in the present study. Mortality in both POS and ISF litters was similar to that which others have reported occurring during PRRSV infection (Kranker et al., 1998). Supplementation with dietary isoflavones did not alter the percentage of pigs born alive, number of pigs weaned per litter, or survivability during the suckling period. Overall, PRRSV negatively impacted gilt feed disappearance and litter outcomes and although dietary isoflavones may help in reducing viral load, supplementation did not rescue reproductive performance. It is important to note for both gilt and pig outcomes that relatively low replication was achieved in the present study, and therefore, caution must be used when interpreting outcomes.

Due to the anorexia observed in infected gilts and the detrimental impact of inflammation on secondary fetal muscle development (GD 54-90, Wigmore and Stickland, 1983), it was hypothesized that pigs exposed to maternal infection would have reduced body and muscle weights. However, at PND 3, there were no differences in muscle and organ size between groups with the exception of relative liver weight. POS pigs had a larger relative liver weight compared to CON pigs, with ISF pigs being intermediate. The liver plays an essential role in the immune response, helping mitigate pathogenic spread from the portal vein (Gao et al., 2008; Zhan and An, 2010). The liver also synthesizes the majority of proteins central to the innate immune response and acts as a central hub for a variety of innate immune cells such as macrophages and natural killer cells (Gao et al., 2008; Zhan and An, 2010). Reflective of this, the liver is heavily involved in inflammatory processes in both the healthy and ill states (Robinson et al., 2016). In humans, the fetal liver is seeded with immune cells beginning in the first trimester in order to disperse resident immune cells and develop innate effector cells (Park et al., 2020). In pigs, hematopoietic activity can be detected in the liver as early as GD 20. Although the fetal immune system does not become competent until after parturition, cells vital to the innate immune response can be detected in umbilical cord blood and the fetal spleen as early as GD 45 (Šinkora and Butler, 2009). These cells have the ability to secrete associated effector molecules such as inflammatory cytokines. Further, the porcine fetal liver becomes increasingly laden with lymphocytes beginning at approximately GD 60 and peaking at GD 100 (Šinkora and Butler, 2009). Further, PRRSV infection has been reported to increase fetal production of IFN-γ and TNF-α at GD 85 (Pasternak et al., 2020), indicating an ability to mount (albeit weak) immune response to PRRSV in utero. This, combined with the fact that PRRSV can cross the placental barrier (Rowland, 2010; Karniychuk et al., 2011; Karniychuk and Nauwynck, 2013; Harding et al., 2017), implies that the increased relative liver weight could be attributed to an increased immune response in POS pigs in utero.

Interestingly, ISF pig’s relative liver weight did not differ from CON or POS pigs. Coupled with the reduction in viral load of ISF gilts, this intermediary response is suggestive of a possible protective effect of isoflavones against PRRSV. However, this protective effect was not apparent during the postnatal period. Although isoflavones are known to cross the placental barrier (Degen et al., 2002; Balakrishnan et al., 2010), maternal supplementation did not aid in reducing viral load perinatally, as evidenced by the low Ct count observed at PND 3. At PND 21, POS pigs had a numerically increased Ct count compared to ISF pigs. Although not significant, this increase is clinically relevant and suggests that POS pigs were more effective in reducing viral load during the suckling period. From PND 28-63, ISF pigs had numerically increased Ct counts compared to POS pigs. Again, although not significant, these increases are clinically relevant and suggest that ISF were better able to reduce viral load post-weaning. Further, this decrease in Ct count from PND 21-28 implies that POS pigs were more poorly equipped to handle the stress associated with weaning, resulting in an increase in viral load that POS pigs were unable to recover from during the trial period. However, despite having an increased Ct count for the remainder of the trial period, ISF pigs continued to test positive for PRRSV throughout the entire trial period, indicating an inability to clear infection despite the reduction in viral load.

Contradictory to PND 3, at PND 21 ISF pigs had reduced body weight and increased relative liver size compared to POS pigs. Further, although relative weights did not differ, absolute weights of the longissimus dorsi, semitendinosus, and psoas major were reduced in ISF pigs compared to CON, with POS being intermediate, whereas there were no differences in absolute weights of the heart and lungs. This is indicative of a shift in energy partitioning in ISF pigs away from muscle growth towards maintaining essential organ growth. Overall, the lack of difference in body, muscle, and organ weights between treatments at PND 3 indicates that maternal PRRSV infection did not stunt offspring development as hypothesized. However, transference of maternal infection postnatally to offspring does impair growth, as evidenced by reduced body and muscle weights in POS and ISF pigs at PND 21. Although maternal isoflavone supplementation may have a protective effect against PRRSV in utero, this defensive ability disappears during postnatal infection.

