Content area
Abstract
Porcine reproductive and respiratory syndrome (PRRS) modified live virus (MLV) vaccines likely confer partial protection against heterologous wild type challenge through broadly reactive T cells. Therefore, the efficacy of Ingelvac PRRS® MLV alone and reconstituted with Ingelvac CircoFLEX® was assessed by comparing the ability to induce a robust cell-mediated immune response and confer protection against challenge. This study utilized both classic immune assays and machine learning software for measuring the cell-mediated immune response to PRRSV vaccination and challenge. Both vaccine groups had significantly reduced viremia, lung viral load, gross and microscopic lung lesions and improved average daily gain compared to the mock vaccinated group. The T cell analysis in the lung revealed a robust response to PRRSV infection leading to marked clearance of virus from the lung in the absence of neutralizing antibodies.
Full text
1 Introduction
Porcine reproductive and respiratory syndrome virus (PRRSV) poses a significant threat to swine health and well-being both in the United States and globally. In individual pigs, PRRSV infects CD163+ macrophages and causes a systemic infection that can result in reproductive failure and/or interstitial pneumonia [ 1–4]. It has also been reported to cause vasculitis, myocarditis, meningoencephalitis, conjunctivitis, and lymphoid necrosis [ 5–8]. On a herd level, a PRRSV outbreak can be devastating and result in abortions in gilts and sows, a marked increase in mortality, anorexia, and weight loss. The resulting economic losses inflicted by this virus are staggering and have recently been estimated at $1.2 billion/year in the USA [ 9]. Furthermore, antibiotic usage to combat secondary bacterial infection in herds experiencing a PRRSV outbreak are at least three times that of a PRRSV naïve herd [ 10]. Therefore, effective PRRSV control measures, such as vaccination, are important for reducing antibiotic usage in swine and limiting the development of antibiotic-resistant bacteria, such as Streptococcus suis, a zoonotic pathogen of swine [ 11].
PRRSV is a highly mutable enveloped single-stranded RNA virus classified in the order Nidovirales and family Arteriviridae. There are two recognized species of the virus: Betaarterivirus europensis (virus name PRRSV-1 or European PRRSV) and Betaarterivirus americense (PRRSV-2 or North American PRRSV) [ 12, 13]. PRRSV-2 is the dominant species circulating in the US and Asia [ 14, 15]. The mechanisms by which PRRSV vaccination induces protective immunity may provide valuable insights for studying other nidoviruses, such as coronaviruses.
PRRSV has been studied extensively by researchers in many different disciplines to better understand its basic virology as well as the resulting immune response and pathologic changes. Despite these concerted efforts and decades of vaccine development, there is no single intervention that provides broadly protective immunity against genetically diverse strains of the virus. However, modified live virus vaccination can result in partial protection from infection with wild type virus by limiting shedding and clinical disease [ 16–18]. While neutralizing antibodies eventually appear after infection or vaccination, they are often frustratingly delayed and at low titers [ 19–21]. The duration of this delay and magnitude of neutralizing antibody titer response may be PRRSV isolate dependent [ 22]. Much of the heterologous partial immune protection is thought to be derived from cross-reactive T cells [ 23–25]. Several studies have shown that a rise in PRRSV-specific interferon gamma (IFN-γ) producing T cells in the blood correlates with a reduction in viremia or transplacental infection [ 26–28]. Additional studies have displayed the importance of a strong cytotoxic T cell response, typified by granzyme B expression, in response to PRRSV infection [ 29, 30]. Therefore, a vaccine capable of inducing a robust and broadly cross-reactive T cell response against wild type PRRSV infection is important for protection from disease.
Effective combinatorial vaccines are beneficial for animal health and wellbeing due to their ability to reduce the number of injections per pig while targeting multiple infectious agents, including synergistic pathogens, at the same time. PRRSV and porcine circovirus-2 (PCV2) have long been known to synergistically promote viral replication and exacerbate clinical disease and lesions [ 31, 32]. The objective of this study was to evaluate if combining a subunit PCV2 vaccine (Ingelvac CircoFLEX®) with PRRSV MLV (Ingelvac® PRRS MLV) enhanced the T cell response elicited by PRRSV MLV. The T cell response was evaluated both in the blood and in the primary tissue of interest, the lung, after wild type challenge. Using machine learning software, PRRSV RNA and T cell response in the lung were quantified to determine the importance of T cells for PRRSV clearance and the resolution of disease.
