Establishment and evaluation of a prognostic model for sodium overload necrosis in acute myeloid leukemia

Introduction

Background

Acute myeloid leukemia (AML) is a malignant hematological disorder, characterized by the uncontrolled proliferation of myeloid blasts that suppress normal hematopoiesis (1), and poses life-threatening risks to patients. Advances in novel targeted therapies, optimized supportive care, and reduced transplant-related mortality have improved the 5-year overall survival (OS) rate to 30% since 2000 (2), yet outcomes remain suboptimal, which is largely attributable to the genomic heterogeneity and clonal evolution of AML. Current risk stratification—classifying patients into favorable-, intermediate-, and adverse-risk groups based on European LeukemiaNet (ELN) 2022 molecular genetics and cytogenetics (3)—guides prognosis and treatment. However, this stratification poses clinical limitations, particularly in patients classified as intermediate-risk who exhibit significant heterogeneity in outcomes. It is worthwhile to investigate novel predictive models that may complement the ELN system and potentially provide additional guidance for therapeutic strategies in AML.

Rationale and knowledge gap

Dysregulation of cell death is a potential etiological cause of AML and chemotherapy resistance. Transient receptor potential melastatin 4 (TRPM4) is a calcium-activated, nonselective, monovalent cation channel. Opening of TRPM4 channels can lead to sodium ion influx, resulting in cellular swelling, plasma membrane rupture, and necrotic cell death (4). This process is accompanied by the release of damage-associated molecular patterns (DAMPs), thereby eliciting antitumor immune responses. Necrosis by sodium overload (NECSO), a form of regulated necrosis rather than an independent programmed cell death pathway, primarily depends on sodium overload instead of iron accumulation or RIPK signaling pathways (5). Notably, TRPM4 overexpression increases the susceptibility of tumor cells to NECSO. TRPM4 is widely expressed in immune cells (6,7) and in various organs (8,9) and is involved in the regulation of multiple physiological functions. TRPM4 is highly expressed in multiple cancers, including diffuse large B-cell lymphoma (10), prostate cancer (11), and breast cancer (12). High expression is closely associated with tumor aggressiveness, disease recurrence, and shorter OS. Building on this evidence, we propose developing a model based on genes related to sodium overload-induced necrosis to predict survival outcomes in AML.

Objective

Using The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases, we constructed a prognostic model using genes related to sodium overload-induced necrosis. We investigated prognostic and immunological characteristics between subgroups to complement the limitations of the ELN risk stratification. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2897/rc).

Methods

Data download and processing

This study systematically analyzed RNA-sequencing data from patients with AML and normal controls based on TCGA and GEO databases. First, the TCGA-Acute Myeloid Leukemia (LAML) dataset, consisting of 151 AML samples, was downloaded from TCGA. Count data were converted to transcripts per million (TPM) format. We also downloaded the GSE12417 dataset from the GEO database. The data corresponding to the GPL570 platform were selected, resulting in 78 AML samples. Samples with incomplete survival data were excluded, clinical variables with substantial missingness were not included in multivariable analyses, and no imputation was performed. All data were processed in a uniform manner for subsequent validation. Furthermore, based on previous literature reports (4), TRPM4 was identified as a key gene that regulates sodium channel-induced cell death. To search for genes related to sodium overload-induced necrosis that are associated with TRPM4, a co-expression analysis was performed comparing the expression data of TCGA_LAML and the TRPM4 gene using the R package “limma”. Filtering criteria were set to a correlation coefficient cutoff of 0.4 and a P value of 0.001. This analysis identified a set of genes related to sodium ion overload-induced necrosis. Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were performed to explore the potential biological functions and signaling pathways associated with the identified gene set. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Construction of the sodium overload-induced necrosis prognostic risk model

