Abstract

   Accumulating evidence suggests that berberine (BBR) exhibits
   anti‐cancer effects in hepatocellular carcinoma (HCC). However, the
   mechanisms by which BBR regulates the immunological microenvironment in
   HCC has not been fully elucidated. In this study, a mouse model of
   orthotopic HCC is established and treated with varying doses of BBR.
   BBR showed effectiveness in reducing tumor burden in mice with HCC.
   Cytometry by time‐of‐flight depicted the alterations in the tumor
   immune landscape following BBR treatment, revealing the enhancement in
   the T lymphocytes effector function. In particular, BBR decreased the
   proportion of TCRb^hiPD‐1^hiCD69^+CD27^+ effector CD8^+ T lymphocytes
   and increased the proportion of Ly6C^hiTCRb^+CD69^+CD27^+CD62L^+
   central memory CD8^+ T lymphocytes. Single‐cell RNA sequencing further
   elucidates the effects of BBR on transcriptional profiles of liver
   immune cells and confirms the phenotypical heterogeneity of T
   lymphocytes in HCC immune microenvironment. Additionally, it is found
   that BBR potentially regulated the antitumor immunity in HCC by
   modulating the receptor‐ligand interaction among immune cells mediated
   by cytokines. In summary, the findings improve the understanding of
   BBR's impact on protecting against HCC, emphasizing BBR's role in
   regulating intrahepatic T cell heterogeneity. BBR has the potential to
   be a promising therapeutic strategy to hinder the advancement of HCC.

   Keywords: Berberine, CyTOF, Hepatocellular carcinoma, scRNA‐seq, T
   cells
     __________________________________________________________________

   BBR treatment effectively reduces tumor burden in mice with HCC,
   thereby protecting against HCC progression. Functionally, BBR
   potentially regulates the differentiation status and phenotypical
   heterogeneity of intrahepatic T lymphocytes in the immune
   microenvironment of HCC. Moreover, BBR may enhance antitumor immunity
   in HCC by modulating cell–cell communications mediated by
   receptor‐ligand interactions.

   graphic file with name ADVS-11-2405182-g005.jpg

1. Introduction

   Liver cancer ranks as the third leading cause of cancer‐related death,
   with hepatocellular carcinoma (HCC) being the most prevalent type of
   primary liver cancer.^[ [30]^1 , [31]^2 ^] Over the past decade,
   systemic therapies including immune checkpoint inhibitors (ICIs),
   tyrosine kinase inhibitors, and monoclonal antibodies have transformed
   the traditional treatment paradigm for HCC.^[ [32]^3 ^] Single‐agent
   ICIs are estimated to provide substantial clinical benefit for 15–20%
   of responders, but face the challenge of biomarkers failing to identify
   this population.^[ [33]^4 , [34]^5 ^] Therefore, it is necessary to
   explore novel therapeutic strategies to improve the landscape of HCC
   management.

   Liver immune microenvironment plays a vital role in the tumorigenesis
   of HCC. For example, tumor‐infiltrating T cells and B cells have been
   observed interacting, leading to enhanced local immune activation and
   control of HCC progression.^[ [35]^6 ^] A study utilizing single‐cell
   RNA sequencing (scRNA‐seq) revealed that exhausted CD8^+ T cells and
   infiltrating Tregs may undergo clonal expansion in the HCC
   microenvironment.^[ [36]^7 ^] Unlike the typical exhausted state
   observed in primary HCC, CD8^+ T cells in early‐relapse HCC exhibit an
   innate‐like reduced cytotoxicity and clonal expansion, potentially
   explaining the compromised antitumor immune function and poor prognosis
   associated with HCC.^[ [37]^8 ^] Therefore, regulating the tumor immune
   microenvironment (TIME) may contribute to developing effective
   therapies for HCC.^[ [38]^9 ^]

   Berberine (BBR), an isoquinoline alkaloid derived from the roots, bark,
   and rhizomes of plants like Coptis chinensis, exhibits diverse
   pharmacological effects.^[ [39]^10 ^] A previous study demonstrated
   that BBR alleviates non‐alcoholic fatty liver disease by reducing
   gluconeogenesis and lipogenesis.^[ [40]^11 ^] Additionally, BBR was
   reported to inhibit HCC through activating PPARδ to trigger apoptotic
   death and promoting gut microbiota‐derived butyric acid production.^[
   [41]^12 ^] In recent years, researchers have increasingly focused on
   the immunomodulatory properties of BBR. Liu and colleagues found that
   BBR exerted its antitumor effect in non‐small cell lung cancer by
   reducing PD‐L1 expression and enhancing tumor‐infiltrating T‐cells
   immunity.^[ [42]^13 ^] Another study revealed that BBR promoted the
   transformation of immunosuppressive M2 macrophages into tumoricidal M1
   macrophages, restoring the antitumor cytotoxicity of T cells.^[ [43]^14
   ^] However, the precise mechanism of BBR in regulating the immune
   response in HCC remains undiscovered.

   In this study, we investigated the effects of BBR on HCC and elucidated
   its immunoregulatory role in the TIME of HCC. We found that BBR could
   notably decrease tumor weight and volume in a mouse model of orthotopic
   HCC. Utilizing cytometry by time‐of‐flight (CyTOF), we identified
   immune cell landscape in HCC. Additionally, the single‐cell
   transcriptional profile of CD45^+ immune cells was constructed through
   scRNA‐seq. BBR administration led to an increase in central memory
   CD8^+ T (Tcm) cells and a decrease in effector T (Teff) cells
   expressing high levels of inhibitory receptors, thereby alleviating T
   cell exhaustion. Cytokine detection revealed the regulatory effect of
   BBR on T cell function. Our findings comprehensively delineate the
   impact of BBR on immune cell landscape of HCC and suggest the potential
   efficacy of BBR as an immune modulatory strategy in vivo.