2 Materials and methods
2.1 Ethics statement
The study was approved by Veterinary Resources Inc. (VRI) (Ames, Iowa) Animal Care and Use Committee and carried out under their supervision at their research facilities (Maxwell, Iowa). Conventional PRRSV negative mixed breed pigs ( n = 134) were weaned from dams at 21 ± 5 days of age and transported to VRI. Upon arrival at VRI, pigs were individually identified, randomly assigned to one of four treatment groups, treated with a metaphylactic dose of Baytril® (Elanco US, Greenfield, IN) following manufacturer recommendations to prevent systemic bacterial disease following the stress of weaning and transport, and acclimated for three days prior to onset of the study. Pigs were initially housed in separate rooms based on vaccine treatment until PRRSV challenge when all inoculated pigs were co-mingled. Strict negative controls were housed in a separate barn for the duration of the study.
2.2 Animal study and tissue collection
Group 1, comprised of 8 pigs, was designated as strict negative controls that were not vaccinated or challenged with PRRSV. Groups 2, 3, and 4, each assigned 42 pigs, were intramuscularly injected with one of the following: 2 mL of Ingelvac® PRRS MLV (Boehringer Ingelheim Vetmedica, Inc., Duluth, GA, Lot #7774826 – T01); 2 mL FLEX CircoPRRS, a combination of Ingelvac® PRRSV MLV and PCV2 subunit vaccines (Boehringer Ingelheim Vetmedica, Inc., Duluth, GA, Lot #7774826 – T02); or 2 mL 1× Wash Phosphate Buffered Saline (wPBS). Groups 1, 2, 3, and 4 were subsequently designated as Control, MLV, FLEX, and Mock, respectively (
Fig. 1
Pigs were observed for clinical signs of disease each day. Serum and rectal temperatures were collected and recorded for each animal at DPV 0, 7, 14, 21, 28, 29, 31, 33, 35, 42, and 49. Body weights were collected on DPV 0, 28, 42, and 49.
Samples collected at necropsy included blood in EDTA tubes, serum, bronchoalveolar lavage (BAL) fluid, left and right cranioventral lung tissue, and tracheobronchial lymph node. Samples were transferred to the Iowa State University Veterinary Diagnostic Laboratory (ISU VDL) for processing, testing and/or storage. Sections of left and right cranioventral lung and tracheobronchial lymph node were fixed in a 10 % neutral-buffered formalin solution (3.7 % formaldehyde) at room temperature (RT) and delivered to the ISU VDL for hematoxylin and eosin (H&E) staining, immunohistochemistry (IHC), and in situ hybridization (ISH) evaluation.
Two pigs were removed from all aspects of study analysis. One pig from the Mock group had irreconcilable inconsistencies in direct detection (IHC and ISH) reactivity in the lung. The other pig was a negative control pig that had a false positive PRRSV PCR result on lung tissue. This pig never seroconverted to PRRSV, was negative on PRRSV ISH, and was uniformly negative for PRRSV viremia via PCR.
2.3 Gross and histopathologic scoring of lung tissues
Gross pathological scoring of lungs was performed by VRI personnel at the time of necropsy utilizing a previously described PRRSV gross pathological lung scoring method [ 3].
Formalin-fixed paraffin-embedded (FFPE) tissue sections of left and right cranioventral lung tissue from each animal were stained with H&E and assessed for the degree of interstitial pneumonia by a board-certified veterinary anatomic pathologist blinded to treatment status utilizing the method described in Halbur et al.: 0 = No histologic findings of significance, 1 = Multifocal and mild lymphohistiocytic infiltration of alveolar septa, 2 = Multifocal and moderate lymphohistiocytic infiltration of alveolar septa, 3 = Diffuse and moderate lymphohistiocytic infiltration of alveolar septa, 4 = Diffuse and severe lymphohistiocytic infiltration of alveolar septa [ 3]. Additionally, necrotic alveolar macrophages were scored based on the following criteria: 0 = No necrotic alveolar macrophages, 1 ≤10 % of alveolar spaces contain necrotic alveolar macrophages, 2 = 10–30 % of alveolar spaces contain necrotic alveolar macrophages, 3 ≥30 % of alveolar spaces contain necrotic alveolar macrophages. Both lung sections were assessed for each animal resulting in one interstitial pneumonia and necrotic macrophage score per pig.
