ABSTRACT
-
Background/Aims
Transmembrane 4 L six family member 1 (TM4SF1) is highly expressed and contributes to the progression of various malignancies. However, how it modulates hepatocellular carcinoma (HCC) progression and senescence remains to be elucidated.
-
Methods
TM4SF1 expression in HCC samples was evaluated using immunohistochemistry and flow cytometry. Cellular senescence was assessed through SA-β-gal activity assays and Western blot analysis. TM4SF1-related protein interactions were investigated using immunoprecipitation-mass spectrometry, co-immunoprecipitation, bimolecular fluorescence complementation, and immunofluorescence. Tumor-infiltrating immune cells were analyzed by flow cytometry. The HCC mouse model was established via hydrodynamic tail vein injection.
-
Results
TM4SF1 was highly expressed in human HCC samples and murine models. Knockdown of TM4SF1 suppressed HCC proliferation both in vitro and in vivo, inducing non-secretory senescence through upregulation of p16 and p21. TM4SF1 enhanced the interaction between AKT1 and PDPK1, thereby promoting AKT phosphorylation, which subsequently downregulated p16 and p21. Meanwhile, TM4SF1-mediated AKT phosphorylation enhanced PD-L1 expression while reducing major histocompatibility complex class I level on tumor cells, leading to impaired cytotoxic function of CD8+ T cells and an increased proportion of exhausted CD8+ T cells. In clinical HCC samples, elevated TM4SF1 expression was associated with resistance to anti-PD-1 immunotherapy. Targeting TM4SF1 via adeno-associated virus induced tumor senescence, reduced tumor burden and synergistically enhanced the efficacy of anti-PD-1 therapy.
-
Conclusions
Our results revealed that TM4SF1 regulated tumor cell senescence and immune evasion through the AKT pathway, highlighting its potential as a therapeutic target in HCC, particularly in combination with first-line immunotherapy.
-
Keywords: Transmembrane 4 L six family member 1; Hepatocellular carcinoma; Senescence; Immunotherapy
Study Highlights
• TM4SF1 is highly expressed in HCC and correlates with poor patient prognosis.
• TM4SF1 drives tumor proliferation and inhibits tumor cell senescence by mediating AKT phosphorylation through its interaction with AKT1 and PDPK1.
• TM4SF1 impaired CD8+ T cell anti-tumor function; targeting TM4SF1 induces senescence and sensitizes immunotherapy in preclinical HCC model, indicating a novel therapeutic approach for HCC.
Graphical Abstract
INTRODUCTION
Liver cancer, particularly hepatocellular carcinoma (HCC) which accounts for about 90% of cases, is a significant global health challenge with a high fatality rate. Despite significant advancements in locoregional and systemic therapies, a substantial proportion of patients fail to respond adequately and ultimately succumb to the disease [
1]. Therefore, the development of more effective treatment options remains imperative. Transmembrane 4 L six family member 1 (TM4SF1), known as tumor-associated antigen L6, has the capacity to engage in interactions with integrins, receptor tyrosine kinases, collagens and other proteins, facilitating the formation of tetraspanin-enriched microdomains. These interactions enable TM4SF1 to drive tumor angiogenesis, cell proliferation, and migration, underscoring its critical role in tumor progression [
2,
3]. TM4SF1 is significantly upregulated across multiple cancer types, including HCC [
4]. Although extensively studied in epithelial malignancies, its precise role and underlying mechanisms in HCC have yet to be thoroughly elucidated.
Cellular senescence represents a state of proliferation arrest in response to endogenous or exogenous stressors. This cessation of cell cycle is predominantly facilitated by p16INK4a (encoded by
CDKN1A) and p21CIP1 (encoded by
CDKN2A) [
5]. In addition, senescent cells often exhibit upregulation of p19ARF, p27, p53,and the loss of lamin B1. A subset of senescent cells, known as the senescence-associated secretory phenotype (SASP), actively secretes growth factors, cytokines, chemokines and proteases. Senescence in tumor cells is a double-edged sword. It halts tumor proliferation and SASP-associated cytokines, such as tumor necrosis factor-alpha (TNF-α), can induce reactive oxygen species-dependent apoptosis in tumor cells [
6]. Moreover, senescent tumor cells can induce senescence in neighboring cells through SASPs and direct cell-to-cell interactions [
7,
8]. However, certain SASP-factors including IL-6, IL-8, CXCL5 and MMP7 contribute to tumorigenesis, invasion, migration and immune evasion [
9-
11]. In HCC, senescence in tumor cells has been shown to suppress proliferation and is modulated by various factors like ammonia metabolism, methionine metabolism and histone melthyltransferase [
12-
14]. Conversely, senescence in hepatic stellate cells and hepatocytes contributed to HCC promotion by SASP factors [
15,
16]. These studies suggest the ambiguous role of senescence in HCC, highlighting the need for further investigation to elucidate its context-dependent contributions to tumor progression.
Given the complicated intrinsic features of senescent tumor cells, they simultaneously influence tumor behavior and reshape the immune microenvironment. Through SASP, senescence can drive tumors toward either progression or elimination. However, the impact of non-secretory phenotypic senescence in tumor microenvironment remains poorly understood. In clinical practice, immunotherapies such as anti-programmed death 1 (anti-PD-1) have been widely utilized. The impact of senescence on immunotherapy efficacy remains debatable. Some studies revealed that the efficacy of immunotherapy was significantly affected by senescent tumor cells in a tissue-specific manner [
17,
18]. Notably, senescence in tumor cells has been reported to enhance immune surveillance and the efficacy of immunotherapy in HCC [
19]. However, the interplay between immunotherapy and non-secretory phenotypic senescence, particularly in the context of HCC, has rarely been explored. In this study, we demonstrate the progressive role of TM4SF1 in HCC from the perspective of tumor senescence and immunotherapy for the first time. TM4SF1 inhibits non-secretory phenotypic senescence in liver cancer by recruiting PDPK1 and AKT1 to interact with each other, inducing upregulation of programmed cell death ligand 1 (PD-L1), downregulation of major histocompatibility complex class I (MHC I) in tumors, and promoting immune escape in liver cancer. Targeting TM4SF1 inhibits tumor progression and alleviates the suppressive immune microenvironment in HCC, enhancing immunotherapy efficacy.
MATERIALS AND METHODS
Public datasets and patient samples
GSE14520, GSE10143, GSE76297 and GSE32879 were obtained from the Gene Expression Omnibus. Non-HCC tumor samples were excluded. ERP117672 was obtained from European Nucleotide Archive. RNA-seq and clinical data from The Cancer Genome Atlas (TCGA) were obtained from Genome Data Commons data portal. All samples were collected postoperatively from patients with a confirmed pathological diagnosis of primary HCC who had not undergone any preoperative treatment. The samples were obtained from the Hepatic Surgery Center of Tongji Hospital. The first cohort (Tongji Cohort) consists of 121 clinical patients between 2012 and 2016, in which cancer and adjacent tissues were used to create tissue microarrays for immunohistochemical (IHC) analysis. All these patients had complete follow-up data. The second cohort (prospective cohort 1) includes 50 postoperative samples from HCC patients between 2019 and 2023. Tumor and adjacent tissue were used for IHC analysis and flow cytometry. The third cohort (prospective cohort 2) includes 14 postoperative samples from HCC patients between 2023 and 2024. Tumors and adjacent tissues were used for IHC analysis. Clinical data and magnetic resonance imaging (MRI)/computed tomography (CT) images were acquired from medical record system and classified. Clinicopathological characteristics of patients in the Tongji and prospective cohorts are provided in
Supplementary Tables 1–
3. Sample collection, utilization, storage, and destruction processes were supervised by the Medical Ethics Committee of Tongji Hospital (TJ-IRB20221136).
Male BALB/c nude mice and C57BL/6 mice (both 6–8 weeks old) were purchased from GemPharmatech Co., Ltd (Jiangsu, China). For subcutaneous model, nude mice were injected with 1×106 MHCC97H-luc or HepG2 cells. For orthotopic model, nude mice were injected with 1×10
6 MHCC97H-luc cells in the left lobe of liver. Subcutaneous tumors were measured every 3 days. Nude mice for subcutaneous were sacrificed on day 21. Nude mice orthotopically model were sacrificed on day 30. C57BL/6 mice were used to establish hydrodynamic tail vein injection (HTVi) HCC model. Plasmids were diluted in 2 mL warmed sterile PBS and rapidly injected into tail veins. For HTVi model, mice were sacrificed at day 30 [
20]. In treatment experiment, self-complementary adeno-associated virus (scAAV) were injected intravenously. Anti-PD-1 or IgG were administered intraperitoneally (BE0146 and BE0089; BioXcell, West Lebanon, NH, USA). Before bioluminescent imaging, mice were anesthetized with isoflurane and injected peritoneally with D-Luciferin. Bioluminescence was imaged by IVIS Spectrum (PerkinElmer, Waltham, MA, USA) or LAGO system (Spectral Instruments Imaging, Tucson, AZ, USA). All animal experiments are approved by the Committee on the Ethics of Animal Experiments of Tongji Hospital (TJH-202304029).
