ScRNA-seq Identifies High CXCL8 Linked to Liver Fibrosis in HBeAg-negative Chronic Hepatitis B Patients With Low HBsAg

Abstract

Background and Aims

Chemokine ligand 8 (CXCL8) is suggested as a potential biomarker in assessing the progression and severity of hepatitis B e antigen–negative treatment-naive patients with “chronic hepatitis”. Analysis of CXCL8 expression in the liver and blood of such patients with low and high hepatitis B surface antigen (HBsAg) levels could provide insights into the immunopathogenesis of hepatitis B virus–related liver disease.

Methods

Thirty-five hepatitis B e antigen–negative treatment-naive patients with “chronic hepatitis” patients were classified into Group I with HBsAg levels below 2000 IU/mL (HBsAg^low, n = 12) and Group II with HBsAg levels > 2000 IU/mL (HBsAg^high, n = 23); both groups had elevated alanine aminotransferase levels (>1.2 × upper limit of normal). Histopathological analysis and plasma cytokine profiling were conducted in all 35 patients, while single-cell RNA sequencing was performed on peripheral blood mononuclear cells and intrahepatic immune cells in 6 patients.

Results

HBsAg^low patients had higher median fibrosis scores 2(0.75–3)) compared to HBsAg^high 0(0–1). HBsAg^low were associated with increased plasma CXCL8 levels (P = .027) and higher CXCL8 expression, which correlated strongly with fibrosis scores (P = .008). Single-cell RNA sequencing of blood and liver biopsies also showed increased CXCL8, CCL2, and IL1β-expressing monocytes and macrophages in HBsAg^low. In addition, HBsAg^low patients showed distinct peripheral CD8_C2 cluster expressing GNLY, GZMH, GZMB, GZMA, ITGB1, CX3CR1, ADGRG1, CCL4, and TGFBR3 contributing to activation, cytotoxicity, and fibrosis.

Conclusion

Elevated CXCL8 expression in plasma, peripheral, and intrahepatic mono-mac populations, along with distinct CD8 T-cell clusters, is associated with greater fibrosis in patients with low HBsAg levels.

Graphical abstract

Introduction

Over 296 million people globally are chronically hepatitis B virus (HBV)–infected with the highest prevalence and approximately 820,000 deaths annually in the Asia-Pacific region and Africa.¹

The progression of chronic HBV infection follows a highly intricate and nonlinear course, transitioning through 5 distinct phases: 1) hepatitis B e antigen (HBeAg)–positive chronic infection, 2) HBeAg-positive chronic hepatitis, 3) HBeAg-negative chronic infection, 4) HBeAg-negative chronic hepatitis, and 5) the hepatitis B surface antigen (HBsAg)–negative phase.^2,3 However, one-third of patients do not fit in these classically defined phases and the phase remains indeterminate, which can then transition between phases and experience concurrent liver injury.⁴

Serum HBsAg level has been shown to vary with phase of chronic HBV infection, to be clinically useful for distinguishing HBeAg negative chronic infection from chronic hepatitis,^5,6 and predict disease progression in patients with low HBV loads.⁷ Among HBeAg negative patients with HBV DNA level below 2000 IU/ml, subjects with HBsAg >1000 IU/ml are at higher risk of hepatocellular carcinoma and chronic hepatitis than the group with low HBsAg.⁸

Nevertheless, about two-thirds of HBeAg-negative chronic HBV infection have an HBsAg level >1000 IU/ml. However, HBV DNA level fluctuates easily, whereas HBsAg level remains relatively stable.⁹ Serum HBsAg levels reflect the interplay between the virus and the immune system.^8,10 HBsAg levels decline during active hepatic inflammation. Lower serum HBsAg levels are therefore considered to reflect a longer duration of HBV-related liver disease.¹¹ Previous studies have suggested that lower HBsAg levels are associated with more severe hepatic fibrosis; the effect being more pronounced in HBeAg-positive subjects.¹⁰ Most adult patients in India are HBeAg negative, and these patients may have varying HBsAg levels and liver fibrosis stages.¹²

