CD8+ T cells in hepatitis B cirrhosis exhibited an altered phenotype
To investigate the changes of CD8+ T cell phenotype in patients with LC, CD8+ T cell subset mature markers (CD45RA and CCR7), activated markers (CD38 and HLA-DR), immunosuppressed markers (PD-1 and Tigit), cytotoxic molecules (perforin and granzyme B) and level of apoptosis were analyzed in HC, CHB and LC groups. A total of 50 patients with hepatitis B cirrhosis, 76 patients with chronic hepatitis B and 56 healthy subjects were enrolled. Firstly, the results showed that LC group displayed a substantially reduced proportion of CD8+ T cells in comparison to HC group (Figure 1A). Additionally, according to the expression of CD45RA and CCR7, CD8+ T cells were divided into naive cells (CD45RA+CCR7+), central memory cells (TCM; CD45RA−CCR7+), effector memory cells (TEM; CD45RA−CCR7−), and terminally differentiated effector memory cells (TEMRA; CD45R+CCR7−), then it was noted that CD8+ T cells in LC and CHB groups primarily exhibited TCM, whereas HC group predominantly had naïve CD8+ T cells (Figure 1B). Furthermore, in CD8+ T cells, LC group showed elevated levels of CD38, HLA-DR, granzyme B, perforin, PD-1, Tigit, and markers of apoptosis compared with those in HC and CHB groups (Figure.1C, D, E, F). Lastly, the expression of the proliferation marker Ki-67 was further examined, and although the results were not statistically significant, the expression of Ki-67 in CD8+ T cells was higher in LC patients than that in HC and CHB patients (Figure.1G). The secretion of IFN-γ, TNF-α and IL-2 in LC group were significantly higher than those in the HC and CHB groups. These findings collectively suggested significant alterations of CD8+ T cell phenotype in LC group compared to HC and CHB groups.
CD38+HLA-DR+CD8+ T Cells Correlated with Liver Injury in Hepatitis B Cirrhosis
As phenotypic changes in CD8+ T cells were correlated with the activation status in LC groups, we directed our analysis towards CD38+HLA-DR+CD8+ T cells, also known as bystander-activated T cells, (6-10). Firstly, the proportion of CD38+HLA-DR+CD8+ T cells in peripheral blood of these three cohorts was compared. The results showed that the proportion of CD38+HLA-DR+CD8+ T cells was significantly elevated in LC group when compared to CHB and HC groups (HC vs. CHB vs. LC: 7.23%vs. 23.35% vs.34.8%) (Figure 2 A, B, C). Analogous bystander-activated CD8+ T cells (CD38+MHC-II+CD8+ T cells) in mice showed similar results in a mouse liver fibrosis model (Figure S1).
Subsequently, the relationship between CD38+HLA-DR+CD8+ T cells and liver injury was investigated. Our findings suggested a significant correlation between the proportion of CD38+HLA-DR+CD8+ T cells and serum ALT and AST levels (Figure2 D, E). Additionally, the association between percentage of CD38+HLA-DR+CD8+ T cell and inflammation was assessed by analyzing the relationship between these T cells and C-reactive protein (CRP) levels, revealing a higher percentage of CD38+HLA-DR+CD8+ T cells in LC group with abnormal CRP levels (Figure 2F). However, the percentage of CD38+HLA-DR+CD8+ T cells did not correlate with the levels of γ-GT and ALP (Figure S2A, B). In contrast, the proportion of CD38+HLA-DR+CD8+ T cells was significantly correlated with the value of TBIL and albumin (Figure S2C, D). Taken together, these data implied the heightened expression of CD38+HLA-DR+CD8+ T cells in LC and their association with liver injury.
