Attenuated effector T cells are linked to control of chronic HBV infection

Heim, Kathrin; Sagar; Sogukpinar, Özlem; Llewellyn-Lacey, Sian; Price, David A.; Emmerich, Florian; Kraft, Anke R. M.; Cornberg, Markus; Kielbassa, Sophie; Knolle, Percy; Wohlleber, Dirk; Bengsch, Bertram; Boettler, Tobias; Neumann-Haefelin, Christoph; Thimme, Robert; Hofmann, Maike

doi:10.1038/s41590-024-01928-4

Download PDF

Article
Open access
Published: 28 August 2024

Attenuated effector T cells are linked to control of chronic HBV infection

Nature Immunology volume 25, pages 1650–1662 (2024)Cite this article

4247 Accesses
2 Citations
19 Altmetric
Metrics details

Subjects

Abstract

Hepatitis B virus (HBV)-specific CD8⁺ T cells play a dominant role during acute-resolving HBV infection but are functionally impaired during chronic HBV infection in humans. These functional deficits have been linked with metabolic and phenotypic heterogeneity, but it has remained unclear to what extent different subsets of HBV-specific CD8⁺ T cells still suppress viral replication. We addressed this issue by deep profiling, functional testing and perturbation of HBV-specific CD8⁺ T cells during different phases of chronic HBV infection. Our data revealed a mechanism of effector CD8⁺ T cell attenuation that emerges alongside classical CD8⁺ T cell exhaustion. Attenuated HBV-specific CD8⁺ T cells were characterized by cytotoxic properties and a dampened effector differentiation program, determined by antigen recognition and TGFβ signaling, and were associated with viral control during chronic HBV infection. These observations identify a distinct subset of CD8⁺ T cells linked with immune efficacy in the context of a chronic human viral infection with immunotherapeutic potential.

Identification of human progenitors of exhausted CD8⁺ T cells associated with elevated IFN-γ response in early phase of viral infection

Article Open access 07 December 2022

Memory-like HCV-specific CD8⁺ T cells retain a molecular scar after cure of chronic HCV infection

Article 04 January 2021

Autophagy-enhancing ATG16L1 polymorphism is associated with improved clinical outcome and T-cell immunity in chronic HIV-1 infection

Article Open access 28 March 2024

Main

Chronic HBV infection is a highly relevant disease with approximately 257 million infected people worldwide resulting in 887,000 deaths per year due to the development of progressive liver disease and hepatocellular carcinoma¹. Clinically, chronic HBV infection seems quite variable in different clinical phases. The clinical phases of chronic HBV infection can be principally divided into two categories: (1) chronic hepatitis B (CHB) characterized by the presence of liver pathology and (2) chronic HBV infection without signs of liver injury (endogenous control). The World Health Organization and clinical guidelines recommend that patients who are chronically HBV-infected with high viral loads (VLs) and/or liver injury receive nucleos(t)ide analog (NUC) treatment to abrogate viral replication and thus immediately reduce viral burden^1,2; however, with NUC treatment a cure is not achieved and therefore long-term therapy is required. Hence, chronic HBV infection with a low VL can either be endogenously controlled or mediated by NUC treatment (NUC control).

HBV-specific CD8⁺ T cells play a dominant role in the outcome of HBV infection and have been linked to both viral control and disease pathogenesis during acute infection^3,4. During chronic HBV infection, altered HBV-specific T cell responses have been reported in the different clinical phases. They are characterized by low abundance, impaired cytokine production, limited expansion capacity, high TOX expression, low T-BET expression and an altered metabolism, which are all characteristics of exhausted T (T_EX) cells ^{5,6,7,8,9,10,11}. Of note, T cell exhaustion represents a distinct T cell differentiation program of functional adaptation to persisting antigen recognition. The molecular mechanisms underlying T cell exhaustion have been mainly established in mouse models of chronic lymphocytic choriomeningitis virus (LCMV) infection^12,13; however, many features of T cell exhaustion could be translated to the human system, for example expression of inhibitory receptors such as PD1 and a transcriptional program reflected by high TOX and EOMES but low T-BET expression, as shown for virus-specific CD8⁺T cells in chronic HBV and hepatitis C virus (HCV) infection^11,14,15. Still, it is not fully elucidated whether the T cell exhaustion program covers all facets of T cell responses in chronic viral infections in humans.

In mice and humans, there is also growing evidence that exhausted T cells do not represent a uniform but rather heterogeneous population consisting of several subsets within two main groups: PD1^int progenitor/precursor/memory-like subsets and PD1^hi terminal exhausted subsets¹⁶. Progenitor/precursor/memory-like subsets are characterized by the transcription factor TCF1 and give rise to the terminal exhausted T cells through antigen-driven proliferation¹⁷. In humans, especially HBV-specific CD8⁺ T cells have also been shown to be heterogeneous with respect to their phenotypic, functional and metabolic characteristics^{18,19,20,21,22}. One determinant of HBV-specific CD8⁺ T cell heterogeneity is the targeted antigen with HBV envelope-specific CD8⁺ T cells being hardly detectable in comparison to HBV core- and HBV polymerase-specific CD8⁺ T cells that are more abundant in the blood^18,19. Additionally, HBV core- and HBV polymerase-specific CD8⁺ T cells differ in phenotype and function^18,19. The relevance of T cell heterogeneity has been underpinned by an association of certain HBV-specific CD8⁺ T cell responses with successful therapy²³; however, the contribution of HBV-specific T cell heterogeneity to the clinical course and thus a distinct clinical phase of chronic infection has not been analyzed in detail. Here, by using single-cell RNA sequencing (scRNA-seq) we analyzed the molecular heterogeneity of HBV-specific CD8⁺ T cells in different clinical phases. We identified a program of T cell attenuation with its TGFβ-involving mechanistic underpinnings in parallel to classical T cell exhaustion in a chronic human viral infection. T cell attenuation was linked to endogenous viral control highlighting the clinical relevance of different facets of HBV-specific T cell programming in chronic viral infection.

Results

HBV-specific CD8⁺ T cell heterogeneity

First, to analyze the transcriptional heterogeneity of HBV-specific CD8⁺ T cells obtained from patients who are chronically HBV-infected (hepatitis B e-antigen negative (HBeAg−) disease phases/treated and nontreated, low VL; n = 14) in an unbiased approach, we performed single-cell RNA sequencing (scRNA-seq) analysis using the mCEL-seq2 protocol²⁴ (Extended Data Fig. 1a,b). As a reference population of human T_EX cells, we included HCV-specific CD8⁺ T cells obtained from patients with HCV infection, either chronically infected (n = 6) or cured by directly acting antivirals (DAAs) (n = 2)²⁵. Cluster analysis applying the Seurat and Harmony algorithm to single-cell transcriptomes^26,27 of 1,954 HBV-specific and 818 HCV-specific CD8⁺ T cells revealed the existence of five different cell clusters (Fig. 1a). Differentially expressed genes (DEGs) comparing these five clusters included IL7R, TCF7, GZMB, TGFBR3, NR4A2, CD38, EZH2 and TOX thus covering cytokine responsiveness, epigenetic remodeling, activation, proliferative and functional characteristics (Fig. 1b). Specifically, cluster 2 was characterized by a memory-like signature, whereas cluster 4 exhibited a profile of terminal T cell exhaustion as also confirmed by gene set enrichment analysis (GSEA) applying gene signatures of T_EX, memory T cells and memory-like T cells obtained from the LCMV mouse model¹⁷ (Fig. 1b,c). Additionally, cluster 1 showed a gene signature that classifies cytotoxic effector cells (GZMB, GNLY, PRF1, KLRG1, CX3CR1, TBX21 and ZEB2) distinct from the recently described STAT5-dependent, KLRC/B1-expressing effector-like T_EX cells^28,29 (Fig. 1b and Extended Data Fig. 1c). Clusters 0 and 3 displayed overlaps with effector/memory signatures (ID2, ID3, TNF and ZEB1) according to the human T cell differentiation Atlas³⁰ (Fig. 1d). Next, we linked these clusters to CD8⁺ T cells targeting HBV versus HCV and observed clearly differential cluster distribution with a unique pattern of heterogeneity of HBV-specific CD8⁺ T cells (Fig. 1e). In particular, cluster 4 was strictly dominated by HCV-specific CD8⁺ T_EX cells obtained from the chronic phase of HCV infection (Fig. 1f). HBV- and HCV-specific CD8⁺ T cells contributed nearly equally to clusters 2 and 3 reflecting memory-like profiles as observed in HCV infection after DAA-mediated cure (Fig. 1f); however, HBV-specific CD8⁺ T cells additionally dominate the cytotoxic cluster 1 and cluster 0 with hardly any HCV-specific CD8⁺ T cell found within these two clusters. Hence, HBV-specific CD8⁺ T cell heterogeneity exceeds HCV-specific CD8⁺ T cell heterogeneity in chronic infection by the presence of cell clusters distinct from terminal T_EX and memory-like T_ML cells.

**Fig. 1: Subset diversification of virus-specific CD8⁺ T cells in chronic viral hepatitis.**

Cytotoxic CD8⁺ T cells target HBV pol in chronic infection

We and others have previously shown that the targeted antigen is a major determinant of HBV-specific CD8⁺ T cell phenotype and function^18,19,21. To investigate the impact of the targeted antigen on the transcriptional heterogeneity and subset diversification, we performed clustering analysis to single-cell transcriptomes of 1,056 HBV core₁₈-specific (n = 7 patients who were chronically HBV-infected) and 1,284 HBV pol₄₅₅-specific CD8⁺ T cells (n = 7 patients who were chronically HBV-infected). We identified four clusters (Fig. 2a) that again revealed differential expression of, for example, NR4A2, GZMB, TGFBR3 and IL7R and thus cover memory and cytotoxic characteristics and different cytokine responsiveness. Cluster 1 was enriched in genes indicative of cytotoxicity, whereas cluster 2 showed an overlapping effector/memory gene signature and clusters 3 and 0 were enriched for signature genes of memory-like/memory T cells, respectively (Fig. 2b). Clusters 3, 2 and 0 were rather equally composed of HBV core₁₈- and HBV pol₄₅₅-specific CD8⁺ T cells; however, HBV pol₄₅₅-specific CD8⁺ T cells dominated the cytotoxic cluster 1 (Fig. 2c). In addition to the cytotoxic signature²⁵, cluster 1 also revealed an enrichment of genes involved in T cell cytokine production (Fig. 2d) complementing the effector cell potential of this HBV pol₄₅₅-specific CD8⁺ T cell subset. Of note, single cytokines were only lowly expressed at the transcript level in our scRNA-seq dataset of virus-specific CD8⁺ T cells (Extended Data Fig. 1d). Different subset diversification of HBV core₁₈- versus HBV pol₄₅₅-specific CD8⁺ T cells was confirmed at the protein level after concatenating flow cytometric expression analysis of GZMB, PRF1, KLRD1, NKG7, TOX, PD1, T-BET, EOMES and CD127 of all HBV-specific CD8⁺ T cells and dimensionality reduction using t-distributed stochastic neighbor embedding (t-SNE) (Fig. 2e). Again, a large fraction of HBV pol₄₅₅-specific CD8⁺ T cells expressed marker molecules of cytotoxicity and effector cell differentiation. Of note, protein expression analysis also revealed a marked expression of CD127 required for T cell maintenance in these cells (Fig. 2e). Taken together, these results identified a cytotoxic HBV pol₄₅₅-specific CD8⁺ T cell subset during chronic infection with typical features of effector cells devoid of hallmarks of T cell exhaustion.