2. Results

2.1. BBR Treatment Alleviates the Tumor Load in Mice with HCC

   To investigate the potential of BBR in the treatment of HCC, we
   established a mouse model of orthotopic HCC. Following one week of
   adaptive feeding, mice were injected with the Hepa 1–6 cells into the
   left lobe of liver and subsequently received intraperitoneal BBR
   administration. The mice were categorized into three groups based on
   BBR dosage: HC group (vehicle administration), LB group (low‐dose BBR,
   10 mg kg^−1), and HB group (high‐dose BBR, 30 mg kg^−1). BBR was
   administered every 2 days for a total of 14 days. Upon euthanasia,
   mouse livers were harvested and tumor tissues were excised for further
   analysis (Figure [44]1A). The results revealed no significant
   differences in mouse body weight among the three groups
   (Figure [45]1B). However, mice treated with BBR exhibited obviously
   diminished tumor volume (Figure [46]1C). In particular, compared to the
   LB group, the high‐dose BBR treatment displayed a more significant
   reduction in tumor volume and weight (p < 0.05), indicating a
   dose‐dependent trend of BBR in alleviating the burden of HCC in mice
   (Figure [47]1D,E). Taken together, BBR demonstrated efficacy in
   reducing tumor burden in mice model, suggesting it's potential function
   in protecting from HCC progression.

Figure 1.

   Figure 1
   [48]Open in a new tab

   BBR treatment reduces tumor burden in HCC bearing mice. A) Experimental
   design. B) Mouse body weight of the tumor‐bearing mice from HC, LB, and
   HB group. C) Gross examination of liver tumors of the HC, LB, and HB
   group. D) Tumor volume of the tumor‐bearing mice from HC, LB, and HB
   group. E) Tumor weight of the tumor‐bearing mice from HC, LB, and HB
   group. ns, non‐significant; ^∗ p < 0.05; ^∗∗ p < 0.01; ^∗∗∗ p < 0.001.

2.2. CyTOF Depicted the Alterations in the Tumor Immune Landscape Following
BBR Treatment

   The tumor tissues collected from the three groups of the murine HCC
   model were processed into single‐cell suspensions, followed by CyTOF
   procedure to analyze alterations in the immune landscape of the HCC
   tumor microenvironment (TME) in response to BBR treatment. Subsequent
   to initial data processing, we conducted a cluster analysis of the
   immune cells using the Phenograph algorithm, resulting in the
   classification of these immune cells into 28 clusters based on
   classical markers (Table [49]S1, Supporting Information). The 28
   clusters were further categorized into 10 primary cell types, namely
   innate lymphoid cells (ILCs), CD8^+ T lymphocytes, CD4^+ T lymphocytes,
   B lymphocytes, natural killer (NK) cells, classical dendritic cells
   (cDCs), monocytes, neutrophils, Kupffer cells, and macrophages (Figure
   [50]2A). In addition, we used t‐SNE plots to illustrate the
   dimensionality reduction clustering outcomes of the 28 immune cells
   clusters (Figure [51]2B) and depict the expression of canonical cell
   surface markers in various immune cells (Figure [52]2C).

Figure 2.

   Figure 2
   [53]Open in a new tab

   CyTOF depicted the alterations in the tumor immune landscape following
   BBR treatment. A) Heatmaps depicting the expression of 42 cell surface
   markers across 28 immune cell subsets. B) t‐SNE plot visualizes the 28
   immune cell subsets. C) t‐SNE plot presents the expression of
   characteristic surface markers across immune cell subsets. D)
   Differential analysis in frequency of eleven immune cell types after
   treatment with BBR. E–G) The marker expression levels of (E) CD44, (F)
   TIGIT, and (G) PD‐1. ^∗ p < 0.05; ^∗∗ p < 0.01.

   A comparison of immune cell subsets between the HC, LB, and HB groups
   revealed notable alterations in the HCC immune landscape following BBR
   treatment. The most significant changes occurred in the T lymphocyte
   and macrophage. Compared to HC group, the proportion of CD4^+ and CD8^+
   T lymphocytes decreased in the high‐dose group, while such changes were
   not significant in the low‐dose group. Macrophages and T lymphocytes
   exhibited opposite trends, as the proportion of macrophages in the
   high‐dose group increased compared to the low‐dose group
   (Figure [54]2D; Figure [55]S1, Supporting Information). In mice treated
   with BBR, in addition to changes in their proportions, T lymphocyte
   subsets showed notable changes in functional characteristics. Compared
   with the control group, the high‐dose group exhibited a significant
   increase in the expression abundance of CD44, a marker that represents
   the functional activation of T lymphocyte (Figure [56]2E), indicating
   enhanced T lymphocyte antitumor activity. Moreover, markers of T
   lymphocyte exhaustion, TIGIT, and PD‐1, exhibited decreased expression
   after BBR treatment, indicating attenuation of T cell dysfunction and
   robust activation (Figure [57]2F,G).