The expression of genes related to sodium overload-induced necrosis were extracted from patients with AML in TCGA and GEO datasets. The TCGA_LAML dataset served as the training set, and the GSE12417 dataset served as the test set. The intersecting genes between the two datasets were identified to obtain the corresponding expression matrices for both the training and test cohorts. The ComBat method was used to remove batch effects between platforms. Within the training set, univariate Cox regression analysis was performed on related genes using the R package “survival”. Genes with a P value <0.05 were considered prognostic and selected for a subsequent least absolute shrinkage and selection operator (LASSO) regression analysis to construct the sodium overload-induced necrosis prognostic risk model. The risk score was calculated: ∑(Coef × xi), where xi symbolizes the normalized expression of target gene i, and Coef symbolizes the regression coefficient (13). Samples from TCGA-LAML and GPL570-GSE12417 were stratified into high- and low-risk groups based on the median risk score calculated from the model. Kaplan-Meier survival analysis was conducted in the high-risk and low-risk groups using the R package “survival”, with statistical significance of group differences assessed by the log-rank test. Risk curves, risk score distribution plots, and risk heat maps for the TCGA-LAML cohort were generated using the R packages “survival” and “pheatmap”. The receiver operating characteristic (ROC) curves for the TCGA-LAML cohort and GPL570-GSE12417 cohort were plotted using the R package “timeROC” to evaluate the predictive performance of the model. Univariate and multivariate Cox regression analyses were performed to assess the independent prognostic value of the risk score by incorporating the risk score and clinical prognostic factors from the TCGA-LAML data set. Stratified analysis was also conducted by grouping based on the risk score combined with clinical factors to explore the prognosis in different subgroups. Gene Set Enrichment Analysis (GSEA) was performed in the high-risk TCGA-LAML group to investigate possible functional mechanisms and enriched pathways.

Constructing a nomogram

Using the R packages “survival” and “regplot”, clinical variables (age and sex) were combined with the RiskScore to develop a nomogram. This enables visual prediction and quantitative assessment of the patient’s prognosis.

Differential and correlation analysis of the RiskScore with immune cell infiltration

Immune cell infiltration was evaluated using the CIBERSORTx algorithm. The CIBERSORT R package, together with the LM22 reference gene signature matrix, was applied to quantify the relative proportions of 22 immune cell types in each sample from the TCGA-LAML cohort. The estimated fractions of the 22 immune cell subsets in each sample summed to 1. Only samples with a CIBERSORT output P value <0.05 were considered for subsequent analyses. In the TCGA-LAML training cohort, a differential analysis of immune cell infiltration levels between the high- and low-risk groups was performed using the R packages “limma” and “vioplot”. The correlation between the RiskScore and various types of immune cells was further assessed to explore potential associations between the RiskScore and immune microenvironment.

Correlation analysis between RiskScore and AML mutation burden

Mutation data from 131 patients with AML were downloaded from the TCGA portal. The R package “maftools” was employed to visualize the overall mutational landscape of AML, including mutation frequencies and types. Patients were stratified into high- and low-risk groups based on their RiskScore, and oncoplots (mutation waterfall plots) were generated to demonstrate the distribution patterns and differences in the top 20 most frequently mutated genes between the two groups.

Drug sensitivity

Drug sensitivity scores were calculated for each sample in the TCGA training cohort using the R packages “oncoPredict”, “limma”, and “parallel”. Differential analysis of drug sensitivity between the high- and low-risk groups was performed, with statistical significance set at P<0.05, indicating differential drug responses across risk groups.

Statistical analysis

Univariate Cox regression analysis was used to identify genes related to the prognosis. Survival curves were generated using the Kaplan-Meier method and differences between-groups were assessed using log-rank tests. All analyzes were performed using the R software (version 4.4.1), and statistical significance was set at P<0.05.

Results

Identification of necrosis-related genes associated with sodium overload

Based on TCGA expression data, co-expression analysis was performed using TRPM4, and 521 genes that were significantly co-expressed with TRPM4 were identified. A co-expression network diagram (Figure 1) was generated, where red indicates a positive correlation with TRPM4, and blue indicates a negative correlation. The GO enrichment analysis of these 521 co-expressed genes revealed that they were mainly enriched in pathways related to the cellular response to copper ions, detoxification of copper ions, the aerobic electron transport chain for adenosine triphosphate (ATP) synthesis, calcium ion secretion, and atrial cardiac muscle cells (Figure 2). Further analysis of the KEGG pathway showed that these genes were mainly enriched in pathways, including insulin secretion, folate biosynthesis, oxidative phosphorylation, and cytokine-cytokine receptor interaction (Figure 3).