**Fig. 2: Distinct subsets of HBV core₁₈- and HBV pol₄₅₅-specific CD8⁺ T cells.**

Cytotoxic CD8⁺ T cells are linked to endogenous HBV control

To further investigate the biological relevance of the identified cytotoxic CD8⁺ T cell population in chronic HBV infection, we compared the transcriptional profile of HBV pol₄₅₅-specific CD8⁺ T cells in patients with chronic HBV infection who endogenously control the virus (HBeAg− cHBV infection) and do not require antiviral therapy versus patients with CHB requiring antiviral NUC therapy (NUC-treated HBeAg− CHB/; NUC control). Cluster analysis to single transcriptomes of 595 HBV pol₄₅₅-specific CD8⁺ T cells obtained from three patients with endogenous control and of 689 HBV pol₄₅₅-specific CD8⁺ T cells obtained from four patients with NUC control revealed the existence of three different cell clusters, including a prominent cytotoxic cluster 1 characterized by GNLY, GZMB, PRF1 and NKG7 expression (Fig. 3a). These HBV pol₄₅₅-specific CD8⁺ T cells with a cytotoxic signature were strongly associated with endogenous control (Fig. 3b). In contrast, clusters 0 and 2 primarily consisted of HBV pol₄₅₅-specific CD8⁺ T cells obtained from NUC control patients (Fig. 3b), expressing memory markers such TCF1 and interleukin (IL)-7R (Extended Data Fig. 2a). To confirm that the absence of the cytotoxic signature in HBV pol₄₅₅-specific CD8⁺ T cells obtained from NUC-treated patients is indeed linked to the absence of viral control before therapy, we flow cytometrically analyzed HBV pol₄₅₅-specific CD8⁺ T cells from patients with HBeAg− CHB and high VLs before initiation of NUC therapy (high VL/chronic HBV). Similar to NUC-treated patients and in contrast to patients with endogenous control, HBV pol₄₅₅-specific CD8⁺ T cells obtained from patients before NUC treatment exhibited reduced expression of GZMB, PRF, NKG7, KLRD1 and KLRG1 (Fig. 3c) as it was also the case for HBV core₁₈-specific CD8⁺ T cells in all tested clinical phases of chronic HBV infection (Extended Data Fig. 2b). It is important to note that this association of cytotoxicity and endogenous control is protein (polymerase/pol)- and not epitope-specific as a similar differential expression of cytotoxic molecules was also present in HBV pol₁₇₃-specific CD8⁺ T cells (Fig. 3c). In addition, we conducted diffusion pseudotime (DPT) analysis³¹ of concatenated flow cytometric data, including CD127, PD1, TOX, GZMB, PRF1, KLRD1, NKG7, ZEB2, T-BET and EOMES expression of HBV pol_455/173-specific CD8⁺ T cells from HBeAg− CHB, NUC control and endogenous control patients (Fig. 3d and Extended Data Fig. 2c). The subsets were residing on a two-dimensional manifold with two branches reflecting diverging trajectories in endogenous control versus CHB/NUC control phases of chronic HBV infection (Fig. 3d). Next, to gain further insights into the differential transcriptional regulation of the HBV pol₄₅₅-specific CD8⁺ T cell response in endogenous versus NUC control patients, we performed DPT analysis of the scRNA-seq data. Again, HBV pol₄₅₅-specific CD8⁺ T cells allocated in a two-dimensional manifold with two branches separating subsets obtained from endogenous versus NUC control patients (Fig. 3e). Gene regulatory network inference using SCENIC³² revealed that cytotoxic HBV pol₄₅₅-specific CD8⁺ T cells from endogenous control patients were characterized by TBX21, BDP1, ZNF246, POLR2A, ATF4 and CREM activity, whereas HBV pol₄₅₅-specific CD8⁺ T cells from NUC control patients exhibited higher activity of EGR1, KLF2 and AP-1 transcription factors (Fig. 3f and Extended Data Fig. 2d). These data indicate a memory-like/progenitor exhausted state of circulating HBV pol₄₅₅-specific CD8⁺ T cells in NUC control patients. In contrast, cytotoxic HBV pol₄₅₅-specific CD8⁺ T cells showed a signature of transcriptionally active effector T cells. Hence, the selective presence of cytotoxic HBV pol₄₅₅-specific CD8⁺ T cells in patients with endogenous control strongly supports a biological relevant link between this subset and immune-mediated HBV control in chronic infection.

**Fig. 3: Cytotoxic HBV pol₄₅₅-specific CD8⁺ T cells are associated with endogenous control in chronic HBV infection.**

Attenuated cytotoxic CD8⁺ T cell response in the chronic HBV phase

The presence of a clinically relevant cytotoxic CD8⁺ T cell subset during endogenously controlled chronic HBV infection prompted us to investigate whether this HBV-specific T cell response is similar to an effector T cell response in acute-resolving infection. We compared the HBV pol₄₅₅-specific CD8⁺ T cell response in 7 acutely HBV-infected versus 26 chronically infected, endogenously controlled patients versus 5 patients who spontaneously resolved an acute HBV infection. The frequency (Fig. 4a) and activation status as reflected by higher Ki67 and CD38 expression (Fig. 4b) of HBV pol₄₅₅-specific CD8⁺ T cells was higher in patients with acute HBV versus endogenously controlled chronic infection; however, the HBV pol₄₅₅-specific CD8⁺ T cell response in patients with endogenous control was marked by an increase in CD127-expressing cells compared to the bona fide virus-specific T cell effector response during the acute phase of HBV infection (Fig. 4c). To compare the functional effector capacity, we first compared the protein expression of cytotoxic molecules in HBV pol₄₅₅-specific CD8⁺ T cells in acute versus endogenously controlled chronic HBV infection. We did not observe a differential relative expression of the cytotoxic molecules GZMB and PRF1 or the cytolytic regulator NKG7 (Fig. 4d) in contrast to HBV core₁₈-specific CD8⁺ T cells obtained from acute compared to endogenous control/chronic HBV infection (Extended Data Fig. 3a). Next, we tested the ex vivo capacity to produce interferon (IFN)γ, tumor necrosis factor (TNF) or to degranulate in response to PMA/ionomycin stimulation (Fig. 4e). Again, no significant differences were evident comparing HBV pol₄₅₅-specific CD8⁺ T cells from endogenous control versus acute infection. Additionally, the effector cell markers PD1 and KLRG1 were also not different (Fig. 4f). Hence, while the strength of the HBV pol₄₅₅-specific CD8⁺ T cell response seems to be attenuated in patients with endogenous control these cells retain the capacity to elicit antiviral effector functions similar to bona fide effector CD8⁺ T cells. This attenuated HBV pol₄₅₅-specific CD8⁺ T cell response is clearly different from a bona fide HBV pol₄₅₅-specific memory CD8⁺ T cell response in patients who spontaneously resolved an acute HBV infection (Extended Data Fig. 3a,b), for example with respect to frequency (Fig. 4a), GZMB, PRF1, KLRG1 and PD1 expression (Fig. 4d,f).

**Fig. 4: Attenuated cytotoxic HBV pol₄₅₅-specific CD8⁺ T cell response.**

Transcriptional regulation of CD8⁺ T cell attenuation

To investigate the transcriptional regulation and fate of the attenuated cytotoxic HBV pol₄₅₅-specific CD8⁺ T cell response in chronic HBV infection, we performed comparative scRNA-seq of HBV pol₄₅₅-specific CD8⁺ T cells obtained from patients with endogenous control (n = 3; 1,921 cells), acute infection (n = 3; 1,900 cells) and resolution of acute infection (n = 2; 240 cells) using the 10x Genomics platform. Clustering analysis of integrated single-cell transcriptomes of HBV pol₄₅₅-specific CD8⁺ T cells identified five cell clusters (Fig. 5a). Of note, HBV pol₄₅₅-specific CD8⁺ T cell cluster identification from scRNA-seq data obtained with the 10x Genomics platform was consistent with the mCEL-seq2 protocol (Extended Data Fig. 4a,b). DEGs comparing these five clusters included NKG7, CX3CR1, KLRG1, PRF1, GZMB, PTMA, IL7R and the transcription factors TBX21, RUNX3, NR4A2, MYBL1, ZNF683, HMGB1 and TCF7 (Fig. 5b). More precisely, HBV pol₄₅₅-specific CD8⁺ T cells from patients with endogenous control could be found in all five clusters with the majority being in cluster 1 that is characterized by cytotoxic effector-like genes (Fig. 5a,b). In contrast, HBV pol₄₅₅-specific CD8⁺ T cells obtained from patients who were acutely HBV-infected were found in clusters 0, 2, 3 and 4 almost completely omitting cluster 1 (Fig. 5a,b) clearly showing different transcriptional mapping of effector/attenuated cells in acute versus chronic endogenous control HBV infection. After viral resolution, clusters 3 and 4 dominated, probably reflecting T cell memory subsets (Fig. 5a,b). Distinct expression patterns of transcriptional regulators in HBV pol₄₅₅-specific CD8⁺ T cells from endogenously controlled/chronic, acute and acute resolved phases of HBV infection were confirmed on the protein level (Fig. 5c and Extended Data Fig. 4c). These differences in ZEB2, T-BET, EOMES, RUNX3 and BLIMP1 expression point toward distinct T cell differentiation programs in acute versus chronic HBV infection irrespective of different VLs (Extended Data Fig. 4d,e). Next, to directly compare effector/effector-like subsets of HBV pol₄₅₅-specific CD8⁺ T cells obtained from acute-resolving HBV infection versus chronic endogenous control, we compared GZMB ⁺ HBV pol₄₅₅-specific CD8⁺ T cells as surrogates. Calculated frequencies of GZMB⁺ HBV pol₄₅₅-specific CD8⁺ T cells were highest during the acute phase, followed by endogenous control and were only rare in the steady state of acutely resolved HBV infection (Fig. 5d). Furthermore, t-SNE analysis of combined PRF1, NKG7, KLRD1, PD1, CD127, ZEB2, T-BET, EOMES and TOX expression (Fig. 5e and Extended Data Fig. 5a) confirmed topographic separation based on distinct transcription factor patterns. Lower ZEB2, T-BET, EOMES (Fig. 5c) and higher TCF1 expression (Extended Data Fig. 5b) also indicated effector T cell attenuation in endogenous control. In addition to analyzing the transcription factor expression, we also assessed transcription factor activity using SCENIC analysis (Fig. 5f and Extended Data Fig. 5c). Differential transcriptional activity of bona fide effector T cells (cluster 0) in patients who were acutely HBV-infected compared to attenuated T cells (cluster 1) obtained from endogenously controlled/chronically HBV-infected patients further highlighted distinct transcriptional programs (Fig. 5f and Extended Data Fig. 5c). In particular, increased activity of STAT1, TBX21 and IRF9 was detectable in effector T cells (cluster 0), whereas KLF6, FOS, JUN and MYBL1 were highly active in attenuated T cells (cluster 1), highlighting that bona fide effector T cells and attenuated T cells not only differ in transcription factor expression but also in their activity. Taken together, attenuated HBV pol₄₅₅-specific CD8⁺ T cells exhibit a unique transcriptional program compared to classical effector T cells but are still linked to endogenous viral control in chronic HBV infection.

