Article Text
PDF
Abstract
Background Histone deacetylase (HDAC), a kind of protease that regulates gene expression by modifying protein acetylation levels, is usually aberrantly activated in tumors. The approved pan-HDAC inhibitors (HDACi) have exhibited clinical benefits for hematopoietic malignancies. Recently, HDACis have emerged as enhancers of antitumor immunity. However, the effect of HDACs on the tumor immune microenvironment of lung adenocarcinoma (LUAD) and the underlying mechanism is largely unknown.
Methods C57BL/6J and BALB/c nude mice with subcutaneous tumors were used for in vivo therapeutic effects and mechanistic investigations. Flow cytometry was used to measure the toxicity and exhaustion of human CD8+T cells after co-culturing with tumor cells and to determine the immunophenotype of tumor-infiltrating CD8+T cells. A series of experimental techniques, including RNA sequencing, quantitative PCR, western blot, ELISA, mass spectrometry, co-immunoprecipitation, chromatin immunoprecipitation and immunohistochemistry, were used to explore the underlying molecular mechanism.
Results The pan-HDACi vorinostat (SAHA) promoted CD8+T cell infiltration and effector function in LUAD through suppressing FGL1, a newly identified major ligand of LAG-3. Mechanistically, SAHA inhibited the activity of HDAC1, an essential deacetylase of JAK1. This increased the acetylation level of JAK1 at lysine 1109, thus promoting its proteasomal degradation and subsequently reducing STAT3-driven FGL1 transcription. The combination regimen of SAHA and anti-LAG-3 therapy was further explored in an immunocompetent LUAD mouse model. Compared with those receiving control or single agent treatments, mice receiving combination therapy exhibited a lower tumor burden and superior CD8+T-cell-killing activity.
Conclusions Our results revealed a novel mechanism by which the HDACi SAHA potentiates CD8+T-cell-mediated antitumor activity through the HDAC1/JAK1/FGL1 axis, providing a rationale for the combined use of HDACis and immunotherapy.
- Immunotherapy
- Lung Neoplasms
- Tumor Microenvironment
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Overexpression or hyperactivation of histone deacetylases (HDACs) is linked to the initiation of cancer and the emergence of resistance to cancer treatments. Recently, HDAC inhibitors have emerged as promising cancer therapeutics, exhibiting efficacy in decelerating cancer progression, triggering cell cycle arrest and apoptosis, and reshaping the tumor microenvironment.
WHAT THIS STUDY ADDS
The first Food and Drug Administration-approved pan-HDAC inhibitor, SAHA, has been shown to enhance the cytotoxicity of CD8+T cells and inhibit their exhaustion both in vitro and in vivo. Mechanistically, SAHA increased the acetylation of JAK1 at K1109 by inhibiting HDAC1, thereby promoting the degradation of JAK1 through a proteasome-dependent pathway and inhibiting the phosphorylation of STAT3 and the transcriptional regulation of FGL1, the ligand of LAG-3.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study revealed the feasibility and underlying molecular mechanisms of a combined immunotherapy strategy for lung adenocarcinoma (LUAD) treatment using HDAC inhibitors in conjunction with the immune checkpoint blockade agent anti-LAG-3, providing a promising new avenue for enhancing the response to LUAD immunotherapy.
Introduction
Lung cancer, particularly non-small cell lung cancer (NSCLC), remains the leading cause of cancer mortality globally, with lung adenocarcinoma (LUAD) as the predominant histological type.1 The emergence of immune checkpoint blockade (ICB) has led to long-term survival benefits for patients with LUAD, especially those without driver gene alterations,2 but response rates remain suboptimal, with an average response rate of less than 30%.3 Favorable survival and durable responses to ICB rely on the infiltration of cytotoxic T cells (CTL) which exhibit low expression of coinhibitory receptors such as Programmed Death-1 (PD-1), Cytotoxic T-Lymphocyte-Associated Antigen-4 (CTLA-4), and Lymphocyte Activation Gene-3 (LAG-3) and high expression of effector molecules such as perforin and granzyme B (GZMB) in the tumor immune microenvironment (TIME).4 5 Therefore, extensive research has focused on exploring combination therapies aiming at activating or recruiting CTLs into the TIME to overcome resistance to ICB.
The histone deacetylase (HDAC) superfamily includes 18 members, which can be divided into four classes. Class I (HDAC1, −2 to –3, and −8), Class IIa (HDAC4, −5 to –7, and −9), Class IIb (HDAC6 and −10) and Class IV (HDAC11) HDACs are Zn2+-dependent enzymes, while Class III (SIRT1-SIRT7) depends on NAD+ and is referred to as the sirtuin family. HDACs play a fundamental role in epigenetically regulating gene expression and chromatin remodeling by removing acetyl groups from histones or non-histone proteins.6 Abnormal HDAC expression and dysregulated chromatin acetylation have been implicated in the pathogenesis of cancer as well as resistance to cancer therapy.7 Consequently, HDAC inhibitors (HDACi) have evolved into attractive antitumor epigenetic therapies.8 To date, five pan-HDACis, such as vorinostat (SAHA), have been approved by the Food and Drug Administration (FDA) as anticancer drugs.9 Intriguingly, in addition to stimulating apoptosis and inducing cell cycle arrest in tumor cells, an increasing number of studies have shown that HDACis exert influence on diverse facets of the TIME, such as promoting the infiltration and function of CTLs,10 11 inhibiting the immunosuppression ability of regulatory T cells12 13 and myeloid-derived suppressor cells.14 15 Specifically, HDACis mediate the recruitment of immune cells by modulating the expression of chemokines such as CXCL9, 10, and 13.16 17 Furthermore, HDACis refine the immune landscape by regulating the expression of major histocompatibility complex (MHC) class I and II molecules,18 co-stimulatory molecule CD86, and other immunogenic tumor-associated antigens.19 These alterations facilitate the transmission of stimulatory signals and, in turn, augment the infiltration and functionality of CTLs.20 21 Notably, HDACis augment ICB responses in preclinical cancer models through their upregulation of PD-L1 expression.22 23 However, combining HDACis with anti-PD-1/Programmed Death-Ligand 1 (PD-L1) therapy for NSCLC in real-world studies has shown acceptable safety but limited effectiveness.24 Thus, alternative immunotherapy combination regimens and the potential molecular mechanisms through which HDACis reshape the TIME in NSCLC, especially CD8+T-cell-mediated tumor immunity, need to be explored.
