Glutamine-driven upregulation of NPDC1 promotes colorectal cancer progression through PI3K/AKT signaling

Abstract

Background

Colorectal cancer (CRC) ranks as one of the leading causes of cancer-related mortality globally. NPDC1 is a novel regulator involved in cell proliferation and is upregulated in CRC. However, the biological function and mechanism of NPDC1 driving CRC progression have not been investigated.

Methods

We integrated single-cell RNA-seq data and bulk RNA-seq cohorts to identify prognostic epithelial gene clusters. The R package “ClusterGVis” was employed to categorize six distinct gene clusters within epithelial cells, following Cox regression identifying poor prognosis genes (HR > 1) in the C1 cluster showing progressive upregulation across the four stages. NPDC1 expression was validated by quantitative real-time polymerase chain reaction (qRT-PCR), immunohistochemistry (IHC) and immunofluorescence (IF). Functional impacts on proliferation, metastasis, and immune microenvironment were assessed using CCK8 assays, EdU staining, colony formation, transwell assays and flow cytometry. Additionally, gene set enrichment analysis (GSEA) based on KEGG terms was performed to investigate the potential signaling pathways and biological functions associated with NPDC1 in CRC. The regulatory role of NPDC1 in tumor progression was assessed establishing subcutaneous xenograft tumor model and lung metastasis model of mouse CRC.

Results

NPDC1 is significantly upregulated in KRAS mutant CRC and correlates with poor prognosis. Functional experiments demonstrated that NPDC1 drives CRC proliferation in vitro and in vivo but does not affect apoptosis, migration, or invasion. Mechanistically, KRAS mutation-induced glutamine metabolism elevates NPDC1 expression via JUND, activating the PI3K-AKT pathway to promote tumor growth independently of immune modulation.

Conclusion

Collectively, our results reveal NPDC1 as a KRAS-glutamine axis effector that specifically regulates CRC proliferation via PI3K/AKT signaling, suggesting that NPDC1 could serve as a potential therapeutic target for CRC treatment, particularly in KRAS mutant CRC.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-026-07733-x.

Introduction

According to the latest global cancer statistics, colorectal cancer (CRC) remains one of the leading causes of cancer-related morbidity and mortality worldwide [1]. Although significant progress has been made in CRC treatment, including surgical resection, chemotherapy, radiotherapy, and targeted therapies, severe challenges persist in clinical practice [2]. Specifically, metastasis, recurrence, and treatment resistance in advanced CRC patients result in a low overall survival rate [3], highlighting the urgent need for novel molecular targets to improve patient outcomes.

Oncogene and tumor suppressor gene alterations critically drive CRC development. Landmark studies have identified several key oncogenic drivers in CRC, such as KRAS and BRAF, alongside tumor suppressors like APC and TP53, whose mutations disrupt cellular proliferation, apoptosis, and DNA repair mechanisms [4, 5]. Notably, significant progress has been made in the clinical application of targeted therapies aimed at oncogenes. For example, KRAS G12C inhibitors such as adagrasib and sotorasib have emerged as breakthrough agents for the treatment of KRAS mutant solid tumors, especially CRC and non-small cell lung cancer (NSCLC) [6]. In addition, the TP53 mutant-targeted drug Rezatapopt, which selectively binds to the cleft in TP53 Y220C mutant proteins, can restore normal p53 conformation and transcriptional activity, thus exerting effective antitumor effects [7]. Therefore, identifying new oncogenes remains critical for advancing cancer therapy.

NPDC1 (neural proliferation, differentiation and control 1), initially identified in the nervous system, regulates neuronal proliferation and differentiation. NPDC1 can directly interact with the transcription factors E2F-1 and DP-1, as well as cyclins D1, D2, D3, and cdk2, indicating that NPDC1 may regulate the cell cycle and cell proliferation [8]. Recent research has revealed that aberrant NPDC1 expression is linked to Alzheimer’s disease and schizophrenia through the regulation of neural cell proliferation and differentiation [9, 10]. In addition, NPDC1 has significant oncogenic relevance. NPDC1 is associated with the recurrence of acute myeloid leukemia (AML) and is highly expressed in CRC, where it is correlated with perineural invasion and poor prognosis [11–13]. However, the upstream regulators controlling NPDC1 expression and its downstream effector mechanisms in tumor progression remain undefined. Therefore, further exploration of the underlying mechanisms involved is conducive to a deeper understanding of the tumor regulatory effect of NPDC1 and provides an important theoretical basis for tumor treatment targeting NPDC1.

