Abstract

   Despite significant advances in the diagnosis and treatment of
   hepatocellular carcinoma (HCC), metastasis and recurrence remain two
   major obstacles to improving the clinical outcomes for HCC patients.
   Here, we demonstrate that splenic invariant natural killer T (iNKT)
   cells can significantly inhibit Hepa1-6-mediated intrahepatic HCC
   metastasis. Interestingly, in the HCC metastasis model, iNKT deficiency
   can result in a significant decrease in percentage and absolute number
   of CD4^+ T cell and interleukin-4 level, thus suggesting the
   involvement of the cross-talk between iNKTs and CD4^+ T cells in
   limiting HCC metastasis to the spleen. Transcriptional signatures of
   CD4^+ T cells following iNKT deficiency displaying impairment of their
   cell migration function. During HCC metastasis, splenic iNKT rapidly
   secrete interferon-γ to promote the migration of CD4^+ T cells from the
   marginal zone into the white pulp, thereby triggering subsequent
   migration of splenic B cells to the liver and exerting anti-tumor
   immune effects on Hepa1-6 cells. In conclusion, interactions between
   interferon-γ and its receptor on iNKT and CD4^+ T cells can effectively
   coordinate immune activity between the marginal zone and the white
   pulp, thereby ultimately inhibiting intrahepatic HCC metastasis. These
   findings reveal the mechanism underlying the resistance of splenic iNKT
   to tumor metastasis.

   Subject terms: Imaging the immune system, Innate lymphoid cells
     __________________________________________________________________

   graphic file with name 42003_2025_7798_Figa_HTML.jpg
     __________________________________________________________________

   Splenic iNKT promotes the entry of CD4^+ T cells into the white pulp
   through secretion of IFN-γ. CD4^+ T cells regulate splenic B cells
   migrate to the liver and inhibit HCC metastasis.

Introduction

   Hepatocellular carcinoma (HCC) is the fourth leading cause of
   cancer-related mortality worldwide and the leading cause of death from
   liver cirrhosis^[36]1. With a 5-year survival rate of 18%, HCC is the
   second deadliest tumor after pancreatic cancer^[37]2. Although,
   significant advances have been made both in the diagnosis and treatment
   of cancer over the past few centuries, but metastasis and recurrence
   remain the main obstacle to improving clinical outcomes in cancer
   patients^[38]3. Despite the availability of various treatment options
   for the management of HCC patients, such as surgical resection,
   percutaneous ablation, radiotherapy, and transarterial and systemic
   therapies, these patients are still at a high risk of tumor metastasis,
   especially local intrahepatic invasion^[39]4. Surgical resection, is
   the recommended treatment option, with a 5-year overall survival rate
   of ~70%, but a significant drawback of this approach is the high
   incidence of local invasion by liver tumors^[40]5, occurring in up to
   80% of patients^[41]5,[42]6. The local invasion of cancer is a complex,
   multi-step process involving infiltration of cancer cells into the
   primary lesion, entry into the blood stream, survival in the
   circulation, extravasation and attachment and colonization of the
   metastatic site^[43]7. The means by which splenic iNKTs exert immune
   surveillance of this process remains largely unknown.

   The mouse spleen is divided into red pulp (RP) and white pulp (WP)
   according to function and structure, with the marginal zone (MZ)
   located between these two areas^[44]8. In contrast to lymph nodes, the
   spleen lacks afferent lymphatic vessels and all cells as well as
   antigens can enter the spleen via the blood. Therefore, RP dominates
   the spleen tissue, constituting a significant portion, whereas the WP
   accounts for less than a quarter, but serves as key immunologic region
   of the spleen^[45]9. The structure of the spleen allows it to
   efficiently remove blood-borne microorganisms, older erythrocytes and
   cellular debris from the circulation, earning it the distinction of
   being the body’s largest blood filter^[46]10. Moreover, except for red
   blood cells, all types of blood cells, regardless of the volume can
   pass through the blood vessels in the spleen for ~10 min^[47]11–[48]13.
   This facilitates full contact and interaction between tumor cells and
   immune cells in the red pulp and marginal zone. Interestingly, numerous
   studies have reported the beneficial role of the spleen in
   significantly improving the tumor prognosis. For instance, preservation
   of the spleen during extended lymphadenectomy for gastric cancer can
   effectively reduce various complications and improve overall
   survival^[49]14. Furthermore, the presence of liver metastases leads to
   the suppression of splenic dendritic cells, which poses a challenge to
   anti-tumor vaccination efforts. Measurement of spleen stiffness was the
   only predictor of late HCC recurrence and was directly related to the
   extent of liver disease and portal hypertension, both of which were
   associated with HCC recurrence^[50]15. According to a comment made on
   the article, it was proposed that splenic stiffness during HCC might be
   influenced to some extent by alterations in the immune system within
   the spleen^[51]16. Therefore, the spleen may play an irreplaceable role
   in immune surveillance of local invasion and metastasis in HCC.

   Mouse iNKT is characterized by the expression of the unique
   CD1d-restricted T cell receptor (TCR) Vα14Jα18^[52]17. Interestingly,
   both the frequency and function of iNKTs in the tumor or circulation
   have been found to be highly correlated with overall survival in
   numerous malignancies^[53]18–[54]21. For instance, iNKT can inhibit the
   growth of murine HCC by improving oncogenic beta-catenin-induced
   chronic inflammation in the liver^[55]22,[56]23. Methylcholanthrene
   (MCA) was shown to initiate spontaneous tumor growth more rapidly in
   Ja18^−/− mice (iNKT-deficient) in comparison to the wild-type mice, and
   isolated iNKTs were found to display direct cytotoxicity against
   MCA-induced sarcoma^[57]24. Disruption of commensal gut bacteria in
   mice can induce hepatic iNKT-mediated anti-tumor effects^[58]25.
   Adoptive transfer of autologous iNKTs has demonstrated promising safety
   and efficacy in treating advanced HCC^[59]26,[60]27. Concurrently,
   iNKT-induced anti-tumor effects have been reported to be pleiotropic,
   such as TCR-engineered iNKTs can induce potent anti-tumor responses by
   dual targeting of both cancer and suppressive myeloid cells^[61]28.
   Moreover, in all the above- described studies, iNKTs have demonstrated
   direct and/or indirect anti-tumor immune surveillance effects.
   Nevertheless, the precise role of splenic iNKT, particularly in
   relation to local invasion of HCC, remains predominantly unrevealed.

   Here, the findings highlight the distinctive function of splenic iNKTs
   in controlling immune surveillance in the context of local invasion of
   HCC. The interferon-γ (IFN-γ)/IFN-γR-dependent interaction between iNKT
   and CD4^+ T cells can promote the migration of CD4^+ T cells from the
   marginal zone into the white pulp, a key event in preventing local
   intrahepatic invasion of HCC. This study provides insights into
   understanding of how splenic iNKT drives immune surveillance of local
   invasion of tumor cells and facilitate the development of iNKT
   cell-based anti-cancer immunotherapies.