FGL1 was originally discovered to be secreted by hepatocytes and is related to proliferation and metabolism25 and has been found recently highly expressed in many human cancers.26 Notably, a recent study demonstrated that FGL1 is one of the main functional ligands of LAG-3, forming a novel immune checkpoint pathway to regulate T-cell-mediated tumor immunity beyond PD-L1/PD-1.27 High FGL1 expression contributes to tumor progression by inducing T-cell depletion, dysfunction, and immune evasion and predicts poor outcomes in patients receiving anti-PD-L1/PD-1 therapy.26–28 Hence, FGL1 has risen to prominence as a promising immune checkpoint target in cancer immunotherapy, especially in the targeted treatment of LUAD. Gaining deeper insights into the molecular regulation of FGL1 expression could lead to the development of therapeutic strategies for LUAD.
In this study, we showed that the HDACi SAHA enhanced CD8+T-cell-mediated antitumor immunity by decreasing FGL1 expression both in vitro and in vivo. Mechanistically, SAHA suppressed the transcription of FGL1 by targeting HDAC1. HDAC1 silencing inactivated the JAK1/STAT3 pathway by increasing the K1109 acetylation of JAK1 and promoting its ubiquitination-binding and degradation, thus transcriptionally reducing FGL1 expression. Disruption of the HDAC1/FGL1 axis inhibited immune escape in LUAD. Taken together, the results of this study reveal the regulatory mechanism of FGL1 and provide new insight into treatment strategies for improving the therapeutic efficacy of ICB in patients with LUAD.
Methods
Chromatin immunoprecipitation-quantitative PCR
The chromatin immunoprecipitation (ChIP) assay was conducted using an EZ-ChIP kit (Millipore, USA) according to the manufacturer’s recommended protocol. Briefly, cells were cross-linked with 1% formaldehyde for 10 min and then quenched using 0.125 M glycine. The cells were sonicated with an ultrasonic cell disruptor from Xianou (China). One per cent of the total chromatin was kept as input. Chromatin was precipitated using either an STAT3 primary antibody (#9139, Cell Signaling Technology, USA) or an equivalent amount of control IgG. After unlocking cross-linking and RNA and protein digestion, genomic DNA was purified with a PCR Purification Kit (Beyotime, China). DNA enrichment was detected by PCR. The sequences of primers used for quantitative PCR (qPCR) are listed in online supplemental table 1.
Supplemental material
Flow cytometry analysis
The cell samples were resuspended in ice-cold buffer (1× phosphate-buffered saline (PBS) and 2% fetal bovine serum) and then stained with the antibodies listed in online supplemental table 2. Fixable viability stain (#65-0865-14, eBiosciences, USA) was used to exclude dead cells. Additionally, FcR Blocking Reagent (Miltenyi Biotec, Germany) was used to ensure the specificity and accuracy of the staining process. All samples were acquired with a CytoFLEX flow cytometer (Beckman, USA) and analyzed with FlowJo software.
In vivo mouse experiments
Female C57BL/6J and BALB/c nude mice (6–8 weeks old) were obtained from Nanjing Medical University Animal Center and housed under standard pathogen-free conditions. All animal experiments were approved by the Committee on the Ethics of Animal Experiments at Nanjing Medical University (IACUC-2109036). For the SAHA monotherapy assay, LLC cells (2×106 cells per mouse) in 100 µL of PBS were subcutaneously injected into the right side of each C57BL/6J mouse or BALB/c nude mouse and H1299 cells (3×106 cells per mouse) in 100 µL of PBS were subcutaneously injected into the right side of each BALB/c nude mouse. Seven days after tumor cell injection, the mice were randomly divided into two groups (n=5 for each group) and treated intraperitoneally with SAHA (100 mg/kg per mouse) or vehicle once every other day. For the combination therapy assay, 2×106 LLC cells in 100 µL of PBS were subcutaneously injected into C57BL/6J mice. Seven days after tumor cell injection, the mice were randomly divided into four experimental groups (n=5 for each group)—the vehicle-treated group, SAHA-treated group, anti-LAG-3-treated group, and SAHA+anti-LAG-3-treated group—and treated with an intraperitoneal injection of SAHA (100 mg/kg per mouse, once every other day) or anti-LAG-3 (100 µg per mouse, every 3 days), either individually or in combination. In addition, 3×106 Hdac1 knockdown and/or Fgl1 overexpression and negative control LLC cells in 100 µL of PBS were subcutaneously injected into C57BL/6J mice. Tumor size was measured with a vernier caliper and calculated with the formula width2×length×0.5 every other day. The % tumor growth inhibition (TGI), calculated as (1−(T/C))×100%, quantifies the difference in tumor growth inhibition between treatment (T) and isotype control groups (C). The mice were sacrificed when the tumor volume reached 1,500 mm3 or the ulcer diameter of the tumor reached 1.5 cm. The tumors were resected, weighed and then digested into single cells for flow cytometry. Peripheral blood was collected to detect the abundance of FGL1 in mouse serum in the SAHA monotherapy assay. Kidneys and livers were collected and embedded in paraffin for subsequent histochemical analysis in the combination therapy assay. The antibodies used for in vivo experiments are detailed in online supplemental table 2.
Statistics
All the statistical results are presented as the mean±SD from three independent replicates. The normality of the data distribution was determined by the Shapiro-Wilk test. The statistical significance for normally distributed data was analyzed using Student’s t-test between two groups or analysis of variance followed by Tukey’s post hoc test among multiple groups. All the statistical analyses and graphs were generated using GraphPad Prism V.9.0 (USA). A p value<0.05 was considered statistically significant (ns: not significant, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001).
The detailed methodology is found in the online supplemental file.
