Phenotype of Recently Proliferating CD4 + T Cells in Blood of Acute Dengue Patients and its Dynamicity During the Course of Disease
To understand the phenotype and dynamics of CD4+ T cell response during the course of dengue, we enrolled a cohort of 170 adult acute dengue patients (Table S1). Blood samples were collected during acute phase (6.7\(\pm\)3.2 days of illness) until the discharge from the hospital. Some of the patients (n=32) were longitudinally followed during an extended convalescence phase (33.7\(\pm\)12.9 days post-symptom onset) (Fig. 1A). Dengue was confirmed by detecting DENV NS1-Antigen or immunoglobulin IgM in the patients plasma. Patients with non-confirmed dengue or with comorbidity were excluded from the study. Majority of the patients (~ 2/3) in the cohort experienced secondary DENV infection (Fig. 1B), as evident by the IgM to IgG ratio. We then analysed the memory (CD45RA–) CD4+ T cells in the peripheral blood mononuclear cells (PBMCs) from a sub-group of acute patients (n=21). We detected a sizeable recently proliferating cells expressing Ki67 (Fig. 1C and Fig. S1A). The proliferating cells showed the expression of several activation associated markers including ICOS, CD38 and HLA-DR (Fig. 1D and Fig. S1B). The Ki67+ proliferating cells showed an elevated expression of CXCR3 but not CCR6, suggesting their skewness towards Th1 phenotype. Further analysis revealed the existence of CXCR5 expressing Tfh cells as a minor subset (7.5\(\pm\)1.0%). To better understand the co-expression of these phenotypic markers, we performed non-linear dimensional reduction analysis using t-SNE on memory CD4+ T cells from patients showing heightened proliferation (n=9) (Fig. 1E and Fig. S1C). A distinguished cluster of activated cells was located in the bottom of tSNE plot co-expressing ICOS, CD38, PD-1 and CXCR3 but not CCR6.The CXCR5 showed a minimal occupancy within this activated cluster. Additionally, activated cluster had region of variable Ki67 expression, and a relatively small HLA-DR expressing region (Fig. 1E). To define activation phenotype, we thus used ICOS+Ki67+ double positive cells which include the recently proliferating and activated cells. Moreover, the combinatorial analysis confirmed that all the ICOS+Ki67+ cells are equally represented by ICOS+CD38+ cells but not by HLA-DR+CD38+ cells (Fig. S1D and S1E). These ICOS+Ki67+ activated cells had a heightened expression of CD38, PD-1 and CXCR3 (Fig. 1F) as compared to their resting memory counterparts (ICOS–Ki67–). Next we measured the dynamics of CD4+ T cell activation during the acute phase in primary and secondary dengue. We found a progressive increase in the activated T cells in primary dengue from 3.1% on 4.3\(\pm\)0.9 days to 11.8% on 7.4\(\pm\)1.4 days of illness (Fig. 1G). We also witnessed the similar increase in frequency of activated cells expressing PD-1 during the later phase of acute illness (Fig. 1I). Although secondary dengue showed a significant level of early T cell activation, the magnitude remains unchanged in the acute illness (Fig. 1G and 1I). We then followed a sub-group of patients to the convalescence phase to examine the dynamicity during recovery. A significant 4-fold (median:12.1–3.1%) decline in frequency of activated T cells was observed in convalescent dengue (Fig. 1H), which was also reflected by the decline in PD-1 expressing cells similar to the levels in healthy donors (Fig. 1J). Waning of T cell activation in convalescent dengue was consistent in both primary and secondary dengue (Fig. 1G and 1I). Altogether, acute dengue is associated with a robust activation and expansion of CD4+ T cells during a week after onset of illness in both primary and secondary disease. The expanded T cells in circulation majorly comprise of PD1-expressing activated cells that swiftly undergo a resolution phase where these cells seems to have contracted their size or migrated to the peripheral tissues.
