CD137L promotes immune surveillance in melanoma via HLTF regulation

Liang, Long; Zhu, Lin; Li, Xin; Zhou, Wenbin; Zhang, Yanbin; Hung, Mien-Chie; Zhang, Guanxiong; Chen, Yong; Kuang, Xinwei; Su, Juan; Liu, Jing; Chen, Xiang; Liu, Hong

doi:10.1038/s41467-025-63338-w

Download PDF

Article
Open access
Published: 26 September 2025

CD137L promotes immune surveillance in melanoma via HLTF regulation

Long Liang^1,2^na1,
Lin Zhu^1,3^na1,
Xin Li²^na1,
Wenbin Zhou^1,3,
Yanbin Zhang²,
Mien-Chie Hung ORCID: orcid.org/0000-0003-4317-4740^4,5,
Guanxiong Zhang ORCID: orcid.org/0000-0003-4819-6543^1,3,
Yong Chen ORCID: orcid.org/0000-0002-7188-4566⁶,
Xinwei Kuang^1,3,
Juan Su^1,3,
Jing Liu²,
Xiang Chen ORCID: orcid.org/0000-0001-8187-636X^1,3 &
…
Hong Liu ORCID: orcid.org/0000-0001-9976-2985^1,3

Nature Communications volume 16, Article number: 8478 (2025) Cite this article

1095 Accesses
9 Altmetric
Metrics details

Subjects

Abstract

Immune checkpoint blockers (ICBs) have demonstrated substantial efficacy across various malignancies, yet the benefits of ICBs are limited to a subset of patients. Therefore, it is essential to identify novel therapeutic targets. By integrating multi-omics data from cohorts of patients with melanoma treated with ICBs, a positive correlation is observed between tumor CD137L expression and the efficacy of PD-1 blockade. Functionally, CD137L induction in cancer cells significantly enhances anti-tumor immunity by promoting CD8⁺ T cell survival, both in vivo and in vitro. Mechanistically, helicase-like transcription factor (HLTF) is identified as a pivotal transcriptional regulator of CD137L, controlling its expression through phosphorylation of serine at position 398. Therapeutically, the AMPK agonist AICAR (acadesine) as an inducer of CD137L, exhibiting synergistic effects with PD-1 or CTLA-4 blockade. In summary, our findings elucidate a mechanism controlling CD137L expression and highlight a promising combination therapy to enhance the efficacy of ICBs in melanoma. One Sentence Summary: Inducing co-stimulatory immune checkpoint CD137L expression in melanoma cells enhances T cell-mediated anti-tumor immunity.

Overcoming immune checkpoint blockade resistance in solid tumors with intermittent ITK inhibition

Article Open access 21 September 2023

Cancer cell plasticity defines response to immunotherapy in cutaneous squamous cell carcinoma

Article Open access 24 June 2024

Early memory differentiation and cell death resistance in T cells predicts melanoma response to sequential anti-CTLA4 and anti-PD1 immunotherapy

Article 02 June 2021

Introduction

In recent years, immunotherapy has significantly transformed cancer treatment, markedly extending patient survival¹. Immunotherapies primarily activate anti-tumor immune responses by modulating the signaling pathways of immunostimulatory or immunosuppressive molecules². Immune checkpoint inhibitors, such as anti-PD-1 and anti-CTLA-4 antibodies, have shown clinical efficacy in 30–40% of patients with advanced melanoma³. However, a substantial proportion of patients remain unresponsive to these therapies, underscoring the need to identify additional immune checkpoint molecules as therapeutic targets.

The immune checkpoint CD137 (4-1BB, TNFRSF9), a member of the tumor necrosis factor (TNF) receptor superfamily, is an inducible receptor predominantly expressed on activated T lymphocytes and natural killer (NK) cells⁴. Its only known natural ligand, CD137L (4-1BBL, TNFSF9), is expressed on antigen-presenting cells (APCs)⁵. CD137L, a type II transmembrane protein of the TNF superfamily (TNFSF), exists as a trimer in its membrane-bound form. Upon binding to receptors on T cells, it recruits specific TNF receptor-associated factors (TRAF1, TRAF2, and TRAF3) and initiators⁶. TRAF2 and TRAF3 then trigger downstream signaling cascades, including the NF-κB, ERK, JNK, and p38 MAPK pathways, delivering a potent co-stimulatory signal that enhances T cell activation⁷. Over the past few decades, CD137 has been explored as a potential second-generation immuno-oncology target to further activate tumor-specific T cells⁸. Agonistic antibodies targeting CD137 have been proven to effectively reduce tumor growth in mouse models of melanoma, glioblastoma, and lymphoma⁹. However, clinical trials with two CD137-specific monoclonal antibodies (mAbs) were halted due to intolerable hepatotoxicity (urelumab, BMS) or low efficacy (utomilumab, Pfizer)^10,11. A major challenge in advancing CD137 agonists for clinical use lies in minimizing off-target toxicity while maintaining robust anti-tumor efficacy.

In addition to its role as a ligand for CD137, CD137L can also transduce signals via reverse signaling, a process that enables cells expressing CD137L to receive and transmit signals¹². Consequently, CD137L plays distinct roles in different cell types¹³. Recent studies have revealed that while CD137L is primarily expressed in APCs, its aberrant expression in tumor cells has been observed in various cancers, where it may influence tumor progression^14,15,16,17. For example, CD137L promotes drug resistance in pancreatic cancer and supports metastasis in colon cancer^18,19. Understanding the role of CD137-CD137L interactions in specific malignancies is essential before the clinical application of agonistic anti-CD137 antibodies or blockade of CD137-CD137L signaling. However, the function and regulation of CD137L in melanoma remain poorly understood.

This study analyzed data across multiple tumor types and revealed that high CD137/CD137L expression correlates with better immunotherapy response and prognosis. Notably, CD137L expression was lower in melanoma compared to normal pigmented nevi. Exogenous overexpression of CD137L did not influence melanoma cell proliferation in vivo or in vitro but was found to activate T cells to eliminate tumor cells. Furthermore, HLTF was identified as a key regulator of CD137L expression, which binds to the CD137L promoter and represses its expression. AMPK phosphorylates HLTF at Ser398, which activates CD137L expression. Activation of AMPK upregulates CD137L, leading to significant synergistic effects with other immune checkpoint inhibitors in vivo.

Results

CD137L expression levels were correlated with the efficacy of PD-1 mAb therapy in patients with melanoma

Immune response is co-regulated by both inhibitory and co-stimulatory checkpoints. To investigate whether co-stimulatory checkpoints correlate with their sensitivity to immunotherapy, we analyzed the correlation between the TNF superfamily of co-stimulatory checkpoints and sensitivity to anti-PD-1 therapy. We found that the expression level of ligand-receptor group CD137/CD137L (TNFRSF9/TNFSF9) was associated with better overall survival (OS) or progression-free survival (PFS) in 6 independent immunotherapy cohorts, involving melanoma, lung cancer, head and neck cancer (Fig. 1a and Supplementary Table 1). We further compared the performance of CD137L with known biomarkers (baseline levels of PD-L1 and CD8) for distinguishing non-responders and responders with immunotherapy. We observed that CD137L achieved acceptable performance in more datasets, and in most cases, outperformed CD8 and PD-L1(CD274) in the 10 independent cohorts (Fig. 1b). To investigate the relationship between tumor expressed CD137L and immunotherapy with PD-1 mAb, we analyzed the melanoma patient’s samples with immunotherapy, tumor samples from 29 patients with melanoma treated with the PD-1 antibody (Supplementary Table 2). The patient samples were divided into responders and non-responders (Fig. 1c), and the expression levels of CD137L were determined by immunofluorescence assay. The immunofluorescence showed that high-CD137L expression significantly distinguished patient responsiveness to PD-1 mAb therapy in patients (Fig. 1d). Patients with high-CD137L expression also have significantly longer PFS, which can efficiently distinguish between responders and non-responders (Fig. 1e). Our results suggest that CD137L in tumors may be a predictive indicator of the effectiveness of PD-1 antibody therapy.

**Fig. 1: CD137L expression levels were correlated with the efficacy of PD-1 mAb therapy in melanoma patients.**

The expression pattern of CD137L in melanoma was further evaluated by analyzing single-cell RNA sequencing (scRNA-seq) datasets from patients with melanoma, which allowed for the examination of CD137L expression across individual cell populations. It is noteworthy that the percentage of CD137L-positive cells within the melanoma cell population is minimal, and these cells exhibit a lower expression level compared to other cell types. (Fig. 1f). However, a subset of melanoma cells exhibited elevated CD137L expression, indicating increased tumor immunogenicity (Fig. 1lg). Further analysis of samples from 35 patients with melanoma and 24 benign cases from healthy donors revealed significantly lower CD137L expression in patients with melanoma (Fig. 1h). Higher CD137L expression levels were associated with improved overall survival (Fig. 1i), suggesting that elevated CD137L may enhance anti-tumor immunity.

