Abstract

Background & Aims

   HBV infection is one of the leading causes of liver cirrhosis. However,
   the immune microenvironment in patients with HBV cirrhosis remains
   elusive.

Methods

   Single-cell RNA sequencing was used to analyse the transcriptomes of
   76,210 immune cells in the livers of six healthy individuals and in
   five patients with HBV cirrhosis.

Results

   Patients with HBV cirrhosis have a unique immune ecosystem
   characterised by an accumulation of macrophage-CD9/IL18,
   macrophage-C1QA, NK Cell-JUNB, CD4^+ T cell-IL7R, and a loss of B
   cell-IGLC1 clusters. Furthermore, our analysis predicted enhanced cell
   communication between myeloid cells and all immune cells in patients
   with HBV-related cirrhosis. Pseudo-time analysis of myeloid cells,
   natural killer (NK) cells, and B cells demonstrated a significant
   accumulation of mature cells and a depletion of naive cells in HBV
   cirrhosis. In addition, we observed an increase in antigen processing
   and presentation capacities in myeloid cells in patients with HBV
   cirrhosis, whereas NK cell-mediated cytotoxicity was substantially
   reduced.

Conclusions

   Our results provide valuable insight into the immune landscape of HBV
   cirrhosis, suggesting that HBV cirrhosis is associated with the
   accumulation of activated myeloid cells and impaired cytotoxicity in NK
   cells.

Impact and implications

   The absence of single-cell transcriptome profiling of immune cells in
   HBV cirrhosis hinders our understanding of the underlying mechanisms
   driving disease progression. To address this knowledge gap, our study
   unveils a distinctive immune ecosystem in HBV cirrhosis and represents
   a crucial advancement in elucidating the impact of the immune milieu on
   the development of this condition. These findings constitute
   significant strides towards the identification of more effective
   therapeutic approaches for HBV cirrhosis and are relevant for
   healthcare professionals, researchers, and pharmaceutical developers
   dedicated to combating this disease.

   Keywords: Cirrhosis, Hepatitis B, Immune cell, Liver, Single-cell
   RNAseq

Graphical abstract

   [41]graphic file with name ga1.jpg
   [42]Open in a new tab

Highlights

     * •
       Scar-associated macrophage populations with distinct marker genes
       were expanded in HBV cirrhosis.
     * •
       Exhausted NK cells (NK cell-JUNB) are notably increased in HBV
       cirrhosis.
     * •
       Accumulation of relatively mature cells and loss of relatively
       naive cells was observed in myeloid, NK, and B cell subsets in HBV
       cirrhosis.
     * •
       Induction of antigen processing and presentation capacities was
       observed in myeloid cells in HBV cirrhosis.
     * •
       NK cell-mediated cytotoxicity is substantially diminished in HBV
       cirrhosis.

Introduction

   Liver cirrhosis is one of the most relevant chronic diseases
   endangering public health. Globally, more than one million individuals
   died of liver cirrhosis in 2017,[43]^1 and chronic HBV infection is one
   of the leading causes of cirrhosis, especially in Asia.[44]^2

   The liver functions as an important constituent of the immune system.
   Complex interactions between immune cells contribute to the progression
   of liver cirrhosis. Different immune cell subsets, including
   lymphocytes and myeloid cells, cause inflammation and tissue injury,
   which are critical events during the development of liver
   cirrhosis.[45]^3 Although immunological insights into the pathogenesis
   of liver cirrhosis arising from non-alcoholic fatty liver disease,
   alcoholic injury, and primary biliary cholangitis have been made,[46]^4
   our understanding of HBV-associated cirrhosis remains limited. Existing
   reports have predominantly focused on characterising the immune
   microenvironment in HBV infection and HBV-associated hepatocellular
   carcinoma.[47]^5^,[48]^6

   The development of high-throughput RNA sequencing at the single-cell
   level has revolutionised transcriptome analysis. Single-cell RNA
   sequencing (scRNA-seq) offers novel approaches for resolving the
   complexity and heterogeneity of immune cells, enabling the
   identification of novel cell subsets and the exploration of underlying
   cell lineage relationships.[49]^7 In this study, we used scRNA-seq to
   construct an unbiased and comprehensive atlas of intrahepatic immune
   cells and compared the changes in immune cells in the livers of healthy
   individuals and patients with HBV cirrhosis. The exact dissection of
   the contribution of immune cells to liver cirrhosis could provide new
   clues to develop novel therapeutic targets for HBV-associated liver
   cirrhosis.

