You are here

  • Home
  • Archive
  • Volume 12, Issue 2
  • Salmonella-mediated methionine deprivation drives immune activation and enhances immune checkpoint blockade therapy in melanoma

Article Text

Article menu
Clinical/translational cancer immunotherapy
Original research
Salmonella-mediated methionine deprivation drives immune activation and enhances immune checkpoint blockade therapy in melanoma
  1. Sujin Zhou1,
  2. Shiwei Zhang1,
  3. Kexin Zheng1,
  4. Zixuan Li1,
  5. Enyu Hu1,
  6. Yunping Mu1,
  7. Jialuo Mai2,
  8. Allan Zhao1,
  9. Zhenggang Zhao1 and
  10. Fanghong Li1
  1. 1School of Biomedical and Pharmaceutical Sciences, Guangdong University of Technology, Guangzhou, Guangdong, China
  2. 2Guangzhou Sinogen Pharmaceutical Co Ltd, Guangzhou, Guangdong, China
  1. Correspondence to Zhenggang Zhao; zhaozg{at}gdut.edu.cn; Fanghong Li; fli{at}gdut.edu.cn

Abstract

Background Although immune checkpoint inhibitor (ICI)-based therapy is advantageous for patients with advanced melanoma, resistance and relapse are frequent. Thus, it is crucial to identify effective drug combinations and develop new therapies for the treatment of melanoma. SGN1, a genetically modified Salmonella typhimurium species that causes the targeted deprivation of methionine in tumor tissues, is currently under investigation in clinical trials. However, the inhibitory effect of SGN1 on melanoma and the benefits of SGN1 in combination with ICIs remain largely unexplored. Therefore, this study aims to investigate the antitumor potential of SGN1, and its ability to enhance the efficacy of antibody-based programmed cell death-ligand 1 (PD-L1) inhibitors in the treatment of murine melanoma.

Methods The antitumor activity of SGN1 and the effect of SGN1 on the efficacy of PD-L1 inhibitors was studied through murine melanoma models. Further, The Cancer Genome Atlas-melanoma cohort was clustered using ConsensusClusterPlus based on the methionine deprivation-related genes, and immune characterization was performed using xCell, Microenvironment Cell Populations-counter, Estimation of Stromal and Immune cells in MAlignant Tumor tissues using Expression data, and immunophenoscore (IPS) analyses. The messenger RNA data on programmed death-1 (PD-1) immunotherapy response were obtained from the Gene Expression Omnibus database. Gene Set Enrichment Analysis of methionine deprivation-up gene set was performed to determine the differences between pretreatment responders and non-responders.

Results This study showed that both, the intratumoral and the intravenous administration of SGN1 in subcutaneous B16-F10 melanomas, suppress tumor growth, which was associated with an activated CD8+T-cell response in the tumor microenvironment. Combination therapy of SGN1 with systemic anti-PD-L1 therapy resulted in better antitumor activity than the individual monotherapies, respectively, and the high therapeutic efficacy of the combination was associated with an increase in the systemic level of tumor-specific CD8+ T cells. Two clusters consisting of methionine deprivation-related genes were identified. Patients in cluster 2 had higher expression of methionine_deprivation_up genes, better clinical outcomes, and higher immune infiltration levels compared with patients in cluster 1. Western blot, IPS analysis, and immunotherapy cohort study revealed that methionine deficiency may show a better response to ICI therapy

Conclusions：This study reports Salmonella-based SGN1 as a potent anticancer agent against melanoma, and lays the groundwork for the potential synergistic effect of ICIs and SGN1 brought about by improving the immune microenvironment in melanomas.

  • Combined Modality Therapy
  • Drug Evaluation, Preclinical
  • Drug Therapy, Combination
  • Melanoma

Data availability statement

Data are available in a public, open access repository. No data are available.

http://creativecommons.org/licenses/by-nc/4.0/

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/.

https://doi.org/10.1136/jitc-2023-008238

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • There is a crucial need to identify effective combinations of immune checkpoint inhibitors and develop new therapies for the treatment of melanoma.

    • SGN1, a genetically modified Salmonella typhimurium species that causes the targeted deprivation of methionine in tumor tissues, is currently under investigation in clinical trials and has exhibited excellent inhibitory effects on tumor growth and metastasis in several cancer types.

    WHAT THIS STUDY ADDS

    • Both the intratumoral and the intravenous administration of SGN1 significantly retarded the growth of melanoma, which might be related to an increased intratumoral infiltration of CD8+T cells.

    • Combination therapy of SGN1 with anti-programmed cell death-ligand 1 therapy resulted in better antitumor activity than the individual monotherapies, respectively.

    • Tumor-specific methionine restriction drives immune activation in patients with melanoma, which could increase in response to immune checkpoint inhibitors.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • SGN1 is an effective anticancer agent against melanoma. Meanwhile, it can be used to boost responses to immunotherapies in patients with melanoma.

    Background

    Malignant melanoma is the most aggressive form of skin cancer with increasing incidence worldwide.1 Even though immune checkpoint inhibitor (ICI) therapies have revolutionized the treatment of metastatic melanoma, the majority of patients still suffer from a lack of durable response to therapy.2 3 Therefore, novel therapies as well as novel combination treatments involving programmed death-1 (PD-1)/programmed cell death-ligand 1 (PD-L1) immune blockers, which produce an increased response rate in patients with cancer, are needed urgently.

    Almost all cancer cells, particularly tumor-initiating cells, are strongly dependent on methionine for their growth and metastasis.4–6 This is largely attributed to an impaired methionine metabolism pathway,7 8 including melanoma cells.9 Methionine has been implicated in several crucial cellular activities, such as protein synthesis, purine synthesis, folate trapping,10 and epigenetic changes.11–13 Recently, pharmacological research has focused on the development of inhibitors of methionine metabolic pathways, low-methionine diets, or recombinant L-methioninase as tumor-killing techniques in different cancers.14–16 Each approach has certain clinical and practical disadvantages, such as high toxicity and off-target activity.8 17 18

    Several bacterial strains have been evaluated in animal models and demonstrated preferential targeting of solid tumors, with several of them, such as Salmonella VNP2000919 20 have advanced to clinical trials.21 22 VNP20009 was attenuated by chromosomal deletion of purI and msbB genes, generating a modified lipid A with no functional endotoxin production, and a purine auxotrophic mutation.19 20 To specifically deprive cancer tissues of methionine without affecting systemic methionine, we have developed an attenuated Salmonella-based agent, SGN1, which is a genetically modified strain of VNP20009 that overexpress an L-methioninase.23 SGN1 can target, and preferentially replicate in the tumors, specifically depriving them of methionine, an essential amino acid.23 Currently, SGN1 has been approved for clinical study (NCT05103345 and NCT05038150) by the Food and Drug Administration, the Taiwan Food and Drug Administration, and China’s National Medical Products Administration. However, the inhibitory effect of SGN1 on melanoma remains unknown. Additionally, Salmonella species24 25 as well as methionine restriction have been known to increase antitumor immunity and improve the efficacy of therapies for checkpoint blockade.12 26 The efficacy of the SGN1-ICI combination, on the other hand, is uncertain.