All data analysis and figure plotting were performed using GraphPad Prism (version 8.0.2; GraphPad Software Inc., San Diego, CA, USA) or R software (version 4.1.0; R Foundation for Statistical Computing, Vienna, Austria). Values are represented as the mean±standard deviation. The log-rank test was used to assess the prognostic significance of the Kaplan–Meier curves. Significance in bar chart was determined using the Student’s t-test. Correlations were assessed by Pearson correlation analysis.
RESULTS
TM4SF1 promotes the progression of HCC
To explore the role of TM4SF1 in HCC, we conducted analysis in public databases and found a significant upregulation of TM4SF1 expression in tumor tissues across multiple HCC datasets (
Fig. 1A,
Supplementary Fig. 1A). Additionally, in the GES14520 dataset, patients with high TM4SF1 expression had worse overall survival (OS) and disease-free survival (DFS) (
Fig. 1B), and similar results were confirmed in TCGA-LIHC dataset (
Supplementary Fig. 1B). Immunohistochemistry (IHC) analysis of the HCC microarray from the Tongji cohort revealed that tumor tissues exhibited higher TM4SF1 expression than adjacent non-tumor tissues, with its localization predominantly on the cell membrane (
Fig. 1C1). Meanwhile, in prospective cohort 1 from Tongji hospital, flow cytometry analysis of resected surgical samples confirmed the high expression of TM4SF1 in HCC tissues (
Fig. 1C2). In addition, patients with high TM4SF1 expression indicated poorer OS and DFS (
Fig. 1D). These different detection methods consistently demonstrated the high expression of TM4SF1 in HCC cells. To further understand the role of TM4SF1 in HCC, we examined several cell lines including MHCC97H, HepG2, Huh7 and HLF, which exhibit higher malignancy in HCC [
21,
22]. We confirmed elevated TM4SF1 expression in these cell lines relative to normal liver cells (
Fig. 1E). Based on varying expression of TM4SF1, we selected MHCC97H and HLF for further investigation. A lentivirus was used to overexpress TM4SF1, and successful overexpression was confirmed by flow cytometry (
Supplementary Fig. 1C). The effects of TM4SF1 on HCC were investigated. Overexpression of TM4SF1 significantly enhanced tumor cell proliferation, as shown by CCK-8 assay, colony formation assay, EdU assay and cell cycle analysis (
Supplementary Fig. 1D–
G). Additionally, TM4SF1 enhanced the anti-apoptotic ability of tumor cells (
Supplementary Fig. 1H). This proliferative effect was further validated in a nude mouse subcutaneous tumor model and an orthotopic HCC model injected with luciferase-expressing MHCC97H cells (
Supplementary Fig. 1I1,
1I2). In contrast, TM4SF1 knockdown markedly inhibited HCC cell growth
in vitro and
in vivo (
Supplementary Fig. 2A–
F). Overall, these findings indicate that TM4SF1 is overexpressed in HCC tissues and plays a key role in driving tumor progression.
To investigate the mechanism of TM4SF1-driven HCC progression, we performed transcriptome sequencing on TM4SF1 knockdown cell lines. GESA revealed significant enrichment in the Fridman senescence pathway in TM4SF1 knockdown groups.
CDKN1A (encoding p21) and
CDKN2A (encoding p16), markers of cellular senescence, were significantly upregulated following TM4SF1 knockdown (
Fig. 2A). TM4SF1 overexpression decreased the proportion of cells in the G0/G1 phase (
Supplementary Fig. 1G), whereas TM4SF1 knockdown induced G0/G1 phase arrest, a hallmark of senescence (
Supplementary Fig. 2F). To confirm senescence, senescence-associated beta-galactosidase (SA-β-gal) staining was performed, showing an increase in number of senescent cells in TM4SF1 knockdown HCC cells (
Fig. 2B,
Supplementary Fig. 2G). However, SASP factors, including IL-6, IL-8, CCL-2, IL-1β, and IL-1α showed no significant changes following TM4SF1 knockdown (
Supplementary Fig. 2H). We performed IHC analysis on tissues from the orthotopic HCC model. Consistent with the
in vitro results, we observed increased protein levels of p21 and p16 following TM4SF1 knockdown, along with a decreased Ki-67 index (
Fig. 2C). Additionally, SA-β-gal staining confirmed an increase in senescent tumor cells after TM4SF1 knockdown (
Fig. 2C). Further analysis of TM4SF1 and tumor cell senescence in HCC tissue microarrays from the Tongji cohort revealed a negative correlation between TM4SF1 expression and senescence, as evidenced by p21 and p16 IHC staining (
Fig. 2D,
E). These findings suggest that reduced TM4SF1 expression induces tumor cell senescence, leading to slowed proliferation. However, further studies are required to elucidate the mechanisms by which TM4SF1 mediates this effect.
To investigate how TM4SF1 affects HCC progression and cellular senescence, we performed immunoprecipitation mass spectrometry (IP-MS) to identify TM4SF1-interacting proteins. Silver staining confirmed high TM4SF1 expression, and IP-MS revealed AKT1 and PDPK1 as the top interactors (
Supplementary Fig. 3A). Given the crucial roles of the AKT pathway and the molecular chaperone PDPK1 in cellular proliferation and senescence [
23-
25], we hypothesized that TM4SF1 potentially regulated the AKT pathway and influenced tumor cell senescence. First, we validated the interaction between exogenous TM4SF1 and AKT1, as well as PDPK1, by bimolecular fluorescence complementation (BiFC) assays and co-immunoprecipitation (co-IP) (
Fig. 3A,
Supplementary Fig. 3B). Immunofluorescence (IF) revealed the co-localization of TM4SF1 with AKT1 and PDPK1 in HEK293T cells transfected with exogenous plasmids (
Supplementary Fig. 3C1). To detect endogenous interactions, we used cell lines overexpressing Flag-tagged TM4SF1. IF analysis confirmed the interactions between TM4SF1, AKT1, and PDPK1 (
Supplementary Fig. 3C2).
The phosphorylation of AKT1 at S473 residue, facilitated by its binding to PDPK1, is crucial in the proliferation signaling pathway [
26]. The interaction between TM4SF1, AKT1 and PDPK1 implies a link to AKT1 phosphorylation. Knockdown of TM4SF1 in MHCC97H and HLF cells led to a significant decrease in p-AKT-S473, and a significant upregulation of the senescence-related markers p21 and p16 (
Fig. 3B1). These findings were further confirmed in the orthotopic tumor sections (
Fig. 3B2,
Supplementary Fig. 3D). Meanwhile, we assessed the phosphorylation levels of p38, p65, and STAT3, which are key modulators in SASP factors secretion [
27]. However, no significant changes were observed in these pathways (
Supplementary Fig. 3E). Further investigation of the regulatory relationship between PDPK1 and TM4SF1 in AKT1 phosphorylation revealed that PDPK1 did not affect TM4SF1 expression (
Supplementary Fig. 3F). Knockdown of TM4SF1, even with overexpressed PDPK1, inhibited AKT1 phosphorylation and induced cellular senescence (
Supplementary Fig. 3G1,
3H). Conversely, overexpression of TM4SF1 in cells with PDPK1 knockdown did not lead to changes in AKT1 phosphorylation or senescence (
Supplementary Fig. 3G1,
3H). In addition, CCK-8 assay revealed impaired proliferation following TM4SF1 or PDPK1 knockdown (
Supplementary Fig. 3G2). These findings suggest that the regulation of AKT1 phosphorylation and HCC cell senescence by TM4SF1 requires the involvement of PDPK1.
Further examination of the interaction domains between TM4SF1 and AKT1, as well as PDPK1, revealed that AKT1 interacted with the N-terminus of TM4SF1, whereas PDPK1 interacted with the C-terminus (
Supplementary Fig. 4A–
4C). Based on this, we hypothesized that TM4SF1 might mediate the interaction between AKT1 and PDPK1 through different regions. TM4SF1 was overexpressed exogenously in HEK293T cells. The interaction between AKT1 and PDPK1, as assayed by co-IP, confirmed the positive role of TM4SF1 (
Fig. 3C1). Endogenous co-IP detection further confirmed that the overexpression or knockdown of TM4SF1 enhanced or inhibited the binding of AKT1 to PDPK1, respectively (
Fig. 3C2,
Supplementary Fig. 4D). To further validate the regulation of p-AKT-S473 by TM4SF1, we utilized patient-derived organoids (PDO). Expression of Glypican-3 (GPC3) and TM4SF1 was examined in PDO (
Fig. 3D). Organoids were infected with lentivirus to knock down TM4SF1. Decreased Ki-67 index and organoid area indicated impaired proliferation (
Fig. 3E1). Using multiplex immunohistochemistry (mIHC), we detected lower levels of pAKT-S473 and higher levels of p21 and p16 (
Fig. 3E2). Taken together, these results indicate that TM4SF1 mediates tumor proliferation and senescence through the PDPK1/AKT/p21-p16 pathway.