Patients and methods

Study participants

   The study was approved by the Ethics Review Center of the First
   Affiliated Hospital of Harbin Medical University (No. 2021179). Liver
   tissue was collected from orthotopic liver transplantations conducted
   at the Department of Liver Surgery of the First Affiliated Hospital of
   Harbin Medical University between 1 March 2021 and 30 September 2021.
   The sample cohort consisted of two healthy donors and two patients
   diagnosed with HBV cirrhosis. Healthy samples consisted of small
   specimens used for assessing organ quality before liver
   transplantation. Informed consent was obtained from all the patients
   and donors involved in the study. Single-cell transcriptome data from a
   total of 45,993 immune cells were acquired from these samples.
   Additionally, we acquired single-cell transcript transcriptome data
   from four healthy livers (19,389 cells) sourced from HRA001730 (a study
   on HBV infection)[50]^6 and three HBV cirrhotic livers (10,828 cells)
   sourced from HRA000069 (a study on HBV related cancer
   development),[51]^8 with the permission from China National Center for
   Bioinformation (CNCB). The clinical characteristics of these patients
   are summarised in [52]Table S1. Overall, a total of 76,210 immune cells
   from healthy livers (n = 6) and HBV cirrhotic livers (n = 5) were
   analysed.

Human tissue processing

   During the surgical procedure, wedge biopsies of non-ischemic fresh
   liver tissue (∼2–3 g per liver) were collected before interrupting
   hepatic vascular inflow. The collected liver tissue was immediately
   immersed in a pre-cooled DMEM medium (Gibco, 11965092, New York, USA)
   at 4 °C and transported to a sterile laboratory bench within 5 min for
   further processing. The tissue was washed with phosphate-buffered
   saline (PBS, Gibco, 14190094, New York, USA) three times and cut into
   small pieces of approximately ∼1–3 mm^3 size. The samples were
   suspended in a cryopreservation solution (10% dimethylsulfoxide, Sigma,
   D8418, Shanghai, China) in 90% foetal bovine serum (Gibco, 10100147,
   New South Wales, Australia), and stored at -80 °C until thawed and
   utilised for single-cell sequencing analysis within 5 days.

Tissue dissociation and cell sorting

   Single-cell suspensions from tissues were prepared using a multi-tissue
   dissociation kit (MACS, 130110201, Bergisch Gladbach, Germany)
   according to the manufacturer's instructions. Briefly, the tissue
   samples were thawed at 37 °C and washed twice with DMEM before being
   minced and digested in a gentleMACS™ C tube (MACS, Bergisch Gladbach,
   Germany) containing digestion solution (4.7 ml of DMEM, 200 μl Enzyme
   D, 25 μl Enzyme A, and 100 μl Enzyme R [MACS, Bergisch Gladbach,
   Germany]) for 30 min at 37 °C. The cell suspension was passed through a
   40 μm cell strainer (Corning, 431750, New York, USA) and centrifuged at
   500 × g for 5 min. The cell pellet was resuspended in red blood cell
   lysis buffer (QIAGEN, 8570396, California, USA) and incubated on ice
   for 3 min. The suspension was centrifuged at 500 × g for 5 min at 4 °C,
   resuspended in PBS, and further processed for cell sorting.

   To obtain haematopoietic cells, the suspension was stained with
   PerCP-Cyanine5.5 conjugated CD45 monoclonal antibody (Invitrogen,
   45045941, New York, USA), and propidium iodide (PI) (Invitrogen,
   BMS500PI, New York, USA) dye. The stained cells were washed twice with
   DMEM medium containing 10% serum and sorted by fluorescence-activated
   cell sorting (FACS) (Sony, SH800S, Tokyo, Japan).

ScRNA-seq based on droplet generation

   For scRNA-seq, cells were processed through the 10X Genomics single
   cell platform using the Next GEM Single cell v3.1 library kit (10X
   Genomics, PN-1000157, California, USA), Next GEM Gel Beads v3.1 (10X
   Genomics, PN-2000164, California, USA), and Next GEM chip G (10X
   Genomics, PN-2000177, California, USA). Approximately 16,000 cells for
   each sample were loaded onto 10X Chromium chips, and around
   8,000–10,000 cells were captured. Gel beads with barcodes, single
   cells, and primers were wrapped in oil droplets to form the gel beads
   in emulsion emulsions. cDNA libraries with barcode and unique multiplex
   index (UMI) information for sequencing was generated by reverse
   transcription of the mRNA. The cDNA libraries were subjected to
   sequencing using the MGISEQ-2000 sequencing platform.

Dataset integration and batch effect removal

   We integrated our scRNA-Seq dataset with datasets from two previous
   studies[53]^6^,[54]^8 and removed batch effect using integration
   anchors as implemented in Seurat.[55]^9 Dataset integration, batch
   effect removal, and subsequent downstream analysis followed the
   standard Seurat workflow.

