The UCSC Xena database (https://xenabrowser.net/datapages/) was utilized to obtain mRNA expression data and clinical information from the TCGA pan-cancer cohort and GTEx datasets. Transcriptomic profiles for 21 cancer cell lines were retrieved from the Cancer Cell Line Encyclopedia (CCLE) (https://sites.broadinstitute.org/ccle/). Genomic alterations of USP37 across 33 types of cancer were analyzed using the cBioPortal for Cancer Genomics tool (http://cbioportal.org). The subcellular localization of the USP37 protein was confirmed through data from the Human Protein Atlas (HPA) (https://www.proteinatlas.org/). The abbreviations for various cancer types are provided in Table 1.

The hallmark gene set file (h.all.v7.4.symbols.gmt), consisting of 50 hallmark gene sets, was acquired from the Molecular Signatures Database (MSigDB, https://www.gsea-msigdb.org/gsea/index.jsp). This file was employed to determine the normalized enrichment score (NES) and false discovery rate (FDR) for differentially expressed genes (DEGs) between groups with high and low USP37 expression across different cancer types. Specifically, significance was assessed using permutation-based FDR. The “high USP37 expression” and “low USP37 expression” subgroups based on the median expression value of USP37 across all samples in the analyzed cohort in GSEA. Gene Set Enrichment Analysis (GSEA) was carried out using the R package “clusterProfiler”, and the findings were illustrated with a bubble plot created using the R package “ggplot2” (Yu et al., 2012).

Using the Sangerbox tool platform (http://past20.sangerbox.com/), we conducted a comprehensive Spearman correlation analysis to assess the statistical relationships between USP37 and established immunotherapy biomarkers across various cancer types. Key metrics, such as tumor mutation burden (TMB) and microsatellite instability (MSI), were included in the analysis, alongside the evaluation of correlations with other immune checkpoint-related genes. This investigation provides insights into USP37’s potential role in the tumor immune microenvironment and aids in assessing its viability as a target for immunotherapy (Shen et al., 2022).

The interaction network data involving USP37 comes from PPI research (https://genemania.org/search/). For prognostic analysis, we used data from UCSC (https://xenabrowser.net/). The standardized pan-cancer dataset, sourced from the TCGA prognostic study, includes overall survival (OS) and disease-specific survival (DSS) data. Using the survival and Survminer R software packages, the Cox regression model and Kaplan-Meier analysis were performed to examine the relationship between USP37 expression and prognosis. The log-rank test was used to compare the high-expression and low-expression subgroups, and a p-value less than 0.05 was considered statistically significant.

We investigated the correlation between the expression of five mismatch repair (MMR) genes—MLH1, MSH2, MSH6, PMS2, and EPCAM—and USP37 using expression profile data from TCGA. These MMR genes are essential for preserving genomic stability and repairing DNA errors. Additionally, DNA methylation, a key epigenetic modification, has a significant impact on gene expression and is intricately linked to tumor development and progression (Martisova et al., 2021). To further understand the role of USP37 in relation to MMR gene expression, we explored their potential mechanisms in tumorigenesis and examined the influence of DNA methylation. Specifically, we analyzed the co-expression of USP37 with five DNA methyltransferases (DNMT1, TRDMT1, DNMT3A, DNMT3B, and DNMT3L) to assess the connection between USP37 and DNA methylation. The results of these analyses were visualized using heatmaps generated by the Sangerbox online platform (http://www.sangerbox.com/tool) (Shen et al., 2022).

Normal pancreatic ductal epithelial cells (H6C7, utilized as controls) and pancreatic cancer cell lines (AsPC-1, BxPC-3, SW 1990, PANC-1) were sourced from the American Type Culture Collection (Manassas, VA, USA). For cell culture, we employed Dulbecco’s Modified Eagle Medium (DMEM) provided by Gibco, Thermo Fisher (China), supplemented with 10% fetal bovine serum and 100 U/mL penicillin-streptomycin from Invitrogen (USA). The cells were grown under standard conditions at 37 °C in a humidified incubator with 5% CO₂ to ensure optimal growth and maintenance.