2.4 PRRSV Quantitative PCR
Reverse transcriptase quantitative PRRSV PCR (RT-qPCR) was performed on serum and pooled sections of left and right cranioventral lung from each pig at the ISU VDL. Samples were processed and tested by a commercial PRRSV RT-qPCR kit according to laboratory SOP as previously described [ 34].
2.5 PRRSV seroconversion and neutralizing antibody titers
PRRSV IgG was evaluated in serum using a PRRS X3 ELISA antibody kit (IDEXX Laboratories) at the ISU VDL as previously described [ 35]. Neutralizing antibody titers were evaluated in serum as previously described with the following modification: the challenge virus (PRRSV-2 1–7-4) was used as the test virus [ 35].
2.6 PBMC isolation and PRRSV IFN-γ ELISpot assay
Peripheral blood mononuclear cells (PBMCs) were isolated on DPV 14, 28, 42, and 49 from blood collected in EDTA tubes. Briefly, blood was diluted with an equal volume of PBS mix (1× PBS + 2 % fetal bovine serum (FBS) + 2 mM EDTA) and slowly layered over 15 mL of lymphocyte separation media (LSM) in a 50 mL SepMate™ conical tube. Layered blood tubes were then centrifuged at 1200 x g for ten minutes at room temperature with the brake on. Following centrifugation, the top layer of fluid was poured off into a new 50 mL conical tube and washed with PBS mix followed by a 10-min centrifuge at 300 x g. The cellular pellet was resuspended in 5 mL of ACK lysis buffer and incubated for 3 min with periodic rotation. The ACK was then diluted with 30 mL of PBS mix and followed by another 10-min centrifuge at 300 x g. The resulting cellular pellet was resuspended in 10 mL of complete RPMI w/ 5 % FBS (RPMI) for cell counting prior to freezing in liquid nitrogen.
Porcine IFN-γ ELISpot plates (R&D Systems Inc.®) were pre-incubated with 200 μl of RPMI for 20 min at room temperature. Cells were thawed from liquid nitrogen, washed, and resuspended in RPMI. Cells were enumerated and added into wells at 250,000 live cells per well in 100 μl of RPMI. Each animal had negative (cells without stimulation) and positive control (phytohemagglutinin at 5 μg/μl) wells. Three PRRSV-2 isolates were used to stimulate cells at a multiplicity of infection (MOI) of 0.1 with technical duplicates for each isolate. Following the addition of stimulants, cells were incubated for 24 h at 37 °C with 5 % CO 2. Following incubation, cells were discarded and the kit staining protocol was followed through spot formation and drying of the membrane. Spots were read on a CTL ImmunoSpot® ELISpot reader.
2.7 Dual chromogenic IHC/ISH labeling
Dual chromogenic IHC/ISH labeling was performed to simultaneously detect membrane bound CD3 protein and either interferon gamma (IFN-γ) mRNA or granzyme-B (GZMB) mRNA. RNAscope™ 2.5 VS Probe-Ss-IFNG (ACD catalog #490829) was utilized for IFN-γ in-situ Hybridization (ISH), and BaseScope™ VS probe BA-Ss-GZMB-3zz-st-C1 (ACD catalog #1171519-CT) was utilized for GZMB ISH. CD3 Immunohistochemistry (IHC) utilized CONFIRM anti-CD3 (2GV6) Rabbit Monoclonal (Roche Diagnostics, Indianapolis, IN USA; Catalog number 05278422001) for the primary antibody.
Protocols provided by the manufacturer (Protocols 3061 and 3071) were followed to prepare yellow chromogenic IHC and red chromogenic ISH preparations on FFPE tissues cut into 4 μm sections and placed on Superfrost® Plus slides (VWR™, Radnor, PA). The appropriate ISH assay was performed first, followed by immunohistochemistry (IHC) for CD3 with the Ventana DISCOVERY™ ULTRA System (Roche Diagnostics, Indianapolis, IN). All slides were checked for quality and consistency of staining across preparations by a board-certified veterinary anatomic pathologist.