2.3. BBR Treatment Regulated the Heterogeneity of T Lymphocytes in Mice with
HCC

   T lymphocytes play pivotal roles in targeting and killing specific
   cells, regulating cytokine secretion and immune processes.^[ [58]^15 ^]
   In this study, a decrease in the proportion of T lymphocytes was
   observed in HCC‐bearing mice treated with a high dose of BBR.
   Therefore, we performed a stepwise cluster analysis on the of T
   lymphocytes, aiming to explore the heterogeneity of T lymphocyte
   subsets following treatment with BBR. T lymphocytes were subcategorized
   into 11 clusters based on canonical T cell markers (Figure [59]3A).
   T‐SNE plots were generated to display the distribution and quantity of
   each cell cluster (Figure [60]3B). Following BBR treatment, we visually
   observed alterations in the immune landscape through t‐SNE plots
   (Figure [61]S2A, Supporting Information). Statistical analysis revealed
   significant changes in the proportions of C03, C07, C09, and C11
   subsets post‐BBR treatment (Figure [62]3C,D). Notably, C07,
   characterized by TCRb^hiPD‐1^hiCD69^+CD27^+, represented an CD8^+ Teff
   lymphocytes with high PD‐1 expression (Table [63]S2, Supporting
   Information). Conversely, C11, featured by
   Ly6C^hiTCRb^+CD69^+CD27^+CD62L^+, represented a subset of CD8^+ Tcm
   lymphocytes with potential tumor‐killing characteristics. After
   treatment, the TCRb^hiPD‐1^hiCD69^+CD27^+ CD8^+ Teff decreased in
   proportion in the HB group, while the proportion of
   Ly6C^hiTCRb^+CD69^+CD27^+CD62L^+ CD8^+ Tcm increased, underscoring the
   enhanced ability of the immune system to effectively eradicate tumor
   cells.

Figure 3.

   Figure 3
   [64]Open in a new tab

   BBR treatment regulated the heterogeneity of T lymphocytes in mice with
   HCC. A) T lymphocyte subsets pedigree markers heatmap. B) The t‐SNE
   plot of T lymphocyte subsets. C) The proportional distribution of T
   lymphocyte subsets. D) Variations in the proportion of C03, C07, C09,
   and C11 subsets among HC, LB, and HB group. E) The expression of
   functional marker PD‐1 among the three groups. F) The expression of
   functional marker CD69 among the three groups. ^∗ p < 0.05; ^∗∗ p <
   0.01; ^∗∗∗ p < 0.001.

   A comprehensive analysis of functional markers within T lymphocytes
   revealed decreased expression of markers indicative of functional
   exhaustion, such as PD‐1, TIGIT, and FOXP3 (Figure [65]3E; Figure
   [66]S2B,C, Supporting Information). In contrast, the expression of
   markers suggestive of functional activation, such as CD69 and CD25, was
   increased (Figure [67]3F; Figure [68]S2D, Supporting Information).
   These findings collectively indicate that BBR treatment regulated the
   heterogeneity of T lymphocyte in mice with HCC, leading to an overall
   improvement in the effector function of T lymphocytes. Specifically,
   BBR therapy counteracted T lymphocyte exhaustion by diminishing the
   proportion of CD8^+Teff lymphocytes characterized by high expression of
   inhibitory receptors, while increasing the proportion of CD8^+ Tcm
   lymphocytes.

2.4. scRNA‐seq Elucidated the Global Liver Immune Profiles Respond to BBR
Treatment

   To further elucidate the effects of BBR treatment on liver immune
   profiles of mice with HCC, we collected mouse liver cancer tissue
   samples (n = 6:6:6) from HC group, LB group, and HB group. Single‐cell
   suspension was prepared using the method established in the preceding
   phase of the experiment. Subsequently, scRNA‐seq was performed to
   obtain single‐cell transcriptional profiles. An unsupervised cluster
   analysis of the scRNA‐seq data annotated eleven distinct immune cell
   types based on established marker genes, including monocyte‐derived
   macrophages (MoMFs), cDCs, neutrophils, basophils, Kupffer cells, CD4^+
   T cells, CD8^+ T cells, NK cells, plasma cells, B cells, and
   plasmacytoid dendritic cells (pDCs) (Figure [69]4A). Myeloid cells
   exhibited specific expression of LYZ2 genes, while neutrophils
   specifically expressed S100A9 and S100A8 genes. B cells demonstrated
   specific expression of CD79a genes, whereas T/NK cells specifically
   expressed CD3g and CD3d genes (Figure [70]4B). It is noteworthy that
   MoMFs, T/NK cells, and neutrophils constituted the predominant cell
   types across the three groups (Figure [71]4C). Furthermore, various
   other immune cell populations were present in the liver of HCC mice,
   albeit in smaller proportions, indicating heterogeneity within TIME.
   Compared with that in the HC group, the proportion of MoMFs in the HB
   group exhibited an increasing trend (Figure [72]4D). Conversely, the
   proportion of CD4^+ T cells was decreased in the HB group. After BBR
   treatment, the difference in the proportion of immune cells mainly
   centered on T cells and MoMFs, and the overall trend was consistent
   with the CyTOF results.

Figure 4.

   Figure 4
   [73]Open in a new tab

   scRNA‐seq elucidated the global liver immune profiles respond to BBR
   treatment. A) UMAP plot of the cluster analysis of all samples, with
   each color denoting a distinct cell type. B) Expression profiles of
   specific marker genes for each cell type in UMAP. C) Proportion of
   immune cell types in the HC, LB, and HB groups. D) Alterations in the
   proportions of distinct cell types in the HC, LB, and HB groups. ^∗ p <
   0.05; ^∗∗ p < 0.01; ^∗∗∗ p < 0.001.