    Herein, we have investigated the antitumor potential and immunomodulatory activity of SGN1, and its ability to enhance the efficacy of antibody-based PD-L1 inhibitors in the treatment of murine melanoma. Our findings show that Salmonella SGN1 has strong anticancer potential against melanoma, thereby laying the framework for the synergistic treatment of ICIs with SGN1 in melanoma.

    Results

    SGN1 significantly inhibits the growth of melanoma in vivo

    First, we evaluated the antitumor activity of SGN1 in a B16-F10 melanoma subcutaneous xenograft model. Once the tumors reached a volume of approximately 100 mm3, the mice were randomly assigned to receive a single intratumoral injection of SGN1 (2×105, or 2×106 CFU/mouse), VNP-V (control VNP20009 carrying a non-expressing vector, 2×105, or 2×106 CFU/mouse), or phosphate-buffered saline (PBS). Compared with the PBS or VNP-V-treated groups, SGN1 significantly suppressed tumor growth at a dose of 2×105 or 2×106 CFU at 9 days post-treatment (p<0.001) (figure 1A–C). As compared with the PBS group, the mean tumor size decrease was 49.5% and 61.3%, and 32.7% and 45.6% when compared with the same dose of VNP-V in the SGN1 2×105 and SGN1 2×106 groups, respectively (figure 1A–C).

    Figure 1
    Figure 1

    SGN1 significantly inhibits the growth of melanoma in vivo. A single injection of vehicle (PBS), VNP-V, or SGN1 is administered to C57/B6 mice bearing B16-F10 melanoma xenografts. (A–C) A single intratumoral injection of SGN1 (2×105, or 2×106 CFU/mouse), VNP-V (2×105, or 2×106 CFU/mouse) or PBS. (D–H) A single intravenous injection of SGN1 (2×104, 2×105, or 2×106 CFU/mouse), VNP-V (2×106 CFU/mouse), or PBS after the tumors attain the expected size. (A and D) SGN1 suppresses tumor growth significantly in B16-F10 melanoma xenografts; n=5 mice per group, data are expressed as mean±SEM of the biological replicates; **p<0.01. (B and E) B16-F10 melanoma growth over time in C57/B6 mice. n=5 mice per group. (C and F) Representative images of the subcutaneous tumors at the end of the treatment. (G) H&E staining of the xenograft tumors collected after treatment with PBS, VNP-V, or SGN1. (H and I) Flow cytometric analysis of T cells in B16-F10 melanoma xenografts; flow cytometric scatter plots and gating in the analysis of the cell surface expression of CD45 and CD8 markers. n=5 mice per group, *p<0.05, **p<0.01. N, necrosis; PBS, phosphate-buffered saline.

    Further, we also examined the effect of SGN1 on the growth of melanoma by intravenously administering a single dose of SGN1 (2×104, 2×105, or 2×106 CFU/mouse), VNP-V (2×106 CFU/mouse), or PBS. Results showed that SGN1 inhibited B16-F10 melanoma subcutaneous xenograft growth in a dose-dependent manner (the mean tumor size decrease was 36.1%, 50.1% and 82.1%, as compared with the PBS group in the SGN1 2×104, 2×105, and 2×106 groups, respectively, p<0.001), with significantly smaller tumor volumes in the SGN1 group than in the VNP-V group at 6 days post-treatment (mean tumor size: 0.6382 vs 1.626 cm3 for SGN1 2×106 vs VNP-V 2×106, p<0.001) (figure 1D–F). H&E-staining analysis revealed an extensive necrotic area in the SGN1 treatment group (figure 1G). A study of the bacterial distribution revealed enduring and preferentially bacterial colonization in tumors than in the livers (online supplemental figure S1A). Furthermore, mice receiving SGN1 during treatment did not exhibit any significant changes in body weight or any adverse effects relative to the PBS-treated group (online supplemental figure S1B,C).

    Supplemental material

    Recent studies have demonstrated that a sulfur amino acid-restricted diet or methionine deficiency can lead to an increase in the CD8+ T-cell count.12 27 Therefore, we also evaluated the tumor samples using flow-based immunophenotyping for CD8 and CD45 expression to detect changes in the population and phenotype of the intratumoral CD8+ T cells. Our findings revealed that both groups treated with bacteria experienced a significant increase in the proportion of intratumoral CD8+ T cells, while the group treated with SGN1 showed a significant increase in the infiltration of CD8+ T cells in the tumor (figure 1H,I). These results indicated that SGN1 significantly inhibited the growth of melanoma, which might be related to an increased intratumoral infiltration of CD8+T cells.

    Synergistic inhibitory effect of SGN1 and anti-PD-L1 therapy in vivo

    We next investigate whether SGN1 might enhance the effect of anti-PD-L1 therapy in mice with melanoma. Once the tumors had attained the expected size, SGN1 was administered intravenously at a single dose of 2×104 CFU/mouse on day 0. Subsequently, anti-PD-L1 monoclonal antibody (PD-L1 mAb,150 µg/mouse) was administered intraperitoneally on day 1, 4, 7 of a 9-day cycle) (figure 2A). The results revealed that the mice in the PD-L1 mAb alone or SGN1 alone groups displayed significantly inhibited B16-F10 tumor growth compared with the mice in the PBS group (the mean tumor size decrease was 32.8% or 39.9%, as compared with the PBS group in the PD-L1 mAb alone or SGN1 alone group, respectively, p<0.001) (figure 2B–D). The combination therapy of SGN1 and PD-L1 mAb exhibited a substantial antitumor effect compared with the individual monotherapies, respectively (the mean tumor size decrease was 35.3% or 27.6%, as compared with the PD-L1 mAb alone or SGN1 alone group, respectively, p<0.001) (figure 2B–D). H&E-staining, which was performed on the post-treatment tumor samples, revealed an extensive necrotic area in the SGN1 alone and the SGN1 and PD-L1 mAb combined treatment group (figure 2E). Flow cytometry analysis showed that treatment of mice with a combination of SGN1 and PD-L1 mAb significantly increased the tumor-infiltrating CD8+ T-cell population (figure 2F,G). Moreover, during treatment, the four groups of mice appeared as active as the mice in the two control groups, with no significant weight loss (online supplemental figure S1C). These findings reveal that SGN1 acts synergistically with ICIs and produces an enhanced antitumor effect on melanoma cells compared with either therapy alone.

