
==== Front
Sci Rep
Sci Rep
Scientific Reports
2045-2322
Nature Publishing Group UK London

39232081
71421
10.1038/s41598-024-71421-3
Article
TIMM9 as a prognostic biomarker in multiple cancers and its associated biological processes
Zhang Lisheng
Huang Yan
Yang Yanting
Liao Birong
Hou Congyan
Wang Yiqi
Qin Huaiyu
Zeng Huixiang
He Yanli blhhh@gzucm.edu.cn

Gu Jiangyong gujy@gzucm.edu.cn

Zhang Ren zren@gzucm.edu.cn

https://ror.org/03qb7bg95 grid.411866.c 0000 0000 8848 7685 School of Basic Medical Sciences, Guangzhou University of Chinese Medicine, 232 Outer Ring East Road, Guangzhou University City, Panyu District, Guangzhou, 510006 Guangdong China
4 9 2024
4 9 2024
2024
14 2056811 4 2024
28 8 2024
© The Author(s) 2024
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
TIMM9 has been identified as a mediator of essential functions in mitochondria, but its association with pan-cancer is poorly understood. We herein employed bioinformatics, computational chemistry techniques and experiments to investigate the role of TIMM9 in pan-cancer. Our analysis revealed that overexpression of TIMM9 was significantly associated with tumorigenesis, pathological stage progression, and metastasis. Missense mutations (particularly the S49L variant), copy number variations (CNV) and methylation alterations in TIMM9 were found to be associated with poor cancer prognosis. Moreover, TIMM9 was positively related with cell cycle progression, mitochondrial and ribosomal function, oxidative phosphorylation, TCA cycle activity, innate and adaptive immunity. Additionally, we discovered that TIMM9 could be regulated by cancer-associated signaling pathways, such as the mTOR pathway. Using molecular simulations, we identified ITFG1 as the protein that has the strongest physical association with TIMM9, which show a promising structural complement.

Keywords

Pan-cancer
Mitochondria
Metabolism reprogramming
Oxidative phosphorylation
Drug resistance
Subject terms

Cancer
Computational biology and bioinformatics
National Undergraduate Training Program for Innovation and Entrepreneurship202210572003 Natural Science Foundation of Guangdong Province, ChinaNo. 2022A1515011575 The National Natural Science Foundation of China81873154 issue-copyright-statement© Springer Nature Limited 2024
==== Body
pmcIntroduction

Tumorigenesis has been demonstrated to be a complicated, metabolism-related process, including abnormalities in glycolysis and oxidative phosphorylation, which reshape the tumor microenvironment (TME)1. Mitochondria have been identified as mediators of essential metabolic and immunologic processes2,3, so investigating the association between mitochondria and tumorigenesis is likely to be meaningful.

TIMM9 is a molecular chaperone localized in mitochondrial intermembrane space (IMS)4. It plays an essential role in the transport of proteins that are destined for the mitochondrial inner membrane5. As a member of small TIM family6. TIMM9 has been shown to interact with TIMM107. Further studies have demonstrated that an arrangement of 3 TIMM9 and 3 TIMM10 molecules forms a heterohexamer8, which mediates the substrate specificity of the TIM22 mitochondrial import pathway9. Reportedly, overexpression of TIMM9 is associated with a negative prognosis in gastric cancer10. TIMM9 is among the genes involved in oxidative phosphorylation that have been used as signatures for predicting prognosis11. Additionally, TIMM9 is selectively localized in breast neoplasm rather than peripheral tissues12. This indicates an association between TIMM9 and tumorigenesis. However, it is necessary to conduct a comprehensive investigation of the expression variations of TIMM9 and their relationship with prognosis in pan-cancer, at various detection levels.

In this study, we aimed to investigate the expression levels of TIMM9 in various types of cancer, and its correlations with prognosis, tumorigenesis, metastasis, proliferation, differentiation, tumor microenvironment (TME), drug resistance and other phenotypes. Metabolic reprogramming, including alterations in oxidative phosphorylation and glycolysis, has been extensively studied in cancer. In addition, we have analyzed TIMM9-associated signaling pathways as well as possible TIMM9 regulation strategies.

Results

TIMM9 expression is significantly elevated in cancerous tissues

We used TIMER 2.0 to investigate the differential RNA expression of TIMM9 between cancer tissues and normal tissues. Our analysis revealed significant overexpression of TIMM9 in 13 different types of cancers (Fig. 1A), including CHOL, COAD, ESCA, GBM, HNSC, KICH, KIRC, LIHC, LUAD, LUSC, PRAD, READ and STAD. However, TIMM9 was significantly down-regulated in BRCA, KICH, THCA, and UCEC. The box plot generated by TNMplot illustrates TIMM9 expression levels in various normal and tumor tissues across multiple cancer types. TIMM9 expression is consistently upregulated in various tumors compared to corresponding normal tissues, as analyzed by the TNMplot database (Fig. S1). Experimental data from RT-qPCR trials further demonstrated that TIMM9 expression in tumor cell lines was higher than in corresponding normal epithelial cells in lung cancer and liver cancer (Fig. 1B), indicating that the overexpression of TIMM9 is a common phenomenon in tumorigenesis. Using the GSCA database, we found that TIMM9 was upregulated in specific pathological stages of cancer, such as between Stage I and Stage IV in HNSC, Stage I and Stage II in LIHC, Stage I and Stage III in LUSC, and Stage II and Stage III in LUSC (Fig. 1C). Compared to primary tumors, TNMplot database analysis revealed that TIMM9 was overexpressed in metastatic states of colon cancer, liver cancer, lung cancer and skin cancer (Fig. 1D).Fig. 1 TIMM9 is overexpressed in various types of cancer. (A) Differential RNA expressions of TIMM9 in TCGA, analyzed by TIMER 2.0 database. (B) RT-qPCR assays detected differential expressions of TIMM9 between tumor cell lines (lung: PC9, NCI-H460, NCI-H1299; liver: HepG2) and normal epithelial cells (lung: BEAS-2B, liver: LO2). (C) RNA expression levels of TIMM9 in pathological stages. (D) Differential RNA expressions of TIMM9 among normal, tumor and metastatic groups. (E) Differential expressions of TIMM9 in proteomics levels. *, **, ***, **** correspond to p < 0.05, p < 0.01, p < 0.001, and p < 0.0001 respectively.

The DepMap Portal analysis elucidated a significant positive correlation between TIMM9 gene expression and protein levels (p = 3.08E−8, Fig. S13B). TIMM9 was found to be upregulated at the protein level in colon cancer, ovarian cancer, UCEC, lung adenocarcinoma, lung squamous cell carcinoma, and hepatocellular carcinoma, while it was downregulated in breast cancer, clear cell renal cell carcinoma, glioblastoma multiforme, head and neck squamous carcinoma, and pancreatic adenocarcinoma (Fig. 1E and Fig. S2). Immunohistochemical analysis from the HPA database revealed that TIMM9 protein levels are significantly upregulated in various cancers, including lung cancer, liver cancer, colorectal cancer, ovarian cancer, prostate cancer, and endometrial cancer (Figs. S3, S4).

TIMM9 is overexpressed in cancer cells within tumor tissues

Diagrams were obtained from the TISIDB database to visualize the expression localization of TIMM9 (Fig. 2A). TIMM9 showed a higher expression level in malignant cell clusters compared to other adjacent cell clusters in CESC, KIRC, CHOL, LIHC, ESCA, NSCLC, LSCC and PRAD. Additionally, cell differentiation trajectories in the CellTracer database demonstrated that TIMM9 maintained a high expression level in pseudo-times of malignant cells (Fig. 2B), Indicating that the overexpression of TIMM9 may influence tumor differentiation and contribute to the formation of tumor heterogeneity across various cancer types.Fig. 2 Differential expressions of TIMM9 at single-cell transcriptome levels. (A) The expression levels of TIMM9 among multiple types of cells in TME. (B) The expression levels of TIMM9 in pseudo-time trajectories of tumor cells differentiations. TME tumor microenvironment.

High expression of TIMM9 indicates poor cancer prognosis

We used the GEPIA 2 database to investigate the correlation between TIMM9 expression and clinical outcomes by plotting Kaplan–Meier plots. Our analysis revealed a significant association between high expression of TIMM9 and poor overall survival (OS) in ACC, BLCA, HNSC, KICH, LIHC and LUAD and disease-free survival (PFS) in ACC, BLCA, KICH, KIRP, LIHC and LUAD (Fig. 3A,B). Interestingly, high TIMM9 expression was unexpectedly associated with a more favorable prognosis in the OS of KIRC. This trend, although not statistically significant, was similarly observed in the GSCA database (Table S1), indicating that the relationship between TIMM9 expression and prognosis is more complex and may vary depending on the cancer type. We then used the OncoLnc database to confirm these findings by generating Kaplan–Meier plots (Fig. 3C), which showed a relationship between high expression of TIMM9 and poor OS prognosis in BLCA, CHOL, EAC, HNSCC and LIHC. Altogether, in most cases, high expression of TIMM9 is associated with poor prognosis outcomes.Fig. 3 Correlation analyses between expression of TIMM9 and prognosis of tumors. (A,B) Overall survival (A) and disease-free survival (B) were analyzed by GEPIA 2 database. (C) Overall survival of various cancers was analyzed by LOGpc database.

TIMM9 exhibits genomic instabilities in tumors

The cBioPortal database analysis revealed that TIMM9 has genomic alterations in the majority of types of cancers (Fig. 4A). We analyzed the mutation landscape (Fig. 4C) and collected data on missense mutations (Table S2). The structure of the TIMM9–TIMM10 complex was obtained from the RCSB PDB database (pdb code: 7cgp, chain D: TIMM9, chain H: TIMM10). We performed in silico site-directed mutagenesis simulations using PyMOL 2.4.0 software (Fig. S5A) to obtain the structures after missense mutations. We calculated the binding energies of TIMM9 and TIMM10 using the PDBePISA platform and found that the binding energy of S49L was largely decreased (Table S2), indicating an improvement in binding capability (interface of TIMM9 with SER-49: − 20.0 kcal/mol; interface of TIMM9 with LEU-49: − 21.1 kcal/mol). S49L, a missense mutation, improves binding capability between TIMM9 and TIMM10. The 3D (Fig. 4B) and 2D (Fig. 4D) structures of WT and S49L were visualized. SER-49 only formed a van der Waals interaction with MET-65 of TIMM10, whereas LEU-49 formed one with ILE-61 and MET-65 of TIMM10. Independent Gradient Modeling (IGM) also confirmed that LEU-49 can form a stronger interaction with TIMM10 (Fig. 4E), indicating S49L may positively affects TIMM9-associated functions by forming a hydrophobic nucleus.Fig. 4 TIMM9 exhibited genomic instabilities in tumors. (A) The visualization of genomic alterations in TIMM9. Red, blue, and green represent amplification, deep deletion, and mutation, respectively. (B) The site-directed mutagenesis simulation and structural exhibition of S49L missense mutation, in TIMM9. The original structure was obtained from PDB database (PDB code: 7cgp, green: TIMM9, cyan: TIMM10). (C) Landscape of concrete mutation sites in TIMM9 (green dots: missense mutations, brown dots: splice mutations). (D) 2-D visualization of molecular micro-environment of S49L. (E) Weak interactions around 49th residue of TIMM9 were analyzed using IGM method, the strength of interaction increases from green to blue. (F) CNV landscape of TIMM9 (Hete heterozygous, Homo homozygous, Amp amplification, Del deletion). (G) Correlation between CNV and expression of TIMM9 and TIMM10 in tumors. (H) Association between differentiation in CNV level and prognosis. (I) Differential methylation of TIMM9. **, ***, correspond to p < 0.01, p < 0.001, respectively. (J) Correlation between methylation and expression of TIMM9 and TIMM10 in tumors. (K) Association between methylation and prognosis. CNV copy number variation.