Results

Splenic iNKT deficiency exacerbates metastasis of hepatocellular carcinoma

   To validate the potential role of splenic iNKT in hepatocellular
   carcinoma metastasis, we established a mouse model of intrahepatic
   metastasis of hepatocellular carcinoma (HCC-m) mouse model. 5 × 10^5
   Hepa1-6 cells were injected orthotopically into the mouse liver and the
   liver tissues at the injection site were subsequently removed. After 2
   weeks, Hepa1-6 cells in the circulation were found to recolonize in the
   liver parenchyma to form intrahepatic metastases. In addition, compared
   with the control mice, the spleen volume as well as weight of HCC-m the
   mice were significantly increased (Fig. [62]1A). Splenectomized mice
   displayed more severe HCC metastasis, with a significant increase in
   the tumor weight (Fig. [63]1B). To gain a deeper understanding of how
   the spleen can suppress HCC metastasis, we analyzed both the
   percentages and absolute numbers of the different immune cell
   populations in the spleens of HCC-m mice at 2 weeks. We identified that
   both the percentage and absolute number of splenic iNKTs were
   significantly elevated in HCC-m mice at 2 weeks (Fig. [64]1C–E). We
   observed that the percentage and absolute number of splenic CD4^+ T
   cells decreased, whereas the percentage of splenic CD8^+ T cells
   increased in HCC-m mice (Fig. [65]1F–H). However, there were no
   significant differences noted in the percentages and absolute numbers
   of splenic B cells and NK cells among these two groups of mice
   (Fig. [66]S1A–F). We further examined the possible effect of iNKT on
   HCC metastases in Ja18^−/− mice, HCC-m models for wild-type (WT) and
   Ja18^−/− mice were constructed and tumor weights were analyzed after 2
   weeks. Interestingly, iNKT deficiency significantly aggravated the HCC
   metastasis, but adoptive transfer of iNKT in the spleen effectively
   inhibited this process (Fig. [67]1I, J). These findings confirmed that
   splenic iNKTs have the ability to inhibit HCC metastasis.

Fig. 1. iNKT deficiency is beneficial for the metastasis of hepatocellular
carcinoma.

   [68]Fig. 1
   [69]Open in a new tab

   A Representative appearance (left) and spleen weight (right) of spleens
   from the control mice and mice with metastasis (n = 7). B 5 × 10^5
   Hepa1-6 cells were injected orthotopically into the mouse liver, and
   the liver tissues at the injection site was subsequently removed. The
   representative appearance of livers from the control and splenectomized
   mice (left) and tumor weight (right) (n = 6). C The proportion of the
   CD45^+ TCR-β^+ CD1d-PBS57-Tetramer^+ iNKT obtained from the spleens of
   control mice and mice with metastasis were analyzed using the flow
   cytometry. The isotype Ig of each antibody was used as a negative
   control. D Quantification of the percentage and E absolute number of
   iNKT in the indicated group (n = 5). F Flow cytometry analysis of the
   percentages of CD45^+ CD3^+ CD4^+ T cell and CD45^+ CD3^+ CD8^+ T cell
   in the spleens of control and HCC-m mice. G The percentage and H
   absolute number of CD4^+ T cells and CD8^+ T cells in both groups of
   mice were statistically analyzed. The variance has been marked on the
   graph (n = 4). I Representative appearance of the liver metastasis
   tumors in WT+Hepa1-6, Jα18^−/−+Hepa1-6 and Jα18^−/−+Hepa1-6+iNKT
   adoptively transferred mice. For the establishment of the splenic iNKT
   adoptive transfer mouse model, 5 × 10^5 iNKTs were orthotopically
   injected into the spleen of the recipient mouse. Tumor tissue is shown
   within the dotted box. J Statistical analysis of the liver tumor
   weights of mice in each group (n = 5).

The presence of iNKT contributes to the maintenance and activation of CD4^+ T
cells

   We performed additional experiments to provide a clearer understanding
   of the crucial role played by iNKTs - which make up only about 2% of
   non-parenchymal cells in the spleen - in suppressing HCC metastasis.
   iNKT serves as a bridge between innate immunity and adaptive immunity
   and plays a crucial role in the initial stage of the disease^[70]29.
   Therefore, we focused towards determining whether impairments in iNKTs
   have any impact on immune cell populations 3 days after injection of
   Hepa1-6 cells. We thus analyzed the differences in the percentages of
   the major immune cell populations in the spleen of Ja18^−/− mice. We
   observed that at steady state, Ja18^−/− mice exhibited decreased basal
   CD4^+ T cell number in comparison to the control mice (Fig. [71]2A–C).
   The reduced number of splenic CD4^+ T cells caused by HCC metastasis
   was further worsened by iNKT deficiency, resulting in a further
   decrease in both the number and percentage of CD4^+ T cells
   (Fig. [72]2A–C). However, iNKT deficiency had no impact on the
   percentages of splenic B cells (Fig. [73]S2A, B) and NK cells
   (Fig. [74]S2C, D). Adoptive transfer of iNKTs to the spleen can
   effectively restore the reduction in splenic CD4^+ T cell percentage
   caused HCC metastasis (Fig. [75]2D, E). This observation led us to
   investigate whether iNKTs have the ability to promote the proliferation
   of CD4^+ T cells. Consistent with our expectations, splenic CD4^+ T
   cells from both Ja18^−/− mice and Ja18^−/− mice with HCC-m exhibited
   diminished proliferation ability, indicated by the low level of Ki67
   expression (Fig. [76]2F, G). However, iNKT deficiency did not affect
   the apoptosis of CD4^+ T cell (Fig. [77]S2E, F). The expression of CD1d
   is crucial for the cross-talk between antigen-presenting cells and
   iNKTs. We analyzed the expression levels of CD1d on CD4^+ T cells in
   the spleens of HCC-m and control mice by flow cytometry. Both groups of
   mice showed almost no expression of CD1d on splenic CD4^+ T cells
   (Fig. [78]S2G, H). Similarly, splenic CD4 + T cells of iNKT-deficient
   mice also lacked expression of CD1d (Fig. [79]S2G, H). In addition, the
   expression of the T cell activation marker CD69 and the level of
   interleukin-4 (IL-4) were both reduced in splenic CD4^+ T cells of
   Ja18^−/− mice in comparison to the control mice (Fig. [80]2H, I).
   Interestingly, the expression levels of IFN-γ, IL-6 and TNF-α in CD4^+
   T cells were not affected by iNKT deficiency (Fig. [81]2H, I).
   Therefore, the cross-talk between iNKT and CD4^+ T cells can play a
   vital role in inhibiting HCC metastasis.

Fig. 2. The cross-talk between iNKT and CD4^+ T cells helps to control the
metastasis of hepatocellular carcinoma.