Analysis software

   The following software was used for analysis: Cell Ranger 4.0, for data
   comparison, cell identification, and gene quantification; Seurat
   3.2,[56]^9 for cell clustering and marker gene identification:
   Doubletfinder 2.0.3,[57]^10 for double cell removal; Velocyto
   0.17.17,[58]^11 for RNA rate change analysis; Reactome 1.26.0,[59]^12
   for gene enrichment analysis; Gsva 1.37.5,[60]^13 gene variation
   analysis; GSEA 4.1.0,[61][14], [62][15] for gene enrichment analysis;
   CellChat 1.6.1,[63]^16 for cell–cell communication analysis; monocle
   2.22.0,[64][17], [65][18] for pseudo-time analysis.

Immunofluorescence

   The slides with liver sections were placed in an oven (Haiyiheng
   Scientific Instruments, PH-070A, Shanghai, China) and baked at a
   constant temperature of 63 °C for 2 h. Deparaffinisation was then
   performed using xylene, followed by gradual rehydration with a series
   of ethanol solutions. Antigen retrieval was performed by microwave
   heating in antigen retrieval solution (DAKO, Cambridgeshire, UK).
   Subsequently, sections were blocked with 10% goat serum in PBS for
   10 min. After washing with PBS, the sections were incubated with
   primary antibodies ([66]Table S2) at 4 °C overnight, followed by
   incubation with Cy3 or FITC conjugated secondary antibodies for 20 min
   at room temperature. The sections were counterstained with DAPI (Thermo
   Scientific, 62248, New York, USA) for 10 min at room temperature and
   mounted with the VECTASHIELD hardest antifade mounting medium (Vector
   Labs, H-1400, California, USA). Full scans were visualised with a
   Tissue FAXS viewer (Tissue Gnostics, Vienna, Austria), and five
   different images were randomly selected from each sample. The number of
   positive cells in each image was counted and averaged for statistical
   analysis.

Results

Single-cell atlas of immune cells in healthy and HBV cirrhotic human livers

   We isolated CD45^+ leucocytes by FACS and obtained single-cell
   transcriptome data from 45,993 immune cells. By integrating these
   datasets with those from two previous studies[67]^6^,[68]^8 using
   Seurat software, we analysed a total of 76,210 immune cells derived
   from healthy livers (n = 6) and livers with HBV cirrhosis (n = 5).

   Principal component analysis (PCA), followed by dimensionality
   reduction using the Uniform Manifold Approximation and Projection
   (UMAP) algorithm with 15 principal components was performed.
   Twenty-three cell clusters were identified based on the most
   significant marker genes ([69]Fig. 1A and [70]Fig. S1A) and classified
   into four major cell subtypes: T cells (CD4+/CD8A+/CD8B+), myeloid
   cells (CD14+/FCGR3A+), NK cells (KLRF1+), and B cells (CD19+)
   ([71]Fig. 1B). To assess potential differences between healthy and
   HBV-cirrhotic liver tissues, a comparative analysis was conducted, as
   illustrated in [72]Fig. 1C. No significant variations were observed in
   the proportions of the four major cell subtypes between the two groups.
   Furthermore, [73]Fig. 1D displays the most significant marker genes
   associated with each subcluster, providing valuable insights into their
   distinct molecular signatures. The contamination rate of
   non-haematopoietic cells, encompassing cycling cells, mesenchymal
   cells, and hepatocytes, was determined to be low, representing a mere
   2.8% of the total cell count.

Fig. 1.

   [74]Fig. 1
   [75]Open in a new tab

   Single-cell atlas of HBV cirrhotic and healthy livers.

   (A) Clustering analysis of all CD45^+ leucocytes. (B) UMAP plot showing
   the expression of CD4, CD8A, CD8B, CD14, CD19, FCGR3A(CD16), and KLRF1.
   (C) The proportion of T cells, NK cells, myeloid cells, and B cells in
   healthy liver and HBV cirrhotic liver samples. (D) Gene expression
   heatmap in each cell cluster. NK, natural killer; UMAP, Uniform
   Manifold Approximation and Projection.

Enhanced antigen processing and presentation in myeloid cells during HBV
cirrhosis

   Myeloid cells are involved in the progression and sustenance of hepatic
   disorders across multiple stages.[76]^19 We performed differential
   expression analysis of myeloid cells comparing healthy and HBV
   cirrhotic liver samples. A total of 489 genes were upregulated and 718
   genes were downregulated in myeloid cells from HBV cirrhotic livers
   ([77]Table S3 and [78]Fig. S1B). GSEA analysis of myeloid cells with
   annotation pathway from Gene Ontology (GO) database showed that
   chromatin and cellular response to reactive oxygen species was
   inhibited, whereas antigen binding, positive regulation of T cell
   mediated immunity, and antigen processing and presentation of peptide
   or polysaccharide antigen via MHC class II were promoted in HBV
   cirrhosis ([79]Fig. S1C). Similarly, utilising annotation pathway from
   Kyoto Encyclopaedia of Genes and Genomes (KEGG) database, GSEA analysis
   demonstrated that antigen processing and presentation were induced, and
   the MAPK signalling pathway was inhibited in HBV cirrhosis myeloid
   cells ([80]Fig. S1C).