A global exhibition of copy number variations (CNV) of TIMM9 was visualized by a pie chart (Fig. 4F). High CNV levels of TIMM9 and TIMM10 are associated with their high expression (Fig. 4G). Kaplan–Meier plots revealed that high CNV (amplification or deletion) of TIMM9 is significantly associated with poor prognosis in OS of KIRC, KIRP, LGG, MESO, PCPG, UCEC, and UCS (Fig. 4H), PFS of KIRC, KIRP, LGG, LIHC, MESO, READ, and UCEC (Fig. S5D), and DSS of KIRC, KIRP, LGG, MESO, and UCEC (Fig. S5E).

The methylation of TIMM9 was also investigated and the methylation level of TIMM9 is significantly decreased in KIRP, LIHC and PRAD (Fig. 4I). Coherently, the SMART database analysis showed that TIMM9 methylation (CpG site: cg16020706) is significantly lower in tumor tissues compared to normal tissues in BLCA, BRCA, ESCA, HNSC, KIRC, KIRP, LIHC, LUSC, PAAD, PRAD, SARC, and UCEC (Fig. S6). Methylation levels of TIMM9 and TIMM10 are negatively correlated with their expression (Fig. 4J). We also found that low TIMM9 methylations is associated with poor prognosis in OS of BLCA, GBM, LAML, and LUAD (Fig. 4K), PFS of GBM, KIRP, PRAD and UCS (Fig. S5B), and DSS of BLCA and GBM (Fig. S5C).

To summarize, the genomic instabilities of TIMM9, including missense mutations, CNV, and methylation, could serve as biomarkers for oncogenesis and tumor prognosis.

TIMM9 is strongly related with cancer-associated biological functions

GSEA enrichments were used to gain insight into biological functions associated with TIMM9. Critical signaling pathways, including “Cell cycle”, “DNA replication”, “Oxidative Phosphorylation”, “Nucleotide excision repair” and “PI3K-Akt signaling pathway” were enriched in cancers (Fig. 5A). TIMM9 expression was positively associated with “Cell Cycle”, “Cell Division”,and “DNA repair”, but was negatively associated with “Innate Immune Response” and “T Cell Mediated Immunity” across multiple cancer types in ssGSEA analysis (Fig. 5B). The heat maps generated by the CellTracer database showed that TIMM9 expression levels were strongly correlated with “Cell Cycle”, “DNA Damage”, “DNA Repair” and “invasion”, all along pseudo-times in single-cell sequencing (Fig. S7A). Scatter plots generated by the CellTracer database provided consistent results, showing that TIMM9 expression levels were positively correlated with “Cell Cycle”, “DNA Repair” and “Proliferation” phenotypes (Fig. S7B–D). DepMap Portal analysis demonstrated that a consistent negative correlation exists between TIMM9 knockout and cell survival across various cell lines, including melanoma, non-small cell lung cancer, hepatocellular carcinoma, and ovarian epithelial tumors (Table S3). The negative gene effect scores indicate that the absence of TIMM9 significantly impairs cell viability, suggesting its critical role in cancer cell survival. Similarly, mTOR knockout also impairs cell survival, indicating its essential role in supporting cancer cell viability (Fig. S13A). More detailed GSEA visualizations correlated TIMM9 with “Cell cycle”, “DNA replication”, “Ribosome”, “Oxidative phosphorylation”, “Respiratory electron transport”, “The citric acid cycle”, and “Metabolism of xenobiotics by cytochrome P450” (Figs. S8, S9), indicating that TIMM9 is positively associated with cell proliferation, ribosomes, and mitochondria. ssGSEA analyses clarified that TIMM9 mediates variations in glycolysis, oxidative phosphorylation, the TCA cycle and respiratory chain complex (Fig. S10). Enrichment analysis of TIMM9 expression-related genes from the DepMap Portal reveals significant involvement in various biological processes. Key processes include cell migration, cell–cell junction assembly, cell adhesion mediated by cadherin, DNA damage response, DNA repair, cell motility, angiogenesis, the G1/S transition of the mitotic cell cycle, and positive regulation of the G2/M transition of the mitotic cell cycle (Fig. S14, Table S4).Fig. 5 Associations between phenotypes and mechanisms of cancers. (A) GSEA enrichments indicate association between TIMM9 and signaling pathways. (B) Associations between differential expression of TIMM9 and cancer-associated phenotypes. (C) RT-qPCR analysis of TIMM9 differential expression with or without 10 µM of OSI-027 (mTOR inhibitor). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

The ssGSEA analysis indicates that TIMM9 expression levels are significantly upregulated in various cancers when the mTOR signaling pathway is elevated (Fig. S11). Specifically, higher TIMM9 expression was observed in BRCA, CRC, KICH, KIRC, LGG, LUAD, LUSC, NSCLC, and SKCM cohorts with elevated mTOR pathway activity. This suggests a strong correlation between TIMM9 expression and mTOR pathway activation across multiple cancer types. The UALCAN analysis reveals an interesting trend: TIMM9 expression levels in the mTOR Pathway-altered group consistently show greater deviation from the Normal group compared to the Others group (Fig. S12). This suggests that changes in TIMM9 expression are more pronounced with alterations in the mTOR pathway. The DepMap Portal analysis demonstrated a significant positive correlation between TIMM9 and mTOR in cell activity function. Our RT-qPCR assays demonstrated that OSI-027, an mTOR inhibitor, regulates TIMM9 expression in NCI-H1299 and NCI-H460 cell lines, suggesting that TIMM9 may support the findings from the above research (Fig. 5C).

Overall, TIMM9 can be considered a biomarker for cancer-related biological functions and signaling pathway alterations.

TIMM9 is associated with tumor microenvironment (TME) variation

TIMM9 was associated with variations in immune cell clusters (Fig. 6A), including BRCA, KIRC, KIRP, LIHC, LUAD, LUSC, PRAD and UCEC (p < 0.001). Cells with a lymphocyte-depleted phenotype had higher expression of TIMM9 in BRCA, KIRC, KIRP, LIHC, PRAD and UCEC (Fig. 6B), suggesting a possible relationship between TIMM9 and lymphocytes exhaustion. TIMM9 expression was negatively correlated with infiltration levels of CD4+ effector memory cells (Fig. S15), but positively correlated with Macrophage M2 cells and Myeloid-derived suppressor cells (MDSC) (Fig. S16). Additionally, TIMM9 expression showed extensively negative correlations with immunostimulators at the expression level (Fig. 6C). TIMM9 expression was positively associated with immune-related genomic scores, such as MANTIS, Tumor Mutation Burden (TMB) and neoantigen loads (Fig. 6D–F).Fig. 6 Analysis of TIMM9 participating in immune process. (A) Associations between TIMM9 expression and immune subtypes across various cancers. (B) Association between expression variation of TIMM9 and subtypes of cancers (C1 = wound healing; C2 = IFN-gamma dominant; C3 = inflammatory; C4 = lymphocyte depleted; C5 = immunologically quiet; C6 = TGF-b dominant). (C) Association between expression of TIMM9 and immunostimulators. (D–F) Relation between TMB (D), MANTIS (E) and Neoantigen Loads (F). *, **, *** represent p < 0.05, p < 0.01, and p < 0.001, respectively.

TIMM9 expression in KIRC shows a distinctive correlation with immune cells. It is positively associated with immune-activating cells like NK, NKT, and MAIT, and negatively associated with immunosuppressive cells such as iTreg, nTreg, and Tr1 (Fig. S17). This pattern contrasts with other cancers, where TIMM9 does not show a similar correlation with immune activation and suppression. This unique interaction may explain the better prognosis seen in KIRC with high TIMM9 expression, differing from its role in other cancers like LUAD.

Despite this, TIMM9 could also serve as a biomarker for good prognosis in overall survival (OS) and progression-free survival (PFS) of immunotherapies (Fig. S18A). TIMM9 expression levels were upregulated after the treatment with IFNγ, IFNβ, or TNFα in multiple types of mouse cancer cell lines (Fig. S18B), and upregulations of TIMM9 was associated with good responses to immunotherapies (Fig. S18C). The DRMref database analysis demonstrated that TIMM9 is upregulated after immunotherapy in Acute Lymphoblastic Leukemia (Table S5). Similarly, the ICBatlas database analysis revealed that TIMM9 expression is notably upregulated following anti-PD1 treatment in various cancers, particularly melanoma (Table S6). Additionally, there is a consistent trend of increased TIMM9 expression in several cancer types, although not all changes reached statistical significance. Furthermore, the TISIDB database analysis showed that TIMM9 expression is significantly higher in responders compared to non-responders in melanoma (Table S7).

Altogether, TIMM9 is associated with variations in the TME and oncogenesis, and may could involve in associated biofunctions. TIMM9 could be seem as a biomarker in TME and immunotherapies.