   [82]Fig. 2
   [83]Open in a new tab

   A Proportion of the CD45^+ CD3^+ CD4^+ T cells derived from the spleens
   of WT mice, WT mice with HCC metastasis, Jα18^−/− mice and Jα18^−/−
   mice with metastasis were analyzed using flow cytometry. B
   Quantification of the percentage and C absolute number of CD4^+ T cell
   in the indicated group. D Proportion of the CD4^+ T cells obtained from
   the spleens of WT mice, Jα18^−/− mice, Jα18^−/− mice with HCC
   metastasis and Jα18^−/−+Hepa1-6+iNKT adoptively transferred mice were
   analyzed using flow cytometry. E Quantification of the percentage of
   CD4^+ T cell in the indicated group. F Flow cytometry analysis of the
   percentage of Ki67^+ CD4^+ T cell in WT, WT + HCC-m, Jα18^−/−,
   Jα18^−/− + HCC-m mice. G Statistical analysis of the percentage of
   Ki67^+ CD4^+ T cell in each group. H Analysis of CD69, IL-4, IFN-γ,
   IL-6, and TNF-α in CD4^+ T cells of the indicated group through flow
   cytometry. I Statistical analysis of CD69, IL-4, IFN-γ, IL-6, and TNF-α
   mean fluorescence intensity (MFI) of CD4^+ T cells of the indicated
   group (n = 5). The data shown was obtained from three independent
   experiments with n = 5 mice per group. The data has been displayed as
   the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

CD4^+ T cells in Ja18^−/− mice exhibited unique mRNA expression patterns
associated with impaired cell migration function

   To investigate the impact of iNKT cell interactions with CD4^+ T cells
   on HCC metastasis in vivo, we isolated CD4^+ T cells from both WT HCC-m
   and Ja18^−/− HCC-m mice and performed RNA sequencing analysis.
   Interestingly, 648 DEGs were identified in the splenic CD4^+ T cells of
   WT HCC-m and Ja18^−/− HCC-m mice. Among them, 292 genes were found to
   be upregulated, whereas 356 genes were downregulated (Ja18^−/− vs Ctrl,
   Fig. [84]3A). GO enrichment of the DEGs (Ja18^−/− vs Ctrl, Down) was
   conducted in major functional classes of biological processes (BP),
   cellular components (CC), and molecular functions (MF). The
   downregulated DEGs were significantly annotated in the processing of T
   cell activation including protein phosphorylation, TCR V(D)J
   recombination, positive regulation of the TCR signaling pathway, as
   well as stress activated protein kinase signaling pathway
   (Fig. [85]3B). In addition, consistent with the previous descriptions,
   signaling pathways related to the cell cycle and cell proliferation
   were downregulated in CD4^+ T cells from Ja18^−/− mice (Fig. [86]3B).
   Furthermore, downregulation of signaling pathways related to
   cytoskeleton organization and cell adhesion was also noted, thus
   suggesting changes in CD4^+ T cell motility in Ja18^−/− mice
   (Fig. [87]3B). Annotation of the cellular components revealed that
   histone methyltransferase complex, cytoskeleton and TCR complex were
   also significantly downregulated (Fig. [88]3C). As for the molecular
   functions, the downregulated DEGs were significantly annotated for the
   transferase activity, protein binding, DNA binding, histone
   methyltransferase activity (Fig. [89]3D). These molecules can promote
   the specific gene expression as well as regulate genome stability, cell
   maturation, DNA methylation, and cell mitosis. Moreover, the Kyoto
   encyclopedia of genes and genomes (KEGG) enrichment analysis indicated
   that the downregulated DEGs were associated with the processes such as
   cell growth and death, cell motility, amino acid metabolism, glycan
   biosynthesis as well as metabolism, which have been linked to the
   anti-tumor effect of T cells (Fig. [90]3E). Gene set variation analysis
   (GSEA) pathway enrichment analysis revealed that CD4^+ T cells in
   Ja18^−/− mice were able to significantly down-regulate inflammatory
   response and cell surface receptor-related signaling pathways, thereby
   suggesting that their activation was effectively inhibited.
   Furthermore, CD4^+ T cells in Ja18^−/− mice downregulated the various
   mRNA-encoding proteins associated with cell migration/chemotaxis
   functions, including cell adhesion, chemotaxis, leukocyte migration and
   lymphocyte chemotaxis (Fig. [91]3F).

Fig. 3. Transcriptional characteristics of splenic CD4^+ T cells in
iNKT-deficient HCC-m mice.

   [92]Fig. 3
   [93]Open in a new tab

   A Volcano map depicting DEGs in CD4^+ T cells from WT+Hepa1-6 and
   Jα18^−/−+Hepa1-6 mice. Each group contained 3 mice. B GO enrichment of
   the various downregulated genes (Jα18^−/− mice vs WT mice) in the
   biological processes, C cellular components, and D molecular functions.
   E KEGG enrichment analysis of the different downregulated genes. F Gene
   set enrichment analysis of the various downregulated genes.

   Gene Ontology (GO) enrichment of the DEGs (Ja18^−/− vs Ctrl, Up) was
   conducted in the GO categories of cellular component (CC), molecular
   function (MF), and biological process (BP). iNKT deficiency led to
   significant upregulation of DEG annotations in extracellular space,
   endothelial reticulum, chaperone complex, peroxidase activity,
   antioxidant activity, and cellular oxidation detoxification-related
   genes in CD4^+ T cells (Fig. [94]S3A–C). In addition, protein
   processing in the endothelial reticulum, ferroptosis, and
   apoptosis-related pathways were significantly upregulated in the CD4^+
   T cells of Ja18^−/− mice, suggesting that iNKT cells play an important
   role in the function of CD4^+ T cells (Fig. [95]S3D).

HCC metastasis can promote the migration of splenic CD4^+ T cells from the
marginal zone into the WP

   To observe the behavioral pattern of CD4^+ T cells in HCC-m mice, we
   examined the distribution of iNKTs and CD4^+ T cells in the spleen of
   CXCR6^GFP/+ mice. We employed intravital microscopy to observe the
   behavior of iNKTs and CD4^+ T cells at two different time points (0 h
   and 4 h following orthotopic injection of Hepa1-6 cells into the
   liver). At steady state, CD4^+ T cells were found to colocalize with
   iNKTs in the marginal zone without entering the splenic WP
   (Fig. [96]S4A). However, 4 h after injection of Hepa1-6 cells into the
   liver, CD4^+ T cells were observed to gradually entere the splenic WP
   (Fig. [97]4A, B). CD4^+ T cells in the splenic WP of HCC-m mice were
   also verified using Z-stack analysis (Fig. [98]4C, supplementary
   Video [99]S1). To validate this finding, we next intravenously
   administered APC-conjugated anti-CD45 antibodies to identify CD4^+ T
   cells present in the marginal zone. We noticed that following injection
   of the Hepa1-6 cells, the proportion of CD4^+ T cells within the WP
   began to increase after 4 h, whereas the percentage of CD4^+ T cells in
   the marginal zone exhibited a significant decline (Fig. [100]4D, E).
   Furthermore, we also confirmed the characteristic changes in dynamics
   of splenic CD4^+ T cell during HCC metastasis (Fig. [101]4F). It was
   observed that in comparison to the the control group, the crawling
   distance as well as crawling velocity of splenic CD4^+ T cells in HCC-m
   mice were significantly increased, which facilitated their rapid entry
   into the WP (Fig. [102]4G, H). At the same time, HCC-m mice increased
   the number of contacts between splenic iNKT and CD4^+ T cells
   (Fig. [103]4I, supplementary Video [104]S2). Simultaneously, iNKTs in
   HCC-m mice exhibited tight interactions with CD4^+ T cells, whereas
   there was no intercellular contact observed between iNKTs and CD4^+ T
   cells in the control mice (Fig. [105]4J, K).