The role of chemokine ligand 8 (CXCL8) was found to be crucial in mediating liver damage in chronic hepatitis B (CHB) patients.¹³ CXCL8 is a proinflammatory chemokine that facilitates the recruitment of immune cells to sites of infection or tissue injury, including the liver. It is produced by various liver cell types, thereby contributing to both inflammation and the fibrotic process in chronic liver diseases.¹⁴ Interleukin-8 is activated in patients with chronic liver diseases and associated with hepatic macrophage (Mac) accumulation in human liver fibrosis. Furthermore, CXCL8 promotes fibrosis by activating hepatic stellate cells, which are pivotal in producing extracellular matrix components, thus playing a central role in the progression of liver fibrosis.¹⁵

Despite advancements in our understanding of the liver environment during different chronic phases, there remains a notable gap in investigating disease-associated cell states specifically in treatment-naive HBeAg-negative chronic hepatitis with different HBsAg levels. This highlights a critical need for detailed, high-resolution studies focused on identifying key cell types, cytokines, and their interactions that could elucidate the underlying mechanisms responsible for fibrosis and cirrhosis in treatment-naive HBeAg-negative chronic hepatitis patients.

Methods

Patient Recruitment

From 2019 to December 2021, 35 HBeAg-negative treatment-naive patients with “chronic hepatitis” (nCHBV) (with no history of any antiviral treatment) were enrolled in this study at the Institute of Liver and Biliary Sciences, New Delhi. The inclusion criteria included the presence of detectable HBsAg and HBV DNA for ≥ 6 months before recruited in patients between 18-60 years and consented to the study. Exclusion criteria involved individuals with significant liver-related complications such as decompensated liver cirrhosis, hepatocellular carcinoma, a history of liver transplantation, or coinfections with hepatitis C virus or HIV.

All details of methods are mentioned in the supplementary information.

Results

Demographic and Clinical Parameters

Thirty-five nCHBV (31 males and 4 females, median age 32 [27–41] years) were grouped based on their HBsAg levels; group I: HBsAg^low (n = 12, HBsAg < 2000 IU/ml) and group II: HBsAg^high (n = 23, HBsAg > 2000 IU/ml). Both the groups had elevated alanine aminotransferase levels (>1.2 × upper limit of normal [ULN]).

The clinical parameters of the patients are listed in Supplementary Table 2. There was no significant difference in both groups except HBsAg levels and fibrosis scores (P < .05). Platelets were found to be low in HBsAg^low patients compared to HBsAg^high, however, this difference was not significant (Supplementary Table 2). All patients were of HBV genotype D. All HBsAg^low and HBsAg^high patients were HBeAg negative with raised alanine aminotransferase >30IU/ml (ULN) for males, 19 IU/mL (ULN) for female’s IU/mL.

Increased Levels of CXCL8, CXCL11, and CXCL12 in HBsAg^low Patients

In HBsAg^low patients with nCHBV, hematoxylin and eosin staining showed increased immune infiltration and Masson's trichrome stain highlights fibrotic changes in liver tissues, with significant liver fibrosis score (Ishak score) (Figure 1A–1C) than HBsAg^high patients (Supplementary Table 3). The proportion of HBsAg^low patients with mild inflammation and severe fibrosis was more than HBsAg^high.

Figure 1.

Histopathology score and levels of circulating plasma cytokines in naive chronic HBV patients. (A) HE staining highlighting tissue architecture and MT staining of liver tissue showing collagen deposition in Lo and Hi patients. (B) Quantification of MT staining or fibrosis. (C) Principal component analysis showing a clear separation between Lo and Hi. (D) Volcano plot showing the significantly up (red) and downregulated (blue) cytokines in Lo patients. (E) Bar plot showing the levels of plasma analytes in Lo patients compared to Hi. Statistical significance is indicated by asterisks: ∗P < .05, ∗∗P < .01. Data are presented as mean ± standard error of mean, with Mann-Whitney U test conducted for group comparisons. HE, Hematoxylin-eosin; Hi, HBsAg^high; I-TAC, interferon-inducible T-cell alpha chemoattractant; IL-9, interleukin 9; IL-8, interleukin 8; Lo, HBsAg^low; MCP-3, monocyte chemoattractant protein 3; MT, Masson's trichrome.