Macrophage-C1QA and macrophage-CD9/IL18 expanded in HBV cirrhosis

   We conducted an in-depth analysis of myeloid subsets. A total of 13
   subclusters were identified, including macrophages, monocytes, and
   dendritic cells ([81]Fig. 2C), and the proportion of each myeloid cell
   subcluster was compared between healthy liver and HBV cirrhosis
   specimens ([82]Fig. 2D and E, [83]Fig. S1D).

Fig. 2.

   [84]Fig. 2

   [85]Fig. 2
   [86]Open in a new tab

   The subtypes of myeloid cells in healthy and HBV cirrhotic livers.

   (A) Clustering analysis of all myeloid cells. (B) Dot plot illustrating
   the expression of marker genes in each cluster. (C) The expression of
   CD14, FCGR3A(CD16), and CD9 in the UMAP plot. (D) The proportion of
   different clusters in myeloid cells in healthy and HBV cirrhotic
   livers. ∗p <0.05 (Mann–Whitney U test). (E) UMAP plots of the
   distribution of major cell clusters in healthy and HBV cirrhotic liver
   samples. (F) Immunofluorescence images showing C1QA, CD9, and IL18
   expression in healthy and HBV cirrhotic livers samples (left panel).
   Quantification of positive cells from immunofluorescent images, ∗∗p
   <0.01 (right panel). (G) GO and KEGG enrichment analysis of
   Monocyte-C1QA and Macrophage-CD9/IL18 marker genes. (H) Differential
   expression of pro-inflammatory and anti-inflammatory genes between
   healthy and HBV cirrhosis in clusters of macrophages and monocytes. ∗p
   <0.05, ∗∗p <0.01, ∗∗∗p <0.001 (Mann-Whitney U test). GO, Gene Ontology;
   KEGG, Kyoto Encyclopaedia of Genes and Genomes; NK, natural killer;
   UMAP, Uniform Manifold Approximation and Projection.

   Three monocyte subclusters were identified according to their crucial
   marker genes: Monocyte-S100A9, Monocyte-S100A6, and Monocyte-FCGR3A
   ([87]Fig. 2A and B). Monocyte-S100A9 and Monocyte-S100A6 were
   considered as CD14+ classical monocyte, enriched in the expression of
   S100A4, S100A6, S100A9, and S100A11 ([88]Fig. 2C). Classical monocytes
   are known for their phagocytic activity.[89]^20 Monocyte-FCGR3A,
   characterised by high FCGR3A (CD16) expression, was classified as
   non-classical CD16+ monocyte ([90]Fig. 2C).

   Macrophages were further clustered into seven subpopulations according
   to characteristic marker gene expression, and defined as
   Macrophage-CD9/IL18, Macrophage-CD9/IFI6, Macrophage-NAMPT,
   Macrophage-CCL5, Macrophage-PLCG2, Macrophage-THBS1, and
   Macrophage-C1QA ([91]Fig. 2A and B). Macrophage-CD9/IL18 and
   Macrophage-CD9/IFI6 shared the same marker gene CD9. Previous studies
   have identified a CD9-enriched macrophage subcluster termed as
   scar-associated macrophages, which expands and resides in the fibrotic
   niche in liver cirrhosis unrelated to viral hepatitis.[92]^4 Consistent
   with this, we found that Macrophage-CD9/IL18 increased in HBV cirrhosis
   samples (p <0.05) ([93]Fig. 2D and E). C1QA encodes the A-chain
   polypeptide of serum complement subcomponent C1q.[94]^21 The proportion
   of Macrophage-C1QA was increased in HBV cirrhosis samples (p <0.05)
   ([95]Fig. 2D and E). Immunofluorescence staining validated the
   expansion of C1QA, CD9, and IL18 positive cells in HBV cirrhosis
   ([96]Fig. 2F). To elucidate the functional characteristics of these two
   upregulated macrophage subclusters in HBV cirrhosis,
   Macrophage-CD9/IL18, and Macrophage-C1QA, we performed KEGG and GO
   pathway enrichment analysis on their marker genes. The results revealed
   that the phenotype of Macrophage-CD9/IL18 and Macrophage-C1QA was
   characterised by MHC class II, immune receptor activity, as well as
   antigen processing and presentation, suggesting the critical anti-viral
   immunity roles of these two elevated macrophage subclusters
   ([97]Fig. 2G, [98]Tables S4 and S5).