    Figure 2
    Figure 2

    Synergistic inhibitory effect of SGN1 and anti-PD-L1 on B16-F10 melanoma subcutaneous xenograft model. (A) Schematic representation of the protocols performed and their timelines, in the treatment of B16-F10 melanoma using SGN1 combined with anti-PD-L1. (B) Tumor volume in different groups at the end of the treatment (n=5 mice per group). (C) Tumor growth curves in different groups. Data are presented as mean±SEM (n=5 mice per group). (D) Representative images of the subcutaneous tumors at the end of the treatment. (E) H&E staining of the xenograft tumors collected post-treatment. (F and G) Percentage of CD8+ and CD3+ T cells in tumor infiltrating lymphocyte. n=3 mice per group. *p<0.05, **p<0.01, ***p<0.001. N, necrosis; PD-L1, programmed cell death-ligand 1.

    Methionine deprivation boosts immune activation

    To identify the relationship between the methionine deprivation-related genes and T-cell activation in melanoma, cluster analysis was performed on patients with melanoma whose data were acquired from The Cancer Genome Atlas database (TCGA) using ConsensusClusterPlus package28 based on the genes retrieved from the gene sets in the molecular signatures database (MSigDB) V.2023.1. Hs including KOKKINAKIS_METHIONINE_DEPRIVATION_96 hours_UP and KOKKINAKIS_METHIONINE_DEPRIVATION_48 hours_UP (online supplemental table S1). Cluster analysis revealed two distinct clusters, cluster 1 (n=239) and cluster 2 (n=232) (figure 3A,B, online supplemental table S2). In cluster 2, the expression of the 144 methionine_deprivation_up genes was relatively higher compared with that in cluster 1 (figure 3C). In addition, the patients in cluster 2 had significantly better clinical outcomes than those in cluster 1 (figure 3D, online supplemental figure S2A). However, no significant difference was noted in the clinical–pathologic characteristics between the two clusters (online supplemental figure S2A-E). To further understand the association between methionine deprivation gene expression and antitumor response, we conducted the functional enrichment analysis using differential gene analysis between the two clusters, followed by the Gene Set Enrichment Analysis (GSEA) using the Kyoto Encyclopedia of Genes and Genomes (KEGG) gene sets. The results demonstrated that the most represented downregulated gene sets in cluster 2 were associated with DNA_Replication, Aminoacyl_tRNA_Biosynthesis, and Oxidative_Phosphorylation, whereas most of the upregulated gene sets in cluster 2 were associated with immune and inflammatory responses, such as Cytokine_cytokine_receptor_interaction, TGF_beta_signaling_pathway, and Intestinal_immune_network (figure 3E,F).

    Supplemental material

    Supplemental material

    Figure 3
    Figure 3

    Methionine deprivation boosts immune cell activation. (A) Consensus clustering matrix when k=2. (B) Consensus clustering CDF with k=2–9. (C) Heatmap depicting the expression of the 144-methionine deprivation-related genes between the two clusters. (D) The Kaplan-Meier survival curve showing the survival difference between cluster 1 and cluster 2. (E, F) Gene Set Enrichment Analysis based on the differential genes between cluster 1 and cluster 2 in TCGA-melanoma patients. The most represented upregulated or downregulated gene sets were displayed. (G) An overview of xCell scores for immune cell types between the two clusters of patients with TCGA-melanoma. (H, I) immune and stromal scores using xCell (H) or ESTIMATE (I). (J, K) CD8+ T-cell (J) and cytotoxic lymphocyte (K) population between the two clusters using Microenvironment Cell Populations-counter analyses. *p<0.05, **p<0.01, ***p<0.001. aDC, activated dendritic cells; cDC,conventional dendritic cells ; CDF, cumulative distribution function; DC,dendritic cell; ESTIMATE, Estimation of Stromal and Immune cells in MAlignant Tumor tissues using Expression data; KEGG, Kyoto Encyclopedia of Genes and Genomes; NK, natural killer; TCGA, The Cancer Genome Atlas.

    We next investigated the characteristics of the immune microenvironment in melanoma of the two clusters by performing xCell and Estimation of Stromal and Immune cells in MAlignant Tumor tissues using Expression data analysis (ESTIMATE) to profile immune characteristics. The xCell algorithm reveals a pronounced upregulation of CD4+ T cells, CD8+ T cells, dendritic cells, and other immune cells in cluster 2 (figure 3G). Both the xCell and ESTIMATE algorithms showed that the patients in cluster 2 displayed significant overall infiltration of immune cells (immune score) and a higher percentage of stromal cells (stromal score) compared with those in cluster 1 (figure 3H,I). In addition, Microenvironment Cell Populations-counter (MCPcounter) analyses showed that antitumor immune cell populations such as CD8+ T cells and cytotoxic lymphocytes were enriched within cluster 2 compared with cluster 1 (figure 3J,K). These results suggested that methionine deprivation was associated with increased immune activation in patients with melanoma.

    Methionine deprivation could increase in response to ICIs

    Since CD8+ T-cell infiltration is strongly associated with the level of PD-L1 in tumor cells,29 we evaluated the PD-L1 (CD274 genes) expression levels in the two clusters and found that cluster 2 showed higher expression of PD-L1 compared with cluster 1 (figure 4A). We treated B16F10 cell lines with methionine-deficient medium (Met-) to deprive methionine levels and measured PD-L1 expression using the fluorescence‐activated cell sorting analysis. A comparison of PD-L1 expression of cells treated with methionine medium (Met+) and (Met−) demonstrated that the latter exhibited significantly higher expression (the mean percentage expression of PD-L1 was 48.2% vs 24.5% for Met− vs Met+ on days 3 post-treatment, p<0.001) (figure 4B,C). A high expression of PD-L1 was observed in the melanoma tissues of mice intravenously injected with SGN1 compared with the PBS or VNP-V group (2.030 vs 0.67 or 0.916; p<0·001) (figure 4D,E). The immunophenoscore (IPS), a superior predictor of response to ICIs in patients with cancer,30 was used to further evaluate the response to immunotherapy. The results demonstrated that the patients in cluster 2 had a higher IPS than those in cluster 1 (figure 4F,G). We further analyzed messenger RNA (mRNA) expression and clinical response data from the Gene Expression Omnibus (GEO) database in pretreatment with anti-PD-1 checkpoint inhibition therapy. The GSEA results revealed that the methionine_deprivation_up pathway was significantly elevated in the responders of ICIs among patients with melanoma compared with that in non‐responders (figure 4H). Similar results were observed in non-small cell lung carcinoma, and patients with head and neck squamous cell carcinoma (figure 4I,J). These results indicate that the relative probability of responding to ICIs is higher in the methionine deficiency group.