This hypothesis was validated by treating TM4SF1-overexpressing cell lines with MK2206, an AKT1 inhibitor. The results showed that inhibition of the AKT pathway reversed the characteristics of TM4SF1 promoting HCC progression and inducing cellular senescence, in both control and TM4SF1-overexpressing HCC cells (
Supplementary Fig. 5A–
5C). Conversely, the restoration of AKT phosphorylation by exogenous AKT1 supplementation completely reversed the impaired proliferation and senescence induced by TM4SF1 knockdown (
Supplementary Fig. 5D–
5F). These findings were also validated in the orthotopic tumor model (
Fig. 3F1,
3F2,
Supplementary Fig. 5G). In conclusion, our data demonstrates that TM4SF1 recruits PDPK1 and AKT1 to promote tumor cell proliferation while inhibiting senescence.
In tumor progression, the immune microenvironment plays a pivotal role. Given the complex regulatory role of tumor cell senescence in the immune microenvironment, we further investigated TM4SF1 in immunocompetent mouse. First, we found that
Tm4sf1 expression was elevated in tumor tissues compared to that in normal liver tissues across multiple mouse HCC models, with the highest expression level detected in HTVi-induced model by injection of oncogenes (
MYC+
CCND1) (
Supplementary Fig. 6A). Based on HCC model of this combination, we designed a strategy using three shRNA constructs to target and knock down
Tm4sf1 (
Fig. 4A). Bioluminescence intensity and gross tumor images indicated a reduced tumor burden in the
Tm4sf1 knockdown group (
Fig. 4B). mRNA analysis showed that the suppression of
Tm4sf1 triggered an increase in
Cdkn1a and
Cdkn2a expression (
Supplementary Fig. 6B). IHC and SA-β-gal staining further confirmed that silencing TM4SF1 induced senescence in the tumors and slowed tumor progression, consistent with previous findings (
Fig. 4C,
Supplementary Fig. 6C).
Since the role of tumor-derived TM4SF1 within the tumor microenvironment remains unclear, we performed flow cytometry analysis of tumor-infiltrating immune cells. Our results revealed that TM4SF1 knockdown increased the proportion of T cells among the total immune cells (
Fig. 4D1). The proportions of major immune cell subgroups, including myeloid-derived suppressor cells (MDSCs), macrophages and conventional dendritic cells (cDCs), were partially increased. However, the proportions of B cells and plasmacytoid dendritic cells (pDCs) showed no significant changes between the control and knockdown groups (
Fig. 4D1). Analysis of T cell subgroups revealed no significant changes in the ratio of CD8+ or CD4+ T cells. However, CD8+ T cells in the knockdown group exhibited higher proportions of Granzyme B+ and IFN-γ+ T cells, along with fewer PD-1+ T cells, indicating enhanced cytotoxic function and reduced exhaustion (
Fig. 4D2). In TM4SF1 knockdown group, CD8+ T cells transitioned from naive state to effectors (
Fig. 4D2). The immunosuppressive Foxp3+CD4+ subgroup did not show significant changes (
Fig. 4D2). This implies that suppression of TM4SF1 may hinder tumor progression by increasing the number of CD8+ T cells while preserving their cytotoxic function.
Previous studies have reported that elevated AKT phosphorylation promotes PD-L1 expression while downregulating MHC I [
28,
29]. Both are closely related to immune cell recognition and exhaustion, particularly in CD8+ T cells. Therefore, we further investigated whether AKT1 phosphorylation, which is regulated by TM4SF1, affected the expression of PD-L1 and MHC I. In MHCC97H and HLF cell lines, TM4SF1 knockdown reduced PD-L1 expression and increased MHC I expression, whereas TM4SF1 overexpression had the opposite effect (
Fig. 4E,
Supplementary Fig. 6D). In TM4SF1 knockdown HCC cells, the exogenous AKT1 supplementation reversed the effect on PD-L1 and MHC I expression (
Fig. 4F,
Supplementary Fig. 6E). These findings suggest that TM4SF1 not only inhibits tumor senescence and promotes tumor proliferation but also influences the tumor microenvironment by modulating the expression of PD-L1 and MHC I in tumor cells through the AKT-senescence pathway.
The expression of PD-L1 and MHC I in tumor cells directly affects CD8+ T cell function. Therefore, we assessed the impact of tumor-derived TM4SF1 on CD8+ T cells
in vitro. Human CD8+ T cells were isolated and co-cultured with HCC cells (
Fig. 5A). The results demonstrated that, compared to the control group, the survival rate of TM4SF1-overexpressed HCC cells was significantly improved, while the reduction in survival was found in TM4SF1-knockdown HCC cells (
Fig. 5B). These findings suggested that tumor cells with high TM4SF1 expression evaded CD8+ T cell-mediated cytotoxicity. The overexpression of TM4SF1 in tumor cells significantly impaired CD8+ T cell cytotoxicity, evidenced by reduced GZMB, TNF-α, and IFN-γ, and induced exhaustion, while inducing exhaustion, marked by increased PD-1 and TIM-3 expression (
Fig. 5C). Additionally, it notably inhibited the proliferation of CD8+ T cells (
Fig. 5C,
Supplementary Fig. 7A,
7B). These findings demonstrate that TM4SF1 promotes tumor progression by directly impairing the function of CD8+ T cells. Finally, HCC tissues from prospective cohort 1 (n=50) were analyzed by IHC. Patients with low TM4SF1 expression (P-092 and P-081) exhibited higher levels of p16, p21, and MHC I, but decreased p-AKT-S473 and PD-L1 expression. In contrast, patients with high TM4SF1 expression (P-067 and P-096) exhibited the opposite pattern (
Fig. 5D). Furthermore, TM4SF1 expression was positively correlated with proliferation (
Fig. 5E), while p-AKT-S473 level was negatively correlated with p21, p16, and MHC I expression and positively correlated with PD-L1 expression (
Fig. 5F). These findings corroborate our previous results. Therefore, we propose that TM4SF1 compromises the anti-tumor immune response of CD8+ T cells through the PDPK1/AKT1/PD-L1-MHC I pathway.
Previous results elucidated the critical role of TM4SF1 in HCC. To further explore its therapeutic potential, we employed both immunodeficient and immunocompetent mouse models. Firstly, an orthotopic HCC model was established in nude mice. The scAAV-DJ serotype was designed to decrease the TM4SF1 expression in tumors (
Fig. 6A). The results showed that TM4SF1 was successfully knocked down
in vivo (
Supplementary Fig. 8A1). The tumor burden in AAV-treated mice was significantly reduced compared to that in control mice (
Fig. 6B). IHC staining of tumor tissue showed decreased TM4SF1 expression, along with reductions in p-AKT-S473, Ki-67 and PD-L1 levels. Meanwhile, senescence markers p21 and p16 were upregulated and SA-β-gal staining further confirmed the induction of tumor senescence in the treatment group (
Fig. 6C,
Supplementary Fig. 8A2). These findings in the nude mouse model validated the feasibility of AAV-based treatment. Next, we utilized the HTVi-induced HCC model in immunocompetent mice. Two weeks after injecting oncogenes (
MYC+
CCND1), we administered scAAV-8 to knock down TM4SF1 and combined it with anti-PD-1 treatment (
Fig. 6D). Analysis of bioluminescence images (
Fig. 6E1) and tumor burden (
Fig. 6E2) revealed that both anti-PD-1 and AAV treatment alone could suppress tumor progression, with the combined treatment group demonstrating the most potent anti-tumor effect. Serum alanine aminotransferase and aspartate aminotransferase levels indicated a reduction in tumor burden following treatment, with the combination therapy group showing the greatest decrease (
Supplementary Fig. 8B). Successful AAV infection was confirmed by zsGreen staining (
Fig. 6F), and
Tm4sf1 expression was effectively knocked down at the mRNA level. All treatment groups demonstrated increased
Cdkn1a and
Cdkn2a expression, with the combination therapy group showing the most pronounced increase (
Supplementary Fig. 8C). AKT phosphorylation level and Ki-67 index decreased across all treatment groups, with the most pronounced reduction in the combined group. p21 and p16 levels were markedly higher in the combination treatment group than in the monotherapy groups (
Fig. 6F,
Supplementary Fig. 8D). Notably, tumor senescence was observed exclusively in AAV treatment group, and not in immunotherapy alone (
Fig. 6F). These results indicate that targeting TM4SF1 effectively induces senescence to inhibit HCC progression, while the combined therapy with anti-PD-1 enhances this effect synergistically.