Total RNA was extracted from cells or tissues using Trizol reagent (Invitrogen, USA). The RNA was then reverse-transcribed into complementary DNA (cDNA) using a reverse transcription kit. Subsequently, cDNA was amplified through quantitative real-time PCR (qRT-PCR) using SYBR Green qPCR Master Mix (Clontech Laboratories, USA), which contains DNA polymerase, primers, and SYBR Green dye. SYBR Green dye binds specifically to double-stranded DNA, emitting fluorescence proportional to the amount of DNA synthesized during amplification. The amplification was conducted on an Applied Biosystems® 7900HT Fast Real-Time PCR System (Thermo Fisher Scientific, USA). The PCR process involved denaturation, annealing, and extension phases. Fluorescence intensity was monitored in real-time during each cycle, enabling the quantification of the target gene relative to a housekeeping gene. Data were analyzed using the comparative Ct (ΔΔCt) method, which compares the threshold cycle (Ct) values of the target gene to those of the reference gene to determine relative expression levels. The following primers were used: for USP37, the forward primer sequence was 5′-TCTCTATTGACCTTCCTCGTAGG-3′ and the reverse primer sequence was 5′-TGCCTGACAAGAGCACACTTCC-3′; for GAPDH, the forward primer sequence was 5′-GTCTCCTCTGACTTCAACAGCG-3′ and the reverse primer sequence was 5′-ACCACCCTGTTGCTGTAGCCAA-3′.

Western blotting was utilized to evaluate the expression levels of target proteins in cell lysates. Cells were first harvested and lysed using RIPA buffer (Thermo Fisher Scientific, USA) supplemented with protease and phosphatase inhibitors (Roche, Switzerland). Protein concentrations in the lysates were determined with the BCA Protein Assay Kit (Pierce, USA). Equal quantities of protein (30 µg) were resolved by SDS-PAGE on 10% polyacrylamide gels and then transferred to PVDF membranes (Millipore, USA) using a semi-dry transfer apparatus (Bio-Rad, USA). The membranes were blocked with 5% non-fat dry milk in TBST (Tris-buffered saline with Tween 20) for 1 h at room temperature to reduce non-specific binding. Membranes were then incubated overnight at 4 °C with primary antibodies specific to the target proteins, diluted in 5% BSA in TBST. After thorough washing, the membranes were exposed to HRP-conjugated secondary antibodies (Thermo Fisher Scientific, USA) for 1 h at room temperature. Protein bands were visualized using ECL Western blotting Substrate (Bio-Rad, USA) and detected with a chemiluminescence imaging system (GE Healthcare, USA). Band intensities were quantified using ImageJ software (NIH, USA). The antibodies used were: USP37 (1:500, Proteintech, 18465-1-AP), α-Tubulin (1:1000, Proteintech, 66031-1-Ig).

The USP37 human shRNA lentivirus and the USP37 overexpression lentivirus were both acquired from Genechem Company (Shanghai, China). The sequences for the USP37 human shRNA lentivirus are as follows: shUSP37#1: 5′-GCAGAAGATGATATTCCAGAA-3′; shUSP37#2: 5′-CGCTTTGCAAACCTGCTTATT-3′; and shNC: 5′-UUCUCCGAACGUGUCACGUTT-3′. For cell infection, target cells were initially seeded in 6-well plates at a density of 2 × 10⁵ cells per well. The following day, the cells were subjected to a 6-h starvation period by replacing the medium with serum-free DMEM to enhance viral uptake. After the starvation period, the lentiviral supernatant containing the virus was added to the cells. The cells were then incubated with the viral supernatant in a humidified incubator at 37 °C with 5% CO2 for 8 h. Subsequently, the viral-containing medium was removed and replaced with a fresh complete medium supplemented with 10% fetal bovine serum to facilitate cell recovery and proliferation. After a total incubation period of 48 h, the cells were selected using puromycin (2 μg/ml, Sigma-Aldrich, USA) to isolate those successfully transduced. Selection was maintained for 1–2 weeks to establish stable cell lines. The efficiency of transduction was assessed through fluorescence microscopy or PCR analysis, depending on the reporter gene used.

After processing, the cells were inoculated into a 6-well plate at a density of 2 × 10³ cells per well and cultured in a complete serum-containing medium for 2 weeks. Following this incubation period, the cells were washed three times with PBS to remove any residual culture medium. Subsequently, the cells were fixed with 4% paraformaldehyde (P1110, Solarbio) for 15 min to preserve cell morphology. The cells were then stained with crystal violet (C8470, Solarbio) for 2 h. Once staining was complete, the cells were washed again with PBS, allowed to air dry, and subsequently photographed and counted under a microscope.