Through scRNA transcriptomic profiling analysis, we identified NPDC1 as being progressively upregulated throughout the progression from normal tissue to inflammatory bowel disease (IBD) and eventually to CRC. We found that the expression of NPDC1 in CRC was correlated with KRAS mutations. Glutamine, a KRAS-associated metabolite [14], potently upregulates NPDC1 in MC38 cells via JUND. Furthermore, NPDC1 appears to drive tumor progression primarily through PI3K-AKT-mediated cell proliferation rather than through immunomodulatory mechanisms. Overall, our research initially clarified the role and molecular mechanism of NPDC1 in the progression of CRC, establishing therapeutic support for NPDC1-targeted intervention.

Results

Single-cell atlas reveals NPDC1 as a key regulator driving the progression in CRC

To investigate novel key regulators of CRC progression, we collected and analyzed published single-cell RNA-sequencing data 47 samples: healthy colorectal tissues, non-inflammatory and inflamed tissues from patients with ulcerative colitis (UC), and CRC tissues. After the gene expression matrix was integrated into a Seurat object, cells with more than 7,000 expressed genes or containing more than 20% mitochondrial genes were excluded, and a high-quality colorectal tissue map consisting of 117,978 single cells was constructed. Harmony was utilized to correct batch effects across various samples. Dimensionality reduction and clustering identified ten major cell types: epithelial cells, B cells, endothelial cells, fibroblasts, myeloid cells, CD8⁺ T/NK cells, CD4⁺ T cells, plasma IgA-producing cells, plasma IgG-producing cells, and glial cells (Fig. 1A-C). To explore stage-specific gene expression changes in epithelial cells, which constitute the major component of tumor tissue, we performed single-cell subset analysis, and volcano plots highlighted significantly differentially expressed genes across inflammation-to-tumor transition stages (Fig. 1D). The expression trend clustering of epithelial genes revealed six distinct clusters, with cluster 1 genes showing progressive upregulation across the four stages (Fig. 1E). Cox regression analysis of TCGA-COAD survival data revealed that NPDC1 and TIMP1 were significant prognostic markers (HR > 1) (Fig. 1F). Given that the role and mechanism of TIMP1 in tumor progression have been clarified [15], we focused on NPDC1 to explore its role and mechanism in the progression of CRC (Fig. 1G).

Fig. 1.

Single-cell atlas reveals NPDC1 as a key regulator driving the progression of CRC. (A) Schematic workflow of NPDC1 screening in colorectal cancer: Integrated scRNA-seq datasets (SCP259, GSE132465) from 10 healthy controls, 7 inflammatory bowel disease (IBD) patients (non-inflamed/inflamed), and 23 CRC patients underwent initial cell clustering, epithelial subset extraction, differential expression analysis, gene module identification, and Cox regression of the C1 module to identify NPDC1 as a key prognostic driver. (B) UMAP plot showing the tissue origins of the cells: healthy (n = 10), IBD noninflamed (n = 7), IBD inflamed (n = 7), and CRC (n = 23) (Left). UMAP analysis of major cell types across 117,978 cells from these 47 samples (right). (C) Dot plot showing the expression levels of marker genes across distinct cell types. (D) Volcano plot showing DEGs of epithelial cells across health-to-CRC transition stages. (E) Heatmap showing six gene clusters (C1–C6) identified in epithelial cells via unsupervised clustering. (F) Forest plot of Cox regression results identifying NPDC1 and TIMP1 as key survival regulators in the C1 gene cluster. (G) Quantification of NPDC1⁺ epithelial cells across healthy, IBD, and CRC stages

NPDC1 upregulation in CRC correlates with poor prognosis

To clarify the expression level of NPDC1 in tumors, pan-cancer analysis via the TIMER 2.0 database revealed significant upregulation of NPDC1 in five cancer types compared with matched normal tissues, especially CRC (Fig. 2A). Additionally, paired-sample analyses of the TCGA-COAD and GSE8671 datasets revealed significantly higher NPDC1 expression in CRC tissues compared to adjacent normal tissues (Fig. 2B-C). The survival analysis of the TCGA-COAD data indicated that patients with high expression levels of NPDC1 had a poorer prognosis. (Fig. 2D). Consistent results were observed in GSE143985 from the K-M plotter database (Fig. 2E). Furthermore, to verify the high expression of NPDC1 in CRC, we collected several paired CRC and adjacent tissue samples and conducted comprehensive analyses. Real-time PCR and immunohistochemistry (IHC) assays consistently demonstrated NPDC1 overexpression in CRC (Fig. 2F-G). Immunofluorescence staining of consecutive clinical sections for Ki67 and NPDC1 showed significant co-localization (Fig. 2H), further confirming simultaneous upregulation of NPDC1 in cancer tissues relative to adjacent normal tissues.

