Bluetongue (BT) is an infectious disease transmitted by Culicoides midges caused by Bluetongue virus (BTV), which is the prototype member of the Orbivirus genus within the Sedoreoviridae family (1). BTV genome consists of 10 double-stranded RNA segments encoding 7 structural proteins (VP1-VP7) and at least 4 non-structural proteins (NS1-NS4) (2, 3). This segmented genome favors the appearance of new serotypes due to the possibility of segment reassortment during co-infection with different serotypes (4, 5). So far up to 29 serotypes have been described (6), although the coexistence of multiple serotypes in the same territory leads to the ongoing emergence of new strains (7). Similarly, the use of live attenuated vaccines could contribute to the appearance of new serotypes, as vaccine strains can reassort with wild type circulating virus (8, 9).

BTV is currently spread through Asia, the Middle East, Australia, South and Central America and Southern Europe, making it a disease of compulsory notification to the WOAH (World Organization for Animal Health) (10). Sheep, goats, cattle and wild ruminants are the main hosts of the disease, which causes high rates of mortality in naïve herds, abortions, weight loss and decrease in fertility and milk production that result in great economic losses for the livestock industry (11–13).

Following the bite of an infected midge, BTV initially replicates in endothelial cells before infecting lymphocytes and cells of the monocytic lineage (14–17). The virus is then transported through the lymph, colonizing the draining lymph nodes and spreading to peripheral tissues (18).BTV is also capable of “hiding” in erythrocytes, which promotes the spread to other tissues (19). BTV has the particularity to produce prolonged infections in ruminants, which probably favors the transmission of the virus back to the Culicoides spp. vector. The host can be infectious for biting midges for at least 21 days; infectious virus can be isolated from blood for up to 49 days; and in some cases, viral RNA can be detected in blood for more than 200 days (20). The prolonged viremia indicates that BTV has mechanisms that allow it to evade immunity to extend its circulation in the host.

BTV can produce leucopenia during the early stages of experimental infections (21–24) and this could be due to virus-related apoptosis of peripheral blood mononuclear cells (23). The infection not only affects leucocytes in the periphery but also in the lymph nodes (25). Indeed, BTV is known to be transported to the draining lymph nodes by conventional DC (26). Once in the lymph node, BTV disrupts follicular DC activity, which results in delayed antibody responses (27). BTV effects on cellular immunity are less well characterized, but early evidence suggested that BTV could produce transient T cell immunosuppression (22). It is therefore critical to understand the interaction of BTV with the host immune system to further our knowledge of BTV infections.

Humoral and cell-mediated immunity play a critical role to overcome the disease. Although neutralizing antibodies can confer long-lasting protection against reinfection with a homologous BTV serotype (28), cellular immunity can confer protection against homologous and heterologous BTV infection in the absence of neutralizing antibodies (29–32). CD4⁺ and CD8⁺ T-lymphocytes, which are activated upon antigen presentation, govern cell-mediated immune responses to BTV that typically result in CD8⁺ T-cell expansion two weeks after BTV infection (21, 24).

Upon BTV infection, dsRNA is detected by pattern recognition receptors (PRRs), in particular toll-like receptors and retinoic acid inducible gene (RIG-1)-like family receptors (33), triggering the production of interferon and other proinflammatory cytokines to activate an antiviral response to combat the infection. Despite the studies on transcriptome analysis of BTV infected peripheral blood mononuclear cells (PBMCs) confirming that part of the differentially expressed genes (DEGs) during infection are involved in PRR activation and induction of cytokine signaling (34), little information is known about the transcriptomic characterization of the immune response against BTV. To date, there have been no in vivo reports of transcriptome profiling of PBMC subpopulations in BTV-infected sheep. We have therefore carried out in the present study a transcriptomic analysis of B, CD14⁺, CD4⁺ and CD8⁺ cells from four sheep infected with BTV-8. In addition, transcriptomic data have been contrasted with functional analysis by flow cytometry and interferon production assays. The present study furthers our understanding of BTV immunomodulatory activity on the host immune cells.