Results
SAHA treatment activated CD8+ T-cell-mediated cancer immunity
Given the involvement of HDACs in modulating immune responses, we first evaluated the expression pattern of the HDAC family and the association of the HDAC family with the immune microenvironment in The Cancer Genome Atlas (TCGA)-LUAD database and found that the expression of most HDACs was greater in tumor tissues (online supplemental figure S1A) and negatively correlated with the immune score (online supplemental figure S1B). To further confirm the role of HDACis in regulating antitumor immunity in LUAD, we subcutaneously injected H1299 or LLC cells into immunodeficient BALB/c nude mice or immunocompetent C57BL/6J mice, respectively, and then intraperitoneally administered SAHA, the first FDA-approved pan-HDACi (figure 1A). SAHA treatment induced TGI of about 48.1% in C57BL/6J mice bearing LLC tumors (figure 1C). In contrast, no such statistically significant reduction in tumor growth was observed in immune-deficient nude mice inoculated with either H1299 cells (online supplemental figure S1C) or LLC cells (figure 1B), suggesting that the antitumor activity of SAHA depended largely on adaptive immune responses. To study the effect of SAHA on CD8+T cells, we analyzed the frequency and phenotype of CD8+T cells in the tumor tissues of vehicle-treated and SAHA-treated C57BL/6J mice. SAHA treatment augmented CD8+T cell infiltration, with an approximate mean fold change of 2 compared with the vehicle group (figure 1D) and elevated the percentage of cytotoxic GZMB+CD8+ T cells (figure 1E). Moreover, the proportion of exhausted CD8+T cells, as assessed by PD-1 and LAG-3 staining (figure 1F,G), was decreased following SAHA treatment. These findings suggested that SAHA treatment triggered antitumor immunity against LUAD in a CD8+T-cell-mediated manner.
SAHA treatment suppressed tumor growth in immunocompetent mice and enhanced CD8+T cell function both in vivo and in vitro. (A) Experimental timeline for in vivo xenograft studies in immunodeficient BALB/c nude mice and immunocompetent C57BL/6J mice. (B) The effect of SAHA on LLC cells growth in immunodeficient BALB/c nude mice (vehicle and SAHA, n=5 mice per group). (C) The effect of SAHA on LLC cells growth in immunocompetent C57BL/6J mice (vehicle and SAHA, n=5 mice per group). (D) Statistical analysis of the proportions of CD8+T cells among CD45+CD3+ T cells in the indicated LLC tumors from treated (C) C57BL/6J mice (n=5 mice per group). (E–G) Representative flow cytometry dot plots and statistical analysis of the percentages of (E) GZMB+, (F) PD-1+, and (G) LAG-3+CD8+ T cells in the indicated LLC tumors from treated (C) C57BL/6J mice (n=5 mice per group). (H) Schematic diagram of the co-culture system. (I–K) Statistics on the proportions of (I) GZMB+, (J) PD-1+, and (K) LAG-3+CD8+ T cells after co-culturing with SAHA-pretreated H1299 cells for 48 hours. (L) Representative images of T-cell-mediated cancer cell killing between the SAHA (3 µM)-treated and vehicle-treated groups of H1299 cells with or without activated CD8+T cell co-culture. Crystal violet staining was used for quantification. The OD values were standardized to the average of the vehicle group without CD8+T cell co-culture. GZMB, granzyme B; LAG-3, Lymphocyte Activation Gene-3; OD, Optical Density; PD-1, Programmed Death-1.
Since HDACs mainly play a functional role in suppressing the acetylation level of histones, we found that the histone acetylation levels in several LUAD cell lines were decreased compared with those in normal human bronchial epithelial cell lines (online supplemental figure S1E), indicating that HDACs are aberrantly activated in LUAD. Next, we selected the classic LUAD cell lines A549 and H1299, which have relatively low histone acetylation levels, for further investigation. Notably, HDAC inhibition with SAHA significantly inhibited the proliferation of H1299 and A549 cells, with IC50 values of 2.94±0.15 µM and 2.99±0.77 µM, respectively (online supplemental figure S1F). To explore the role of SAHA in regulating CD8+T cell activity in vitro, we co-cultured H1299 and A549 cells pretreated or not pretreated with the IC50 of SAHA with activated CD8+T cells (figure 1H). Interestingly, pretreatment with SAHA increased the cytotoxicity of CD8+T cells and inhibited the exhaustion of CD8+T cells (figure 1I–K, online supplemental figure S2A–C). Meanwhile, in the absence of CD8+T cells, SAHA inhibited cell proliferation by approximately 30%, but the presence of CD8+T cells enhanced this inhibitory effect (approximately 70% of the cells were killed) (figure 1L, online supplemental figure S2D). These results indicated that SAHA primarily exerted its cytotoxic effect through CD8+T cells and could enhance CD8+T-cell-mediated tumor cell killing in LUAD.
The involvement of FGL1 in SAHA-induced CD8+ T-cell activation
We then aimed to elucidate how SAHA enhanced CD8+T cell-mediated immune activation. Using RNA sequencing (RNA-seq), we observed that SAHA treatment broadly modulated the gene expression profiles of H1299 cells (online supplemental figure S3A), impacting pathways associated with immune modulation29 30 (online supplemental figure S3B). Given that HDACis extensively regulate gene expression and are involved in the immune editing of tumor neoantigens,20 21 our focus then narrowed down to SAHA-induced alterations in immune checkpoint genes. The heatmap revealed regulation of several immune checkpoint-related genes by SAHA (figure 2A), including MHC class I molecules (HLA-A, HLA-B, HLA-F) and MHC class II molecules (HLA-DPA1, HLA-DPB1), as also verified by Real-time quantitative PCR (RT-qPCR) (figure 2B).