The processed cells were seeded into a 96-well plate at a density of 1 × 10⁴ cells per well. Following inoculation, the cells were incubated under appropriate conditions for 24 h to ensure optimal adhesion and growth. The experiment was then conducted according to the manufacturer’s instructions for the CCK-8 reagent kit. Specifically, CCK-8 reagent was added to each well, and after a defined incubation period, the absorbance was measured at 450 nm using a microplate reader. This procedure assesses cell survival and proliferation, with changes in absorbance indicating variations in cell growth.

To conduct EdU (5-ethynyl-2′-deoxyuridine) detection, treated cells were initially seeded into 6-well plates and cultured to achieve 70%–80% confluence, ensuring stable cell growth during the treatment. EdU solution was then added to the cell culture medium at a concentration of 10 µM and incubated for 6 h at 37 °C in a 5% CO2 environment. This incubation allows EdU to be effectively incorporated into newly synthesized DNA. Following incubation, cells were washed three times with PBS to remove any unbound EdU. The cells were then fixed with 4% paraformaldehyde for 15 min to preserve cellular morphology and subsequently washed three times with PBS. To enhance EdU detection sensitivity, the cells were permeabilized with 0.5% Triton X-100 for 5 min, which removes lipids from the cell membrane and facilitates better penetration of staining reagents. After permeabilization, EdU staining was performed according to the manufacturer’s instructions, with incubation at room temperature for 60 min. Post-staining, the cell nuclei were counterstained with DAPI for 5 min to allow for visual differentiation under a microscope. Cells were then washed with PBS to remove excess staining solution. Finally, images were captured using a microscope, and the percentage of EdU-positive cells was calculated through image analysis to assess cell proliferation.

Tumorigenicity assays and bioluminescence imaging were performed using pancreatic cancer cells that were stably transduced with the firefly luciferase gene, allowing for consistent monitoring of tumor growth through bioluminescent imaging. For the in vivo tumorigenicity experiments, 2 × 10⁶ cells suspended in 100 µL of PBS were subcutaneously injected into the flanks of male BALB/c-nu/nu nude mice. To detect bioluminescent signals in vivo, the mice were anesthetized with isoflurane and imaged using the IVIS Lumina Series III system (PerkinElmer, MA, USA). All animal procedures were approved by the Ethics Committee for Animal Experiments at First Hospital of Putian City.

Pancreatic cancer tissue and adjacent non-tumor tissue samples were collected from patients at the First Hospital of Putian City. In accordance with the Declaration of Helsinki, all participants provided written informed consent. This study’s use of animals was approved by the Ethics Committee of First Hospital of Putian City. (Ethics No. 2024–192). All methods were performed in accordance with the relevant guidelines and regulations.

Tissue sections were initially placed in a 60 °C oven for an hour to dry, followed by cooling to ambient temperature. The paraffin was then removed using a dewaxing solution, and the sections were rehydrated through a series of ethanol gradient solutions. After washing with PBS buffer, endogenous peroxidase activity was blocked by treating the sections with 3% hydrogen peroxide for 10 min. To prevent non-specific binding, the sections were incubated with 5% normal goat serum. Next, the sections were exposed overnight at 4 °C to an anti-USP37 antibody (1:250). Following the antibody incubation, the sections were washed with PBS to eliminate any unbound antibodies. A secondary antibody solution was then applied, and the sections were incubated at room temperature for 30 min. After another PBS wash, the sections were stained with DAB (3,3′-diaminobenzidine) until the desired intensity was achieved. The staining was followed by counterstaining with hematoxylin and dehydration through a gradient of ethanol solutions. Finally, the sections were mounted using a mixture of neutral resin and xylene. Stained sections were examined and photographed under a microscope for further analysis.

All results are expressed as mean ± standard deviation (SD) and were analyzed using GraphPad Prism 8.0, based on at least three independent experiments. Differences between groups were assessed using two-tailed Student's t-tests and analysis of variance (ANOVA). Statistical significance was determined at a p-value of less than 0.05.