    Figure 4
    Figure 4

    Methionine deprivation could increase in response to immune checkpoint inhibitor. (A) CD274 expression is significantly increased in cluster 2. (B, C) PD-L1 levels, as measured by flow cytometry, on B16-F10 cells cultured in methionine-free medium (Met-) and supplemented with 100 µM methionine (Met+) on days 1, 2, and 3 (n=3 of biological replicates). (D, E) Western blot analysis of PD-L1 in tumors intravenous injected with 2×106 CFU/mouse SGN1 (n=3 of biological replicates). (F, G) Average immunophenogram in each cluster. The inner circle is a plot of each of the four IPS components, with higher values representing a more immunogenic phenotype. The outer circle is a plot of the expression of the markers used to compute each of the IPS components. (H–J) Gene Set Enrichment Analysis was performed in response patients versus non‐response patients to ICBs from data set GSE93157. Immunotherapy: anti‐PD-1 therapy . *p<0.05, **p<0.01, ***p<0.001. APC,antigen-presenting cells; CTLA, cytotoxic T-lymphocyte–associated protein 4; CR, complete response; HNSC, head and neck squamous cell carcinoma; HLA, human leukocyte antigens; ICOS,inducible co-stimulator; IDO1, indoleamine 2,3-dioxygenase 1; IPS, immunophenoscore; MDSC, myeloid-derived suppressor cell; LAG3, lymphocyte activation gene 3; MHC, major histocompatibility complex; NES, normalized enrichment score; NSCLC, non-small cell lung carcinoma; PBS, phosphate-buffered saline; PD, progressive disease; PD-1,programmed death 1; PD-L1, programmed cell death-ligand 1; PR, partial response; SD, stable disease; TAP, transporter associated with antigen processing;TCGA, The Cancer Genome Atlas; Tem, effector memory T cells; TIGIT, cell immunoglobulin and ITIM domain; TIM3,T cell immunoglobulin and mucin domain 3; Treg, regulatory T cells.

    Discussion

    Melanoma is one of the most common malignancies worldwide, and the most aggressive and fatal form of skin cancer. Although favorable results have been obtained from immunotherapy, in general, resistance and relapse are frequent.31 In the past decade, efforts have been made to overcome resistance and improve treatment outcomes by combining PD-1/PD-L1 inhibitors with other medicine.32 In this study, we found that SGN1, a genetically modified Salmonella typhimurium species, significantly retarded the growth of melanoma, which could be related to an increased intratumoral infiltration of CD8+ T cells. A low dose of SGN1 was used to improve the response rate and therapeutic efficacy of PD-L1 blockers in a murine melanoma model. In addition, we found that methionine deprivation may affect T cells and other immune cells in patients with melanoma, which could increase in response to ICIs. These findings suggest that SGN1 is an effective anticancer agent against melanoma and that it can be used to boost responses to immunotherapies in patients with melanoma.

    The scientific world has recently witnessed an increasing interest in the use of attenuated bacteria as cancer therapeutic tools over the past 10 years.33 This is largely due to the development of bacterial strains maintaining good antitumor efficacy but with reduced potential to cause toxicities to the host. The bacteria carrier of SGN1, attenuated Salmonella VNP20009, was shown to be non-pathogenic to mice and at least 50,000×less virulent than the parental Salmonella in mice.34 In the present study, mice receiving SGN1 did not exhibit any significant changes in body weight or any adverse effects (such as fever or anything related to systemic active infection) relative to the PBS-treated group. Our previous study reported that when immunodeficient mice (nude mice) were injected with SGN1 at a dose of 2×108/mouse (100–10,000 times the dose used in this study), no noticeable adverse effects were found.23

    Salmonella-mediated antitumor effects are exerted via their capability to target tumors and preferentially colonize the core area of the tumor, inducing direct tumor cell killing.35 A pilot clinical trial and a phase I clinical trial of genetically engineered VNP20009 reported intratumor evidence of bacterial colonization in patients with cancer that lasted at least 15 days following the initial injection.22 36 Our previous published data demonstrated that a single dose of SGN1 preferentially accumulated in tumors for at least 20 days.23 Although our previous published data demonstrated that repeated injections of VNP20009 induced a stronger anticancer effect than a single injection.37 Moreover, animals could tolerate well at 0.5×108 CFU/kg in a 4-week repeated dose toxicity study of SGN1,23a single administration of SGN1 could be sufficient to provide clinical benefit.

    In addition to the excellent safety and high tumor targeting efficacy, Salmonella can shift the tumor microenvironment from an immunosuppressive to an immunogenic environment.38 This success could be achieved through increased infiltration and reprograming of antitumor immune cells, upregulating the expression of proinflammatory cytokines and inducing a shift in the phenotypic and functional characteristics of immune cells.38 Hence, SGN1 might undoubtedly induce immunological sensitization. Nevertheless, SGN1 performed substantially better than the control VNP-V in activating an antitumor immune response and decreasing melanoma growth, implying that SGN1-specific methionine restriction plays an important role in this process. Recent research has shown that a sulfur amino acid-restricted diet or methionine deficiency can result in an increase in CD8+T-cell count.12 27 These observations are consistent with our finding that SGN1 can significantly inhibit the growth of melanoma, which could be related to an increased intratumoral infiltration of CD8+ T cells due to the restriction of methionine. However, there is no direct evidence relating SGN1 to enhanced intratumoral infiltration of CD8+T cells. On the other hand, methionine is also critical for T-cell activation and differentiation.39 40 Initiating dietary methionine restriction 3–4 weeks prior to tumor formation decreases T-cell abundance in immunocompetent mice.41 To identify the relationship between the methionine deprivation and T-cell activation in melanoma, we divided the TCGA-melanoma cohorts into two different clusters using consensus clustering based on the 144 methionine_deprivation_up genes. Patients in cluster 2 showed higher expression levels of the genes upregulated after methionine deprivation, and had higher levels of immune cell infiltration compared with those in cluster 1. These results indicated that tumor-specific methionine deprivation was associated with increased immune activation in patients with melanoma. Undoubtedly, the direct evidence supporting the link between methionine deprivation and T-cell activation in melanoma should be systematically obtained in the future.