Figure 1 Co-expression network diagram of TRPM4-related genes. TRPM4, transient receptor potential melastatin 4.

Figure 2 Gene Ontology enrichment analysis. (A) Bar chart. (B) Bubble plot. (C) Circular plot. (D) Tree diagram. ATP, adenosine triphosphate; BP, biological process; CC, cellular component; MF, molecular function.

Figure 3 Kyoto Encyclopedia of Genes and Genomes enrichment analysis. (A) Bar chart. (B) Bubble plot. (C) Tree diagram.

Construction of a prognostic model for sodium overload-induced necrosis

The expression of sodium overload necrosis-related genes was extracted from TCGA_LAML and GSE12417_GPL570 datasets. First, a univariate COX regression analysis was performed on the TCGA-LAML training set, identifying 150 prognostic genes with a P value of <0.05. Subsequently, the LASSO regression analysis was applied to these 150 genes. Based on the optimal λ value, nine genes were selected for the construction of the model. A plot of the coefficient profile was generated, and the RiskScore was calculated (Figure 4).

Figure 4 Least absolute shrinkage and selection operator regression. (A) Cross-validation curve. (B) Coefficient plot.

Model evaluation

The low-risk group exhibited significantly better survival outcomes than the high-risk group in both training and test sets, with statistically significant differences (P<0.05; Figure 5). In TCGA-LAML training set, the RiskScore yielded P values of <0.001 in both univariate and multivariate analyses, indicating the robustness of the model and the strong independent predictive capacity. The ROC curves for 1-, 3-, and 5-year survival in the TCGA-LAML training set showed area under the curve (AUC) values of 0.838, 0.811, and 0.774, respectively, The ROC curves for 1- and 3-year survival in the test set showed AUC values of 0.724 and 0.646, respectively, indicating that the model maintained satisfactory predictive performance at different time points. In the TCGA-LAML training cohort, comparative ROC analyses incorporating age and sex demonstrated that the RiskScore achieved an AUC of 0.838, which was markedly superior to that of the other clinical variables. In the test set, although sex information was unavailable, comparison with age alone showed that the RiskScore still achieved an AUC of 0.724, outperforming age as a prognostic factor (Figure 6). Additionally, risk curve plots, risk score distributions, and risk heat maps for the training set demonstrated a progressive increase in patient mortality with increasing RiskScore values, further validating the correlation between the RiskScore and prognosis. Regarding the prognostic value of the RiskScore stratified by age, among low-risk patients, those aged <65 years had the most favorable prognosis, whereas patients aged >65 years had the worst outcomes. When the RiskScore was stratified by sex, male patients in the low-risk group exhibited a survival advantage, whereas both sexes in the high-risk group showed similarly poor survival prognoses (Figure 7).

Figure 5 Survival curves. (A) The Cancer Genome Atlas training set. (B) Gene Expression Omnibus testing set.

Figure 6 Additional model evaluation using The Cancer Genome Atlas training set and Gene Expression Omnibus testing set. (A) Multivariate analysis of training set. (B) Univariate analysis of training set. (C) ROC curves of clinical indicators for training set. (D) 1-, 3-, and 5-year ROC curves of training set. (E) ROC curves of clinical indicators for testing set. (F) 1- and 3-year ROC curves of training set. AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic.

Figure 7 Additional model evaluation using The Cancer Genome Atlas training set. (A) Risk score distribution plot. (B) Stratified prognostic analysis of risk score by age. (C) Stratified prognostic analysis of risk score by sex.

Analysis of clinical nomograms

A nomogram model was constructed by integrating the RiskScore with clinical traits to predict patient survival probabilities. As shown in Figure 7, the points were assigned according to sex, RiskScore, and age. The total points were then calculated to predict the survival probabilities at different time points. For example, a patient with a total score of 113 had 1-, 3-, and 5-year survival probabilities of 0.71, 0.438, and 0.306, respectively (Figure 8).

Figure 8 Nomogram analysis. ***, P<0.001.