TIMM9 affects drug sensitivities of medical treatments

Correlations between TIMM9 and drug sensitivities of chemotherapies or targeted therapies were calculated using the RNAactDrug platform. We found that high CNV of TIMM9 was positively correlated with drug sensitivities to Lestaurtinib (FDR = 5.93E−07), NG-25 (FDR = 4.90E−09), Nutlin-3a (−) (FDR = 1.08E−07), NVP-BHG712 (FDR = 6.29E−08), and TL-1-85 (FDR = 5.68E−09) (Fig. 7A), while the expression level of TIMM9 was negatively correlated with drug sensitivities to 5-Fluorouracil (5-FU) (FDR = 1.12E−07), and TAK-715 (FDR = 6.81E−09) (Fig. 7B). The methylation level of TIMM9 was positively correlated with drug sensitivities to Dabrafenib (FDR = 5.35E−09), PLX-4720 (FDR = 4.57E−12) and SB590885 (FDR = 3.79E−09) (Fig. 7C). Prognosis analyses showed positive associations between high expression of TIMM9 and poor outcomes in Gastric cancer treatments with 5-Fluorouracil, but with good outcomes in Ovarian cancer treatments with Platin (Fig. 7D). ROC diagrams showed that TIMM9 is related to pathological response, and boxplots exhibited an association between high TIMM9 expression and non-response to treatments (Fig. 7E). Additionally, GSCA platform analysis showed that both TIMM9 and TIMM10 were negatively correlated with drug sensitivities in most therapies (Fig. 7F). Our CCK8 assay also demonstrated that NCI-H1299 (highly expressed TIMM9) was insensitive to 5-FU compared to other cell lines with low TIMM9 expression (NCI-H460 and PC9) (Fig. 7G).Fig. 7 Association between TIMM9 expression and therapeutic responses. (A–C) Association between between drug sensitivity and CNV (A), expression (B) and methylation (C). (D) Association between TIMM9 expression level and patients’ survival after chemotherapies. (E) ROC curves and boxplots indicating the expression of TIMM9 is associated with responses of chemotherapies. (F) Correlation between CTRP drug sensitivity and the mRNA expression variation of TIMM9 as well as TIMM10. (G) CCK8 experiment of cell activity after adding 5-FU (*** means the cell viability of NCI-H1299 is significantly higher than PC9 (p < 0.001), ### means the cell viability of NCI-H1299 is significantly higher than NCI-H460 (p < 0.001)). CNV copy number variations, ROC receiver operating characteristic.

Lestaurtinib, NG-25, Nutlin-3a (−), NVP-BHG712, and TL-1-85 were selected as candidates to target the TIMM9–TIMM10 complex since their drug sensitivity was associated with high CNV in TIMM9. The druggability of the TIMM9–TIMM10 complex was evaluated using CavityPlus software, which identified two significant druggable pockets at the binding interface between TIMM9 and TIMM10 (Fig. S19, Table S8). The pocket with the highest rank of druggability was used for molecular simulations. Molecular docking was performed using Autodock Vina 1.1.2 and NVP-BHG712 showed the best docking affinity of − 11.2 kcal/mol (Fig. S20A, Table S9). The complex of TIMM9–TIMM10 binding with NVP-BHG712 was subjected to 50 ns molecular dynamics simulation using Gromacs 2020.4 software. The RMSD curve demonstrated the stability of the complex (Fig. S20C, D). The conformations of the ligand showed only minor differences (Fig. S20B). B-factor analysis showed that the binding site of the protein was stable (Fig. S20G). A 2-D diagram and NCI analysis of the last frame (50 ns) clarified the crucial role of van der Waals interactions (Fig. S20 F, H), but not hydrogen bonds (2.017982E−01 ± 1.438853E−02 on average, Fig. S20E).

Analysis using the CREAMMIST database identified several drugs with significant Spearman correlations between TIMM9 expression and IC50 values across different cancer cell lines (Table S10). Notably, Z-LLNle-CHO exhibited the highest correlation (r = 0.20009, p = 8.59E−05) across 380 cell lines, indicating a potential association with TIMM9 activity. Other drugs such as alpha-cyano-4-hydroxycinnamic acid and oxythiamine also showed significant correlations. These findings suggest that higher TIMM9 expression may be associated with reduced therapeutic response to these drugs, as indicated by the increased IC50 values.

TIMM9 mediates cancer-associated gene networks

GeneMANIA and BioGRID databases were used to define the gene networks mediated by TIMM9 (Fig. 8A, Table S11). The Metascape database identified two gene networks associated with oxidative phosphorylation and the regulation of T cell differentiation (Fig. 8B). GO enrichments of the two gene sets indicated a strong relationship between the gene sets and mitochondria-associated biological functions (Fig. 8C,D). Multiple enrichments by databases comprehensively demonstrated that the gene set obtained from the BioGRID database is correlated with chemical carcinogenesis, oxidative phosphorylation, ATP synthesis, respiratory electron transport and the TCA cycle (Fig. 8E). All these results revealed that the gene sets could admirably explain the biological function of TIMM9.Fig. 8 Comprehensive insight into the signaling pathways and networks mediated by TIMM9. (A) Analysis of gene networks mediated by GeneMANIA and BioGRID (yellow lines: physical edges, green lines: genetic edges, purple lines: physical/genetic edges) databases. (B) Metascape enrichment of genes in networks. (C,D) GO enrichment of genes in networks of GeneMANIA (C) and BioGRID (BP biological process, CC cellular component, MF molecular function) (D). (E) KEGG, INTERPRO, and REACTOME enrichment analyses of genes in BioGRID.

The Metascape database was utilized to identify hub genes in the gene set (BioGRID database), which physically interacted with TIMM9 (Fig. 9A). Protein–protein docking by the ZDOCK Server was utilized to evaluate binding affinities between these proteins and TIMM9 (Fig. S21A), and the TIMM9-ITFG1 complex achieved the lowest binding energy (Table S12). A 2-D diagram was plotted to visualize the specific residues in the interface (Fig. 9C). 35 ns molecular dynamics by Gromacs 2020.4 software were utilized to further expound the details of the interaction between TIMM9 and ITFG1. The RMSD (Fig. S21B) and Rg (Radius of gyration, Fig. S21C) consistently indicated that the trajectory of the simulation undoubtedly converged from 25 to 35 ns, and the secondary structure remained stable (Fig. S21D). Surprisingly, the average number of total weak interactions within 0.35 nm was up to 38.22517 pairs, and the average number of hydrogen bonds was 8.91578 (Fig. 9B). The Gibbs Energy Landscape was utilized to further hunt for the most stable conformation of the TIMM9-ITFG1 complex (Fig. S21E). The result of IGM further demonstrated that weak interactions broadly formed between TIMM9 and ITFG1 in the conformation with the lowest Gibbs Energy (Fig. 9D), indicating promising structural matching. PCA analysis was used to identify functional movements, and allosteric changes were found in the first and second principal components (PC1, PC2) (Fig. 9E). We preliminary found an “open-close” synergistical movement of ITFG1 in residues near the N-terminal of TIMM9 (Fig. 9F).Fig. 9 Protein–protein interaction networks mediated by TIMM9. (A) Hub genes selected by Metascape from gene set obtained from BioGRID. (B) Hydrogen Bonds curve plot after MD simulation of TIMM9-ITFG1 complex. (C) The 2-D intermolecular interactions analysis of last frame structure (35 ns) in MD trajectory (Chain A: TIMM9, Chain B: ITFG1, green lines: hydrogen bonds; residues except for which forming hydrogen bonds formed van der Waals interactions). (D) IGM analysis to determine the strength of interactions (the strength becomes stronger from green to blue, green usually represents van der Waals interactions and pi-pi stacking, blue usually represents hydrogen bonds, red represents steric hindrances). (E) PCA analysis of TIMM9-ITFG1 complex. (F) Structural analysis of movement in first principal component (PC1). RMSD root mean square deviation, Rg radius of gyration, MD molecular dynamics, IGM independent gradient model.

Comprehensive analysis of TIMM9 expression and its implications in lung adenocarcinoma (LUAD)

The LUAD expression matrix from the TCGA dataset was obtained since LUAD has been shown to correlate with differential expression (Fig. 1), poor prognosis (Fig. 3), TIMM9-related phenotypes (Fig. 5), and expression data (Fig. 6A,B). Twenty tumor samples with the highest or lowest expression levels of TIMM9 were obtained separately, totaling 40 samples. The R package DESeq2 was used to obtain differential expression genes between the “low expression of TIMM9 group” and the “high expression of TIMM9 group”. These differential expression genes were utilized to perform WGCNA analysis. We set the soft threshold to 3 with R2 > 0.85 and high average connectivity (Fig. 10A). Genes were separated into nine modules (Fig. 10B). TIMM9 was positively correlated with the “MEblue” (r = 0.902925, p = 1.64E−15), “MEyellow” (r = 0.755647, p = 1.74E−08), and "MEgreen" (r = 0.680308665, p = 1.37E−06) modules; negatively correlated with the "MEturquoise" (r = − 0.587936468, p = 6.63E−05), "MEpink" (r = − 0.533037917, p = 0.000398256), "MEred" (r = − 0.5248887, p = 0.000506638), "MEbrown" (r = − 0.485311155, p = 0.001502772), and "MEblack" (r = − 0.47096988, p = 0.002160406) modules; and not correlated with the "MEmagenta" (r = − 0.119299994, p = 0.463425268) and "MEgrey" (r = 0.111817245, p = 0.49212794) modules. Visualizations of the relationships among each module (Fig. 10C) and each gene (Fig. 10D) were performed, and the correlation between the expression of TIMM9 and each gene module was visualized as a heatmap (Fig. 10E). Modules positively associated with TIMM9 expression ("MEblue", "MEyellow", "MEgreen") were correlated with different biological processes, such as "cell adhesion", "ribosome", and "cell cycle" (Fig. 10F–H). These results may explain the phenotypes mediated by TIMM9.Fig. 10 WGCNA gene co-expression network. (A) Relationship of soft threshold and scale-free topology model fit and mean connectivity. (B) Gene classification into multiple colors indicating different modules in the clustering tree. (C) Heat map of module feature genes. Red color means a high correlation, and blue color indicates a low correlation. (D) Clustering dendrogram of module feature genes. Light-colored areas show a strong correlation. (E) Correlations between each module and expression variation of TIMM9. The left triangle presents the correlation between TIMM9 and different modules. Red shows a high correlation. The right triangle indicates the p value, *p < 0.05, **p < 0.01, ***p < 0.001. (F–H) GO enrichment of each module, “MEblue” (F), “MEyellow” (G), and “MEgreen” (H), separately.