SAHA treatment suppressed FGL1 expression and FGL1-dependent CD8+T cell immunosuppression. (A) Heatmap of the differential expression of immune checkpoint genes between the DMSO and SAHA groups. (B) Relative mRNA levels of MHC-I and MHC-II genes induced by SAHA treatment detected by RT-qPCR in H1299 and A549 cells. (C) Correlations between FGL1 expression and CD8+T cell (left) or CD8+central memory T-cell (right) frequencies in LUAD samples analyzed by TIMER V.2.0. (D) Correlations between the expression of FGL1 and CD8A (left) or CD8B (right) in LUAD cells analyzed by TIMER V.2.0. (E–G) Relative mRNA and protein levels of FGL1 in H1299 and A549 cells after treatment with the indicated concentrations of SAHA for 48 hours, as detected by (E) RT-qPCR, (F) western blot and (G) ELISA. (H) Serum Fgl1 protein levels in (figure 1C) C57BL/6 mice treated with SAHA or vehicle, as detected by ELISA (n=5 mice per group). (I) The transfection efficiency of FGL1 overexpression plasmids was detected by western blot in H1299 and A549 cells. (J–L) H1299 cells were transfected with FGL1 overexpression plasmids or negative controls and then pretreated with SAHA (3 µM) or vehicle for 8 hours. Statistics on the proportions of (J) GZMB+, (K) PD-1+ and (L) LAG-3+CD8+ T cells after co-culturing with different groups of H1299 cells. (M) Representative images of the T-cell-mediated cancer cell killing assay. The indicated H1299 cells were co-cultured with activated CD8+T cells for 48 hours. Crystal violet staining was used for quantification. The OD values were standardized to the average of the vector group treated with DMSO. DMSO, Dimethyl sulfoxide; GZMB, granzyme B; LAG-3, Lymphocyte Activation Gene-3; LUAD, lung adenocarcinoma; MHC, major histocompatibility complex; mRNA, messenger RNA; OD, Optical Density; PD-1, Programmed Death-1; RT-qPCR, Real-time quantitative PCR.
To pinpoint the crucial immune checkpoint gene regulated by SAHA that may significantly impact LUAD progression, we conducted an intersection analysis of differential expressed genes between cancerous and non-cancerous tissues from the TCGA-LUAD and the Gene Expression Omnibus dataset GSE30219 with those 16 immune checkpoint genes regulated by SAHA, yielding FGL1 as a potential key regulator (online supplemental figure S3C). FGL1 is upregulated in LUAD (online supplemental figure S3D, E), while suppressed by SAHA treatment. Moreover, high FGL1 expression was negatively associated with CD8+T cell infiltration (figure 2C,D). Additionally, we found that genes coexpressed with FGL1 were inversely enriched in the “response to interferon-gamma” and “T-cell activation” pathways (online supplemental figure S3F–H). Consequently, we prioritized FGL1 as a crucial target in mediating the immunomodulatory effects of SAHA on CD8+T cells in LUAD. Moreover, qPCR (figure 2E) and western blot (figure 2F) assays further confirmed that SAHA treatment suppressed the mRNA and protein levels of FGL1 in a dose-dependent manner. Considering that FGL1 is a secreted protein, we performed an ELISA assay to detect FGL1 protein secretion and found that SAHA strikingly inhibited FGL1 in the supernatant of H1299 and A549 cells (figure 2G) as well as in the serum of the above C57BL/6J mice (figure 2H). More importantly, we transfected FGL1-overexpressing plasmids into H1299 and A549 cells and examined the efficiency of the plasmids through western blot (figure 2I). Ectopic expression of FGL1 impaired SAHA-induced CD8+T cell activation in the co-culture system (figure 2J–L, online supplemental figure S4A–C), while SAHA therapy restrained the immunosuppressive effect of FGL1 overexpression on the CD8+T-cell-mediated tumor cell-killing (figure 2M, online supplemental figure S4D). Overall, we concluded that FGL1 is a key modulator of SAHA-activated CD8+T-cell-mediated antitumor immunity in LUAD.
HDAC1 was responsible for SAHA-mediated FGL1 expression and regulated FGL1-induced CD8+ T-cell malfunction
We next attempted to clarify the underlying molecular mechanism by which SAHA induced the transcriptional inhibition of FGL1. Regarding SAHA is a broad-spectrum HDACi that has specificity for most of the 11 metal-dependent HDAC isoforms,31 we first detected the effect of three selective HDACis, MGCD0103, which targets HDAC isotypes 1, 2, 3, and 1132; the Class IIa HDACi TMP19533; and the selective HDAC6 inhibitor ACY-1215, on the regulation of FGL1 expression using qPCR and western blot (figure 3A,B). Surprisingly, the changes in the mRNA and protein levels of FGL1 following MGCD0103 and SAHA treatment were consistent. To further identify which specific HDAC family member contributes to the inhibition of FGL1. We individually knocked down HDAC1, HDAC2, HDAC3, and HDAC11 in H1299 and A549 cells with siRNAs. Notably, only HDAC1 silencing downregulated FGL1 at the mRNA and protein levels (figure 3C–G, online supplemental figure S5A–D). In line with that, knocking down HDAC1 with two other siRNAs suppressed FGL1 protein levels (figure 3H). In contrast, overexpression of HDAC1 led to an increase in FGL1 protein expression (figure 3I). Moreover, the inhibitory effect of SAHA on FGL1 expression was reversed by ectopic expression of HDAC1 (figure 3J), again indicating that HDAC1 may be responsible for the regulation of FGL1 expression mediated by SAHA.
HDAC1 was responsible for SAHA-mediated FGL1 expression in lung adenocarcinoma. (A–B) Relative (A) mRNA and (B) protein levels of FGL1 in H1299 and A549 cells treated with different HDAC inhibitors. (C–F) Relative mRNA levels of the FGL1 and HDAC genes in H1299 cell lines with or without knockdown of HDAC1, HDAC2, HDAC3 or HDAC11. (G) Relative protein levels of FGL1, HDAC1, HDAC2, HDAC3, and HDAC11 in H1299 and A549 cells with knockdown of HDAC1, HDAC2, HDAC3 or HDAC11. (H) Relative protein levels of FGL1 and HDAC1 in H1299 and A549 cells transfected with two different HDAC1 siRNAs. (I) Relative protein levels of FGL1 and HDAC1 in H1299 and A549 cells with or without HDAC1 overexpression. (J) Representative western blot showing FGL1 and HDAC1 protein expression levels in HDAC1 overexpression plasmid-transfected or negative control-transfected cells with or without SAHA (3 µM) treatment for 48 hours. HDAC, histone deacetylase; mRNA, messenger RNA.
In addition, a positive correlation between HDAC1 and FGL1 in LUAD was observed at the mRNA level (online supplemental figure S5E, F). We then examined the association between HDAC1 and FGL1 proteins in a human tissue array by multiplex immunohistochemical assay. Compared with adjacent non-tumor tissues, upregulation of HDAC1 and FGL1 protein expression was observed in LUAD tissues (figure 4A–C). Furthermore, a positive correlation between HDAC1 and FGL1 protein expression was detected in the tissue array, revealing the regulatory relationship between these two proteins in LUAD (figure 4D).