Overexpression of CD137L enhanced cancer immune surveillance both in immune-competent mice and melanoma cells

To investigate the role of CD137L in melanoma, CD137L was overexpressed in A375, SK-MEL-28, and B16F10 cells. However, overexpression did not affect melanoma cell growth (Supplementary Fig. 1A, B). In vivo tumorigenic assays using immunodeficient (SCID) nude mice showed no significant impact of CD137L overexpression on tumor cell proliferation (Fig. 2a, b and Supplementary Fig. 1C). As CD137L serves as a ligand for the co-stimulatory immune checkpoint CD137, its overexpression may modulate immunity, including T cell activation. Tumorigenesis experiments in immunocompetent mice revealed that CD137L overexpression significantly suppressed tumor growth without affecting body weight, and prolonged survival (Fig. 2c–e and Supplementary Fig. 1D). Further analysis of immune cell infiltration in melanoma revealed a positive correlation between CD137L and CD8⁺ T cells (Fig. 2f). Flow cytometric analysis of intratumoral immune cells showed that the CD3⁺ and CD8⁺ T cell populations increased, and the proportion of GZMB⁺ CD8⁺ T cells in the CD137L overexpression group increased significantly, while other cell types did not change significantly, with activated T cells (Ki67⁺) increasing and macrophages decreasing (Fig. 2g–j and Supplementary Fig. 1E, F). Immunofluorescence staining of tumor tissues further confirmed a significant increase in activated CD8⁺ T cell infiltration in the CD137L overexpression group (Fig. 2k, l). Additionally, an in vitro T cell killing assay demonstrated that CD137L knockdown suppressed T cell cytotoxicity against tumor cells (Fig. 2m), whereas CD137L overexpression significantly enhanced T cell-mediated tumor cell killing (Fig. 2n). These results highlight the critical role of CD137L in tumor immunity in melanoma, suggesting that CD137L upregulation may offer a promising therapeutic strategy.

**Fig. 2: Induction of CD137L enhanced cancer immune surveillance both in immune-competent mice and melanoma cells.**

HLTF functioned as a negative regulator controlling CD137L transcription

The results above indicate that CD137L expression is low in melanoma, and its upregulation could represent a promising therapeutic strategy. To investigate the regulatory mechanism underlying CD137L expression in melanoma, this study first assessed the methylation status of the CD137L gene promoter in melanoma versus adjacent non-tumor tissue but found no significant difference (Supplementary Fig. 2A). Then, a DNA pull-down assay was performed to identify proteins that bind to the CD137L promoter region (Supplementary Fig. 2B). Specific protein bands were detected in the CD137L promoter group using coomassie staining, and these proteins were subsequently identified by mass spectrometry (Fig. 3a). A total of 179 proteins were identified in the two bands (Supplementary Dataset 1). Further analysis of TCGA data revealed that HLTF was the most significantly negatively correlated transcription factor with CD137L expression in both the TCGA-SKCM and Xiangya cohorts among the 179 candidates (Fig. 3b, c and Supplementary Fig. 2C, D). This negative correlation was also observed across nearly all cancer types in the TCGA database (Fig. 3d). Experimental validation confirmed that HLTF directly binds to the CD137L promoter in melanoma cell lines (Fig. 3e). Luciferase reporter assays showed that HLTF overexpression inhibited CD137L promoter-driven dual-luciferase expression in a dose-dependent manner (Fig. 3f). Moreover, ChIP assays confirmed endogenous binding of HLTF to the CD137L promoter (Fig. 3g, h). Overexpression of HLTF in melanoma cells resulted in decreased CD137L mRNA and protein levels, as well as reduced membrane-bound CD137L content. Conversely, knockdown or knockout of HLTF in melanoma cells significantly upregulated CD137L expression and increased membrane-associated CD137L levels (Fig. 3i–l, Supplementary Fig. 2E). To further explore the role of HLTF in regulating CD137L-mediated tumor immunity, the relationship between HLTF expression and tumor immune cell infiltration was analyzed. A significant decrease in CD8⁺ T cell levels was observed in the presence of both HLTF and CD137L, other immune cell populations remained unaffected, suggesting that HLTF may play a negative regulatory role in CD8⁺ T cell activation (Fig. 3m and Supplementary Fig. 2F).

To further investigate the regulatory role of HLTF in tumor immunity, in vitro T cell killing assays were conducted. HLTF knockdown significantly enhanced T cell-mediated cytotoxicity against tumor cells, while HLTF overexpression protected tumor cells from T cell-induced killing (Fig. 3n, o). In vivo tumorigenesis assays in immunocompetent mice revealed that knockdown or knockout of HLTF in the B16F10 cell line significantly inhibited tumor growth without affecting body weight and prolonged the survival of mice (Fig. 3p–s and Supplementary Fig. 3A–H). Flow cytometry analysis demonstrated that CD137L expression was upregulated in both the HLTF knockdown and knockout groups (Fig. 3t and Supplementary Fig. 3i). Additionally, CD8⁺ T cell infiltration in tumors was significantly increased in the HLTF knockdown group, and the proportion of GZMB⁺ activated T cells was notably higher (Fig. 3t and Supplementary Fig. 3J). Immunofluorescence analysis further revealed significantly enhanced infiltration of CD3⁺, CD4⁺, CD8⁺, and GZMB⁺ T cells in the HLTF knockout group (Fig. 3u). These results suggest that HLTF acts as a negative regulator of CD137L expression, and its reduction leads to upregulation of CD137L, thereby enhancing anti-tumor immunity and inhibiting melanoma growth.

HLTF regulates CD137L expression via phosphorylation of serine at position 398

To further explore the mechanisms underlying HLTF regulation of CD137L expression, this study analyzed public mass spectrometry data from the PhosphoSitePlus database (https://www.phosphosite.org/) and identified several phosphorylation sites on HLTF, which were predicted to be regulated by different kinases (Fig. 4a, PhosphoNET, http://www.phosphonet.ca/). Subsequently, these kinases were targeted with specific agonists and inhibitors. AMPK kinase agonists, GSK621 and AICAR (5-Aminoimidazole-4-carboxamide ribonucleoside), significantly increased CD137L expression, while the AMPK inhibitor dorsomorphin downregulated CD137L levels of membrane location (Fig. 4b). To determine whether HLTF phosphorylation is regulated by AMPK, immunoprecipitation assays were performed, confirming that phosphorylated AMPK interacts with HLTF (Fig. 4c, d). Further analysis showed that AMPK binds to HLTF at the SNF2-N and ZFC domains (Fig. 4e). Immunofluorescence microscopy revealed co-localization of HLTF and AMPK in the nucleus, and treatment with AICAR enhanced this co-localization (Fig. 4f). Moreover, the binding of phosphorylated AMPK to HLTF was strengthened upon AICAR treatment (Fig. 4g), and HLTF phosphorylation levels were elevated after AICAR administration (Fig. 4h). To pinpoint the phosphorylation site of HLTF by AMPK, serine residues at positions 397, 398, 397/398, 400, 421, and 635 were mutated to alanine to mimic the dephosphorylated state. Notably, mutation of serine at position 398 significantly inhibited AICAR-induced phosphorylation of HLTF (Fig. 4i). Furthermore, DNA pull-down and ChIP assays revealed reduced binding of HLTF to the CD137L promoter following AICAR treatment (Fig. 4j, k). These results suggest that AMPK upregulates CD137L expression by phosphorylating HLTF, thereby diminishing its inhibitory binding to the CD137L promoter.