**Fig. 5: Unique transcriptional regulation of attenuated HBV pol₄₅₅-specific CD8⁺ T cells.**

TGFβ regulates CD8⁺ T cell attenuation

Next, we set out to address the underlying mechanisms of HBV pol₄₅₅-specific CD8⁺ T cell attenuation in endogenous control during chronic HBV infection. We focused on TGFβ as it has previously been shown in the mouse model of LCMV infection³³ to lead to dampening of effector T cell differentiation, including downregulation of ZEB2 as observed in our analysis (Fig. 5c). In addition, among the top 30 DEGs, we also detected increased TGFBR3 expression specifically in attenuated CD8⁺ T cells targeting HBV pol₄₅₅ in chronic HBV/endogenously controlled infection and not in HBV core₁₈-specific CD8⁺ T cells (Figs. 1b and 2b and Extended Data Fig. 6a). Notably, increased TGFBR3 and TGFBR1 expression levels were also evident comparing attenuated with bona fide effector or memory HBV pol₄₅₅-specific CD8⁺ T cells (Fig. 6a) and in a previously published dataset³⁴ of GZMB⁺CD8⁺ T cells (Extended Data Fig. 6b) sampled from the liver of chronically HBV-infected patients with endogenous control by fine-needle aspiration (FNA). These observations support the hypothesis of increased TGFβ responsiveness in CD8⁺ T cell attenuation. In line with this, we observed a significant enrichment of TGFβ and not IL-10 signaling pathway genes in HBV pol₄₅₅-specific CD8⁺ T cells in endogenous control/chronic versus acute and acute resolved HBV infection (Fig. 6b and Extended Data Fig. 6c). These results argue against a general responsiveness of attenuated HBV pol₄₅₅-specific CD8⁺ T cells toward suppressive cytokines and rather support a distinct role for TGFβ signaling. To further determine the direct impact of TGFβ signaling on the effector T cell program in the context of HBV infection, we performed peptide-specific expansion of HBV pol₄₅₅-specific CD8⁺ T cells in the presence or absence of the TGFβ RI kinase inhibitor II CAS 446859-33-2 (Fig. 6c). Blocking of TGFβ signaling led to an increase in the expression of ZEB2, T-BET, EOMES, GZMB and PRF1 in HBV pol₄₅₅-specific CD8⁺ T cells of patients with endogenous control, whereas ZEB2 and T-BET expression remained stable in HBV pol₄₅₅-specific CD8⁺ T cells obtained from patients acutely infected with HBV (Fig. 6d and Extended Data Fig. 6d,e). Accordingly, inhibition of TGFβ signaling resulted in an increase of HBV pol₄₅₅-specific IFNγ and TNF production of CD8⁺ T cells in all but one tested patient with endogenous control, along with a significant boost of IFNγ and TNF co-producing cells in all tested patients and an elevated HBV pol₄₅₅-specific degranulation (indicated by CD107a expression) of CD8⁺ T cells in three out of six tested patients (Extended Data Fig. 7). These data indicate an augmented function of HBV pol₄₅₅-specific CD8⁺ T cells after TGFβ inhibition. Moreover, we also detected higher TGFβ concentrations in the plasma of patients with endogenous control/chronic (n = 21) compared to acute (n = 7) and acute resolved (n = 6) HBV infection of whom we have also analyzed the HBV pol₄₅₅-specific CD8⁺ T cell responses previously (Extended Data Fig. 6f).

**Fig. 6: TGFβ signaling is linked to HBV-specific CD8⁺ T cell attenuation.**

Next, to test which factor may also impact on HBV pol₄₅₅-specific CD8⁺ T cell attenuation and may be used to boost or even release these cells, we conducted GSEA of the IL-2, IL-12 and IFNα signaling pathways and performed peptide-specific expansion of HBV pol₄₅₅-specific CD8⁺ T cells this time in the presence of either IL-12 or IFNα. Based on expression of cytokine signaling pathway genes, we observed an enrichment of the IL-12 pathway in attenuated HBV pol₄₅₅-specific CD8⁺ T cells versus bona fide effector and memory subsets, but a decrease in IFNα and IL-2 pathway gene expression compared to acute and acutely resolved HBV infection (Extended Data Fig. 8a). Stimulation with IL-12 resulted in increased expression of T-BET, EOMES and perforin (Fig. 6e,f and Extended Data Fig. 8b), whereas stimulation with IFNα only led to an augmented T-BET expression (Fig. 6e,g and Extended Data Fig. 8c). Given these different effects of cytokine interventions on the T cell program, we next wondered which treatment redirected/boosted attenuated HBV pol₄₅₅-specific CD8⁺ T cells most similar toward bona fide effector cells detectable during acute HBV. Dimensional reduction analyses by t-SNE and multidimensional scaling (MDS) revealed that TGFβ blockade directed attenuated HBV pol₄₅₅-specific CD8⁺ T cells most efficiently toward bona fide effector cells accompanied by the largest boost in expansion (Fig. 6h and Extended Data Fig. 8d). Hence, while IL-12 and IFNα can partly redirect attenuated HBV pol₄₅₅-specific CD8⁺ T cells, TGFβ signaling seems to be a key regulator of HBV-specific CD8⁺ T cell attenuation present in the endogenous control phase of chronic HBV infection.

Discussion

Here, we report the identification of an attenuated cytotoxic HBV-specific CD8⁺ T cell population that is present during chronic human viral infection and that is linked to endogenous viral control. Attenuated HBV-specific CD8⁺ T cells exhibited similar expression of effector cell markers such as PD1, KLRG1, CX3CR1, NKG7 and of cytotoxic molecules and cytokine production capacities when compared to effector HBV-specific CD8⁺ T cells obtained from acute-resolving infection underlining their potent effector functions; however, concomitantly, they displayed higher levels of markers for T cell maintenance, such as CD127 and TCF1 and lower expression of transcriptional regulators of the cytotoxic effector T cell program, such as EOMES and ZEB2 (refs. ^28,35,36), indicating an adapted attenuated T cell differentiation program in chronic infection. Still, despite the presence of chronic infection, the attenuated HBV-specific CD8⁺ T cell population does not show typical features of T cell exhaustion. Indeed, scRNA-seq analysis revealed unique clusters of cells that did not overlap with the molecular core signature underlying exhausted CD8⁺ T_EX cells present during chronic human HCV²⁵ and LCMV¹⁷ infection including the STAT5-dependent effector-like T_EX cells^28,29,37. Accordingly, they followed a trajectory that separates from effector and exhausted cells and are clearly distinct from a memory T cell response supporting a unique program and fate of this cell subset in chronic HBV infection.

Key determinants of T cell tuning, adaptation and, thus, fate are antigen recognition dynamics (for example, dose and duration) as well as co-stimulation and cytokine milieu^{38,39,40,41,42,43,44}. Attenuated HBV-specific CD8⁺ T cells were mainly targeting the HBV polymerase antigen and hardly any HBV core antigen, highlighting the impact of specific antigen recognition in the induction of the attenuated T cell fate. Of note, HBV polymerase and HBV core exhibit a mosaic pattern of antigen presentation in the infected liver with lower quantities of HBV polymerase compared to core⁴⁵. In addition to antigen recognition, T cell attenuation is also shaped by a TGFβ cytokine milieu, with ZEB2 being a relevant rheostat. In particular, we observed significantly higher TGFβ plasma levels in patients who were chronically HBV-infected (in line with findings by others⁴⁶) and an increase in the expression of ZEB2 in HBV pol₄₅₅-specific CD8⁺ T cells after blockade of TGFβ. This is in line with previous findings in the LCMV mouse model showing that ZEB2 is selectively inhibited by TGFβ³³ and that TGFβ signaling as well as ZEB2 deletion suppress the effector-associated program of CD8⁺ T cells⁴⁷. Notably, HBV-specific T cell attenuation could only be partly overcome by co-stimulatory cytokine stimulations, including IL-2, IL-12 and type I IFNs with no effect on ZEB2 expression. Combined, these results support an important role of antigen recognition and TGFβ signaling linked to ZEB2 expression in HBV-specific CD8⁺ T cell attenuation in chronic infection. Whether TGFβ inhibition represents a feasible treatment option to boost the HBV-specific CD8⁺ T cell response in patients chronically infected with HBV requires further investigation. The clinical application of TGFβ inhibition remains challenging due to pleiotropic effects and adverse events. Still, therapeutic potential of TGFβ inhibition has been and is currently tested in several clinical trials⁴⁸, especially in the context of cancer treatment. For example, combination with checkpoint blockade inhibition revealed an enhanced increase in virus-specific CD8⁺ T cells in HPV-related cancer associated with an objective treatment response⁴⁸. Hence, TGFβ inhibitors seem to be feasible and promising in clinical application of virus-associated diseases.

The emergence of attenuated HBV-specific CD8⁺ T cells was strictly associated with the so termed HBeAg− chronic HBV infection phase of chronic HBV infection. Of note, testing of the HBeAg⁺ chronic HBV infection and HBeAg⁺ hepatitis B phases warrant further investigations as we restricted our analyses to the HBeAg− disease phases of chronic HBV infection to cover the majority of patients in advanced countries. The clinical phase of HBeAg− chronic HBV infection is characterized by low or even hardly detectable VL and normal levels of transaminases reflecting endogenous viral control and limited immunopathology. This observation proposes a clear link of the T cell program and differentiation to the clinical course in chronic HBV infection. Furthermore, it reveals that HBV-specific CD8⁺ T cells are not ‘nonfunctional’ but rather adapted during chronic infection strongly highlighting the important antiviral efficiency of HBV-specific CD8⁺ T cells also during the chronic phase of HBV infection. Subsequently, CD8⁺ T cells involving viral containment come along with reduced HBV-triggered immunopathology. Functional adaptation within the HBV-specific CD8⁺ T cell attenuation program also comprises the maintenance of these cells in the context of persistent antigen stimulation guided by TCF1 and CD127. This ensures ongoing HBV-specific CD8⁺ T cell surveillance by attenuated subsets in chronic HBV infection, contributing to endogenous control. Functional HBV-specific T cells have recently been reported as candidate immunological biomarkers for safe therapy discontinuation in patients with chronic HBV^49,50 and in spontaneously controlling patients with HIV infection⁵¹ underpinning the translational value of this concept with respect to therapeutic approaches and beyond chronic HBV infection. In our study, we have not confirmed the presence of attenuated HBV-specific CD8⁺ T cells at the site of infection, the liver; however, circulating CD8⁺ T cells adhere and/or infiltrate the HBV-infected liver⁵², thereby comprising an important part of the T cell-mediated surveillance in HBV infection. Moreover, they represent the T cell compartment that can be clinically monitored. All in all, our results unravel a clinically relevant new layer of functional adaptation of antiviral CD8⁺ T cell immunity in chronic HBV infection next to the previously reported phenotypic and functional heterogeneity that is shaped by the underlying viral infection (HBV versus HCV), the targeted antigen (core versus polymerase)^18,19, ongoing replication²¹, duration of infection, age or the tissue distribution⁴⁵. By this, we showed the existence of a specific T cell programming beyond T cell exhaustion in humans that has a clinically relevant role by the potential to contribute to viral containment during chronic infection.