In the present work, we characterized the transcriptome of B cells, CD4⁺ and CD8⁺ T cells, and monocytes in response to in vivo BTV infections in the natural host. To perform this analysis, we experimentally infected 4 sheep with BTV-8. All animals developed severe BT disease, with high fever, apathy, and lack of appetite that peaked at day 7pi. Viremia was still readily detectable by RT-qPCR at the end of the experiment (day 17pi), which is characteristic of the prolonged presence of virus in circulation during BTV infections (20). B cell and monocyte transcriptome indicated that these cells are involved in viral recognition at early timepoints of the infection. We observed a severe inhibition of polyclonal CD4⁺ and CD8⁺ T cell responses at day 7pi, which coincided with the upregulation of the PD1/PD-L1 axis in CD4⁺ T cells, and the downregulation of TCR signaling in CD8⁺ T cells. Functional analysis of T cells responses showed that both CD4⁺ and CD8⁺ T cells activation was impaired at day 7pi, a hallmark of immune checkpoint activation pathways in T cells. Our data also show that BTV induces PD-L1 expression in T cells and monocytes, a process that could mediate T cell response attenuation.

Previous transcriptomic analysis of in vitro BTV infection of ovine and caprine PBMCs pointed at the activation of genes related to immunity, which belonged, among others, to PRR and cytokine/chemokine signaling pathways (34). In the present work, we present the transcriptomic analysis of 4 PBMC subpopulations after in vivo BTV infection in sheep. We found that IFN-related genes DHX58 (LGBP2), IFIT2, IFIT3, ISG15, MX2, OAS2, and XAF1 were upregulated at day 3 and 7pi in all 4 PBMC subpopulations. Singh et al. also reported that IFIT2 and IFIT3 are upregulated in sheep and goat PBMCs infected in vitro with BTV-16 (34). IFIT expression is unconventional among IFN-stimulated genes. IFITs are typically endogenously expressed at low levels since they can act as sensor for the infection by recognizing viral RNA motifs (62). Indeed, IFIT2 can recognize dsRNA while IFIT3 modulates IFIT2-driven apoptosis (63), thus it is plausible that IFIT2 could participate in BTV RNA recognition. A recent study also showed that IFIT1 is important in restricting BTV replication in bovine and ovine primary cells (64). IFIT expression is also induced upon IFN signaling. IFITs can inhibit viral protein translation and promote antiviral signal transduction. Our data therefore confirms that IFIT could play an important role in the host innate immune response to BTV. Further work analyzing IFIT contribution to BTV immunity is nonetheless required to address the precise role of these IFN-stimulated genes in BTV infection. The upregulation of IFN-related genes in B, CD14⁺, CD4⁺, and CD8⁺ cells is indicating that the host is detecting the presence of the viral infection and consequently activating the IFN response. This is a predictable outcome since BTV has been described as a good inducer of IFN, and infection leads to IFN production in cattle and sheep (65–67). Differences in the IFN response kinetics between cattle and sheep could actually be responsible for host susceptibility to disease (64).

Our comprehensive transcriptomic analysis also indicated that B cells and monocytes are involved in the sensing of the infection from day 3pi. Pathway enrichment analysis indicated that cell activation occurs for B cells and monocytes, at least in the periphery, following BTV infection. Upregulation of genes involved in B cell activation, such as CD19 or TNFSF13B, occurred by day 3pi, although this was not maintained at day 7pi. We have previously described that BTV can target follicular DC to delay antibody responses (27). Although our transcriptomic analysis did not detect a significant impairment of B cell activity in the periphery, we found that STAT5B was upregulated at day 7pi. STAT5B has recently been identified as a repressor of the immunoglobulin class switch and the transcriptional program that leads to plasma cell differentiation (68). Thus, it could be speculated that transcriptional mechanisms could be promoted by the infection to delay antibody production in spite of the apparent activation of B cells. Alternatively, follicular DC disruption by the infection could be sufficient to impair B cell activation locally and delay the antibody response. Further characterization of the B cell response to BTV will be required to better elucidate the intricate regulation of antibody production in this infection.