To further investigate the synergistic effect of targeting TM4SF1 and anti-PD-1, our analysis focused on T cells, based on previous results that TM4SF1 accelerated CD8+ T cell exhaustion and impaired cytotoxic function. The results revealed that the treatment group had a significantly higher proportion of tumor-infiltrating T cells compared to the control group, with the combined therapy showing a more pronounced increase than either monotherapy (
Fig. 7A1). In other subgroups, the proportions of B cells, macrophages, and cDCs increased following AAV monotherapy, whereas the proportion of pDCs remained unchanged across all treatment groups. Only MDSCs demonstrated a proportional increase following various treatments (
Supplementary Fig. 8E). No significant changes were observed in the proportion of CD8+ or CD4+ T cells across various treatments, which is consistent with previous findings (
Supplementary Fig. 8F). Besides, there was a notable reduction in naïve CD8+ T cells, with a trend of differentiation towards effector T cells in treatment groups, and the most pronounced shift was observed in the combined therapy group (
Fig. 7A1). Moreover, the combination therapy group exhibited enhanced cytotoxic function and a reduced exhaustion status compared to monotherapy groups (
Fig. 7A2). mIHC staining revealed that the combination therapy group had the highest proportion of GZMB+CD8+ cells and the lowest proportion of PD-1+CD8+ cells (
Fig. 7B,
Supplementary Fig. 9A). Collectively, these results demonstrate that the combined treatment enhances anti-PD-1 immunotherapy by increasing the total infiltrating CD8+ T cells and preserving their cytotoxic function.
To translate these findings into clinical practice, we examined prospective cohort 2 (n=14), consisting of HCC patients who underwent radical resection followed by anti-PD-1 immunotherapy. Based on relapse, patients were categorized into responsive and non-responsive groups. Follow-up MRI or CT images were collected (
Fig. 7C,
Supplementary Fig. 9B). mIHC revealed that tumors in the responsive group had a higher density of CD8+ and GZMB+CD8+ T cells, coupled with a lower presence of PD-1+CD8+ T cells (
Fig. 7D,
Supplementary Fig. 9C). The expression of TM4SF1 and its downstream pathways was assessed in both responsive and non-responsive patient groups using IHC. The results showed that responsive patients had lower expression levels of TM4SF1, p-AKT-S473, and Ki-67 (as reported pathologically), while p21 and p16 expression elevated (
Fig. 7E,
Supplementary Fig. 8D). Additionally, PD-L1, which is closely associated with immunotherapy efficacy, was overexpressed in non-responsive patients, whereas MHC I levels were lower (
Fig. 7E). These findings confirm that TM4SF1 plays a critical role in modulating the response to immunotherapy. Furthermore, the ERP117672 data confirmed that non-responsive patients exhibited higher TM4SF1 expression (
Supplementary Fig. 9E). In conclusion, TM4SF1 impairs anti-tumor immunity by modulating CD8+ T cell function; targeting TM4SF1 may offer a novel strategy to sensitize tumors to anti-PD-1 immunotherapy.
DISCUSSION
TM4SF1 has been extensively investigated in epithelial tumors such as colorectal cancer, breast cancer, and pancreatic cancer. TM4SF1, in conjunction with DDR1, reactivated metastasis in breast cancer by a non-canonical pathway [
30]. TM4SF1 activated the JAK2-STAT3 signaling, enhancing cell stemness in colorectal cancer [
4]. Tumor self-seeded cells with high TM4SF1 expression showed enhanced ability of invasion, migration and recolonization [
31]. Unfortunately, due to the difference between HCC and other tumors, similar roles of TM4SF1 have not been well investigated in HCC. In this study, we confirmed elevated TM4SF1 expression in HCC tissues across multiple datasets. Through interactions involving AKT1 and PDPK1, TM4SF1 activated the classical AKT1-PDPK1 pathway and promoted AKT phosphorylation, which subsequently promoted tumor progression. Targeting TM4SF1 induced non-secretory phenotypic senescence and enhanced anti-tumor immunity. Although we successfully demonstrated TM4SF1 as a therapeutic target for HCC, this study has certain limitations. First, while we elucidated the mechanisms through which TM4SF1 promotes HCC progression, the upstream regulatory mechanisms of TM4SF1, such as transcriptional regulation and post-translational modifications, require further exploration, which we plan to address in future studies. Second, regarding model selection, we employed subcutaneous tumors, orthotopic tumors, and a hydrodynamic model with high TM4SF1 expression. Utilizing diverse combinations of hydrodynamic models may provide a more comprehensive understanding of TM4SF1’s functional roles. Third, the impact of tumor-derived TM4SF1 on other components of the tumor microenvironment, beyond CD8+ T cells, requires further investigation.
Senescence has emerged as a novel therapeutic target in cancer treatment. Senescent cells suppress tumor proliferation directly and influence tumor microenvironment by various means including paracrine, autocrine, and cell to cell interaction. AKT hyperactivation-induced senescence by cysteine import and trans-sulfuration protected gastric cancer cells from ROS-induced cell death [
32]. In colorectal cancer, senescent cells at invasive margins promoted epithelial-mesenchymal transition (EMT) and metastasis [
9]. Micro-RNA family prevented EMT-associated senescence in gastric cancer [
33]. Similarly, chemotherapy-induced senescence in breast cancer facilitated immune evasion and subsequent tumor relapse by upregulating PD-L1 and CD80, while doxorubicin-induced senescence in brain metastases recruited exhausted T cells [
17,
34]. These findings underscore the varied impacts of senescence on tumor progression across cancer types. In HCC, senescent tumor cells upregulated IFN-γ receptor, remodeled tumor microenvironment and became hypersensitized to IFN-γ [
20]. Other studies have demonstrated similar findings: glutaryl-CoA dehydrogenase induced HCC senescence by regulating the pentose phosphate pathway and glycolysis, benefiting immunosurveillance through SASP [
35]. The decreased phosphorylation of mitochondrial protein in HCC cells induced senescence and impaired tumor proliferation as well as EMT process [
36]. Interestingly, IL-17-induced AKT dependent IL-6 secretion has been reported to promote HCC progression without inducing senescence [
37]. p21-activated secretory phenotype induced chemokine CXCL14 secretion, which attracted and polarized macrophages into M1, eliminating preneoplastic cells [
38]. Pre-malignant senescent hepatocytes escaped from immunosurveillance by recruiting immunosuppressive myeloid cells on the early onset of HCC, underscoring context-dependent outcomes of senescence [
39,
40]. While similarities exist, our findings uniquely highlight non-secretory phenotypic senescence in HCC mediated by TM4SF1/AKT/p21-p16, which enhanced anti-PD-1 efficacy. In contrast to studies linking senescence to EMT or metastasis, our research found that senescent HCC cells induced by TM4SF1 knockdown demonstrated no discernible changes in SASP factors, including IL-6, IL-8, CCL-2, IL-1β, and IL-1α, which are implicated in tumor invasion, immune evasion and angiogenesis [
27]. Likewise, the phosphorylation levels of pivotal modulators of SASP secretion, including p38, p65, and STAT3, remained unaltered. Within the tumor microenvironment, senescent HCC cells strengthened the cytotoxic function of CD8+ T cells by regulating PD-L1 and MHC I via the AKT pathway. The unaltered phosphorylation levels of p65 and STAT3 suggest a novel AKT-dependent pathway for the regulation of MHC I and PD-L1, which we plan to explore further in future studies. For other components in the tumor microenvironment, increased MDSCs following treatments require further exploration. We observed an increased proportion of CD4+ Tregs following anti-PD-1 monotherapy compared to that in the control group, consistent with other studies [
41,
42]. However, there was no significant change in the Treg subgroup after TM4SF1-targeted AAV treatment, and combination therapy further reduced Tregs compared to anti-PD-1 monotherapy. This phenomenon partially explained why anti-TM4SF1 therapy sensitized immunotherapy, but its underlying mechanism required further investigation (
Supplementary Fig. 8F).
Our findings indicate that TM4SF1 may serve as a therapeutic target in HCC. In studies of other malignancies, TM4SF1-specific antibodies or antibody-drug conjugates have been used to suppress tumor stemness, growth, and vascularization [
4,
43]. TM4SF1-chimeric antigen receptor-T cell therapy exhibited anti-tumor capacity
in vivo [
44]. These systematic targeting strategies may also benefit HCC treatment. In this study, we employed systemic administration of AAV to knock down TM4SF1 in tumors transcriptionally. It can bypass post-translational modification of the target, which potentially compromises the therapeutic effect. However, AAV carries potential risks. In animal studies, it has been reported that intravenous administration of AAV induces hepatotoxicity, endothelial vessel injury, and microvascular thrombosis [
45]. Neurotoxicity and insertional mutagenesis by AAV are also mentioned. In some cases, the immune system recognizes AAV, reduces its efficiency [
45]. AAV also carries the potential risk of off-target effects, which may affect normal cells. These limitations could partly explain the restricted use of AAV in solid tumors, as evidenced by the limited number of clinical trials involving AAV. In our study, we utilized highly efficient and less immunogenic scAAV serotypes to target liver, incorporating a zsGreen tag to monitor infection efficiency. To mitigate these concerns, the precise administration, modification and optimization of AAV to enhance its specificity could help reduce adverse effects and off-target risks [
46].