CD38+HLA-DR+CD8+T cells comprised a subset of T cells that contained both HBV specific T cells and non-HBV specific T cells
To clarify whether the increase of CD38+HLA-DR+CD8+ T cells in patients with chronic hepatitis B cirrhosis was specifically activated by the antigen, HLA-A2-restricted pentamer staining was performed to detect HBV-specific cells in peripheral blood CD38+HLA-DR+CD8+ T cells from CHB and LC groups (Figure 3A). Although the results were not statistically significant, HBV-specific cells were higher in LC than in CHB (CHB vs. LC: 1.2% vs. 1.89%) (Figure 3B). Additionally, the activation status of virus specific CD8+ T cells other than HBV in CD38+HLA-DR+CD8+ T cells, including EBV and CMV was detected (Figure 3C, E). Our findings revealed that CD38+HLA-DR+CD8+ T cells from LC group expressed elevated levels of EBV-specific (HC vs. CHB vs. LC: 0.38% vs. 1.03% vs.2.03%) and CMV-specific cells (HC vs. CHB vs. LC:0.57% vs.1.44% vs.2.48%) in comparison to HC and CHB groups (Figure 3D, F). This observation suggested that these cells underwent antigen-independent activation in the absence of cognate antigen during LC. In addition, there was no significant difference between pentamer+CD38+HLA-DR+CD8+ T cells and pentamer+CD38-HLA-DR-CD8+ T cells for both HBV-specific and non-HBV-specific CD8+ T cells, indicating that the increase of CD38+HLA-DR+CD8+ T cells in LC patients may not contribute much to the virus specific CD8+ T cell pool (Figure 3G, H, I). Taken together, HBV specific CD8+ T cells and HBV unrelated virus specific CD8+ T cells exhibited an activated phenotype during LC.
Transcriptional analysis of CD38+HLA-DR+CD8+ T cells and CD38-HLA-DR-CD8+ T cells from LC group
In order to further analyze the functional differences between CD38+HLA-DR+CD8+ T cells and CD38-HLA-DR-CD8+ T cells, a comparative analysis of transcriptome alterations was conducted. We sorted these two types of cells from LC groups and conducted RNA seq analysis. Transcriptional analysis revealed 451 significantly DEGs between paired CD38+HLA-DR+ and their corresponding CD38-HLA-DR-, among which 177 genes were upregulated and 274 genes were downregulated (Figure 4A). Further analysis revealed that CD38+HLA-DR+CD8+ T cells exhibited high expression of activation markers (CD38, HLA-DR, CD74, CD156, FLNA, ITGAM, LEF1), exhaustion markers (TIM3), proliferation markers (MKI67) and effector molecules (PRF1, TGFB1, CD107a). Notably, transcriptional regulators RUN3, TBX21 and EOMES, which regulated the secretion of perforin and Granzyme B, were up-regulated in CD38+HLA-DR+CD8+ T cells. In additional, upregulation of natural killer (NK) cell cytotoxicity genes, such as PRF1 and ITGAL, were observed in CD38+HLA-DR+CD8+ T cells, which might provide them with innate immunity (Figure 4B upper). Cytokines and related receptors genes, including TGFB1, TNFRSF1B and FASLG, were upregulated in CD38+HLA-DR+CD8+ T cells (Figure 4B middle). The higher expression of chemokine and related receptors such as CCR5 and CCL5 suggested enhanced migration to inflamed tissues of CD38+HLA-DR+CD8+ T cells (Figure 4B lower).
GO analysis consisted of biological process, cellular component and molecular function. For our study, we only enriched the biological process of GO analysis between these two subsets of T cells. Our results revealed that CD38+HLA-DR+CD8+ were mainly enriched of “positive regulation of cell activation, “cell activation”, “regulation of secretion”, “secretion”, “regulation of exocytosis”, and “T cell mediated immunity” (Figure 4C). Given that CD38+HLA-DR+CD8+ T cells exhibited a significant increase in activity and effector molecular markers in comparison to CD38-HLA-DR-CD8+ T cells, our subsequent experiments would be primarily directed towards studying the effector function of CD38+HLA-DR+CD8+ T cells.
Function of CD38+HLA-DR+CD8+ T cells in patients with LC
Next, we investigated the function of CD38+HLA-DR+CD8+ T cells. Firstly, CD38+HLA-DR+CD8+ T cells were predominantly expressed as TEM and TEMRA cells in LC patients, whereas in HC groups, they were mostly expressed as naive cells (Figure 5A, B).
Then, we evaluated levels of apoptosis and proliferation of CD38+HLA-DR+CD8+ T cells and found that LC patients expressed higher levels of Ki-67 and had a higher proportion of apoptosis (Figure 5C, D).
Subsequently, the expression of immunosuppressive markers, including PD-1 and Tigit, was explored. It was observed that the proportion of PD-1 and Tigit in CD38+HLA-DR+CD8+ T cells of LC group was significantly elevated when compared to that of HC group (Figure 5E).