Enrichment analysis

GSEA on high-risk group samples from TCGA-LAML cohort revealed significant enrichment in pathways related to antigen presentation and processing, chemokine signaling, immunoglobulin A (IgA) production in the intestinal immune network, and systemic lupus erythematosus. These findings suggested the dysregulation of immune regulation and inflammation-associated pathways in high-risk patients (Figure 9).

Figure 9 Gene Set Variation Analysis of high-risk group. KEGG, Kyoto Encyclopedia of Genes and Genomes.

RiskScore and immune cells

Analysis of the differences in immune cell infiltration between the high- and low-risk groups in the TCGA-LAML cohort revealed significant variations in monocytes, resting mast cells, eosinophils, plasma cells, and M2 macrophages (P<0.05; Figure 10A). The correlation analysis between RiskScore and immune cells demonstrated that the RiskScore was positively correlated with monocytes and M2 macrophages but negatively correlated with plasma cells, mast cells, and eosinophils (Figure 10B). These findings suggested that the RiskScore may contribute to the progression of AML by modulating the cellular composition of the immune microenvironment.

Figure 10 Immune infiltration analysis. (A) Differential analysis of immune cells in high-/low-risk groups. (B) Correlation analysis between risk score and immune cells. NK, natural killer.

Risk score and AML mutation burden

Mutation data for 131 AML patients were obtained from TCGA database (Figure 11A). Patients were stratified into high- and low-risk groups based on the RiskScore, and the distribution of the top 20 most frequently mutated genes was visualized for each group. In the high-risk group, the most prevalent mutations occurred in NPM1, RUNX1, and TP53, whereas the low-risk group showed frequent mutations in IDH2, CALR, and DNMT3A (Figure 11B,11C). This distinct mutational landscape suggests differential mutation profiles between the risk groups.

Figure 11 Mutation analysis. (A) Mutation landscape. (B) Top 20 mutated genes in high-risk group. (C) Top 20 mutated genes in the low-risk group. DEL, deletion; INS, insertion; NK, natural killer; SNP, single nucleotide polymorphism; SNV, single nucleotide variant; TMB, tumor mutational burden.

Drug sensitivity analysis

Analysis of TCGA training cohort revealed significant differences in sensitivity to multiple drugs between the high- and low-risk groups (P<0.05). Specifically, a higher sensitivity to 5-fluorouracil was observed in the low-risk group, whereas an increased sensitivity to ABT-737, AZD6482, and entinostat was observed in the high-risk group (Figure 12). These findings suggested the existence of different drug response profiles between risk groups, potentially providing preliminary insights for the development of personalized therapeutic strategies.

Figure 12 Drug sensitivity. (A) 5-fluorouracil. (B) ABT-737. (C) AZD-6482. (D) Entinostat (HDAC inhibitor). HDAC, histone deacetylase; INS, insertion.

Discussion

Key findings

Using TCGA as the training set and the GEO datasets for validation, we identified nine genes related to sodium overload-induced necrosis to construct a prognostic model. The results demonstrated significantly superior survival outcomes in the low-risk group in both the training and the testing cohorts. Univariate and multivariate Cox regression analyses confirmed that the developed model was an independent prognostic predictor.

Strengths and limitations

This study is believed to make a significant contribution to the literature because the RiskScore model developed addresses the limitations of ELN risk stratification, particularly among patients classified as intermediate risk who demonstrate substantial heterogeneity in clinical outcomes. The findings also provide a theoretical basis for improving risk stratification management and guiding targeted therapy. Nevertheless, several limitations should be acknowledged. Chief among these is the reliance on publicly available biological databases, and further experimental validation is required to confirm the robustness and applicability of the findings. Another limitation is the lack of complete ELN 2022 classification and cytogenetic data in public cohorts, which precluded a full assessment of the RiskScore’s incremental value beyond ELN stratification. Prospective studies with comprehensive clinical annotation are warranted to further validate its clinical utility.

Comparison with similar research

AML is a hematological malignancy characterized by distinct heterogeneity, with considerably variability in survival outcomes due to different genetic mutations. Although current prognostic stratification systems have significantly improved the risk assessment and treatment planning for AML, clonal diversity and evolution contribute to primary drug resistance or disease relapse or progression in some patients (14). Therefore, it is imperative to explore novel therapeutic strategies for this subset of patients.