Fig. 4. Splenic CD4^+ T cells in mice with HCC-m enhanced their interaction
with iNKT and subsequently facilitated migration into the white pulp.

   [106]Fig. 4
   [107]Open in a new tab

   A Intravital microscopic observation of the spleen was carried out 4 h
   after the Hepa1-6 cell was injected orthotopically into the spleen of
   CXCR6^Gfp/+ mice. Fluorescein-conjugated anti-mouse CD4 and CD169
   antibodies were employed to label CD4^+ T cells (Red) and marginal
   zones (Blue). Scale bar, 100 µm. White pulps have been indicated with a
   white dashed line. B Statistical analysis of the number of CD4^+ T
   cells in white pulp in each group. C Representative Z-Stack
   fluorescence co-localization images (X = 240 µm, Y = 240 µm, Z = 40 µm)
   were obtained to analyze the potential interactions between CD4^+ T
   cells and white pulp at the different time points. D Flow cytometry
   analysis of the proportion of CD4 + T cells in the spleen white pulp of
   mice with HCC metastasis. Anti-mouse CD45-APC antibody was injected
   into the tail vein of the mice 1 h before the experiment. APC^- CD4^+ T
   cells were CD4^+ T cells located in the white pulp. E Statistical
   analysis of the percentage of CD4^+ T cell in the white pulp. F
   Determination of the migration trajectories of CD4^+ T cells in the
   control and HCC-m mice after adjusting the starting position. The data
   was collected from 5 mice, and scanned for 10 min at 2-s intervals. A
   red dot represents one CD4^+ T cell. G Statistical analysis of the
   crawling distance and H crawling velocity of mouse splenic CD4^+ T
   cells 4 h after orthotopic injection of Hepa1-6 cells into the liver. I
   Statistical analysis of the number of contacts formed between splenic
   iNKT and CD4^+ T cells within 10 min. J Representative Z-Stack
   fluorescence co-localization images (X = 240 µm, Y = 240 µm, Z = 40 µm)
   were collected to examine the potential interactions between CD4^+ T
   cells and iNKTs in different time points. K Statistical analysis of the
   contact volume between CD4^+ T cells and iNKT in each group. The data
   shown was obtained from three independent experiments with n = 5 mice
   per group. The data has been displayed as the mean ± SEM. *P < 0.05,
   **P < 0.01, ***P < 0.001.

The entry of CD4^+ T cells into the WP depends on IFN-γ secreted by iNKT

   To investigate the role of iNKTs in CD4^+ T cell migration, we
   initially examined the expression of iNKT cell activation markers and
   cytokines in the spleens of HCC-m mice and control mice. Splenic iNKT
   of HCC-m mice was found to significantly upregulate the expression of
   CD69, a well characterized marker of iNKT activation (Fig. [108]S5A,
   B). Moreover, iNKT in HCC-m mice was able to significantly upregulate
   the expression levels of IFN-γ and TNF-α, while displaying no impact on
   the expression of IL-4 (Fig. [109]5A, B). Other types of immune cells
   such as NK cells, B cells, CD4^+ T cells, and macrophages do not
   express IFN-γ (Fig. [110]S5C, D). Interestingly, CD8^+ T cells
   expressed a certain level of IFN-γ under homeostasis, whereas in HCC-m
   mice, CD8^+ T cells were no longer able to express IFN-γ
   (Fig. [111]S5E, F). We also noticed that CD4^+ T cells highly expressed
   IFN-γ receptor α chain, but did not express TNFR I (Fig. [112]5C, D).
   In contrast to control mice, CD4^+ T cells from HCC-m mice exhibited
   increased expression of the IFN-γ receptor α chain (Fig. [113]5C, D).
   We further confirmed that NK cells can also express a certain level of
   IFN-γ α chain (Fig. [114]S5G, H), but only 2-3% of the B cell
   population can express IFN-γ α chain (Fig. [115]S5G, H), and CD8^+ T
   cells and macrophages were not found to express IFN-γ α chain
   (Fig. [116]S5G, H). There was no difference observed in expression of
   IFN-γ α chain by NK cells and B cells between HCC-m mice and the
   control mice (Fig. [117]S5G, H).

Fig. 5. The migration of CD4^+ T cells into the spleen white pulp was
primarily dependent on the IFN-γ/IFN-γ receptor signaling pathway.

   [118]Fig. 5
   [119]Open in a new tab

   A Investigation of IFN-γ, TNF-α, and IL-4 in iNKTs of indicated group
   through flow cytometry. B Statistical analysis of IFN-γ, TNF-α, and
   IL-4 MFI of iNKTs of indicated group. C Determination of expression of
   IFN-γ receptor and TNFR I in CD4^+ T cells of indicated group through
   flow cytometry. D Statistical analysis of IFN-γ receptor and TNFR I MFI
   of CD4^+ T cells of the indicated group. E We periplasmically
   transferred primary iNKT cells (1 × 10^5) derived from WT or IFN-γ^−/−
   mice into the spleens of Ja18^−/− mice, followed by establishment of
   the HCC-M mouse model. Three days later, the fluorescein-conjugated
   anti-mouse CD45 antibody was injected into the mouse tail veins 15 min
   before the mice were euthanized. CD4^+ T cells in the red pulp were
   able to bind to the antibody, while CD4^+ T cells in the white pulp
   were unable to bind to the antibody. Flow cytometry was performed to
   analyze the proportion of CD4^+ T cells located in the white and red
   pulp. F Statistical analysis of (E). G Intravital microscopic
   observation of the spleen was performed 4 h after Hepa1-6 cell were
   injected orthotopically into the spleen of CXCR6^Gfp/+, Jα18^−/− and
   CXCR6^Gfp/+ IFN-γ^−/− mice. Fluorescein-conjugated anti-mouse CD4 and
   CD169 antibodies were then used to label CD4^+ T cells (Red) and
   marginal zones (Blue). Scale bar, 100 µm. White pulps have been
   indicated with a white dashed line. H Statistical analysis of the
   number of CD4^+ T cells in white pulp in indicated group. I Expression
   levels of Stat1 and J CD40L mRNA in CD4^+ T cells of Jα18^−/− mice and
   WT mice with HCC metastasis. K Flow cytometry analysis of expression
   levels of IL-4 in CD4^+ T cells of HCC-m IFN-γ^−/− mice. L Statistical
   analysis of the percentage of IL-4^+ CD4^+ T cells in the indicated
   group. The data shown was obtained three independent experiments with
   n = 5 mice per group. The data has been displayed as the mean ± SEM.
   *P < 0.05, **P < 0.01, ***P < 0.001.