    The effect of methionine deprivation on immune cells other than T cells remains poorly understood. We found that methionine deprivation induced a favorable immune profile in patients with melanoma. Apart from CD8+ T cells, the number of macrophages and some other immune cells were significantly increased in cluster 2. IPS analysis comprehensively quantifies tumor immunogenicity by merging the expression of antigen processing cells (major histocompatibility complex (MHC)), checkpoints, immunomodulators, effector cells (activated CD8, active CD4, effector memory T cells) (Tem CD4, Tem CD8), and suppressor cells (regulatory T cells (Treg) and myeloid-derived suppressor cell (MDSC)).30 We found that the patients in cluster 2 possessed significantly higher IPS than those in cluster 1. This work is the first to demonstrate that methionine restriction may affect immune cells other than T cells; however, this requires further study.

    Although some combination therapies can increase the response rates of patients with cancer to PD-1/PD-L1 immune blockers, the low response rate to α-PD-1/PD-L1 therapy remains unresolved.42 In this study, we found that the patients with melanoma in cluster 2 possessed significantly higher IPS than those in cluster 1, which means methionine deprivation may produce a better response to anti-PD-L1 therapy than those in cluster 1. These results are also consistent with previously published findings that methionine restriction enhances antitumor immune response and improves the efficacy of anti-PD-L1 treatments.12 27 Importantly, we found that treatment with an anti-PD-1 antibody combined with SGN1 was associated with significant tumor regression. It also showed that the number of tumor-infiltrating CD8+ T cells was higher in the combination group compared with the monotherapy groups, indicating the effectiveness of the combination in boosting antitumor immunity. Our results suggest that SGN1 could be a potential candidate to combine with anti-PD-L1 antibody. Notably, we found that such synergistic therapeutic effects may also be observed in other types of cancers.

    Conclusion

    In summary, our data has shown a potential novel treatment strategy or synergistic modality with ICIs for melanoma. In addition to demonstrating the potent nature of SGN1, or SGN1 coupled with an ICI on melanoma, for the first time, the findings also reveal that this novel treatment modality is not restricted to only melanoma but can also be extended to other types of malignancies as well.

    Methods

    Cell culture

    The melanoma cell line, B16F10, was purchased from the Cell Bank of the Typical Cultures Preservation Committee, Chinese Academy of Sciences. Cells were cultured in the medium recommended by the manufacturer. For the methionine restriction protocol, the cells were cultured in a methionine-free medium (Gibco, 21013024) and supplemented with 100 µM methionine (Sigma M5308-25G).

    Bacterial strains

    The S. typhimurium strain SGN1 and VNP20009-V were donated by Guangzhou Sinogen Pharmaceutical (Guangzhou, China). Required concentrations of the bacterial culture were obtained by diluting the culture in normal saline, estimating the optical density of the culture at 600 nm, and performing a colony count using the flat colony counting method.

    Mice and tumor cell implantation

    Six-to-seven-week-old female C57BL/6 mice were obtained from the Model Animal Research Center of Nanjing University. In accordance with the “3Rs policy”, experiments were reviewed and approved by the Guangdong Pharmaceutical University Animal Care Committee (gdpulacspf2017064). The tumor cells were harvested post-trypsinization, washed in culture medium, counted, and diluted to appropriate concentrations for inoculation. Subsequently, the cells were injected subcutaneously into the flanks of mice. The tumor volume was calculated using the following formula: tumor volume=width2×length×0.52. The mice were euthanized at approximately 6-9 days after initial treatment, and their tumors were harvested and weighted using double-blinded evaluation. Partial tumor tissues were fixed with 4% paraformaldehyde and embedded in paraffin for immunohistochemical staining.

    Mice treatments

    After the tumors had grown to approximately 100 mm3, the tumor-bearing mice were divided randomly into different groups (n=5 per group). To test the synergistic effect of SGN1 and anti-PD-L1 therapy in vivo, mice were administered 2×104 CFU/mouse SGN1 intravenously or PD-L1 mAb (150 µg/mouse, Bio X Cell, BE0101-50MG) intraperitoneally as monotherapies (n=5 per group). The mice in the SGN1-PD-L1 mAb group received a single intravenous injection of SGN1 (2×104 CFU/mouse SGN1 on day 0) that combined intraperitoneal injection of PD-L1 mAb (150 µg/mouse, on day 1, 4, 7 of a 9-day cycle). The specific schematic view of the treatment plans was described in the figure 4A. In each group of mice, all mice were used to observe the tumor phenotype changes. The mice were euthanized at the end of the treatment, and their tumors were harvested and weighted using double-blinded evaluation. Single cell suspension of implanted tumors was prepared for flow cytometry analysis. Meanwhile, partial tumor samples were also for further western blot and histology.

    Tumor cell preparation and multicolor staining analyses using flow cytometry

    The extracted tumors were placed on ice and thoroughly sliced (1–3 mm cubes) using Roswell Park Memorial Institute (RPMI) 1640 dissociation solution containing 1 mg/mL collagenase IV (11088858001 Roche BR) and 1 mg/mL DNase I (10104159001 Roche BR). The tumor fragments were digested for 40 min in a dissociation solution (approximately 1 g of tumor fragments in 10 mL of dissociation solution). The digested tissue suspension was aspirated into a 20 mL syringe and the clumps were triturated 15 times and then passed through a strainer to obtain a single cell suspension, which was stained with the following antibodies: CD45 (E-ABF1136D, BD Bioscience), CD8 (553030, BD Pharmingen), and PD-L1 (13 684S, Cell Signaling Technology). Flow cytometry data were collected using BD FACSDiva (V.8.0.1) software and analyzed using the FlowJo software (V.10.7.1).

    Analysis of bacteria in tissues

    About 100 mg of tumors or livers were aseptically removed, weighed, and homogenized in 1 mL of ice-cold sterile PBS. Afterward, the homogenate was serially diluted and plated onto a modified LB/agar plate containing antibiotics for 20 hours at 37°C. The bacterial titer (CFU/g tissue) was calculated by counting the colonies and dividing the total weight of the tissue.