Fig. 2.

NPDC1 upregulation in CRC is correlated with poor prognosis. (A) Expression of the NPDC1 gene across various cancer types based on analysis of the TIMER2.0 database. (B-C) Paired-sample analysis of NPDC1 in the TCGA-COAD dataset (n=41 paired samples) and GSE8671 dataset (n=31 paired samples). (D-E) Survival analysis of NPDC1 expression in CRC was performed via TCGA-COAD (D) and GSE143985 from Kaplan‒Meier plotter (E). Patients were stratified by median NPDC1 expression into high and low groups. (F) Real-time PCR quantification of NPDC1 mRNA levels in 4 paired CRC and adjacent normal tissue samples. (G) Representative IHC images of NPDC1 in CRC and adjacent normal tissues from patients, with statistical results of H-score quantification (n = 5 per group, paired t test). (H) Representative images of immunofluorescence showing the expression of NPDC1 and Ki67 in CRC tissue, two consecutive sections from the same patient. Scale bars, 50μm. Data are presented as mean±SD (*p < 0.05, **p < 0.01, ***p < 0.001, and ns, not significant; SD, standard deviation; paired-sample Wilcoxon test for B and C, paired-sample t test for F and G)

NPDC1 promotes the proliferation of MC38 cells in vitro

To further explore the role of NPDC1 in CRC, we found that NPDC1 is closely related to various cell biological activities on the basis of scoring of 14 characteristic tumor gene sets (Fig. 3A). Subsequently, we established stable-expressing MC38 cells with knockdown NPDC1 gene (named sh-NPDC1 cells) and cells with overexpressed NPDC1 gene (named OE-NPDC1 cells) to further verify their effects on cellular biological activities (Fig. 3B-C). CCK-8 proliferation assay demonstrated that sh-NPDC1 markedly impaired the proliferative capacity of MC38 cells (Fig. 3D), which was corroborated by reduced proportions of EdU-positive cells and decreased colony formation efficiency (Fig. 3F, H). Conversely, OE-NPDC1 enhanced both cellular proliferative potential and colony formation capacity (Fig. 3E, G, I). However, neither sh-NPDC1 nor OE-NPDC1 did not affect MC38 cells apoptosis, migration or invasion (Fig. 3J-M). Collectively, these findings demonstrate that NPDC1 primarily affects the proliferative potential of MC38 cells.

Fig. 3.

NPDC1 promotes the proliferation of MC38 cells in vitro. (A) The tumor hallmark scores of NPDC1-positive and NPDC1-negative Epi cells. The AddModuleScore algorithm was used to score hallmark gene sets obtained from the CancerSEA database. (B) Western blotting and real-time PCR were used to assess the knockdown efficiency of NPDC1. (C) Western blotting and real-time PCR were used to assess the overexpression efficiency of NPDC1. (D) CCK-8 assay of sh-NPDC1 MC38 cells. (n = 5 per group). (E) CCK-8 assay of OE-NPDC1 MC38 cells. (n = 5 per group). (F) EdU incorporation assay of sh-NPDC1 MC38 cells (left). Quantification of the results of the EdU incorporation assay (right). Scale bar: 50 μm. (n = 3 per group). (G) EdU incorporation assay of OE-NPDC1 MC38 cells (left). Quantification of the results of the EdU incorporation assay (right). Scale bar: 50 μm (n = 3 per group). (H) Colony formation of sh-NPDC1 MC38 cells (left). Quantification of colony formation (right) (n = 3). (I) Colony formation of OE-NPDC1 MC38 cells (left). Quantification of colony formation (right) (n = 3 per group). (J) Apoptosis rate of NC and sh-NPDC1 MC38 cells (n = 4 per group). (K) Apoptosis rate of Vector and OE-NPDC1 MC38 cells (n = 4 per group). (L) Representative images of the migration and invasion of sh-NPDC1 MC38 cells (left). Quantification of migration and invasion (right). sh-NPDC1 or control MC38 cells were seeded in serum-free medium (5*10⁴/chamber, upper chamber, 8 μm pore membrane) and allowed to migrate toward 10% FBS for 24 h. Non migrated cells were removed, and migrated cells were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet, and imaged. Scale bar: 75 μm (n = 3 per group). (J) Representative images of the migration and invasion of OE-NPDC1 MC38 cells, methodology as described in (L) (left). Quantification of migration (right). Scale bar: 75 μm (n = 3 per group). Data are presented as mean ± SD. (*p < 0.05, **p < 0.01, ***p < 0.001, and ns, not significant; SD, standard deviation; Wilcoxon test for A, Student’s t test for all others)

NPDC1 accelerates tumor growth in vivo

Given that NPDC1 can directly regulate cell proliferation ability, we further explored its regulatory effect on tumor growth. We found that sh-NPDC1 significantly inhibited the growth of subcutaneous tumors in mice. (Fig. 4A), whereas OE-NPDC1 accelerated tumor progression (Fig. 4B). However, neither sh-NPDC nor OE-NPDC1 affected metastatic progression, as determined by the number of metastatic nodules (Fig. 4C, E) and H&E-stained images (Fig. 4D, F).