PCA analysis of TIMM9-related genes significantly positively correlated with TIMM9 expression in LUAD (from GEPIA, Table S13) reveals distinct clustering of tumor and normal samples (Fig. 11A,B). GSVA scores are higher in tumor tissues compared to normal tissues (Fig. 11C). Higher GSVA scores are associated with poorer overall survival (OS), progression-free survival (PFS), and disease-specific survival (DSS) in LUAD patients (Fig. 11D–F). Additionally, GSVA scores show a positive correlation with cell cycle pathway activity (Fig. 11G) and EMT pathway activity (Fig. 11H).Fig. 11 GSCA database analysis of TIMM9-related gene set from the GEPIA database. (A) PCA analysis results executed by GEPIA, showing the variance explained by each principal component. (B) The first two principal components from the PCA dimensionality reduction, distinguishing between tumor (red) and normal (yellow) samples in LUAD. (C) Visualization of GSVA score differences between tumor and normal groups in LUAD. (D) Overall survival (OS) analysis comparing patients with higher and lower GSVA scores in LUAD. (E) Progression-free survival (PFS) analysis for patients with different GSVA scores in LUAD. (F) Disease-specific survival (DSS) analysis for patients with different GSVA scores in LUAD. (G) Spearman correlation between GSVA scores and cell cycle pathway activity in LUAD, showing a positive correlation. (H) Spearman correlation between GSVA scores and EMT pathway activity in LUAD, showing a positive correlation. The gene set used in these analyses was obtained from the GEPIA database and includes genes significantly positively correlated with TIMM9 expression in LUAD.

The GSCA database analysis of the TIMM9-related gene set from the BioGRID database reveals significant findings in LUAD. The GSVA scores, calculated for the gene set significantly positively correlated with TIMM9 expression, are higher in tumor tissues compared to normal tissues (Fig. S22A). Patients with higher GSVA scores exhibit poorer overall survival (OS) and disease-specific survival (DSS) compared to those with lower scores (Fig. S22B,C). Additionally, a positive correlation is observed between GSVA scores and cell cycle pathway activity (Fig. S22D), indicating a potential role of TIMM9 in cell cycle regulation and cancer progression.

In summary, TIMM9 is significantly associated with various phenotypes and survival outcomes in LUAD, highlighting its potential role in cancer biology and as a therapeutic target.

Discussion

An overview of TIMM9

TIMM9, a small protein that mainly consists of alpha-helix motifs and a loop, has been proven to mediate the transport of mitochondrial proteins8. It could bind with TIMM10 to constitute a heterohexameric molecular chaperone13. The TIMM9–TIMM10 complex is the functional form of TIMM913. Multiple diseases are associated with the participation of TIMM99. Several researchers have reported partial association between TIMM9 and tumors12. Nonetheless, our study pioneered the research revealing the association between TIMM9 and prognosis. TIMM9-related functions in oncogenesis, as well as TIMM9-related pan-cancer oncotherapies, were also first reported in our study. A systematic and comprehensive insight has been proposed to elucidate the role of TIMM9 in oncology.

Research logic

TIMM9 is highly expressed in various cancers, prompting an investigation into the regulatory factors such as CNV and methylation, which influence its expression. These genetic and epigenetic alterations suggest TIMM9’s potential as a broad prognostic marker. Given the frequent association of high TIMM9 expression with poor prognosis, we explored its role in malignant phenotypes, including cell proliferation, tumor immunity, and cancer-associated pathways like mTOR. TIMM9 emerged as a potential biomarker across these phenotypes. To further elucidate the pathways linked to TIMM9, we conducted gene set enrichment analysis. Notably, in LUAD, where TIMM9 is overexpressed and correlates with poor prognosis, the association with immune subtypes is the most significant. Thus, we focused our WGCNA and GSVA analyses on LUAD, with results aligning with our prior findings.

TIMM9 is associated with expression, genomic and epigenetic variations in pan-cancer

In our current study, we have validated the overexpression pattern of TIMM9 across various types of cancer. Despite exceptions, such as the significantly downregulated expression of TIMM9 in BRCA and the inconsistency between mRNA and protein expression trends in UCEC, this suggests the possibility of more complex regulatory mechanisms. We also clarified that TIMM9 expression is associated with tumorigenesis, metastasis, and progression in pathological stages. Experimental data first demonstrated that TIMM9 is overexpressed in multiple cancer cell lines, and TIMM9 is expressed highest in NCI-H1299. Data from single-cell RNA sequencing (scRNA-seq) technologies revealed that the expression variation of TIMM9 is located in malignant cells, rather than adjacent cells. This opinion is consistent with previous conclusions by other researchers12. Our research also revealed that TIMM9 maintains high expression during the pseudo-times of cell differentiation trajectories in scRNA-seq. It has been reported that tumor heterogeneity, which is associated with differentiation of tumor cells, causes mortality and drug resistance14. All these pieces of evidence clarify that TIMM9 participates in the emergence, progression, and metastasis of neoplasms. Prognostic evidence consistently showed that high expression of TIMM9 is associated with a poor prognosis in overall survival (OS) and progression free survival (PFS). Particularly, the OS and PFS of ACC, BLCA, KICH, KIRC, and LUAD showed outstanding consistency, indicating that TIMM9 may play a significant role in the development of these cancers, indicating that TIMM9 may play a significant role in the development of these cancers, this suggests it may be involved in oncogenesis. Encouragingly, TIMM9 has been found to be associated with unfavorable prognosis in both hepatocellular carcinoma and gastric cancer10,11, revealing good consistency between our work and that of other researchers. Interestingly, the OS of KIRC showed that high expression of TIMM9 is associated with a positive prognosis in KIRC, hinting that the relationship between TIMM9 and pan-cancer is not simply a one-to-one correspondence, this may due to difference in tumor micro-environment of different cancers. Genomic instability is a common phenotype in cancers. Our research provided a comprehensive insight into the association between genomic variation and TIMM9 in pan-cancer. In silico research of site-mutation simulation clarified that S49L reduces the binding energy between TIMM9 and TIMM10, indicating that S49L may promote the function of TIMM9 by enhancing the binding capability with TIMM10. Excitingly, it has been proved that TIMM9–TIMM10 complex is the functional form of molecular chaperone4. The high CNV and low methylation level of TIMM9 correlated with poor prognosis outcomes. Interestingly, as a collaborator of TIMM9, TIMM10 differential expressions exhibited the same tendency with CNV and methylation level, suggesting that the synergy of TIMM9 and TIMM10 may be not only structural but also genomic. We considered that variation in genomics and epigenetics of TIMM9 may promote oncogenesis by improving its expression level, since methylation is negatively correlated with expression, while CNV is positively correlated with expression in TIMM9. Our work demonstrated that mutagenesis of TIMM9 could be an effective biomarker in prognosis outcome prediction.

TIMM9 is associated with multiple cancer-associated phenotypes and signaling pathways

GSEA and ssGSEA methods have been utilized to clarify tumor-associated phenotypes and mechanisms regulated by TIMM9. Enrichment results of “DNA replication”, “Cell cycle”, “Cell adhesion molecules” inspired us that TIMM9 may be involved in the regulation of cell proliferation and metastasis, since variation in adhesion molecules contributes significantly to metastasis15. High enrichment of oxidative phosphorylation, glycolysis and TCA cycle strongly suggested the association between TIMM9 and energy-associated metabolism and mitochondria. It has been reported that glycolysis is significant in metabolism reprogramming16. Recently, various researchers have focused on positive association between tumorigenesis and oxidative phosphorylation1. All our evidence points to the possibility that TIMM9 mainly had influences on mitochondrial metabolism to mediate oncogenesis. Interestingly, another published research clarified that TIMM9 is one of the biomarkers in oxidative phosphorylation, our research supports and expands this viewpoint11. Surprisingly, both GSEA and WGCNA demonstrated that “ribosome” is significantly enriched in almost all types of cancers, revealing an intense relationship between TIMM9 and ribosomes. Our computational and experimental data further support that TIMM9 is associated with variation in TME and drug resistance. Our gene network and WGCNA analysis successfully explained relevant phenotypes. Overall, our study elucidates the strong association between TIMM9 and cancer-related phenotypes.

Drug resistance of 5-Fluorouracil (5-FU) is a common phenomenon17,18, which leads to mortality in patients. Our results clarified that TIMM9 could be a biomarker in 5-FU drug sensitivity prediction. Higher expression of TIMM9 is usually correlated with poor prognosis outcomes, indicating the significance of TIMM9 in oncotherapy.

Our research into protein-protein interaction network could effectively clarify the function of TIMM9. Interestingly, ITFG1 and TIMM9 show fantastic structural complementary, indicating that the physical interaction of TIMM9 and ITFG1 might be worth further research. Further research on TIMM9 could focus on specific aspects of metabolic and immune regulatory networks, and how TIMM9 participates in the regulation of ribosome, which is complicated and absorbing.

Conclusions

Altogether, this study has provided evidence that TIMM9 is a biomarker for oncogenesis and oncotherapy, which had not been previously established. TIMM9 was first demonstrated to be up-regulated in various cancers, by bioinformatics and RT-qPCR assays. Furthermore, we demonstrated that CNV and methylation is associated with expression of TIMM9, and S49L is an important missense mutation. We first demonstrated that TIMM9 is significantly related to multiple cancer-associated phenotypes, such as the variation of metabolism, immune system, ribosomes, and mitochondria, drug resistance and cancer-related pathways. We also demonstrated that TIMM9 is associated with networks of oxidative phosphorylation, glycolysis, TCA cycle, and electron respiratory chain. WGCNA analysis further demonstrated that these phenotypes are related to TIMM9 in LUAD. Our research has provided insights into the significance of the physical interaction between ITFG1 and TIMM9, and structural complementarity could form properly. TIMM9 could be seen as a new biomarker of tumorigenesis, metastasis, drug resistance, and variation in TME. TIMM9 may serve as a potential target for both glycolysis and oxidative phosphorylation. Our preliminary findings also suggest a positive correlation between the mTOR pathway and TIMM9 expression.

Material and methods

Gene expression analysis of TIMM9

We used “Gene_DE” module of TIMER 2.019 database (http://timer.cistrome.org/) and “pan-cancer” module of TNMplot20 database (https://tnmplot.com/analysis/) to obtain differential expression information of TIMM9 between tumor groups and adjacent normal tissue groups. “Proteomics” module of UALCAN21 database (http://ualcan.path.uab.edu/analysis.html) and “PATHOLOGY” module of the Human Protein Atlas database22 (HPA, https://www.proteinatlas.org/ENSG00000100575-TIMM9/pathology) were used to analyze the differential expression of TIMM9 at the proteomics level20. “Expression” module of GSCA23 platform (http://bioinfo.life.hust.edu.cn/GSCA/#/expression) was used to clarify the differential expression of TIMM9 among pathological stages21. “Target Discovery” module of DepMap24 portal (https://depmap.org/portal/tda/) was used to analyze the TIMM9 expression correlation between mRNA level and protein level. “Gene expression comparison” module of TNMplot20 database (https://tnmplot.com/analysis/) was utilized to analyze TIMM9 differential expression among normal, tumor, and metastatic stages.