**Fig. 4: HLTF regulated CD137L expression via phosphorylation of serine at position 398.**

Given AMPK’s regulation of multiple substrates and its close link to immunity, its action through CD137L and HLTF was further investigated. Melanoma cells were treated with AICAR, and transcriptomic analysis revealed increased activation of AMPK-related pathways, including MAPK and cAMP signaling (Supplementary Fig. 4A, B). Analysis of immune checkpoint expression revealed CD137L as the most significantly upregulated checkpoint (Fig. 4l and Supplementary Dataset 2). Treatment of melanoma cell lines with varying concentrations of AICAR showed a dose-dependent increase in CD137L expression at both the protein and mRNA levels, as well as enhanced membrane localization of CD137L, with no effect on the immortalized keratinocyte HaCaT cell line (Fig. 4m–o). Other AMPK agonists, Metformin, and GSK621 also increased CD137L protein of membrane levels (Supplementary Fig. 4C–E). Different concentrations of AICAR-treated immunocompetent tumor-bearing mice significantly inhibited tumor growth and promoted T cell infiltration and activation, while having no significant effect on the content of other cells in the tumor (NK, B, DC, M2 macrophage cells) (Supplementary Fig. 4F–J). Furthermore, T-cell killing assays demonstrated that AICAR significantly enhanced T cell-mediated cytotoxicity (Supplementary Fig. 4K). To further elucidate the function of CD137L in the mechanism of AICAR, CD137L knockout mouse cell lines were utilized to generate tumors and were treated with or without AICAR. The results demonstrated that tumor growth accelerated in the absence of CD137L, the anti-tumor effect of AICAR was ablated in the CD137L knockout group (Fig. 4p, q), and the infiltration and activation of T cells in the tumor diminished in the CD137L knockout group with or without AICAR (Fig. 4r and supplementary Fig. 4L–O). Concurrently, the content of CD137L in B cells, M2 macrophages, and dendritic cells remained largely unaltered following AICAR treatment (Supplementary Fig. 4P). In vitro experimentation yielded analogous outcomes. Pretreatment with AICAR following CD137L knockdown abolished the ability of AICAR to enhance T-cell cytotoxicity (Fig. 4s). Furthermore, HLTF overexpression followed by AICAR treatment revealed that both wild-type HLTF and the S398A mutant inhibited AICAR-enhanced T-cell killing of tumor cells. In contrast, the S398D mutation, where serine 398 was substituted with aspartate to mimic HLTF phosphorylation, restored the ability of AICAR to enhance T cell cytotoxicity, similar to the effect of AICAR treatment alone (Fig. 4t). These results suggest that AMPK-mediated phosphorylation of HLTF plays a pivotal role in alleviating HLTF’s inhibitory effect on CD137L expression.

Synergistic effect of CD137L inducer and PD-1 mAb or CTLA-4 mAb therapy in melanoma preclinical mouse models

The application of CTLA-4 and PD-1 monoclonal antibodies has marked significant progress in tumor immunotherapy, however, there remains potential for further enhancing therapeutic efficacy. Combining different therapeutic agents is a promising strategy to enhance treatment outcomes. As a co-stimulatory checkpoint independent of both the CTLA-4 and PD-1 pathways, CD137L emerges as a potential target for combination therapy. AICAR, a specific AMPK agonist, is currently in clinical use for the treatment and protection against cardiac ischemic injury. To assess the potential of AICAR in melanoma immunotherapy, a low dose of AICAR was combined with clinically approved immune checkpoint inhibitors, anti-CTLA-4 and anti-PD-1 monoclonal antibodies. Results demonstrated that AICAR alone suppressed tumor growth, and, notably, when combined with the anti-CTLA-4 and anti-PD-1 antibodies, tumor growth was significantly more inhibited than with the single-agent treatments in immunocompetent mice (Fig. 5a–c). Histological analysis of major organs (heart, liver, spleen, lung, kidney, and skin) and liver, kidney function, and blood cell counts showed no obvious toxic side effects in the combination treatment group (Supplementary Fig. 5). No significant differences in body weight were observed (Supplementary Fig. 6A). Tumor volume was also markedly reduced in the combination treatment group (Fig. 5d), and the survival of mice was prolonged compared to those receiving single-agent therapies (Fig. 5e). Additionally, immunofluorescence analysis revealed a significant increase in CD3⁺, CD4⁺, CD8⁺ T cells and GZMB⁺ CD8⁺ T cells within the tumor microenvironment in the combination therapy group (Fig. 5f, g). Flow cytometry also showed that CD3⁺, CD4⁺, CD8⁺ T cells and active GZMB⁺ CD8⁺ T cells were significantly increased in the combination of AICAR with CTLA-4 mAb or PD-1 mAb (Fig. 5h). We also tested other types of T cells and other immune cells, and the results showed that M2 macrophages, Treg cells, and exhausted T cells decreased when used in combination, while proliferating T cells increased (Ki67⁺), and there was no significant change in NK cells (Supplementary Fig. 6B–L). These results indicate that AICAR, when combined with CTLA-4 and PD-1 blockade, enhances melanoma inhibition.

Discussion

CD137L, primarily expressed on APCs, co-stimulates T-cell activation through binding to CD137. While it has been identified in tumor cells and plays varying roles across different cancer types, its function in melanoma remains unclear. In this study, activated AMPK can phosphorylate HLTF to enhance CD137L expression, thereby activating T cells to eliminate melanoma cells (Fig. 6). The combination of the AMPK agonist AICAR with CTLA-4 or PD-1 immune checkpoint-blocking antibodies significantly inhibited melanoma growth in immunocompetent mice. Furthermore, CD137L expression signatures may serve as predictive biomarkers for patient responsiveness to anti-PD-1 therapy.

**Fig. 6: HLTF binding at the CD137L promoter inhibits CD137L expression in melanoma.**

Human CD137 and CD137L expression extends beyond immune cells; some tumors exploit the CD137-CD137L pathway to promote tumor progression^16,17,18,19. In certain cancers, such as hepatocellular carcinoma, CD137L activates reverse signaling in tumor cells, inducing IL-8 production upon binding to CD137²⁰. In acute myeloid leukemia, CD137L signaling stimulates the release of the immunosuppressive cytokine IL-10²¹. Additionally, in cutaneous T-cell lymphoma, leukemia, Hodgkin lymphoma, and colon cancer, CD137L expression promotes tumor cell proliferation through reverse signaling initiated by the CD137-CD137L interaction^17,19,21,22. These findings suggest that CD137L may have a dual role in tumor progression: it can promote tumor development in high-CD137L-expressing tumors while having a different functional impact in tumors with low CD137L expression. In our melanoma model, CD137L expression was lower in melanoma cells compared to nevus cells, and overexpression of CD137L did not impact melanoma growth in vitro or in immunodeficient mice. Interestingly, however, overexpression of CD137L in melanoma cells significantly inhibited tumor growth and activated T cells in immunocompetent mice. Additionally, elevated CD137L levels enhanced T-cell-mediated tumor cell killing in vitro. These results suggest that the low expression of CD137L in melanoma cells contributes to immune escape. In contrast to other tumor types, the reverse signaling pathway mediated by CD137L does not appear to play a significant role in melanoma. Thus, boosting CD137L expression in melanoma cells represents a promising therapeutic strategy. However, this approach may not be universally beneficial across all cancers, and the dual nature of CD137L signaling, particularly its immune-related effects, should be carefully considered in therapeutic strategies.

HLTF, a member of the SWItch/Sucrose Non-Fermenting (SWI/SNF) family, plays a critical role in chromatin remodeling²³. It has been implicated in gene transcription, DNA repair, and the maintenance of genome stability, all of which support its tumor suppressor function²³. Identified as transcription factors due to their ability to bind specific DNA sequences, HLTF target genes in humans typically harbor a conserved cis-element: (A/G)G(T/C)(G/T)G and (C/A)C(T/A)TN(T/G)^24,25. This conserved motif is also present in the CD137L promoter. Across species, HLTF exhibits similar roles in DNA repair and chromatin remodeling, with its deletion severely impairing post-replication and double-stranded DNA repair mechanisms. Such disruptions in genomic stability trigger apoptosis^{26,27,28,29,30}, further validating HLTF as a tumor suppressor gene. Hypermethylation of HLTF is frequently observed in cancers such as colorectal cancer, gastric cancer, and non-small cell lung cancer. However, its role varies across tumor types. In hypopharyngeal squamous cell carcinoma and hepatocellular carcinoma, elevated HLTF expression correlates with poorer prognosis. Despite these insights, its function in tumors within the context of immune system involvement, particularly in melanoma, remains unexplored. In this study, HLTF was found to bind the CD137L promoter, suppressing its expression. Knockdown or knockout of HLTF resulted in increased CD137L expression and enhanced T-cell activation. These findings suggest that elevated HLTF expression may contribute to immune evasion in melanoma, indicating that HLTF does not function as a tumor suppressor in this context.