Methods

Study cohort

Blood samples of participants with viral hepatitis were recruited at the Department of Medicine II of the University Hospital Freiburg, Germany and at the Department of Gastroenterology, Hepatology and Endocrinology and at Hannover Medical School, Hannover, Germany. In total, 80 HLA-A*02:01-positive individuals and 14 HLA-B*35:01-positive individuals with chronic HBV infection were included in this study. Individuals were classified into different clinical phases of chronic HBV infection according to EASL guidelines that consider the presence of HBeAg, HBV DNA levels, transaminase levels (alanine transaminase (ALT) and aspartate transferase (AST)) and the presence or absence of liver inflammation². Throughout the entire study, the HBeAg− cHBV infection phase is referred to as endogenous control/chronic HBV. The HBeAg− CHB phase is specified as high VL/chronic HBV and individuals within the HBeAg− CHB phase and under nucleos(t)ide analogs (NUC) treatment are indicated as NUC control/chronic HBV. In addition, eight HLA-A*02:01-positive individuals with acute HBV infection as well as 11 HLA-A*02:01-positive individuals who resolved an acute HBV infection were enrolled in this study. Characteristics of included individuals are summarized in Supplementary Tables 1–6. HBeAg, serum HBV DNA and AST/ALT values were determined as part of the clinical diagnostics at the University Hospital Freiburg, Germany. Confirmation of HLA-A*02:01 was performed by four-digit HLA-typing by next-generation sequencing on an MiSeq system using commercially available primers (GenDx). Due to limitations of sample material per patient we were not able to perform all analyses with cells from all patients. Detailed information about the inclusion of the patients in the different analyses are given in the Source data.

Ethics

Written informed consent was obtained from all participants before blood donation. Consent covers data publication after anonymization. The study was conducted according to federal guidelines, local ethics committee regulations of Albert-Ludwigs-Universität, Freiburg, Germany (no. 474/14, 383/19, 384/18), Hannover Medical School, Hannover, Germany (no. 3374-2016) and the Declaration of Helsinki (1975).

PBMC isolation

Venous blood samples were collected in EDTA-anticoagulated tubes. Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using lymphocyte separation medium (PAN Biotech) and stored at −80 °C until usage. Frozen PBMCs were thawed in complete medium (RPMI 1640 supplemented with 10% fetal calf serum, 1% penicillin/streptomycin and 1.5% 1 M HEPES buffer (all additives from Thermo Scientific)) containing 50 U ml⁻¹ Benzonase (Sigma).

Peptides and tetramers

Peptides of immunodominant HBV-derived epitopes (HLA-A*02:01/HBV core₁₈, FLPSDFFPSV; HLA-A*02:01/HBV pol₄₅₅, GLSRYVARL; HLA-B*35:01/HBV pol₁₇₃, SPYSWEQEL) were synthesized with standard Fmoc chemistry and a purity of >70% from Genaxxon. HLA class I epitope-specific tetramers were generated through conjugation of biotinylated peptide/HLA class I monomers with PE-conjugated streptavidin (Agilent) at a peptide/HLA I: streptavidin molar ratio of 5:1. Of note, targeted epitopes of HBV-specific CD8⁺ T cells were previously analyzed for viral sequence mutations. T cell responses of patients harboring viral sequence mutations in the targeted epitope were excluded.

HBV viral sequencing

Viral DNA was extracted from 0.75–1.0 ml patient’s plasma at the used bleed date or latest available HBV DNA-positive time point using QIAamp UltraSens technology (QIAGEN) according to the manufacturer’s protocol. A two-step nested PCR approach was used to amplify HBV DNA fragments for viral sequence analyses of HBV epitopes. The following specific primers were used: HBV core₁₈/13F1 (5′-CACCTCTGCCTAATCATCTC-3′), HBV core₁₈/499R1 (5′-CCGGAAGTGTT GATAAGATAGG-3′), HBV core₁₈/46F2 (5′-ACTGT TCAAGCCTCCAAGCTG-3′), HBV core₁₈/₄₅₅R2 (5′-GAGGAGTGCGAATCCACACTC-3′), HBV pol₄₅₅/ 1384F1 (5′-ACCAAACT CTGCARGATCCCAG-3′), HBV pol₄₅₅/2467R1 (5′-TGGTGGCTCCAGTTCAGGAAC-3′), HBV pol₄₅₅/1430F2 (5′-TGGTGGCTCCAGTTCAGGAAC-3′), HBV pol₄₅₅/2324R2 (5′-ATCAATAGG CCTGTTAACAGGAAG-3′), HBV pol₁₇₃/454F1 (5′-GGAGTGTGGATTCGCACTCC-3′), HBV pol₁₇₃/ 3146R1 (5′-CCAATTTATGCCTACAGCCTC-3′), HBV pol₁₇₃/484F2 (5′-AGACCACCAA ATGCCCCTATC-3′) and HBV pol₁₇₃/3143R2 (5′-CAATTTATGCCTACAGCCTCCTA-3′). Purified PCR products were sequenced via Sanger sequencing (Eurofins). Sequences and chromatograms were evaluated with the program Geneious (Biomatters), allowing comparison with wild-type reference sequences (NCBI accession codes: genotype A X02763, genotype D X02496 and genotype E X75657).

Magnetic bead-based enrichment of HBV epitope-specific CD8⁺ T cells

Enrichment of HBV epitope-specific CD8⁺ T cells was performed. In brief, 1 × 10⁷ to 2 × 10⁷ PBMCs were labeled for 30 min with PE-coupled peptide-loaded HLA class I tetramers. Enrichment was then carried out with anti-PE microbeads using magnetic-activated cell sorting technology (Miltenyi Biotec) according to the manufacturer’s instructions. Subsequently, enriched HBV epitope-specific CD8⁺ T cells were used for multiparametric flow cytometric analysis. Ex vivo frequency of HBV epitope-specific CD8⁺ T cells were calculated as follows. First, the number of total CD8⁺ T cells within the analyzed sample was determined by multiplying the (counted) total number of PBMCs in the pre-enriched sample with the frequency of single, live, non-dump CD8⁺ T cells in the pre-enriched sample. The absolute number of virus-specific CD8⁺ T cells was the number of HBV epitope-specific CD8⁺ T cells within the single, live, non-dump CD8⁺ T cell gate present in the enriched fraction. The frequency of circulating HBV epitope-specific CD8⁺ T cells was defined as the absolute number of HBV epitope-specific CD8⁺ T cells/absolute number of CD8⁺ T cells.

Ex vivo functional CD8⁺ T cell capacity assay

After magnetic bead-based enrichment of HBV pol₄₅₅-specific CD8⁺ T cells, enriched samples were stimulated with PMA (50 ng ml⁻¹) and ionomycin (1 μg ml⁻¹) (all Sigma) in the presence of anti-CD107a (H4A3, 1:100, cat. 560664, BD Bioscience) to detect the degranulation surrogate CD107a. To assess the intracellular cytokine production, brefeldin A (GolgiPlug, 0.5 μl ml⁻¹) and monensin (GolgiStop, 0.5 μl ml⁻¹) (all BD Biosciences) were added after 1 h. After an additional 3 h of incubation, surface and intracellular staining were performed. One-third of the enriched fractions were used as unstimulated controls for background subtraction. Of note, the low frequency of HBV-specific CD8⁺ T cells hinders testing of the functional activity in response to peptide stimulation; however, stimulation with PMA/ionomycin enables testing of the functional capacity of CD8⁺ T cells, bypassing TCR signaling and thus allowing previous peptide-loaded HLA class I-based enrichment of HBV-specific CD8⁺ T cells to increase the sensitivity.

TGFβ signaling blockade assay

Approximately 2 × 10⁶ PBMCs were stimulated with HBV pol₄₅₅-specific peptide (GLSRYVARL, 10 μg ml⁻¹) and anti-CD28 monoclonal antibody (0.5 μg ml⁻¹, BD Bioscience) and expanded at 37 °C for 14 days in complete culture medium containing rIL-2 (20 IU ml⁻¹, STEMCELL Technologies). Then, 50 ng ml⁻¹ TGF-β RI kinase inhibitor II CAS 446859-33-2 (Sigma-Aldrich), was added every 3–4 days. Expanded HBV pol₄₅₅-specific CD8⁺ T cells without inhibition of TGF-β RI kinase were used as control.

Treatment of HBV pol₄₅₅-specific CD8⁺ T cells with IL-12 or IFNα

Approximately 1–2 × 10⁶ PBMCs were stimulated with HBV pol₄₅₅-specific peptide (10 μg ml⁻¹) and anti-CD28 monoclonal antibody (0.5 μg ml⁻¹, BD Bioscience) and expanded at 37 °C for 14 days in complete culture medium containing rIL-2 (20 IU ml⁻¹, STEMCELL Technologies). Then, 10 ng ml⁻¹ rIL-12 (Sigma) or 1,000 U ml⁻¹ IFNα (STEMCELL Technologies) were added every 3–4 days to the respective culture. Expanded HBV pol₄₅₅-specific CD8⁺ T cells without any treatment were used as control.

TGFβ ELISA

Human TGFβ1 ELISA (Invitrogen) was performed according to the manufacturer’s protocol. In brief, standards and plasma samples or supernatant of 4 h PMA (50 ng ml⁻¹) and ionomycin (1 μg ml⁻¹) (all Sigma)-stimulated enriched HBV pol₄₅₅-specific CD8⁺ T cells were pipetted into preassigned wells. After 2 h of incubation and washing, a biotin-conjugated anti-human TGFβ1 antibody was added and bound to human TGFβ1 captured by the first antibody. Following 1 h incubation, unbound biotin-conjugated anti-human TGFβ1 antibody was removed during a wash step. Then, streptavidin–horseradish peroxidase (HRP) was added. After 1 h incubation, unbound streptavidin–HRP was removed during a wash step and a substrate solution reactive with HRP was applied to the wells for around 15 min. A colored product was formed in proportion to the amount of human TGFβ1 present in the sample or standard. The reaction was terminated by addition of the stop solution. The absorbance was measured at 450 nm with a TECAN Spark with SparkControl megellan software v.2.2. (Tecan). A standard curve was prepared from seven human TGFβ1 standard dilutions and human TGFβ 1 sample concentration were determined using Magellan software (Tecan).