   To investigate the potential relationship between iNKTs and CD4^+ T
   cells in the mouse spleen, we isolated WT and IFN-γ^−/− mouse-derived
   iNKT cells and transferred them into the spleens of Ja18^−/− mice.
   After establishing the HCC-M model, flow cytometry was used to analyze
   the proportion of splenic CD4^+ T cells entering the white pulp. We
   injected fluorescein-conjugated anti-CD45 antibodies into the
   circulatory system through the tail vein. The antibody is able to bind
   to CD4^+ T cells in the red pulp but is unable to contact CD4^+ T cells
   in the white pulp. It was found that IFN-γ-deficient iNKT cells did not
   allow entry of CD4^+ T cells into the white pulp in contrast to WT-iNKT
   (Fig. [120]5E, F). Further, we orthotopically injected Hepa1-6 cells
   into Ja18^−/− mice and subsequently analyzed the behavior of CD4^+ T
   cells in the spleen. After 4 h of injection of Hepa1-6 cells, iNKT
   deficiency caused a significant decrease in the number of CD4^+ T cells
   entering into the WP (Fig. [121]5G, H). Moreover, CD4^+ T cell entry
   into the WP was also inhibited in the spleens of CXCR6^GFP/+ IFN-γ^−/−
   mice (Fig. [122]5G, H). iNKT deficiency also contributed to reduced
   expression of Stat1 in CD4^+ T cells, which is a key transcription
   factors in IFN-γ signaling pathway (Fig. [123]5I). iNKT deficiency also
   caused a substantial decrease in the mRNA level of CD40L in CD4^+ T
   cells (Fig. [124]5J), which can interact with CD86 on the surface of B
   cells to mediate the activation of B cell. We further analyzed the
   protein levels of STAT1 phospho and CD40L on CD4^+ T cells in the
   spleens of Jα18^−/−-HCC-m and WT-HCC-m mice using flow cytometry. As
   expected, Jα18^−/− mice showed lower protein levels of STAT1 phospho
   and CD40L on CD4^+ T cells compared to the WT mice (Fig. [125]S5I, J).
   IFN-γ deficiency also significantly reduced expression of IL-4 by CD4^+
   T cells (Fig. [126]5K, L), which has been reported to be crucial for
   both the survival and function of B cells. Therefore, it was concluded
   that in HCC-m mice, IL-4 secretion by CD4^+ T cells depends primarily
   on the IFN-γ signal provided by iNKTs. However, IFN-γ deficiency did
   not affect the expression of CD86 on B cells (Fig. [127]S5K, L).
   Collectively, these results highlighted the pivotal role of IFN-γ
   signaling pathway in regulating the migration of CD4^+ T cells into the
   WP.

Splenic CD4^+ T cell depletion leads to increased metastatic burden of
hepatocellular carcinoma

   To clarify whether CD4^+ T cells contribute to the inhibition of HCC
   metastasis, we injected a CD4^+ T cell-depletion antibody in situ into
   the mouse spleens (50 µg/3 days, a total of 7 injections). Subsequent
   flow cytometry analysis indicated depletion of ~80% of splenic CD4^+ T
   cells (Fig. [128]6A, B). Splenic CD4^+ T cell depletion exacerbates
   intrahepatic metastasis of hepatocellular carcinoma (Fig. [129]6C, D).
   Furthermore, depletion of splenic CD4^+ T cells in Jα18^−/− mice led to
   more severe hepatocellular carcinoma metastasis (Fig. [130]6C, D).

Fig. 6. Splenic CD4^+ T cell depletion leads to increased metastatic burden
of hepatocellular carcinoma.

   [131]Fig. 6
   [132]Open in a new tab

   A Flow cytometry analysis of the proportion of splenic CD4^+ T cells in
   WT and CD4^+ T cell-depletion antibody intrasplenically injected mice.
   B Statistical analysis of (A) (n = 3). C Representative appearance of
   metastatic liver tumors in WT + Hepa1-6, WT + Hepa1-6 + CD4^+ T cell
   depletion antibody, and Jα18^−/− + Hepa1-6 + CD4^+ T cell
   depletion-antibody mice. Tumor tissue is shown within the dotted boxes.
   D Statistical analysis of (C) (n = 6), ***P < 0.001. **P < 0.01.

CD4^+ T cells regulate splenic B cells from HCC-m mice migrate to the liver
and exert cytotoxic effects on Hepa1-6 cells

   First, 100 µl of EZ-Link Sulfo-NHS-LC-Biotin was injected in situ into
   the spleens of Ctrl and HCC-m mice. This biotinylates only surface
   proteins of whole cells as the negatively charged reagent does not
   permeate cell membranes. The biotin on the surfaces of spleen-derived
   cells was able to bind to fluorescein-conjugated streptavidin but not
   to liver-derived cells. Only 5% of the splenic CD4^+ T cells migrated
   to the liver in both WT and HCC-m mice (Fig. [133]S6A, B) and there was
   no statistical difference between the two groups. Similarly, splenic
   CD8^+ T and iNKT cells showed minimal migration to the liver
   (Fig. [134]S6C–F). To verify whether CD4^+T cells exhibited
   cytotoxicity toward Hepa1-6 cells, we purified CD4^+T cells from the
   spleens of WT and HCC-m mice and co-cultured them with Hepa1-6 cells.
   We used Zombie UV™ dye to analyze the mortality rate of Hepa1-6 cells.
   Zombie UV™ is an amine reactive fluorescent dye that does not penetrate
   live cells but is taken up by cells with compromised membranes. Thus,
   it can be used to assess the live vs. dead status of mammalian cells.
   The use of the dye showed that CD4^+T cells from both groups of mice
   had no direct cytotoxic effects on Hepa1-6 cells (Fig. [135]S6G, H).
   Meanwhile, the expression of CD69, an activation marker, is upregulated
   by hepatic CD4^+ T cells in HCC-m mice (Fig. [136]S6I, J). Although
   hepatic CD4^+ T cells showed basal expression of IL-4, no difference in
   expression was found between HCC-m and Ctrl mice (Fig. [137]S6I, J).
   Furthermore, hepatic CD4^+ T cells did not express either IFN-γ or
   CD40L (Fig. [138]S6I, J).