Previous studies have shown that HLTF can be regulated by ubiquitination³¹, but its phosphorylation regulation remains unexplored. Our data identified multiple phosphorylation sites in HLTF, with Ser398 being specifically regulated by AMPK. Phosphorylation of HLTF by AMPK releases the inhibition of CD137L expression, while AMPK agonists also significantly upregulate CD137L and synergize with CTLA-4 or PD-1 blocking antibodies. AMPK, a key regulator of energy metabolism and immunity^32,33, has been implicated in tumor immunity. For instance, metformin, a non-specific AMPK agonist, has been recognized as an immune stimulator that enhances the efficacy of immune checkpoint blockade through multiple mechanisms³⁴. AMPK activation by metformin induces phosphorylation and abnormal glycosylation of PD-L1, promoting its endoplasmic reticulum-associated degradation (ERAD) and reducing membrane localization^35,36, thereby boosting cytotoxic T lymphocyte activity. Metformin can also promote CD8 T cell proliferation and cytokine production by altering T cell metabolism^37,38. Preclinical studies demonstrate that metformin synergizes with anti-CTLA-4 and anti-PD1 antibodies, improving tumor control and survival in murine models^36,39,40,41. However, metformin has many other effects that directly affect the immunomodulatory properties in related diseases through AMPK-dependent and AMPK-independent mechanisms^41,42. Therefore, whether its combination with ICB is effective in clinical practice requires further investigation. Our findings indicate that CD137L knockdown in melanoma cells, in the absence of IFNG-induced PD-L1 expression, significantly attenuates AICAR-induced T-cell cytotoxicity against tumor cells, CD137L knockout also attenuates the effect of AICAR on increasing T-cell infiltration and activation in vivo, suggesting that AICAR activates T-cell anti-tumor immunity by upregulating CD137L. AICAR, a cAMP analog, serves as a more specific AMPK agonist compared to metformin in treating cardiovascular diseases⁴³. AICAR significantly enhances the efficacy of CTLA-4 or PD-1 antibodies, indicating its potential as a potent immunomodulator to boost anti-tumor immunity. However, activating AMPK to upregulate CD137L may not be universally applicable across all tumor types. AMPK activation or loss of function can have diverse effects on tumor progression. For instance, in melanoma, mutated B-RAF (V600E) inhibits LKB1, leading to AMPK inhibition and promoting transformation⁴⁴. In breast cancer, AMPK activation phosphorylates and inactivates ACC1/2, maintaining NADPH homeostasis and promoting cell survival and progression⁴⁵. AMPK phosphorylation of PDHA is essential for PDHc activation and the TCA cycle, promoting breast cancer metastasis⁴⁶. Therefore, the use of AMPK activation for anti-tumor therapy must consider its broader effects on both tumor progression and the immune microenvironment.

Despite the dramatic improvement in patient survival with the clinical application of PD-1 antibodies, the overall response rate remains suboptimal, highlighting the need for reliable biomarkers to predict treatment efficacy. AICAR-induced upregulation of CD137L significantly enhances the anti-tumor effect of PD-1 antibodies, suggesting a correlation between CD137L expression and PD-1 treatment response. Tumor analysis further revealed that patients with elevated CD137L expression exhibited better prognoses after PD-1 mAb therapy. While PD-1 mAb prevents T-cell exhaustion, CD137L promotes T-cell activation and survival, potentially contributing to improved patient outcomes. However, clinical trials evaluating CD137-specific monoclonal antibodies, and first-generation agonists have been halted due to adverse effects like intolerable hepatotoxicity (urelumab) and low efficacy (utomilumab)^10,11. Urelumab was discontinued due to hepatotoxicity, primarily caused by the induction of S100A4⁺ macrophage infiltration in the liver. Soluble S100A4 activates the Akt pathway, prolonging CD8⁺ T-cell survival and leading to liver damage⁴⁷. Effective signaling of TNFRSF receptors typically requires ligand binding and subsequent receptor hyperaggregation. Since utomilumab inhibits the 4-1BB/4-1BBL interaction, it fails to induce 4-1BB hyperaggregation, resulting in functional inefficacy and poor monotherapy outcomes⁴⁸. Nevertheless, utomilumab has shown promise when combined with other antibodies, with improved tolerability and efficacy⁴⁹. The clinical development of 4-1BB agonists for cancer immunotherapy has evolved significantly. Second-generation agents, over 40 of which entered Phase 1/2 trials since 2017, employ innovative strategies to overcome these limitations. These include bispecific/trispecific formats targeting tumor antigens (e.g., HER2, PD-L1) or stromal markers (e.g., FAP) to enable localized 4-1BB activation^50,51,52, Fc-silencing to reduce systemic toxicity, and epitope optimization (e.g., non-4-1BBL-competing CRD4 binders)^53,54. Some utilize conditional activation (e.g., ATP-dependent binding) or uneven binding-site ratios to enhance clustering^55,56,57. The combination of PD-1 mAb and CD137 mAb could enhance therapeutic efficacy and response rates, but it may also introduce significant side effects. A promising clinical approach could involve using low-toxicity small molecules or drugs, such as AICAR or metformin, to activate AMPK or selectively upregulate CD137L expression in tumor cells. This strategy, when combined with immune checkpoint inhibitors, may offer a viable therapeutic option for tumors by reversing CD137L signaling and counteracting AMPK inactivation.

Methods

Ethical statement

All human and animal studies were conducted in compliance with the relevant ethical guidelines and approved by the appropriate ethics committees. The study protocol (protocol number: 202103617) was approved by the Institutional Review Board of Xiangya Hospital, and all tissue samples were collected in accordance with the informed consent policy. Written informed consent was obtained from all participants, and the study was conducted in compliance with the principles outlined in the Declaration of Helsinki.

Cell culture and treatment

All cell lines obtained from ATCC (American Type Culture Collection, Manassas, VA, USA) were confirmed by STR DNA fingerprinting and mycoplasma were negative. The human melanoma cell lines A375 and SK-MEL-28 were cultured in DMEM medium supplemented with 10% FBS, 100U of penicillin and 100 μg/mL streptomycin. Mouse melanoma cell line B16F10 and human immortalized keratinocyte cell line HaCaT were cultured in RPMI1640 medium. AICAR was added to the complete medium at the indicated concentrations and time.

Regents and antibody

Anti-p-172 Thr AMPK (2535), anti-AMPK (2532) from Cell Signaling Technology. anti-HLTF (14286-1-AP) and GAPDH (60004-1-Ig) from Proteintech. Anti-Granzyme B antibody (ab4059), anti-CD8 (ab217344), anti-CD4 (ab183685), and anti-human CD137L (ab254385) from Abcam. Anti-mouse CD137L (14-9056-82) from Invitrogen. AICAR was purchased from Selleck (S1802), in vivo mAb anti-mouse CTLA-4 (BE0164), anti-PD-1(BE0146) and isotype antibody (BE0086 and BE0089) from Bioxcell.

RNA isolation and Q-PCR assay

Total RNA was isolated from cultured cells by using a standard protocol of Trizol (Invitrogen). A total of 1 µg RNA was reverse-transcribed using the SuperScript III First-Strand cDNA synthesis system (Life Technologies) according to the manufacturer’s instructions. Standard Quantitative PCR was performed using SYBRGreen and CD137L primer (F:5’CCTGAGCTACAAAGAGGACAC3’, R: 5’CAAGTGAAACGGAGCCTGA3’) or GAPDH primer. All mRNA expression levels were normalized to GAPDH and calculated using the 2^−△△CT method.

Dual-luciferase reporter assay

A375 or SK-MEL-28 cells were seeded in 24-well plates and transfected with 0.5 μg/well luciferase reporter plasmids, 10 ng of pRL-CMV (Renilla luciferase) co-transfected with or without HLTF WT or mutant, after 48 h transfection, the luciferase activity was detected by using Dual-Luciferase Reporter Assay System Kit (E1910, Promega, Madison, WI, USA) according to the manufacturer’s instruction.

DNA pull-down assay

The CD137L promoter sequence was obtained from the UCSC database, 2000 bp before the translation start site (chr19:6, 529, 037-6, 531, 037), biotin-labeled double-stranded DNA probe was synthesized by PCR with biotin-labeled primers (forward primer) (F:5’CCTCCAGCAGGCCAGAGCAG3’, R:5’TGCGGGAAGACAC AGCGCGC3’). Nuclear proteins were extracted from both melanoma cell lines. The biotin-labeled CD137L promoter probe (2 μL) was added to the nuclear lysate 500 μg, 500 μL) with 20 μL streptavidin-agarose beads (S1638, Sigma, St. Louis, MO) and a control was added with the same sequence promoter without biotin labeled. After incubation overnight at 4 °C with rotation, the beads were washed three times with lysis buffer and were eluted by boiling at 100 °C for 10 min. The pull-down proteins were separated by polyacrylamide gel electrophoresis and detected by silver staining.

ChIP assay

The ChIP assay was conducted using Simple ChIP@ Enzymatic Chromatin IP kit (9003, Cell Signaling Technology) following the manufacturer’s instructions. Briefly, about 1 × 10⁷ cells were harvested, then cross-linked by 1% formaldehyde and digested chromatin with micrococcal nuclease. Next, the break nuclear membrane is used by several ultrasonic pulses. Chromatin was immunoprecipitated by either control IgG or HLTF (2 μg), or Flag (14793, Cell Signaling Technology, 2 μg) or Myc (2267, Cell Signaling Technology, 4 μg) primary antibody. After washing and reverse cross-linking, the eluted DNAs were quantified by qPCR. The primer sequences were listed below: CD137L primer #1 (F:5’GGGAATCAGAGACAGAGAGAAAC3’, R:5’CTCCCTCTCTTTATGCATCTCTG3’); primer #2 (F:5’GCGACAGAG ATAACAGAAGCA3’, R:5’CTCTGACTCTCTCTCCCTATCT3’); primer #3 (F: 5’CAGAGGCACGCATAGACATAA3’, R:5’CCCGTTAACGTCTTTCTCTGT3’).