Multiparametric flow cytometry

The following antibodies were used for multiparametric flow cytometry: anti-CD14-APC-eFluor780 (61D3, 1:100 dilution), anti-CD19-APC-eFluor780 (HIB19, 1:100 dilution), anti-EOMES-PE-Cy5.5 (WD1928, 1:800 dilution), anti-KLRG1-SuperBright702 (13F12F2, 1:50 dilution), anti-T-BET-PE-Cy7 (4B10, 1:200 dilution) and anti-TOX-eFluor660 (TRX10, 1:100 dilution) (all eBioscience). Anti-BLIMP1-PE-Dazzle (6D3, 1:25 dilution), anti-CD107a-APC (H4A3, 1:100 dilution), anti-CD38-BUV737 (HB7, 1:200 dilution), anti-CD45RA-BUV496 (HI100,1:200 dilution), anti-CD8-BUV395 (RPA-T8,1:200 dilution), anti-CD8-BUV496 (RPA-T8, 1:00 dilution), anti-CD8-BB515 (RPA-T8,1:100 dilution), anti-GZMB-BV510 (GB11, 1:100 dilution), anti-IFNγ-FITC (4 S.B3, 1:8 dilution), anti-Ki67-BV510 (B56, 1:20 dilution), anti-KLRD1-BV421 (HP-3D9, 1:200 dilution), anti-TNF-PE-Cy7 (MAb11, 1:50 dilution), anti-PD1-BV786 (EH12.1, 1:33 dilution) (all BD Bioscience). anti-CCR7-Ax700 (G043H7, 1:25 dilution), anti-CCR7-BV786 (G043H7, 1:33 dilution), anti-CD127-BV605 (A019D5, 1:33 dilution), anti-CD45RA-BV510 (HI100, 1:200 dilution), anti-rabbit IgG- FITC (Poly4064, 1:200 dilution), anti-rabbit IgG-PE-CF594 (Poly4064, 1:200 dilution) and anti-PRF1-BV711 (dG9, 1:200 dilution) (all BioLegend). Anti-ZEB2-AF647 (923328, 1:25 dilution, R&D), anti-NKG7-pure (polyclonal, 1:50 dilution, Biorybt) and anti-RUNX1/2/3-pure (SD0803, 1:800 dilution, Thermo Fisher). Fixable Viability Dyes APC-eFluor780 (1:200 dilution, eBioscience) was used for live/dead discrimination. Fixation/Permeabilization Solution kit (BD Biosciences) and FoxP3/Transcription Factor Staining Buffer Set (eBioscience) were applied according to the manufacturer’s instructions to stain for cytoplasmic and intranuclear molecules, respectively. Cells were fixed with 2% paraformaldehyde (PFA; Sigma). Analyses were performed using FACS Canto II or LSRFortessa with FACSDiva software v.10.6.2 (BD). We performed a machine and measurement standardization procedure applying the CS&T system (BD Biosciences) to optimize and standardize cytometer setup and to receive reproducible data. Data were evaluated with FlowJo v.10 (Treestar). Only individuals harboring ≥80% non-naive HBV epitope-specific CD8⁺ T cells and with at least ten cells were included in further analyses.

Dimensionality reduction of multiparameter flow cytometry data

The visualization of multiparametric flow cytometry data was conducted with R v.4.3.3. using the Bioconductor (release 3.18) CATALYST package (v.1.26.0)⁵³. The analyses were performed on gated non-naive HBV epitope-specific CD8⁺ T cells as well as non-naive GZMB⁺ HBV pol₄₅₅-specific CD8⁺ T cells. Analysis for Figs. 2e, 3d and 6h and Extended Data Figs. 2c and 8d included the markers GZMB, PRF1, NKG7, KLRD1, T-BET, EOMES, ZEB2, CD127, TOX and PD-1. Analysis for Fig. 5e and Extended Data Fig. 5a was performed on PRF1, NKG7, KLRD1, T-Bet, EOMES, ZEB2, CD127, TOX and PD-1. Downsampling of cells to 1,000 was performed before dimensionality reduction to facilitate the visualization of different samples. Marker intensities were transformed by arcsinh (inverse hyperbolic sine) with a cofactor of 150. Dimensionality reduction on the transformed data was achieved by t-SNE, MDS and diffusion map visualization.

Single-cell sorting for scRNA-seq by mCEL-seq2 approach

HBV epitope-specific CD8⁺ T cells were enriched by magnetic bead-based enrichment, and surface staining was performed. Live HBV epitope-specific CD8⁺ T cells were sorted in 384-well plates (Bio-Rad) containing lysis buffer and mineral oil using FACS Melody Cell Sorter with BD FACSChorus software v.3.4.1 (BD). Naive CD45RA⁺CCR7⁺ T cells were excluded. The sorter was run on single-cell sort mode. After the completion of sorting, the plates were centrifuged for 1 min at 2,200g at 4 °C, snap frozen in liquid nitrogen and stored at −80 °C until processing.

scRNA amplification and library preparation of mCEL-seq2 approach

scRNA-seq was performed using the mCEL-seq2 protocol, an automated and miniaturized version of CEL-seq2 on a mosquito nanoliter-scale liquid-handling robot (TTP LabTech)^24,54. As reported previously, the CEL-seq2 protocol was established and optimized to analyze the heterogeneity of rare cells⁵⁴. Libraries with 96 cells each were sequenced per lane on an Illumina HiSeq 3000 sequencing system (pair-end multiplexing run) at a depth of ~130,000–200,000 reads per cell.

Quantification of transcript abundance of mCEL-seq2 approach

Paired-end reads were aligned to the transcriptome using bwa (v.0.6.2-r126) with default parameters⁵⁵. The transcriptome contained all gene models based on the human whole-genome ENCODE v.24 release. All isoforms of the same gene were merged to a single gene locus. Furthermore, gene loci overlapping by >75% were merged to larger gene groups. The right mate of each read pair was mapped to the ensemble of all gene loci and to the set of 92 External RNA Control Consortium spike-ins in the sense direction⁵⁶. Reads mapping to multiple loci were discarded. The left read contains the barcode information: the first six bases corresponded to the unique molecular identifier (UMI) followed by six bases representing the cell-specific barcode. The remainder of the left read contains a polyT stretch. For each cell barcode, the number of UMIs per transcript was counted and aggregated across all transcripts derived from the same gene locus. Based on binomial statistics, the number of observed UMIs were converted into transcript counts⁵⁷.

scRNA-seq data analysis of mCEL-seq2 approach

Clustering analysis and visualization were performed using Seurat v.4.3.0²⁶. Only protein coding genes were included in the downstream data analysis. Mitochondrial genes and genes encoding for small and large ribosomal subunits (Supplementary Fig. 1a) were excluded before clustering. Cells with a total number of transcripts <1,000 (<500 for HBV core₁₈ data) were discarded, and count data of the remaining cells were normalized. As published previously for mCEL-seq2 approaches^{24,25,58,59,60}, cells expressing >1.5% of KCNQ1OT1, a potential marker for low-quality cells⁶¹, were not considered for analysis. In brief, KCNQ1OT1 serves as a non-cell-type-specific marker in several datasets analyzed across different tissues in mice and humans using mCEL-seq2 approach. Cells, also HBV-specific CD8⁺ T cells, with a higher percentage of KCNQ1OT1 transcripts mainly represent low-quality cells with a significantly lower number of transcripts per cell (Supplementary Fig. 2). In particular, the genes correlating with KCNQ1OT1 were primarily long noncoding RNAs (AC006548.28, AC092669.3, ASTN2, ATP9B, BLOC1S6, C4orf32, HELLPAR, KIAA1328, LINC00504, LINC00506, LINC00976, MAB21L3, ORC4, RASEF, RP11-142G1.3, RP11-17A4.2, RP11-259O2.3, RP11-346D14.1, RP11-382B18.1, RP11-609N14.4, RP11-624A21.1, RP11-703G6.1, RP3-323A16.1, RP4-598P13.1, RPAP2, SOD2, SULT1B1, TBC1D8B, TMEM212, UGDH-AS1, WDPCP and Xbac-BPG154L12.4) and their inclusion or exclusion would not affect the results and findings of the study. Additionally, transcripts correlating to KCNQ1OT1 with a Pearson’s correlation coefficient > 0.65 were removed. We observed heterogeneity between participants in the dataset (Supplementary Fig. 3), which was corrected using the RunHarmony function of the Harmony package (v.0.1.1), which runs Harmony, a scRNA-seq data integration algorithm with a Seurat workflow²⁷. Default parameters were used to run Seurat, except for cluster resolution (set to 0.5 for the integrative analysis of HCV and HBV data as well as the combined analysis HBV core₁₈ and HBV pol₄₅₅ data and 0.4 for the analysis of HBV pol₄₅₅ data). Cells from each participant were considered a batch. Dimensionality reduction for visualization was performed using the RunUMAP function where reduction parameter was set to ‘harmony’. DEGs were calculated using the FindMarkers function of Seurat algorithm. Diffusion maps were created using the destiny R package (v.3.16.0)³¹. The top 2,000 variable genes obtained using the FindVariableFeatures function of Seurat were included in the diffusion map calculation. The number of nearest neighbors k was set to 10 and 50 principal components were used. Gene regulatory network inference on HBV pol₄₅₅ data was performed using SCENIC (v.1.3.1) with default settings. The weightThreshold was set to 0.006 in the runSCENIC_1_coexNetwork2modules function³².

scRNA-seq by 10x Genomics approach

scRNA-seq was performed using 10x Genomics with feature barcoding technology to multiplex cells from different individuals so that they could be loaded on one well to reduce costs and minimize technical variability. Hashtag oligonucleotides were obtained as purified and already oligonucleotide-conjugated in TotalSeq-C (5′ chemistry) format from BioLegend. HBV pol₄₅₅-specific CD8⁺ T cells were enriched by magnetic bead-based enrichment and surface staining was performed. Cells were stained with barcoded antibodies together with the surface staining solution. The antibody concentrations used were 1 µg per million cells, as recommended by the manufacturer (BioLegend). After staining, cells were washed three times in PBS (PAN Biotech) containing 2% BSA (AppliChem) and 0.01% Tween 20 (Merck KGaA) followed by centrifugation (300g for 5 min at 4 °C) and supernatant exchange. After the final wash, the cells were resuspended in PBS, filtered through 40-µm cell strainers and processed for sorting. Live HBV pol₄₅₅-specific CD8⁺ T cells were sorted in 1.5-ml microcentrifuge tubes (Eppendorf) containing 20 µl PBS using BD Aria Fusion with FACSDiva software v.10.6.2 (BD). The 1.5-ml microcentrifuge tubes were coated with 5% BSA overnight and rinsed twice with 1 ml PBS before usage. Naive, CD45RA⁺CCR7⁺ T cells were excluded. The sorter was run on purity sort mode. Sorted cells were processed through the 10x Genomics single-cell workflow according to the manufacturer’s instructions. Libraries were pooled to desired quantities to obtain appropriate sequencing depths as recommended by 10x Genomics and sequenced on a NovaSeq 6000 flow cell.

Quantification of transcript abundance of 10x Genomics approach

Quantification of gene expression and cell surface protein abundance counts was performed using Cell Ranger v.6.0.0 using the count command, which performs alignment, filtering, barcode counting and UMI counting as well as processing feature barcoding data. Alignment was performed using a prebuilt Cell Ranger human reference GRCh38 (GENCODE v.32/Ensembl 98).

scRNA-seq data analysis of 10x Genomics approach

Clustering analysis and visualization were performed using Seurat v.4.3.0 (ref. ²⁶). Demultiplexing of the hashtag oligonucleotides was performed using HTODemux function with positive.quantile set to 0.99. Mitochondrial genes and genes encoding for small and large ribosomal subunits (Supplementary Fig. 1b) as well as T cell receptor α, β, γ and δ (TRA, TRB, TRG and TRD) were excluded before clustering. Cells with a total number of transcripts <500 and expressing >10% of mitochondrial genes were discarded and count data of the remaining cells were normalized. We observed heterogeneity between participants in the dataset, which was corrected using the RunHarmony function of the harmony package (v.0.1.1) which runs Harmony, a scRNA-seq data integration algorithm with the Seurat workflow²⁷. Default parameters were used to run Seurat, except that cluster resolution was set to 0.6. Cells from each participant were considered as a batch. Dimensionality reduction for visualization was performed using RunUMAP function where reduction parameter was set to ‘harmony’. DEGs were calculated using the FindAllMarkers function of Seurat algorithm. Gene regulatory network inference was performed using SCENIC (v.1.3.1) with default settings. The weightThreshold was set to 0.004 in the runSCENIC_1_coexNetwork2modules function³².