Gene expression analysis on single-cell transcriptome level

Expression variation information among different clusters of cells in cancers was obtained from the “Dataset” module of the TISCH225 database (http://tisch.comp-genomics.org/gallery/). CellTracer26 database (http://bio-bigdata.hrbmu.edu.cn/CellTracer/CellTracer_index.jsp) was utilized to analyze variation of TIMM9 expression during cell the trajectory of pseudo-time to illustrate the association between TIMM9 and malignant cells in breast cancer.

Prognosis analysis of TIMM9

In order to explain the correlation between TIMM9 and prognosis in pan-cancer, “Survival Analysis” module of GEPIA227 database (http://gepia2.cancer-pku.cn/#index) was utilized to obtain Kaplan–Meier curves and heatmaps to visualize prognosis outcomes in overall survival (OS) and progression free survival (PFS). The analysis parameters utilized the median value to define the high and low expression groups, each comprising 50% of the samples. The Hazards Ratio (HR) and 95% Confidence Interval (CI) were included in the calculations, with the axis units set to months. The high and low expression groups were color-coded in red and blue, respectively, to facilitate clear comparison of survival outcomes. Kaplan–Meier curves from LOGpc28–31 database (https://bioinfo.henu.edu.cn/DatabaseList.jsp) was utilized as a supplement, which obtained raw data from TCGA and GEO database (https://www.ncbi.nlm.nih.gov/geo/). The division of the two groups was determined by the default parameters of the database, and the cutoff value is indicated on the figure. The GSCA23 database (https://guolab.wchscu.cn/GSCA/#/expression) was used to further clarify the association between TIMM9 expression levels and KIRC prognosis.

Genomic instabilities of TIMM9

Overall mutation landscape and mutation site visualization were analyzed by cBioPortal32 database (https://www.cbioportal.org/). Structure of TIMM9–TIMM10 complex was obtained from RCSB PDB platform (https://www.pdbus.org/, pdb code: 7cgp, Missense mutations simulations were performed by PyMOL 4.6.0 software (https://pymol.org/2/) particularly, and concrete missense mutation sites were obtained from the result of cBioportal analysis. PDBePISA webtool (https://www.ebi.ac.uk/msd-srv/prot_int/) was utilized to calculate binding free energy changes before and after mutation simulation. 2-D visualization of the conformation that represented a binding energy reduction was performed by LigPlot 2.2 software33 (https://www.ebi.ac.uk/thornton-srv/software/LigPlus/download2.html). Multiwfn 3.834 (dev) software (http://sobereva.com/multiwfn/) was utilized to perform Independent Gradient Model (IGM) analysis to characterize the interaction strength between mutation site and molecular environment, calculating based on promolecular electron density approximation. “CNV summary”, “CNV and Expression” and “CNV and Survival” modules in GSCA23 platform were respectively used in plotting the CNV percentage landscape, correlations of CNV with TIMM9 and TIMM10 mRNA expressions, associations of CNV with prognosis outcomes of overall survival (OS), progression free survival (PFS) and disease-specific survival (DSS). Similarly, “Differential methylation”, “Methylation and Survival”, “Methylation and Expression” modules in GSCA platform were individually analyzed for differential methylations of TIMM9 and correlations of TIMM9 and TIMM10 methylations with their expressions. The association between differential methylations of TIMM9 and prognosis outcomes in OS, PFS, and DSS was researched. The methylation level between normal group and tumor group was also analyzed by “Quick Start” module of Shiny Methylation Analysis Resource Tool35 (SMART, http://www.bioinfo-zs.com/smartapp/).

Analysis of phenotypes and pathways associated with TIMM9

In order to provide a detailed analysis and speculate biological functions mediated by TIMM9 in pan-cancer, GSEA analyses were performed by “Pathway Enrichment” module of CAMOIP36 platform (http://www.camoip.net/), which automatically compared samples with high-expression of TIMM9 and samples with low-expression of TIMM9. In addition, ssGSEA was also performed by “Pathway Enrichment” module of CAMOIP database, which was utilized to analyze associations of TIMM9 with significant phenotypes and pathways in important organelles. CellTracer26 database was utilized to analyze correlation between TIMM9 expressions and specific biological functions. Variations of correlations between cancer-associated phenotypes and TIMM9 along pseudo-time trajectories were visualized by heat maps. The relationship between TIMM9 and mTOR signaling pathway was analyzed by “Proteomics” module of UALCAN database (https://ualcan.path.uab.edu/). Using DepMap24, gene effect scores for TIMM9 and mTOR were analyzed through CRISPR knockout studies across various cell lines to determine whether these genes are functionally linked in supporting cell viability. The DepMap24 database was used to identify genes with expression patterns similar to TIMM9. Genes with the highest co-dependency scores were considered for further analysis, indicating a significant functional relationship with TIMM9. The obtained gene set was used to perform Gene Ontology (GO) enrichment by DAVID37 database (https://david.ncifcrf.gov/).

Analysis of correlations between TIMM9 and tumor microenvironments (TME)

TISIDB38 database (http://cis.hku.hk/TISIDB/index.php) was utilized to explore the relations between TIMM9 expression and immunocells phenotypes. CAMOIP36 platform was utilized to obtain correlations of TIMM9 with scores of TMB (Tumor Mutation Burder), MANTIS (an evaluation method of Microsatellite Instability) and Neoantigen Loads in order to analyze relations between genomics instabilities and differential expressions of TIMM9. Scatter plots in TIMER 2.019 database (http://timer.cistrome.org/) was used to research the level of tumor-related immune cell infiltration affected by TIMM9. TIMER2.0 database was used to clarify correlations between expression (exp) of immunostimulators and TIMM9.

Association between TIMM9 expression and overall survival (OS) and relapse-free survival (RFS) in anti-PD-1 treatment were tracked by using “Immunotherapy” module of KM-plotter39 database (http://kmplot.com/analysis/). All potential cutoff values between the lower and upper quartiles are calculated, and the threshold that performs the best is selected as the cutoff. “Gene” module of TISMO40 database (http://tismo.cistrome.org) was used to correlate TIMM9 expressions with responders of immunotherapies and cytokine therapies. The Wald test in DESeq2 was used for statistical evaluation of differences between groups (*FDR ≤ 0.05, **FDR ≤ 0.01, ***FDR ≤ 0.001), and the comparison results are summarized in the box plot. We analyzed changes in TIMM9 expression levels before and after cancer immunotherapy and targeted therapy using the DRMref41 (https://ccsm.uth.edu/DRMref/) and ICBatlas42 (https://guolab.wchscu.cn/ICBatlas/#!/) databases. Additionally, we examined the differences in TIMM9 expression levels between responders and non-responders to immunotherapy using the TISIDB38 (http://cis.hku.hk/TISIDB/) database, which was based on R package limma. “Immune” module of GSCA23 database was used to search the correlation between TIMM9 expression and infiltration.

Analysis of relations between TIMM9 and drug resistances

RNAactDrug43 database (http://bio-bigdata.hrbmu.edu.cn/RNAactDrug) was utilized to research correlations between variations of TIMM9, including in CNVs, methylations and expressions, to drug sensitivities. The KM plotter39 database was used to research correlations between TIMM9 and prognosis after chemotherapies. The ROC Plotter database (https://www.rocplot.org/) was used to analyze pathological complete responses mediated by TIMM9, using forms of receiver operating characteristic (ROC) curves and boxplots to show associations between the expressions of TIMM9 and responses to chemotherapy. Correlations between CTRP drug sensitivities and TIMM9 and TIMM10 mRNA expressions were visualized by the GSCA database.

Candidates for inhibitors were selected from compounds whose drug sensitivities were positively correlated with TIMM9. Initial structures of compounds were optimized by the “optimization” module of Gaussian 16 software, using b3lyp/6-311g(d,p) (http://gaussian.com/g16main/). Molecular docking between these compounds and the TIMM9–TIMM10 complex (pdb code: 7cgp) was performed by Autodock vina 1.1.2 software. The docking pocket was found by CavityPlus51 (http://www.pkumdl.cn:8000/cavityplus/index.php) database. The protein–ligand complex with the best docking affinity was selected to perform MD simulation by Gromacs 2020.4 software for 50 ns at 300 K, and the AMBER99SB force field was applied. Other parameters were assumed based on the official tutorial (http://www.mdtutorials.com/gmx/lysozyme/index.html). The RMSD curve, hydrogen number curve, and structure with B-factor were obtained from Gromacs. The last frame of MD was obtained, which was visualized in 2-D diagram by LigPlot 2.2 software, and Multiwfn 3.8 (dev) software was utilized to analyze the intermolecular interactions between the protein and ligand, using the method of noncovalent interaction (NCI) analysis. VMD44 1.9.3 software was utilized to clarify the result of the NCI analysis.

The CREAMMIST52 database (https://creammist.mtms.dev) was utilized to analyze the relationship between TIMM9 expression and drug sensitivity across multiple cancer cell lines. The "Associated Drugs (Gene Expression)" module was employed to calculate the Spearman correlations between TIMM9 gene expression and IC50 values for various drugs.

Analysis of relations between TIMM9 and gene interaction networks

Two gene sets associated with TIMM9 were separately obtained from GeneMANIA45 (http://genemania.org/) and BioGRID (https://thebiogrid.org/) databases. The Metascape46 (https://metascape.org/gp/index.html#/main/step1) and DAVID database (https://david.ncifcrf.gov/) were used to perform GO enrichments of the above two gene sets. KEGG, INTROPRO and REACTOME enrichments were performed for the gene set from the BioGRID47 database by the DAVID platform. Genes that physically (protein–protein) interacted with TIMM9 were analyzed by the Metascape database, and hub genes were obtained automatically. Protein structures of hub genes were obtained from the Alphafold47 database (https://alphafold.ebi.ac.uk/), since most crystal structures have never been reported. Physical interactions were simulated by the ZDOCK48 Server (https://zdock.umassmed.edu/). The structure of the complex with the lowest binding energy was first visualized in a 2-D diagram by LigPlot 2.2 software and secondly utilized in molecular dynamics (MD), which was performed by Gromacs 2020.4 software (https://www.gromacs.org/). AMBER99SB force field was adapt in 35 ns of MD simulation after energy minimum. Root means square deviation (RMSD), radius of gyration (Rg), number of hydrogen bonds and Gibbs energy landscape were analyzed by Gromacs 2020.4. Variations in secondary structures were identified by DSSP algorithm. Intermolecular interactions were analyzed by Multiwfn 3.8 (dev) software by NCI method. R package of “Bio3D”49 was used to perform PCA analysis of the trajectory, and movement pattern of PC1 was obtained.