As inflammation has been reported to be antagonistic to muscle development (Du et al., 2010), it was hypothesized that maternal infection would be damaging to both offspring muscle fiber number and size. However, there were no statistical differences between treatments in either outcome at both PND 3 or 21. This contradicts other authors who report that sustained inflammation during gestation results in reduced offspring muscle fiber size and differentiation (Tong et al., 2009; Cadaret et al., 2019). This discrepancy may result from differences in model (rodent v. pig) or method of inducing inflammation. Tong et al. (2009) created low-level chronic inflammation using maternal obesity, whereas Cadaret et al. (2019) utilized repeated injections of bacterial endotoxins. Both of these models use different mechanisms to trigger inflammatory cytokine production compared to PRRSV infection. For example, maternal obesity has been associated with elevated levels of IL-6, c-reactive protein, IL-2, IL-8, and TNF-α, but the resultant inflammation is generally low grade compared to inflammation induced by pathogens (Ramsay et al., 2002; Madan et al., 2009; Schmatz et al., 2009; Pantham et al., 2015). Further, a review of maternal obesity and inflammation found that TNF-α and IL-6 are not consistently elevated in obese subjects (Pendeloski et al., 2017). This differs greatly from PRRSV infection, which induces high levels of TNF-α and IL-6 (Liu et al., 2010; Zhang et al., 2013; Li et al., 2017; An et al., 2020; Lu et al., 2020). Similarly, Cadaret et al. (2019) utilized repeated injections of LPS, which also differs in its induced inflammatory response compared to PRRSV infection. LPS signals through TLR4, resulting in the rapid, severe production of TNF-α and IL-6 (Miller et al., 2005; Lu et al., 2008; Płóciennikowska et al., 2015). Further, there are several TLR4 regulatory pathways that can shut off inflammatory cytokine production, resulting in relatively rapid recovery from peak inflammation when LPS is introduced independently of a bacterial infection (Lu et al., 2008). Conversely, PRRSV is recognized by TLR3 (Luo et al., 2008; Sang et al., 2008) resulting in delayed, less robust inflammation (Krasowska-Zoladek et al., 2006). These differences in inflammatory induction among studies may contribute to the discrepancy in effect on fetal muscle development. Alternatively, inflammation induced by PRRSV may not have been potent enough to cause the expected detrimental effects. Infection with PRRSV is known to be immunosuppressive (Song et al., 2013; Lunney et al., 2015; Li et al., 2018), largely through the induction of regulatory cytokine IL-10 (Song et al., 2013; Lunney et al., 2015). This feature of PRRSV may have mitigated the gilts immune response, preventing them from producing enough inflammation to induce changes in fetal muscle development. Numerically, POS pigs had approximately 60,467 fewer fibers than ISF pigs, and 86,915 fewer fibers than CON pigs. This reduction in muscle hyperplasia, though not statistically significant, may still pose a negative impact on the growth potential of the pigs.

Given their immunomodulatory capacity in a PRRSV challenge (Rochell et al., 2015; Smith et al., 2019; Smith et al., 2020), it was expected that isoflavone supplementation would curb the severity of infection in the dam, meaning a reduced maternal inflammatory response. Although isoflavone supplementation may have expedited viral clearance in gilts, it did not result in protection of the offspring from inflammatory stimuli. Prior to LPS injection (0 h post-injection) ISF pigs had increased TNF-α concentration, indicating a baseline inflammatory state. Once challenged with LPS, ISF pigs continue to have the greatest TNF-α concentration at 4- and 8-h post-injection compared to POS and CON pigs who did not differ. The area under each curve [9,710, 6,698, and 4,660 TNF-α (pg/mL) × hours post-injection for ISF, POS, and CON pigs, respectively] indicates that maternal immune activation results in an exaggerated inflammatory state, although some of this inflammation can likely be attributed to postnatal infection.