Exorbitant Accumulation of CXCR5PD-1 CD4 T Cells in Dengue with Warning Signs and Severe Dengue
We next determined the frequency of subsets expressing PD-1 in co-occurrence with the peripheral Tfh (pTfh) cells (Fig. 2A). The frequency of pTfh cells (CXCR5+PD-1+) was significantly increased (HD: 1.4%, DENV: 3.6%, Fig. 2B) in dengue. In addition, peripheral CXCR5–PD-1+ subset was highly accumulated (HD: 11.7%, DENV: 24.7%, Fig. 2B). Interestingly, CXCR5–PD-1+ was the predominant subset in the activated T-cell pool (~ 75%, Fig. S2 A), which also showed a greater activation and expansion compared to pTfh cells (CXCR5+PD-1+: 15.9%,CXCR5–PD-1+: 32.3%, Fig. 2C). To determine the association of these two activated subsets in disease outcome, we classified the acute dengue patients into mild (Dengue Fever, DF without warning signs), moderate (DF with warning signs) and severe dengue (severe bleeding, leakage syndromes and organ involvement) according to WHO guidelines 30 (Table S1). Severe disease was more apparent in secondary dengue (~ 76%) with comparable moderate (~ 60%) and mild (~ 49%) dengue like primary infection (Fig. 2D). Severe dengue patients had prolonged illness than mild or moderate disease. The frequency of total pTfh and CXCR5–PD-1+ cells was higher in moderate dengue than mild patients. However, consistently higher magnitude of CXCR5–PD-1+ cells was retained in severe dengue, which was not the case with pTfh cells (Fig. 2E-G). Of note, the magnitude of these subsets was not different between primary and secondary dengue in any disease outcome (Fig. S2 B-C). Unlike pTfh cells that remains unchanged in different disease outcome (Fig. 2H and Fig. S2D), CXCR5–PD-1+cells showed strikingly higher activation in moderate and severe dengue than mild disease (Fig. 2I). In addition, CXCR5–PD-1+ cells were slightly more activated in mild and moderate secondary dengue than in primary, but remains at similar levels in severe outcome (Fig. S2E). Thus, CXCR5–PD-1+ cells with an activated phenotype were associated more with moderate and severe disease outcomes. Further to confirm that the CXCR5–PD-1+ cells do not contain the regulatory T (Treg) cells, we analysed CD25+FOXP3+ cells in dengue (Fig. S2F-G). Although the activated (FOXP3hi) Treg cells were frequently observed during the acute phase, with decline during recovery (Fig. S2F), the CXCR5–PD-1+ cells with increased PD-1 expression were mainly present in the non-Treg compartment (Fig. S2G). Collectively, the data provided the evidence for excessive expansion of the peripheral CXCR5–PD-1+CD4+ T-cell subset in pathologically progressive dengue with warning signs and severity.
CXCR5 – PD-1 + Subset Showed Enrichment in DENV-Specific CD4+ T Cells and Remain Detectable in Dengue Recovered Individuals
We further examined the antigen-specific T-cell pool in dengue to understand the dynamics and stability of above discussed T-cell subsets. We analysed DENV-specific T cells in PBMCs from dengue patients stimulated with the megapool of 180 different CD4 T-cell reactive peptides (DENVpep) derived from all four DENV (DENV1-4) serotypes, in Activation Induced Marker (AIM) assay. AIM analysis of a subset of acute patients (n = 44) found an average of 0.78% of DENV-specific cells in the total pool of memory CD4+ T cells (Fig. 3A-B). Antigen-independent stimulation with mitogen Concanavalin A (ConA) induced an average response of 2.59% cells (Fig. 3C). The DENV-specific T cells increased from 0.78–1.01%, corresponding to an average of 4 and 7 days of acute illness. Only 0.28% of DENV-specific T cells remained detectable in convalescent dengue around 34 days after symptom onset (Fig. 3D–3E). We next determined the enrichment of antigen-specific T cells in the CXCR5 and PD-1 expressing subsets (Fig. S3A). Interestingly, PD-1 expressing CXCR5+ and CXCR5– subsets showed more enrichment of antigen-specific cells than their PD-1-negative counterparts. Strikingly, higher enrichment of antigen-specific T cells was found in the CXCR5–PD-1+ subset than that of pTfh (CXCR5+PD-1+: 1.35%, CXCR5–PD-1+: 2.88%, Fig. S3A), which was not the case in ConA stimulation (Fig. S3B). Although both the subsets slightly reduced the PD-1 expression after recovery and followed similar dynamic pattern, the expansion of DENV-specific pool was more apparent in the CXCR5–PD-1+ subset (Fig. 3F-G). The reduction of activated phenotype (ICOS+PD-1+) from acute (71.0%) to convalescent (33.8%) dengue indicated a shift of antigen-specific T cells from an activated effector to a memory phenotype (Fig. S3C–S3D), which was also reflected in the Th1 (CXCR3+CCR6–) skewed phenotype (Fig. S3E–S3F). We further examined individuals recovered from dengue (~ 10 months after symptom onset) to investigate whether antigen-specific CD4+ T cells acquire memory. The stimulation with DENVpep revealed around 0.17% of antigen-specific circulating T cells (Fig. 3H), which comprised of CXCR5–PD-1+ subset marked with a Th1 memory phenotype (Fig. 3I–3J). To determine if this memory subset also exist in the lymphoid organs we examined the lymph node tissues of DENV-seropositive individuals from dengue endemic region, who were previously exposed to the DENV. Although total Tfh subsets, CXCR5+PD-1+ and GC-Tfh (CXCR5++PD-1++) were highly enriched in secondary lymphoid organ (Lymph Node), interestingly, we found almost similar frequency of total CXCR5–PD-1+ subset in paired LN tissue and blood (Fig. S3G-S3H). Antigen-specific stimulation with DENVpep detected DENV-specific CXCR5–PD-1+ cells in the LN tissues of these virus exposed individuals, but it existed in peripheral blood in a higher magnitude as compared to LN tissues (Fig. S3I-S3J).