    Western blot and histology

    Total cell proteins were extracted from the mice in each group and quantified. Protein lysates were electrophoretically separated on 12% SDS-polyacrylamide gels and transferred to 0.2 µm nitrocellulose membranes. Primary antibodies against PD-L1 (13,684S, Cell Signaling Technology), and actin (3,700, Cell Signaling Technology) were used in the analysis. All tissues were embedded in paraffin, and the tissue sections were stained with H&E stain and subjected to histological analysis. Images were captured using a Nikon microscope.

    Data source

    The mRNA expression data and corresponding clinicopathological features of melanoma were obtained from the TCGA databases (https://tcgadata.nci.nih.gov).43 Cancer data sets with ICI therapy were collected from the GSE9315 data set44 of the GEO database (https://www.ncbi.nlm.nih.gov/geo/). Because the current study was an analysis based on publicly available databases with pre-existing institutional review board approval, ethics statement and informed consent were waived.

    Consensus molecular clustering by “ConsensusClusterPlus”

    For the consensus clustering analysis, read counts were converted to log2-counts-per-million and the ConsensusClusterPlus package was employed. Methionine deprivation-related genes (online supplemental table S1) were retrieved from the gene sets in the MSigDB V.2023.1. Hs including KOKKINAKIS_METHIONINE_DEPRIVATION_96 hours_UP and KOKKINAKIS_METHIONINE_DEPRIVATION_48 hours_UP. We used the k-means method and set 80% sampling each time along with 1,000 iterations to ensure the consistency of clustering. The optimal number of clusters was determined by the consensus heatmap and cumulative distribution function curves. The number of clusters was set as K=2.

    Survival analysis

    Survival distribution and significance were evaluated using Kaplan-Meier curves, and log-rank test analyses were performed using the survival and survminer packages in R software (V.4.2.2, http://www.r-project.org).

    Association between the clusters and immunity

    The immunological landscape was assessed using the xCell algorithm and ESTIMATE analysis.45 46 The xCell and ESTIMATE package in R was used to compute the immunological scores, and stromal scores. The ratios of the immune cell subtypes were assessed using the MCPcounter package based on the expression data.47 Comparison of the variations in the immune cell subtypes among the different clusters was performed using the Mann-Whitney U test.

    Detection of cancer neoantigens and immunophenoscores

    The IPS was calculated for each patient in the two clusters, and immunophenograms were created to display the various immunophenotypes in each tumor. The IPS scores (range: 0–10) were based on the expression of MHC molecules, cytotoxic T-lymphocyte–associated protein 4 (CTLA-4), lymphocyte activation gene 3 (LAG3), cell immunoglobulin and ITIM domain (TIGIT), hepatitis A virus cellular recep- tor 2 gene (HAVCR2), inducible co-stimulator (ICOS), indoleamine 2,3-dioxygenase 1 (IDO1), Programmed Cell Death 1 Ligand 2 Protein (PCDC1LG2), CD274,and CD27 (immunomodulators), activated CD8+ and CD4+ T cells (effector cells), Tem CD8+ and CD4+ T cells and Tregs, and MDSCs (immune suppressor cells).30

    Differentially expressed genes and functional enrichment analysis

    The mRNA data from the TCGA database was used as the input data, and differentially expressed gene (DEG) analysis was conducted using the DEGseq or limma R package. GSEA-based KEGG were conducted using the clusterProfiler R package.48 The “KOKKINAKIS_METHIONINE_DEPRIVATION_UP” curated gene sets were downloaded from the MSigDB.

    Quantification and statistical analysis

    All graphs and statistical procedures were performed using either GraphPad or or R software (version 4.2.2, http://www.r-project.org). A two-tailed Student’s t-test or one-way analysis of variance was performed followed by Tukey’s test, unless otherwise stated in the figure legend. Data are expressed as mean±SEM, and p<0.05 was considered as statistically significant.

    Data availability statement

    Data are available in a public, open access repository. No data are available.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Not applicable.