Our flow cytometry analysis of monocytes also showed that the CD14⁺CD16⁺ monocyte subpopulation expanded by day 7pi. Expansion of this monocyte subset has also been reported during PPRV infection in sheep (69). This monocyte population may therefore play a role in the ovine immune response to viral infections. Full characterization of this CD14⁺ CD16⁺ monocyte population would require the introduction of other surface markers, such as CD172a as previously done in cattle (49), which would allow for a clearer differentiation between classical, activated and intermediate/non-classical monocytes. Although monocyte subset functions have not been fully characterized in sheep (48), their bovine counterparts appear to share similar activity to the human equivalent monocyte population (70, 71). Recent studies have further confirmed this shared phenotype of monocytes in cattle and humans at the transcriptomic level (49, 72). Broadly speaking in human and cattle, classical monocyte subsets can drive a pro-inflammatory response while non-classical monocyte subsets are thought to be anti-inflammatory and involved in vascular repair. Along with the upregulation of IFN-related genes in monocytes, we found that proinflammatory cytokine genes IL23A, TNF and IL6 were upregulated at day 7pi, which coincided with the increased presence of CD14⁺CD16⁺ cells in the monocyte subset. This could indicate that an inflammatory response is promoted by monocytes at day 7pi and that activation of classical monocytes is occurring. Nonetheless, we concomitantly detected a consistent upregulation of the IL27 gene throughout the infection. IL-27 is typically associated with the induction of the regulatory cytokine IL-10 and the promotion of T regulatory cell activity, although it can also be critical for CD8⁺ T cell activation (reviewed in (73)). CD14⁺ cell transcriptomic analysis also revealed that the genes expressing the immune checkpoint proteins TIM-3 and PD-L1 (HAVCR2 and CD274) are upregulated at day 7pi. We confirmed the increased expression of PD-L1 protein on monocyte surface by flow cytometry at this timepoint. Expression of these immune checkpoint molecules could represent a counterpoint to immune activation and the need to dampen inflammation to prevent immunopathology. These data indicate that this CD14⁺CD16⁺ monocyte subset could also include intermediate/non-classical monocytes, which could be important to repair the vascular damage produced by BTV infection. Interestingly, PD-L1 expression in non-classical monocytes has been linked to immunoregulatory functions (74), and transcriptomic studies in cattle have found that this monocyte subset expresses higher levels of PD-L1 and IL-27 (49). In future work it will be interesting to determine the specific PD-L1 expression on CD14⁺CD16⁺ monocytes, as this would help characterize this monocyte subset. Talker et al. have proposed a model of continuous differentiation for monocytes in cattle (49). They suggest that the CD14⁺CD16⁺ population could represent an heterogeneous intermediate monocyte subset that is transitioning from a classical inflammatory profile towards a non-classical anti-inflammatory profile. Our bulk transcriptomic analysis of CD14⁺ cells during viral infection also indicates that the enrichment in circulating CD14⁺CD16⁺ coincided with the upregulation of both inflammatory and anti-inflammatory gene sets in monocytes, which could fit this model. Overall, it appears that a complex response to the infection is taking place in monocytes. On the one hand, upregulation of PRR and TNF signaling pathways indicate that monocytes participate in the recognition of the infection and the initiation of the immune response, but on the other hand, expression of potent inhibitors of immunity such as IL-27, TIM-3 (75, 76), or PD-L1 (74) could also limit the stimulatory activity of these cells and promote an immunoregulatory environment. The balance between these signals will probably modulate the response of surrounding cells. It should be noted that our current transcriptomic analysis was based on CD14⁺-sorted cells, and as a result, this will encompass several monocyte subsets. Transcriptomic profiling of the different monocyte populations in sheep could shed some light on their function, and consequently help interpret the transcriptomic response of monocytes in infection.