Altogether, these findings clearly establish the expansion of DENV-specific CD4+ T cells in acute dengue that were enriched in CXCR5–PD-1+ subset. Although the frequency of CXCR5–PD-1+ was robustly contracted during recovery, it acquired a memory-like phenotype that persisted for a long time in blood and lymphoid tissues.
Activated CXCR5PD-1 CD4 T Cells Positively Associate with Expanded Plasmablast and Antibody response in Dengue
As the PD1-expressing subset showed the dominance and co-occurrence with the pTfh subset, we attempted to determine their association with antibody response in dengue. We thus measured the anti-NS1 and anti-prM/M/E IgG in acute and convalescent dengue plasma. Like T cells, primary dengue showed a gradual increase in the anti-NS1 IgG, while an early and significantly higher response was detected in secondary dengue, at all time-points (Fig. 4A and Fig. S4A). Similar results were observed for the anti-prM/M/E IgG titers (Fig. S4B). Anti-NS1 IgG continued to increase during recovery as observed in both cross-sectional (Fig. 4A) and longitudinal (Fig. 4B) analyses, which was not the case with anti-prM/M/E IgG (Fig. S4B–S4C), consistent with longer availability of NS1 antigen but not virion particles during disease progression 31. Notably, antibody response to both the antigens was significantly higher in the severe dengue than that of the moderate disease (Anti-NS1; moderate AUC: 0.26 x 106, severe AUC: 0.39 x 106, p = 0.04, Fig. 4C), (Anti-prM/M/E; moderate AUC: 0.74 x 106, severe AUC: 1.78 x 106, p = 0.03, Fig. 4D). CXCL13 play a critical role in T-B cross-talk and may indicate a productive germinal center (GC) response 32. We found significant levels of CXCL13 in both primary and secondary patients at the early onset (~ 4 days of illness) of acute disease (primary: 260.1 pg/mL, secondary: 270.0 pg/mL, Fig. S4D). A gradual and significant increase in CXCL13 levels was found only in primary dengue in the later stages (~ 7 days of illness) of disease (primary: 400.5 pg/mL, secondary: 192.2 pg/mL, p = 0.01, Fig. S4D), which declines to basal levels during recovery (Fig. S4D–S4E). Thus, CXCL13 levels were found to follow the same trend as CD4+ T cell responses. As an output of T-B cross-talk, we next measured the antibody producing B cells called plasmablast in the acute disease. We found a strong evidence of plasmablasts differentiation in acute patients (p < 0.0001, Fig. 4E and Fig. S4F). Most of these plasmablast expressed Ki67, CD27, CD71 and CXCR3 but down-regulated expression of CD19, CD20 indicating their effector phenotype (Fig. 4F–4G and Fig. S4F-S4G). To establish the biological relevance of the expanded CD4+ T-cell subsets in relation to humoral response, we performed correlation analysis (Fig. 4H). Unlike activated pTfh (CXCR5+PD-1+) that showed a positive association only with plasmablasts, the expanding activated CXCR5–PD-1+ T cells showed strong correlation with both antibodies (anti-NS1 and anti-prM/M/E IgG) as well as plasmablasts.