   To define the inflammatory state of the myeloid subclusters, we
   examined the differential expression of ten pro-inflammatory genes and
   20 anti-inflammatory genes. The fold change (log2) in gene expression
   levels in HBV cirrhotic livers compared to healthy livers was
   visualised using a heatmap ([99]Fig. 2H). In the non-classical monocyte
   subcluster, Monocyte-FCGR3A, we observed downregulation of four
   pro-inflammatory genes and no upregulation, indicating that the
   pro-inflammatory effect of non-classical monocytes declined in HBV
   cirrhosis. In contrast, the anti-inflammatory effect of
   Macrophage-CD9/IL18 and Macrophage-C1QA were attenuated.

Enhanced CD74/CXCR4 signalling pathway on myeloid cells in HBV cirrhosis

   To investigate the potential intercellular crosstalk among immune cells
   in HBV cirrhosis, cellular communication was predicted by using
   Cellchat. Notably, an increase in myeloid cell communication with all
   immune cells was predicted in HBV cirrhosis ([100]Fig. 3A). The
   communication of different myeloid subsets was then examined, and the
   results revealed that increased interactions with other immune cells
   was largely evident in Macrophage-C1QA and Macrophage-CCL5 subsets
   ([101]Fig. 3B). Communication with other immune cells was largely
   reduced in plasmacytoid dendritic cells (pDCs) and slightly decreased
   in NK cells ([102]Fig. 3B), which was in accordance with earlier
   findings that chronic HBV infection causes functional defects in pDCs
   and NK cells.[103]^22^,[104]^23 We also found that CD74/CXCR4 receptors
   in myeloid cells were activated by macrophage migration inhibitory
   factor (MIF) in HBV cirrhosis (p <0.05) ([105]Fig. 3C). MIF binds to
   the CD74/CXCR4 complex, activating the ERK1/2 and AKT pathways via
   G-protein and promoting chemotaxis.[106]^24^,[107]^25

Fig. 3.

   [108]Fig. 3
   [109]Open in a new tab

   Cell communication analysis of myeloid cells.

   (A) The strength of cell communication of immune cells in the healthy
   liver and HBV cirrhosis. (B) The strength of cell communication of
   myeloid cell subclusters in the healthy liver and HBV cirrhosis. (C)
   Cell communication of myeloid cell receptors and immune cell ligands ∗p
   <0.05, ∗∗p <0.01, ∗∗∗p <0.001 (Wilcoxon test).

Myeloid cells differentiated towards activated cell subsets in HBV cirrhosis

   To explore the ontogeny and potential route of differentiation of
   myeloid cells in HBV cirrhosis, we performed pseudo-time and RNA
   velocity analyses ([110]Fig. 4A and B). The results demonstrated that
   Monocyte-S100A6, Monocyte-S100A8, and Monocyte-FCGR3A exhibited a
   relatively immature phenotype, whereas Macrophages, pDC, cDC1, and cDC2
   were classified as mature subsets ([111]Fig. 4C and D). Among the
   various subsets of macrophages, Macrophage-PLCG2 and Macrophage-NAMPT
   represent the final differentiation of myeloid cells. Pseudo-time and
   RNA velocity analyses revealed a shift from immature (Monocyte-S100A6,
   Monocyte-S100A8, and Monocyte-FCGR3A) to mature (macrophage and
   dendritic cell subsets) in HBV cirrhosis ([112]Fig. 4E and
   [113]Fig. S1E), suggesting that myeloid cells differentiated into
   mature subgroups as HBV cirrhosis progressed.

Fig. 4.

   [114]Fig. 4
   [115]Open in a new tab

   Pseudo-time and RNA velocity analysis of myeloid cells.

   (A) Pseudo-time analysis of myeloid cells in healthy and HBV cirrhotic
   livers. (B) RNA velocity analysis of myeloid cells in healthy and HBV
   cirrhotic livers. (C) Cellular trajectory of myeloid cells. (D) The
   distribution of myeloid cell subtypes on the pseudo-time trajectory.
   (E) Pseudo-time trajectory plots of myeloid cell subtypes in healthy
   and HBV cirrhotic livers. (F) Gene expression dynamics of myeloid cells
   along the pseudo-time. Genes cluster into four sets, and each set had a
   specific expression profile, as depicted by a selection of
   characteristic genes.

   We identified four distinct gene sets associated with myeloid cells
   differentiation as HBV cirrhosis progressed ([116]Fig. 4F). Set 1
   consists of markers for complement activation (C1QA, C1QB, C1QC) and
   antigen presentation (HLA-DQB1), which are macrophage and DC cell
   markers. Set 2 contained ribosomal protein (RPS10, RPS11, RPL31) and
   eukaryotic translation elongation factor (EEF1B2, EEF1A1), which are
   involved in cellular function like mRNA transcription and translation.
   The marker gene for final differentiations of myeloid cells, PLCG2, was
   also included in Set 3. Set 4 included myeloid cell precursor and
   immature monocyte marker genes, such as S100A8, S100A9, and S100A12.