This study conducted a systematic analysis of USP37 across various cancer types using several public databases, including TCGA, to investigate its potential role in tumor development and prognosis. The investigation covered gene expression differences among different cancers, examined the relationship between USP37 expression and patient survival, explored its mutation profile, and assessed its link to immune infiltration and key cellular signaling pathways. By integrating these analyses, the study seeks to clarify USP37’s biological functions in cancer progression and its potential as a therapeutic target and prognostic marker. The research workflow is depicted in Figure 1.

We began by examining USP37 gene expression across various tissues using the GTEx dataset, which revealed considerable heterogeneity. The highest expression of USP37 was observed in testicular tissue (Figure 2A). Similarly, CCLE data analysis demonstrated significant variability in USP37 expression among different cell lines (p < 0.001), consistent with the GTEx findings (Figure 2B). To further explore USP37 expression in cancers, we analyzed mRNA levels from TCGA and GTEx datasets across multiple tumor types. High USP37 expression was detected in GBM, LGG, BRCA, ESCA, COAD, STAD, HNSC, KIRC, LUSC, PAAD, LAML, and CHOL, while lower expression was observed in KIRP, SKCM, BLCA, THCA, OV, TGCT, UCS, ACC, and KICH (Figure 2C). Notably, both USP37 mRNA levels were significantly higher in PAAD tissues compared to normal controls (Figure 2D). Immunofluorescence (IF) imaging revealed that the USP37 protein predominantly localized in the nuclei of U2-OS and U-251MG tumor cells, with minimal cytoplasmic expression (Figure 2E). Finally, using the “Pathological Stage Plot” function in HEPIA2, we analyzed the correlation between USP37 expression and pathological stages in ACC, KIRC, LIHC, OV, and PAAD (Figure 2F). In summary, USP37 shows varying expression across tissues and cancers, with higher levels in several tumors, especially PAAD, and is mainly localized in the nucleus of tumor cells.

Analysis of USP37 genomic alterations revealed its mutations in several tumor types. The highest mutation frequency was observed in UCEC, affecting over 6% of patients, with “mutation” being the most frequent alteration, followed by “amplification” (Figure 2G). Figure 2H provides a detailed overview of the types, locations, and frequencies of USP37 genetic alterations. Of the 152 identified genetic variations, “Missense” mutations were the most prevalent, particularly the A491V/X491_splice mutation in the UCH domain. The X491_splice mutations, found in cancers such as GBM and LUSC led to a splice mutation at position 491 of the USP37 protein. Additionally, A491V mutations were identified in UCEC and COAD, resulting in a missense mutation at the same position within the USP37 gene (Figure 2I).

We utilized the Kaplan-Meier method along with univariate Cox regression analysis to assess the prognostic significance of USP37 in a pan-cancer context. The forest plot (Figure 3A) indicated that lower USP37 expression was linked to improved overall survival (OS) in certain cancers, including ACC (HR = 1.95, 95% CI [1.35–2.82], p = 0.00046), KIRP (HR = 1.29, 95% CI [1.03–1.62], p = 0.02), LGG (HR = 1.45, 95% CI [1.07–1.97], p = 0.02) and LIHC (HR = 1.19, 95% CI [1.03–1.37], p = 0.02). In contrast, higher USP37 expression was associated with better OS in KIRC (HR = 0.83, 95% CI [0.74–0.93], p = 0.00099). The analysis also extended to disease-specific survival (DSS), where reduced USP37 expression corresponded to better DSS outcomes in ACC (HR = 1.97, 95% CI [1.35–2.89], p = 0.00058), LGG (HR = 1.47, 95% CI [1.07–2.02], p = 0.02), and LIHC (HR = 1.22, 95% CI [1.02–1.46], p = 0.03). Conversely, USP37 overexpression was related to worse DSS in KIRC (HR = 0.78, 95% CI [0.69–0.90], p = 0.00032) (Figure 3B). Kaplan-Meier survival curves for ACC, LGG, LIHC, KIRC showed that higher USP37 expression correlated with poorer OS in ACC, LGG, and LIHC, while being associated with better OS in KIRC (Figure 3C). Similarly, elevated USP37 expression was linked to shorter DSS in ACC, KIRP, and LGG, but prolonged DSS in KIRC patients (Figure 3D). Univariate and multivariate Cox regression analyses of overall survival are presented in Table 2. Overall, these results suggest that USP37 may serve as a prognostic predictor across various cancers.