Importantly, targeting TM4SF1 in HCC not only suppresses tumor progression by inducing tumor cell senescence but also enhances the efficacy of immunotherapy. In clinical practice, assessing TM4SF1 expression in biopsy or surgical specimens could have a significant positive impact on subsequent treatment strategies. TM4SF1 may serve as a marker for predicting immunotherapy response. Patients with low TM4SF1 expression indicate a better immunotherapy response. For patients with high TM4SF1 expression, various approaches to downregulate its levels could be employed to simultaneously inhibit tumor progression and enhance sensitivity to immunotherapy. HCC treatment could benefit from systemic drug delivery as well as locoregional approaches such as transarterial chemoembolization or hepatic arterial infusion chemotherapy. With optimized AAV, it is feasible to utilize TM4SF1-targeted AAV-loaded beads in transarterial chemoembolization and integrate AAV into hepatic arterial infusion chemotherapy to downstage advanced HCC. These strategies provide opportunities to implement AAV-based targeting of TM4SF1 in HCC with greater precision and versatility. Recent advances in multiple cancers have led to the development of therapies targeting senescent cells, which aim to suppress tumors by enhancing the clearance of senescent cells [
47]. We envision that combining TM4SF1 inhibition with therapies that eliminate senescent tumor cells, alongside immunotherapy, may provide a more effective strategy for HCC treatment.
FOOTNOTES
-
Authors’ contribution
WFZ, FRL, YCL and ZBL conceived and designed the study. WFZ, FRL and YCL performed most experiments, assisted by ZZ, HFH, SWN, and HWZ. Patient samples were collected by ZZ, HFH, and SWN and analyzed by WFZ. WFZ, FRL, and YCL contributed to data analysis. WFZ, FRL, and YCL wrote the manuscript. ZGZ, ZBL, and HWZ revised the manuscript. ZGZ, ZBL, and XPC provide administrative help.
-
Acknowledgements
This study was funded by National Natural Science Foundation of China (No.81502530, No.82172976), the Young Scientists Fund of the National Natural Science Foundation of China (No. 82303628, No. 82303494, No. 82103597), China Postdoctoral Science Foundation (2023M741275) and Natural Science Foundation of Hubei Province, China (2024AFA076). Sample collection, utilization, storage, and destruction processes were supervised by Medical Ethics Committee of Tongji Hospital (TJ-IRB20221136). All animal experiments are approved by the Committee on the Ethics of Animal Experiments of Tongji Hospital (TJH-202304029). The authors express gratitude to Ruiqi Mao from Union Hospital, Tongji Medical College, Huazhong University of Science and Technology for providing technical support in confocal and BiFC experiments. The authors appreciate the help of Jingjiao Song from the Experimental Medicine Center, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, China.
-
Conflicts of Interest
The authors have no conflicts to disclose.
SUPPLEMENTAL MATERIAL
Supplementary material is available at Clinical and Molecular Hepatology website (
http://www.e-cmh.org).
Supplementary Figure 1.
TM4SF1 promotes HCC progression. (A) Expression of TM4SF1 in GSE32879. (B) Overall survival and progression-free survival analysis based on differential TM4SFS1 expression in TCGA, using the R survminer package. (C) Confirmation of TM4SF1 overexpression in MHCC97H and HLF by flow cytometry. (D) 1) CCK-8 assay in MHCC97H and HLF after TM4SF1 overexpression. 2) CCK-8 assay in HepG2 and Huh7 transfected with TM4SF1 overexpression plasmid. (E) Colony formation assay in MHCC97H and HLF following TM4SF1 overexpression. (F) EdU assay in MHCC97H and HLF following TM4SF1 overexpression. (G) Cell-cycle analysis of MHCC97H, HLF and Huh7 following TM4SF1 overexpression. (H) Apoptosis assay of MHCC97H and HLF treated with sorafenib (2 μM) for 48 hours following TM4SF1 overexpression, by flow cytometry. (I) 1) Tumor images and tumor volumes following TM4SF1 over-expression in MHCC97H subcutaneous model (n=5); 2) Representative bioluminescence and tumor images with statistical analysis in orthotopic tumor model (n=5). Data are represented as means±standard deviation. The P-values (B) were calculated using log-rank test. The rest of P-values were calculated using the Student’s t-test (unpaired). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; HCC, hepatocellular carcinoma.
Supplementary Figure 2.
Knocking down TM4SF1 suppresses tumor proliferation and induces senescence. (A) Confirmation of TM4SF1 knockdown in MHCC97H and HLF by flow cytometry. (B) 1) CCK-8 assay of MHCC97H and HLF after TM4SF1 knockdown. 2) CCK-8 assay of HepG2 and Huh7 transfected with TM4SF1 siRNAs. (C) Colony formation assay in MHCC97H and HLF after TM4SF1 knockdown. (D) EdU assay in MHCC97H and HLF after TM4SF1 knockdown. (E) Tumor images and tumor volumes of MHCC97H and HepG2 subcutaneous model after TM4SF1 knockdown (n=5). (F) Cell-cycle analysis of MHCC97H, HLF and HepG2 following TM4SF1 knockdown. (G) SA-β-gal staining in HepG2 and Huh7 transfected with TM4SF1 siRNAs. (H) Relative senescence-associated cytokine secretion (IL-6, IL-8, CCL-2, IL-1β, and IL-1α) following TM4SF1 knockdown. Data are represented as means±standard deviation. P-values were calculated using the unpaired Student’s t-test. ns>0.05, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; ns, not significant.
Supplementary Figure 3.
TM4SF1 interacts with AKT1 and PDPK1, and promotes AKT phosphorylation. (A) 1) Verification of co-immunoprecipitation products by silver staining; 2) Top interactors of TM4SF1 identified by mass spectrometry and the peptide fragments of AKT1 and PDPK1. (B) The interaction of TM4SF1-AKT1 and TM4SF1-PDPK1, detected by co-immunoprecipitation in HEK293T cells following transfection with the corresponding plasmids. (C) 1) The co-localization of TM4SF1-AKT1 and TM4SF1-PDPK1 in HEK293T cells following transfecting plasmids, using confocal immunofluorescence demonstrating; 2) The co-localization of endogenous TM4SF1-AKT and TM4SF1-PDPK1 in MHCC97H and HLF cells by confocal immunofluorescence imaging. (D) IHC intensity of p-AKT-S473 in the orthotopic tumor model, related to Figure 3B2. (E) The changes in the phosphorylation levels of p38, p65 and STAT3 following TM4SF1 knockdown. (F) MFI of TM4SF1 following PDPK1 siRNA transfection. (G) 1) Expression level of p-AKT-S473, PDPK1, p21 and p16 in MHCC97H and HLF transfected with TM4SF1 or PDPK1 siRNAs, followed by overexpression of PDPK1 or TM4SF1, respectively; 2) CCK-8 assay of MHCC97H and HLF transfected with TM4SF1 or PDPK1 siRNAs, followed by overexpression of PDPK1 or TM4SF1, respectively. (H) SA-β-gal staining of MHCC97H and HLF transfected with TM4SF1 or PDPK1 siRNAs, followed by overexpression of PDPK1 or TM4SF1, respectively. Data are represented as means±standard deviation. P-values were calculated using the Student’s t-test (unpaired). ns>0.05, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; PDPK1, 3-phosphoinositide dependent protein kinase 1; MFI, Mean fluorescence intensity; IHC, immunohistochemistry; ns, not significant.
Supplementary Figure 4.
TM4SF1 interacts with AKT and PDPK1. (A) Schematic diagram of constructions of truncated AKT1, PDPK1 and TM4SF1. (B) Interactions between truncated TM4SF1 and AKT1. (C) Interactions between truncated TM4SF1 and PDPK1. (D) Interaction between AKT1 and PDPK1 in HLF following the knockdown or overexpression of TM4SF1. TM4SF1, transmembrane 4 L six family member 1; PDPK1, 3-phosphoinositide dependent protein kinase 1.
Supplementary Figure 5.