We started detecting anti-BTV cellular immunity by day 15pi. This is in line with other reports that show that humoral and cellular anti-BTV immunity can be detected from day 9 post-infection (24), and that CD8⁺ T and B cell expansion occurs by day 10-12 at the lymph node proximal to the infection site (25). In spite of the mounting immune response, BTV is still capable of replication in the host as shown by the presence of virus in peripheral blood throughout our experimental infection. Several mechanisms of immune escape have been described for BTV. We have shown that BTV can target follicular DC to delay antibody responses (27). BTV is also known to suppress IFN responses via some of its non-structural and structural proteins (77–81). Although these mechanisms contribute to BTV immune evasion, they are unlikely to fully account for BTV capacity to prolong its circulation in the host.

Transcriptomic analysis of CD4⁺ and CD8⁺ T cells indicated that T cell responses could be disturbed at day 7pi. This coincided with the decreased number of CD8⁺ T cells in the periphery at this timepoint. Functional studies confirmed the transcriptomic data, and we found that CD4⁺ and CD8⁺ T cell activation was severely impaired at day 7 pi when compared to day 0. The limited capacity of T cell mitogens and activators to trigger a response in day 7pi cells is therefore indicative that the infection has imprinted in the periphery a program in T cells that impairs activation. These impaired T cell responses correlated with the activation of the PD-1/PD-L1 axis according to our transcriptomic data. The PD-1/PD-L1 axis is an immune checkpoint that controls T cell exhaustion, and its inhibition has proved useful for antitumor immunotherapy. Activation of the PD-1/PD-L1 axis appears to be a natural process in acute viral infections that controls excessive T cell activation, prevents immunopathogenicity, and maintains peripheral tolerance (82). The activation of the PD-1/PD-L1 axis in T cells is associated with an increase in the activation threshold required to trigger TCR signaling (83), and thus this could account for the impaired T cell phenotype observed upon BTV infection. In β-coronavirus infections, such as murine hepatitis virus and SARS-Cov2, the PD1/PD-L1 axis is involved in T cell dysfunction (84). Indeed, viruses that produce chronic infections can exploit the PD1/PD-L1 axis, among other immune checkpoints, to exhaust virus-specific T cells (82).

Although the canonical PD-1/PD-L1 exhaustion pathway involves engagement of PD-1 on T cells, our transcriptomic analysis showed it was actually PD-L1 (CD274) expression that was upregulated both in CD4⁺ and CD8⁺ T cells at day 7pi. Flow cytometry analysis confirmed an increase in the subsets of PD-L1 expressing-CD4⁺ and -CD8⁺ T cells at day 7pi. Upregulation of PD-L1 in CD4⁺ T cells has previously been reported in in vitro HIV infections (85). PD-L1 expression is upregulated in immune cells in response to pro-inflammatory cytokines or type I IFN (82). PD-L1 can be expressed in activated T cells and studies in PD-L1-deficient mice have shown that its presence on T cells limits activation (61). Indeed, in cancer PD-L1 activation on T cells weakens antitumor immunity (86), while PD-L1 signaling on human memory CD4⁺ T cells induces a regulatory phenotype (87). PD-L1 engagement on T cells promotes self-tolerance and suppresses the activity of neighboring effector T cells and macrophages (86). PD-L1⁺ T cells can impair T cell activity by interacting with PD-1 on effector cells. PD-L1 also reduces Th1 polarization and promotes Th17 differentiation in CD4⁺ T cells (86). Our transcriptomic pathway analysis showed that Th17 signaling was upregulated in CD4⁺ T cells at day 7pi, and thus it could be speculated that PD-L1 engagement could contribute to this finding. LAG3 upregulation is taking place concomitantly to PD-L1 expression in both T cell subsets at day 7pi. LAG3 also belongs to the immune checkpoint family, and probably works synergistically with PD-1 to promote T cell exhaustion (88). LAG3 upregulation on CD4⁺ and CD8⁺ T could therefore also contribute to the immunosuppressed phenotype of T cells at this timepoint. It is noteworthy that IL10 expression is also upregulated in CD4⁺ T cells at day 7pi, and this could also participate in the immunosuppression since this cytokine is key in immune modulation (89). Taken together, our data indicate that BTV infection is triggering the activation of several immune regulatory processes.