Western blotting

Cells were lysed with RIPA buffer (Beyotime, China) supplementary with protease and phosphatase inhibitor (Roche, Isere, France) after PBS washing. The protein concentration was determined using a Pierce BCA protein assay kit (Thermo Fisher Scientific, MA, USA). Equal amounts of proteins were loaded onto polyacrylamide gels. Whole-cell extracts were subjected to 10% sodium dodecylsulfate-polyacrylamide gel electrophoresis and transferred to a nitrocellulose(NC) membrane,followed by blocking with 5% nonfat milk in PBS and then probed with the indicated primary antibodies overnight at 4 ℃.After washing with PBS-Tween 20, the membrane was incubated with horseradish peroxidase-conjugated goat anti-mouse IgG(Jackson ImmunoResearch,cat. no.111-005-003)or goat anti-mouse IgG(Jackson ImmunoResearch, cat.no.115-005-003). Then, an enhanced chemiluminescence assay was carried out to display specific protein bands.

Mass spectrometry analysis

Protein samples were extracted from silver-stained electrophoretic bands of DNA-pull-down assays. Two biological replicates (n = 2) were analyzed, with bead control samples included. Excised gel bands were destained in 50% acetonitrile/25 mM ammonium bicarbonate, reduced with 10 mM DTT (56 °C, 30 min), alkylated with 55 mM iodoacetamide (room temperature, 20 min), and digested with sequencing-grade trypsin (Promega, 1:20 w/w, 37 °C, 16 h). Peptides were extracted with 60% acetonitrile/1% trifluoroacetic acid, lyophilized, and resuspended in 0.1% formic acid.

The lyophilized peptide fractions were resuspended in dd H₂O containing 0.1% formic acid, and 2 μL aliquots of which were loaded into a nanofiber C18 (Acclaim PepMap 100, 75 μm × 2 cm) trap column. The online Chromatography separation was performed on the Easy nLC 1200 system (Thermo Fisher). The trapping and desalting procedure was carried out with a volume of 20 μL 100% solvent A (0.1% formic acid). Then, an elution gradient of 5-38% solvent B (80% acetonitrile, 0.1% formic acid) in 60 min was used on an analytical column (Acclaim PepMap RSLC, 75 μm × 25 cm C18-2 μm 100 Å). The LC/MS data were analyzed for protein identification and quantification using PEAKS Studio 8.5.

Immunofluorescence

Cells were fixed with 4% paraformaldehyde for 10 min, after three times PBS wash, the cell was permeabilized by 0.1% Triton X-100 for 10 min. After washing the slide blocked with bovine serum albumin for 1 h. The indicated primary antibodies were incubated overnight at 4 °C, followed by incubation with FITC or Cy3-conjugated secondary antibody for 1 h at RT. The nuclei were stained with DAPI (Sigma). Viewer images were visualized using a confocal microscope (Zeiss LSM 510, Germany).

Fluorescent multiplex immunohistochemistry

For mouse tumor tissues, 4 mm paraffin sections of patient samples were baked for 120 min at 60 °C and then deparaffinized. Antigen was retrieved at EDTA antigen retrieval buffer (pH 9.0) and maintained at a sub-boiling temperature for 15 min in a microwave oven. After spontaneous fluorescence quenching, the samples were blocked in 3% BSA, PBS with 0.25% Triton X-100 for 1 h at room temperature. Primary antibodies targeting (followed CD3: ab16669 1:800, CD8a: ab217344 1:800, Granzyme B: ab4059 1:400, or CD137L: ab64912) was incubated 1 h at RT in the blocking solution and the following incubated with MACH2 Rb HRP-Polymer (RHRP520H) or MACH2 M HRP-Polymer (MHRP520H) for 10 min at RT. After washing in TBST, was followed by incubation with fluorophore antibody (CD3 Opal690 1:200, CD8 Opal620 1:200, and GZMB Opal520 1:200), respectively. After extensive washing in PBS-0.25% Triton X-100, antigen was retrieved at EDTA antigen retrieval buffer (pH 9.0) and maintained at a sub-boiling temperature for 15 min in a microwave oven. Then incubate the second primary antibody, repeat until all antibodies have been incubated. Next, the slides were mounted with Prolong Diamond Antifade Mountant with DAPI (p36962, Thermo Fisher). Images were detected and captured by PhenoImager HT (Akoya Biosciences).

Immunoprecipitation

For immunoprecipitation (IP), cells were harvested and lysed in cold IP lysis buffer (P0013, Beyotime, China) supplemented with protease and phosphatase inhibitor (Roche, France), the lysis was incubated with primary antibody at 4 °C overnight then add protein A/G agarose beads (Invitrogen) for 2–4 h at 4 °C. After three washes (PBS with 0.1% Triton X-100), bound proteins were eluted by boiling with a 2×SDS loading buffer.

For Co-IP, the cells were first transfected with plasmids with tagged (Flag or GFP) protein genes. After indicated treatment, the cells were harvested and lysed. The supernatants were incubated with anti-tag-agarose beads (anti-Flag, A4596, Sigma, St. Louis, MO, USA; or anit-GFP Biolink, 20 µL,) overnight at 4 °C, then the precipitates were washed three times (PBS with 0.1% Triton X-100) and analyzed by Western blotting.

Generate knockout/knockdown or overexpression stable cells

Using pLentiCRISPRv2 to generate HLTF Knockout. The targeting sequences are as follows. HLTF KO mouse sgRNA: (#1) GACTGCGAATACTTCCAAAC, (#2) TATTCGCAGTCTGTCCAGTA.

shRNA purchased from Sigma-Aldrich, CD137L human shRNA: (#1) AGCTACAAAGAGGACACGA, (#2) TGGACAGAGTCCGAATCCT. HLTF human shRNA: (#1) CAGTTTGGACAACATATAA, (#2) CCGCTTTCTGTGTTAAGCA.

HLTF mouse shRNA : (#1) CCTATATAATGGACAACAA, (#2) TCCAAGACTTTCATACTCA.

For overexpression, CD137L (human CCDS12169.1, mouse CCDS28926.1) forms an Ensembl database, constructed onto the pCDH vector.

To package lentivirus, the overexpression or LentiCRISPR KO constructs were transfected into HEK 293T cells with two packaging plasmids (psPAX2 and the pMD2.G plasmid). The medium was changed at 16 h after transfection. The supernatant was collected at 24 and 48 h after transfection. A375, SK-MEL-28, or B16F10 cells (60% confluency) were incubated in a lentivirus-containing medium with polybrene (10 μg/mL; Selleck Chemicals). Cells were selected by using puromycin (1 μg/mL, Selleck Chemicals) after 48 h of infection for 7 days to test the efficiency.

T cell-mediated tumor cell killing assay

Human peripheral blood mononuclear cells (PBMC) were cultured in ImmunoCult™-XF T cell expansion medium (10981, STEMCELL Technologies) with IL-2 (10 ng/mL, PeproTech, Rocky Hill, NJ, USA) for 1 week before treatment the ImmunoCult™ Human CD3/CD28/CD2 T cell activator (10970, STEMCELL Technologies) add-in culture medium for 48 h. Cancer cells were allowed to adhere to the plates and treated with the indicated concentration of AICAR overnight and then incubated with activated T cells for 48 h. The ratios between cancer cells and activated cells is 1:2.5 or 1:5. T cells and cell debris were removed and washed with PBS. The living cells were then quantified by a spectrometer at OD (570 nm) followed by crystal violet staining.