To further elucidate the functional attributes of CD38+HLA-DR+CD8+ T cell, we investigated the expression of cytotoxic molecules, including perforin, granzyme B, IFN-γ, IL-2 and TNF-α. It was evident that the proportion of granzyme-B+, perforin+ in CD38+HLA-DR+CD8+ T cells of LC group was significantly higher than that of CHB and HC groups (Figure 5F). Furthermore, the secretion of IFN-γ, TNF-α and IL-2 in LC group were significantly higher than those in the HC and CHB groups (Figure 5G). Besides this, CD38+HLA-DR+CD8+ T cells demonstrated a notably higher capacity to produce IFN-γ than CD38-HLA-DR-CD8+ T cells (Figure 5G). Collectively, these results suggested that CD38+HLA-DR+CD8+ T cells were mainly TEM and TEMRA and exhibited heightened secretion of cytotoxic-related cytokines.
CD38+HLA-DR+CD8+ T cells were activated by IL-15 and expressed NKG2D
Next, we investigated the factors contributing to the increased frequency of CD38+HLA-DR+CD8+ T cells in LC group. We hypothesized that known cytokines in the inflammatory environment might drive the activation of CD38+HLA-DR+CD8+ T cells, which occurred independently of antigens. Consequently, an in vitro investigation to discern which cytokines could activate CD8+ T cells from HC group in the absence of antigenic stimulation was conducted. The expression of CD38+HLA-DR+CD8+ T cells from healthy human CD8+ T cells treated with or without IL-2, IL-15, IL-18 or LPS for 1 day, 2 days, 3 days was examined, and the results showed that the longer the IL-15 stimulation, the more CD38+HLA-DR+CD8+ T cells were expressed (Figure 6A, B). These results highlighted that IL-15 as the main cytokines driving non-HBV specific CD8+ T cell activation. However, the presence of IL-15 was not found in the plasma, which possible was due to the patients not being in the acute phase. Beside this, an augmented activation level in CD8+ T cells stimulated with IL-2 was observed, which might be related to the proliferative effect of IL-2 on effector T cells.
Our previously results identified that these cells exhibited innate immune-like functions through RNAseq and phenotypic expression research.
Then, the expression of NKG2D and NKP30, as the natural cytotoxicity receptors, in CD38+HLA-DR+CD8+ T cells stimulated with IL-15 was examined in HC group. We found that mean florescence intensity (MFI) and frequency of NKG2D but not NKP30 were elevated, although not statistically significant, following IL-15 stimulation (Figure 6C, D).
We further examined the expression of NKG2D and NKP30 on CD38+HLA-DR+CD8+ T cells and CD38-HLA-DR-CD8+ T cells in LC group (Figure.6E, G). Although there was no significant difference in the frequency of NKP30+ cells between the two groups, the frequency of NKG2D+ cells, MFI of NKG2D or NKP30 on CD38+HLA-DR+CD8+ T cells was higher than that on CD38-HLA-DR-CD8+ T cells (Figure 6F, H). We also found that LC patients expressed higher levels of NKG2D or NKP30 (Figure 6I, J). Taken together, these results suggested that IL-15 could induce NKG2D-dependent bystander activation of CD8+ T cells in LC.
NKG2D expression contributed to the TCR-independent cell-killing function of bystander-activated CD8+ T cells
Our previous data demonstrated that CD38+HLA-DR+CD8+ T cells exhibited a phenotype of innate immunity and overexpress NKG2D, which mediated the natural killer cell cytotoxicity. Based on this, we evaluated whether CD38+HLA-DR+CD8+ T cells stimulated with IL-15 in HC group triggered cytolytic activity in the absence of TCR involvement. To this end, K562 cells were selected as the target cells for CD8+ T cell-mediated cytotoxicity assay since they lacked HLA Class I molecule (Figure 7A). We found that K562 cells were lyzed by IL-15-activated CD8+ T cells, suggesting that these CD8+ T cells exhibited NK cell-like killing activity that did not require TCR involvement (Figure 7B).
Subsequently, to evaluate whether the cytotoxicity of CD38+HLA-DR+CD8+ T cells targeting the target cells depended on direct cell-to-cell contact or other soluble factors, we employed a 0.4 μm transwell plate, which allowed cytokines to pass through while preventing the passage of cells. Physically separating T cells and target cells abolished cytotoxicity, indicating that the killing activity of CD38+HLA-DR+CD8+ T cells were dependent on direct cell-to-cell contact rather than soluble factors (Figure 7C, D).
Lastly, the antibodies were used to block the killing effect of K562, thereby identifying the receptor responsible for the innate killing activity of CD8+ T cells activated by IL-15. Anti-NKG2D, NKG2DL or anti-perforin antibodies abrogated the K562 cytolytic activity of CD8+ T cells activated by IL-15 (Figure 7E). The results showed that CD8+ T cells activated by IL-15 exerted natural cytotoxic effects in an NKG2D dependent manner.