Downregulation of HDAC1 inhibited tumor growth and enhanced the cytotoxicity of CD8+T cells by regulating FGL1. (A) Representative mIHC images of HDAC1 (green), FGL1 (pink) and DAPI (blue) in tumor and adjacent normal tissues from patients with lung adenocarcinoma (adjacent non-tumor tissues, n=48; tumor, n=48). Scale bar=50 µm. (B–C) Statistical analysis of average optical density (AOD) of (B) HDAC1 or (C) FGL1 between tumor tissues and adjacent normal tissues in A. The p values were calculated using the Wilcoxon matched-pairs signed-rank test. (D) Correlation analysis of HDAC1 and FGL1 for A. The p value was calculated using the Spearman correlation analysis. (E–G) Stable expressing non-targeting control (NTC) or Hdac1 knockdown LLC cell lines were further overexpressing lentiviral vector (LV) or Fgl1. (E) Tumor images, (F) tumor volume, and (G) tumor weight in C57BL/6J mice injected subcutaneously with the above cell lines from different groups (NTC+LV, shHdac1+LV, NTC+Fgl1, shHdac1+Fgl1, n=5 mice per group). (H) The effect of Hdac1 knockdown and/or Fgl1 overexpression on CD8+T cells infiltration in the LLC tumors from C57BL/6J mice (n=5 mice per group). (I–K) Representative flow cytometry dot plots and statistical analysis of (I) GZMB+, (J) PD-1+, and (K) LAG-3+CD8+ T cells in the LLC tumor tissues with Hdac1 knockdown and/or Fgl1 overexpression (n=5 mice per group). GZMB, granzyme B; HDAC, histone deacetylase; LAG-3, Lymphocyte Activation Gene-3; mIHC, multiplex immunohistochemical; PD-1, Programmed Death-1.
To further clarify the effect of the HDAC1-FGL1 axis on tumor growth and the TIME in vivo, we overexpressed Fgl1 in LLC cells in which Hdac1 was stably knocked down (online supplemental figure S5G) and implanted the above cell lines or control cells into C57BL/6J mice for xenograft experiments (figure 4E). Overexpression of Fgl1 promoted tumor growth, resulting in an approximately threefold increase in tumor mass. In contrast, the knockdown of Hdac1 nearly abrogated tumor growth compared with the empty vector control group. Furthermore, within the Hdac1 knockdown groups, an increase in Fgl1 expression attenuated the antitumor effect of Hdac1 inhibition, suggesting that Fgl1 plays a crucial role in Hdac1-mediated tumor growth (figure 4F,G). Using flow cytometry, we found that the knockdown of Hdac1 in LLC cells resulted in an increase in intratumoral CD8+T cell infiltration (figure 4H) and enhanced the percentage of the GZMB+CD8+ T-cell subgroup. Decreased PD-1+ and LAG-3+CD8+ T-cell populations were also observed in the Hdac1-knockdown group compared with the control group. In contrast, Fgl1 overexpression exhibited an opposing effect on CD8+T cell infiltration and antitumor activity. Moreover, ectopic expression of Fgl1 in Hdac1 knockdown tumors weakened the Hdac1 downregulation-mediated enhancement of CD8+T cell infiltration, reducing it to levels comparable to those observed in control groups, and also restored the inhibitory effect of Hdac1 silencing on CD8+T cell exhaustion (figure 4I–K). Together, these data demonstrated that inhibition of HDAC1 potentiated CD8+T cell infiltration and activity by regulating FGL1 expression, thereby activating the TIME of LUAD.
HDAC1 regulated FGL1 expression by suppressing JAK1 protein level
Histone acetylation is commonly linked to chromatin de-condensation and facilitating gene activation.6 Surprisingly, our previous results have uncovered a positive correlation between the activity of HDAC1, the enzyme typically known to suppress histone acetylation and gene transcription, and the transcription level of FGL1, suggesting the existence of other complex interaction mechanisms between these two factors. Furthermore, we found that the knockdown of EP300, a key histone acetyltransferase, led to a significant downregulation of histone H3 and H4 acetylation levels, without eliciting notable changes in mRNA or protein expression levels of FGL1 (online supplemental figure S6A, B). Based on these findings, we postulated that HDAC1 inhibition might repress FGL1 transcription through modulation of the activity of its upstream transcription factor, rather than by directly modulating histone acetylation levels at the FGL1 locus. We then screened six databases for the transcription factors of FGL1 and identified STAT3 as the key mediator (figure 5A). Additionally, our previous RNA-seq data showed that the target genes of STAT3 were significantly impacted by SAHA (enrichment False Discovery Rate <0.05) (online supplemental figure S6C). Hence, we assumed that STAT3 could be key to HDAC1-mediated FGL1 expression. As expected, the mRNA expression of FGL1 was observed to undergo a substantial decline on STAT3 silencing, with a per cent decrease of about 50% (figure 5B). Furthermore, three predicted STAT3 binding sites on the promoter of FGL1 were pinpointed using the JASPAR database (figure 5C). ChIP-qPCR further confirmed that STAT3 directly bound to these three sites in the FGL1 promoter (figure 5D). However, interleukin-6, which induces the transcriptional activity of STAT3, dramatically upregulated FGL1 expression in a dose-dependent manner (figure 5E) and counteracted the reduction in FGL1 transcription caused by HDAC1 knockdown (figure 5F). Therefore, these findings demonstrated that HDAC1 inhibition decreased FGL1 transcription through the STAT3 pathway.
HDAC1 regulated the mRNA transcription of FGL1 via the JAK1/STAT3 pathway. (A) Venn diagram showing the common predicted transcription factors of FGL1 among the six databases. (B) Relative mRNA levels of FGL1 in H1299 and A549 cell lines with or without STAT3 knockdown. (C) The consensus DNA-binding motif of STAT3 (upper), the three predicted binding sites (middle) and the binding sequences (bottom) of STAT3 at the FGL1 promotor from the JASPAR database. (D) The three binding sites of STAT3 on the FGL1 promoter in the H1299 and A549 cell lines were verified via chromatin immunoprecipitation-quantitative PCR analysis. (E) Relative mRNA levels of FGL1 in H1299 and A549 cells treated with the indicated concentrations of IL6. (F) A549 and H1299 cells were transfected with HDAC1 siRNAs or the corresponding negative controls for 48 hours and then treated with 100 µM IL6 for 24 hours. The relative mRNA levels of FGL1 were analyzed. (G) MS fragment ion spectra of JAK1 pulled down from H1299 cell lysates transfected with Flag-HDAC1 by a Flag antibody. (H) The endogenous interaction between HDAC1 and JAK1 was verified by co-IP in H1299 cells. (I) Representative western blot showing JAK1, STAT3, p-STAT3, and HDAC1 protein levels in H1299 cells with or without HDAC1 knockdown. (J) Representative western blot showing JAK1, STAT3, p-STAT3, and HDAC1 protein levels in H1299 cells with or without HDAC1 overexpression. co-IP, co-Immunoprecipitation; HDAC, histone deacetylase; IL, interleukin; mRNA, messenger RNA; MS, Mass spectrometric.