TM4SF1 promotes AKT phosphorylation. (A) EdU assay in MHCC97H and HLF which were treated with MK2206 (150 nM) for 24 hours on the basis of TM4SF1 overexpression, with statistical analysis. (B) SA-β-gal staining in MHCC97H which were treated with MK2206 (150 nM) for 24 hours on the basis of TM4SF1 overexpression, with statistical analysis. (C) Expression levels of p-AKT-S473, p21 and p16 in MHCC97H and HLF which were treated with MK2206 (150 nM) for 24 hours on the basis of TM4SF1 overexpression. (D) EdU assay in MHCC97H and HLF with AKT1 supplementation on the basis of TM4SF1 knockdown, with statistical analysis. (E) SA-β-gal staining in HLF with AKT1 supplementation on the basis of TM4SF1 knockdown, with statistical analysis. (F) Expression levels of p-AKT-S473, p21 and p16 in MHCC97H and HLF with AKT1 supplementation on the basis of TM4SF1 knockdown. (G) IHC intensity of orthotopic tumor model with AKT1 supplementation (n=5). Data are represented as means±standard deviation in the bar graphs. P-values were calculated using the Student’s t-test (unpaired). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; SA-β-gal, senescence-associated beta-galactosidase; IHC, immunohistochemistry.
Supplementary Figure 6.
TM4SF1 inhibits senescence in vivo and regulates PD-L1 and MHC I expression in HCC cells. (A) Expression of Tm4sf1 in different HCC models. (B) Expression of Tm4sf1, Cdkn1a, and Cdkn2a in different groups (n=5). (C) Changes of Ki-67 index, p-AKT-S473, p16 and p21 in HTVi model, based on IHC staining (n=5). (D) PD-L1 and MHC I expression level in HLF on the basis of TM4SF1 knockdown or overexpression, under control (PBS) or IFN-γ stimulation (40 ng/mL, 24 hours). (E) PD-L1 and MHC I expression level in IFN-γ stimulated HLF (40 ng/mL, 24 hours) with AKT1 supplementation following TM4SF1 knockdown. Data are represented as means±standard deviation. P-values were calculated using the Student’s t-test (unpaired). ns>0.05, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; PD-L1, programmed cell death ligand 1; MHC I, major histocompatibility complex class I; HCC, hepatocellular carcinoma; HTVi, hydrodynamic tail vein injection; IFN-γ, interferon gamma; ns, not significant.
Supplementary Figure 7.
TM4SF1 affects CD8+ T cytotoxic function and induces exhaustion. (A) Cytotoxic function and exhaustion status of CD8+ T cells following co-culture with HLF for 48 hours. (B) Statistical analysis of cytotoxic function and exhaustion status markers of CD8+ T cells. Data are represented as means±standard deviation. P-values were calculated using the Student’s t-test (unpaired). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; GZMB, granzyme B; TNF-α, tumor necrosis factor-alpha; PD-1, programmed death 1.
Supplementary Figure 8.
Treatment targeting TM4SF1 suppresses tumor progression and sensitizes immunotherapy. (A) 1) Validation of TM4SF1 knockdown by scAAV in orthotopic model, by flow cytometry; 2) Statistical analysis of IHC staining and Ki-67 index in AAV-treated orthotopic model (n=6). (B) Serum ALT and AST level in HTVi-induced HCC model with different treatments (n=5). (C) Expression of Tm4sf1, Cdkn1a, and Cdkn2a in different groups (n=5). (D) Statistical analysis of IHC score and Ki-67 index in HTVi-induced HCC model with different treatments (n=5). (E) Proportions of main immune cell subsets in HTVi-induced HCC model with different treatments (n=5). (F) Proportions of CD4+ and CD8+ T cells and Tregs in HTVi-induced HCC model with different treatments (n=5). Data are represented as means±standard deviation. P-values were calculated using the Student’s t-test (unpaired). ns>0.05, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; scAAV, self-complementary adeno-associated virus; IHC, immunohistochemistry; AAV, adeno-associated virus; ALT, alanine aminotransferase; AST, aspartate aminotransferase; HTVi, hydrodynamic tail vein injection; HCC, hepatocellular carcinoma; ns, not significant.
Supplementary Figure 9.
TM4SF1 is negatively correlated with immunotherapy efficacy. (A) Statistical analysis of CD8+ T cell counts and proportion of GZMB+ or PD-1+ cells among CD8+ T cells in HTVi-induced HCC model with different treatments (n=5). (B) Preoperative, postoperative, and post-anti-PD-1 treatment MRI or CT images of the rest of patients in prospective cohort 2. (C) Statistical analysis of CD8+ T cell counts in patients’ tumor slides (n=10). (D) Ki-67 index of patients who received immunotherapy, based on pathological reports (n=7). (E) TM4SF1 mRNA expression in dataset ERP117672. Data are represented as the means±standard deviation. P-values for (A), (C) and (D) were calculated using the unpaired Student’s t-test. P-values for (E) were calculated using unpaired Mann-Whitney test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; GZMB, granzyme B; PD-1, programmed death 1; HTVi, hydrodynamic tail vein injection; HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging; CT, computed tomography.
Figure 1.TM4SF1 promotes HCC progression. (A) Differential expression of TM4SF1 in HCC and non-tumor samples from public datasets. (B) Kaplan -Meier plots of the overall survival and disease-free survival based on differential TM4SFS1 expression, with expression data sourced from GSE14520. (C) 1) TM4SF1 expression assessed by IHC in tumor and adjacent tissues from Tongji cohort with statistical analysis; 2) TM4SF1 expression evaluated by flow cytometry in tumor and adjacent tissues from prospective cohort 1, with statistical analysis. (D) Kaplan-Meier survival curves for overall survival and disease-free survival based on TM4SF1 expression in the Tongji cohort. (E) MFI of TM4SF1 in HCC cell lines and human-liver cell lines. (F) Representative bioluminescence and tumor images of MHCC97H orthotopic tumor model following TM4SF1 knockdown (n=5). Data are represented as means±standard deviation. For TM4SF1 expression in GSE14520 and GSE76497, P-values were calculated using the Student’s t-test (paired). For TM4SF1 expression in GSE10143, P-values were calculated using the Student’s t-test (unpaired). P-values for (E) and (F) were calculated using the Student’s t-test (unpaired). P-values for (C) was calculated using the Student’s t-test (paired). P-values for (B) and (D) were calculated using log-rank test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; HCC, hepatocellular carcinoma; IHC, immunohistochemistry; MFI, Mean fluorescence intensity.
Figure 2.Knockdown of TM4SF1 induces tumor cell senescence both in vitro and in vivo. (A) GSEA enrichment analysis of RNA-seq data and qPCR validation of senescence related genes. (B) SA-β-gal staining of MHCC97H and HLF after TM4SF1 knockdown. (C) Representative images of H&E and IHC staining of p16, p21 and Ki-67, together with SA-β-gal staining in tumor in orthotopic tumor model with statistical analysis of the staining intensity. (D) Representative images of IHC staining of p21 and p16 in Tongji cohort. (E) Correlation analysis of p16 and p21 expression levels with TM4SF1 in Tongji cohort. Data are represented as means±standard deviation. P-values for bar graphs were calculated using the Student’s t-test (unpaired). P-values for (E) were calculated using the Pearson correlation analysis. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; GSEA, gene set enrichment analysis; SA-β-gal, senescence-associated beta-galactosidase; TM4SF1, transmembrane 4 L six family member 1; IHC, immunohistochemistry.
Figure 3.TM4SF1 promotes AKT phosphorylation via interaction with AKT1 and PDPK1. (A) Schematic diagram and representative images of BiFC assay validating interactions among TM4SF1, AKT1 and PDPK1. (B) 1) Western blots of p-AKT-S473 and senescence related protein (p16, p21, γ-H2AX) in MHCC97H and HLF following TM4SF1 knockdown; 2) Representative images of IHC staining of p-AKT-S473 in MHCC97H orthotopic tumor model. (C) 1) Western blots demonstrating the effect of exogenous TM4SF1 on the interaction between AKT1 and PDPK1 in HEK293T cells; 2) Western blots showing the impact of endogenous TM4SF1 on the AKT1-PDPK1 interaction in MHCC97H after TM4SF1 overexpression and knockdown. (D) Representative images of PDO (bright field and H&E) and mIHC staining of GPC3 and TM4SF1. (E) 1) Representative bright field images and IHC staining for Ki-67 in PDOs following TM4SF1 knockdown, with statistical analysis; 2) Representative images of mIHC showing changes of p-AKT-S473, p16 and p21 expression following TM4SF1 knockdown. (F) 1) Representative bioluminescence and tumor images of MHCC97H orthotopic tumor model with AKT1 overexpression after TM4SF1 knockdown with statistical analysis (n=5); 2) Representative images of H&E and IHC staining of p16, p21 and Ki-67, together with SA-β-gal staining in orthotopic tumor. Data are represented as means±standard deviation. P-values were calculated using the Student’s t-test (unpaired). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; AKT1, AKT serine/threonine kinase 1; PDPK1, 3-phosphoinositide dependent protein kinase 1; IHC, immunohistochemistry; PDO, patientderived organoid; GPC3, Glypican-3.