Tumor immune cell profile analysis by FACS

The tumor of mice was excised and divided into two parts for histopathological analysis and flow cytometry analysis. Samples were then filtered through a 70 μm cell strainer to obtain single-cell suspensions. Cells were first blocked with TruStain FcX™ (anti-mouse CD16/32, BioLegend, 101320) and stained with Zombie Aqua™ Fixable Viability Kit (BioLegend, 423102) to exclude dead cells. For surface staining, cells were incubated with the following antibodies for 20 min at 4 °C: APC-Cy7 anti-mouse CD45 (BD Biosciences, 557659), APC anti-mouse CD3 (BioLegend, 100236), BB700 anti-mouse CD4 (BD Biosciences, 566407), PE-Cy7 anti-mouse CD8a (BD Biosciences, 552877), PE anti-mouse CD137L (BioLegend, 107105), BV421 anti-mouse CD279/PD-1 (BD Biosciences, 562584), BV785 anti-mouse CD366/TIM-3 (BioLegend, 119725), BV421 anti-mouse/human CD19 (BioLegend, 115538), BV605 anti-mouse NK1.1 (BioLegend, 108740), APC anti-mouse CD25 (BioLegend, 102012), FITC anti-mouse CD11c (BioLegend, 117306), PerCP-Cy5.5 anti-mouse I-A/I-E (BioLegend, 107626), BV605 anti-mouse/human CD11b (BioLegend, 101235), APC anti-mouse F4/80 (BioLegend, 123116), and PE-Cy7 anti-mouse CD206/MMR (BioLegend, 141720). For intracellular staining, cells were fixed and permeabilized using True-Nuclear™ Transcription Factor Buffer Set (BioLegend, 424401), followed by staining with PE anti-mouse FOXP3 (eBioscience, 12-5773-82), PE/Dazzle594 anti-human/mouse Granzyme B (BioLegend, 372216), BV605 anti-mouse/human Ki67 (BioLegend, 151215), AF488-TCF1/TCF7 (CST, 6444), and TOX antibody (eBioscience, 12-6502-82). Flow cytometry analysis was performed using FACSymphony™ A3, and data were analyzed using FlowJo software (version 10.8).

Clinical data and sample collection

Paraffin sections of tumor tissue from 29 melanoma patients treated with anti-PD-1 mAb at Xiangya Hospital between March 2018 to January 2022. All patients were followed for more than 6 months. The study protocol was approved by the Institutional Review Board of Xiangya Hospital. All tissue samples were collected in compliance with the informed consent policy. Patients were stratified into response groups based on RECIST (response evaluation criteria in solid tumors) 1.1 criteria, patients with a complete response (CR), partial response (PR), or stable disease (SD) with progression-free survival (PFS) longer than 6 months were classified as responders, while patients with SD with PFS shorter than or equal 6 months and PD were categorized as non-responders.

Mouse tumor generation and implantation

Five to six weeks female C57BL/6 mice were purchased from Hunan STA Laboratory Animal Co., LTD and quarantined for one week before use. Animal care and experiments involved in this study were performed following the Accreditation of Laboratory Animal Care International guidelines and approved by the Animal Care and Use Committee of the Xiangya Hospital of Central South University (Changsha, Hunan, China). About 5 × 10⁵ B16F10 cells were suspended in 100 µL PBS, and inoculation of tumor cells was performed on the dorsal or lateral side of the mouse. Nearly 1 week later, mice were pooled and randomly divided into several groups. The tumor sizes were measured twice a week. Tumor volume (TV) was calculated as TV (mm³) = (length × width²)/2. Animals were sacrificed, and the tumor burden was greater than 2500 mm³ according to the Animal Care Guidelines. These mice were humanely euthanized after measurement. When a mouse needed to be euthanized immediately to terminate the experiment, we also made a record on that day. The combination therapy group was treated with AICAR (300 mg/kg, ip), anti-mouse CTLA-4 mAb (initial dose 200 µg/mouse, followed by 100 µg/mouse every 3 days), anti-mouse PD-1 mAb (initial dose 200 µg/mouse, followed by 100 µg/mouse every 3 days) combination therapy, or a single drug only. IgG isotype was given to the control group every three days for a total of four injections.

RNA-seq analysis

Tumor cell line SK-MEL-28 with or without AICAR treatment were collected and performed with polyA+ RNA sequencing. Construction of RNA libraries was generated by QuantSeq Rev 3’ mRNA-Seq Library Prep Kit for Illumina® (NEB, USA) according to the manufacturer’s recommendations. Then, PCR products were purified (AMPure XP system), and library quality was assessed on an Agilent Bioanalyzer 2100 system. Finally, the library preparations were sequenced with 150 bp paired-end reads on the Novaseq 6000 platform. After FastQC, reads satisfied data quality were aligned to the human reference genome (GRCh38) using hisat2⁵⁸ with the default parameters. Gene expression was quantified using stringtie⁵⁹ with gencode_v28 annotation. Differential analysis was performed by DESeq2⁶⁰.

Survival analysis

Kaplan–Meier survival analysis and multivariable cox proportional hazards regression was performed by R package survival¹⁹. The optimal cutoff point between gene expression and survival of patient was determined using Maxstat method in R package survminer. The area under the receiver operating characteristic curve (AUC) was calculated to evaluate the performance of predicting immunotherapy response based on the expression levels of TNFSF9/TNFRSF9, PD-L1(CD274) and CD8. Gene set expression levels were calculated by GSVA³⁰. Only tumor samples collected prior to immunotherapy treatment were included.

DNA methylation analysis

Tumor and matched normal samples for 41 melanoma patients from Xiangya Hospital were collected and stored at −80 °C. The genome-wide DNA methylation was measured using the Illumina Infinium Human Methylation EPIC Bead Chip following the standard protocol provided by Illumina. The methylation data were processed using the Chip Analysis Methylation Pipeline (ChAMP) package⁶¹. After loading the raw image data from the IDAT files by ChAMP, we further filtered the probes based on the following criteria: (1) probes with a detection p value of >0.01 in at least one sample; (2) probes with a bead count of <3 in at least 5% of samples per probe; and (3) multi-hit probes. The beta values were calculated and normalized by beta mixture quantile dilation (BMIQ) method⁶². For comparing methylation of CD137L promoter between tumor and control samples, we first calculated the average signal of probes in the promoter, including, and then performed Wilcoxon’s rank-sum test.

Single-cell RNA-seq analysis

We perform single-cell transcriptomic analysis using the R package Seurat (4.1.1)⁶³. Data were filtered with below criteria: cells with more than 20% reads in mitochondria or with less than 200 features. Normalization and scaling were performed with default parameters. Uniform manifold approximation and projection (UMAP) was used in scRNA-seq for dimension reduction and visualization. We annotated the cell clusters with SingleR (1.8.1) and followed by curation. Further, clusters with small cells (<50) were removed.

Quantification and statistical analysis

All quantitative data were expressed as mean ± standard deviation (SD) derived from a minimum of three independent experimental replicates (mouse tumor growth curve using data ± standard error of the mean, SEM). Statistical comparisons were performed using appropriate methods as specified in respective figure legends: Wilcoxon rank-sum test, unpaired two-tailed Student’s t-test, or one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Survival outcomes were assessed through Kaplan–Meier curve analysis with log-rank testing for group differences, complemented by univariate Cox regression modeling to determine hazard ratios (HR). All reported p values are explicitly annotated in the corresponding figures. Experimental designs employed three to six biological replicates (samples/animals) per group, with detailed replication schemes provided in figure legends. No datasets were excluded during analysis. To ensure objectivity, investigators were blinded during both data acquisition and analytical phases. All statistical computations and graphical representations were generated using GraphPad Prism 10.0 (GraphPad Software, Inc.) and the R programming language (v3.6.0; https://www.r-project.org).

Reporting summary

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

Data availability

Transcriptomics data of 33 cancer types were downloaded from The Cancer Genome Atlas (TCGA) data portal (https://portal.gdc.cancer.gov/). In-house data of melanoma were available from our previous study (He_SignalTransductTargetTher_2022 and Farshidfar_NatCommun_2022 cohort)^64,65. Transcriptomics data of immunotherapy cohorts were collected from multiple public databases (Supplementary Table 1). Processed single-cell transcriptome data for melanoma were retrieved from GEO under accession number GSE189889(43). The infiltrations of immune cells for TCGA were calculated by ImmuCellAI⁶⁶. Source data are provided with this paper.

Code availability

All scripts used to analyze the data in the MATERIALS AND METHODS section are available via Code Ocean at https://codeocean.com/capsule/2891918/tree/v3.