We next investigated the role of HDAC1 in regulating the STAT3/FGL1 axis. Surprisingly, we found that HDAC1 did not directly interact with STAT3 (data not shown) or affect STAT3 protein levels. Thus, we conducted quantitative proteomics to identify specific HDAC1-binding proteins that regulate STAT3 signaling. Among them, JAK1, an upstream kinase that phosphorylates STAT3, emerged as a potential intermediary in HDAC1/STAT3/FGL1 signaling (figure 5G). Co-Immunoprecipitation (co-IP) assays verified that HDAC1 interacted with JAK1 endogenously in H1299 (figure 5H) and A549 (online supplemental figure S6D) cells. AlphaFold 3 and PyMOL were also used to predict the direct interaction between HDAC1 and JAK1 and to identify potential amino acid pairs involved in their interplay (online supplemental figure S6E). To investigate the potential reciprocal regulation between HDAC1 and JAK1, we detected the expression of JAK1 at both the transcriptional and translational levels on HDAC1 silencing. The results indicated that HDAC1 knockdown almost did not affect JAK1 mRNA levels (online supplemental figure S6F), but significantly decreased JAK1 protein levels and subsequent STAT3 phosphorylation (figure 5I, online supplemental figure S6G), while HDAC1 overexpression had the opposite effect (figure 5J, online supplemental figure S6H). To sum up, our findings supported that HDAC1 directly interacted with JAK1, thus suppressing the protein level of JAK1 and subsequent STAT3/FGL1 axis.
HDAC1 deacetylated JAK1 at K1109 and promoted its ubiquitination-dependent degradation
We next investigated the molecular mechanisms by which HDAC1 influenced the expression of JAK1 protein. Consistent with the previous reports that HDAC1 deacetylates non-histone substrates,34 co-IP assays revealed that HDAC1 knockdown caused a marked increase in JAK1 lysine acetylation (figure 6A, online supplemental figure S7A), while the opposite effect was detected in HDAC1-overexpressing cells (figure 6B, online supplemental figure S7B), demonstrating that JAK1 is a deacetylation substrate of HDAC1. Furthermore, K269 and K1109 were identified as two putative acetylated lysine residues of JAK1 (figure 6C) according to the results from the “MusiteDeep” and “PhosphoSitePlus” databases. Then, we mutated each of these two putative acetylated lysine (K) sites to arginine (R) to generate acetylation-deficient mutants of JAK1. Our results showed that HDAC1 silencing failed to enhance the acetylation level of the JAK1 K1109R mutant but not the K269R mutant, indicating that HDAC1 specifically deacetylated K1109 of JAK1 (figure 6D).
HDAC1 deacetylated JAK1 at K1109 and promoted JAK1-ubiquitin binding. (A) JAK1 acetylation level in H1299 cells with or without HDAC1 knockdown detected by co-IP assay. (B) JAK1 acetylation level in H1299 cells with or without HDAC1 overexpression detected by co-IP assay. (C) Venn diagram depicting the common predicted acetylation sites on the lysine residue of JAK1 between the MusiteDeep and PhosphoSite databases. (D) JAK1 acetylation level in H1299 cells after cotransfection with HDAC1 siRNA and HA-tagged wild-type JAK1 or JAK1 mutants (K269R and K1109R) or negative control plasmids for 48 hours detected by co-IP assay. (E) Western blot analysis of JAK1 in H1299 cells after transfection with siNC or siHDAC1 for 48 hours and treatment with 20 µg/mL CHX for the indicated times. (G) Western blot analysis of JAK1 in A549 cells after transfection of vector or HDAC1 overexpression plasmid for 48 hours and treatment with 20 µg/mL CHX for the indicated times. (F, H) Statistical analysis of JAK1 protein expression in the (F) E and (H) G groups. (I) Representative western blot showing JAK1 protein levels in H1299 cells treated with the indicated concentrations of MG132 for 8 hours. (J) Representative western blot showing JAK1 protein levels in H1299 cells stimulated with 10 µM MG132 for the indicated times. (K) Western blot analysis of JAK1 in H1299 cells transfected with siNC or siHDAC1 for 48 hours and treated with 10 µM MG132 for 8 hours. (L) Western blot analysis of JAK1 ubiquitination levels in H1299 cells transfected with the indicated constructs for 48 hours and stimulated with MG132 for 8 hours. (M) Relative mRNA levels of FGL1 in H1299 and A549 cells cotransfected with WT JAK1 or the JAK1 K1109R mutant and siNC or siHDAC1 for 48 hours, as measured by RT-qPCR. CHX, cycloheximide; co-IP, co-Immunoprecipitation; HDAC, histone deacetylase; mRNA, messenger RNA; RT-qPCR, Real-time quantitative PCR.