It remains plausible that the heightened inflammatory response of pigs born to ISF-supplemented dams can be partially explained by the postnatal PRRSV infection experienced by ISF pigs. It is expected that pigs responding to a PRRSV infection will have a greater concentration of TNF-α compared with uninfected contemporaries (Xiao et al., 2010). However, both ISF and POS pigs experienced a postnatal PRRSV infection throughout the study period. So, effects from this underlying postnatal challenge would be observed in both ISF and POS groups. Therefore, it is surprising that ISF pigs consistently had increased TNF-α concentration compared to POS pigs, and suggests that maternal supplementation with isoflavones antagonizes offspring inflammation. However, even though they had the greatest TNF-α concentration at all time-points, ISF pigs exhibited a similar response to LPS stimulation as POS pigs, and both ISF and POS pig responses were less drastic than that of CON pigs. From 0- to 4-h post-injection, ISF and POS pigs had 9- and 10.6-fold increases in TNF-α concentration, whereas CON pigs had an 18-fold increase. This implies that there may be differences in offspring programming due to infection status, meaning that maternal immune activation may dampen the offspring’s postnatal innate immune response. This supposition contradicts previous findings (Zager et al., 2013; Onore et al., 2014; Rose et al., 2017), but many agree that offspring exposed to maternal immune activation are consistently more inflammatory with or without a postnatal immune challenge (Zager et al., 2013; Onore et al., 2014, Rose et al., 2017; Carlezon et al., 2019; Garcia-Valtanen et al., 2020).

Interestingly, all treatments were similar in their recovery from peak inflammation, with ISF, POS, and CON pigs having 2-, 2.7-, and 2.5-fold decreases from 4- to 8-h post-injection. This suggests that although maternal immune activation resulted in a stunted inflammatory response, it did not subsequently increase rate of recovery. Nevertheless, it is important to view differences in rate of increase/recovery in context of the absolute concentration of TNF-α of each treatment. Although ISF pigs had a lower rate of increase and similar rate of recovery compared with CON pigs, ISF pigs still had the greatest TNF-α concentration at each time-point. This suggests that maternal immune activation and supplementation with isoflavones predisposes offspring to an inflammatory state. Meanwhile, the similarity between CON and POS TNF-α concentrations insinuates that maternal immune activation alone does not result in this differential programming of baseline inflammation. This conclusion is in agreement with other maternal immune activation studies. Antonson et al. (2017) also infected pregnant gilts with PRRSV at the beginning of the third trimester (GD 76) and challenged the offspring with LPS on PND 14. They report maternal immune activation status had no effect on offspring circulating TNF-α concentration at 0 or 4 h post-injection, supporting the conclusion that maternal immune activation does not inherently result in a pro-inflammatory phenotype. Overall, the similarity in rates of change, but differences in absolute TNF-α concentration, between POS and ISF pigs implies two types of differential programming may be at play. Based on the reduced rate of increase in POS/ISF pigs compared to CON pigs, maternal immune activation dampens the offspring’s innate immune response. However, based on absolute TNF-α concentration, maternal immune activation does not impact offspring baseline inflammatory state unless coupled with isoflavone supplementation.

There is little consensus in the literature regarding offspring immune response differences attributed to maternal environment. Some studies report that maternal immune activation results in impaired offspring inflammatory responses (Beloosesky et al., 2010), whereas others report that offspring of immune activated mothers have a greater proportion of activated macrophages and inflammatory markers (Onore et al., 2014). However, Beloosesky et al. (2010) used LPS, a bacterial mimetic, to challenge the dams, whereas Onore et al. (2014) used polyI:C, a viral mimetic. The insult resulting in maternal immune activation may be critical in determining which arm of the offspring’s immune response undergoes the most programming. It has been reported that maternal immune activation lowers offspring IFN-γ responsiveness to postnatal stressors (Rymut et al., 2021). In this case, both the trigger for maternal immune response and postnatal stress were viral in nature [PRRSV inoculation and poly(I:C) injection, respectively]. It could be concluded that maternal immune activation results in a protective phenotype in the offspring specific to the type of insult experienced by the dam. This may be attributed to the variance in the innate immune response to bacterial and viral insults (Carrillo et al., 2017). Toll-like receptor (TLR)-3 is implicated in response to viral insults, whereas TLR-4 is implicated in response to bacterial insults (Allhorn et al., 2008; El-Zayat et al., 2019). TLR-4 is located on the cell membrane, and once bound to a ligand activates a MyD88-dependent pathway resulting in the synthesis of pro-inflammatory cytokines by NF-κB. In contrast, TLR-3 is located on the endosome and activates a Toll/interleukin 1 receptor domain containing adaptor inducing interferon β (TRIF)-dependent pathway resulting in type 1 interferon (IFN) production. Although TLR-4 can activate the TRIF-dependent pathway, TLR-3 is unable to stimulate the MyD88-dependent pathway (El-Zayat et al., 2019). Therefore, the discrepancy between the current study and that done by Beloosesky et al. (2010) may be due to the different insults used to stimulate a maternal immune response. Both Onore et al. (2014) and the current study use a viral insult, a challenge that results in NF-κB activation, and agree that offspring from immune challenged dams are more inflammatory than their control counterparts. Hence, changes in offspring innate immune responses due to maternal inflammation may be specific to the type of pathogen that the dam encounters.