NK cell-mediated cytotoxicity was suppressed in HBV cirrhosis

   NK cells elicit cytotoxic effects, especially in virus-infected cells,
   and drive the elimination of HBV-infected hepatocytes.[117]^26 We
   performed a differential expression analysis between healthy and HBV
   cirrhotic NK cells ([118]Table S6, [119]Fig. S2A). GSEA analysis with
   GO annotation revealed that the functions of NK cell such as immune
   effector process and secretory pathways were suppressed in HBV
   cirrhosis samples ([120]Fig. S2B). Similarly, with annotation pathway
   from KEGG database, GSEA analysis showed that NK cell mediated
   cytotoxicity was extensively suppressed in HBV cirrhosis samples
   ([121]Fig. S2B).

Exhausted NK cells expanded in HBV cirrhosis

   To further understand the roles of NK cells in the progression of HBV
   cirrhosis, we conducted an in-depth clustering of NK cells and
   identified nine distinct subclusters: NK Cell-JUNB, NK Cell-FCER1G, NK
   Cell-GZMH, NK Cell-TRAC, NK Cell-MYOM2, NK Cell-ANXA1, NK Cell-TIPARP,
   NK Cell-PLCG2, and NK Cell-GPR183 ([122]Fig. 5A and B). Compared with
   healthy liver, the proportion of NK Cell-JUNB increased (p <0.001) in
   HBV cirrhosis ([123]Fig. 5C and D, [124]Fig. S2C). NK Cell-JUNB was
   enriched by marker genes TNF, LAG3, and IL32 ([125]Table S7). LAG-3
   represents an important indicator for NK cell exhaustion.[126]^27 To
   better characterise the function of NK Cell-JUNB, the marker genes were
   analysed by KEGG and GO pathway enrichment analysis, with the results
   suggesting that NK Cell-JUNB was associated with cell–cell adhesion and
   regulation of T cell activation ([127]Fig. 5E). Pseudo-time analysis
   was then performed to explore the differentiation trajectory of NK
   cells ([128]Fig. 5F–I, [129]Fig. S2D). Interestingly, a distinct
   trajectory of NK cell differentiation was observed in HBV cirrhosis
   samples, where NK cells at the initial stage were nearly absent.

Fig. 5.

   [130]Fig. 5
   [131]Open in a new tab

   The subtypes of NK cells in HBV cirrhosis and healthy liver samples.

   (A) Clustering analysis of NK cells. (B) Dot plot showing the marker
   genes in each cluster. (C) The proportion of NK cells subclusters in
   healthy and HBV cirrhotic liver samples. ∗p <0.05, ∗∗p <0.01, ∗∗∗p
   <0.001 (Mann-Whitney U test). (D) UMAP plots of the distribution of NK
   cell subclusters in healthy and HBV cirrhotic livers. (E) GO and KEGG
   enrichment analysis of NK cell-JUNB marker genes. (F) Pseudo-time
   analysis of NK cells in healthy and HBV cirrhotic livers. (G)
   Pseudo-time of NK cells in healthy and HBV cirrhotic livers. (H)
   Pseudo-time analysis of NK cells in healthy and HBV cirrhotic livers.
   (I) The distribution of NK cell subtypes on the pseudo-time trajectory.
   GO, Gene Ontology; KEGG, Kyoto Encyclopaedia of Genes and Genomes; NK,
   natural killer; UMAP, Uniform Manifold Approximation and Projection.

CD4^+ T cell-IL7R expanded and naive B cells disappeared in HBV cirrhosis

   We performed an in-depth analysis of T cell subclusters and identified
   nine T cell subclusters ([132]Fig. 6A–C). Compared with the healthy
   liver, the proportion of CD4^+ T cell-IL7R increased (p <0.01) in HBV
   cirrhosis ([133]Fig. 6D and E). KEGG pathway and GO enrichment analysis
   with CD4^+ T cell-IL7R marker genes suggested that the function of
   CD4^+ T cell-IL7R was associated with the ribosome, the NF-kappa B
   signalling pathway, and virus receptor activity ([134]Fig. 6F). In
   pseudo-time analysis, CD4^+ T cell-IL7R exhibited a relatively mature
   phenotype ([135]Fig. S3A–D).

Fig. 6.

   [136]Fig. 6

   [137]Fig. 6
   [138]Open in a new tab

   The subtypes of T and B cells in HBV cirrhosis and healthy liver
   samples.