Further, proinflammatory plasma chemokine and cytokines like CXCL8 (interleukin 8), I-TAC (CXCL11), and SDF-1α (CXCL12) associated with fibrosis were significantly high but interleukin 9 and monocyte chemoattractant protein 3 were low in HBsAg^low compared to HBsAg^high (Figure 1D and 1E).

Bulk Transcriptome of Peripheral Blood Mononuclear Cells and Intrahepatic Mononuclear Immune Cells Revealed Immune Cell Activation in HBsAg^low Patients Compared to HBsAg^high

To decipher the fibrosis-associated transcriptomic alterations and immune cell composition, we have performed single-cell transcriptome profiling of peripheral blood mononuclear cells (PBMCs) and intrahepatic immune cells (Supplementary Figure 1).

After quantitative and qualitative assessments (Supplementary Figures 2 and 3), we successfully captured 28,000 PBMCs and 4296 intrahepatic mononuclear cells. Bulk RNA sequencing identified 1160 differentially expressed transcripts in PBMCs and 1357 in intrahepatic immune cells (Figure 2A). In PBMCs, 7 genes (PRIM1, FOS, MNDA, KLF4, TXNIP, DUSP6, and SELL) were significantly upregulated, while 20 genes were downregulated in HBsAg^low compared to HBsAg^high patients (Figure 2B). Similarly, the intrahepatic immune transcriptome revealed 31 upregulated genes, including DUSP6, FOS, IL1RN, IL1β, CCL4, CXCL3, CCL3, CXCL2, and CXCL8, while 8 genes (TAGLN2, EIF5A, CD37, IGHM, SH3BGRL3, and CD79A) were downregulated (Figure 2B and 2C, Supplementary Table 4).

Figure 2.

Transcriptome signatures of PBMCs and intrahepatic cells in HBsAg^low patients. (A) Bar graph depicting number of genes and average log fold change of genes in (B) PBMCs and (C) intrahepatic cells in HBsAg^low patients compared to HBsAg^high. (D, E) Violin plot showing fold change of commonly expressed genes in PBMCs and intrahepatic cells.

A comparative analysis of PBMCs and intrahepatic immune cells revealed FOS and DUSP6 as commonly upregulated genes, while IGHM, SH3BGRL3, CD37, EIF5A, TAGLN2, and LMNA were downregulated in both compartments (Figure 2D and 2E, Supplementary Table 4). This analysis reveals immune alterations in nCHBV across peripheral and intrahepatic immune compartments.

Increased Myeloid and CD8 T-Cell Population in HBsAg^low Patients

Leiden-based clustering determined 22 distinct major clusters in PBMCs, representing CD4, CD8, other T, B, natural killer (NK), monocytes, and dendritic cells (DCs) on the behalf of cell-specific markers (Figure 3A–3D). In healthy, most abundant cells were CD4 (52%), followed by B (13%), monocytes (13%), CD8 (11%), and NK (9%). DC (<1%) (Figure 3E). However, in nCHBV, this proportion was altered in context of HBsAg load. In HBsAg^low patients, there was a striking rise in monocyte population in HBsAg^low (healthy 13% vs HBsAg^low 33% vs HBsAg^high 29%), CD8 T cell population 11% vs 25% vs 19% and other T cells 1% vs 2% vs 6% (Figure 3E). But there was marked decrease of CD4 cell population 52% vs 28% vs 26%, B cell population 13% vs 4% vs 11% and NK cells population 9% vs 8% vs 7% (Figure 3E).

Figure 3.

Characterization and proportion of peripheral and intrahepatic mononuclear immune cells in nCHBV. (A) t-SNE plot displaying the condition-wise annotation of single-cell RNA data from the HC (gray), Lo (red) and Hi (blue) in PBMCs and liver (B) t-SNE plot projection of c1–c22 clusters in PBMCs and c1-c8 in intrahepatic. (C) Annotation of Leiden clusters based on the reference-based markers identified immune cell types like B cells, CD4 T, CD8 T, Other T cells, NK cells, mono-mac and DC. (D) Percentage-based proportionality bar graph representing each cell type in PBMCs and liver in each group. Black arrows indicate notable changes in the proportion of specific immune cells. (E) Heat map of gene expression in different cell types in PBMCs and intrahepatic. HC, healthy control; Hi, HBsAg^high; Lo, HBsAg^low; t-SNE, t-distributed stochastic neighbor embedding.