Comprehensive analysis of TIMM9 expression and its implications in lung adenocarcinoma (LUAD)

Using the R packages “TCGAbiolinks” and “SummarizedExperiment”, we downloaded the expression count matrix of LUAD, using only samples of cancer patients for analysis. The CPM method was utilized to normalize the expression matrix. The WGCNA50 package was used for the modularization of gene networks. Automated networks were used to create co-expression gene networks. Modules were identified using hierarchical clustering and dynamic tree cutting functions. To connect modules with clinical characteristics, module membership (MM) as well as gene significance (GS) were estimated. Hub modules were identified by Pearson module membership correlation (MM) and a p < 0.05. High module connectivity and clinical importance were denoted by MM > 0.85 as well as GS > 0.2, respectively. GO enrichments were used to link each module with biological processes.

We utilized the GSCA database (https://guolab.wchscu.cn/GSCA/#/expression) to analyze TIMM9-related genes that were significantly positively correlated with TIMM9 expression in LUAD, obtained from the GEPIA and Biogrid database. Principal component analysis (PCA) was conducted to distinguish between tumor and normal samples. Gene set variation analysis (GSVA) scores were calculated to compare differences between tumor and normal tissues. Overall survival (OS), progression-free survival (PFS), and disease-specific survival (DSS) analyses were performed to assess the prognostic significance of GSVA scores. Spearman correlation analyses were conducted to evaluate the association between GSVA scores and pathway activities, specifically cell cycle and EMT pathways.

Cell lines and reagents

Cell lines (BEAS-2B, PC9, NCI-H460, NCI-H1299, LO2, HepG2) were obtained from the American Type Culture Collection (Manassas, VA, USA). Complete 1640 medium was prepared by adding 10% fetal bovine serum (ExCell Bio Inc., Shanghai, China) and 1% penicillin–streptomycin (Gibco, Grand Island, NY, USA). Cell lines were cultured in complete medium at 37 °C and 5% CO2. 5-FU (aladdin, China) and OSI-027 (aladdin, China) were dissolved in PBS (Sangon Biotech, China). We obtained CCK8 from Shanghai Beibo Biological Co. (Shanghai, China).

RT-qPCR assay

TRIzol (Sangon Biotech, China) was utilized to obtain total RNA from cells. We used the OD260/280 ratio to confirm the purity and concentration of RNA, which was then used to reverse transcribe RNA to cDNA. The primer sequences of GAPDH were 5′–3′ (Forward Primer: AATGGGCAGCCGTTAGGAAA) and 5′–3′ (Reverse Primer: GCCCAATACGACCAAATCAGAG). The primer sequences of TIMM9 were 5′–3′ (Forward Primer: AGGGACGTGGGATTAGGTAAGA) and 5′–3′ (Reverse Primer: TGTGCAGCCATATTCTTCTGGT). The sequences were synthesized by Sangon Biotech (Shanghai, China). The reverse transcription kit and qPCR kit were obtained from Vazyme (Nanjing, China). We used the 2(−∆∆CT) method to calculate expression levels. We conducted RT-qPCR assays to determine whether TIMM9 was upregulated in tumor cell lines (p-values calculated using one-way ANOVA, with comparisons made between each group of cancer cells and epithelial cells after normality and homogeneity of variance tests) and whether TIMM9 expression was downregulated after 24-h treatment with the mTOR inhibitor OSI-027 (p-values calculated using t-tests, with comparisons made between treated tumor cells and normal cells in each group after normality and homogeneity of variance tests).

CCK8 assay

2 × 103 cells were seeded into 96-well plates in advance. The assay was started with 5-FU concentration of 0, 100, 200, 400, 800, and 1600 μM. Then, we removed the medium. CCK-8 reagent was mixed with serum-free medium in an Eppendorf tube, and 100 μL was added to each well in the 96-well plates. After incubating the plates at 37 °C for 1 h, we detected the OD value at a wavelength of 450 nm (p value was calculated by one-way ANOVA and comparisons were made between each group, after normality tests and homogeneity of variance tests).

Supplementary Information

Supplementary Information.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-71421-3.

Acknowledgements

Authors thank Zinan Xie in Jiangsu University for his help in arranging of diagrams, thank Fengyi Ding in Sun Yat-Sen University for her help in partial data acquisition, thank Tian Lu in Beijing Kein Research Center for Natural Sciences for his unselfish guidance in molecular simulations, thank academic and spiritual supports from Xinqi Chen in Guangzhou University of Chinese Medicine, thank editors and reviewers for reviewing the manuscript.

Author contributions

R.Z., J.G. and Yanli He provided crucial thoughts for the design of research. L.Z. designed the research proposal. All authors participated in research design. L.Z., Y.W., C.H. performed qPCR and CCK8 assays. L.Z., B.L., Yan Huang, Y.Y., H.Q., performed the bioinformatics researches. Computational chemistry researches were performed by L.Z., Yan Huang, B.L., Y.Y., B.L., Yan Huang contributed in data and pictures collation. L.Z., B.L., and Yan Huang completed the paper writing. H.Z., C.H., provided critical suggestions for article writing. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by Natural Science Foundation of Guangdong Province, China [grant number No.2022A1515011575]; the National Natural Science Foundation of China [grant number 81873154]; National Undergraduate Training Program for Innovation and Entrepreneurship [grant number 202210572003].

Data availability

Data is provided within the manuscript or supplementary information files.

Competing interests

The authors declare no competing interests.

Publisher's note

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

These authors contributed equally: Lisheng Zhang, Yan Huang, Yanting Yang and Birong Liao.
==== Refs
References