Prognostic models based on programmed cell death mechanisms provide critical predictive values for AML, with current research focusing primarily on pyroptosis, autophagy, and ferroptosis. However, sodium overload-induced necrosis has received little attention in patients with AML. To our knowledge, this is the first study to incorporate sodium overload-induced necrosis into an AML prognostic model, thus complementing conventional ELN risk stratification. As reported previously (4), TRPM4 is currently the only known gene associated with sodium overload-induced necrosis. Using gene expression data from TCGA database, we performed a co-expression analysis with TRPM4 to identify potential genes linked to this necrotic pathway. Enrichment analysis of these genes revealed their predominant involvement in copper ion homeostasis and calcium ion signaling pathways. Elevated copper levels in the serum and tumor cells in patients with cancer are correlated with tumor cell proliferation and disease burden (15). Similarly, massive sodium influx triggers membrane depolarization, disrupting store-operated calcium entry (SOCE), a process modulated by calcium release-activated channels. Altered SOCE activity enhances cellular proliferation and migration, while suppressing apoptosis (16-20), collectively promoting oncogenic phenotypes.

Explanations of findings

ROC curve analysis and risk stratification plots indicated a high predictive precision for patient survival, with elevated RiskScores strongly correlating with increased mortality risk. In particular, the expression of genes included in the model increased proportionally with increasing RiskScores, identifying them as high-risk biomarkers. Tregs establish immunosuppressive microenvironments in tumors via CCL22-dependent mechanisms (21); NADH:ubiquinone oxidoreductase subunit C1 (NDUFC1) is upregulated in gastric cancer (22), and RAB36 (member of the RAS oncogene family) is overexpressed in bladder cancer (23). Collectively, these findings suggest that the genes included in our model are potential therapeutic targets with prognostic value. Stratified analysis integrating the RiskScore with age and sex revealed optimal survival in patients aged <65 years with a low RiskScore, whereas those aged >65 years with high RiskScore exhibited the poorest results. Male patients in the low-risk group had a modest survival advantage. The clinical nomogram effectively predicted survival by combining the RiskScore with clinical features. Furthermore, GSEA revealed a significant enrichment of high-risk patients in immune pathways, including antigen presentation, chemokine signaling, IgA production in the intestinal immune network, and systemic lupus erythematosus, implicating immune microenvironment dysregulation in the progression of AML.

Further analysis exploring the relationship between our prognostic model and the immune microenvironment revealed significant disparities in immune cell infiltration, particularly involving monocytes, M2 macrophages, and mast cells, between the high- and low-risk groups. The RiskScore exhibited a positive correlation with monocytes and M2 macrophages, but a negative correlation with mast cells and other immune subsets. Current evidence indicates that tumor-associated macrophages, which often resemble the M2 phenotype, promote tumor growth and proliferation while suppressing antitumor immunity (24). In contrast, mast cells play a dual role in the tumor microenvironment; they can exert antitumor effects interacting with innate (e.g. neutrophils and eosinophils) and adaptive immune cells (25).

The prognosis of AML is significantly correlated with the burden of tumor mutations. In this study, the high-risk group predominantly harbored mutations in NPM1, RUNX1, and TP53, whereas the low-risk group exhibited frequent mutations in IDH2, CALR, and DNMT3A. In particular, although NPM1 mutations are classified as favorable risk factors in conventional ELN stratification, our model identified them as enriched in high-risk patients. This apparent contradiction likely stems from the co-occurrence of adverse mutations in high-risk cases, such as RUNX1 and TP53, which are frequently associated with complex karyotypes, and co-mutations in genes such as TET2 and SF3B1 (26). In low-risk patients, targeting IDH2 and DNMT3A mutations with IDH2 inhibitors and hypomethylating agents has resulted in improved outcomes. Given the limited sample size, the mutation analyses in this study are exploratory and hypothesis-generating, and these findings require validation in larger cohorts. Drug sensitivity analysis revealed a greater response to ABT-737, buparlisib, and entinostat in high-risk patients. ABT-737 is a synthetic BH3 mimetic that potently binds BCL-2, displacing pro-apoptotic Bax/Bak to induce tumor apoptosis (27). The limited oral bioavailability of ABT-737 prompted the development of venetoclax (ABT-199), which exhibits improved pharmacokinetics and clinical efficacy, leading to its broad adoption in AML therapy. Buparlisib is a highly selective PI3K inhibitor that suppresses proliferation and promotes apoptosis through the blockade of the PI3K-AKT pathway (28). Entinostat, an HDAC1 (histone deacetylase 1) inhibitor, induces leukemic cell apoptosis and reduces tumor burden (29). Thus, high-risk patients exhibit reduced sensitivity to conventional chemotherapy but may benefit from these novel targeted agents.