   (A) Clustering analysis of all the T cells. (B) The expression of CD4,
   CD8A, and CD8B in UMAP plot (C) Dot plot showing the marker genes in
   each cluster. (D) UMAP plots of the distribution of T cell subclusters
   in healthy and HBV cirrhotic livers. (E) The proportion of T cell
   subclusters in healthy and HBV cirrhotic livers. ∗∗p <0.01
   (Mann-Whitney U test). (F) GO and KEGG enrichment analysis of CD4^+ T
   cell-IL7R marker genes. (G) Clustering analysis of B cells. (H) Dot
   plot showing the marker genes in each cluster. (I) UMAP plots of the
   distribution of B cell subclusters in healthy and HBV cirrhotic liver
   samples. (J) The proportion of B cell subclusters in healthy and HBV
   cirrhotic liver samples. ∗p <0.05 (Mann-Whitney U test). (K) GO and
   KEGG enrichment analysis of B cell-IGLC1 marker genes. GO, Gene
   Ontology; KEGG, Kyoto Encyclopaedia of Genes and Genomes; UMAP, Uniform
   Manifold Approximation and Projection.

   In-depth clustering of B cells was performed to determine the potential
   contribution of B cells ([139]Fig. 6G). We identified nine B cell
   subclusters ([140]Fig. 6G and H). Compared with the healthy liver
   samples, the proportion of B cell-IGLC1 decreased (p <0.05) in HBV
   cirrhosis samples ([141]Fig. 6I and J). We also performed the KEGG
   pathway and GO enrichment analysis of the marker genes of B cell-IGLC1.
   The functions of B cell-IGLC1 were associated with B cell activation
   and the B-cell receptor signalling pathway ([142]Fig. 6K). In
   pseudo-time analysis, we found that naive B cells almost disappeared in
   HBV cirrhosis ([143]Fig. S3E–H).

Discussion

   Our single-cell RNA sequencing analysis revealed a unique immune
   ecosystem in HBV cirrhosis, characterised by an accumulation of
   Macrophage-CD9/IL18, Macrophage-C1QA, NK Cell-JUNB, CD4^+ T cell-IL7R,
   and a loss of B cell-IGLC1 cell clusters. Our results suggest that
   changes in immune cell subsets contribute, at least in part, to the
   progression of HBV cirrhosis. Currently, there is a lack of single-cell
   transcriptome profiling studies focused on immune cells in HBV
   cirrhosis, which limits our understanding of the underlying mechanisms
   involved in disease progression. Our study provides an essential
   advance in understanding the impact of the immune environment on HBV
   cirrhosis. Communication among different types of immune cells in the
   liver is vital for regulating local inflammatory responses and
   preventing the propagation of inflammatory signals. Cirrhosis disrupts
   the equilibrium between adaptive immunity and tolerance to pathogens,
   resulting in tissue damage and systemic inflammation.[144]^3

   Myeloid cells are believed to be essential to the pathogenesis of
   chronic liver injury,[145]^28 which was confirmed in our analysis.
   Previous studies have identified a subset of scar-associated
   macrophages (TREM2+CD9+) in non-hepatic cirrhosis.[146]^4 Although we
   did not observe the TREM2+ subpopulation, we identified two CD9+
   subpopulations of macrophages (Macrophages-CD9/IL18 and
   Macrophage-CD9/IFI6) in HBV cirrhosis. It is possible that variations
   in the aetiology of cirrhosis may contribute to differences in the
   characteristic genes of scar-associated macrophages. The proportion of
   Macrophage-CD9/IL18 and Macrophage-C1QA increased from healthy to HBV
   cirrhosis. The pool of hepatic myeloid cells expands dramatically after
   liver injury because of infiltrating monocyte-derived
   macrophages.[147]^29 Accordingly, although we observed the maturation
   of macrophages in HBV cirrhosis, there was a significant accumulation
   of macrophages and loss of monocytes in HBV cirrhosis. The mobilisation
   of monocytes in cirrhosis has been previously described and is
   consistent with the notion that systemic inflammation is one of the
   major drivers of disease progression in
   cirrhosis.[148]^3^,[149]^28^,[150]^30

   NK cells are a prominent subset of lymphocytes abundant in the healthy
   liver. Classical NK cells elicit cytotoxic effects, especially against
   virus-infected cells, thereby contributing to the elimination of
   HBV-infected hepatocytes.[151]^26 Additionally, NK cells exhibit
   antibacterial activities by secreting cytokines such as
   interferon-gamma (IFN-γ), tumour necrosis factor alpha (TNF-α), and
   IL-17.[152]^3 Understanding whether changes in NK cells within the
   liver of HBV cirrhosis are primarily driven by HBV infection or liver
   injury can be challenging. We observed an expansion of the NK cells
   involved in regulating T-cell activation and leucocyte cell–cell
   adhesion. Therefore, we propose that functional changes in NK cells may
   be attributed to cirrhosis-related factors rather than solely to HBV
   infection.