Taken together, the data provided an overall understanding of antibody and plasmablast responses during acute dengue. Certainly, these responses were more strongly associated with the CXCR5–PD-1+ subset indicating the role of this subset in regulating the humoral immunity in dengue.
Antigen Specific CXCR5 – PD-1 + CD4+ T Cells Robustly Secrete B-Cell Helper Cytokines and Capable of Promoting B-cell Responses via IL21 signalling axis
The strong association of the CXCR5–PD-1+ subset with in vivo antibody and plasmablast response intrigued for a detailed investigation of its B cell help potential. We first examined the production of CD40L and IFNγ in ex vivo stimulated PBMCs with DENVpep for a short time (~ 6 h) in ICS assay. DENVpep stimulation induced significant number of CD40L and IFNγ producing cells in memory CD4+ T cells (Fig. 5A). Average of 1.36% and 1.04% DENV-reactive IFNγ+ and CD40L+ cells were present in acute patients, respectively (Fig. 5B). DENV-reactive T cells co-expressing IFNγ and CD40L were present in significantly higher frequency (p = 0.0002, Fig. 5C). Moreover, majority of these cytokine-producers (CD40L, IFNγ and IL10) were enriched in T cells expressing PD-1 and CXCR3, indicating the presence of functionally potent DENV-specific Th1 effector cells (Fig. 5D). We next analysed the CXCR5+ and CXCR5– subsets in PD-1 expressing T cells for their cytokine producing capability and found that the CXCR5– cells contained significantly more DENV-specific cytokine-producers (Fig. 5E–5F). Even in TCR-independent stimulation the CXCR5– subset expressed significantly higher IFNγ than CXCR5+ cells, although they showed equal ability to express CD40L (Fig. 5E). We further confirmed the cytokine secretory function in the sort-purified CXCR5+ and CXCR5– subsets, which showed a similar finding like IFNγ-ICS analysis (Fig. S5A). Because IL21 cytokine is important for B-cell differentiation and function, we also examined IL21 secretion in these sort-purified subsets. ELISpot analysis showed that the CXCR5– subset can produce and secrete significant levels of IL21 comparable to the CXCR5+ subset upon stimulation with both antigen-specific and mitogen condition (DENVpep CXCR5+: 73 SFUs/million, DENVpep CXCR5–: 128 SFUs/million, Fig. S5B). To determine whether cytokine-producer CXCR5– subset can also engage B cells in a productive cross-talk, we utilized our recently described efficient autologous T-B co-culture assay 33. Here, the B cell helper activity of CXCR5– subset was measured in autologous cocultures with sort-purified B cells in DENV-specific settings. Activated T cells from acute disease showed a poor survival in long-term (up to 9 days) cocultures (Data not shown). Therefore, we used the memory CD4+ T cells from DENV-seropositive individuals with significant levels of DENV-reactive IgG titers (Fig. S5C). Naive T cells failed to elicit a B-cell response due to insufficient peptide-reactive T cells (Fig. 5G–5H). On the other hand, in DENVpep stimulation, both CXCR5+PD-1+ and CXCR5–PD-1+ cells strongly promoted the plasmablast differentiation. Unlike memory CXCR5+ cells where both PD1+/PD1− subsets elicited similar B-cell responses, only CXCR5–PD-1+ cells showed significantly higher B cell output than their PD-1– counterparts. Similarly, like pTfh, CXCR5–PD-1+ cells elicited detectable antibody responses in the coculture supernatant in the majority of donors tested (5/8 donors, Fig. S5D–S5E), against both the DENV NS1 and prM/M/E antigens. Notably, only memory B cells, but not naive, responded preferentially to B cell help provided by CXCR5–PD-1+ and pTfh cells (Fig. 5I and Fig. S5F). As CXCR5−PD-1+ cells secreted high levels of B-cell helper cytokines, we further attempted to identify the dominant cytokine regulating the help-axis of CXCR5−PD-1+ cells by blocking the activity of IL21, IL10 and IL4 cytokines in cocultures. In CXCR5−PD-1+ T cell cocultures, we observed a strong reduction in plasmablast output in IL21-blocking condition (~ 60%, Fig. 5J) and a moderate reduction by blocking IL10 (~ 25%), but not by inhibiting IL4. This finding was consistent with the help-axis utilized by pTfh cells (Fig. S5G).
