Equine arteritis virus long-term persistence is orchestrated by CD8+ T lymphocyte transcription factors, inhibitory receptors, and the CXCL16/CXCR6 axis

Abstract

Equine arteritis virus (EAV) has the unique ability to establish long-term persistent infection in the reproductive tract of stallions and be sexually transmitted. Previous studies showed that long-term persistent infection is associated with a specific allele of the CXCL16 gene (CXCL16S) and that persistence is maintained despite the presence of local inflammatory and humoral and mucosal antibody responses. Here, we performed transcriptomic analysis of the ampullae, the primary site of EAV persistence in long-term EAV carrier stallions, to understand the molecular signatures of viral persistence. We demonstrated that the local CD8+ T lymphocyte response is predominantly orchestrated by the transcription factors eomesodermin (EOMES) and nuclear factor of activated T-cells cytoplasmic 2 (NFATC2), which is likely modulated by the upregulation of inhibitory receptors. Most importantly, EAV persistence is associated with an enhanced expression of CXCL16 and CXCR6 by infiltrating lymphocytes, providing evidence of the implication of this chemokine axis in the pathogenesis of persistent EAV infection in the stallion reproductive tract. Furthermore, we have established a link between the CXCL16 genotype and the gene expression profile in the ampullae of the stallion reproductive tract. Specifically, CXCL16 acts as a "hub" gene likely driving a specific transcriptional network. The findings herein are novel and strongly suggest that RNA viruses such as EAV could exploit the CXCL16/CXCR6 axis in order to modulate local inflammatory and immune responses in the male reproductive tract by inducing a dysfunctional CD8+ T lymphocyte response and unique lymphocyte homing in the reproductive tract.

Fig 1. Analysis workflow carried out in this study.

Transcriptome analysis of the ampullae (RNAseq) was performed along with subsequent extensive GO and pathway analysis, IHC and ISH. EAV, equine arteritis virus; DGEA, differential gene expression analysis; WGCNA, weighted gene co-expression network analysis; LT, EAV long-term carrier stallion; ST, EAV short-term carrier stallion; N, naïve stallion; TF, transcription factor; GO, gene ontology; IHC, immunohistochemistry; ISH, in situ hybridization.

Fig 2. Differential gene expression profile in the ampullae of long-term (n = 3) and short-term (n = 6) EAV carrier stallions compared to the naïve group (n = 3).

(A) The Venn diagram depicts the differentially expressed genes (DEGs) common between long-term and short-term carrier stallions, with a clear majority being upregulated. (B and C) GO analysis (biological process and molecular function) of commonly upregulated genes between long-term and short-term carrier stallions reveal involvement in biological processes associated with cell adhesion, extracellular matrix organization, response to wounding, integrin signaling, among others, with a high proportion of genes presenting binding and catalytic activities. (D) Significant pathways associated with commonly upregulated genes.

Fig 3. Differential gene expression profile in the ampullae of long-term EAV carrier stallions (n = 3) reveal significant differences compared to short-term carrier stallions (n = 6).

(A) Heatmap depicting the expression pattern of the differentially expressed genes (DEGs) between long-term and short-term carrier stallions (n = 390 genes). The majority of the DEGs between these two groups were upregulated. Darker reds are indicative of a higher expression. The interactive heatmap can be found here (B) Representative heatmaps showing changes in gene expression levels during long and short-term viral persistence. DEGs (n = 390) were classified based on their Gene Ontology (GO) terms and selected categories are depicted. Differential gene expression analysis revealed a significant upregulation of genes involved in adaptive (the interactive heatmap can be found here) and innate (the interactive heatmap can be found here) immune responses, response to virus infection (the interactive heatmap can be found here), chemotaxis (the interactive heatmap can be found here), apoptosis (the interactive heatmap can be found here) and adhesion (the interactive heatmap can be found here). In addition, perturbation in the gene expression of several transcription factors was observed (the interactive heatmap can be found here). Darker reds are indicative of a higher expression. N, naïve; LT, long-term carrier stallions; ST, short-term carrier stallions. (C) Top 25 canonical pathways associated with the DEGs (n = 390) observed between long-term and short-term carrier stallions as determined by Ingenuity Pathway Analysis (IPA). Canonical pathways were predominantly associated with T lymphocyte pathways and signaling mechanisms.

Fig 4. Further characterization of the inflammatory response at the site of persistence during EAV long-term persistent infection.

(A-C) Presence of scattered granzyme B⁺ cells (C, arrowheads) in inflammatory infiltrates of long-term carrier stallions. (D-F) The nuclear expression of AR was predominant in the glandular epithelia and scattered stromal cells regardless of infection status. Interestingly, cells within inflammatory infiltrates showed AR expression (F, arrowheads). Granzyme B and AR-specific immunostaining. DAB. 400X. Bar = 20 μm. (G-I) Inflammatory infiltrates in long-term carrier stallions presented scattered CTLA-4⁺ T lymphocytes (I, arrowheads). CTLA-4-specific immunostaining. Fast Red. 400X. Bar = 20 μm.