Fig. 4.

NPDC1 promotes the growth of CRC in vivo. (A) Images of subcutaneous tumors derived from MC38 cells transfected with NC or sh-NPDC1 lentivirus. 6–8 weeks male C57BL/6 mice (n = 5 per group) were injected subcutaneously into the right groin with 1 × 10⁶ cells resuspended in 100µL PBS. Tumors were harvested 14 days post-injection(left). Quantification of tumor weight and volume. The tumor volume was measured at harvest via the following formula: volume = (length × width²)/2. (B) Images of subcutaneous tumors derived from MC38 cells transfected with Vector or OE-NPDC1 lentivirus, methodology as described in (A) (left). Quantification of tumor weight and volume, as described in (A) (right). (C) Representative images of distal lung metastases from 6-8-week-old male C57BL/6 mice injected intravenously with NC or sh-NPDC1 MC38 cells (1 × 10⁶ cells in 100µL of PBS). The mice (n = 5 per group) were sacrificed on day 14 postinjection(left). Quantification of lung metastasis nodules counted under a stereomicroscope (right). (D) Representative HE images of the lung metastases from (C). Scale bar, 200 μm (left); Scale bar, 50 μm (right). (E) Representative images of distal lung metastases from mice injected intravenously with Vector or OE-NPDC1 MC38 cells (1 × 10⁶ cells in 100µL of PBS)0.6–8 weeks male C57BL/6 mice (n = 5 per group) were sacrificed on day 14 postinjection(left). Quantification of lung metastasis nodules counted under a stereomicroscope (right). (F) Representative HE images of the lung metastases from (E). Scale bar, 200 μm (left); Scale bar, 50 μm (right). Data are presented as mean ± SD (**p < 0.01, ***p < 0.001, and ns, not significant; SD, standard deviation; Student’s t test)

KRAS mutant-mediated glutamine metabolism upregulates the expression of NPDC1 in CRC cells

Given the high expression of NPDC1 and its tumor regulatory effect, we further explored specific mechanisms that drive high NPDC1 expression in CRC. scRNA-seq analysis of epithelial cell clusters (Fig. 5A-B) and bubble plots showed NPDC1 was predominantly upregulated in CMS3 epithelial cells (Fig. 5C), which was confirmed in TCGA-COAD and GSE39582 bulk cohorts (Fig. 5D-E). Since KRAS mutations are linked to CMS3 CRC [16], we found significantly higher NPDC1 expression in KRAS mutant samples from the two cohorts (Fig. 5F-G). TCGA-COAD mutational landscape further showed higher KRAS mutation frequency in the high-NPDC1 group (sFig. 1 A). Stratified survival analysis revealed high NPDC1 correlated with worse outcomes in KRAS mutant patients (Fig. 5H), but not in KRAS wild-type patients (Fig. 5I), implying KRAS mutation may drive NPDC1 upregulation.

Fig. 5.