1. Sica V Bravo-San Pedro JM Stoll G Kroemer G Oxidative phosphorylation as a potential therapeutic target for cancer therapy Int. J. Cancer 2020 146 10 17 10.1002/ijc.32616 31396957
Sica, V., Bravo-San Pedro, J. M., Stoll, G. & Kroemer, G. Oxidative phosphorylation as a potential therapeutic target for cancer therapy. Int. J. Cancer 146, 10–17. 10.1002/ijc.32616 (2020).31396957 10.1002/ijc.32616
2. Klein K He K Younes AI Role of mitochondria in cancer immune evasion and potential therapeutic approaches Front. Immunol. 2020 11 573326 10.3389/fimmu.2020.573326 33178201
Klein, K. et al. Role of mitochondria in cancer immune evasion and potential therapeutic approaches. Front. Immunol. 11, 573326. 10.3389/fimmu.2020.573326 (2020).33178201 10.3389/fimmu.2020.573326
3. Missiroli S Perrone M Genovese I Pinton P Giorgi C Cancer metabolism and mitochondria: Finding novel mechanisms to fight tumours EBioMedicine 2020 59 102943 10.1016/j.ebiom.2020.102943 32818805
Missiroli, S., Perrone, M., Genovese, I., Pinton, P. & Giorgi, C. Cancer metabolism and mitochondria: Finding novel mechanisms to fight tumours. EBioMedicine 59, 102943. 10.1016/j.ebiom.2020.102943 (2020).32818805 10.1016/j.ebiom.2020.102943
4. Ceh-Pavia E Spiller MP Lu H Folding and biogenesis of mitochondrial small Tim proteins Int. J. Mol. Sci. 2013 14 16685 16705 10.3390/ijms140816685 23945562
Ceh-Pavia, E., Spiller, M. P. & Lu, H. Folding and biogenesis of mitochondrial small Tim proteins. Int. J. Mol. Sci. 14, 16685–16705. 10.3390/ijms140816685 (2013).23945562 10.3390/ijms140816685
5. Paschen SA Neupert W Protein import into mitochondria IUBMB Life 2001 52 101 112 10.1080/15216540152845894 11798021
Paschen, S. A. & Neupert, W. Protein import into mitochondria. IUBMB Life 52, 101–112. 10.1080/15216540152845894 (2001).11798021 10.1080/15216540152845894
6. Baker MJ Mooga VP Guiard B Impaired folding of the mitochondrial small TIM chaperones induces clearance by the i-AAA protease J. Mol. Biol. 2012 424 227 239 10.1016/j.jmb.2012.09.019 23036860
Baker, M. J. et al. Impaired folding of the mitochondrial small TIM chaperones induces clearance by the i-AAA protease. J. Mol. Biol. 424, 227–239. 10.1016/j.jmb.2012.09.019 (2012).23036860 10.1016/j.jmb.2012.09.019
7. Webb CT Gorman MA Lazarou M Ryan MT Gulbis JM Crystal structure of the mitochondrial chaperone TIM9.10 reveals a six-bladed alpha-propeller Mol. Cell 2006 21 123 133 10.1016/j.molcel.2005.11.010 16387659
Webb, C. T., Gorman, M. A., Lazarou, M., Ryan, M. T. & Gulbis, J. M. Crystal structure of the mitochondrial chaperone TIM9.10 reveals a six-bladed alpha-propeller. Mol. Cell 21, 123–133. 10.1016/j.molcel.2005.11.010 (2006).16387659 10.1016/j.molcel.2005.11.010
8. Weinhäupl K Wang Y Hessel A Architecture and assembly dynamics of the essential mitochondrial chaperone complex TIM9·10·12 Structure 2021 29 1065 1073.e4 10.1016/j.str.2021.04.009 33974880
Weinhäupl, K. et al. Architecture and assembly dynamics of the essential mitochondrial chaperone complex TIM9·10·12. Structure 29, 1065-1073.e4. 10.1016/j.str.2021.04.009 (2021).33974880 10.1016/j.str.2021.04.009
9. Rampelt H Sucec I Bersch B The mitochondrial carrier pathway transports non-canonical substrates with an odd number of transmembrane segments BMC Biol. 2020 18 2 10.1186/s12915-019-0733-6 31907035
Rampelt, H. et al. The mitochondrial carrier pathway transports non-canonical substrates with an odd number of transmembrane segments. BMC Biol. 18, 2. 10.1186/s12915-019-0733-6 (2020).31907035 10.1186/s12915-019-0733-6
10. Lin C Fang C Sun D High expression of mitochondrial intermembrane chaperone TIMM9 represents a negative prognostic marker in gastric cancer J. Formos. Med. Assoc. 2017 116 476 483 10.1016/j.jfma.2016.08.007 27720672
Lin, C. et al. High expression of mitochondrial intermembrane chaperone TIMM9 represents a negative prognostic marker in gastric cancer. J. Formos. Med. Assoc. 116, 476–483. 10.1016/j.jfma.2016.08.007 (2017).27720672 10.1016/j.jfma.2016.08.007
11. Chen W Yang Z Chen Y A novel oxidative phosphorylation-associated gene signature for prognosis prediction in patients with hepatocellular carcinoma Dis. Mark. 2022 2022 3594901 10.1155/2022/3594901
Chen, W., Yang, Z. & Chen, Y. A novel oxidative phosphorylation-associated gene signature for prognosis prediction in patients with hepatocellular carcinoma. Dis. Mark. 2022, 3594901. 10.1155/2022/3594901 (2022).10.1155/2022/3594901
12. Sotgia F Whitaker-Menezes D Martinez-Outschoorn UE Salem AF Tsirigos A Lamb R Sneddon S Hulit J Howell A Lisanti MP Mitochondria, "fuel" breast cancer metabolism: Fifteen markers of mitochondrial biogenesis label epithelial cancer cells, but are excluded from adjacent stromal cells Cell Cycle 2012 11 4390 4401 10.4161/cc.22777 23172368
Sotgia, F. et al. Mitochondria, “fuel” breast cancer metabolism: Fifteen markers of mitochondrial biogenesis label epithelial cancer cells, but are excluded from adjacent stromal cells. Cell Cycle 11, 4390–4401. 10.4161/cc.22777 (2012).23172368 10.4161/cc.22777
13. Ivanova E Jowitt TA Lu H Assembly of the mitochondrial Tim9–Tim10 complex: A multi-step reaction with novel intermediates J. Mol. Biol. 2008 375 229 239 10.1016/j.jmb.2007.10.037 18022191
Ivanova, E., Jowitt, T. A. & Lu, H. Assembly of the mitochondrial Tim9–Tim10 complex: A multi-step reaction with novel intermediates. J. Mol. Biol. 375, 229–239. 10.1016/j.jmb.2007.10.037 (2008).18022191 10.1016/j.jmb.2007.10.037
14. Zhang A Miao K Sun H Deng CX Tumor heterogeneity reshapes the tumor microenvironment to influence drug resistance Int. J. Biol. Sci. 2022 18 3019 3033 10.7150/ijbs.72534 35541919
Zhang, A., Miao, K., Sun, H. & Deng, C. X. Tumor heterogeneity reshapes the tumor microenvironment to influence drug resistance. Int. J. Biol. Sci. 18, 3019–3033. 10.7150/ijbs.72534 (2022).35541919 10.7150/ijbs.72534
15. Läubli H Borsig L Altered cell adhesion and glycosylation promote cancer immune suppression and metastasis Front. Immunol. 2019 10 2120 10.3389/fimmu.2019.02120 31552050
Läubli, H. & Borsig, L. Altered cell adhesion and glycosylation promote cancer immune suppression and metastasis. Front. Immunol. 10, 2120. 10.3389/fimmu.2019.02120 (2019).31552050 10.3389/fimmu.2019.02120
16. Li Z Zhang H Reprogramming of glucose, fatty acid and amino acid metabolism for cancer progression Cell. Mol. Life Sci. 2016 73 377 392 10.1007/s00018-015-2070-4 26499846
Li, Z. & Zhang, H. Reprogramming of glucose, fatty acid and amino acid metabolism for cancer progression. Cell. Mol. Life Sci. 73, 377–392. 10.1007/s00018-015-2070-4 (2016).26499846 10.1007/s00018-015-2070-4
17. Blondy S David V Verdier M Mathonnet M Perraud A Christou N 5-Fluorouracil resistance mechanisms in colorectal cancer: From classical pathways to promising processes Cancer Sci. 2020 111 3142 3154 10.1111/cas.14532 32536012
Blondy, S. et al. 5-Fluorouracil resistance mechanisms in colorectal cancer: From classical pathways to promising processes. Cancer Sci. 111, 3142–3154. 10.1111/cas.14532 (2020).32536012 10.1111/cas.14532
18. Sethy C Kundu CN 5-Fluorouracil (5-FU) resistance and the new strategy to enhance the sensitivity against cancer: Implication of DNA repair inhibition Biomed. Pharmacother. 2021 137 111285 10.1016/j.biopha.2021.111285 33485118
Sethy, C. & Kundu, C. N. 5-Fluorouracil (5-FU) resistance and the new strategy to enhance the sensitivity against cancer: Implication of DNA repair inhibition. Biomed. Pharmacother. 137, 111285. 10.1016/j.biopha.2021.111285 (2021).33485118 10.1016/j.biopha.2021.111285
19. Li T Fu J Zeng Z Cohen D Li J Chen Q Li B Liu XS TIMER2.0 for analysis of tumor-infiltrating immune cells Nucleic Acids Res. 2020 48 W1 W509 W514 10.1093/nar/gkaa407 32442275
Li, T. et al. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res. 48(W1), W509–W514. 10.1093/nar/gkaa407 (2020).32442275 10.1093/nar/gkaa407
20. Bartha Á Győrffy B TNMplot.com: A web tool for the comparison of gene expression in normal, tumor and metastatic tissues Int. J. Mol. Sci. 2021 22 2622 10.3390/ijms22052622 33807717
Bartha, Á. & Győrffy, B. TNMplot.com: A web tool for the comparison of gene expression in normal, tumor and metastatic tissues. Int. J. Mol. Sci. 22, 2622. 10.3390/ijms22052622 (2021).33807717 10.3390/ijms22052622
21. Chandrashekar DS Bashel B Balasubramanya SAH Creighton CJ Ponce-Rodriguez I Chakravarthi BVSK Varambally S UALCAN: A portal for facilitating tumor subgroup gene expression and survival analyses Neoplasia 2017 19 8 649 658 10.1016/j.neo.2017.05.002 28732212
Chandrashekar, D. S. et al. UALCAN: A portal for facilitating tumor subgroup gene expression and survival analyses. Neoplasia 19(8), 649–658. 10.1016/j.neo.2017.05.002 (2017).28732212 10.1016/j.neo.2017.05.002
22. Rockberg J Uhlén M Prediction of antibody response using recombinant human protein fragments as antigen Protein Sci. 2009 18 11 2346 2355 10.1002/pro.245 19760667
Rockberg, J. & Uhlén, M. Prediction of antibody response using recombinant human protein fragments as antigen. Protein Sci. 18(11), 2346–2355. 10.1002/pro.245 (2009).19760667 10.1002/pro.245
23. Liu CJ Hu FF Xia MX Han L Zhang Q Guo AY GSCALite: A web server for gene set cancer analysis Bioinformatics 2018 34 21 3771 3772 10.1093/bioinformatics/bty411 29790900
Liu, C. J. et al. GSCALite: A web server for gene set cancer analysis. Bioinformatics 34(21), 3771–3772. 10.1093/bioinformatics/bty411 (2018).29790900 10.1093/bioinformatics/bty411
24. Tsherniak A Vazquez F Montgomery PG Weir BA Kryukov G Cowley GS Gill S Harrington WF Pantel S Krill-Burger JM Meyers RM Ali L Goodale A Lee Y Jiang G Hsiao J Gerath WFJ Howell S Merkel E Ghandi M Defining a cancer dependency map Cell 2017 170 3 564 576.e16 10.1016/j.cell.2017.06.010 28753430
Tsherniak, A. et al. Defining a cancer dependency map. Cell 170(3), 564-576.e16. 10.1016/j.cell.2017.06.010 (2017).28753430 10.1016/j.cell.2017.06.010
25. Han Y Wang Y Dong X TISCH2: Expanded datasets and new tools for single-cell transcriptome analyses of the tumor microenvironment Nucleic Acids Res. 2023 51 D1425 D1431 10.1093/nar/gkac959 36321662
Han, Y. et al. TISCH2: Expanded datasets and new tools for single-cell transcriptome analyses of the tumor microenvironment. Nucleic Acids Res. 51, D1425–D1431. 10.1093/nar/gkac959 (2023).36321662 10.1093/nar/gkac959
26. Guo Q Wang P Liu Q Hao Y Gao Y Qi Y Xu R Chen H Xin M Wu X Sun R Zhi H Zhang Y Ning S Li X Cell Tracer: A comprehensive database to dissect the causative multilevel interplay contributing to cell development trajectories Nucleic Acids Res. 2023 51 1 861 869 10.1093/nar/gkac892