Figure 4.TM4SF1 diminishes anti-tumor immunity mainly through CD8+ T cells. (A) Schematic diagram of the structures of oncogenic and knockdown plasmids injected into mice (B) Representative bioluminescence and tumor images of shNC group and sh3in1 groups, with statistical analysis (n=5). (C) Representative images of H&E and IHC staining of HNF4α, p-AKT-S473, p16, p21 and Ki-67, together with SA-β-gal in HTVi-induced HCC model. (D) 1) Percentages of immune cell subsets within CD45+ cells (n=5); 2) Percentages of CD4+, CD8+ T cells in all T cells and percentages of PD-1+, IFN-γ+, GZMB+, naïve and effector T cells in CD8+ T cells, along with Treg in CD4+ T cells (n=5). (E) PD-L1 and MHC I expression level in MHCC97H after TM4SF1 overexpression or knockdown, under stimulation of control (PBS) or IFN-γ. (F) PD-L1 and MHC I expression level in IFN-γ stimulated MHCC97H (40 ng/mL, 24 hours) with AKT1 supplementation after TM4SF1 knockdown. Data are represented as means±standard deviation. P-values were calculated using the Student’s t-test (unpaired). ns>0.05, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; IHC, immunohistochemistry; SA-β-gal, senescence-associated beta-galactosidase; HTVi, hydrodynamic tail vein injection; HCC, hepatocellular carcinoma; PD-1, programmed death 1; IFN-γ, interferon gamma; GZMB, granzyme B; PD-L1, programmed cell death ligand 1; MHC I, major histocompatibility complex class I; TM4SF1, transmembrane 4 L six family member 1; AKT1, AKT serine/threonine kinase 1.
Figure 5.TM4SF1 induces exhaustion of CD8+ T cells by regulating PD-L1 and MHC I expression. (A) Schematic diagram of the workflow of co-culture experiment. (B) CCK-8 assay of tumor cell viability expressing with varying TM4SF1 levels, following co-culture with CD8+ T cells at different timepoints. (C) Cytotoxic function and exhaustion status of CD8+ T cells following co-culture with MHCC97H. (D) Representative IHC staining images of TM4SF1, p-AKT-S473, p16, p21, PD-L1 and MHC I in prospective cohort 1. (E) Correlations analysis of Ki-67 and p-AKT-S473 expression with TM4SF1 expression in prospective cohort 1 (n=50). (F) Correlation analysis of p21, p16, PD-L1 and MHC I expression with p-AKT-S473 expression in prospective cohort 1 (n=50). Data are represented as means±standard deviation. P-values for bar graphs were calculated using the Student’s t-test (unpaired). P-values for (E) and (F) were calculated using the Pearson correlation analysis. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; PD-L1, programmed cell death ligand 1; MHC I, major histocompatibility complex class I; IHC, immunohistochemistry.
Figure 6.Targeting TM4SF1 suppresses tumor progression in vivo. (A) Schematic diagram of AAV administration in MHCC97H orthotopic model and the structure of scAAV. (B) Representative bioluminescence and tumor images of shNC group and shTM4SF1 group with statistical analysis (n=6). (C) Validation of scAAV-DJ infection (zsGreen) and representative images of H&E and IHC staining for TM4SF1, p-AKT-S473, p16, p21, PD-L1, MHC I, and Ki-67, together with SA-β-gal staining in different groups. (D) Schematic diagram of scAAV-8 and anti-PD-1 administration in HTVi-induced HCC model. (E) 1) Representative bioluminescence images in different treatment groups with statistical analysis (n=5); 2) Representative tumor images in different treatment groups with statistical analysis (n=5). (F) Validation of scAAV-8 infection (zsGreen) and representative images of H&E and IHC staining for p-AKT-S473, p16, p21, and Ki-67, along with SA-β-gal staining in different treatment groups. Data are represented as means±standard deviation. P-values were calculated using the Student’s t-test (unpaired). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; AAV, adeno-associated virus; IHC, immunohistochemistry; PD-1, programmed death 1; HTVi, hydrodynamic tail vein injection; HCC, hepatocellular carcinoma; SA-β-gal, senescence-associated beta-galactosidase.
Figure 7.TM4SF1 impairs anti-PD-1 efficacy via CD8+ T cell function. (A) 1) Percentage of T cells in CD45+ cells and percentages of naïve and effector T cells in CD8+ T cells in different treatment groups (n=5); 2) Percentages of PD-1+, GZMB+ and IFN-γ+ T cells in CD8+ T cells in different treatment groups (n=5). (B) Representative mIHC images of PD-1+, GZMB+ T cells in CD8+ T cells in different treatment groups (n=5). (C) Preoperative, postoperative, and post-anti-PD-1 treatment images of responsive (R) and NR patients. (D) 1) Representative mIHC images of PD-1+ and GZMB+ T cells in CD8+ T cells in R and NR patients; 2) Statistical analysis of Figure 7D1. (E) 1) Representative images of H&E and IHC staining of TM4SF1, p-AKT-S473, p16, p21, PD-L1, and MHC I in R and NR patients; 2) Statistical analysis of IHC scores in R and NR patients (n=7). (F) A schematic model illustrating the mechanism by which TM4SF1 promotes HCC progression and impairs anti-tumor immunity, created with BioRender.com. Data are represented as means±standard deviation. P-values were calculated using the Student’s t-test (unpaired). *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. TM4SF1, transmembrane 4 L six family member 1; PD-1, programmed death 1; NR, non-responsive; mIHC, multiplex immunochemistry; IHC, immunohistochemistry; HCC, hepatocellular carcinoma.
Abbreviations
AKT serine/threonine kinase 1
bimolecular fluorescence complementation
gene set enrichment analysis
hydrodynamic tail vein injection
immunoprecipitation-mass spectrometry
Mean fluorescence intensity
major histocompatibility complex class I
multiplex immunochemistry
3-phosphoinositide dependent protein kinase 1
programmed cell death ligand 1
senescence-associated secretory phenotype
senescence-associated beta-galactosidase
self-complementary adeno-associated virus
transmembrane 4 L six family member 1
tumor necrosis factor-alpha
REFERENCES
- 1. Vogel A, Meyer T, Sapisochin G, Salem R, Saborowski A. Hepatocellular carcinoma. Lancet 2022;400:1345-1362.
- 2. Shih SC, Zukauskas A, Li D, Liu G, Ang LH, Nagy JA, et al. The L6 protein TM4SF1 is critical for endothelial cell function and tumor angiogenesis. Cancer Res 2009;69:3272-3277.
- 3. Yáñez-Mó M, Barreiro O, Gordon-Alonso M, Sala-Valdés M, Sánchez-Madrid F. Tetraspanin-enriched microdomains: a functional unit in cell plasma membranes. Trends Cell Biol 2009;19:434-446.
- 4. Chen G, She X, Yin Y, Ma J, Gao Y, Gao H, et al. Targeting TM4SF1 exhibits therapeutic potential via inhibition of cancer stem cells. Signal Transduct Target Ther 2022;7:350.
- 5. Calcinotto A, Kohli J, Zagato E, Pellegrini L, Demaria M, Alimonti A. Cellular senescence: aging, cancer, and injury. Physiol Rev 2019;99:1047-1078.
- 6. Kim JJ, Lee SB, Park JK, Yoo YD. TNF-alpha-induced ROS production triggering apoptosis is directly linked to Romo1 and Bcl-X(L). Cell Death Differ 2010;17:1420-1434.
- 7. Acosta JC, Banito A, Wuestefeld T, Georgilis A, Janich P, Morton JP, et al. A complex secretory program orchestrated by the inflammasome controls paracrine senescence. Nat Cell Biol 2013;15:978-990.
- 8. Kuilman T, Michaloglou C, Vredeveld LC, Douma S, van Doorn R, Desmet CJ, et al. Oncogene-induced senescence relayed by an interleukin-dependent inflammatory network. Cell 2008;133:1019-1031.
- 9. Park SS, Lee YK, Choi YW, Lim SB, Park SH, Kim HK, et al. Cellular senescence is associated with the spatial evolution toward a higher metastatic phenotype in colorectal cancer. Cell Rep 2024;43:113912.
- 10. Pereira BI, Devine OP, Vukmanovic-Stejic M, Chambers ES, Subramanian P, Patel N, et al. Senescent cells evade immune clearance via HLA-E-mediated NK and CD8(+) T cell inhibition. Nat Commun 2019;10:2387.
- 11. Chen C, Xu ZQ, Zong YP, Ou BC, Shen XH, Feng H, et al. CXCL5 induces tumor angiogenesis via enhancing the expression of FOXD1 mediated by the AKT/NF-κB pathway in colorectal cancer. Cell Death Dis 2019;10:178.