KRAS mutant-mediated glutamine metabolism promotes the expression of NPDC1 in CRC cells. (A) UMAP plot depicting unsupervised clustering of epithelial cells from integrated scRNA-seq datasets (SCP259, GSE132465), with cell type annotations following the original dataset authors’ definitions. (B) UMAP plot of NPDC1 expression in Epi cells. (C) The proportion of Epi cells expressing NPDC1 and the normalized expression levels of positive cells in each Epi cell type. (D) Violin plot of NPDC1 mRNA expression in the GSE39582 dataset stratified by CMS classification (CMS1: n = 84, CMS2: n = 167, CMS3: n = 88, CMS4: n = 157). (E) Violin plot of NPDC1 mRNA expression in the TCGA-COAD cohort stratified by the predicted CMS classification (CMS1: n = 82, CMS2: n = 128, CMS3: n = 68, CMS4: n = 126). (F) Violin plot showing NPDC1 mRNA expression in the GSE39582 dataset grouped by KRAS mutation status (Mut, mutant; WT, wild-type; Mut: n = 213; WT: n = 322). (G) Violin plot showing NPDC1 mRNA expression in TCGA-COAD samples grouped by KRAS mutation status (Mut, mutant; WT, wild-type; Mut: n = 174; WT: n = 164). (H) Overall survival analysis within KRAS mutant patients in the TCGA-COAD cohort. Patients were stratified by median NPDC1 expression into high and low groups. (I) Overall survival analysis within KRAS wild-type patients in the TCGA-COAD cohort. Patients were stratified by median NPDC1 expression into high and low groups. (J) Representative western blot showing NPDC1 protein levels in MC38 cells treated with glutamine at 200µM for the indicated time points. (K) Representative western blot showing NPDC1 protein levels in the MC38 cells treated with glutamine inhibitor(V-9302) at 20µM for 48 h. (L) CCK-8 assay of sh-NPDC1 MC38 cells treated with glutamine at 200µM (n = 5 per group). (M) GSEA revealed that L-glutamine import gene sets were significantly upregulated in the KRAS mutant samples relative to the wild-type samples in the TCGA-COAD cohort. (N) GSEA revealed that glutamine transport gene sets were significantly upregulated in the KRAS mutant samples relative to the wild-type samples in the GSE39582 cohort. (O) Representative western blot showing the KRASG12D -induced upregulation of NPDC1 was abolished in MC38 cells treated with glutamine inhibitor(V-9302). (P) Venn diagram depicting overlapping transcription factors predicted to regulate NPDC1. Predictions were integrated from five databases (ChIP-Atlas, hTFtarget, KnockTF, ENCODE, JASPAR) using the TFTF R package. (Q) Pearson’s correlation of NPDC1 with JUND expression in epithelial cells. Analysis based on integrated scRNA-seq data (corresponding to Fig. 5A). Gene expression values were averaged across epithelial cells per sample prior to Pearson correlation analysis. (R) Pearson’s correlation of JUND expression with glutamine transport activity in epithelial cells. Integrated scRNA-seq data (corresponding to Fig. 5A) were analyzed using the AddModuleScore algorithm to quantify “GOBP_GLUTAMINE_TRANSPORT” activity. Per-sample epithelial cell averages were subjected to Pearson correlation analysis. (S) The sequence logo of JUND obtained from the JASPAR database (top). Predicted JUND binding site in the promoter region of NPDC1 in the JASPAR database (bottom). (T) Real-time PCR quantification of NPDC1 mRNA levels in the MC38 cells treating combined glutamine (200µM) and JUND inhibitor (T-5224, 10µM). Data are presented as mean ± SD (***p < 0.001, and ns, not significant; SD, standard deviation; Wilcoxon test for D-G, Student’s t test for S)

Considering that the tumor microenvironment (TME) can also regulate gene expression, we exposed MC38 cells to several representative TME conditions: hypoxia, high glutamine, and high lactate. Interestingly, only glutamine treatment progressively upregulated NPDC1 at both the mRNA and protein levels (Fig. 5J, sFig. 1B-C). In addition, glutamine inhibitor (V-9302) significantly downregulates the expression of NPDC1(Fig. 5K, sFig. 1D). Concurrently, V-9302 treatment also effectively suppressed cellular proliferation (sFig. 1E). Furthermore, the knockdown of NPDC1 rescued the glutamine-induced increase in NPDC1 expression (sFig. 1 F) and the associated acceleration of cell proliferation (Fig. 5L).

Coincidentally, KRAS mutation is significantly associated with glutamine metabolism [14], and GSEA enrichment analysis based on GO terms also revealed that glutamine transport pathway was significantly enriched in KRAS mutant tumors across both bulk cohorts (Fig. 5M-N). Functional validation via transfection of KRAS^wt or KRAS^G12D plasmids into MC38 cells demonstrated V-9302 rescues KRAS^G12D-induced NPDC1 upregulation and effect in promoting cell proliferation (Fig. 5O, sFig.1G-H).

To identify the signaling pathways and transcriptional regulators mediating glutamine-induced NPDC1 upregulation, we performed multi-database screening for transcription factors associated with NPDC1 expression. The results revealed JUND as a potential transcriptional regulator of NPDC1 expression (Fig. 5P-Q). Furthermore, glutamine transport also showed a strong positive correlation with JUND expression level. (Fig. 5R). Using the JASPAR database, we predicted a putative JUND binding site within − 1460 to -1450 of the NPDC1 promoter (Fig. 5S). In addition, glutamine was found to enhance JUND expression in MC38 cells in a time-dependent manner (sFig. 1I). More importantly, JUND inhibitor (T-5224) reverses the effect of glutamine in upregulating NPDC1(Fig. 5T). Collectively, these data indicate KRAS mutation-enhanced glutamine metabolism mediates NPDC1 upregulation in CRC via transcription factor JUND.