To identify USP37-associated cancer hallmarks, we performed gene set enrichment analysis (GSEA) using differentially expressed genes (DEGs) between low and high USP37 expression subgroups for each cancer type. Our analysis revealed that USP37 expression was significantly associated with oxidative phosphorylation, myogenesis, mitotic spindle, G2M checkpoint, E2F targets, and allograft rejection pathways. In the pathway enrichment analysis, USP37 expression was found to be positively correlated with key cell cycle-related pathways, including the mitotic spindle, G2M checkpoint, and E2F targets in most majority of cancers (Figure 4).

MMR (mismatch repair) is an essential cellular process that ensures the correction of DNA replication errors, maintaining genetic stability in all organisms. When this repair mechanism fails, it can result in mutations and genetic abnormalities, contributing to the development of various diseases, particularly cancer (Pecina-Slaus et al., 2020; Olave and Graham, 2022; Emiloju et al., 2023). A study was conducted to investigate the relationship between USP37 expression levels and mutations in five key MMR genes to explore the role of USP37 in tumorigenesis. The findings demonstrated a significant positive association between USP37 and MLH1, MSH2, MSH6, PMS2, and EPCAM across most cancer types. However, it is important to note that, compared to the other four MMR genes, USP37 exhibits a weaker positive correlation with EPCAM and shows no correlation in certain tumors, including ACC, DLBC, ESCA, GBM, LGG, LUAD, MESO, SARC, SKCM, TGCT, UCS, and UVM (Figure 5A). Abnormal DNA methylation is one of the key factors in tumor development, which promotes tumor progression by regulating gene expression. It can lead to silencing of tumor suppressor genes or activation of oncogenes, thereby promoting cell proliferation, inhibiting apoptosis, and enhancing invasiveness. Therefore, DNA methylation, as a common epigenetic variation, has become an important biomarker for cancer diagnosis, treatment, and prognosis evaluation (Papanicolau-Sengos and Aldape, 2022; Zhu et al., 2023). The results indicated that USP37 was significantly associated with at least one of the four methyltransferase genes (DNMT1, DNMT2, DNMT3A, and DNMT3B) in various cancer types, except UCS (Figure 5B).

Next, we utilized the GeneMANIA online program to create a PPI network for USP37, which is displayed in Figure 5C, to investigate the probable processes by which USP37 played a role in cancer carcinogenesis. USP37 demonstrated significant physical interactions with KRT84, CDT1, FZR1, and ALB, as illustrated in Figure 5C. Research has shown that reducing CDT1 expression can significantly inhibit the malignant behavior of HCC cells (Yao et al., 2023), which indicates that USP37 may influence tumor progression by modulating the activity of CDT1 and other interacting proteins, requiring further investigation.

To explore the mechanisms by which USP37 influences the tumor immune microenvironment, public single-cell RNA-seq datasets were analyzed to assess USP37 expression across various immune cell types. The heatmap revealed that USP37 is expressed in immune cells, particularly T lymphocytes—including CD4 Tconv, Treg, CD8 T, and CD8 Tex cells—followed by innate immune cells such as NK cells, dendritic cells (DCs), and monocytes/macrophages. The expression and distribution of USP37 across different immune cells in various tumors are shown in Supplementary Figure S1. Specifically focusing on gliomas and considering the level of USP37 expression, we demonstrated its distribution in immune cells. Our analysis indicated that in gliomas_GSE102130, USP37 is primarily expressed in Ac-like malignant cells, OC-like malignant cells, OPC-like malignant cells, oligodendrocytes, and monocytes/macrophages (Figures 5D–F). In gliomas_GSE1311928_Smartseq2, USP37 is expressed in Ac-like malignant cells, CD8Tex, MES-like malignant cells, monocytes/macrophages, NPC-like malignant cells, OPC-like malignant cells, and oligodendrocyte (Figures 5G–I). This implies that USP37 might play a role in regulating the interactions between immune cells and tumor cells, potentially affecting glioma progression and immune evasion. Nevertheless, additional research is needed to clarify the specific functional roles of USP37 in these cellular processes.