   Recent research by Nkongolo et al.[153]^31 uncovered an enriched
   population of hepatotoxic CD8+ T cells that displayed a highly
   activated immune signature during liver damage. In our analysis, CD8+ T
   cell-NKG7 possesses similar marker genes (IFNG, GZMA, GZMK, TNFRSF9,
   HLA-DR, CD74, PDCD1, LAG3, and TIGIT) and may represent hepatotoxic
   CD8+ T cells. The proportion of hepatotoxic CD8+ T cells correlated
   with alanine aminotransferase (ALT) levels.[154]^31 It is worth noting
   that the patients with HBV cirrhosis we included had received antiviral
   therapy, resulting in better control of liver damage, with ALT levels
   increasing by a mean of 3.8-fold ([155]Table S1). As a result, our
   study demonstrated a slight but not statistically significant increase
   in hepatotoxic CD8+ T cell-NKG7 in HBV cirrhotic liver samples
   ([156]Fig. 6E).

   Although our study provides valuable data for understanding the
   contribution of immune cells to the progression of HBV cirrhosis, a few
   limitations should be noted. Firstly, our data sets did not allow us to
   distinguish whether the observed changes are specifically related to
   the HBV infection, cirrhosis, or a combination of both. Future studies
   could address this by including samples from HBV-infected but
   non-cirrhotic livers and cirrhosis samples from non-HBV causes. Because
   of the challenges in collected cirrhosis tissue, the cirrhosis tissue
   samples we included from previously published data were from patients
   with hepatocellular carcinoma,[157]^8 which may have introduced
   potential influence from the tumour background on the immune cells.
   Another limitation is the inability to identify neutrophil granulocytes
   in our analysis. The fragility of neutrophils during cell isolation
   procedures might have led to their destruction. Optimised protocols may
   be utilised in future study to comprehend the role of neutrophil
   granulocytes in HBV cirrhosis. Moreover, cirrhosis may progress to
   acute-on-chronic liver failure (ACLF),[158]^32 an advanced disease
   stage that typically leads to distinct immune cell patterns.[159]^30
   Better understanding of the immune ecosystem will provides specific
   therapeutic modalities to prevent and overcome ACLF.

   In conclusion, our scRNA-seq analysis of immune cells in healthy livers
   and HBV cirrhosis livers revealed a unique immune ecosystem associated
   with HBV cirrhosis. HBV cirrhosis promotes monocyte activation and
   differentiation and is associated with functional changes in NK cells.
   These findings contribute to a better understanding of the immune
   mechanisms underlying HBV cirrhosis and may aid in the development of
   novel therapeutic strategies for treatment.

Financial support

   This study was financially supported by Shenzhen Science and Technology
   Program (JCYJ20210324121802008), the National Natural Science
   Foundation of China (32270833), the Natural Science Foundation of
   Fujian Province (2022J01014, 2022R1001002), and the Natural Science
   Foundation of Xiamen City (3502Z20227164).

Authors’ contributions

   Contributed to the conception and design of the study: QB, XH, TH, FT,
   CE. Provided administrative, study supervision, and obtained funding:
   XH, TH, FT, CE. Provided human specimens: LH. Performed experiments and
   substantially contributed to the acquisition of data and its analysis:
   QB, XH, RL, XH, HB, MH. Interpretation of data: All authors. Drafted
   the manuscript: QB, CE. Revised the manuscript critically for important
   intellectual content: All authors.

Data availability statement

   Data are available at reasonable request to the corresponding author.
   Single-cell RNA sequencing raw data published in NCBI (RPRJNA833766).

Conflicts of interest statement

   The authors declare no conflicts of interest that pertain to this work.

   Please refer to the accompanying ICMJE disclosure forms for further
   details.

Footnotes

   Author names in bold designate shared co-first authorship

   Supplementary data to this article can be found online at
   [160]https://doi.org/10.1016/j.jhepr.2023.100883.

Contributor Information

   Frank Tacke, Email: frank.tacke@charite.de.

   Cornelius Engelmann, Email: cornelius.engelmann@charite.de.

   Tianhui Hu, Email: thu@xmu.edu.cn.

Supplementary data

   The following are the supplementary data to this article:
   Multimedia component 1
   [161]mmc1.pdf^ (3MB, pdf)
   Multimedia component 2
   [162]mmc2.docx^ (40.2KB, docx)
   Multimedia component 3
   [163]mmc3.pdf^ (225KB, pdf)
   Multimedia component 4
   [164]mmc4.pdf^ (14.2MB, pdf)

References