Statistics

Statistical analysis was performed with GraphPad Prism v.10 (GraphPad Prism Software) and R v.4.3.3. Statistical significance was assessed by a Kruskal–Wallis test, including Dunn’s multiple comparisons test, two-tailed Mann–Whitney comparison tests, simple linear regression analysis and two-tailed Wilcoxon matched-pairs signed-rank test. Analyses were performed in independent experiments.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The expression count files for the scRNA-seq data of HBV epitope-specific CD8⁺ T cells reported in this paper are available to download from the Gene Expression Omnibus under accession number GSE259231. Processed data can be downloaded from https://github.com/sagar161286/HBV2024. The scRNA-seq data of HCV-specific CD8⁺ T cells used for comparison were downloaded under accession number GSE150305 (https://doi.org/10.1038/s41590-020-00817-w). For the alignment of scRNA-seq data, Cell Ranger human reference GRCh38 (GENCODE v.32/Ensembl 98) was used for the 10x Genomics approach. The gene sets for GSEA were downloaded from ref. ¹⁷ (exhaustion (chronic TCF-1⁻), memory-like (chronic TCF-1⁺) and memory signature), ref. ²⁵ (cytotoxicity signature) or GO:0002369 (https://www.gsea-msigdb.org/gsea/msigdb/mouse/geneset/GOBP_T_CELL_CYTOKINE_PRODUCTION.html and IFNG, TNF and IL2 were added), for IL-10 pathway genes (https://www.gsea-msigdb.org/gsea/msigdb/cards/BIOCARTA_IL10_PATHWAY), for IL-2 pathway genes (https://www.gsea-msigdb.org/gsea/msigdb/cards/BIOCARTA_IL2_PATHWAY), for IL-12 pathway genes (https://www.gsea-msigdb.org/gsea/msigdb/human/geneset/BIOCARTA_IL12_PATHWAY.html) and for type I IFN pathway genes (https://www.gsea-msigdb.org/gsea/msigdb/cards/ST_TYPE_I_INTERFERON_PATHWAY). The scRNA-seq data of liver FNAs and blood used for comparison were downloaded under accession number GSE234241 (ref. ³⁴). HBV wild-type reference sequences used for analyzing the HBV viral sequencing data were downloaded with NCBI accession codes: genotype A X02763, genotype D X02496 and genotype E X75657. Raw data in this study are provided in Source data. Source data are provided with this paper.

Code availability

Codes to reproduce the scRNA-seq data analysis and the analyses of multiparametric flow cytometry data and figures are available at https://github.com/sagar161286/HBV2024.

References

World Health Organization. Hepatitis B. https://www.who.int/news-room/fact-sheets/detail/hepatitis-b (2024).
EASL. 2017 Clinical Practice Guidelines on the management of hepatitis B virus infection. J. Hepatol. 67, 370–398 (2017).
Article Google Scholar
Thimme, R. et al. CD8(+) T cells mediate viral clearance and disease pathogenesis during acute hepatitis B virus infection. J. Virol. 77, 68–76 (2003).
Article CAS PubMed PubMed Central Google Scholar
Maini, M. K. et al. The role of virus-specific CD8(+) cells in liver damage and viral control during persistent hepatitis B virus infection. J. Exp. Med. 191, 1269–1280 (2000).
Article CAS PubMed PubMed Central Google Scholar
Webster, G. J. et al. Longitudinal analysis of CD8⁺ T cells specific for structural and nonstructural hepatitis B virus proteins in patients with chronic hepatitis B: implications for immunotherapy. J. Virol. 78, 5707–5719 (2004).
Article CAS PubMed PubMed Central Google Scholar
Boni, C. et al. Characterization of hepatitis B virus (HBV)-specific T-cell dysfunction in chronic HBV infection. J. Virol. 81, 4215–4225 (2007).
Article CAS PubMed PubMed Central Google Scholar
Das, A. et al. Functional skewing of the global CD8 T cell population in chronic hepatitis B virus infection. J. Exp. Med. 205, 2111–2124 (2008).
Article CAS PubMed PubMed Central Google Scholar
Kurktschiev, P. D. et al. Dysfunctional CD8⁺ T cells in hepatitis B and C are characterized by a lack of antigen-specific T-bet induction. J. Exp. Med. 211, 2047–2059 (2014).
Article CAS PubMed PubMed Central Google Scholar
Schurich, A. et al. Distinct metabolic requirements of exhausted and functional virus-specific CD8 T cells in the same host. Cell Rep. 16, 1243–1252 (2016).
Article CAS PubMed PubMed Central Google Scholar
Fisicaro, P. et al. Targeting mitochondrial dysfunction can restore antiviral activity of exhausted HBV-specific CD8 T cells in chronic hepatitis B. Nat. Med. 23, 327–336 (2017).
Article CAS PubMed Google Scholar
Heim, K. et al. TOX defines the degree of CD8⁺ T cell dysfunction in distinct phases of chronic HBV infection. Gut 70, 1550–1560 (2020).
Article PubMed Google Scholar
Wherry, E. J. & Kurachi, M. Molecular and cellular insights into T cell exhaustion. Nat. Rev. Immunol. 15, 486–499 (2015).
Article CAS PubMed PubMed Central Google Scholar
McLane, L. M., Abdel-Hakeem, M. S. & Wherry, E. J. CD8 T cell exhaustion during chronic viral infection and cancer. Annu. Rev. Immunol. 37, 457–495 (2019).
Article CAS PubMed Google Scholar
Bengsch, B. et al. Coexpression of PD-1, 2B4, CD160 and KLRG1 on exhausted HCV-specific CD8⁺ T cells is linked to antigen recognition and T cell differentiation. PLoS Pathog. 6, e1000947 (2010).
Article PubMed PubMed Central Google Scholar
Alfei, F. et al. TOX reinforces the phenotype and longevity of exhausted T cells in chronic viral infection. Nature 571, 265–269 (2019).
Article CAS PubMed Google Scholar
Blackburn, S. D., Shin, H., Freeman, G. J. & Wherry, E. J. Selective expansion of a subset of exhausted CD8 T cells by αPD-L1 blockade. PNAS 105, 15016–15021 (2008).
Article CAS PubMed PubMed Central Google Scholar
Utzschneider, D. T. et al. T cell factor 1-expressing memory-like CD8(+) T cells sustain the immune response to chronic viral infections. Immunity 45, 415–427 (2016).
Article CAS PubMed Google Scholar
Schuch, A. et al. Phenotypic and functional differences of HBV core-specific versus HBV polymerase-specific CD8⁺ T cells in chronically HBV-infected patients with low viral load. Gut 68, 905–915 (2019).
Article CAS PubMed Google Scholar
Hoogeveen, R. C. et al. Phenotype and function of HBV-specific T cells is determined by the targeted epitope in addition to the stage of infection. Gut 68, 893–904 (2019).
Article CAS PubMed Google Scholar
Winkler, F. et al. Enolase represents a metabolic checkpoint controlling the differential exhaustion programmes of hepatitis virus-specific CD8(+) T cells. Gut 72, 1971–1984 (2023).
Article CAS PubMed Google Scholar
Cheng, Y. et al. Multifactorial heterogeneity of virus-specific T cells and association with the progression of human chronic hepatitis B infection. Science Immunol. 4, eaau6905 (2019).
Article CAS Google Scholar
Aliabadi, E. et al. Impact of HBsAg and HBcrAg levels on phenotype and function of HBV-specific T cells in patients with chronic hepatitis B virus infection. Gut 71, 2300–2312 (2022).
Article CAS PubMed Google Scholar
Bertoletti, A. & Ferrari, C. Adaptive immunity in HBV infection. J. Hepatol. 64, S71–s83 (2016).
Article CAS PubMed Google Scholar
Herman, J. S., Sagar & Grün, D. FateID infers cell fate bias in multipotent progenitors from single-cell RNA-seq data. Nat. Methods 15, 379–386 (2018).
Article CAS PubMed Google Scholar
Hensel, N. et al. Memory-like HCV-specific CD8(+) T cells retain a molecular scar after cure of chronic HCV infection. Nat. Immunol. 22, 229–239 (2021).
Article CAS PubMed Google Scholar
Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e3529 (2021).
Article CAS PubMed PubMed Central Google Scholar
Korsunsky, I. et al. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat. Methods 16, 1289–1296 (2019).
Article CAS PubMed PubMed Central Google Scholar
Giles, J. R. et al. Shared and distinct biological circuits in effector, memory and exhausted CD8(+) T cells revealed by temporal single-cell transcriptomics and epigenetics. Nat. Immunol. 23, 1600–1613 (2022).
Article CAS PubMed PubMed Central Google Scholar
Beltra, J. C. et al. Stat5 opposes the transcription factor Tox and rewires exhausted CD8(+) T cells toward durable effector-like states during chronic antigen exposure. Immunity 56, 2699–2718.e2611 (2023).
Article CAS PubMed Google Scholar
Giles, J. R. et al. Human epigenetic and transcriptional T cell differentiation atlas for identifying functional T cell-specific enhancers. Immunity 55, 557–574.e557 (2022).
Article CAS PubMed PubMed Central Google Scholar
Haghverdi, L., Büttner, M., Wolf, F. A., Buettner, F. & Theis, F. J. Diffusion pseudotime robustly reconstructs lineage branching. Nat. Methods 13, 845–848 (2016).
Article CAS PubMed Google Scholar
Aibar, S. et al. SCENIC: single-cell regulatory network inference and clustering. Nat. Methods 14, 1083–1086 (2017).
Article CAS PubMed PubMed Central Google Scholar
Guan, T. et al. ZEB1, ZEB2, and the miR-200 family form a counterregulatory network to regulate CD8(+) T cell fates. J. Exp. Med. 215, 1153–1168 (2018).
Article CAS PubMed PubMed Central Google Scholar
Genshaft, A. S. et al. Single-cell RNA sequencing of liver fine-needle aspirates captures immune diversity in the blood and liver in chronic hepatitis B patients. Hepatology 78, 1525–1541 (2023).
PubMed Google Scholar
Dominguez, C. X. et al. The transcription factors ZEB2 and T-bet cooperate to program cytotoxic T cell terminal differentiation in response to LCMV viral infection. J. Exp. Med. 212, 2041–2056 (2015).
Article CAS PubMed PubMed Central Google Scholar
Omilusik, K. D. et al. Transcriptional repressor ZEB2 promotes terminal differentiation of CD8⁺ effector and memory T cell populations during infection. J.f Exp. Med. 212, 2027–2039 (2015).
Article Google Scholar
Daniel, B. et al. Divergent clonal differentiation trajectories of T cell exhaustion. Nat. Immunol. 23, 1614–1627 (2022).
Article CAS PubMed PubMed Central Google Scholar
Iezzi, G., Karjalainen, K. & Lanzavecchia, A. The duration of antigenic stimulation determines the fate of naive and effector T cells. Immunity 8, 89–95 (1998).
Article CAS PubMed Google Scholar
Shedlock, D. J. & Shen, H. Requirement for CD4 T cell help in generating functional CD8 T cell memory. Science 300, 337–339 (2003).
Article CAS PubMed Google Scholar
Kolumam, G. A., Thomas, S., Thompson, L. J., Sprent, J. & Murali-Krishna, K. Type I interferons act directly on CD8 T cells to allow clonal expansion and memory formation in response to viral infection. J. Exp. Med. 202, 637–650 (2005).
Article CAS PubMed PubMed Central Google Scholar
Mescher, M. F. et al. Signals required for programming effector and memory development by CD8⁺ T cells. Immunol. Rev. 211, 81–92 (2006).
Article CAS PubMed Google Scholar
Williams, M. A., Tyznik, A. J. & Bevan, M. J. Interleukin-2 signals during priming are required for secondary expansion of CD8⁺ memory T cells. Nature 441, 890–893 (2006).
Article CAS PubMed PubMed Central Google Scholar
Joshi, N. S. et al. Inflammation directs memory precursor and short-lived effector CD8(+) T cell fates via the graded expression of T-bet transcription factor. Immunity 27, 281–295 (2007).
Article CAS PubMed PubMed Central Google Scholar
Ruby, C. E., Redmond, W. L., Haley, D. & Weinberg, A. D. Anti-OX40 stimulation in vivo enhances CD8⁺ memory T cell survival and significantly increases recall responses. Eur. J. Immunol. 37, 157–166 (2007).
Article CAS PubMed Google Scholar
Khakpoor, A. et al. Spatiotemporal differences in presentation of CD8 T cell epitopes during hepatitis B virus infection. J. Virol. 93, e01457–18 (2019).
Article CAS PubMed PubMed Central Google Scholar
Wiegand, S. B. et al. Soluble immune markers in the different phases of chronic hepatitis B virus infection. Sci. Rep. 9, 14118 (2019).
Article PubMed PubMed Central Google Scholar
Dahmani, A. & Delisle, J. S. TGF-β in T cell biology: implications for cancer immunotherapy. Cancers 10, 194 (2018).
Article PubMed PubMed Central Google Scholar
Metropulos, A. E. & Munshi, H. G. & Principe, D.R. The difficulty in translating the preclinical success of combined TGFβ and immune checkpoint inhibition to clinical trial. eBioMedicine 86, 104380 (2022).
Article CAS PubMed PubMed Central Google Scholar
Rivino, L. et al. Hepatitis B virus-specific T cells associate with viral control upon nucleos(t)ide-analogue therapy discontinuation. J. Clin. Investig. 128, 668–681 (2018).
Article PubMed PubMed Central Google Scholar
Tian, Y., James Ou, J. H., Hepatitis, B. & Virus-Specific, T. Cells as a biomarker for discontinuation of nucleos(t)ide analogue therapy for chronic hepatitis B. Hepatology 69, 1342–1344 (2019).
Article PubMed Google Scholar
Collins, D. R., Gaiha, G. D. & Walker, B. D. CD8(+) T cells in HIV control, cure and prevention. Nat. Rev. Immunol. 20, 471–482 (2020).
Article CAS PubMed PubMed Central Google Scholar
Pallett, L. J. & Maini, M. K. Liver-resident memory T cells: life in lockdown. Semin. Immunopathol.https://doi.org/10.1007/s00281-022-00932-w (2022).
Crowell H. L., Zanotelli, V. R. T., Chevrier S., Robinson M. CATALYST: Cytometry dATa anALYSis Tools. R package version 1.15.0 https://github.com/HelenaLC/CATALYST (2020).
Hashimshony, T. et al. CEL-Seq2: sensitive highly-multiplexed single-cell RNA-seq. Genome Biol. 17, 77 (2016).
Article PubMed PubMed Central Google Scholar
Li, H. & Durbin, R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics 26, 589–595 (2010).
Article PubMed PubMed Central Google Scholar
Baker, S. C. et al. The External RNA Controls Consortium: a progress report. Nat. Methods 2, 731–734 (2005).
Article CAS PubMed Google Scholar
Grun, D., Kester, L. & van Oudenaarden, A. Validation of noise models for single-cell transcriptomics. Nat. Methods 11, 637–640 (2014).
Article PubMed Google Scholar
Aizarani, N. et al. A human liver cell atlas reveals heterogeneity and epithelial progenitors. Nature 572, 199–204 (2019).
Article CAS PubMed PubMed Central Google Scholar
Sankowski, R. et al. Mapping microglia states in the human brain through the integration of high-dimensional techniques. Nat. Neurosci. 22, 2098–2110 (2019).
Article CAS PubMed Google Scholar
Nusser, A. et al. Developmental dynamics of two bipotent thymic epithelial progenitor types. Nature 606, 165–171 (2022).
Article CAS PubMed PubMed Central Google Scholar
Grun, D. et al. De novo prediction of stem cell identity using single-cell transcriptome data. Cell Stem Cell 19, 266–277 (2016).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank all patients for participating in the current study and the FREEZE-Biobank Center for biobanking (Freiburg University Medical Center) and the Medical Faculty for support. Furthermore, we thank J. Wersing and the Lighthouse Core Facility for help with cell sorting. The study was supported by grants from the German Research Foundation (DFG; TRR 179 project no. 272983813; CRC 1160 project no. 256073931; CRC 1479 project no. 441891347): IMPATH CRC 1160 project A02 to R.T. and M.H., IMPATH CRC 1160 project A03N to B.B., IMPATH CRC 1160 project A06 to C.N.-H., CRC/TRR 179 project TP04 to T.B. and CRC 1479 project 10 to M.H. M.H. was further supported by the DFG Heisenberg program.