Figure 6A illustrates the associations between USP37 and 47 immune regulators across various cancer types. Our analysis revealed that USP37 exhibits a strong positive correlation with most immune regulators in cancers such as HNSC, KIRC, LIHC, PRAD, and READ while showing a strong negative correlation with most immune regulators in SARC. Additionally, USP37 demonstrated a significant positive relationship with ADORA2A, CD160, TNFSF15 and NRP1 across the majority of cancers. To assess USP37’s potential as a predictor of immune checkpoint inhibitor (ICI) efficacy, we evaluated the correlation between USP37 expression and both TMB and microsatellite instability (MSI).

Positive correlations between tumor mutation burden TMB and USP37 expression were observed in UCEC, THYM, STAD, SKCM, LUAD, LAML, and COAD. In contrast, negative correlations were identified in THCA, LIHC, KIRP, and BRCA (Figure 6B). For microsatellite instability (MSI), positive associations with USP37 expression were found in UCEC, READ, MESO, COAD, and CESC, while negative correlations were observed in DLBC, THCA, SKCM, PRAD, HNSC, DLBC, and BRCA (Figure 6C). These results suggest that USP37 could be a potential predictor of ICI efficacy in these cancers. Furthermore, the top three tumors showing the strongest correlations with USP37 expression were KIRC, SARC, and HNSC based on the StromalScore; SARC, UCEC, and BRCA based on the Estimated Immune Score; and SARC, UCEC, and TGCT based on the ESTIMATE Score (Figure 7). Overall, these findings highlight a robust association between USP37 expression and immune infiltration levels in these cancers.

Next, we performed qRT-PCR and Western blot assays on clinical PAAD samples to validate the expression levels of USP37 in comparison to adjacent non-tumor tissues. The results showed that USP37 mRNA and protein expression levels are significantly increased in PAAD, consistent with the bioinformatics analysis. (Figures 8A–C). Additionally, IHC analyses demonstrated that USP37 was highly expressed in PAAD tissues (Figures 8D,E). These findings confirmed the elevated expression of USP37 in PAAD tissues, supporting its potential role in tumor progression and aligning with bioinformatics predictions.

The consistent overexpression of USP37 in PAAD and cell lines prompted us to investigate its potential oncogenic role in PAAD. To explore the role of USP37 in pancreatic cancer (PC) cell proliferation, we conducted a series of assays. First, we performed qRT-PCR and Western blot analyses to measure the mRNA and protein levels of USP37 in the human pancreatic duct epithelial cell line H6C7 and four PC cell lines: AsPC-1, BxPC-3, SW 1990, and PANC-1. The results showed that USP37 expression was highest in SW1990 cells and lowest in PANC-1 cells compared to H6C7 (Figures 9A,B, Supplementary Figure S2A). We then used USP37-targeting shRNA vectors to knock down USP37 expression in SW1990 cells, which exhibit high endogenous levels of USP37 (Figures 9C,D, Supplementary Figure S2B). Conversely, we generated USP37-overexpressing PANC-1 cells, which have relatively low endogenous USP37 levels (Figures 9E,F, Supplementary Figure S2C). Silencing USP37 in SW1990 cells significantly suppressed cell growth (Figure 9G), while USP37 overexpression in PANC-1 cells significantly enhanced cell proliferation, as demonstrated by CCK-8 assays (Figure 9H). Consistent with these findings, colony formation and EdU assays revealed that USP37 knockdown reduced cell proliferation in SW1990 cells (Figures 9I,L), whereas USP37 overexpression increased cell proliferation in PANC-1 cells (Figures 9J,K). Consistently, RTCA data indicated that USP37 knockdown significantly inhibited the proliferation of SW1990 cells, whereas its overexpression enhanced the proliferation of PANC-1 cells (Figures 9M,N). In conclusion, these findings suggest that USP37 can promote the proliferation of PC cells in vitro.

Finally, to evaluate the effect of USP37 on tumorigenesis in vivo, we performed subcutaneous xenograft assays in nude mice using USP37-knockdown SW1990 cells. After 4 weeks of growth, a significant reduction in tumor size and weight was observed (Figure 9O). In contrast, USP37-overexpressing PANC-1 cells showed markedly increased tumor growth and weight compared to controls (Figure 9P). Collectively, these data indicate that USP37 plays a crucial oncogenic role in promoting PC cell growth and tumorigenicity.