NPDC1 promotes tumor growth via the PI3K/AKT pathway

To further explore the potential mechanism by which NPDC1 regulates cell proliferation and tumor progression, we conducted differential analysis of 11 oncogenic pathways across two independent bulk RNA-seq cohorts. These findings revealed that the PI3K/AKT, NOTCH, MYC and RTK/RAS pathways were significantly upregulated in the NPDC1-high group (Fig. 6A-D). Functional assessment revealed that among inhibitors targeting PI3K/AKT, NOTCH, MYC, or RTK/RAS signaling, only PI3K/AKT pathway inhibition (Alpelisib) was uniquely effective in suppressing NPDC1-driven cell proliferation (Fig. 6E, sFig. 2 A-D). In addition, PI3K/AKT pathway activation was observed in both the NPDC1-positive Epi and KRAS mutant Epi populations (Fig. 6F, sFig. 2E), indicating that PI3K/AKT signaling drives NPDC1-mediated tumor proliferation and growth acceleration. We also found that NPDC1 knockdown significantly inhibits the protein levels of p-AKT but not AKT and PI3K (Fig. 6G). The PI3k inhibitors (Alpelisib and Buparlisib) did not alter the total protein levels of NPDC1, PI3K, or AKT in NPDC1 overexpression cells, but both significantly suppressed AKT phosphorylation (Fig. 6H). Furthermore, Buparlisib, another PI3K inhibitor, and AKT silencing similarly abrogated the pro-proliferative effect induced by NPDC1 overexpression (Fig. 6I-L). Similarly, Buparlisib also eliminated the tumor growth acceleration mediated by NPDC1 overexpression (Fig. 6M). These consistent findings across both pharmacological and genetic interventions confirm that NPDC1 promotes proliferation primarily through activation of the PI3K/AKT pathway.

Fig. 6.

NPDC1 drives tumor proliferation via the PI3K/AKT pathway. (A) Limma differential analysis of 11 oncogenic pathway scores in the TCGA-COAD cohort between the NPDC1-high and NPDC1-low groups. (B) Histogram of correlation analysis between PI3K/AKT, NOTCH, MYC and RTK/RAS molecules and NPDC1 expression in the TCGA-COAD cohort. (C) Limma differential analysis of 11 oncogenic pathway scores in the GSE39582 cohort between the NPDC1-high and -low groups. (D) Histogram of correlation analysis between PI3K/AKT, NOTCH, MYC and RTK/RAS molecules and NPDC1 expression in the GSE39582 cohort. (E) CCK-8 proliferation assay of OE-NPDC1 MC38 cells treated with PI3K inhibitor (Alpelisib).MC38 cells were seeded at 1 × 10³ cells/well and treated with Auparlisib (10µM) (n = 5 per group). (F) KEGG pathways enriched in NPDC1-positive Epi cells according to the integrated scRNA-seq data (Fig. 5A). NPDC1-positive cells (n = 18,217) were defined by NPDC1 normalized expression > 0. Differentially expressed genes (DEGs) were identified via Seurat v4 (log2FC > 0.25, FDR < 0.05), and KEGG enrichment was performed via the gseKEGG function ofClusterProfiler package. (G) Representative western blot showing knockdown NPDC1 reduced activation of p-AKT. (H) Representative western blot showing the PI3K inhibitors (Alpelisib and Buparlisib) reverse the activation of p-AKT induced by NPDC1 overexpression. MC38 cells were seeded at 5 × 104 cells/well and treated with Buparlisib or Auparlisib (10µM) for 48 h. (I) CCK-8 proliferation assay of OE-NPDC1 MC38 cells treated with PI3K inhibitor (Buparlisib).MC38 cells were seeded at 1 × 10³ cells/well and treated with Buparlisib (10µM) (n = 5 per group). (J) EdU incorporation assay of OE-NPDC1 MC38 cells with PI3K inhibitors (Buparlisib) (n = 3 per group). (K) Representative western blot showing knocking down AKT reverses the activation of p-AKT induced by NPDC1 overexpression. MC38 cells were seeded at 3 × 10⁴ cells/well and transfected with siAKT for 48 h. (L) CCK-8 proliferation assay of OE-NPDC1 MC38 cells following AKT knockdown. MC38 cells were seeded at 1 × 10³ cells/well following transfected with siRNA for 20 h (n = 5 per group). (M) Images of subcutaneous tumors from 6–8 weeks-old C57BL/6 male mice injected with 1 × 10⁶ NC or sh-NPDC1 MC38 cells into the left groin. The mice were treated with Buparlisib (50 mg/kg) or DMSO via intraperitoneal injection every 2 days starting 3 days post-injection. The tumors were harvested 14 days later(left). Quantification of tumor weight and volume (right). The tumor volume was measured at harvest via the following formula: volume = (length × width²)/2. (N) Surface diagram of the docking model an interfacing residues between NPDC1 and p85α protein (hydrogen bond interaction, dotted line). (O) Immunoprecipitation analysis of the interaction between NPDC1 and p85α in MC38 cells. (P) Representative images of immunofluorescence showing the expression and colocalization of NPDC1 and p85. Scale bars,25 μm. (Q) Representative western blot showing the KRAS^G12D-induced activation of p-AKT and upregulation of NPDC1 was abolished in MC38 cells treated with glutamine inhibitor (V-9302). (R) CCK-8 assay of KRAS^G12D MC38 cells with with PI3K inhibitor (Alpelisib) (n = 5 per group). (S) CCK-8 assay of KRAS^G12D MC38 cells with with PI3K inhibitor (Buparlisib) (n = 5 per group). Data represent as mean ± SD (**p < 0.01, ***p < 0.001, and ns, not significant; SD, standard deviation; Student’s t test)