2.8 Chromogenic PRRSV ISH staining and quantification
To detect and localize PRRSV RNA, chromogenic RNAscope™ 2.5 Assay (ACD Inc.) was performed on prepared 4 μm FFPE tissue sections with target probe PRRSV-GB-WT-ORF1a-C1 (ACD catalog #1251221). Manufacturer guidelines (document numbers 322,452 and 322,360-USM) and established procedures were followed for manual preparation with red chromogenic ISH labeling [ 36].
Quantification of PRRSV ISH labeling within the lung was performed using the HALO® image analysis platform (Indica Labs®, v3.4.2986). All slide preparations were digitized by an Aperio® GT 450 slide scanner (Leica Biosystems®, Deer Park, IL). Regions of interest were manually annotated in two sections of lung tissue. The luminal space and non-alveolar tissue immediately surrounding bronchi, the lumina and walls of adjacent arteries and veins, and any artifacts of tissue or slide processing were excluded from analysis during the annotation process. PRRSV ISH labeling was quantified as a percentage of labeled surface area over the total surface area of the annotated regions with the Area Quantification algorithm (v2.3.1). For data handling and visualization purposes, tissue area coverage by PRRSV ISH staining was multiplied by 1000.
2.9 T cell IHC and ISH Segmentation and Quantification
Whole slide images (SVS format) were imported into QuPath where two rectangular regions of interest (ROIs) were drawn per image, one per lung tissue. ROIs were drawn to encompass lung parenchyma and exclude bronchi. The portion of the images selected by the ROI were then exported to TIFF format using ImageJ.
NIS-Elements software (Nikon Instruments) was then utilized to segment the three structures (cell nuclei, CD3-positive reactivity, and ISH-positive reactivity (GZMB or IFN-γ)). All three structures were stained with different chromogenic stains. Thus, the NIS.ai artificial intelligence toolbox was employed to generate three AI classifiers that were trained to recognize the structures in the RGB images. Briefly, 400 × 400 pixel training images were cropped from a subset of the ROI TIF images. Training images came from a mix of both CD3 +GZMB and CD3 +IFN-γ stained images from 8 pigs representing all treatment groups and across both necropsy time points. Nuclei and CD3 classifiers were trained on a total of 10 images over two iterative rounds. ISH classifiers were ultimately trained on a total of 14 images. These training images represented a diversity of cellular distributions and staining intensities. The training images were then manually annotated in NIS-Elements to mark ground truth features for AI training. The AI training functions SegmentObjects.ai and Segment.ai were then trained on this training data for 1500 iterations in each round, achieving a training loss of less than 0.005 for each classifier.
For some of the classifiers, multiple rounds of re-training were performed to improve the segmentation accuracy. Briefly, a new set of training images was re-segmented with each AI classifier and then manually corrected as done in the initial annotation. The resultant training data was then used to train the AI classifiers in another round. Each subsequent round used different training images to increase the diversity and robustness of the classifier performance.
The final AI classifiers were then applied to the full set of ROI images, followed by batch image processing (fill holes, size exclusion for nuclei and CD3-positive staining) and feature extraction (counting of nuclei, CD3-positive cells, ISH foci). Quantification of CD3-positive, ISH-containing cells was performed by intersecting the respective segmentation masks for the different markers of interest.
2.10 Statistical analysis
Ordinary one-way Analysis of Variance (ANOVA) were performed to compare means among different groups within data sets using GraphPad Prism 10 (version 10.2.3). P < 0.05 were considered significant (∗ <0.05, ∗∗ <0.01, ∗∗∗ <0.001). Unpaired t-tests were conducted to compare groups within data sets found to have a significant difference by ANOVA. P < 0.05 were considered significant (∗ <0.05, ∗∗ <0.01, ∗∗∗ <0.001). Pearson or Spearman correlations were performed when appropriate between selected pairs of data collected from the same individuals using XY charts with GraphPad Prism 10 (version 10.2.3).