Author information

These authors contributed equally: Kathrin Heim, Sagar.
These authors jointly supervised this work: Robert Thimme, Maike Hofmann.

Authors and Affiliations

Department of Medicine II, Medical Center – University of Freiburg, Freiburg, Germany
Kathrin Heim, Sagar, Özlem Sogukpinar, Sophie Kielbassa, Bertram Bengsch, Tobias Boettler, Christoph Neumann-Haefelin, Robert Thimme & Maike Hofmann
Faculty of Medicine, University of Freiburg, Freiburg, Germany
Kathrin Heim, Sagar, Özlem Sogukpinar, Florian Emmerich, Sophie Kielbassa, Bertram Bengsch, Tobias Boettler, Christoph Neumann-Haefelin, Robert Thimme & Maike Hofmann
Division of Infection and Immunity, Cardiff University School of Medicine, University Hospital of Wales, Cardiff, UK
Sian Llewellyn-Lacey & David A. Price
Systems Immunity Research Institute, Cardiff University School of Medicine, University Hospital of Wales, Cardiff, UK
David A. Price
Institute for Transfusion Medicine and Gene Therapy, University Medical Center, University of Freiburg, Freiburg, Germany
Florian Emmerich
Department of Gastroenterology, Hepatology, Infectious Diseases and Endocrinology, Hannover Medical School, Hannover, Germany
Anke R. M. Kraft & Markus Cornberg
German Centre for Infection Research (DZIF), partner site Hannover-Braunschweig, Braunschweig, Germany
Anke R. M. Kraft & Markus Cornberg
Centre for Individualised Infection Medicine (CiiM), Hannover, Germany
Anke R. M. Kraft & Markus Cornberg
Cluster of Excellence Resolving Infection Susceptibility (RESIST; EXC), Hannover Medical School, Hannover, Germany
Anke R. M. Kraft & Markus Cornberg
Institute of Molecular Immunology, School of Medicine, Technical University of Munich (TUM), Munich, Germany
Percy Knolle & Dirk Wohlleber
German Center for Infection Research, Munich, Germany
Percy Knolle
Institute of Molecular Immunology, School of Life Science, TUM, Munich, Germany
Percy Knolle
Signaling Research Centres BIOSS and CIBSS, University of Freiburg, Freiburg, Germany
Bertram Bengsch

Authors

Kathrin Heim
View author publications
You can also search for this author in PubMed Google Scholar
Sagar
View author publications
You can also search for this author in PubMed Google Scholar
Özlem Sogukpinar
View author publications
You can also search for this author in PubMed Google Scholar
Sian Llewellyn-Lacey
View author publications
You can also search for this author in PubMed Google Scholar
David A. Price
View author publications
You can also search for this author in PubMed Google Scholar
Florian Emmerich
View author publications
You can also search for this author in PubMed Google Scholar
Anke R. M. Kraft
View author publications
You can also search for this author in PubMed Google Scholar
Markus Cornberg
View author publications
You can also search for this author in PubMed Google Scholar
Sophie Kielbassa
View author publications
You can also search for this author in PubMed Google Scholar
Percy Knolle
View author publications
You can also search for this author in PubMed Google Scholar
Dirk Wohlleber
View author publications
You can also search for this author in PubMed Google Scholar
Bertram Bengsch
View author publications
You can also search for this author in PubMed Google Scholar
Tobias Boettler
View author publications
You can also search for this author in PubMed Google Scholar
Christoph Neumann-Haefelin
View author publications
You can also search for this author in PubMed Google Scholar
Robert Thimme
View author publications
You can also search for this author in PubMed Google Scholar
Maike Hofmann
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

K.H. planned, performed and analyzed the experiments except for the computational analysis of scRNA-seq data. S. performed the scRNA-seq experiments together with K.H. and analyzed the scRNA-seq data. S.K. and Ö.S. contributed to the experiments. S.L.-L., D.A.P., A.R.M.K. and M.C. provided resources. F.E. performed four-digit HLA-typing by next-generation sequencing. D.W., P.K., T.B., B.B. and C.N.H. contributed to data interpretation. M.H. and R.T. designed and supervised the study, contributed to experimental design and planning, interpreted data and wrote the manuscript, together with K.H.

Corresponding authors

Correspondence to Robert Thimme or Maike Hofmann.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Immunology thanks Matteo Iannacone, Mala Maini, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editor: P. Jauregui, in collaboration with the Nature Immunology team. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Experimental setup of scRNAseq analysis using mCEL-seq2.

(a) Study design is illustrated. Comprehensive analysis of the single-cell transcriptomes of HCV-specific CD8 ⁺ T cells from individuals infected with chronic HCV (n = 6) and after DAA-mediated cure of chronic HCV infection (n = 2) versus HBV-specific CD8 T cells from chronic HBV-infected individuals (n = 14) were performed by using the mCEL-seq2 technology. (b) Gating strategy of flow cytometry-based cell sorting of HBV-specific CD8 ⁺ T cells for scRNA-seq: after lymphocyte gating and two-way doublet exclusion, live CD8⁺ T cells were gated. Naïve CD8⁺ T cells (CD45RA + CCR7 + ) were excluded and HLA-A*02:01-restricted HBV-specific CD8⁺ T cells were identified by peptide-loaded HLA class I tetramer and then sorted into 384-well plates. (c) Feature plots showing gene expression of STAT5, KLRC1 and KLRB1 genes in C0–4. Scaled expression is depicted (gray, low; red, high). (d) Heatmap showing the expression of cytokines, cytotoxic markers and activation-related genes within the virus-specific CD8 + T cell clusters. Scaled expression levels are color coded (blue, low; brown, high).

Source data

Extended Data Fig. 2 Phenotypic and transcriptional characteristics of HBV-specific CD8⁺ T cells.