References

Carlino, M. S., Larkin, J. & Long, G. V. Immune checkpoint inhibitors in melanoma. Lancet 398, 1002–1014 (2021).
Article CAS PubMed Google Scholar
Mellman, I., Coukos, G. & Dranoff, G. Cancer immunotherapy comes of age. Nature 480, 480–489 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Seidel, J. A., Otsuka, A. & Kabashima, K. Anti-PD-1 and anti-CTLA-4 therapies in cancer: mechanisms of action, efficacy, and limitations. Front. Oncol. 8, 86 (2018).
Article PubMed PubMed Central Google Scholar
Etxeberria, I., Glez-Vaz, J., Teijeira, Á & Melero, I. New emerging targets in cancer immunotherapy: CD137/4-1BB costimulatory axis. ESMO Open 4, e000733 (2020).
Article PubMed PubMed Central Google Scholar
Melero, I., Hervas-Stubbs, S., Glennie, M., Pardoll, D. M. & Chen, L. Immunostimulatory monoclonal antibodies for cancer therapy. Nat. Rev. Cancer 7, 95–106 (2007).
Article CAS PubMed Google Scholar
Choi, B. K. & Lee, H. W. The murine CD137/CD137 ligand signalosome: a signal platform generating signal complexity. Front. Immunol. 11, 553715 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zapata, J. M. et al. CD137 (4-1BB) signalosome: complexity is a matter of TRAFs. Front. Immunol. 9, 2618 (2018).
Article PubMed PubMed Central Google Scholar
Hashimoto, K. CD137 as an attractive T cell co-stimulatory target in the TNFRSF for immuno-oncology drug development. Cancers 13, 2288 (2021).
Chu, D. T. et al. An update on anti-CD137 antibodies in immunotherapies for cncer. Int. J. Mol. Sci. 20, 1822 (2019).
Segal, N. H. et al. Results from an integrated safety analysis of urelumab, an agonist anti-CD137 monoclonal antibody. Clin. Cancer Res. 23, 1929–1936 (2017).
Article ADS CAS PubMed Google Scholar
Segal, N. H. et al. Phase I study of single-agent utomilumab (PF-05082566), a 4-1BB/CD137 agonist, in patients with advanced cancer. Clin. Cancer Res. 24, 1816–1823 (2018).
Article CAS PubMed Google Scholar
Kim, H. J. et al. Reverse signaling through the costimulatory ligand CD137L in epithelial cells is essential for natural killer cell-mediated acute tissue inflammation. Proc. Natl Acad. Sci. USA 109, E13–E22 (2012).
CAS PubMed Google Scholar
Zeng, Q., Soe, Y. M., Lim, Y., Sobota, R. M. & Schwarz, H. CD137 ligand interacts with CD32a to trigger reverse CD137 ligand signaling. Cell Mol. Immunol. 17, 1188–1189 (2020).
Article CAS PubMed PubMed Central Google Scholar
Kang, S. W. et al. Anti-CD137 suppresses tumor growth by blocking reverse signaling by CD137 ligand. Cancer Res. 77, 5989–6000 (2017).
Article CAS PubMed Google Scholar
Qian, Y. et al. CD137 ligand-mediated reverse signaling inhibits proliferation and induces apoptosis in non-small cell lung cancer. Med. Oncol. 32, 44 (2015).
Article PubMed Google Scholar
Cheng, K. et al. CD137 ligand signalling induces differentiation of primary acute myeloid leukaemia cells. Br. J. Haematol. 165, 134–144 (2014).
Article CAS PubMed Google Scholar
Kamijo, H. et al. Aberrant CD137 ligand expression induced by GATA6 overexpression promotes tumor progression in cutaneous T-cell lymphoma. Blood 132, 1922–1935 (2018).
Article CAS PubMed Google Scholar
Glorieux, C. & Huang, P. Regulation of CD137 expression through K-Ras signaling in pancreatic cancer cells. Cancer Commun.39, 41 (2019).
Google Scholar
Grimmig, T. et al. Expression of tumor-mediated CD137 ligand in human colon cancer indicates dual signaling effects. Oncoimmunology 8, e1651622 (2019).
Article PubMed PubMed Central Google Scholar
Wang, Q. et al. Analysis of CD137 and CD137L expression in human primary tumor tissues. Croatian Med. J. 49, 192–200 (2008).
Article CAS Google Scholar
Baessler, T. et al. CD137 ligand mediates opposite effects in human and mouse NK cells and impairs NK-cell reactivity against human acute myeloid leukemia cells. Blood 115, 3058–3069 (2010).
Article CAS PubMed Google Scholar
Ho, W. T. et al. Expression of CD137 on Hodgkin and Reed-Sternberg cells inhibits T-cell activation by eliminating CD137 ligand expression. Cancer Res. 73, 652–661 (2013).
Article CAS PubMed Google Scholar
Dhont, L., Mascaux, C. & Belayew, A. The helicase-like transcription factor (HLTF) in cancer: loss of function or oncomorphic conversion of a tumor suppressor?. Cell Mol. Life Sci. 73, 129–147 (2016).
Article CAS PubMed Google Scholar
Ding, H. et al. Characterization of a helicase-like transcription factor involved in the expression of the human plasminogen activator inhibitor-1 gene. DNA Cell Biol. 15, 429–442 (1996).
Article CAS PubMed Google Scholar
Hayward-Lester, A. et al. Cloning, characterization, and steroid-dependent posttranscriptional processing of RUSH-1 alpha and beta, two uteroglobin promoter-binding proteins. Mol. Endocrinol. 10, 1335–1349 (1996).
CAS PubMed Google Scholar
Castro, M. et al. Multiplexed methylation profiles of tumor suppressor genes and clinical outcome in lung cancer. J. Transl. Med. 8, 86 (2010).
Article PubMed PubMed Central Google Scholar
Kim, J. J. et al. Promoter methylation of helicase-like transcription factor is associated with the early stages of gastric cancer with family history. Ann. Oncol. 17, 657–662 (2006).
Article CAS PubMed Google Scholar
Capouillez, A. et al. The helicase-like transcription factor is a strong predictor of recurrence in hypopharyngeal but not in laryngeal squamous cell carcinomas. Histopathology 55, 77–90 (2009).
Article PubMed Google Scholar
Zhang, X. et al. Loss of heterozygosity and methylation of multiple tumor suppressor genes on chromosome 3 in hepatocellular carcinoma. J. Gastroenterol. 48, 132–143 (2013).
Article ADS CAS PubMed Google Scholar
Xu, Y. et al. HLTF promotes hepatocellular carcinoma progression by enhancing SRSF1 stability and activating ERK/MAPK pathway. Oncogenesis 12, 2 (2023).
Article CAS PubMed PubMed Central Google Scholar
Qing, P., Han, L., Bin, L., Yan, L. & Ping, W. X. USP7 regulates the stability and function of HLTF through deubiquitination. J. Cell Biochem. 112, 3856–3862 (2011).
Article CAS PubMed Google Scholar
González, A., Hall, M. N., Lin, S. C. & Hardie, D. G. AMPK and TOR: The Yin and Yang of cellular nutrient sensing and growth control. Cell Metab. 31, 472–492 (2020).
Article PubMed Google Scholar
Herzig, S. & Shaw, R. J. AMPK: guardian of metabolism and mitochondrial homeostasis. Nat. Rev. Mol. Cell Biol. 19, 121–135 (2018).
Article CAS PubMed Google Scholar
Ma, R., Yi, B., Riker, A. I. & Xi, Y. Metformin and cancer immunity. Acta Pharm. Sin. 41, 1403–1409 (2020).
Article CAS Google Scholar
Cha, J. H. et al. Metformin promotes antitumor immunity via endoplasmic-reticulum-associated degradation of PD-L1. Mol. Cell 71, 606–620.e607 (2018).
Article CAS PubMed PubMed Central Google Scholar
Dai, X. et al. Energy status dictates PD-L1 protein abundance and anti-tumor immunity to enable checkpoint blockade. Mol. Cell 81, 2317–2331 e2316 (2021).
Article CAS PubMed PubMed Central Google Scholar
Finisguerra, V. et al. Metformin improves cancer immunotherapy by directly rescuing tumor-infiltrating CD8 T lymphocytes from hypoxia-induced immunosuppression. J. Immunother. Cancer 11, e005719 (2023).
Huang, X. et al. Metformin reprograms tryptophan metabolism to stimulate CD8+ T-cell function in colorectal cancer. Cancer Res. 83, 2358–2371 (2023).
Article CAS PubMed Google Scholar
Afzal, M. Z., Mercado, R. R. & Shirai, K. Efficacy of metformin in combination with immune checkpoint inhibitors (anti-PD-1/anti-CTLA-4) in metastatic malignant melanoma. J. Immunother. Cancer 6, 64 (2018).
Article PubMed PubMed Central Google Scholar
Vazquez-Martin, A., Oliveras-Ferraros, C., Del Barco, S., Martin-Castillo, B. & Menendez, J. A. The anti-diabetic drug metformin suppresses self-renewal and proliferation of trastuzumab-resistant tumor-initiating breast cancer stem cells. Breast Cancer Res. Treat. 126, 355–364 (2011).
Article CAS PubMed Google Scholar
Akce, M. et al. Phase II trial of nivolumab and metformin in patients with treatment-refractory microsatellite stable metastatic colorectal cancer. J. Immunother. Cancer 11, e007235 (2023).
Foretz, M., Guigas, B. & Viollet, B. Metformin: update on mechanisms of action and repurposing potential. Nat. Rev. Endocrinol. 19, 460–476 (2023).
Article CAS PubMed Google Scholar
Shirwany, N. A. & Zou, M. H. AMPK in cardiovascular health and disease. Acta Pharm. Sin. 31, 1075–1084 (2010).
Article CAS Google Scholar
Zheng, B. et al. Oncogenic B-RAF negatively regulates the tumor suppressor LKB1 to promote melanoma cell proliferation. Mol. Cell 33, 237–247 (2009).
Article CAS PubMed PubMed Central Google Scholar
Jeon, S. M., Chandel, N. S. & Hay, N. AMPK regulates NADPH homeostasis to promote tumour cell survival during energy stress. Nature 485, 661–665 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Cai, Z. et al. Phosphorylation of PDHA by AMPK drives TCA cycle to promote cancer metastasis. Mol. Cell 80, 263–278.e267 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zhang, J. et al. S100A4 blockage alleviates agonistic anti-CD137 antibody-induced liver pathology without disruption of antitumor immunity. Oncoimmunology 7, e1296996 (2018).
Article PubMed PubMed Central Google Scholar
Li, Y. et al. Limited cross-linking of 4-1BB by 4-1BB ligand and the agonist monoclonal antibody utomilumab. Cell Rep. 25, 909–920.e904 (2018).
Article PubMed Google Scholar
Hong, D. S. et al. Utomilumab in patients with immune checkpoint inhibitor-refractory melanoma and non-small-cell lung cancer. Front. Immunol. 13, 897991 (2022).
Article CAS PubMed PubMed Central Google Scholar
Claus, C. et al. Tumor-targeted 4-1BB agonists for combination with T cell bispecific antibodies as off-the-shelf therapy. Sci. Transl. Med. 11, eaav5989 (2019).
Peper-Gabriel, J. K. et al. The PD-L1/4-1BB bispecific antibody-anticalin fusion protein PRS-344/S095012 elicits strong T-cell stimulation in a tumor-localized manner. Clin. Cancer Res. 28, 3387–3399 (2022).
Article CAS PubMed PubMed Central Google Scholar
Jeong, S. et al. Novel anti-4-1BBxPD-L1 bispecific antibody augments anti-tumor immunity through tumor-directed T-cell activation and checkpoint blockade. J. Immunother. Cancer 9, e002428 (2021).
Eskiocak, U. et al. Differentiated agonistic antibody targeting CD137 eradicates large tumors without hepatotoxicity. JCI Insight 5, e133647 (2020).
Qi, X. et al. Optimization of 4-1BB antibody for cancer immunotherapy by balancing agonistic strength with FcgammaR affinity. Nat. Commun. 10, 2141 (2019).
Article ADS PubMed PubMed Central Google Scholar
Kamata-Sakurai, M. et al. Antibody to CD137 activated by extracellular adenosine triphosphate is tumor selective and broadly effective in vivo without systemic immune activation. Cancer Discov. 11, 158–175 (2021).
Article CAS PubMed Google Scholar
Warmuth, S. et al. Engineering of a trispecific tumor-targeted immunotherapy incorporating 4-1BB co-stimulation and PD-L1 blockade. Oncoimmunology 10, 2004661 (2021).
Article PubMed PubMed Central Google Scholar
Muik, A. et al. Preclinical characterization and phase I trial results of a bispecific antibody targeting PD-L1 and 4-1BB (GEN1046) in patients with advanced refractory solid tumors. Cancer Discov. 12, 1248–1265 (2022).
Article CAS PubMed PubMed Central Google Scholar
Kim, D., Paggi, J. M., Park, C., Bennett, C. & Salzberg, S. L. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat. Biotechnol. 37, 907–915 (2019).
Article CAS PubMed PubMed Central Google Scholar
Kovaka, S. et al. Transcriptome assembly from long-read RNA-seq alignments with StringTie2. Genome Biol. 20, 278 (2019).
Article CAS PubMed PubMed Central Google Scholar
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Article PubMed PubMed Central Google Scholar
Tian, Y. et al. ChAMP: updated methylation analysis pipeline for Illumina BeadChips. Bioinformatics 33, 3982–3984 (2017).
Article CAS PubMed PubMed Central Google Scholar
Teschendorff, A. E. et al. A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics 29, 189–196 (2013).
Article CAS PubMed Google Scholar
Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e3529 (2021).
Article CAS PubMed PubMed Central Google Scholar
He, Y. et al. Multi-omics characterization and therapeutic liability of ferroptosis in melanoma. Signal Transduct. Target Ther. 7, 268 (2022).
Article PubMed PubMed Central Google Scholar
Farshidfar, F. et al. Integrative molecular and clinical profiling of acral melanoma links focal amplification of 22q11.21 to metastasis. Nat. Commun. 13, 898 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Miao, Y. R. et al. ImmuCellAI: a unique method for comprehensive T-cell subsets abundance prediction and its application in cancer immunotherapy. Adv. Sci. 7, 1902880 (2020).
Article CAS Google Scholar