2.5. scRNA‐seq Confirmed the Phenotypical Heterogeneity of T Lymphocytes in
the Immune Microenvironment of HCC

   To discern phenotypical heterogeneity of T cells, we conducted
   unsupervised cluster analysis on scRNA‐seq data comprising 8626 T cells
   from three sample sets (Figure [74]5A,B). According to the
   characteristic genes, 11 distinct T lymphocyte clusters were
   identified: C0 (TNFSF11^+CD4^+Teff), C1 (ICOS^+CD4^+Teff), C2
   (KLRC1^+CD8^+Teff), C3 (Cycling CD8^+T), C4 (LEF1^+CD8^+Tcm), C5
   (GZMC^+CD8^+Teff), C6 (CD4^+Treg), C7 [T‐cell stress response state
   (TSTR) CD4^+T], C8 (CD4^+Th17), C9 (Cycling CD4^+Treg), and C10
   (EOMES^+CD8^+Tn) (Table [75]S3, Supporting Information). Compared to
   the HC group, high‐dose BBR treatment resulted in reduced frequency of
   immunosuppressive TNFSF11^+CD4^+ Teff cell expressing IL1R2 and
   increased frequency of LEF1^+CD8^+ Tcm cell potentially undergoing
   differentiation (Figure [76]5C). Moreover, the TSTR CD4^+ T cell and
   pro‐inflammatory CD4^+ Th17 cell expressing IL23R were also observed to
   be significantly elevated in the HB group.

Figure 5.

   Figure 5
   [77]Open in a new tab

   scRNA‐seq confirmed the phenotypical heterogeneity of T lymphocytes in
   the immune microenvironment of HCC. A) UMAP plot of T lymphocyte
   clusters in all samples. B) UMAP plot of marker CD4 and CD8a. C)
   Differences in the proportion of distinct T lymphocyte clusters in the
   HC, LB, and HB groups. D) Heatmap illustrating the characteristic genes
   across 11 clusters of T lymphocytes. E) Violin plot displaying
   characteristic genes in C0, C4, C7, and C8 clusters. F) Violin plot
   displaying the expression of characteristic genes in C0, C4, C7, and C8
   clusters, respectively. ^∗ p < 0.05; ^∗∗ p < 0.01.

   To further investigate the functional alterations within the T
   lymphocytes, we analyzed gene expression levels concerning naive
   markers, effector molecules, inhibitory receptors, cytokine or cytokine
   receptors, and co‐stimulatory molecules (Figure [78]5D). The results
   showed that the inhibitory receptors gene LAG3 and PDCD1 were highly
   expressed in TNFSF11^+CD4^+ Teff cell. Characteristic genes indicating
   T cell differentiation potential, such as LEF1, exhibited robust
   expression in LEF1^+CD8^+ Tcm cells. The co‐stimulatory molecule CD28
   was significantly expressed on TSTR CD4^+ T cells, while the gene of
   cytokines or cytokine receptors were obviously expressed on CD4^+ Th17
   cells. Next, our investigation revealed significant differences in the
   gene expression profiles within these four clusters (Figure [79]5E).
   Specifically, expressions of TNFSF11, IL1R2, ECE1, CD4, and FURIN were
   upregulated in TNFSF11^+CD4^+ Teff cells, while expressions of SELL,
   CCR7, TCF7, LEF1, and BCL2 were elevated in LEF1^+CD8^+ Tcm cells. TSTR
   CD4^+ T cells significantly expressed FURIN, LEF1, BCL2, HSPA1a,
   HSPA1b, and IFIT1, whereas CD4^+ Th17 cell showed significant
   expression of FURIN, TRDC, and IL23R.

   In addition, we investigated the effects of BBR treatment on the
   expression levels of characteristic genes in C0, C4, C7, and C8
   (Figure [80]5F). IL2RA, a characteristic gene of TNFSF11^+CD4^+ Teff
   cells, was significantly upregulated following treatment with a high
   dose of BBR. TNFRSF18, which can significantly reverse CD4^+ Treg
   cell‐mediated immunosuppression as a T cell co‐stimulatory signal, was
   highly expressed in LEF1^+CD8^+ Tcm and TSTR CD4^+ T cells in HB group.
   GZMB, which can inhibit aberrant IL‐17 production in CD4^+ T cells,
   thereby impeding angiogenesis and subsequent tumor development, was
   upregulated in CD4^+ Th17 cells in HB group. These findings suggest
   that T cells exhibit heterogeneity in the immune microenvironment of
   HCC, and BBR treatment regulate antitumor immunity by affecting the
   expression of characteristic genes in T cells.

2.6. BBR Treatment Potentially Regulated the Level of Cytokines Secreted by
Heterogeneous T Cells

   To further explore the potential regulatory effect of BBR on the murine
   HCC immune microenvironment, we quantitatively analyzed 40 cytokines in
   mouse serum (Table [81]S4, Supporting Information). The principal
   component analysis (PCA) plot indicated clear distinctions among
   samples from HC, LB, and HB groups (Figure [82]6A). Next, we analyzed
   the differential expression of cytokines between the HC and HB groups
   (Figure [83]6B). Cluster heatmap showed that most cytokines were
   significantly down‐regulated by high‐dose BBR treatment, such as IL‐1α,
   IL‐1β, IL‐7, IL‐13, IL‐15, TNF‐α, monocyte chemotactic protein 1
   (MCP‐1), and leptin (Figure [84]6C). GO and KEGG enrichment analysis
   revealed that the differentially expressed cytokines proteins primarily
   involved in biological processes related to cytokine and
   chemokine‐mediated signaling pathways (Figure [85]6D,E).

Figure 6.

   Figure 6
   [86]Open in a new tab

   BBR treatment potentially regulates the immune microenvironment by
   modulating cytokine secretion. A) PCA plot of cytokines in the HB, LB,
   and HC groups. B) Differential expression of cytokines between the HB
   and HC groups. C) Heatmap illustrating the differences in cytokines
   between the HB and HC groups. D) Dot plot depicting the alterations of
   cytokines among the HB, LB, and HC groups. E) GO‐BP enrichment
   analysis. F) KEGG pathway enrichment analysis.