3 Results
3.1 PRRSV MLV vaccines limited viremia and improved average daily gain
Evaluation of the effect of the addition of a PCV2 subunit vaccine to PRRSV MLV began with the assessment of clinical parameters. The average daily gain (ADG), temperature, and viremia of pigs were measured and compared between treatment groups throughout the study. Pigs that received PRRSV MLV or FLEX vaccines had significantly higher ADG (DPV 0–42) compared to pigs in the Mock group (
Fig. 2A
3.2 PRRSV vaccination reduced pathological changes in the lung
PRRSV causes a systemic infection frequently resulting in interstitial pneumonia. Grossly, this manifests as diffusely non-collapsing, edematous, and often tan to hyperemic mottled lungs. On DPC14, Mock vaccinated pigs had a significantly higher percentage of the gross lung affected compared to MLV and FLEX vaccinated pigs (
Fig. 3A
3.3 PRRSV vaccination limited virus in the lung
The marked decrease in the severity of gross and histopathologic lesions from DPC14 to DPC21 warranted investigation into the presence of PRRSV in the lungs of challenged pigs across both timepoints with the working hypothesis that the noted resolving disease followed a reduction of virus in the lung. Quantitative PRRSV RT-PCR revealed significantly more nucleic acid in the lung of Mock pigs compared to FLEX vaccinated pigs on DPC14 (
Fig. 4A
PRRSV ISH and quantitative image analysis were employed to assess the abundance of viral nucleic acid within sections of cranioventral lung complimentary to lung tissue sections collected for PCR ( Fig. 4B-D). While there was a strong linear correlation with PRRSV RT-PCR ( r = 0.83, p < 0.0001), there were notable differences in results between ISH and RT-PCR ( Fig. 4E). Using an ISH probe able to detect our challenge PRRSV but not MLV, there was significantly less PRRSV detected in the lungs of both MLV and FLEX pigs compared to Mock pigs on DPC14. No significant differences were found between the levels of PRRSV hybridization between the two vaccine groups at either time point.
3.4 Vaccinated groups had similar numbers of circulating PRRSV-specific T cells
To investigate potential differences in immune response after vaccination, the T cell response was evaluated in PBMCs using a PRRSV-stimulated IFN-γ ELISpot. Three PRRSV isolates were used for stimulation: the MLV virus that was in both vaccines, the challenge isolate (1–7-4), and a heterologous 1–4-4 virus circulating in the midwestern region of the United States in 2022. There were no significant differences between ELISpot results observed when comparing the MLV and FLEX groups across the different time points or viruses used for cellular stimulation (
Fig. 5
In the ELISpot assay, the MLV vaccine virus did not stimulate as strong of an IFN-γ response compared to the two wild type PRRSV isolates. For example, on DPV14, there were several pigs in both the MLV and FLEX groups that exhibited high numbers of IFN-γ spots in response to the 1–7-4 challenge virus even though they had yet to be exposed to that isolate ( Fig. 5A). Comparatively, the MLV vaccine virus elicited a very low number of spots in those pigs at that time point even though the same MOI was used for all three viruses.
3.5 Mock vaccinated pigs had a robust T cell response in the lung
The T cell response in the lung, the primary tissue of interest in PRRSV infection, was evaluated with direct detection techniques for T cells (CD3 IHC) and effector molecule expression (IFNγ and GZMB ISH) (
Fig. 6A-D
Next, T cells expressing either IFN-γ or GZMB were evaluated: double positive CD3 and IFN-γ or GZMB cells were enumerated and divided by the area of evaluated tissue. Results were similar between both IFN-γ and GZMB expressing T cells, apart from approximately 4–5-fold more GZMB + T cells compared to IFN-γ + T cells ( Fig. 6F&G). Similar to the T cell infiltration results, on DPC14, there were significantly more T cells expressing IFN-γ or GZMB in Mock pigs compared to the two vaccinated groups, and there was a significant reduction in the number of these cells in the Mock group one week later (DPC21). There was not a statistically significant difference between any of the challenged groups on DPC21, and there was no difference between the MLV and FLEX groups for T cells in the lung at either time point.