    References

    1. ↵
      1. Menzies AM,
      2. Gershenwald JE
      . Richard A Scolyer Cutaneous melanoma. Lancet 2023;402:01581–7. doi:10.1016/S0140-6736(23)00821-8
      OpenUrl
    2. ↵
      1. Wolchok JD,
      2. Chiarion-Sileni V,
      3. Gonzalez R, et al
      . Overall Survival with Combined Nivolumab and Ipilimumab in Advanced Melanoma. N Engl J Med 2017;377:1345–56. doi:10.1056/NEJMoa1709684
      OpenUrlCrossRefPubMed
    3. ↵
      1. Placke J-M,
      2. Kimmig M,
      3. Griewank K, et al
      . Correlation of tumor PD-L1 expression in different tissue types and outcome of PD-1-based immunotherapy in metastatic melanoma – analysis of the DeCOG prospective multicenter cohort study ADOREG/TRIM. eBioMedicine 2023;96:104774. doi:10.1016/j.ebiom.2023.104774
      OpenUrl
    4. ↵
      1. Wang Z,
      2. Yip LY,
      3. Lee JHJ, et al
      . Methionine is a metabolic dependency of tumor-initiating cells. Nat Med 2019;25:825–37. doi:10.1038/s41591-019-0423-5
      OpenUrlCrossRefPubMed
    5. ↵
      1. El-Kenawi A,
      2. Berglund A,
      3. Estrella V, et al
      . Elevated Methionine Flux Drives Pyroptosis Evasion in Persister Cancer Cells. Cancer Research 2023;83:720–34. doi:10.1158/0008-5472.CAN-22-1002
      OpenUrl
    6. ↵
      1. Mecham JO,
      2. Rowitch D,
      3. Wallace CD, et al
      . The metabolic defect of methionine dependence occurs frequently in human tumor cell lines. Biochem Biophys Res Commun 1983;117:429–34. doi:10.1016/0006-291x(83)91218-4
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Sanderson SM,
      2. Mikhael PG,
      3. Ramesh V, et al
      . Nutrient availability shapes methionine metabolism in p16/MTAP-deleted cells. Sci Adv 2019;5:eaav7769. doi:10.1126/sciadv.aav7769
    8. ↵
      1. Sanderson SM,
      2. Gao X,
      3. Dai Z, et al
      . Methionine metabolism in health and cancer: a nexus of diet and precision medicine. Nat Rev Cancer 2019;19:625–37. doi:10.1038/s41568-019-0187-8
      OpenUrlCrossRefPubMed
    9. ↵
      1. Liteplo RG
      . Altered methionine metabolism in metastatic variants of a human melanoma cell line. Cancer Lett 1989;44:23–31. doi:10.1016/0304-3835(89)90103-1
      OpenUrlPubMed
    10. ↵
      1. Sullivan MR,
      2. Darnell AM,
      3. Reilly MF, et al
      . Methionine synthase is essential for cancer cell proliferation in physiological folate environments. Nat Metab 2021;3:1500–11. doi:10.1038/s42255-021-00486-5
      OpenUrl
    11. ↵
      1. Hosios AM,
      2. Vander Heiden MG
      . The redox requirements of proliferating mammalian cells. J Biol Chem 2018;293:7490–8. doi:10.1074/jbc.TM117.000239
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Li T,
      2. Tan Y-T,
      3. Chen Y-X, et al
      . Methionine deficiency facilitates antitumour immunity by altering m6A methylation of immune checkpoint transcripts. Gut 2023;72:501–11. doi:10.1136/gutjnl-2022-326928
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Meng J,
      2. Wang L,
      3. Wang J, et al
      . METHIONINE ADENOSYLTRANSFERASE4 Mediates DNA and Histone Methylation. Plant Physiol 2018;177:652–70. doi:10.1104/pp.18.00183
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Gao X,
      2. Sanderson SM,
      3. Dai Z, et al
      . Dietary methionine influences therapy in mouse cancer models and alters human metabolism. Nature 2019;572:397–401. doi:10.1038/s41586-019-1437-3
      OpenUrlCrossRefPubMed
    15. ↵
      1. Kalev P,
      2. Hyer ML,
      3. Gross S, et al
      . MAT2A Inhibition Blocks the Growth of MTAP-Deleted Cancer Cells by Reducing PRMT5-Dependent mRNA Splicing and Inducing DNA Damage. Cancer Cell 2021;39:209–24. doi:10.1016/j.ccell.2020.12.010
      OpenUrl
    16. ↵
      1. El-Sayed AS
      . Microbial L-methioninase: production, molecular characterization, and therapeutic applications. Appl Microbiol Biotechnol 2010;86:445–67. doi:10.1007/s00253-009-2303-2
      OpenUrlCrossRefPubMed
    17. ↵
      1. Tan Y,
      2. Zavala J,
      3. Xu M, et al
      . Serum methionine depletion without side effects by methioninase in metastatic breast cancer patients. Anticancer Res 1996;16:3937–42.
      OpenUrlPubMed
    18. ↵
      1. Quinlan CL,
      2. Kaiser SE,
      3. Bolaños B, et al
      . Targeting S-adenosylmethionine biosynthesis with a novel allosteric inhibitor of Mat2A. Nat Chem Biol 2017;13:785–92. doi:10.1038/nchembio.2384
      OpenUrlCrossRefPubMed
    19. ↵
      1. Low KB,
      2. Ittensohn M,
      3. Le T, et al
      . Lipid A mutant Salmonella with suppressed virulence and TNFalpha induction retain tumor-targeting in vivo. Nat Biotechnol 1999;17:37–41. doi:10.1038/5205
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Kenneth BL,
      2. Martina I,
      3. Xiang L, et al
      . Construction of VNP20009: a novel, genetically stable antibiotic-sensitive strain of tumor-targeting Salmonella for parenteral administration in humans. Methods Mol Med 2004;90:47–60.
      OpenUrlPubMed
    21. ↵
      1. Toso JF,
      2. Gill VJ,
      3. Hwu P, et al
      . Phase I Study of the Intravenous Administration of Attenuated Salmonella typhimurium to Patients With Metastatic Melanoma . JCO 2002;20:142–52. doi:10.1200/JCO.2002.20.1.142
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Heimann DM,
      2. Rosenberg SA
      . Continuous intravenous administration of live genetically modified salmonella typhimurium in patients with metastatic melanoma. J Immunother 2003;26:179–80. doi:10.1097/00002371-200303000-00011
      OpenUrlCrossRef
    23. ↵
      1. Zhou S,
      2. Lin Y,
      3. Zhao Z, et al
      . Targeted deprivation of methionine with engineered Salmonella leads to oncolysis and suppression of metastasis in broad types of animal tumor models. Cell Reports Medicine 2023;4:101070. doi:10.1016/j.xcrm.2023.101070
      OpenUrl
    24. ↵
      1. Howell LM,
      2. Forbes NS
      . Bacteria-based immune therapies for cancer treatment. Seminars in Cancer Biology 2022;86:1163–78. doi:10.1016/j.semcancer.2021.09.006
      OpenUrl
    25. ↵
      1. Duong MT-Q,
      2. Qin Y,
      3. You S-H, et al
      . Bacteria-cancer interactions: bacteria-based cancer therapy. Exp Mol Med 2019;51:1–15. doi:10.1038/s12276-019-0297-0
      OpenUrlCrossRefPubMed
    26. ↵
      1. Fang L,
      2. Hao Y,
      3. Yu H, et al
      . Methionine restriction promotes cGAS activation and chromatin untethering through demethylation to enhance antitumor immunity. Cancer Cell 2023;41:1118–33. doi:10.1016/j.ccell.2023.05.005