   Next, we investigated the relationship between heterogeneous T cells
   and their expression of cytokines gene based on scRNA‐seq data
   (Figure [87]6F). The results showed that high‐dose BBR administration
   resulted in downregulated expression of C‐C chemokine ligand (CCL) 2,
   colony‐stimulating factor‐1 (CSF‐1), IL‐1β, and C‐X‐C motif chemokine
   ligand (CXCL) 1. Conversely, the TNFSF11^+CD4^+ Teff cells exhibited
   upregulated expression of Interferon‐γ and IL‐13, while CCL3 expression
   was downregulated. IL‐15, known for inducing T cell activation and
   maintain T cell effector function, was predominantly expressed by TSTR
   CD4^+ T cells in the HB group. Additionally, CD4^+ Th17 in the HB group
   exhibited substantial secretion and expression of TNF. This indicates
   that the secretion of cytokines is linked to heterogeneous T cells and
   is influenced by high‐doses BBR.

2.7. BBR Enhanced Cell–Cell Interactions and Communications of Immune Cells
in HCC

   CellChat was employed to determine the ligand‐receptor interactions and
   communication networks among various immune cell types (Figure [88]7A;
   Figure [89]S3A, Supporting Information). Compared to the HC group, we
   observed an overall increase in the number of intercellular
   interactions after high‐dose BBR treatment (Figure [90]7B).
   Ligand‐receptor analysis elucidated that high‐dose BBR predominantly
   regulated the signal transduction from liver immune cell to CD4^+ T
   cells through chemokines ligand‐receptor pairs, such as CCL3–CCR5,
   CCL5–CCR5, CXCL16–CXCR6, as well as SPP1–(ITGA+ITGB) ligand/receptor
   pairs, such as SPP1–(ITGA4+ITGB1), SPP1–(ITGAV+ITGB1) (Figure [91]7C).
   In addition, BBR mainly mediates intercellular communication from CD4^+
   T cells to immune cells such as MoMF and Kupffer cell through
   ligand‐receptor pairs including CCL5–CCR1, CCL5–CCR5, macrophage
   migration inhibitory factor (MIF)–(CD74+CD44), LGALS9– CD44,
   LGALS9–CD45, LGALS9–HAVCR2, and SPP1–(ITGA+ITGB) (Figure [92]7D).

Figure 7.

   Figure 7
   [93]Open in a new tab

   BBR enhances cell–cell interactions and communications of immune cells
   in HCC. A) The number of intercellular interactions in the HC and HB
   groups, respectively. B) Bar plot presents the interaction number of
   the HC and HB groups, the heatmap represents the differential number of
   interactions comparing HB groups to HC group. C) Bubble map presenting
   the comparison of significant ligand–receptor pairs in the HB, LB, and
   HC groups, when CD4^+ T cell as receptor cells. D) Bubble map
   presenting the comparison of significant ligand–receptor pairs in the
   HC, LB, and HB groups, when CD4^+ T cell as ligand cells.

   Additionally, high‐dose BBR regulated the interactions from liver
   cancer immune cells to CD8^+ T cells via ligand‐receptor pairs such as
   CCL3–CCR5, CCL5–CCR5, CXCL6–CXCR6, and SPP1–(ITGA4+ITGB1) (Figure
   [94]S3B, Supporting Information). BBR mainly mediates the communication
   from CD8^+ T cells to other immune cells through ligand‐receptor pairs
   such as CCL5–CCR1, CCL5–CCR5, and SPP1–(ITGA+ITGB) (Figure [95]S3C,
   Supporting Information). Moreover, high‐dose BBR treatment can also
   induce specific signals in CD8^+ T cells, including TGFβ, IL1, and GDF
   (Figure [96]S3D, Supporting Information). Collectively, the anti‐liver
   cancer effect of high‐dose BBR may be achieved via modulating the
   ligand‐receptor interactions among immune cells mediated by cytokines
   and chemokines, thereby regulating the liver immune microenvironment.

2.8. Pseudotime Analysis and Development Trajectories of T Lymphocyte

   The Monocle algorithm was utilized to obtain the pseudotime trajectory
   of CD4^+ T cells (Figure [97]8A). The trajectory map revealed that Th17
   cells were mainly distributed at the beginning of the pseudotime path.
   Subsequently, Th17 cells developed into TSTR CD4^+ T cells before
   differentiating into CD4^+ Treg cells, with effector CD4^+ T cells
   emerging early in the differentiation process (Figure [98]8B). A
   comparison of the pseudotime trajectories of CD4^+ T cells between the
   HB and HC groups demonstrated that high‐dose BBR treatment might
   regulate the development and differentiation of CD4^+ T cells
   (Figure [99]8C). To dissect the CD8^+ T cell developmental trajectories
   that occurs during high‐dose BBR treatment, we arranged single CD8^+T
   cells into pseudotime trajectory (Figure [100]8D). We found that
   central memory CD8^+ T cells mainly existed in the initial
   developmental stages, indicative of their potential for differentiation
   and their role as reserve cells for antitumor immunity
   (Figure [101]8E). In addition, the distribution characteristics of
   effector CD8^+ T cells on the pseudotime path suggested that they seem
   to be in the middle stage of development. Further analysis indicated
   that high‐dose BBR treatment may regulate the heterogeneity of CD8^+ T
   cells by affecting their differentiation status (Figure [102]8F).

Figure 8.

   Figure 8
   [103]Open in a new tab

   Pseudotime analysis and development trajectories of T lymphocyte
   clusters. A) Pseudotime trajectory of CD4^+ T cells. B) Mapping of each
   CD4^+ T cell cluster on a locus diagram. C) Cell trajectories of CD4^+
   T cells in the HC, LB, and HB groups. D) Pseudotime trajectory of CD8^+
   T cells. E) Mapping of each CD8^+ T cell cluster on a trajectory
   diagram. F) Cell trajectories of CD8^+ T cells in the HC, LB, and HB
   groups.