Although not statistically significant, at DPV 0 ISF pigs had a shift in T cell populations when compared to CON and POS pigs (Table 6), suggesting differential programming in the adaptive immune system. ISF pigs produced on average 1,057 fewer lymphocytes compared to CON and POS pigs, with an approximate 11% reduction in the proportion of T cells. As peripheral blood mononuclear cells (PBMC) can include lymphocytes, monocytes, natural killer cells, and dendritic cells (Kleiveland, 2015), this means that ISF pigs had a greater proportion of these other immune cells than CON pigs. This may be attributed to the postnatal infection experienced by these animals. Natural killer cells can be used in viral defense by recognizing infected cells and inducing apoptosis while also producing the antiviral cytokine IFN-γ (Vivier et al., 2008). Dendritic cells also aid in viral defense by presenting antigens to T and B cells (Bell et al., 1999). However, both natural killer and dendritic cells belong to the innate arm of the immune system and are intended to activate the adaptive arm. This activation could explain why ISF pigs also had an approximate 12% reduction in undifferentiated T cell count compared to CON and POS pigs, implying that of the T cells produced, a greater proportion of them would be differentiated. Indeed, ISF pigs had a significantly greater percentage of dual positive T cells when compared to CON and POS pigs, and a numerical increase of 7% in cytotoxic T cells. These data suggest that ISF pigs have a more mature adaptive response at DPV 0 than both POS and CON pigs, although this did not result in clearance of the postnatal infection.

Table 6.

The effect of maternal PRRSV infection and dietary supplementation with soy isoflavones on offspring proportional lymphocyte populations

	DPV 0			DPV 7					DPV 14
Item	CON	POS	ISF	SEM	CON	POS	ISF	SEM	CON	POS	ISF	SEM	Treatment P-value	Time P-value	Treatment × Time P-value
Pigs1, n	16	15	12		16	15	12		16	15	12
Lymphocytes2, n	5654	5021	4281	719	4881	5055	5004	738	6022	5065	4605	720	0.61	0.56	0.12
Lymphocytes3, %	56.54	50.21	42.81	7.20	48.81	50.55	50.04	7.38	60.22	50.65	46.05	7.20	0.61	0.56	0.12
Total T cells4, %	64.82xy	65.77x	53.75	5.10	61.77x	66.55xy	55.21	5.20	66.9y	71.76y	56.07	5.10	0.18	0.03	0.54
Undifferentiated T cells5, %	53.47xy	52.05x	41.27x	4.51	49.06x	51.95x	41.12x	4.63	56.73y	61.06y	47.61y	4.51	0.15	< 0.01	0.56
Helper T cells6, %	21.12x	18.29x	18.40x	1.34	17.96y	17.68x	18.43x	1.42	12.57z	12.15y	15.52y	1.34	0.60	< 0.01	0.02
Cytotoxic T cells7, %	21.24	22.78	29.28	3.42	25.95	21.85	29.91	3.53	25.03	19.77	28.91	3.42	0.22	0.51	0.25
Dual Positive T cells8, %	4.18ax	6.79ab	11.05bx	1.80	7.03y	8.27	10.12x	1.84	5.72xy	6.88	7.96y	1.80	0.25	0.02	0.02
Helper:Cytotoxic T cells	1.12x	0.92x	0.75	0.16	0.80y	0.86x	0.66	0.16	0.60z	0.66y	0.57	0.16	0.64	< 0.01	0.06

Least squares means within a day across treatments lacking a common superscript letter differ (P < 0.05).