Fig 5. Gene expression analysis of selected genes, transcription factor enrichment analysis and molecular networks associated with long-term EAV persistence.

(A) Heatmap depicting gene expression profiling by RT-qPCR of selected genes. Upregulation of specific chemokines/cytokines and chemokine receptors (including CXCL16 and CXCR6), selected transcription factors and inhibitory receptors was observed. The heatmap was generated using -ΔCt values. (B) Analysis of putative transcription factor site enrichment in DEGs between long-term and short-term carrier stallions using CiiiDER. Color and size of circles reflect p-value of enrichment. Over-represented transcription factors of potential interest are depicted. Those differentially expressed in long-term carrier stallions are depicted in bold. (C) Molecular network associated with DEGs observed between long-term and short-term carrier stallions. The network is driven by specific transcription factors, some of which were over-represented as determined by transcription factor binding site enrichment analysis. Upregulated and downregulated genes (log₂ fold-change compared to short-term carrier stallions) are depicted in red and green color, respectively. The degree in color intensity reflects the magnitude of the fold-change, where intense red or green indicate a higher or lower fold-change, respectively. Only direct relationships are shown.

Fig 6. T-helper 1 (Th1) pathway analysis based on DEGs observed between long-term and short-term carrier stallions.

Numerous transcription factor genes associated with the Th1 pathway are upregulated in long-term persistently infected stallions along with other related genes including cytokines. Upregulated genes (log₂ fold-change compared to short-term carrier stallions) are depicted in red. Non-differentially expressed genes within the pathway are depicted in grey. The degree in color intensity reflects the magnitude of the fold-change, where intense red indicates a higher fold-change. Direct (solid arrows) and indirect (dashed arrows) relationships are shown.

Fig 7. Analysis of a subset of transcription factors in the ampullae of naïve, short-term and long-term carrier stallions by immunohistochemistry determined the predominance of EOMES and NFATC2 in inflammatory infiltrates during EAV persistence.

(A-C) EOMES-specific immunostaining. (D-F) NFATC2-specific immunostaining. (G-I) TBX21-specific immunostaining. (J-L) PRDM1-specific immunostaining. (M-O) Phosphorylated Akt-specific immunostaining. DAB. 200X. Bar = 250 μm.

Fig 8. Quantitative immunohistochemical analysis of specific transcription factors (EOMES, NFATC2, PRDM1 and TBX21) in inflammatory infiltrates of long-term persistently infected stallions.

The number of positive cells was determined in a total of five equivalent inflammatory infiltrates per stallion. Blue dots represent values for individual infiltrates, while the mean number of positive cells/inflammatory infiltrate ± standard error of the mean are represented in red, respectively. Letters (a, b and c) indicate statistically significant differences between transcription factors (p-values < 0.001).

Fig 9. Putative genes regulated by or interacting with the transcription factors analyzed in the ampullae of long-term EAV carrier stallions.

(A) EOMES. (B) NFATC2. (C) TBX21. (D) PRDM1. Gene lists were generated from the immune-database Immuno-Navigator, the Ingenuity Knowledgebase and literature search. Transcription factors of interest are shown in circles, where the size of the circle is indicative of their predominance in inflammatory infiltrates as determined in this study. DEGs identified during EAV persistence are depicted in orange while those that were not differentially expressed are shown in light blue.

Fig 10. CXCL16 and CXCR6 expression in the ampullae of naïve (n = 3), EAV short-term (n = 6) and long-term carrier stallions.

(A-C) CXCL16 is significantly upregulated in the ampullae of EAV long-term persistently infected stallions (C) compared to short-term carrier (B) and naïve stallions (A) as determined by ISH. CXCL16 is predominantly expressed in the mucosal epithelium and lymphocytic infiltrates (C, inset), especially in the luminal area. CXCL16-specific mRNA RNAscope ISH, Fast Red, 400X. Bar = 20 μm. (D-F) CXCR6 is significantly upregulated in the ampullae of EAV long-term persistently infected stallions (F) compared to short-term carrier (E) and naïve stallions (D) as determined by ISH. CXCR6 was expressed in lymphocytic infiltrates and co-localized with CXCL16 expression in inflammatory cells (F, inset). CXCR6-specific mRNA RNAscope ISH, Fast Red, 400X. Bar = 20 μm. (G) Higher expression of CXCL16 and CXCR6 in the ampullae of long-term carrier stallions (n = 3) was quantitatively determined by (left to right) TaqMan RT-qPCR and RNAscope ISH. Blue dots represent values for individual stallions, while mean 2^-ΔCt values ± standard error of the mean and median scores are represented in red, respectively. Letters indicate statistically significant differences between groups (p-value<0.05).