3. Discussion

   In recent years, significant advancements have been made in the
   immunotherapy of HCC.^[ [104]^16 ^] Although ICIs have emerged as the
   first‐line treatment for patients with advanced HCC, only fewer than
   30% of patients achieve complete or partial remission. Therefore,
   exploring effective strategies to enhance its efficacy has become a key
   element in the development of HCC immunotherapy. Studies have shown
   that natural products like saponins, polysaccharides, and flavonoids
   hold substantial potential in tumor immunotherapy, primarily due to
   their role in the remodeling of tumor‐immunosuppressive
   microenvironment.^[ [105]^17 , [106]^18 ^] In this study, we
   demonstrate that BBR exerts a tumor‐suppressive effect in mouse model
   of orthotopic HCC, possibly by modulating the distribution pattern and
   function of immune cells in TIME. Further analysis revealed a decline
   in exhausted T (Tex) cells and an augmentation in functionally
   activated T cells. To the best of our knowledge, this study is the
   first to elucidate that how BBR could regulate intrahepatic T cell
   heterogeneity to protect against HCC. Our findings may provide new
   mechanistic insights into the natural product BBR as a promising
   anti‐HCC candidate.

   Exhaustion of tumor‐infiltrated lymphocytes, especially CD8^+ T cells,
   is one of the physiological mechanisms for HCC to evade antitumor
   immune responses.^[ [107]^19 ^] Tex cells exhibit a loss of effector
   function, characterized by the expression of multiple immune
   checkpoints, such as cytotoxic T lymphocyte‐associated antigen 4
   (CTLA4), PD‐1, T cell immunoglobulin and mucin‐domain containing‐3
   (TIM3), TIGIT, and FOXP3.^[ [108]^20 ^] Blockading inhibitory receptors
   is expected to overcome T cell exhaustion, thus eliciting effective
   antitumor immunity.^[ [109]^21 ^] Our study found that BBR
   administration reduced the expression level of PD‐1, TIGIT, and FOXP3
   on T cells, suggesting that BBR has the potential to restore T cell
   exhaustion. A recent clinical trial demonstrated that lenvatinib plus
   anti‐PD‐1 antibodies is effective and safety conversion therapy regimen
   for unresectable HCC, and the responders showed significant enrichment
   of CD8^+ T cells.^[ [110]^22 ^] Banta et al. found that dual blockade
   of PD‐1 and TIGIT coinhibitory receptors on T cells restored
   costimulatory receptor CD226 signaling and optimized antitumor CD8^+
   T‐cell responses.^[ [111]^23 ^] FOXP3 is a key transcription factor of
   Treg and plays an important role in maintaining the immunosuppressive
   function of Treg.^[ [112]^24 ^] Depletion of FOXP3^+ Tregs and PD‐1^+ T
   cells can restore the antitumor capabilities of T cells and expand Teff
   cell populations.^[ [113]^25 ^] However, the mechanism by which BBR
   inhibits the expression of immune checkpoints is still not well
   understood and requires further exploration.

   Our data revealed that BBR increased the percentage of Tcm with high
   CD69 and CD25 expression. Similarly, Li et al. demonstrated the notable
   effect of BBR in promoting the memory and effector differentiation of
   CD8^+ T cells, mainly achieved through activating AMPK and STAT5.^[
   [114]^26 ^] Another study showed that BBR antagonized the β‐catenin
   pathway by inhibiting β‐catenin translation and mTOR activity, thereby
   reducing the survival rate of HCC cells.^[ [115]^27 ^] During
   tuberculosis, Pahuja and colleagues found that BBR enrich lung‐resident
   memory T cells by modulating NOTCH3/AKT signaling pathway, leading to
   enhanced anti‐tubercular immunity.^[ [116]^28 ^] These studies suggest
   that BBR can effectively regulate different cellular signaling
   pathways, providing new insights into the immune regulatory mechanism
   of BBR. Moreover, it was reported that BBR exhibited non‐toxicity to
   normal hepatocytes and easily penetrated tissues and tumors.^[ [117]^29
   ^] Therefore, BBR holds promise as an immunomodulator for HCC.

   In this study, high‐dose BBR was found to protect against HCC
   progression, with decreased serum concentrations of IL‐1α, Leptin,
   MCP‐1, and TNF‐α. Similarly, BBR was reported to reduce the secretion
   of TNF‐α and IL‐10 by Th1 and Th17 cells in mice with autoimmune
   neuritis.^[ [118]^30 ^] It is well established that HCC arises from
   precancerous, oncogene‐induced transformation of senescent cells, which
   can secrete IL‐6, IL‐8, IL‐1α, Leptin/Leptin receptor, MCP‐1, etc. IL‐6
   and TNF‐α have been reported to promote HCC through the interaction of
   signal transducer and activator of transcription protein 3 and nuclear
   factor kappa B. Besides, Qin et, al. found that BBR can hinder the
   function of antigen‐presenting cells by downregulating the
   co‐stimulatory molecules CD80 and CD86 and the cytokines IL‐6 and
   IL‐12.^[ [119]^31 ^] On the other hand, our findings revealed that the
   transcript profiles related to cytokines was significantly altered in T
   cells of the HB group, with downregulation of genes encoding CSF‐1 and
   IL‐1β. TNFSF11^+CD4^+ Teff cells with high expression of TNFSF11, which
   positively regulates IL‐1β, TNF‐α, and IL‐6, were significantly
   decreased in the HB group, leading to a decrease in the aforementioned
   cytokines. Numerous clinical studies have highlighted the
   pro‐tumorigenic effects of IL‐1α and IL‐1β, in which IL‐1β stimulates
   inflammatory mediators, promotes cellular invasion and
   immunosuppression.^[ [120]^32 , [121]^33 ^] Additionally, TNF exhibits
   diverse tumor‐promoting activities, such as stimulating cytokine
   cascade responses, fibrotic responses, and altering adhesion
   receptors.^[ [122]^34 ^] These findings imply that BBR treatment could
   alter the transcriptional profiling of immune cells and prevent HCC
   progression through influencing the secretion of cytokines.