Similarly, liver-infiltrating mononuclear immune cells formed 8 clusters, including CD4, CD8, other T cells, B cells, NK cells, Macs, and DCs (Figure 3A–3D). HBsAg^low showed increased Mac (healthy 3% vs HBsAg^low 11% vs HBsAg^high 2%), CD8(20% vs 27% vs 2%), NK (19% vs 25% vs 14%), DCs (4% vs 12% vs 7%), and B cells (3% vs 4% vs 15%) compared to healthy and HBsAg^high individuals (Figure 3E). CD4 T cells remained similar in healthy (16%) and HBsAg^low (17%) but increased in HBsAg^high (30%), while other T cells were significantly reduced in HBsAg^low (35% vs 4% vs 10%) (Figure 3E). Therefore, we have observed the HBsAg-dependent immune shifts in nCHBV, with increased monocytes and CD8 T cells in PBMCs and intrahepatic compartments.

Heterogenous Myeloid Compartment in HBsAg^low Patients

Leiden-based clustering of mono-mac revealed 15 distinct clusters (M_c0-c14) in the periphery (Figure 4A and 4C, Supplementary Figure 6). Based on cluster of differentiation 14 (monocyte marker) (CD14) and 16, cell typing of these clusters was performed (Figure 4D). Out of the 15 clusters, there was a remarkable difference in clusters 0, 1, 5, and 6 in nCHBV and healthy patients (Figure 4B). Clusters c1 and c6 were present in the HBsAg^low but disappeared in HBsAg^high; on the contrary, clusters c0 and c5 were missing in HBsAg^low. This variation highlights distinct monocyte responses influenced by antigen load.

Figure 4.

Comparative single-cell transcriptomic profiling of mono-mac (monocytes and Macs). (A) t-SNE projection of mono-mac in periphery and intrahepatic immune cells. (B) UMAP projections of combined (HC, HBsAg^high, and HBsAg^low) mono-mac showing distinct cluster in peripheral monocytes (M_c0–c14) and group-wise annotation of mono-mac subclusters with numbers reflecting scanty or missed clusters. (C) Percentage-based proportionality bar graph representing each cell cluster. (D) Annotated UMAP plot using CD14+ CD16+ mono markers (M_c0–c14). (E) Heatmap of gene expression patterns (Log2 fold change) across peripheral monocytes (M_c0,1,5,6) and intrahepatic Macs (Mø_c0–c1), with color intensity, indicating expression levels from high (red) to low (purple). (F) Violin plots showing Log2 fold change of commonly expressed genes in periphery and intrahepatic mono-macs. HC, healthy control; mono, monocytes; UMAP, uniform manifold approximation and projection.

Abundantly, present M_c0 (41%) and M_c5 (8%) clusters in HBsAg^high (Figure 4C) had high expression of NINJ1, IER3, C15ORF48, SERPINB9, and RSRP1, XIST, MALAT1, UBXN11, S100A9, RHOB, CYBA, CD14, PTPN6, RGS2, TMSB10, NEAT1, JUNB, AC103702.2 associated with cell stress and apoptosis (Figure 4E). While M_c1 (30%) and M_c6 (7%) which are remarkably present in HBsAg^low, showed high expression of CCL3, CCL3L1, CCL20, TNFAIP6, TEX14, EREG, CXCL2, CCL4, IL1β, CXCL8, SELL, FOS, PRIM1, CREB1, AL592429.2, ZFP36 genes (Figure 4B and 4C) involved in cell recruitment leading to inflammation, cell apoptosis at the injury site. Intrahepatic Macs (Figure 4A) also showed higher expression of genes like CXCL8, CCL20, IL1β, and TNFAIP6 (Figure 4E).