   Surprisingly, a portion of splenic B cells were observed to migrate to
   the liver, and this proportion was significantly higher in HCC-m mice
   compared to the WT controls (Fig. [139]7A, B). The depletion of CD4^+T
   cells significantly reduced the directed migration of splenic B cells
   to the liver (Fig. [140]7A, B), suggesting that CD4^+ T cells are
   critical for the resistance of splenic B cells to intrahepatic
   metastasis of hepatocellular carcinoma. Meanwhile, the frequency of
   hepatic B cells was significantly increased in HCC-m mice compared with
   control mice, and the depletion of CD4^+ T cells resulted in a
   significant decrease in the frequency of hepatic B cells (Fig. [141]7C,
   D). Co-culture of splenic B cells derived from HCC-m mice and Hepa1-6
   cells resulted in a significant increase in the rate of apoptosis in
   the latter (Fig. [142]7E, F). However, splenic B cells derived from
   control mice did not exhibit cytotoxicity against Hepa1-6 cells,
   strongly suggesting a significant role for CD4^+ T cells in B cell
   resistance to Hepa1-6 cells (Fig. [143]7E, F). Additionally,
   spleen-derived B cells expressed higher levels of IgM compared to B
   cells originating from the liver, suggesting a stronger anti-tumor
   effect (Fig. [144]7G, H). In conclusion, although CD4^+ T cells were
   unable to migrate from the spleen to the liver in large numbers, they
   could regulate the migration of splenic B cells to the liver and were
   able to kill HCC cells through upregulation of the expression of IgM.
   In addition, we observed a significant reduction in CD40, MHC-II, and
   IgM expression in splenic B cells from Jα18^−/− mice (Fig. [145]7I, J).
   These data confirm that B cell activation is closely associated with
   the anti-metastatic cascade orchestrated by iNKT and CD4^+ T cells.

Fig. 7. Splenic B cells from HCC-m mice migrate to the liver and exert
cytotoxic effects on Hepa1-6 cells.

   [146]Fig. 7
   [147]Open in a new tab

   A Flow cytometry was performed to analyze the percentage of
   spleen-derived B cells in the hepatic B cell pool in the Ctrl, HCC-m,
   and HCC-m/CD4^+ T cell-depleted groups. B Statistical analysis of (A).
   C Flow cytometry was used to analyze the frequency of hepatic CD19^+ B
   cells in control mice, HCC-m mice, and CD4^+ T cell depletion HCC-m
   mice. D Statistical analysis of (C). E Hepa1-6 cells (5 × 10^4) were
   co-cultured with splenic B cells (2 × 10^4) from WT or HCC-m mice in
   U-bottomed 96-well plates. After 3 days, the Hepa1-6 cells were
   collected and incubated with Zombie UV^TM dye. The percentage of dead
   Hepa1-6 cells was analyzed by flow cytometry. F Statistical analysis of
   (E). G Flow cytometry was used to analyze the expression levels of IgM
   and MHC-II in liver-derived and spleen-derived B cells. H Statistical
   analysis of (G). I Flow cytometry was used to analyze the expression
   levels of IgM, MHC-II, CD40, and CD25 in B cells in WT and Jα18^−/−
   mice. J Statistical analysis of (I). The data shown was obtained three
   independent experiments with n = 3 mice per group. The data has been
   displayed as the mean ± SEM. ***P < 0.001. **P < 0.01, *P < 0.05. NS:
   non-significant.

Discussion

   Herein, we present evidence that the interaction between splenic iNKT
   and CD4^+ T cells plays a key role in inhibition of HCC metastasis.
   Splenectomy and iNKT deficiency both can exacerbate the metastasis, but
   splenic iNKT cell adoption can significantly decrease the burden of HCC
   metastasis. Further analysis of the phenotypic changes of CD4^+ T cells
   in HCC metastasis model of Jα18^−/− mouse revealed key differences
   between CD4^+ T cells and other immune cell populations. These findings
   indicate that iNKT can interact with CD4^+ T cells in the subtissue
   niche of the spleen and affect the secretion of IL-4 by the latter.
   RNA-seq analysis revealed the impact of iNKT deficiency on the
   transcriptional signature of CD4^+ T cells, which can significantly
   impair the migration of CD4^+ T cells. iNKT and CD4^+ T cells are
   mainly co-localized in the spleen marginal zone, and IFN-γ can serve as
   an important factor in regulating the entry of CD4^+ T cells into the
   WP. Subsequently, splenic B cells migrate to the liver and exert
   cytotoxic effects on Hepa1-6 cells. Collectively, these results
   demonstrated that the cross-talk between iNKT and CD4^+ T cells was a
   critical event in mediating immune signaling between the splenic
   marginal zone and WP, ultimately impacting the progression of local
   invasion in HCC.

   In advanced tumors, both tumor-specific CD4^+ and CD8^+ T cells lack
   effector functions and express inhibitory receptors. This
   dysfunction/exhaustion could be primarily attributed to continued
   exposure (for days to weeks) to various tumor antigens and the
   immunosuppressive tumor microenvironment^[148]30–[149]32. During acute
   infection, 24 h of antigen stimulation was found to be sufficient to
   facilitate CD8^+ T cell proliferation and differentiation into
   functional effector and memory states^[150]33,[151]34. In addition,
   recent studies have also reported that the function of CD8^+ T cells in
   tumor-bearing hosts was severely impaired 12 h after exposure to the
   tumor antigens, and this reduction in effector function can begin as
   early as 6 h^[152]35. Therefore, the initial 12 h after antigen
   exposure can serve as a critical window for determining T cell fate and
   anti-tumor effects.

   iNKT plays a crucial role in mediating CD4^+ T infiltration into the WP
   and IL-4 secretion, which is important for subsequent B cell-associated
   anti-tumor effects. Activated B cells can proliferate and eventually
   differentiate into three distinct functional subpopulations: germinal
   center B cells (GC B cells) which undergo affinity maturation of their
   BCRs, thus ultimately producing the progeny cells with high affinity
   for the corresponding antigens, antibody-secreting B cells (ASCs) and
   memory B cells (MBCs)^[153]36. The characteristic somatic
   hypermutations and relaxed cell cycle checkpoint properties of GC B
   cells increase their susceptibility for the generation of self-reactive
   clones and thereby predispose them to apoptosis as a protective
   mechanism^[154]37. IL-4, (also known as B cell stimulation factor 1)
   has a significant impact on regulating almost every stage of B cell
   fate. IL-4 can effectively stimulate GC B cells by activating the
   BCL-XL—Stat6 axis while protecting them from apoptosis^[155]38,[156]39.
   In addition, IL-4 as also possess the capability to function as a
   potential costimulatory cytokine to induce B cell
   proliferation^[157]40. Exposure of CD40L-stimulated B cells to IL-4 can
   promote their polarization towards antigen-presenting cells^[158]41.
   Although IL-4 is not required for antibody formation, it can serve as
   an important driver of IgE and IgG1 class switching and can thereby
   promote the differentiation of B effector 2 cells^[159]42,[160]43.
   Moreover, IL-4 can profoundly influence the differentiation fate of B
   cells by promoting differentiation of MBC and negatively regulating the
   fate of ASC^[161]44.