      OpenUrl
    27. ↵
      1. Yue T,
      2. Li J,
      3. Zhu J, et al
      . Hydrogen Sulfide Creates a Favorable Immune Microenvironment for Colon Cancer. Cancer Res 2023;83:595–612. doi:10.1158/0008-5472.CAN-22-1837
      OpenUrl
    28. ↵
      1. Wilkerson MD,
      2. Hayes DN
      . ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics 2010;26:1572–3. doi:10.1093/bioinformatics/btq170
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Huang Q,
      2. Wu X,
      3. Wang Z, et al
      . The primordial differentiation of tumor-specific memory CD8+ T cells as bona fide responders to PD-1/PD-L1 blockade in draining lymph nodes. Cell 2022;185:4049–66. doi:10.1016/j.cell.2022.09.020
      OpenUrl
    30. ↵
      1. Charoentong P,
      2. Finotello F,
      3. Angelova M, et al
      . Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell Rep 2017;18:248–62. doi:10.1016/j.celrep.2016.12.019
      OpenUrlCrossRefPubMed
    31. ↵
      1. Topp BG,
      2. Channavazzala M,
      3. Mayawala K, et al
      . Tumor dynamics in patients with solid tumors treated with pembrolizumab beyond disease progression. Cancer Cell 2023;41:1680–8. doi:10.1016/j.ccell.2023.08.004
      OpenUrl
    32. ↵
      1. Ratajczak K,
      2. Grel H,
      3. Olejnik P, et al
      . Current progress, strategy, and prospects of PD-1/PDL-1 immune checkpoint biosensing platforms for cancer diagnostics, therapy monitoring, and drug screening. Biosensors and Bioelectronics 2023;240:115644. doi:10.1016/j.bios.2023.115644
      OpenUrl
    33. ↵
      1. Gurbatri CR,
      2. Arpaia N,
      3. Danino T
      . Engineering bacteria as interactive cancer therapies. Science 2022;378:858–64. doi:10.1126/science.add9667
      OpenUrl
    34. ↵
      1. Lee KC,
      2. Zheng L-M,
      3. Luo X, et al
      . Comparative Evaluation of the Acute Toxic Effects in Monkeys, Pigs and Mice of a Genetically Engineered Salmonella Strain (VNP20009) Being Developed as an Antitumor Agent . Int J Toxicol 2000;19:19–25. doi:10.1080/109158100225006
      OpenUrlCrossRef
    35. ↵
      1. Liang K,
      2. Liu Q,
      3. Li P, et al
      . Genetically engineered Salmonella Typhimurium: Recent advances in cancer therapy. Cancer Letters 2019;448:168–81. doi:10.1016/j.canlet.2019.01.037
      OpenUrl
    36. ↵
      1. Cunningham C,
      2. Nemunaitis J
      . A phase I trial of genetically modified Salmonella typhimurium expressing cytosine deaminase (TAPET-CD, VNP20029) administered by intratumoral injection in combination with 5-fluorocytosine for patients with advanced or metastatic cancer. Protocol no: CL-017. Version: April 9, 2001. Hum Gene Ther 2001;12:1594–6.
      OpenUrlPubMed
    37. ↵
      1. Zhou S,
      2. Zhao Z,
      3. Lin Y, et al
      . Suppression of pancreatic ductal adenocarcinoma growth by intratumoral delivery of attenuated Salmonella typhimurium using a dual fluorescent live tracking system. Cancer Biol Ther 2016;17:732–40. doi:10.1080/15384047.2016.1177683
      OpenUrl
    38. ↵
      1. Chen W,
      2. Zhu Y,
      3. Zhang Z, et al
      . Advances in Salmonella Typhimurium-based drug delivery system for cancer therapy. Adv Drug Deliv Rev 2022;185:114295. doi:10.1016/j.addr.2022.114295
      OpenUrl
    39. ↵
      1. Pandit M,
      2. Kil Y-S,
      3. Ahn J-H, et al
      . Methionine consumption by cancer cells drives a progressive upregulation of PD-1 expression in CD4 T cells. Nat Commun 2023;14:2593. doi:10.1038/s41467-023-38316-9
    40. ↵
      1. Roy DG,
      2. Chen J,
      3. Mamane V, et al
      . Methionine Metabolism Shapes T Helper Cell Responses through Regulation of Epigenetic Reprogramming. Cell Metab 2020;31:250–66. doi:10.1016/j.cmet.2020.01.006
      OpenUrlCrossRefPubMed
    41. ↵
      1. Ji M,
      2. Xu X,
      3. Xu Q, et al
      . Methionine restriction-induced sulfur deficiency impairs antitumour immunity partially through gut microbiota. Nat Metab 2023;5:1526–43. doi:10.1038/s42255-023-00854-3
      OpenUrl
    42. ↵
      1. Sun Q,
      2. Hong Z,
      3. Zhang C, et al
      . Immune checkpoint therapy for solid tumours: clinical dilemmas and future trends. Signal Transduct Target Ther 2023;8:320. doi:10.1038/s41392-023-01522-4
    43. ↵
      1. Cancer Genome Atlas Network
      . Genomic Classification of Cutaneous Melanoma. Cell 2015;161:1681–96. doi:10.1016/j.cell.2015.05.044pmid:26091043
      OpenUrlCrossRefPubMed
    44. ↵
      1. Prat A,
      2. Navarro A,
      3. Paré L, et al
      . Immune-Related Gene Expression Profiling After PD-1 Blockade in Non-Small Cell Lung Carcinoma, Head and Neck Squamous Cell Carcinoma, and Melanoma. Cancer Res 2017;77:3540–50. doi:10.1158/0008-5472.CAN-16-3556pmid:28487385
      OpenUrlAbstract/FREE Full Text
    45. ↵
      1. Aran D,
      2. Hu Z,
      3. Butte AJ
      . xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol 2017;18:220. doi:10.1186/s13059-017-1349-1pmid:29141660
      OpenUrlCrossRefPubMed
    46. ↵
      1. Yoshihara K,
      2. Shahmoradgoli M,
      3. Martínez E, et al
      . Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 2013;4:2612. doi:10.1038/ncomms3612pmid:24113773
      OpenUrlCrossRefPubMed
    47. ↵
      1. Becht E,
      2. Giraldo NA,
      3. Lacroix L, et al
      . Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol 2016;17:218. doi:10.1186/s13059-016-1070-5pmid:27765066
      OpenUrlCrossRefPubMed
    48. ↵
      1. Yu G,
      2. Wang L-G,
      3. Han Y, et al
      . clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012;16:284–7. doi:10.1089/omi.2011.0118pmid:22455463
      OpenUrlCrossRefPubMedWeb of Science

    Supplementary materials

    Footnotes

    • Contributors SZ, ZZ: Conception, data analysis, methodology, validation, wrote the manuscript. SZ: Data collection and database organization. AZ, FL: Manuscript supervision. KZ, ZL, EH, JM: Methodology and validation. ZZ, FL: Guarantors.

    • Funding This work was supported by grants from the Guangdong Innovative Research Team Program (2016ZT06Y432 to AZ, FL, and SZ); Key research and development program of Guangdong Province for "Innovative drug creation" (2019B020201015 to FL); Guangzhou Technology Program of Agriculture and Social Development (2023B03J1291 to AZ); “Guangzhou Sinogen Pharmaceutical Co., Ltd” foundation (21HK0399 to SZ, 21HK0398 to ZZ).

    • Competing interests AZ is a shareholder of Guangzhou Sinogen Pharmaceutical.

    • 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.