- 12. Wang K, Jiang X, Jiang Y, Liu J, Du Y, Zhang Z, et al. EZH2-H3K27me3-mediated silencing of mir-139-5p inhibits cellular senescence in hepatocellular carcinoma by activating TOP2A. J Exp Clin Cancer Res 2023;42:320.
- 13. Wang Y, Cheng C, Lu Y, Lian Z, Liu Q, Xu Y, et al. β-catenin activation reprograms ammonia metabolism to promote senescence resistance in hepatocellular carcinoma. Cancer Res 2024;84:1643-1658.
- 14. Jin C, Li Y, Su Y, Guo Z, Wang X, Wang S, et al. Novel copper complex CTB regulates methionine cycle induced TERT hypomethylation to promote HCC cells senescence via mitochondrial SLC25A26. Cell Death Dis 2020;11:844.
- 15. Huang Y, Yang X, Meng Y, Shao C, Liao J, Li F, et al. The hepatic senescence-associated secretory phenotype promotes hepatocarcinogenesis through Bcl3-dependent activation of macrophages. Cell Biosci 2021;11:173.
- 16. Yamagishi R, Kamachi F, Nakamura M, Yamazaki S, Kamiya T, Takasugi M, et al. Gasdermin D-mediated release of IL-33 from senescent hepatic stellate cells promotes obesity-associated hepatocellular carcinoma. Sci Immunol 2022;7:eabl7209.
- 17. Uceda-Castro R, Margarido AS, Cornet L, Vegna S, Hahn K, Song JY, et al. Re-purposing the pro-senescence properties of doxorubicin to introduce immunotherapy in breast cancer brain metastasis. Cell Rep Med 2022;3:100821.
- 18. Deng Y, Chen Q, Yang X, Sun Y, Zhang B, Wei W, et al. Tumor cell senescence-induced macrophage CD73 expression is a critical metabolic immune checkpoint in the aging tumor microenvironment. Theranostics 2024;14:1224-1240.
- 19. Chen HA, Ho YJ, Mezzadra R, Adrover JM, Smolkin R, Zhu C, et al. Senescence rewires microenvironment sensing to facilitate antitumor immunity. Cancer Discov 2023;13:432-453.
- 20. Liu F, Liao Z, Qin L, Zhang Z, Zhang Q, Han S, et al. Targeting VPS72 inhibits ACTL6A/MYC axis activity in HCC progression. Hepatology 2023;78:1384-1401.
- 21. Li Y, Tang ZY, Ye SL, Liu YK, Chen J, Xue Q, et al. Establishment of cell clones with different metastatic potential from the metastatic hepatocellular carcinoma cell line MHCC97. World J Gastroenterol 2001;7:630-636.
- 22. Dor I, Namba M, Sato J. Establishment and some biological characteristics of human hepatoma cell lines. Gan 1975;66:385-392.
- 23. Aoki M, Batista O, Bellacosa A, Tsichlis P, Vogt PK. The akt kinase: molecular determinants of oncogenicity. Proc Natl Acad Sci U S A 1998;95:14950-14955.
- 24. Foersch S, Sperka T, Lindner C, Taut A, Rudolph KL, Breier G, et al. VEGFR2 signaling prevents colorectal cancer cell senescence to promote tumorigenesis in mice with colitis. Gastroenterology 2015;149:177-189.e10.
- 25. Jiang Q, Zhang X, Dai X, Han S, Wu X, Wang L, et al. S6K1-mediated phosphorylation of PDK1 impairs AKT kinase activity and oncogenic functions. Nat Commun 2022;13:1548.
- 26. Vivanco I, Sawyers CL. The phosphatidylinositol 3-Kinase–AKT pathway in human cancer. Nat Rev Cancer 2002;2:489-501.
- 27. Schmitt CA, Wang B, Demaria M. Senescence and cancer — role and therapeutic opportunities. Nat Rev Clin Oncol 2022;19:619-636.
- 28. Zheng X, Chen J, Deng M, Ning K, Peng Y, Liu Z, et al. G3BP1 and SLU7 Jointly Promote Immune Evasion by Downregulating MHC-I via PI3K/Akt Activation in Bladder Cancer. Adv Sci (Weinh) 2024;11:e2305922.
- 29. Lastwika KJ, Wilson W, Li QK, Norris J, Xu H, Ghazarian SR, et al. Control of PD-L1 expression by oncogenic activation of the AKT–mTOR pathway in non–small cell lung cancer. Cancer Res 2016;76:227-238.
- 30. Gao H, Chakraborty G, Zhang Z, Akalay I, Gadiya M, Gao Y, et al. Multi-organ site metastatic reactivation mediated by non-canonical discoidin domain receptor 1 signaling. Cell 2016;166:47-62.
- 31. Yang H, Wang H, He Y, Yang Y, Thompson EW, Xia D, et al. Identification and characterization of TM4SF1+ tumor self-seeded cells. Cell Rep 2024;43:114512.
- 32. Zhu H, Chan KT, Huang X, Cerra C, Blake S, Trigos AS, et al. Cystathionine-β-synthase is essential for AKT-induced senescence and suppresses the development of gastric cancers with PI3K/AKT activation. Elife 2022;11:e71929.
- 33. Yu L, Cao C, Li X, Zhang M, Gu Q, Gao H, et al. Complete loss of miR-200 family induces EMT associated cellular senescence in gastric cancer. Oncogene 2022;41:26-36.
- 34. Shahbandi A, Chiu FY, Ungerleider NA, Kvadas R, Mheidly Z, Sun MJS, et al. Breast cancer cells survive chemotherapy by activating targetable immune-modulatory programs characterized by PD-L1 or CD80. Nat Cancer 2022;3:1513-1533.
- 35. Lao Y, Cui X, Xu Z, Yan H, Zhang Z, Zhang Z, et al. GlutarylCoA dehydrogenase suppresses tumor progression and shapes an anti-tumor microenvironment in hepatocellular carcinoma. J Hepatol 2024;81:847-861.
- 36. Tan S, Zhang X, Guo X, Pan G, Yan L, Ding Z, et al. DAP3 promotes mitochondrial activity and tumour progression in hepatocellular carcinoma by regulating MT-ND5 expression. Cell Death Dis 2024;15:540.
- 37. Gu FM, Li QL, Gao Q, Jiang JH, Zhu K, Huang XY, et al. IL-17 induces AKT-dependent IL-6/JAK2/STAT3 activation and tumor progression in hepatocellular carcinoma. Mol Cancer 2011;10:150.
- 38. Sturmlechner I, Zhang C, Sine CC, van Deursen EJ, Jeganathan KB, Hamada N, et al. p21 produces a bioactive secretome that places stressed cells under immunosurveillance. Science 2021;374:eabb3420.
- 39. Eggert T, Wolter K, Ji J, Ma C, Yevsa T, Klotz S, et al. Distinct functions of senescence-associated immune responses in liver tumor surveillance and tumor progression. Cancer Cell 2016;30:533-547.
- 40. Oncology T, Translational Medicine OA. Role of chemokines in the hepatocellular carcinoma microenvironment and their translational value in immunotherapy. Oncology and Translational Medicine 2022;8:1-17.
- 41. Geels SN, Moshensky A, Sousa RS, Murat C, Bustos MA, Walker BL, et al. Interruption of the intratumor CD8+ T cell:Treg crosstalk improves the efficacy of PD-1 immunotherapy. Cancer Cell 2024;42:1051-1066.e1057.
- 42. Kamada T, Togashi Y, Tay C, Ha D, Sasaki A, Nakamura Y, et al. PD-1+ regulatory T cells amplified by PD-1 blockade promote hyperprogression of cancer. Proc Natl Acad Sci U S A 2019;116:9999-10008.
- 43. Visintin A, Knowlton K, Tyminski E, Lin CI, Zheng X, Marquette K, et al. Novel anti-TM4SF1 antibody–drug conjugates with activity against tumor cells and tumor vasculature. Mol Cancer Ther 2015;14:1868-1876.
- 44. Shen Y, Liu G, Zhang Q, Tian X, Ouyang L, Zhang L. Construction of CAR-T cells targeting TM4SF1 and its anti-tumor capacity in ovarian cancer. Immunol Lett 2023;255:1-9.
- 45. Wang JH, Gessler DJ, Zhan W, Gallagher TL, Gao G. Adeno-associated virus as a delivery vector for gene therapy of human diseases. Signal Transduct Target Ther 2024;9:78.
- 46. Li A, Mitsunobu H, Yoshioka S, Suzuki T, Kondo A, Nishida K. Cytosine base editing systems with minimized off-target effect and molecular size. Nat Commun 2022;13:4531.
- 47. D’Ambrosio M, Gil J. Reshaping of the tumor microenvironment by cellular senescence: An opportunity for senotherapies. Dev Cell 2023;58:1007-1021.
, Zhibin Liao1,2,3