   This study is the first to characterize the HCC immune landscape and
   constructed the CD45^+ single‐cell transcriptional profile in
   BBR‐treated mice. We elucidate the impacts of BBR on intrahepatic
   immune cells in mice with HCC, specifically heterogeneous T cells.
   However, this study still has some limitations. First, the exact
   molecular mechanisms through which BBR regulates T cell function remain
   elusive and require further explorations. Second, further studies are
   greatly needed to determine how BBR affects the receptor‐ligand
   interactions between T cell and different immune cells subpopulations.
   Third, it is necessary to investigate the combination of BBR and
   emerging immunotherapy for HCC to ensure favorable outcomes in clinical
   applications.

4. Conclusion

   In conclusion, our findings indicated that BBR prevented the
   progression of HCC through regulating the heterogeneity of intrahepatic
   T lymphocytes, which involves increasing the proportion of central
   memory CD8^+ T cells while decreasing effector CD8^+ T cells to reverse
   T cell exhaustion. Additionally, BBR may inhibit HCC by regulating the
   secretion of cytokines by T lymphocytes and affecting their
   differentiation status. Our study offers new insights into the
   immunomodulatory mechanisms of BBR in HCC, indicating potential
   application prospects on HCC management.

5. Experimental Section

Cell line and Cell Culture

   The Hepa 1–6 liver cancer cell line was obtained from the American Type
   Culture Collection (ATCC, USA) and stored in cryopreservation at the
   Institute of Immunology, Zhejiang University. Following established
   protocols, the Hepa 1–6 cells were cultured in Dulbecco's Modified
   Eagle's Medium (Gibco, USA) supplemented with 10% fetal bovine serum
   (Gibco, USA) and 1% penicillin–streptomycin (Gibco, USA).^[ [123]^35 ^]
   Continuous cell culture was conducted in a sterile environment at 37 °C
   with a 5% CO[2] concentration.

Animals

   Wild type C57BL/6 male mice, aged 8–10 weeks, were acquired from the
   Experimental Animal Center of Hangzhou Medical College and fed in the
   Laboratory Animal Center of Zhejiang University under specific
   pathogen‐free conditions. The mice were maintained on a 12‐h dark/light
   cycle at a temperature of 22–25°C with ad libitum access to food and
   water. All animal experiments complied with the guidelines for the Care
   and Use of Laboratory Animals of the National Research Council and were
   approved by the Zhejiang University Animal Ethics Committee (NO.
   ZJU20240129).

BBR Solution

   BBR with a purity of 98% was provided by Meilun Biology (Dalian,
   China), dissolved in dimethylsulfoxide (DMSO) (Aladdin, Shanghai,
   China) as a 10 mm stock solution, and kept at −20°C. For all
   intraperitoneally administration to mice, this stock solution was
   diluted in phosphate buffer saline (PBS) (Servicebio, Wuhan, China) to
   achieve the concentration of BBR working solution.

Mouse Model of Orthotopic HCC

   After a week of adaptive feeding, the wild type C57BL/6 mice were
   subject to construct a mouse model of orthotopic HCC.^[ [124]^36 ^] The
   detailed processes are as follows: the mice were anesthetized by
   intraperitoneal injections of 1% pentobarbital sodium solution at a
   dose of 50 mg kg^−1. The mice were then placed on an experiment table
   in a supine position, and the abdominal cavity was exposed. 1 × 10^6
   Hepa1‐6 cells suspended in 10 µL 20% Corning Matrigel (Matrigel:PBS =
   1:4) were injected into the left lobe of the liver. Subsequently, the
   liver was repositioned, and the peritoneum and skin were sutured. All
   surgical procedures were conducted meticulously to prevent additional
   injuries, and the wound was disinfected post‐operation. The mice were
   then placed in an incubator until they naturally regained
   consciousness.

   Next, the HCC mouse models were randomly divided into three groups (6
   mice in each group): HC group, LB group, and HB group. The HB group
   received a BBR working solution at a dosage of 30 mg kg^−1, while the
   LB group received a BBR working solution at a dosage of 10 mg kg^−1.
   The HC group was given vehicle only (0.01% DMSO in 99.9% PBS). The
   volume of administration was 100ul solution per 10 g of mice.
   Administration was carried out intraperitoneally every other day for a
   period of 14 days.

Preparation of a Single‐Cell Suspension

   Eighteen fresh HCC tissues were completely resected from the three
   groups of HCC mouse models (HC vs LB vs HB; n = 6 each) and preserved
   in tissue storage solution, then transported at a low temperature.
   Following washing with cell culture medium twice, the tissues were cut
   into 1 mm^3 pieces. Then, the tissue pieces were placed in a C tube,
   into which a digestive enzyme mix (Mouse Tumor Dissociation Kit,
   Miltenyi Biotec, Germany) was added, and the tube was filled to a
   volume of 5 mL with cell culture medium. The tissues were dissociated
   by incubating at 37 °C for 1 h and were subsequently filtered through a
   70‐µm cell strainer. The filtered tissues were subjected to
   centrifugation at 300 × g and a temperature of 4 °C to obtain the
   cells. Finally, the supernatant was aspirated, and the isolated single
   cells were resuspended for CyTOF and scRNA‐seq.