Together, these data indicated that the DENV-specific T cells within CXCR5–PD-1+ subset were functionally effective in secreting the effector cytokines, including those required for B-cell help, and the cytokine-producing cells were largely enriched in PD-1 expressing subsets. Moreover, data also revealed that the CXCR5–PD-1+ subset is capable of productive collaboration with B cells in antigen-specific settings, which is mostly driven by IL21 signalling axis.
Single Cell RNAseq and TCRseq Analyses of activated CD4 T cells of acute dengue patients
With the clear indication of the B-cell help potential, we next explored the cellular heterogeneity and transcriptomic landscape of PD1 expressing CXCR5− subset in acutely ill dengue patients. We thus performed a single-cell transcriptome analyses of activated CD4+ T cell population (ICOS+CD38+), which were enriched for the majority of proliferating cells (Fig. 6A). After applying the quality control parameters of the analysis pipeline, a total of 4,361 activated T cells (pooled from 3 patients) were recovered for analyses (Fig. S6A-B). We integrated the datasets using the Seurat CCA-based method 34, and performed graph-based clustering after regression of confounding factors like number of UMIs, genes, percentage of mitochondrial genes, etc. to create a uniform manifold approximation and projection (UMAP) plot (Fig. 6B and Fig. S6C–D). UMAP analysis revealed 10 distinctly positioned clusters in the dengue activated CD4+ T cell population, reflecting their unique transcriptional signatures (Fig. 6B–6E). Clusters 0–6 consisted of the majority of cells, accounting for 95% of the total cells (Fig. 6C and Fig. S6E). Differential gene expression analysis revealed highly enriched transcripts in each cluster, as shown in Fig. 6D–6E and Table S2. The hierarchical ranking of clusters based on enriched transcripts identified cluster 9, 8 and 7 as very distinct rare clusters (Fig. 6D). Cluster 0 was enriched with regulatory T (TREG) cell gene signatures (IL2RA, FOXP3, RTKN2, and IKZF2) (Fig. 6E). Cluster 2 presented gene signatures of T central memory (TCM) cells, (TCF7, CCR7). This cluster was located nearest to cluster 8, which showed naive T (TN) cell like gene signatures, uniquely expressed FHIT and SNED1. Cluster 7 had an enriched signature of TH17 cells, expressed CCR6, CTSH, and KLRB1. Clusters 5, and 6 were highly enriched for effector TH1 cell gene signatures, LAG3, GZMA, GZMK, and PRF1, while the clusters 3, 1, and 4 mainly expressed transcripts associated with the cell cycle genes, MKI67, PCNA, MCM7 and TOP2A that was also confirmed by cell cycle phase analysis (Fig. 6E–6F). Cluster 9 that was located near to cluster 5, expressed the T effector memory (TEM) genes, (CCL4, CCL5) and uniquely upregulated expression of two long non-coding RNAs (lncRNA), FP671120.4 and FP236383.3.
Next, we identified the clusters expressing CXCR5, PDCD1, and CXCR3 transcripts (Fig. 6G–6H and Fig. S6F). We detected expression of PDCD1 transcripts in all the clusters, except clusters 0 (TREG), and 8 (TN), which accounted for ~ 77% of the total cells. Cells co-expressing CXCR5 or CXCR5 and PDCD1 (~ 15%) were mostly present in cluster 2 (TCM). Moreover, highly upregulated expression of CXCR3 was found in clusters 5, 6, 3, 1 and 4. These observations were further supported by expression of cell surface protein (csp) in these clusters (Fig. S6F). Interestingly, single cell trajectory analysis found PD-1 expressing clusters 5, 6, 3 as closely related, with a trajectory branching from proliferative cluster 3 to cluster 5 (TH1 effector) or cluster 6 (TH1 effector memory), suggesting two alternative cell fates (Fig. 6I). Cluster 6 appeared to be an intermediate population extending its trajectory into cluster 2 with less differentiated TCM cells. Furthermore, proliferative clusters 1 and 4 followed a separate trajectory leading to either cluster 0 (TREG) or cluster 5 (TH1 effector).