We found many genes that were commonly expressed in both periphery and liver mono-macs, like CXCL2, CXCL8, CCL20, TNFAIP6, IL1β, RHOB, S100A9, and TEX14. In peripheral monocytes, CCL2, CCL8, and IL1β showed higher expression in HBsAg^low compared to HBsAg^high. Intrahepatic Mac compartment also showed higher expression of CXCL8, CCL20, CCL3, TNFAIP, TEXA14, and IL1B, in HBsAg^low compared to HBsAg^high (Figure 4F). This gene signature reflects immune cell chemotaxis, adhesion, and activation. Notably, CXCL8 is mainly secreted by CD14+ Mac which facilitates CD4, CD8, and NK immune cell trafficking in HBsAg^low patients' livers. Moreover, CCL4, CCL20, IL1β, and CXCL8 genes act as inflammatory mediators that attract and activate immune cells and are significantly associated with fibrosis.

Additionally, we observed that most specialized peripheral antigen-presenting DCs were not much different in HBsAg^low and HBsAg^high, except that there was the appearance of one cluster, DC_c2, which was not present in healthy (Supplementary Figure 5B). DC_c2 cluster (6% in HBsAg^high and 10% in HBsAg^low, Supplementary Figure 5C) specifically had genes related to antigen presentation, immune metabolic regulation, transport functions, and inflammation due to HBV infection (Supplementary Figure 5D). However, we did not observe any difference in intrahepatic DC compartment.

Distinct Change in CD4, CD8, NK, and B Cells in HBsAg^low Patients

When we analyzed CD8 compartment in periphery and intrahepatic (Figure 5A), we observed a remarkable difference in the peripheral CD8 T cell clusters (CD8_c0–CD8_c6) in patient group and healthy (Figure 5B). Out of 7 peripheral CD8 clusters, clusters 0, 2, 3, 4, and 6 had altered abundance in patients and healthy control (Figure 5B and 5C). With CD8 T cell differentiating general markers, clusters were designated as naive, T central memory, and T effector memory (Figure 5D). We have observed an increased T effector memory population in HBsAg^low compared to HBsAg^high (Figure 5D). The most abundant CD8 cluster was CD8_c0 (42%) in HBsAg^low (Figure 5C) with high expression of LEF1, MAL, SELL, CCR7, LRRN3, BACH2, IL6ST, NELL2 and TCF7 genes (Figure 5E) associated with adhesion, T-cell differentiation, and development. Further, CD8_c2 cluster was also more (34%) in HBsAg^low with high expression of granulysin (GNLY), GZMH, GZMB, GZMA, ITGB1, CX3CR1, ADGRG1, CCL4, TGFBR3 associated with cell adhesion, activation cytotoxicity and fibrosis (Figure 5E).

Figure 5.

Comparative single-cell transcriptomic profiling of CD8 T in blood and liver of healthy, Hi and Lo (A) tSNE plot showing CD8 T cells in peripheral and intrahepatic (B) UMAP projections of combined (HC, Hi^, and Lo) CD8 showing distinct cluster in periphery and group-wise annotation of CD8_c0–c6 clusters and circles showing scanty or missed clusters. (C) Percentage-based proportionality bar graph representing each cell cluster. (D) Annotated UMAP plot showed subtype of naive, TCM and TEM CD8 T cells. (E) Heatmap of gene expression patterns (Log2 fold change) across peripheral CD8_c0–c6 and intrahepatic CD8ø_c0–c1 with color intensity, indicating expression levels from high (red) to low (purple). (F) Violin plots showing Log2 fold change of commonly expressed genes in periphery and intrahepatic CD8 T cells. HC, healthy control; Hi, HBsAg^high; Lo, HBsAg^low.

On the contrary, CD8_c3, 4, and 6 clusters (16%,10%, and 5%) were densely presented in HBsAg^high compared to healthy and HBsAg^low with significantly high expression of CCR7, NKG7, CST7, RTKN2, SLC7A5, LDLR, mznf331, LMNA, and MYADM genes involved in immune cell migration, transcriptional regulation, transporters, and metabolism.

Intrahepatic CD8 T cells had high expression of ZFP36, XCL1, NR4A1, PMAIP1, HSP90AA1, NFKBIZ, DUSP2, NFKBIA, FOSB, FGFBP2, GNLY, TXNIP, PDE3B, GZMH, PXN, and SELL genes associated with effector T cells, infiltration, inflammation, and apoptosis (Figure 5E).