To assess whether NPDC1 directly interacts with PI3K/AKT pathway components, we employed a structure-based prediction algorithm and identified potential direct binding between NPDC1 and the alpha regulatory subunit (p85α) of PI3K (Fig. 6N, sFig. 2 F). This interaction was subsequently confirmed through co-immunoprecipitation (Fig. 6O) and immunofluorescence (Fig. 6P) assays.

In addition, we observed that G12D mutant KRAS significantly elevated the protein levels of NPDC1 and phosphorylated AKT, while treatment with the glutamine inhibitor V-9302 completely abrogated these effects (Fig. 6Q). Furthermore, Alpelisib and Buparlisib partially reversed the accelerated proliferation driven by the KRAS mutation (Fig. 6R-S). These findings collectively indicate that the KRAS/Glutamine/NPDC1 axis functions upstream of PI3K/AKT signaling to regulate tumor cell proliferation.

NPDC1 has no effect on the CRC tumor immune microenvironment

Tumor cells and their microenvironment actively shape an immunosuppressive TME during progression, fostering immune evasion and disease advancement [17, 18]. To further explore whether the regulation of tumor progression mediated by NPDC1 is related to tumor immunity, we further established a subcutaneous tumor model in nude mice and found that NPDC1 knockdown still inhibited tumor progression (Fig. 7A). Consistent results were observed upon NPDC1 overexpression in mouse subcutaneous tumor models, which suggests that NPDC1 does not impact immunity (Fig. 7B). Additionally, we found that NPDC1 knockdown increased the total number of T cells in the TME but not in the spleen (Fig. 7C; sFig. 3 A-B). Moreover, knockdown NPDC1 did not affect the frequencies of CD4⁺ T cells, CD8⁺ T cells, IFN-γ⁺ T cells, GZMB⁺ T cells, TIM3⁺ T cells or TNFα⁺ T cells in either the TME or the spleen (Fig. 7D-H; sFig. 3 C-G), which indicating NPDC1 knockdown did not lead to significant changes in the quantity, activation status, or effector functions of T lymphocytes.

Fig. 7.

NPDC1 has no effect on the CRC tumor immune microenvironment. (A) NC or sh-NPDC1 MC38 cells (1 × 10⁶ cells in 100µL of PBS) were subcutaneously injected into the left groin of 6-8-week-old male nude mice. Tumor progression was observed 2 weeks later (n = 4) (left). Quantification of tumor weight (middle) and volume (right). (B) Vector or OE-NPDC1 MC38 cells (1 × 10⁶ cells in 100µL of PBS) were subcutaneously injected into the left groin of 6-8-week-old male nude mice. Tumor progression was observed 2 weeks later (n = 4) (left). Quantification of tumor weight (middle) and volume (right). (C-H) The proportions of T lymphocytes, CD4⁺ T cells, CD8⁺ T cells, IFN-γ⁺ T cells, GZMB⁺ T cells, TIM3⁺ T cells and TNFα⁺ T cells in tumor tissue (n = 4). 6-8-week-old male C57BL/6 mice were subcutaneously injected with MC38 cells (NC or sh-NPDC1), and the percentages of T lymphocytes in tumors were subsequently analyzed via flow cytometry on day 14. Each sample represents an individual mouse (n = 4 per group). (I-J) Percentages of total macrophages (I) and M1/M2 macrophage subtypes (J) within tumor tissues. The methods are described above in (C) (n = 4 per group). (K-L) The percentage of total MDSCs (K) and the proportions of PMN-MDSCs and M-MDSCs (L) within tumor tissues. The method used is as described in (C) (n = 4 per group). Data are presented as mean ± SD (*p < 0.05, **p < 0.01, ***p < 0.001, and ns, not significant; SD, standard deviation; Student’s t test)