We then performed single-cell TCRseq analysis to inspect the clonal relationship and the expansion within these clusters. We found one TRAV13-1.TRBV2 hyperexpanded clonotype (945 clones) and four large clonotypes: TRAV29.TRBV11-2 (58 clones), TRAV12-1.TRBV12-4 (47 clones), TRAV21.TRBV5-1 (46 clones), and TRAV26-1.TRBV19 (44 clones) (Fig. S6G and Table S3). The majority of hyperexpanded and large/medium expanded clonotypes were present in clusters 5, 6, 3, 1 and 4, resulting in an average of at least 1.5 cells per clonotype observed in these clusters. On the other hand, cluster 0 (TREG) and cluster 2 (TCM) cells showed mostly the single/small expanded clonotypes (Fig. 6J–6K and Fig. S6H). Clonotype overlap analysis revealed a significant distribution of TCR clonotypes between the proliferative clusters 1, 4 and other major clusters, including cluster 0 (TREG) (Fig. 6L–6M). Notably, the PD-1 expressing clusters 5, 6, and 3 showed significant clonotype overlap with each other (Fig. S6I), but to a lesser extent with Tfh subset containing cluster 2.
Certainly, these data suggested a sizeable heterogeneity in activated CXCR5–PD-1+ T cells, which were mostly composed of proliferative and TH1 effector/ effector memory subsets with hyper- and large expandability and a significant clonal overlap. Moreover, the CXCR5–PD-1+ T cells in dengue seems to originate from a distinct repertoire than pTfh subset as apparent by the marginal clonal overlap between PD1-expressing clusters and the Tfh containing cluster.
Distinct features of B cell helper subset in CXCR5 – PD-1 + population with tissue migrating patterns and strong association with extrafollicular B cells in dengue
A higher extent of cellular heterogeneity was evident in the activated CXCR5–PD-1+ T cells. Therefore, to delineate the cell subset with B-cell helper signatures, we further inspected the gene signatures of PD-1-expressing clusters 5, 6, and 3 in more detail. Cluster 5 showed the highest expression of PDCD1 together with other activating and co-inhibitory genes - ICOS, CD69, CD44, CD38, CD70, TNFSF9, LAG3 and HAVCR2 (Fig. 7A and S7A). Expression of HLA-DRA, HLA-DRB1, IL2RA, TIGIT and ENTPD1 was limited to cluster 0 (TREG), but not other clusters. In addition, cluster 5 showed a strong enrichment of TH1 effector cytokines including IFNG, GZMA, PRF1, NKG7, GZMB, IL10, CCL4 and CCL5 (Fig. 7A–7B). Most of these cytokines confer cytotoxic functions and upregulated by type 1 interferon signalling. Therefore, we performed the gene set enrichment analysis (GSEA) of cytotoxic CD4 and interferon response signatures in these clusters. Cluster 5 showed the highest positive enrichment of both signatures (Fig. S7B–S7C) expressing TH1 transcriptional regulators which are known to control these functions, TBX21, STAT4, RUNX3, PRDM1, IKZF3, ID2, MAF and KLF6 (Fig. 7A). Notably, cells in clusters 5, 6 and 3 showed strong tissue migration markers, expressing various chemokine receptors and integrins (CCR2, CCR5, CXCR3, CXCR6, CX3CR1, ITGB1, ITGA4, ITGAE, S1PR1 and S1PR4). They also showed expression of adhesion molecules important for differentiation, SH2D1A, SLAMF1, SLAMF6. Besides, cluster 2 (TCM) strongly expressed the activation marker TNFSF8, the transcriptional regulators, TCF7, LEF1, TOX2, BCL6, ID2, and the migration genes CCR7, SELL, S1PR1, CXCR5 (Fig. 7A and S7A). Certainly, the data further confirmed that the cluster 5 was highly TH1-polarized with cytotoxic CD4 signatures, while the cluster 2 had a central memory (TCM) like signature enriched in Tfh-related genes.