Peripheral and intrahepatic CD8 T cells commonly showed the upregulation of SELL, GZMH, GNLY, and FGFBP2 in HBsAg^low patients (Figure 5F). These genes are associated with CD8 T cells activation and cytotoxic function against viral infected cells. Further, single-cell RNA sequencing data revealed a distinct population of NK cell clusters (NK_c0-c4) in PBMCs (Supplementary Figure 8A) with a remarkable difference in cluster 1 which was present in healthy and HBsAg^low but diminished in HBsAg^high (Supplementary Figures 8C–E). NK_c1 cluster of HBsAg^low showed high expression of GBP1, CCL4, XAF1, RSAD2, TIGIT, CCND2, CCL3, PARP9, PRIM1 genes (Supplementary Figure 8F) involved in cytotoxic and fibrosis.

Intrahepatic NK cells exhibited elevated expression of FGFBP2, CX3CR1, GNLY, GZMB, GZMH, MMP3, PRF1, XCL1, NFKB1, XCL2, ZNF331, and IRF8, genes linked to cytotoxic activity, cellular infiltration, inflammation, and apoptosis (Supplementary Figures 8B and 8F).

Additionally, both peripheral and intrahepatic NK cells in HBsAg^low patients consistently showed upregulation of FGFBP2, SPON2, XCL1, and XCL2 (Supplementary Figure 8G). These genes are associated with NK cell adhesion and migration, which may enhance immune cell interactions and contribute to antiviral defense mechanisms.

Discussion

Patients with HBsAg^low exhibit low viral replication with notable immune activation.¹⁶ The distinctive features of the liver cannot be completely captured by only profiling the readily accessible plasma and PBMC samples. Therefore, to fill this knowledge gap, we conducted comprehensive single-cell profiling of both intrahepatic mononuclear cells and matched PBMCs from chronic hepatitis B HBeAg-negative, treatment-naive HBsAg^low and HBsAg^high patients. Increased levels of circulating chemokines such as CXCL8, CXCL11, and CXCL12 contributed concentration gradient to immune cell infiltration and activation in liver injury.^16,17

Total transcriptomic analysis in HBsAg^low PBMCs identified the significant upregulation of FOS, TXNIP, KLF4, MNDA, DUSP6 and PRIM1, which promote immune regulation, inflammation, cellular stress, and cytotoxicity responses.18, 19, 20, 21, 22 These gene expression signatures suggest a heightened state of immune infiltration and activation in HBsAg^low patients.

Immune dysregulation in liver fibrosis is observed in context of LMNA, SH3BGRL3, TAGLN2, EIF5A, CD37, and SLC7A5 gene expressions by various correlative studies. Downregulation of LMNA, SH3BGRL3, TAGLN2, EIF5A, CD37, and SLC7A5 impairs cellular integrity, cytoskeletal dynamics, and metabolic homeostasis, further exacerbating fibrosis.23, 24, 25, 26, 27

At the same time, upregulated expression of FOS, TXNIP, MNDA, DUSP6, and PRIM1 in HBsAg^low patients PBMCs promotes inflammation, oxidative stress, hepatocyte proliferation, immune modulation, and fibrosis.18, 19, 20, 21, 22^,28

Integrated blood and liver tissue analysis revealed intense inflammatory profile, with expression of CXCL8, CCL4/MIP1β, CCL3, CXCL2, CCL3L1, IL1RN, IL1β, and PTGS2 in HBsAg^low patients.29, 30, 31, 32 This inflammatory environment suggests immune cell infiltration, activation, and immune-mediated tissue damage within the liver, a characteristic feature of liver fibrosis in CHB.^15,28, 29, 30 Notably, FOS and DUSP6 were upregulated in both PBMCs and liver-infiltrating cells, implying a central role for these genes in the development of fibrosis in HBsAg^low.^18,21