Immune checkpoint proteins are typically triggered to resolve inflammation and limit tissue damage. Indeed, blockade of PD-L1 and depletion of Tregs in viral infections improves viral clearance, but leads to the development of immunopathology (90). Thus, BTV could be taking advantage of T cell regulatory loops that are in place to prevent tissue damage to expand its window of replication in the host. The activation of an inflammatory response resulting from the vascular damage produced by BTV infection in endothelial cells would need to be countered by potent inhibitors of inflammation such as IL-27. IL-27 promotes PD-1/PD-L1 and IL-10 expression (73), and we found that it was upregulated in the monocyte population throughout the infection. In this context, the virus could take advantage of the regulatory pathways that dampen inflammation to promote T cell immunosuppression at the peak of viral replication, for instance by inducing immune checkpoint protein expression such as PD-L1 and LAG-3 on T cells and by promoting IL-27 and PD-L1 expression in monocytes. Our data indicate that BTV is probably profiting from regulatory mechanisms to limit T cell responses and implicate PD-L1 expression on T cells and monocytes as a possible mechanism to suppress the host immune response.

Ultimately, our data show that the host mounts T cell responses capable of recognizing BTV by day 15pi. This coincided with the expansion of CD8⁺ T cells in the periphery, and the upregulation of IFNG in CD8⁺ cells in our transcriptomic analysis. Transcriptomic analysis of CD8⁺ cells also showed that no gene involved in TCR signaling were downregulated at day 15pi, suggesting that T cell activity has been normalized at this timepoint. Thus, in spite of the potent inhibitory effects of the infection at day 7pi on T cells, this immunosuppression appears to be transient and does not prevent the development of anti-BTV T cells. It remains unclear whether this immunosuppressive window has affected the quality of the T cell repertoire that responds to BTV. Multiple BTV infections reduce the spectrum of T cell epitopes recognized by the host (21), and thus BTV can affect its recognition by T cells. A more in depth analysis of the T cell phenotypes (effector/memory) generated upon BTV infections will need to be carried out to better evaluate the T cell activity during and after the infection.

Taken together, these data highlight the capacity of BTV to manipulate regulatory pathways in T cells, probably to limit its recognition and further its replication since cellular immunity is involved in disease protection (29–32). Further understanding of the interactions between the host immune system and BTV could help better comprehend the mechanisms that allow BTV to prolong its circulation in spite of the mounting immune response. Ultimately, these studies can provide veterinary medicine with better tools to combat the disease.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: GSE233994 (GEO).

The animal study was approved by Committee on the Ethics of Animal Experiments of the Spanish Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA) and the National Animal Welfare Committee. The study was conducted in accordance with the local legislation and institutional requirements.

AL-L: Investigation, Writing – original draft, Writing – review & editing, Data curation, Formal Analysis, Methodology. JR: Formal Analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing, Conceptualization, Supervision. IG-G: Investigation, Writing – review & editing, Data curation, Software. DR-M: Investigation, Writing - review & editing, Methodology. EM: Investigation, Writing - review & editing, Methodology. VM: Funding acquisition, Supervision, Writing – review & editing. NS: Funding acquisition, Supervision, Conceptualization, Investigation, Project administration, Writing – original draft, Writing – review & editing.