The acetylation of proteins influences protein degradation or stabilization partly through interactions with ubiquitination, another lysine modification. For example, acetylation may stabilize proteins by competing with ubiquitination for binding sites.35 Acetylation may also modulate the interaction of E3 ligases or deubiquitinating enzymes with the target protein.36 Since the JAK1 mRNA level did not change on HDAC1 knockdown (online supplemental figure S6E), we speculated that the acetylation of JAK1 may affect its stabilization and ubiquitination. A cycloheximide assay showed that HDAC1 knockdown shortened the half-life of the JAK1 protein compared with that in the control groups (figure 6E,F), whereas HDAC1 overexpression notably extended the half-life of JAK1 (figure 6G,H), implying that HDAC1 silencing impaired its stability. Next, we asked whether HDAC1-mediated JAK1 degradation relies on the ubiquitin-proteasome pathway. Treatment with MG132 (a proteasome inhibitor) notably reduced JAK1 degradation in a dose-dependent (figure 6I, online supplemental figure S7C) and time-dependent (figure 6J, online supplemental figure S7D) manner and reversed HDAC1 silencing-induced JAK1 downregulation (figure 6K). Moreover, HDAC1 silencing increased the ubiquitination-binding of WT JAK1 but not of the JAK1 K1109R mutant (figure 6L), indicating that the K1109 acetylation of JAK1 mediated by HDAC1 knockdown promoted JAK1 degradation in a ubiquitin-proteasome-dependent manner. Moreover, overexpression of WT JAK1 restored the transcription of FGL1 inhibited by HDAC1 silencing, while FGL1 mRNA expression remained relatively unchanged with JAK1 K1109 mutant overexpression (figure 6M). Overall, we concluded that inhibiting HDAC1 specifically acetylated JAK1 at K1109, promoting ubiquitin-proteasome-dependent degradation of JAK1 and thus suppressing the STAT3-mediated FGL1 transcription.
The combination of SAHA and anti-LAG-3 therapy synergistically suppressed tumor growth and promoted CD8+ T-cell activation
LAG-3 has been identified as a promising target for immunotherapy in addition to PD-1 and CTLA-4.37 Several anti-LAG-3 monoclonal antibodies (mAbs) are currently undergoing clinical trials and exhibit encouraging activity against cancers. Combinational therapy has become the mainstream strategy for LAG-3-targeting therapeutics.38 A major milestone in anti-LAG-3 mAb combination therapy came in 2022 with FDA approval of anti-LAG-3 plus anti-PD-1 as a first-line treatment for unresectable or metastatic melanoma.39 Nevertheless, further work is required to maximize the potential of such therapeutic options.
We observed that SAHA modulated the expression of MHC-II and FGL1, two well-recognized ligands of LAG-3.37 Additionally, our prior results revealed that SAHA boosted CD8+T cell infiltration and cytotoxicity in tumors. Thus, we hypothesized that this immune activation induced by SAHA treatment could improve the antitumor immune response to ICB therapy. To further evaluate the therapeutic effect of SAHA combined with the anti-LAG-3 mAb, we inoculated LLC cells into C57BL/6J mice and then treated them with SAHA and/or the anti-LAG-3 mAb (figure 7A). In the C57BL/6J mouse model, we observed average TGI of 46.6% and 57.5% with either SAHA monotherapy or anti-LAG-3 monotherapy, respectively, compared with the control group. Notably, the combination therapy potentiated tumor suppression, yielding an average TGI of 85.1%, outperforming the monotherapy groups (figure 7B). Furthermore, the combined treatment not only increased CD8+T cell infiltration but also enhanced their cytotoxic activity and reduced CD8+T cell exhaustion levels, compared with the control group. However, compared with monotherapy groups, the combination therapy demonstrated a notable advantage only in enhancing CD8+T cell cytotoxic activity, without exhibiting significant superiority in promoting CD8+T cell infiltration or inhibiting their exhaustion levels (figure 7E–H). Therefore, this suggests that the increased antitumor efficacy of the combination therapy may partially rely on enhancing CD8+T cell-mediated tumor-killing capabilities. Moreover, we did not observe any weight loss (figure 7C) and liver or kidney dysfunction (figure 7D) in the mice, suggesting that the combination therapy was well tolerated. Together, these findings demonstrate that SAHA in combination with the anti-LAG-3 mAb exhibits a synergistic tumor suppression and induces potent antitumor immunity partially through activation of CD8+T cell cytotoxicity.
The combination of SAHA and anti-LAG-3 therapy effectively suppressed lung adenocarcinoma tumor growth and augmented CD8+T cell function. (A) Experimental schedule for SAHA alone or in combination with anti-LAG-3 therapy in the C57BL/6J mouse model. (B) The effect of different therapy regimes on LLC cells growth in immunocompetent C57BL/6J mice (vehicle, SAHA, anti-LAG-3, SAHA+anti-LAG-3, n=5 mice per group). (C) Body weight of different groups of C57BL/6J mice (vehicle, SAHA, anti-LAG-3, SAHA+anti-LAG-3, n=5 mice per group). (D) Representative liver and kidney H&E staining images for each group. (E) Statistical analysis of CD45+CD3+CD8+ T-cell infiltration in the indicated tumors from different treatment groups (vehicle, SAHA, anti-LAG-3, SAHA+anti-LAG-3, n=5 mice per group). (F–H) Representative flow cytometry dot plots and statistical analysis of the percentages of (F) GZMB+, (G) PD-1+, and (H) LAG-3+CD8+ T cells in the indicated xenograft tumor tissues (vehicle, SAHA, anti-LAG-3, SAHA+anti-LAG-3, n=5 mice per group). GZMB, granzyme B; LAG-3, Lymphocyte Activation Gene-3; PD-1, Programmed Death-1.
Discussion
Immunotherapy, such as anti-CTLA-4 and anti-PD-1, has significantly improved the outcomes of patients with cancer. Nevertheless, the overall response rate of patients with LUAD is far from satisfactory, which calls for further exploration of other effective immune checkpoint pathways and combination therapies. Recent studies have highlighted the great potential of small-molecule chemotherapeutics for reshaping and activating the TIME. HDACi has demonstrated promising synergistic antitumor potential when combined with anti-PD-1/PD-L1 and anti-CTLA-4 mAbs in preclinical models of breast, ovarian, prostate, colon, liver, pancreatic, lung cancers, as well as melanoma.14 20 40–42 In recent years, HDACis have shown anticancer potential in combination with other emerging ICBs in laboratory models. Specifically, in melanoma mouse models, HDAC6 inhibitor enhanced the in vivo antitumor activity of anti-CD47 by modulating the functional status of macrophages and natural killer cells.43 In this study, we investigated molecular mechanisms underlying the activation of CD8+T cell-based immune activity induced by SAHA and explored the potential benefits of combining SAHA with anti-LAG-3 in LUAD, thereby unlocking novel avenues for the development of combinatorial cancer immunotherapy strategies in LUAD.