Further, from integrated blood and liver, single cell analysis, we observed sharp increase in monocytes in HBsAg^low compared to both healthy and HBsAg^high patients, suggesting a mono-mac-driven immune response in low HBsAg antigen contexts. An increased frequency of mono-mac cells with high expression of FOS, FOSB, ZFP36, KLF4, and CXCL3 were observed in HBsAg^low patients in our study, suggesting mono-mac modulation and infiltration of CD8, NK cells in the liver as documented earlier.^18,20,33

Despite the limited number of patients for blood and tissue integrated analysis, our single cell analysis interestingly showed the appearance of new peripheral mono clusters; M_c1 and M_c6 in HBsAg^low with enriched expression of inflammation-related genes, including CCL3, CXCL8, and IL1β, supporting recruitment and immune cell activation at injury sites. This also supports that lower HBsAg levels may preserve monocyte subsets critical for immune response and cell recruitment.34, 35, 36, 37 In contrast, in HBsAg^high patients, clusters M_c0 and M_c5 exclusively displayed high expression of apoptosis-protective genes such as NINJ1, SERPINB9, and JUNB, indicating a potential role in protecting against cell stress under high HBs antigen exposure.^38,39

This Mac activity, particularly through CXCL8, correlated with increased liver infiltration of CD14+ve Macs and a higher fibrosis association in HBsAg^low patients. Intrahepatic CD14+ve Macs express CCL2, CXCL8 and make a CXCL8 concentration gradient to recruit cytotoxic CD8 and NK-immune subsets, leading to fibrosis in CHB-infected animal models.^40,41 In HBsAg^high patients, CD8 and CD4 T cells predominated, suggesting enhanced T cell-mediated immunity but with increased exhaustion markers, possibly reflected immune exhaustion with high antigen. Such immune remodeling could occur due to increased Mac infiltration, migration, and recruitment of CD8 and NK within the live.⁴²

The inflammatory response, driven by cytokines like CXCL8, aligns with enhanced trafficking and activation of CD4, CD8, and NK cells. However, we have found significantly high CD4 regulatory T cells in HBsAg^high, reflecting immune exhaustion with increased HBsAg load.^42,43

Recently, other groups have observed that HBe antigen and different splice isoforms of HBsAg modulate CD8 T cells gene expression towards exhaustion state.⁴⁴ This could be explained by the CD8 molecular signature with HBsAg load in nCHBV. However, in certain contexts, CD8 T cell gene expression was not correlated with HBsAg level. However, our PBMCs data reflected in naive CD4_c3 cluster expressing MALAT1, LINC00861, and LRRN3 was mainly present in healthy but missing in patients. But there were a dominant CD8 c0 and c2 clusters representing both naive and effector T cells with expression of SELL, CCR7, LRRN3, LEF1, BACH2, MAL, IL6ST, NELL2, and TCF7 critical for T-cell adhesion, differentiation, and development in HBsAg^low patients compared to HBsAg^high.⁴⁵ TEM_CD8_c2 also expressed cytotoxic genes GNLY, GZMH, GZMB, and GZMA. The coexpression of SELL, GZMH, GNLY, and FGFBP2 (Figure 5F) in both peripheral and intrahepatic CD8⁺ T cells of HBsAg^low patients suggests a strong cytotoxic potential, enabling effective clearance of infected cells.^46,47 However, this heightened immune activity may also contribute to liver fibrosis by promoting chronic inflammation and tissue damage.

Conclusion

This study reveals that naive chronic hepatitis B patients with low HBsAg levels (<2000 IU/ml) exhibit increased immune activation and higher fibrosis scores, suggesting a link between lower HBsAg levels and fibrosis progression. Single-cell profiling highlighted the elevated chemotactic and inflammatory mediators like CXCL8, CCL20, and IL1β in HBsAg^low patients, which may also contribute to immune cell recruitment and fibrosis progression.

However, this cross-sectional study design limits the ability to establish causal relationships between HBsAg levels and fibrosis progression. The relatively small sample size may affect the translational strength of the findings. Future longitudinal studies with larger cohorts are needed to validate these findings and explore the mechanisms underlying immune activation and fibrosis in this patient population.

However, at the same time, this study highlights the need of regular follow-ups for chronic hepatitis B virus patients in clinics, as despite low HBsAg levels patients can develop worse pathogenesis.