Download references

Acknowledgements

We thank Jieying Zhang and Jun Mi for providing the shRNAs. Chi-Chung Hui for discussions. This study was supported by grants from the National Key Research and Development Program of China (2018YFA0107800 to J.L.), the National Natural Science Foundation of China (Nos.82100137 to L.L., 82102891 to X.K.), the Natural Science Foundation of Hunan Province for Outstanding Young Scholars (no. 2024JJ4075 to L.L. no. 2022JJ20097 to X.K.), the Natural Science Foundation of Hunan Province (No. 2024JJ6684 to L.Z.), the Scientific Research Program of FuRong Laboratory (No. 2024PT5106 to X.K.), the Central South University Innovation-Driven Research Program (No. 2023CXQD025 to X.K.), the Youth Science Foundation of Xiangya Hospital (No. 2023Q20 to L.Z.), the Natural Science Foundation of Changsha (No.kq2403025 to L.Z.) and the fellowship of China Postdoctoral Science Foundation (No. 2023M743967 to L.Z.), the Central South University Scientific Research Innovation Projects (2024ZZTS0526 to X.L.).

Author information

These authors contributed equally: Long Liang, Lin Zhu, Xin Li.

Authors and Affiliations

Department of Dermatology, Xiangya Hospital & School of Life Sciences, Central South University, Changsha, China
Long Liang, Lin Zhu, Wenbin Zhou, Guanxiong Zhang, Xinwei Kuang, Juan Su, Xiang Chen & Hong Liu
Medical Genetics & School of Life Sciences, Central South University, Changsha, China
Long Liang, Xin Li, Yanbin Zhang & Jing Liu
Hunan Key Laboratory of Skin Cancer and Psoriasis, Hunan Engineering Research Center of Skin Health and Disease, Xiangya Clinical Research Center for Cancer Immunotherapy, Furong Laboratory, Central South University, Changsha, China
Lin Zhu, Wenbin Zhou, Guanxiong Zhang, Xinwei Kuang, Juan Su, Xiang Chen & Hong Liu
Graduate Institute of Biomedical Sciences and Center for Molecular Medicine, China Medical University, Taichung, China
Mien-Chie Hung
Department of Biotechnology, Asia University, Taichung, China
Mien-Chie Hung
Department of Musculoskeletal Oncology, Fudan University Shanghai Cancer Center, Shanghai, China
Yong Chen

Authors

Long Liang
View author publications
Search author on:PubMed Google Scholar
Lin Zhu
View author publications
Search author on:PubMed Google Scholar
Xin Li
View author publications
Search author on:PubMed Google Scholar
Wenbin Zhou
View author publications
Search author on:PubMed Google Scholar
Yanbin Zhang
View author publications
Search author on:PubMed Google Scholar
Mien-Chie Hung
View author publications
Search author on:PubMed Google Scholar
Guanxiong Zhang
View author publications
Search author on:PubMed Google Scholar
Yong Chen
View author publications
Search author on:PubMed Google Scholar
Xinwei Kuang
View author publications
Search author on:PubMed Google Scholar
Juan Su
View author publications
Search author on:PubMed Google Scholar
Jing Liu
View author publications
Search author on:PubMed Google Scholar
Xiang Chen
View author publications
Search author on:PubMed Google Scholar
Hong Liu
View author publications
Search author on:PubMed Google Scholar

Contributions

L.L., Y.C., X.K., J.S., J.L., X.C., and H.L. designed the study and drafted the manuscript. L.L., X.L., L.Z., W.Z., and Y.Z. carried out the experiments. G.Z. performed the statistical and bioinformatics analysis. L.Z., X.K., and J.S. collected the patients’ specimens and information. L.L., J.L., X.C. M.-C.H. revised the manuscript. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Guanxiong Zhang, Yong Chen, Xinwei Kuang, Juan Su, Jing Liu, Xiang Chen or Hong Liu.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Transparent Peer Review file

Reporting Summary

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Liang, L., Zhu, L., Li, X. et al. CD137L promotes immune surveillance in melanoma via HLTF regulation. Nat Commun 16, 8478 (2025). https://doi.org/10.1038/s41467-025-63338-w

Download citation

Received: 30 August 2024
Accepted: 11 August 2025
Published: 26 September 2025
DOI: https://doi.org/10.1038/s41467-025-63338-w