Implications and actions needed

This study underscores the prognostic significance of sodium overload-induced necrosis in AML. The established RiskScore model may enhance risk stratification beyond conventional classification systems and contribute to more precise individualized prognostic evaluation. Further prospective studies are required to validate the robustness and clinical utility of this model and to elucidate the underlying mechanisms linking sodium overload-related necrosis to the immune microenvironment, thereby facilitating the identification of potential therapeutic targets and immunomodulatory strategies.

Conclusions

This study established a risk score-based prognostic model that demonstrates robust predictive power for patient outcomes and suggested the underlying mechanisms of sodium overload-induced necrosis in AML. Our study reveals disease characteristics and differences in biological behavior from a variety of perspectives, thereby providing insight into the identification of novel therapeutic targets.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2897/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2897/prf

Funding: This study was supported by the 2024 Scientific Research Development Program of the Affiliated Hospital of North Sichuan Medical College (No. 2024MPZK012) and Project of Sichuan Province Primary Healthcare Development Research Center (No. SWFZ19-Q-09).

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2897/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. DiNardo CD, Erba HP, Freeman SD, et al. Acute myeloid leukaemia. Lancet 2023;401:2073-86. [Crossref] [PubMed]
2. National Institute of Health. Surveillance, Epidemiology, and End Results (SEER) US county populations 1969–2020; 2022. Available online: http://www.seer.cancer.gov/popdata [accessed November 9, 2022].
3. Lachowiez CA, Long N, Saultz J, et al. Comparison and validation of the 2022 European LeukemiaNet guidelines in acute myeloid leukemia. Blood Adv 2023;7:1899-909. [Crossref] [PubMed]
4. Fu W, Wang J, Li T, et al. Persistent activation of TRPM4 triggers necrotic cell death characterized by sodium overload. Nat Chem Biol 2025;21:1238-49. [Crossref] [PubMed]
5. Mansour HM, Mohamed AF, El-Khatib AS, et al. Kinases control of regulated cell death revealing druggable targets for Parkinson's disease. Ageing Res Rev 2023;85:101841. [Crossref] [PubMed]
6. Barbet G, Demion M, Moura IC, et al. The calcium-activated nonselective cation channel TRPM4 is essential for the migration but not the maturation of dendritic cells. Nat Immunol 2008;9:1148-56. [Crossref] [PubMed]
7. Weber KS, Hildner K, Murphy KM, et al. Trpm4 differentially regulates Th1 and Th2 function by altering calcium signaling and NFAT localization. J Immunol 2010;185:2836-46. [Crossref] [PubMed]
8. Launay P, Fleig A, Perraud AL, et al. TRPM4 is a Ca2+-activated nonselective cation channel mediating cell membrane depolarization. Cell 2002;109:397-407. [Crossref] [PubMed]
9. Nilius B, Prenen J, Droogmans G, et al. Voltage dependence of the Ca2+-activated cation channel TRPM4. J Biol Chem 2003;278:30813-20. [Crossref] [PubMed]
10. Loo SK, Ch'ng ES, Md Salleh MS, et al. TRPM4 expression is associated with activated B cell subtype and poor survival in diffuse large B cell lymphoma. Histopathology 2017;71:98-111. [Crossref] [PubMed]
11. Berg KD, Soldini D, Jung M, et al. TRPM4 protein expression in prostate cancer: a novel tissue biomarker associated with risk of biochemical recurrence following radical prostatectomy. Virchows Arch 2016;468:345-55. [Crossref] [PubMed]