Guo, Q. et al. Cell Tracer: A comprehensive database to dissect the causative multilevel interplay contributing to cell development trajectories. Nucleic Acids Res. 51(1), 861–869. 10.1093/nar/gkac892 (2023).10.1093/nar/gkac892
27. Tang Z Li C Kang B Gao G Li C Zhang Z GEPIA: A web server for cancer and normal gene expression profiling and interactive analyses Nucleic Acids Res. 2017 45 W98 W102 10.1093/nar/gkx247 28407145
Tang, Z. et al. GEPIA: A web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res. 45, W98–W102. 10.1093/nar/gkx247 (2017).28407145 10.1093/nar/gkx247
28. An Y Wang Q Zhang G Sun F Zhang L Li H Li Y Peng Y Zhu W Ji S Guo X OSlihc: An online prognostic biomarker analysis tool for hepatocellular carcinoma Front. Pharmacol. 2020 11 875 10.3389/fphar.2020.00875 32587519
An, Y. et al. OSlihc: An online prognostic biomarker analysis tool for hepatocellular carcinoma. Front. Pharmacol. 11, 875. 10.3389/fphar.2020.00875 (2020).32587519 10.3389/fphar.2020.00875
29. Zhang G Wang Q Yang M Yuan Q Dang Y Sun X An Y Dong H Xie L Zhu W Wang Y Guo X OSblca: A web server for investigating prognostic biomarkers of bladder cancer patients Front. Oncol. 2019 9 466 10.3389/fonc.2019.00466 31275847
Zhang, G. et al. OSblca: A web server for investigating prognostic biomarkers of bladder cancer patients. Front. Oncol. 9, 466. 10.3389/fonc.2019.00466 (2019).31275847 10.3389/fonc.2019.00466
30. Liu J Lichtenberg T Hoadley KA Poisson LM Lazar AJ Cherniack AD Kovatich AJ Benz CC Levine DA Lee AV Omberg L Wolf DM Shriver CD Thorsson V Cancer Genome Atlas Research NetworkHu H An Integrated TCGA pan-cancer clinical data resource to drive high-quality survival outcome analytics Cell 2018 173 2 400 416.e11 10.1016/j.cell.2018.02.052 29625055
Liu, J. et al. An Integrated TCGA pan-cancer clinical data resource to drive high-quality survival outcome analytics. Cell 173(2), 400-416.e11. 10.1016/j.cell.2018.02.052 (2018).29625055 10.1016/j.cell.2018.02.052
31. Wang Q Yan Z Ge L Li N Yang M Sun X Xie L Zhang G Zhu W Wang Y Li Y Li X Guo X OSeac: An online survival analysis tool for esophageal adenocarcinoma Front. Oncol. 2020 10 315 10.3389/fonc.2020.00315 32211334
Wang, Q. et al. OSeac: An online survival analysis tool for esophageal adenocarcinoma. Front. Oncol. 10, 315. 10.3389/fonc.2020.00315 (2020).32211334 10.3389/fonc.2020.00315
32. Gao J Aksoy BA Dogrusoz U Dresdner G Gross B Sumer SO Sun Y Jacobsen A Sinha R Larsson E Cerami E Sander C Schultz N Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal Sci. Signal. 2013 6 269 pl1 10.1126/scisignal.2004088 23550210
Gao, J. et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci. Signal. 6(269), pl1. 10.1126/scisignal.2004088 (2013).23550210 10.1126/scisignal.2004088
33. Laskowski RA Swindells MB LigPlot+: Multiple ligand-protein interaction diagrams for drug discovery J. Chem. Inf. Model. 2011 51 10 2778 2786 10.1021/ci200227u 21919503
Laskowski, R. A. & Swindells, M. B. LigPlot+: Multiple ligand-protein interaction diagrams for drug discovery. J. Chem. Inf. Model. 51(10), 2778–2786. 10.1021/ci200227u (2011).21919503 10.1021/ci200227u
34. Lu T Chen F Multiwfn: A multifunctional wavefunction analyzer J. Comput. Chem. 2012 33 5 580 592 10.1002/jcc.22885 22162017
Lu, T. & Chen, F. Multiwfn: A multifunctional wavefunction analyzer. J. Comput. Chem. 33(5), 580–592. 10.1002/jcc.22885 (2012).22162017 10.1002/jcc.22885
35. Li Y Ge D Lu C The SMART App: An interactive web application for comprehensive DNA methylation analysis and visualization Epigenet. Chromatin 2019 12 1 71 10.1186/s13072-019-0316-3
Li, Y., Ge, D. & Lu, C. The SMART App: An interactive web application for comprehensive DNA methylation analysis and visualization. Epigenet. Chromatin 12(1), 71. 10.1186/s13072-019-0316-3 (2019).10.1186/s13072-019-0316-3
36. Lin A Qi C Wei T Li M Cheng Q Liu Z Luo P Zhang J CAMOIP: A web server for comprehensive analysis on multi-omics of immunotherapy in pan-cancer Brief. Bioinform. 2022 23 3 bbac129 10.1093/bib/bbac129 35395670
Lin, A. et al. CAMOIP: A web server for comprehensive analysis on multi-omics of immunotherapy in pan-cancer. Brief. Bioinform. 23(3), bbac129. 10.1093/bib/bbac129 (2022).35395670 10.1093/bib/bbac129
37. Sherman BT Hao M Qiu J Jiao X Baseler MW Lane HC Imamichi T Chang W DAVID: A web server for functional enrichment analysis and functional annotation of gene lists (2021 update) Nucleic Acids Res. 2022 50 W1 W216 W221 10.1093/nar/gkac194 35325185
Sherman, B. T. et al. DAVID: A web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 50(W1), W216–W221. 10.1093/nar/gkac194 (2022).35325185 10.1093/nar/gkac194
38. Ru B Wong CN Tong Y Zhong JY Zhong SSW Wu WC Chu KC Wong CY Lau CY Chen I Chan NW Zhang J TISIDB: An integrated repository portal for tumor-immune system interactions Bioinformatics 2019 35 20 4200 4202 10.1093/bioinformatics/btz210 30903160
Ru, B. et al. TISIDB: An integrated repository portal for tumor-immune system interactions. Bioinformatics 35(20), 4200–4202. 10.1093/bioinformatics/btz210 (2019).30903160 10.1093/bioinformatics/btz210
39. Lánczky A Győrffy B Web-based survival analysis tool tailored for medical research (KMplot): Development and implementation J. Med. Internet Res. 2021 23 e27633 10.2196/27633 34309564
Lánczky, A. & Győrffy, B. Web-based survival analysis tool tailored for medical research (KMplot): Development and implementation. J. Med. Internet Res. 23, e27633. 10.2196/27633 (2021).34309564 10.2196/27633
40. Zeng Z Wong CJ Yang L TISMO: Syngeneic mouse tumor database to model tumor immunity and immunotherapy response Nucleic Acids Res. 2022 50 D1391 D1397 10.1093/nar/gkab804 34534350
Zeng, Z. et al. TISMO: Syngeneic mouse tumor database to model tumor immunity and immunotherapy response. Nucleic Acids Res. 50, D1391–D1397. 10.1093/nar/gkab804 (2022).34534350 10.1093/nar/gkab804
41. Liu X Yi J Li T Wen J Huang K Liu J Wang G Kim P Song Q Zhou X DRMref: Comprehensive reference map of drug resistance mechanisms in human cancer Nucleic Acids Res. 2024 52 D1 D1253 D1264 10.1093/nar/gkad1087 37986230
Liu, X. et al. DRMref: Comprehensive reference map of drug resistance mechanisms in human cancer. Nucleic Acids Res. 52(D1), D1253–D1264. 10.1093/nar/gkad1087 (2024).37986230 10.1093/nar/gkad1087
42. Yang M Miao YR Xie GY Luo M Hu H Kwok HF Feng J Guo AY ICBatlas: A comprehensive resource for depicting immune checkpoint blockade therapy characteristics from transcriptome profiles Cancer Immunol. Res. 2022 10 11 1398 1406 10.1158/2326-6066.CIR-22-0249 36095221
Yang, M. et al. ICBatlas: A comprehensive resource for depicting immune checkpoint blockade therapy characteristics from transcriptome profiles. Cancer Immunol. Res. 10(11), 1398–1406. 10.1158/2326-6066.CIR-22-0249 (2022).36095221 10.1158/2326-6066.CIR-22-0249
43. Dong Q Li F Xu Y Xiao J Xu Y Shang D Zhang C Yang H Tian Z Mi K Li X Zhang Y RNAactDrug: A comprehensive database of RNAs associated with drug sensitivity from multi-omics data Brief. Bioinform. 2020 21 6 2167 2174 10.1093/bib/bbz142 31799597
Dong, Q. et al. RNAactDrug: A comprehensive database of RNAs associated with drug sensitivity from multi-omics data. Brief. Bioinform. 21(6), 2167–2174. 10.1093/bib/bbz142 (2020).31799597 10.1093/bib/bbz142
44. Humphrey W Dalke A Schulten K VMD: Visual molecular dynamics J. Mol. Graph. 1996 14 33 28 10.1016/0263-7855(96)00018-5 8744570
Humphrey, W., Dalke, A. & Schulten, K. VMD: Visual molecular dynamics. J. Mol. Graph. 14, 33–28. 10.1016/0263-7855(96)00018-5 (1996).8744570 10.1016/0263-7855(96)00018-5
45. Franz M Rodriguez H Lopes C Zuberi K Montojo J Bader GD Morris Q GeneMANIA update Nucleic Acids Res. 2018 46 W1 W60 W64 10.1093/nar/gky311(2018) 29912392
Franz, M. et al. GeneMANIA update. Nucleic Acids Res. 46(W1), W60–W64. 10.1093/nar/gky311(2018) (2018).29912392 10.1093/nar/gky311(2018)
46. Zhou Y Zhou B Pache L Metascape provides a biologist-oriented resource for the analysis of systems-level datasets Nat. Commun. 2019 10 1523 10.1038/s41467-019-09234-6 30944313
Zhou, Y. et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat. Commun. 10, 1523. 10.1038/s41467-019-09234-6 (2019).30944313 10.1038/s41467-019-09234-6
47. Oughtred R Rust J Chang C The BioGRID database: A comprehensive biomedical resource of curated protein, genetic, and chemical interactions Protein Sci. 2021 30 187 200 10.1002/pro.3978 33070389
Oughtred, R. et al. The BioGRID database: A comprehensive biomedical resource of curated protein, genetic, and chemical interactions. Protein Sci. 30, 187–200. 10.1002/pro.3978 (2021).33070389 10.1002/pro.3978
48. Pierce BG Wiehe K Hwang H Kim BH Vreven T Weng Z ZDOCK server: Interactive docking prediction of protein–protein complexes and symmetric multimers Bioinformatics 2014 30 1771 1773 10.1093/bioinformatics/btu097 24532726
Pierce, B. G. et al. ZDOCK server: Interactive docking prediction of protein–protein complexes and symmetric multimers. Bioinformatics 30, 1771–1773. 10.1093/bioinformatics/btu097 (2014).24532726 10.1093/bioinformatics/btu097
49. Grant BJ Skjaerven L Yao XQ The Bio3D packages for structural bioinformatics Protein Sci. 2021 30 20 30 10.1002/pro.3923 32734663
Grant, B. J., Skjaerven, L. & Yao, X. Q. The Bio3D packages for structural bioinformatics. Protein Sci. 30, 20–30. 10.1002/pro.3923 (2021).32734663 10.1002/pro.3923
50. Langfelder P Horvath S WGCNA: An R package for weighted correlation network analysis BMC Bioinform. 2008 9 559 10.1186/1471-2105-9-559
Langfelder, P. & Horvath, S. WGCNA: An R package for weighted correlation network analysis. BMC Bioinform. 9, 559. 10.1186/1471-2105-9-559 (2008).10.1186/1471-2105-9-559
51. Xu Y Wang S Hu Q Gao S Ma X Zhang W Shen Y Chen F Lai L Pei J CavityPlus: A web server for protein cavity detection with pharmacophore modelling, allosteric site identification and covalent ligand binding ability prediction Nucleic Acids Res. 2018 46 W1 W374 W379 10.1093/nar/gky380 29750256
Xu, Y. et al. CavityPlus: A web server for protein cavity detection with pharmacophore modelling, allosteric site identification and covalent ligand binding ability prediction. Nucleic Acids Res. 46(W1), W374–W379. 10.1093/nar/gky380 (2018).29750256 10.1093/nar/gky380
52. Yingtaweesittikul H Wu J Mongia A Peres R Ko K Nagarajan N Suphavilai C CREAMMIST: An integrative probabilistic database for cancer drug response prediction Nucleic Acids Res. 2023 51 D1 D1242 D1248 10.1093/nar/gkac911 36259664
Yingtaweesittikul, H. et al. CREAMMIST: An integrative probabilistic database for cancer drug response prediction. Nucleic Acids Res. 51(D1), D1242–D1248. 10.1093/nar/gkac911 (2023).36259664 10.1093/nar/gkac911