Importantly, the proliferative PD1-expressing clusters, especially cluster 3, showed highest expression of IFNG, CD40LG and IL21 (Fig. 7B). As demonstrated above and as IL21-signaling was the major B-cell help axis utilized by the PD1-expressing subset, we identified IFNG+ cells expressing GZMB and IL21 transcripts to distinguish between TH1 cytotoxic and helper subset, respectively (Fig. 7C). Interestingly, we found that the GZMB+ cells mainly enriched in clusters 5 and 6 had no significant co-expression with IL21+ cells, but IL21+ cells enriched in proliferative clusters 3, 1 and 4 showed some co-expression with GZMB+ cells (Fig. S7D). The polyclonal activation with SEB antigen also confirmed the existence of two cell groups in memory CXCR5–PD-1+ T cells, either expressing IL21 or GZMB cytokines (Fig. S7E). We thus performed a differential gene expression analysis between PD-1-expressing IL21+ and GZMB+ cells after excluding the GZMB+IL21+ cells (Fig. 7D). Cytotoxic GZMB+ cells showing genes related to cytotoxic function, GZMA, GZMH, NKG7, PRF1, GNLY, appeared more terminally differentiated. Strikingly, these cells showed higher expression of HOPX transcription factor, which is induced by T-bet upon repeated antigenic stimulations and enriched in terminally differentiated effector Th1 cells 35. In addition, these cells upregulated CD44 transcripts, that may promote their survival as Th1 effector memory 36. Besides, helper IL21+ cells appeared to be more proliferative than GZMB+ cells as they expressed key cell cycle regulators CDK6 and CCND2 and higher transcripts of glycolytic pathway, PFKL and PKM. The IL21+ cells expressed mitochondrial chaperonins HSPD1 and HSPE1, and antioxidant genes MT1G, SOD1, PRDX1 that can support the metabolic adaptations and protection from oxidative stress. Interestingly, these IL21+ cells strongly expressed TOX2 transcription factor, which is mainly expressed by central memory (TCM) cells might be essential for the maintenance of these helper cells in central memory compartment. Importantly, LGMN is highly expressed in IL21+ helper cells which is a key asparaginyl endopeptidase (AEP) expressed in lysosomes. This enzyme has broad targets of substrates including cathepsins and α1-thymosin and can intrinsically promote Th1 activity in these helper cells 37. Notably, the IL21+ and GZMB+ cells displayed a distinct gene landscape as compared to the double negative or double positive cell types (Fig. 7E). However, these two cell groups, IL21+ and GZMB+ cells, expressed comparable levels of PDCD1, MKI67, CCR2 and CCR5 (Fig. 7F) suggesting that both contained CXCR5–PD-1+ proliferative cells and had similar tissue migrating patterns. We further examined the clonotype diversity and found that only 13 TCR clonotypes were shared between the two cell groups, with IL21+ - 13/63, and GZMB+ cells − 13/136 TCR clonotypes (Fig. 7G). The TCR overlap was observed only in the highly expandable clones. These data indicate the existence of pre-determined repertoire of multifunctional precursors that acquire either a helper or a cytotoxic features.
Circulating CXCR5–PD-1+ T cells showed strong signs of tissue migration, indicating an active inflammatory condition in peripheral tissues 38. This was evident from cytokine analysis data (Fig. S7F), which showed the expression of CXCR3 chemokine receptor ligands, CXCL9 (MIG), CXCL10 (IP-10), type 1 interferon, IFNα2, and various pro-inflammatory cytokines, LIF, IL1Rα, IL1α, IL2Rα and IFNγ. We thus hypothesized that these tissue migrating T cells would have interacted with B cells in extrafollicular sites, including peripheral tissues, and elicited the B cell responses. Therefore, we analysed the extrafollicular B cells in acute dengue patients marked with higher expression of CD11c but reduced CD21 expression 39. CD21–CD11c+ B cells in the blood of acute patients consisted mainly of IgD–CD27– (double negative, DN) and IgD+CD27– (naive) B cells, followed by IgD+CD27– (switched memory) B cells (Fig. S7G). Although the extrafollicular B cells showed a higher enrichment in DN cells than naive B cells (Fig. 7H). The CD21–CD11c+ DN B cells showed increased CD19 expression but reduced CXCR5 expression, maybe to migrate away from B cell follicles (Fig. 7I). Remarkably, these extrafollicular (CD21–CD11c+ DN) B cells showed a strong positive association with activated CXCR5–PD-1+ T cells, but not CXCR5+PD-1+ (pTfh) cells (Fig. 7J), suggesting their coordinated expansion and differentiation in acute dengue.
Collectively, our single cell analysis clearly established substantial heterogeneity in CXCR5–PD-1+ T cells, which majorly consist of specialized subsets with cytotoxic or helper functions. Notably, this data is the first evidence for the presence of extrafollicular B cells in dengue, which seem to be induced by pathogenically accumulating peripheral B-cell helper T-cell subset.