12. Rivas J, Díaz N, Silva I, et al. KCTD5, a novel TRPM4-regulatory protein required for cell migration as a new predictor for breast cancer prognosis. FASEB J 2020;34:7847-65. [Crossref] [PubMed]
13. Zhao B, Tang X. Identification of anoikis-related genes to develop a risk model and predict the prognosis and tumor microenvironment in rectal adenocarcinoma. Front Genet 2025;16:1604541. [Crossref] [PubMed]
14. Ho TC, LaMere M, Stevens BM, et al. Evolution of acute myelogenous leukemia stem cell properties after treatment and progression. Blood 2016;128:1671-8. [Crossref] [PubMed]
15. Shanbhag VC, Gudekar N, Jasmer K, et al. Copper metabolism as a unique vulnerability in cancer. Biochim Biophys Acta Mol Cell Res 2021;1868:118893. [Crossref] [PubMed]
16. Frischauf I, Zayats V, Deix M, et al. A calcium-accumulating region, CAR, in the channel Orai1 enhances Ca(2+) permeation and SOCE-induced gene transcription. Sci Signal 2015;8:ra131. [Crossref] [PubMed]
17. Samakai E, Hooper R, Martin KA, et al. Novel STIM1-dependent control of Ca2+ clearance regulates NFAT activity during T-cell activation. FASEB J 2016;30:3878-86. [Crossref] [PubMed]
18. Frischauf I, Litviňuková M, Schober R, et al. Transmembrane helix connectivity in Orai1 controls two gates for calcium-dependent transcription. Sci Signal 2017;10:eaao0358. [Crossref] [PubMed]
19. Lin YP, Bakowski D, Mirams GR, et al. Selective recruitment of different Ca(2+)-dependent transcription factors by STIM1-Orai1 channel clusters. Nat Commun 2019;10:2516. [Crossref] [PubMed]
20. Yeh YC, Lin YP, Kramer H, et al. Single-nucleotide polymorphisms in Orai1 associated with atopic dermatitis inhibit protein turnover, decrease calcium entry and disrupt calcium-dependent gene expression. Hum Mol Genet 2020;29:1808-23. [Crossref] [PubMed]
21. Röhrle N, Knott MML, Anz D. CCL22 Signaling in the Tumor Environment. Adv Exp Med Biol 2020;1231:79-96. [Crossref] [PubMed]
22. Xu L, Chen X, Jiang H, et al. NDUFC1 Is Upregulated in Gastric Cancer and Regulates Cell Proliferation, Apoptosis, Cycle and Migration. Front Oncol 2021;11:709044. [Crossref] [PubMed]
23. Zhu Y, Liang S, Pan H, et al. Inhibition of miR-1247 on cell proliferation and invasion in bladder cancer through its downstream target of RAB36. J Biosci 2018;43:365-73. [Crossref] [PubMed]
24. Belgiovine C, D'Incalci M, Allavena P, et al. Tumor-associated macrophages and anti-tumor therapies: complex links. Cell Mol Life Sci 2016;73:2411-24. [Crossref] [PubMed]
25. Komi DEA, Redegeld FA. Role of Mast Cells in Shaping the Tumor Microenvironment. Clin Rev Allergy Immunol 2020;58:313-25. [Crossref] [PubMed]
26. Daver NG, Maiti A, Kadia TM, et al. TP53-Mutated Myelodysplastic Syndrome and Acute Myeloid Leukemia: Biology, Current Therapy, and Future Directions. Cancer Discov 2022;12:2516-29. [Crossref] [PubMed]
27. Hu N, Tian Y, Song Y, et al. The inhibition of Beclin1-dependent autophagy sensitizes PTC cells to ABT737-induced death. Genet Mol Biol 2024;47:e20220170. [Crossref] [PubMed]
28. Xing J, Yang J, Gu Y, et al. Research update on the anticancer effects of buparlisib. Oncol Lett 2021;21:266. [Crossref] [PubMed]
29. Richter LE, Wang Y, Becker ME, et al. HDAC1 Is a Required Cofactor of CBFβ-SMMHC and a Potential Therapeutic Target in Inversion 16 Acute Myeloid Leukemia. Mol Cancer Res 2019;17:1241-52. [Crossref] [PubMed]