Fig 11. Differential gene expression and weighted gene co-expression network analysis in stallions homozygous and heterozygous for the CXCL16S allele (CXCL16S/CXCL16S and CXCL16S/CXCL16R; n = 5).

(A) A total of 542 genes were differentially expressed between animals exhibiting different genotype (i.e. between stallions carrying the susceptibility allele [CXCL16S, n = 5] and those who did not [homozygous for CXCL16R, n = 4]) and following EAV infection. A lower proportion (188 genes) were upregulated in stallions carrying the CXCL16S allele. Darker reds are indicative of a higher expression. The interactive heatmap can be found herehttp://rpubs.com/pouyadini/379472. (B and C) Gene module identification as determined by weighted gene co-expression network analysis (WGCNA). A total of 24 gene modules with varying number of genes were identified among 12,303 selected genes. (D) Gene co-expression network construction for selected genes. The heatmap depicts the topological overlap (correlation between the expression profiles of each pair of genes). Each row and column correspond to a gene, where light color denotes low topological overlap with progressively darker color indicating higher topological overlap. Dark squares along the diagonal represent gene modules. (E) Module-trait correlation heatmap based on the correlation analysis of module eigengenes (MEs) and percentage of CD3⁺ T lymphocytes susceptible to in vitro EAV infection in stallions (n = 12). Pearson correlation values (r) are shown for each module along with the respective p-value within parenthesis. The blue and lightyellow modules (arrows) presented a significant positive correlation with CD3⁺ T lymphocyte susceptibility (p-values < 0.05).

Fig 12. “Hub” gene and neighbor identification in the blue and lightyellow modules.

(A) Network visualization of the blue and lightyellow modules (n = 1,195 genes and n = 130 genes, respectively). Highly connected, putative “hub” genes within the module are shown in the squared area. “Hub” genes were selected based on a module membership (MM) value ≥ 0.90, a p-value < 0.05 and gene significance (GS) ≥ 0.5. Node color reflects node degree (number of connections to other genes in the network), where darker reds indicate a higher degree. Network analysis was performed using the NetworkAnalyzer tool in Cytoscape. (B) CXCL16 first neighbors as determined by network analysis of the blue module. CXCL16 was identified as a “hub” gene within the blue module and several of its neighbor genes are involved in immune-related pathways such as CCL2, BATF, CD3G, FASLG, IRF9, CD8A, THEMIS, among others. Node color and size have been mapped to reflect node degree, where darker reds and larger nodes indicate a higher degree. (C) “Hub” genes in the lightyellow module network. (D) “Hub” transcription factor genes belonging the green module derived from the transcription factor network constructed from stallions carrying the CXCL16S allele (homozygous and heterozygous). Node size reflects node degree (number of connections to other genes in the network). Transcription factors identified as differentially expressed between stallions homozygous or heterozygous for CXCL16S and those homozygous for CXCL16R are colored in red. Interestingly, transcription factors that were over-represented during long-term EAV persistence are among the “hub” transcription factors identified in stallions homozygous or heterozygous for CXCL16S.

Fig 13. Schematic representation of the current knowledge of EAV persistence in the reproductive tract of the stallion.

EAV tropism during persistent infection is associated with CD8⁺ T lymphocytes and CD21⁺ B lymphocytes infiltrating the ampullae, fibrocytes, tissue macrophages and dendritic cells. Long-term persistence is associated with a strong inflammatory response primarily mediated by CD8⁺ T lymphocytes and a smaller population of CD4⁺ T helper 1 lymphocytes. Additionally, local plasma cells produce diverse EAV-specific immunoglobulin isotypes with varying neutralizing abilities that are shed into the seminal plasma. Long-term persistence is associated with an enhanced expression of the chemokine CXCL16 in the mucosal epithelium and lymphocytes and with the downregulation of seminal exosome eca-mir-128, a putative modulator of the CXCL16/CXCR6 axis. Expression of the CXCL16 receptor, CXCR6, is also upregulated in lymphocytes, and along with other chemokines and receptors (CXCL9-11, CXCR3) likely mediates specific homing of infected lymphocytes into the reproductive tract of the stallion and migration of these cells across the epithelial lining, and drives an immunologically unique microenvironment that favors viral persistence. The inflammatory process is accompanied by upregulation of inhibitory receptors (PDCD1, CTLA-4, among others) and predominance of the transcription factors EOMES and NFATC2, potential mediators of CD8⁺ T lymphocyte hyporesponsiveness that leads to viral persistence. CD8⁺ Tc, CD8⁺ T lymphocytes; CD4⁺ Th1, CD4⁺ T-helper 1 lymphocytes; CD21⁺ Bc, CD21⁺ B lymphocytes.