   Although immuno-oncology remains T-cell centric, several recent studies
   concur that B-cell-infiltrated tumors have significantly better
   prognosis and enhanced T-cell responses. The focus has been on the B
   cells’ capacity to cross present antigens to T cells when discussing
   their role in enhancing anti-tumor immunity. These B cells are usually
   organized into the tertiary lymphoid structures (TLS) and distributed
   inside or outside the tumor, which are predominantly aggregates of
   immune cells with lymph node-like characteristics that have been
   reported to generate both local and sustained immune
   responses^[162]45,[163]46. B cells in TLS can form germinal centers and
   actively produce different antibodies that can potentially recognize
   tumor-associated antigens^[164]47. Antibodies secreted by B cells can
   effectively recognize extracellular domains present on the surface of
   tumor cells and thereby redirect the cytotoxic activity of both the
   natural killer cells and myeloid cells against the tumor cells^[165]47.
   Interestingly, antibodies also have the potential to facilitate uptake
   of the dying tumor cells by antigen-presenting cells, thereby exposing
   intracellular antigens through antigen presentation^[166]48. In
   addition, tumor cell-derived exosomes can also be effectively
   neutralized by antibodies^[167]49. Moreover, cytokine production by B
   cells can create a microenvironment that could be more permissive to
   effector T cells but detrimental to the tumor growth^[168]50. These
   properties of B cells can enable them to exert substantial anti-tumor
   effects even outside the tumor tissues, which is consistent with our
   current observations.

   Our data reveal that splenic B cells, but not hepatic B cells, play a
   critical role in resisting local invasion of hepatocellular carcinoma.
   On the one hand, the spleen has a larger pool of B cells compared to
   the liver, where hepatic B cells represent only a small fraction of the
   intrahepatic immune cell population. In addition, B cells need to be
   activated in secondary lymphoid organs, such as the spleen and lymph
   nodes^[169]51. Hepatic B cells are predominantly B-2 cells and are
   characterized by the secretion of IgG-type antibodies. A specialized
   subpopulation of B cells exists at the margins of the splenic white
   pulp, termed marginal zone B cells (Mz-B cells). Mz-B cells are the
   main source of preexisting IgM in the body, and are capable of reacting
   to antigens and converting into IgM-secreting plasma cells^[170]52. IgM
   has a greater ability to activate the complement system compared to
   IgG, which in turn exerts a cytotoxic effect on target cells^[171]53.
   Therefore, splenic B cells have a stronger anti-tumor immune effect
   compared to hepatic B cells.

   Overall, our current study proposes a mechanism through which the
   potential interaction of splenic iNKT and CD4^+ T cells can
   significantly limit metastasis of HCC by inducing B cell migration to
   the liver. CD4^+ T cells are primarily responsible for transmitting
   immune signals between the spleen marginal zone and the WP. Additional
   research is required to establish how B cells can effectively clear
   circulating tumor cells, consequently mitigating the development of
   liver metastases Therefore, it is anticipated that therapeutic
   approaches involving the activation of splenic iNKTs or adoptive
   transfer of iNKT to the spleen can effectively suppress the metastasis
   and recurrence of HCC after surgery.

Conclusion

   The findings of this study demonstrated that splenic iNKTs inhibit
   intrahepatic metastasis of hepatocellular carcinoma. Splenic iNKTs
   promote the entry of CD4^+ T cells into the white pulp through the
   secretion of IFN-γ. Subsequently, CD4^+ T cells modulated the migration
   of splenic B cells to the liver where they exert cytotoxic effects on
   Hepa1-6 cells. These findings demonstrating the successful inhibition
   of HCC metastasis by splenic iNKTs are of great significance. These
   results expand the understanding of the immune surveillance mechanism
   of iNKT-mediated metastasis.

Methods

Animal experiments

   All mice were kept in a pathogen-free facility with a 12 h light/dark
   cycle and provided ad libitum access to food and water. All animal
   experiments were conducted according to protocols approved by the
   Animal Ethics Committee of the affiliated Suzhou Hospital of Nanjing
   Medical University (Project number K-2024-047). We have complied with
   all relevant ethical regulations for animal use. Eight-week-old
   wild-type C57BL/6 mice were purchased from ZiYuan Biotechnology Co.
   Ltd. (Hangzhou, China). The IFN-γ^−/− mice were kindly provided by
   Professor Yuanli Chen (Hefei University of Technology, Hefei, China).
   In contrast to other groups who have shown that intrasplenic injection
   of Hepa1-6 cells leads only to the development of liver tumors^[172]54,
   the present study established a mouse model of direct intrahepatic HCC
   metastasis to better simulate the process of HCC metastasis. This model
   was established by the direct orthotropic injection of 5 × 10^5 Hepa1-6
   cells (obtained from the National Collection of Authenticated Cell
   Cultures of China) into the lower left lobe of the liver. After 10 min,
   two-thirds of the liver lobe was removed to ensure that the Hepa1-6
   cells had entered the liver sinusoids but not the liver parenchyma
   (Fig. [173]S7A). Control mice were treated with in situ injection of
   physiological saline into the liver, after which the liver tissue at
   the injection site was excised. The absolute number of mCherry^+
   Hepa1-6 cells within the resected liver tissue was analyzed using flow
   cytometry. The data showed that the resected liver tissue contained
   approximately one-third of the Hepa1-6 cells, indicating that
   approximately two-thirds of the Hepa1-6 cells remained within the mouse
   liver tissue (Fig. [174]S7B). After a specified time, the mice were
   sacrificed for further evaluation. In the mouse splenectomy model, a
   2–3 mm incision was made in the back of the mouse, after which the
   spleen was ligated and excised directly, while in the control mice only
   the skin was incised but the spleen was not removed.

Enrichment of murine splenic iNKTs and B cells

   Single-cell suspensions of spleen cells were prepared as previously
   described^[175]55. For splenic iNKT or B cell enrichment, 5 × 10^6
   splenocytes were collected and incubated with allophycocyanin
   (APC)-conjugated CD1d-PBS57-tetramer or an APC-conjugated anti-mouse
   CD19 antibody for 30 min at 4 °C in the dark. The iNKTs and B cells
   were thereafter isolated using an EasySep™ Mouse APC Positive Selection
   kit (Stem Cell, CA, USA). The purity of the cells after enrichment was
   over 95% (Fig. [176]S7C). For the establishment of the splenic iNKT
   adoptive transfer mouse model, 5 × 10^5 iNKTs were injected
   orthotopically into the spleens of the recipient mice.