IFN-γ is an anti-viral cytokine that can be secreted by more than just T cells and has previously been shown to potently inhibit PRRSV replication [ 39]. However, there was only rare expression of IFN-γ outside of T cells ( Fig. 6B) with no differences noted between treatment groups at any time point. GZMB is an important effector secreted by cytotoxic T cells and NK cells that can cause apoptosis in viral infected cells [ 40]. There were numerous non-T cells with low level expression of GZMB noted in each section ( Fig. 6D). Interestingly, on DPC14, there were significantly more GZMB transcripts (foci) in CD3 − cells in Mock pigs compared to the vaccinated groups ( Fig. 6H). Surprisingly, at this same time point, there were significantly more of these foci in FLEX pigs compared to MLV pigs.
Evaluation of the IHC/ISH-stained slides identified ISH + T cells with one or multiple ISH foci per cell. Each ISH focus corresponds to an individual RNA copy [ 41]. Therefore, we investigated whether treatment impacted the number of copies for IFN-γ and GZMB per T cell. The mean number of ISH copies, for both IFN-γ and GZMB, per positive T cell were evaluated for all time points with no significant differences between groups noted.
3.6 Absence of a neutralizing antibody response against the PRRSV challenge virus
To assess the possible impact of the antibody response on the resolution of PRRSV infection in the lung, neutralizing antibody titers against the challenge virus (1–7-4) were evaluated in MLV vaccinated and Mock pigs on DPC0, 14, and 21 ( Fig. 6I). Only the MLV vaccinated pigs had a titer, which was very low at 1:4. Furthermore, only 3/8 pigs had a titer on DPC14 and 4/8 on DPC21. No Mock vaccinated pigs showed a neutralizing antibody titer against the challenge virus during the trial.
3.7 Associations between PRRSV infection and T cell response in the lung
Within our data sets, there were notable trends between the PRRSV infection and the T cell response in the lung. These were investigated further to determine if temporal associations existed between the two. Principally, at both time points for the Mock group, the high level of PRRSV detected in the lung was compared with the robust T cell response. Interestingly, there was a moderately strong association between the number of IFN-γ
+ T cells in lung parenchyma and PRRSV ISH detected (
r = 0.5321,
p = 0.0438) (
Fig. 7A
The amount of detected PRRSV was also compared to GZMB + T cells in Mock pigs at both necropsy time points. Here, there was a weaker r value ( r = 0.425), and the p value was non-significant at 0.11 ( Fig. 7B). Interestingly, when comparing PRRSV and GZMB + T cells for vaccinated (MLV and FLEX) pigs, there was a weak negative correlation r = −0.3350, but with a p value of 0.0609 ( Fig. 7C).
We also investigated if PBMC IFN-γ ELISpot results were predictive of the IFN-γ response in the lung. However, there was no correlation between ELISpot and lung (IHC/ISH) IFN-γ expression at DPC21.
4 Discussion
PRRSV MLV vaccines and prior infection have previously been shown to provide partial cross protection to heterologous challenge, which was reproduced here [ 16, 21, 42]. When compared to Mock pigs, vaccinated pigs (MLV and FLEX) had significantly lower levels of PRRSV nucleic acid in the blood and lung, along with less severe gross and histologic lesions at the DPC14 time point. Importantly, vaccinated pigs gained significantly more weight during the study. The quantity of PRRSV RNA hybridization in the lung sections and viremia along with pathology results show that the level of protection against PRRSV infection and disease conferred by MLV and FLEX was comparable. Likewise, the magnitude of the T-cell response in both the blood and lung and seroconversion was remarkably similar in both vaccinated groups. Additionally, the vaccinated groups demonstrated nearly identical gross and microscopic lesion scores as well as average daily gain. Collectively, MLV and FLEX vaccines provided partial protection to a heterologous PRRSV infection in a nearly identical fashion.