CyTOF

   Purified antibodies were labeled with metal tags using the MaxPAR
   Antibody Labeling Kit (Fluidigm, USA) and adjusted to the desired
   concentration prior to application (Table [125]S5, Supporting
   Information). To exclude dead cells, the cells were washed with 1 × PBS
   and subsequently stained them with 100 µL of cisplatin solution (250 n
   [m] ) on ice for 5 min. Subsequently, cells were incubated in an Fc
   receptor‐blocking solution before applying a mixture of surface
   antibodies to stain cells. Cells were then washed twice with
   fluorescence‐activated cell sorting (FACS) buffer (1 × PBS containing
   0.5% bovine serum albumin, BD Bioscience, USA) and fixed overnight in
   200 µL of intercalation solution (Maxpar Fix and Perm Buffer containing
   250 nm 191/193Ir, Fluidigm, USA). Following fixation, cells were washed
   with FACS buffer and then with perm buffer (eBioscience, USA) and
   stained with an intracellular antibody cocktail for 30 min on ice.
   Subsequently, cells were washed and resuspended in deionized water,
   mixed with 20% EQ4 Element beads (Fluidigm, USA). The data were
   obtained on a Helios mass cytometer (Fluidigm, USA), then normalized
   and gated in FlowJo (version 10.1, BD, USA) or uploaded to Cytobank
   (Santa Clara, CA) for subsequent analysis. t‐Distributed stochastic
   neighbor embedding (t‐SNE) algorithm was utilized to visualize
   high‐dimensional data.^[ [126]^37 ^]

ScRNA‐seq

   A single‐cell suspension containing over 90% living cells was acquired,
   with the cell concentration adjusted to 1000 cells/µL. According to the
   standard protocol, the single‐cell suspension was loaded onto the
   MobiNova−100 high‐throughput single‐cell controller device (MobiDrop,
   Zhejiang, China) to produce single‐cell gel beads in emulsions
   (GEMs).^[ [127]^38 ^] Then, the captured cell was lysed and the
   released mRNAs were reverse‐transcribed into complementary DNA (cDNA)
   with barcode. The GEMs were then disrupted to mix cDNAs from different
   cells. The resulting cDNAs were used as templates to amplify through
   polymerase chain reaction (PCR), and its quality was assessed using an
   Agilent 2100 (Agilent Technologies, USA). According to the
   manufacturer's instructions, scRNA‐seq libraries were prepared using
   MobiCube RNA‐seq Chip A Single Cell Kit v2.1 (MobiDrop, Zhejiang,
   China), which was then subjected to the MobiNova−100 single‐cell
   sequencing platform (MobiDrop, Zhejiang, China) with a 150 bp
   paired‐end reading strategy. After removing low‐quality reads and
   poly‐A tail and adapter sequences, the qualified raw data were
   processed to construct an expression profile matrix through Mobivision
   (v 1.1). The output gene expression matrices were analyzed through
   Seurat package (version 3.1.2) using the R software (version 4.0.3).
   After all cells were nonlinear dimensionally reduced via uniform
   manifold approximation and projection (UMAP) and projected into 2D
   space, cells were clustered together based on common characteristics.^[
   [128]^39 ^] Individual cells in each cluster were annotated according
   to the gene expression of canonical markers for specific cell types.
   Marker genes exhibiting differential expression across these cell
   clusters were analyzed using the FindMarker function in the Seurat.

Cell–Cell Communication Analysis

   Cell–cell communication analysis was performed using the R package
   CellChat to investigate potential intercellular interaction.^[ [129]^40
   ^] The “netVisual_circle” function was used to calculate the
   interactions number and strength between any two cell types. Visual
   representations illustrating total outgoing and incoming interaction
   probabilities were generated employing the function
   “netAnalysis_signalingRole_scatter”. The expression profiles of
   identified genes pertaining to ligand‐receptor pairs were presented in
   graphical format using the “Dotplot” function of Seurat.

Pseudotime Analysis

   The pseudotime trajectory was inferred using the Monocle (version 2)
   package.^[ [130]^41 ^] Highly variable gene expression profiles
   identified through scRNA‐seq were subjected to dimensionality reduction
   using the DDRTree algorithm for subsequent analysis. Subsequently, the
   function “orderCells” was employed to order each cell along pseudotime
   trajectory.

Cytokine Quantification and Functional Enrichment Analysis

   Following the provided instructions, 40 cytokines were identified from
   murine serum samples (HC vs LB vs HB; n = 6 each) utilizing the
   Quantibody Array Glass Chip (RayBiotech, USA). Raw data underwent
   preprocessing and normalization using RayBiotech software. Normalized
   data were analyzed employing moderated t‐statistics utilizing the limma
   data package from Bioconductor. Additionally, differentially expressed
   proteins were identified, with the cut‐off was set as to fold change >
   1.2 or fold change < 0.83, and adjust p‐value < 0.05. Subsequent
   analysis results were visualized through scatter plots, volcano plots,
   and heatmaps.

Statistical Analysis

   GraphPad Prism (version 8.0) and R (version 4.0.3) were used for
   statistical analysis and graphing. The unpaired Student's t‐test were
   applied for comparisons between two groups. Data are the mean ±
   standard deviation. P value less than 0.05 were considered statistical
   significance. ^* p < 0.05, ^** p < 0.01 and ^*** p < 0.001.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   Q.W. and C.X. conceived and designed the study. J.H. and Q.S. performed
   the experiments and wrote the draft. J.H. participated in data
   acquisition and analysis. Q.W. and C.X. revised the manuscript. All
   authors read and approved the final manuscript.

Supporting information

   Supporting Information
   [131]ADVS-11-2405182-s001.docx^ (18.8MB, docx)

Acknowledgements