   To analyze the distribution of iNKTs after adoptive transfer, 1 × 10^5
   primary iNKT cells were injected in situ into the spleens of Ja18^−/−
   mice. The mice were sacrificed three days later, and the absolute
   number of iNKT cells in the spleens, livers, and peripheral blood was
   analyzed by flow cytometry. The iNKTs in the spleen, liver, and
   peripheral blood accounted for ~61% of the total iNKT cell population
   (~6.1 × 10^4 cells) (Fig. [177]S7D). Additional cells may have been
   lost due to a variety of factors, such as cell death, lung
   sequestration, or random error, among others. Secondly, the spleen
   contained the greatest number of iNKT cells, accounting for ~85% of the
   total (Spleen + Liver + Blood = 100%) (Fig. [178]S7E). The spleen
   retained the vast majority of iNKT cells in our iNKT cell adoption
   transfer model.

Intravital microscopy

   CXCR6^GFP/+ mice (Most iNKT cells express CXCR6, ~76% of GFP^+ cells
   were iNKTs (Fig. [179]S7F)) were anesthetized by intravenous injection
   with 4 mg/kg 2,2,2-tribromoethanol (Sigma-Aldrich, [180]T48402) and
   10 mg/kg xylazine (Sigma-Aldrich, X1126). Under anesthesia, an
   appropriate incision was made in the right posterior part of the mouse
   to expose the spleen. The mice were placed in the right decubitus
   position, spleen was pulled out and exposed on the glass plate. The
   naked abdominal tissues were covered with gauze soaked in saline to
   prevent the dehydration. During the experiment, mice were placed on a
   special plate under a constant temperature control at 37 °C and
   continuously supplemented with the normal saline to prevent the tissue
   dehydration. For intravital imaging of iNKT and CD4^+ T cells, 5 μg of
   fluorescein-conjugated anti-mouse-CD4 and anti-mouse-CD169 antibodies,
   respectively were injected into the tail vein. Thereafter, motion
   trajectories and contact volumes of iNKT and CD4^+ T cells were
   recorded and analyzed using confocal laser scanning microcopy (Zeiss,
   LSM800, Germany). The Z-stack and video data was analyzed and displayed
   through Imaris 9.0.1 and Zen 3.3 software.

Flow cytometry

   Flow cytometry analysis of splenic cells and non-parenchymal cells of
   the liver were performed based on our previously reported
   protocol^[181]56. The mice were euthanized and spleen cells were
   mechanically dissociated with a 70 μm strainer. A single-cell
   suspension was first incubated with anti-CD16 and anti-CD32 antibodies
   (BioLegend, San Diego, CA) to block the Fc receptors. To analyze the
   absolute number of each cell populations, 100 μl of Precision Count
   Beads were added (1 × 10^6/ml beads, BioLegend, San Diego, CA) into the
   samples. The single-cell suspension was then washed and resuspended in
   cold phosphate-buffered saline (PBS), followed by incubation with
   fluorescein-labeled surface marker antibodies for 30 min at 4 °C in
   dark. For intracellular cytokine staining, the various samples were
   pre-treated with PMA/Ionomycin for 4 h (Brefeldin A was added after
   2 h), after which the cells were fixed and permeabilized using the BD
   Fixation/Permeabilization kit (BD, USA) according to the manufacturer’s
   instructions. Sample analysis was performed using a Cytoflex S flow
   cytometer (Beckman Coulter, USA) and the data were analyzed using
   FlowJo V10 software (Stanford University, USA) and Cyto Expert V2.4
   (Beckman Coulter). Details of the antibodies used in the study are
   provided in Table [182]S1. The gating strategy of the cell population
   relevant to this study is demonstrated in Fig. [183]S8.

Splenic cell tracing experiment

   One hundred microliters of EZ-Link Sulfo-NHS-LC-Biotin (Thermo Fisher,
   Cat. # 21335, 1 mg/ml) were injected in situ into the mouse spleens.
   This biotinylates only the surface proteins of whole cells as the
   negatively charged reagent does not permeate cell membranes. The biotin
   on the surfaces of the spleen-derived cells was able to bind to the
   fluorescein-conjugated streptavidin but not to liver-derived cells.
   After 3 days, non-parenchymal liver cells were prepared using the above
   method and incubated with antibodies against surface markers and
   fluorescein-conjugated streptavidin. The percentages of
   streptavidin-positive cells in the populations of interest were then
   determined by flow cytometry.

Mouse CD4^+ T cells isolation and RNA-seq

   CD4^+ T cells were isolated from the mouse spleens (each group
   consisted of 3 mice) using Mojosort^TM CD4 T Cell isolation kit
   according to the manufacturer’s instructions (BioLegend, San Diego,
   CA). We first added appropriate quantity of TRIzol lysis buffer
   (Invitrogen, CA, USA) into the cell sample tube (1 mL TRIzol for
   1 × 10^6 cells) and total RNA was extracted according to the
   manufacturer’s instructions. The transcriptome sequencing and analysis
   were performed by BGI Genomics (Shenzhen, China). Essentially,
   differential expression analysis was carried out using the DESeq2
   (v1.4.5) with Q value ≤ 0.05 (or FDR ≤ 0.001).

Gene Annotation

   To analyze the changes in phenotype, GO and Kyoto Encyclopedia of Genes
   and Genomes (KEGG) enrichment analyses of various annotated
   differentially expressed genes (DEGs) was performed by Phyper based on
   Hypergeometric test. The significant levels of the terms and pathways
   were corrected by P value with a rigorous threshold (P ≤ 0.05). GSEA
   was carried out using GSEA software.

Statistics and reproducibility

   GraphPad Prism software (Version 9.0, USA) was used for all the
   statistical analyses. For the comparisons between the two sets of
   samples, an unpaired t-test with Welch’s correction or an unpaired
   two-tailed Student’s t-test was performed. One-way analysis of variance
   (ANOVA) with a multiple-comparison test was conducted for comparisons
   among the multiple groups of samples. A threshold of P < 0.05 was
   considered as statistically significant.

   For the RNA-seq analysis involved in this study, n = 3 mice were used
   for each group. All samples were independent and groups balanced with
   controls. The sequencing data was filtered with SOAPnuke by (1)
   Removing reads whose low-quality base ratio (base quality less than or
   equal to 15) is more than 20%; (2) Removing reads containing sequencing
   adapter; (3) Removing reads whose unknown base (‘N’ base) ratio is more
   than 5%, afterwards clean reads were obtained and stored in FASTQ
   format. Expression level of gene was calculated by RSEM (v1.3.1). The
   heatmap was drawn by pheatmap (v1.0.12) according to the gene
   expression difference in different samples. Essentially, differential
   expression analysis was performed using the DESeq2 (v1.4.5) with Q
   value ≤ 0.05 (or FDR ≤ 0.001).

Reporting summary

   Further information on research design is available in the [184]Nature
   Portfolio Reporting Summary linked to this article.

Supplementary information

   [185]Supplementary Imformation^ (3.3MB, pdf)
   [186]42003_2025_7798_MOESM2_ESM.pdf^ (77.9KB, pdf)

   Description of Additional Supplementary File
   [187]Supplementary Video 1^ (58.1MB, mp4)
   [188]Supplementary Video 2^ (32.2MB, mp4)
   [189]Reporting Summary^ (2.2MB, pdf)

Acknowledgements