(a) Feature plots showing TCF7 and IL7R gene expression in C0–2 of HBV pol₄₅₅-specific CD8⁺ T cells (endogenous control/chronic HBV n = 3, NUC control/chronic HBV n = 4). Scaled expression is depicted (gray, low; red, high). (b) Expression analysis of cytotoxic molecules (GZMB and PRF1), the cytolytic regulator NKG7 as well as the C-type lectin-like NK cell receptors KLRD1 and KLRG1 was performed in peptide-loaded/HLA class I tetramer-enriched HBV core₁₈-specific CD8⁺ T cells from chronic HBV-infected individuals (for GZMB, PRF1, NKG7, KLRD1: endogenous control/ chronic HBV, low viral load, dark green, n = 14; high viral load, dark green/black, n = 2; NUC control/chronic HBV, low viral load, light green, n = 11; for KLRG1: endogenous control/ chronic HBV, low viral load, dark green, n = 8; high viral load, dark green/black, n = 2; NUC control/chronic HBV, low viral load, light green, n = 12). Representative histograms including gating of the individual markers are displayed. Each dot represents one HBV core₁₈-specific CD8⁺ T cell response and with color coding as indicated. Bar charts show the median value with IQR. (c) Expression levels of indicated markers are plotted on the diffusion map (blue, low; red, high) (endogenous control/ chronic HBV, low viral load, dark green, n = 12; high viral load, dark green/black, n = 4; NUC control/chronic HBV, low viral load, light green, n = 14). (d) Regulon activity matrix obtained from single-cell transcriptomes of HBV pol₄₅₅-specific CD8⁺ T cells of chronically HBV-infected individuals (endogenous control/chronic HBV n = 3, NUC control/chronic HBV n = 4) after applying SCENIC algorithm. Statistical significance was assessed by a Kruskal–Wallis test including Dunn’s multiple comparisons test (b).

Source data

Extended Data Fig. 3 Cytotoxic characteristics of HBV core18-versus HBV pol455-specific CD8⁺ T cells.

(a) Flow cytometric analysis of the cytotoxic molecules GZMB and PRF1, the cytolytic regulator NKG7 as well as the C-type lectin-like NK cell receptor KLRD1 in HBV core₁₈-specific CD8⁺ T cells from these cohorts is shown. Statistical graph depicts frequencies of HBV core₁₈-specific CD8⁺ T cells. Each dot represents one of the respective HBV core₁₈-specific CD8⁺ T cell responses with color coding as indicated: dark green, endogenous control/chronic HBV, n = 14; light green, acute HBV, n = 5; dark gray, acute resolved HBV, n = 3. (b) MDS representation of concatenated flow cytometry data gated on HBV core₁₈-specific and HBV pol₄₅₅-specific CD8⁺ T cells of acute (n = 12; HBV core₁₈: n = 5, HBV pol₄₅₅: n = 7), acute resolved (n = 7; HBV core₁₈: n = 3, HBV pol₄₅₅: n = 4) and endogenous control/chronic HBV infection (n = 26; HBV core₁₈: n = 14, HBV pol₄₅₅: n = 12) is illustrated. Bar charts show the median value with IQR. Statistical significance was assessed by a Kruskal–Wallis test including Dunn’s multiple comparisons test (a).

Source data

Extended Data Fig. 4 Distinct cytotoxic HBV pol455-specific CD8⁺ T cell responses in endogenously controlled versus acute HBV infection.

(a) Comparison of mCEL-seq2 (endogenous control/chronic HBV, n = 3; NUC control/chronic HBV, n = 4) and 10x Genomics scRNA-seq data (endogenous control/chronic HBV, n = 3; acute HBV, n = 3; acute resolved HBV, n = 2) are illustrated. nFeature RNA, nCount RNA and the clustering are depicted. (b) Comparison of C0–4 identified with 10x scRNA-seq with cluster signatures defined in the mCEL-seq2 approach by GSEA. (c) The transcription factors RUNX3 and BLIMP1 were assessed in HBV pol₄₅₅-specific CD8⁺ T cells of different cohorts (endogenous control/chronic HBV, n = 7; acute HBV, n = 5; acute resolved HBV, n = 4). Histograms and statistical graph depicting gating and frequencies of HBV pol₄₅₅-specific CD8⁺ T cells are displayed. (d) Serum HBV DNA in endogenous control/chronic HBV (n = 24) versus acute HBV (n = 8) and acute resolved HBV (n = 11) are depicted. (e) Correlation analyses of viral load and the frequency of HBV pol₄₅₅-specific CD8⁺ T cells expressing GZMB, ZEB2, EOMES obtained from individuals within endogenous control/chronic HBV (for GZMB and EOMES: n = 11; for ZEB2: n = 13) or acute HBV (n = 7) are depicted, respectively. Each dot represents one HBV pol₄₅₅-specific CD8⁺ T cell response with color coding as indicated: dark green, in endogenous control/chronic HBV; light green, acute HBV; dark gray, acute resolved HBV. Bar charts show the median value with IQR. Statistical analyses were performed via Kruskal–Wallis test including Dunn’s multiple comparisons test (b,c,d) and simple linear regression analysis (e).

Source data

Extended Data Fig. 5 Characteristics of HBV pol455-specific CD8⁺ T cells in endogenously controlled vs. acute. vs acute resolved HBV infection.

(a) Protein expression levels of the indicated cytotoxic- and exhaustion-associated markers in GZMB⁺ HBV pol₄₅₅-specific CD8⁺ T cells are plotted on t-SNE (endogenous control/chronic HBV, n = 12; acute HBV, n = 7; acute resolved HBV, n = 4). Scaled expression (blue, low; red, high) is depicted. (b) Expression levels of TCF-1 were examined in HBV pol₄₅₅-specific CD8⁺ T cells t-SNE (endogenous control/chronic HBV, n = 8; acute HBV, n = 4; acute resolved HBV, n = 5). Histograms and statistical graph depicting gating and frequencies of HBV pol₄₅₅-specific CD8⁺ T cells are displayed. (c) Regulon activity matrix obtained from single-cell transcriptomes of HBV pol₄₅₅-specific CD8⁺ T cells in endogenous control/chronic HBV (n = 3), acute HBV (n = 3) and acute resolved HBV infection (n = 2) after applying SCENIC algorithm. Each dot represents one HBV pol₄₅₅-specific CD8⁺ T cell response with color coding as indicated: dark green, in endogenous control/chronic HBV; light green, acute HBV; dark gray, acute resolved HBV. Bar charts show the median value with IQR. Statistical analyses were performed via Kruskal–Wallis test including Dunn’s multiple comparisons test (b).

Source data

Extended Data Fig. 6 TGFß pathway analyses of HBV-specific CD8⁺ T cells.

(a) Heatmap showing TGFβ-receptor gene expression in HBV core₁₈-specific and HBV pol₄₅₅-specific CD8⁺ T cells of endogenous control/chronic HBV (n = 7; HBVcore₁₈: n = 4, HBVpol₄₅₅: n = 3) and NUC control/chronic HBV (n = 7; HBVcore₁₈: n = 3, HBVpol₄₅₅: n = 4) (left). Scaled expression levels are color coded (blue, low; brown, high). GSEA analysis of the TGFβ signaling pathway in HBV core₁₈-specific versus HBV pol₄₅₅-specific CD8⁺ T cells of endogenous control/chronic HBV and NUC control/chronic HBV (right). The aggregated transcript counts are mapped in violin plots (box center line, median; box limits, upper and lower quartiles; box whiskers, IQR). (b) Feature plots showing CD8A, GZMB, TGFBR1 and TGFBR3 co-expression in CD8⁺ T cells sampled from blood and the liver by FNA (data from Genshaft et al., Hepatology, 2023). (c) IL-10 cytokine pathway gene expression analysis of HBV pol₄₅₅-specific CD8⁺ T cells obtained from endogenous control/chronic HBV (n = 3), acute HBV (n = 3) and acute resolved HBV infection (n = 2). The aggregated transcript counts are mapped in violin plots (box center line, median; box limits, upper and lower quartiles; box whiskers, IQR). (d) Dot plots showing T-BET and PRF1 expression of HBV pol₄₅₅-specific CD8⁺ T cells obtained from endogenous control/chronic HBV (n = 6) after 14 days of in vitro culture with (left) and without TGF-β RI Kinase Inhibitor II (right). Statistical graph depicting frequencies of T-BET+ or PRF1 + HBV pol_455-specific CD8⁺ T cells are displayed. (e) ZEB2, T-BET, EOMES, GZMB and PRF1 expression of HBV pol₄₅₅-specific CD8⁺ T cells obtained from acute HBV infection (n = 4) after 14 days of in vitro culture with and without TGF-β RI Kinase Inhibitor II is shown. (f) TGFβ concentration in plasma (EDTA) in endogenous control/chronic HBV (n = 26), acute HBV (n = 7) and acute resolved HBV (n = 7) was determined by ELISA and is depicted as pg/ml. Bar charts show the median value with IQR. Statistical analyses were performed via Kruskal–Wallis test including Dunn’s multiple comparisons test (a,c,f) and two-tailed Wilcoxon matched-pairs signed rank test (d,e).

Source data

Extended Data Fig. 7 Impact of TGFβ signalling inhibition on HBV pol455-specific CD8⁺ T cell function.

(a) Dot plots showing IFNγ-, TNF and CD107a-producing HBV pol₄₅₅-specific CD8⁺ T cells in response to peptide stimulation after 14 days of in vitro culture with (left) and without TGFβ RI Kinase Inhibitor II (middle). Negative control is depicted on the right. The percentages of IFN-γ-, TNF and CD107a-producing HBV pol₄₅₅-specific CD8⁺ T cells were determined in relation to the frequency of HBV pol₄₅₅-specific CD8⁺ T cells for individuals with endogenous control/chronic HBV. (b) The polyfunctionality of HBV pol₄₅₅-specific CD8⁺ T cells were assessed by gating on IFN-γ and TNF positive cells. Each dot represents one of the respective HBV pol₄₅₅-specific CD8⁺ + T cell responses with color coding as indicated: dark green, endogenous control/chronic HBV, n = 6. Statistical analyses were performed via two-tailed Wilcoxon matched-pairs signed rank test (a,b).

Source data

Extended Data Fig. 8 Effects of IL12 and IFNα on the attenuated HBV pol₄₅₅-specific CD8⁺ T cell response.

(a) IL-2, IL-12 and IFNα cytokine pathway gene expression analysis of HBV pol₄₅₅-specific CD8⁺ T cells obtained from endogenous control/chronic HBV (n = 3), acute HBV (n = 3) and acute resolved HBV infection (n = 2). The aggregated transcript counts are mapped in violin plots (box center line, median; box limits, upper and lower quartiles; box whiskers, IQR). (b+c) PRF1 expression of HBV pol₄₅₅-specific CD8⁺ T cells after 14 days of in vitro culture with (b) IL-12 (endogenous control/chronic HBV n = 7) or (c) IFNα treatment (endogenous control/chronic HBV n = 8) or without any treatment is shown (d) Protein expression levels of the indicated cytotoxic and effector-associated markers in HBV pol₄₅₅-specific CD8⁺ T cells are plotted. Scaled expression (blue, low; red, high) is depicted. Each dot represents one of the respective HBV pol₄₅₅-specific CD8⁺ T cell responses with color coding as indicated: dark green, endogenous control/chronic HBV. Statistical analyses were performed via Kruskal–Wallis test including Dunn’s multiple comparisons test (a) and two-tailed Wilcoxon matched-pairs signed rank test (b,c).

Source data

Supplementary information

Supplementary Information

Supplementary Figs. 1–3 and legends and Supplementary Tables 1–6.

Reporting Summary

Peer Review File

Source data

Source Data Fig. 1

Used patient source data.