The absence of an elevation in IFN-γ or GZMB expressing T cells in the lung in both vaccinated groups above that of the control animals at 14 and 21 DPC was unexpected, as a memory T cell response appears detectable in PBMCs 4 weeks after inoculation [ 43, 44]. Furthermore, recent work has shown a robust PRRSV-specific T cell response in the lung of previously vaccinated pigs at DPC14 [ 45]. In our study, a memory T cell response in the lung likely occurred around 5–10 days post heterologous challenge and was absent by our DPC14 sampling point as PRRSV infection in the lung was already significantly reduced compared to Mock [ 42]. The ELISpot data supports the existence of cross-reactive memory T cells in vaccinated pigs, as 14 days after vaccination many pigs had T cells in their blood that recognized the virus they would be challenged with two weeks later. Future studies should focus on the temporal, spatial, and phenotypic characterization of these memory and effector T cells in the lungs and blood of previously vaccinated and experimentally inoculated pigs, as understanding these intricacies appears foundational for the identification of conserved T cell epitopes on PRRSV. This approach has been applied to several other rapidly mutating RNA viruses and has provided insight for the development of experimental vaccines that may elicit greater cross-protection against challenge [ 46–48].
In this study, IFN-γ and GZMB expression in the lung were evaluated with ISH. While recognition of both proteins, versus mRNA, would have been preferable, this was prevented by the lack of IHC validated antibodies for porcine GZMB and IFN-γ. Furthermore, the low background of the ISH for both targets was ideal for individual cell quantification. One limitation is that the observed IFN-γ and GZMB expression in the lung may not all be PRRSV-specific responses, evidenced by the results in the negative control animals. Cellular isolation from the lung, which was not possible in this study, would be necessary to characterize the PRRSV-specific T cell response and would also allow for further phenotyping of T cells with flow cytometry which would have allowed for the differentiation of T cell subset populations. The use of fixed tissue sections for characterization of IFN-γ and GZMB expression has the benefit of capturing expression at the time of tissue fixation, shortly after death; whereas, cellular isolation has the addition variables of tissue dissociation, red blood cell lysis, freezing, thawing, and then culture which may impact results. The utilization of both in situ direct detection and flow cytometry techniques would allow for a thorough evaluation of vaccine induced cell-mediated immunity.
The quantitative analysis of IFN-γ and GZMB expressing T cells and PRRSV ISH assays demonstrated that there is greater T cell activity in the lung when increased amounts of PRRSV nucleic acid were present. This infiltration of lymphocytes partially contributed to the elevated interstitial pneumonia scores, along with a large influx of macrophages [ 49]. This conclusion is best supported by the immunologically naïve PRRSV challenged pigs in this study, which demonstrated an intriguing aspect of PRRSV infection dynamics and immune response. Despite developing the most severe degree of interstitial pneumonia and having the highest levels of PRRSV nucleic acid 14 days after PRRSV challenge, the immunologically naïve pigs reduced the amount of viral nucleic acid within the lung and interstitial pneumonia to levels observed in vaccinated pigs by DPC21. This is not to say that these pigs were completely clinically resolved at this time point, as the virus is known to persist in swine and continue to cause clinical signs in those that survive initial infection [ 27, 50]. However, the presented data show that after challenge, T cells expressing IFN-γ and GZMB infiltrated the lungs of Mock vaccinated pigs. Subsequently, the bulk of PRRSV was cleared from the lung in the absence of detectable neutralizing antibodies against the infecting PRRSV isolate. While the role of other immune cells in the control of PRRSV cannot be ruled out, it seems likely that T cells played an integral role in virus elimination. Furthermore, the possible negative correlation between GZMB+ T cells and PRRSV in the lung of vaccinated pigs suggests that cytotoxic T cells are particularly important for protection, although further investigation is warranted. Taken together, these findings suggest that the CMI response is the most important effector of early PRRSV clearance from the lung and further advocates for the identification of conserved T cell epitopes on PRRSV.
Upon visual assessment of CD3/GZMB stained slides, the abundance of granzyme B transcription in CD3 − cells was perhaps not surprising, as numerous cell types are known to express the pleiotropic GZMB [ 51, 52]. However, this finding became more intriguing when GZMB transcripts were quantified and separated by treatment groups for DPC14. The significant elevation of GZMB transcripts in the cells of Mock pigs compared to vaccinated pigs could be due to non-cytotoxic functions in other immune cells, apoptosis, or possibly NK cells functioning in a cytotoxic manner on viral infected cells [ 53]. The difference between FLEX and MLV groups was unexpected, especially considering the numerous other similarities between these two groups. In any case, further investigation may be warranted to determine if this non-T cell GZMB expression is important for clearance of virus or related to one of its many other functions [ 54].