We chose SAHA, an HDACi that was first used in the clinic and found to be well tolerated, as our research object. There is evidence showing that treatment with SAHA facilitates the infiltration of CD8+T cells in human lung tumors.44 In line with previous studies, our results indicated that SAHA activated CD8+T-cell-mediated antitumor immunity in vitro and in vivo, enhancing both the infiltration and effector function of CD8+T cells and thus yielding an immunostimulatory tumor microenvironment. Based on our RNA-seq data, we found that SAHA treatment modulated the expression of the immune checkpoint molecules MHC-I and MHC-II. Notably, SAHA treatment increased PD-L1 expression but downregulated FGL1 expression. FGL1 is the newly discovered ligand of LAG-3,27 one of the most promising inhibitory immune checkpoints which has been reported to have a more profound effect on the suppression of T-cell proliferation and activation than PD-1 to facilitate immune evasion.45 Although there is controversy over whether FGL1 can surpass MHC-II as the main functional ligand of LAG-3,46 an increasing number of studies have confirmed that there is a strong affinity between FGL1 and LAG-3.28 47 Moreover, FGL1 has been reported to predict the efficacy of PD-L1/PD-1 blockade therapy, suggesting a possible regulatory mechanism between these immune checkpoints.27 In this study, we confirmed that SAHA treatment enhanced the infiltration of CD8+T cells, creating an inflamed TIME. Therefore, we focused on the transcriptional inhibition of inhibitory checkpoints by SAHA and discovered that ectopic expression of FGL1 attenuated CD8+T cell immunity activated by targeting HDAC1 both in vitro and in vivo. These results suggest that targeting FGL1 is a potential strategy for LUAD immunotherapy regardless of PD-L1 expression.
Although most recent studies have investigated the post-translational modifications of FGL1,28 48 our finding that HDAC1 downregulated FGL1 at both the mRNA and protein levels led us to speculate that HDAC1 might affect FGL1 by regulating transcription factors. Through database screening and mass spectrometric analysis, we found that HDAC1 transcriptionally regulated FGL1 expression via the JAK1/STAT3 pathway. Numerous studies have suggested that the JAK1/STAT3 pathway is extensively overactivated in cancer development and plays a critical role in suppressing the expression of key immune activation regulators while increasing the generation of immunosuppressive factors.49 In the present study, we investigated the interaction between HDAC1 and JAK1 and further revealed that silencing HDAC1 enhanced the acetylation at the K1109 site of JAK1, promoting its ubiquitination-mediated degradation and subsequently inhibiting the transcription of FGL1. The detailed molecular mechanism is depicted in online supplemental figure S8. Notably, given the varied effects of different HDAC family members on the JAK1/STAT3 pathway,50 more specific selective HDAC1 inhibitors need to be investigated in depth as a promising strategy for LUAD immunotherapy.
Previous studies on the mechanisms of HDACis combined with ICB in activating the TIME have suggested that HDACi may enhance immune responses by augmenting tumor antigen presentation, promoting T-cell activation and infiltration, and diminishing the activity of immunosuppressive cells. ICB, on the other hand, restores the function of CD8+T cells against tumors by blocking immune checkpoint signals, particularly PD-1/PD-L1. Our study uniquely revealed that HDACi activated CD8+T cell antitumor activity through the novel modulation of the FGL1/LAG-3 immune checkpoint axis, thereby broadening our understanding of HDACi-mediated immune activation and presenting an innovative perspective for immunotherapy strategies of LUAD. Importantly, we also highlight the pivotal role of HDAC1 in modulating the post-translational acetylation status of the JAK1/STAT3/FGL1 signaling cascade via its deacetylase activity, which underscores the complex interplay between epigenetic regulatory genes and immune checkpoint signaling in antitumor immunity of LUAD.
However, this study bears certain limitations. More in vivo tumor-bearing mice models less sensitive to the SAHA and anti-LAG3 monotherapy need to be tested to evaluate the efficacy of combining HDACis with ICBs. Additional mechanisms by which SAHA and anti-LAG-3 synergize to suppress LUAD growth warrant deeper investigation. Moreover, the effects of SAHA and anti-LAG3 on absolute immune cell counts in the TIME remain further exploration to refine our understanding of lymphocyte dynamics with different immunotherapy regimens. Lastly, the lack of direct acetylome sequencing for JAK1 hinders a full grasp of its acetylation regulated by HDAC1, underscoring the need for further research into molecular mechanisms governing JAK1 kinase activity and HDAC1-modulated downstream pathways to develop HDAC1/JAK1-targeted therapies.
Together, our results reveal that the HDACi SAHA improves the immunotherapy response in preclinical models of LUAD and elucidate the role and regulatory mechanism of the HDAC1/JAK1/STAT3/FGL1 axis in tumor progression and immune escape, suggesting that HDAC1 is a potential target for combinational therapeutic strategies to improve ICB response.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
All procedures related to the tissue array were approved by the Life Sciences Ethics Committee of Changsha Yaxiang Biotechnology with the authority granting ethics approval number of Csyayj2024056. Participants gave informed consent to participate in the study before taking part.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (82173347, 82172889), Jiangsu Province Hospital High-level Talent Cultivation Program (Phase I) (CZ0121002010037), Jiangsu Province Capability Improvement Project through Science, Technology and Education (CXZX202204) and Beijing Xisike Clinical Oncology Research Foundation (Y-2022METAZQN-0012). We would like to thank the Core Facility of the First Affiliated Hospital of Nanjing Medical University for its help in the detection of experimental samples.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
TX, YF, YG and DX contributed equally.
Contributors TX, YF, YG and DX are joint first authors. TX designed the experiment. TX, YF, YG and DX performed the experiments and acquired data. TH, TY, Y-YX and H-YS assisted with the data and figure integration. YS and PM supervised the entire work of the article and provided the necessary funding. TX wrote the original manuscript of the article. PM and YS reviewed and edited the original manuscript. All authors read and approved the final manuscript. YS and PM are responsible for the overall content as the guarantor.
Funding This work was supported by the National Natural Science Foundation of China (82173347, 82172889), Jiangsu Province Hospital High-level Talent Cultivation Program (Phase I) (CZ0121002010037), Jiangsu Province Capability Improvement Project through Science, Technology and Education (CXZX202204) and Beijing Xisike Clinical Oncology Research Foundation (Y-2022METAZQN-0012).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.