Least squares means within a treatment across days lacking a common superscript letter differ (P < 0.05).

Data were analyzed using a split-plot design, with litter serving as whole plot and offspring pig serving as split plot.

Absolute count of lymphocytes detected.

Lymphocytes, % = (absolute count of lymphocytes/10,000 [total number of cells counted per sample]) × 100.

Total T cells, % = (Number of CD3+ cells/Number of lymphocytes) × 100.

Undifferentiated T cells, % = 100 − (Helper T cells, % + Cytotoxic T cells, % + Dual Positive T cells, %).

Helper T cells defined as CD3+ CD4+ cells.

Cytotoxic T cells defined as CD3+ CD8+ cells.

Dual Positive T cells defined as CD3+CD4+CD8+ cells.

Abbreviations: CON, control treatment (unsupplemented diet without infection); DPV, days post-vaccination; ISF, isoflavone-supplemented treatment (with infection); POS, positive control (unsupplemented diet with infection); SEM, standard error of the mean.

Given the prolonged nature of the postnatal infection experienced by ISF and POS pigs at this point (59 d; litters tested positive at birth; Table 3), it is expected that the adaptive immune response would have been mounted, meaning a greater proportion of lymphocytes found in PBMC isolates. This was not the case, meaning there may be a failure in ISF and POS pigs in mounting an adequate anti-PRRSV adaptive response. Conversely, ISF pigs had a greater proportion of memory T cells compared to both CON and POS pigs, implying a more developed T cell response. Interestingly, this did not result in viral clearance, as ISF pigs continued to test positive for PRRSV until study termination. PRRSV is known to be immunosuppressive and delays the mounting of a specific neutralizing response (Murtaugh et al., 2002; Song et al., 2013), which may explain the lack of viral clearance despite the mature T cell population. Further, differences in T cell populations persist at DPV 7 with the exception of dual positive cells. From DPV 0 to 7, CON pigs had an increase in the percentage of dual positive T cells, whereas POS and ISF pigs did not differ across days. This implies a limited ability to mount an adaptive response to the vaccine given. Finally, differences in T cell populations at DPV 14 resemble those observed at DPV 0, with ISF pigs having numerically the greatest proportion of cytotoxic T cells and lowest proportion of undifferentiated and total T cells. However, dual positive cell populations do not differ among treatments at DPV 14, indicating no change in CON/POS pigs and a reduction in ISF pigs. Cytotoxic T cells are responsible for killing infected host cells and are therefore considered part of the active adaptive response (Chisari, 1997). To develop immunological tolerance or to halt a response once pathogen clearance is achieved, regulatory T cells have the ability to terminate cytotoxic T cell activity (Mempel et al., 2006). The continued elevation of cytotoxic T cell proportions in ISF pigs compared to POS pigs suggests that either 1) ISF pigs were unable to mount as effective a response to the vaccine as POS pigs or 2) that ISF pigs continued to be under a more severe postnatal challenge than POS pigs. Either of these outcomes indicates a difference between ISF and POS pigs in programmed immunological response.

It is difficult in the current study to determine which changes observed were due to maternal programming versus the postnatal PRRSV challenge experienced by both ISF and POS groups. To elucidate the effects of an inflammatory maternal environment on offspring immune programming, further research should be conducted using a non-infectious inflammation inducer. This would eliminate the confounding effects of postnatal infection on offspring immunological outcomes. Overall, the data presented in the current study conclude that there is evidence for differential immune programming in offspring born to infected mothers, especially when the dam is supplemented with isoflavones during infection.

Glossary

Abbreviations

CD

cluster of differentiation

Ct

cycle threshold

DPI

days post-inoculation

DPV

days post-vaccination

ELISA

enzyme-linked immunosorbent assay

GD

gestational day

IFN

interferon

IL

interleukin

LPS

lipopolysaccharide

MyD88

myeloid differentiation primary response 88

NF-κ

nuclear factor-κB

PBMC

peripheral blood mononuclear cells

PDIFF

probability of differences

PND

postnatal day

Poly(I:C)

polyinosinic:polycytidylic acid

PRRSV

porcine reproductive and respiratory syndrome virus

TCID

tissue culture infective dose

TLR

toll-like receptor

TNF-α

tumor-necrosis factor alpha

Conflicts of Interest Statement

The authors declare no real or perceived conflicts of interest.