In this study, two methods were used for the quantitation of PRRSV in the lung. While the results from both methods correlate very strongly, the discrepancies between the ISH and PCR modalities detected in the MLV group at DPC14 are worthy of consideration. When using only PCR results, one might conclude that MLV vaccinated pigs and Mock naïve pigs showed no difference in the amount of PRRSV in the lung, whereas the pigs in the FLEX group showed lower amounts of viral replication. However, this was not the case when analyzed with ISH, where vaccine groups appeared equivocal, and both had lower levels of PRRSV hybridization in the lung compared to naive counterparts. This suggests the possibility that subtle differences between groups may go unnoticed if using PCR as the sole modality to measure viral nucleic acid in tissue.
The discrepancies in results between the ISH and PCR modalities may be attributed to differences in the specificities of each assay. The ISH probe was designed to detect the challenge virus without detecting the vaccine virus. However, the PRRSV PCR that was used does not distinguish between vaccine and wild type virus and was likely detecting both in the lung samples from our vaccinated and challenged pigs. Additionally, it is known that PCR assays can produce false positive results, and any viral material in contact with the tissue to be tested could be quantified, regardless of its source [ 55, 56].
This study highlights some of the strengths of using direct detection techniques and quantitative image analysis software. During algorithm development and tissue annotation, the informed user makes determinations based on what is visible to them and can visually inspect any outliers or unexpected results. It is important that the user remains blinded to groups and establishes inclusion criteria to avoid bias when training these algorithms or removing artifactual staining, just as it is imperative for PCR technicians to follow standard operating procedures and use of proper aseptic technique.
5 Conclusions
The presented study showed the addition of a PCV2 subunit vaccine to PRRSV MLV vaccine did not affect heterologous immunity or protection against heterologous PRRSV challenge compared to PRRSV MLV alone. The novel utilization of direct detection techniques coupled with machine learning quantitative analysis demonstrated the role of the T cell in the clearance of PRRSV from the lung in the absence of neutralizing antibodies. These results warrant further investigation into the PRRSV-specific T cell response in the lung, as well as the identification and characterization of conserved T cell epitopes on PRRSV that may be targeted for the generation of more broadly protective vaccines.
CRediT authorship contribution statement
Andrew Noel: Writing – original draft, Formal analysis, Data curation. Jianqiang Zhang: Writing – review & editing, Methodology, Investigation, Formal analysis, Conceptualization. Teerawut Nedumpun: Writing – review & editing, Investigation, Formal analysis, Data curation. Panchan Sitthicharoenchai: Writing – review & editing, Investigation, Data curation. Baoqing Guo: Writing – review & editing, Investigation, Data curation. Reid Phillips: Writing – review & editing, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Marius Kunze: Writing – review & editing, Project administration, Investigation, Funding acquisition, Conceptualization. Oliver Gomez-Duran: Writing – review & editing, Funding acquisition, Conceptualization. Jennifer Groeltz-Thrush: Writing – review & editing, Validation, Data curation. Emily Rahe: Writing – review & editing, Investigation, Data curation. Michael C. Rahe: Writing – review & editing, Writing – original draft, Supervision, Project administration, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.
Funding
This study was supported by
Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Michael Rahe reports financial support was provided by Boehringer Ingelheim Vetmedica GmbH. Michael Rahe reports a relationship with Boehringer Ingelheim Vetmedica GmbH that includes: speaking and lecture fees. Teerawut Nedumpun reports a relationship with Boehringer Ingelheim Vetmedica GmbH that includes: speaking and lecture fees. Jianqiang Zhang reports grant support was provided by Boehringer Ingelheim over the past three years. Co-authors Oliver Gomez-Duran and Marius Kunze are employed by Boehringer Ingelheim, the company that makes both the Ingelvac and FLEX vaccines. Reid Phillips has retired from Boehringer Ingelheim. All other authors declare no competing interests. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We thank the group from Nikon Instruments Inc. for their work on the direct detection image quantitation: Henning Mann, Guarav Joshi, and Michael Yang. The authors would also like to thank Yang Qu for his consultation on the statistical analysis of our data. Yang was supported by the North Carolina State University College of Veterinary Medicine Office of Research.
© 2025 The Authors