In addition to adaptive immunity, innate immunity also plays an important role in tumor progression [19, 20]. Thus, we further analyzed the changes of macrophages and myeloid-derived suppressor cells (MDSCs). The results revealed that NPDC1 knockdown significantly decreased the population of macrophages and increased the number of M1-like macrophages in the spleen (sFig. 3 H-J) without influencing their frequency in the tumor (Fig. 7I-J). Similarly, NPDC1 knockdown did not affect the overall proportion of MDSCs, polymorphonuclear MDSCs (PMN-MDSCs) or monocytic MDSCs (M-MDSCs) in the tumor (Fig. 7K-L) or spleen (sFig. 3 K-L). Collectively, these results indicate that tumor immunity is not involved in the tumor-promoting effect mediated by NPDC1.

Discussion

In this study, we identified NPDC1 as a CRC regulatory factor whose expression is consistently elevated across normal colon, IBD, and CRC tissues and revealed its crucial role in promoting CRC progression. Our findings suggest that glutamine metabolism mediated by KRAS mutations may promote high expression of NPDC1 in CRC, which in turn modulates PI3K/AKT signaling to promote CRC growth. This discovery highlights the molecular basis underlying the aggressiveness of KRAS-mutated tumors and reveals that targeting NPDC1 may be a strategy to overcome resistance to KRAS mutation or PI3K inhibition.

KRAS mutations are present in approximately 37–47% of CRC [21, 22], reprogramming cancer cell metabolism to increase glutamine intake and maintaining ATP production and the biosynthesis of intermediates required for growth [23, 24]. Here, we found that NPDC1 expression is significantly elevated in CRC tissues harboring KRAS mutations and validated that glutamine directly upregulates NPDC1 via the transcription factor JUND. These findings indicate that NPDC1, as a downstream effector factor of glutamine metabolism mediated by KRAS, enables CRC cells to adapt to metabolic stress and maintain their proliferative capacity. Given the specific elevation of NPDC1 in KRAS mutant CRC and its role in metabolic adaptation, NPDC1 presents unique therapeutic value for KRAS mutant CRC patients. Importantly, as a potential oncogene, NPDC1 is continuously upregulated during the occurrence and development of CRC, suggesting the universality of targeted therapy. Although we identified JUND as a transcriptional regulator of NPDC1, the molecular mechanisms by which glutamine activates JUND remain unclear. For example, whether glutamine metabolites directly regulate JUND through post-translational modifications or nuclear localization to influence NPDC1 expression remains to be determined.

In addition, our findings have confirmed that NPDC1 can bind to PI3K to activate the PI3K/AKT signaling pathway. The PI3K/AKT pathway, which is well-established as a key regulator of tumor cell proliferation, survival, and resistance to apoptosis, is frequently dysregulated in CRC [25, 26]. Our findings elucidate a previously uncharacterized KRAS-glutamine metabolic axis that activates PI3K/AKT signaling via NPDC1, a mechanism that functionally diverges from canonical PI3K activation pathways such as receptor tyrosine kinase signaling or PTEN loss [27]. This unique axis positions NPDC1 at the critical interface of metabolic and oncogenic signaling, providing a strong rationale for its therapeutic targeting in synergistic combinations of KRAS and PI3K pathway inhibitors. However, the structural basis of the NPDC1-PI3K interaction remains undefined. Elucidating this interface would directly facilitate the development of targeted peptide mimetics or small-molecule inhibitors.

Overall, this study elucidated the undefined role of NPDC1 in CRC progression by revealing its upstream/downstream regulatory mechanisms. Our findings also provide new insights into CRC metabolic reprogramming and signaling networks, establishing the foundation for NPDC1-targeted therapeutic development.

Materials and methods

Dataset source and preprocessing

The data of the TCGA-COAD cohort were downloaded from The Cancer Genome Atlas (TCGA) (http://cancergenome.nih.gov/). The CRC RNA-seq datasets GSE39582 and GSE8671 and the scRNA-seq dataset GSE132465 were obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/). The IBD scRNA-seq dataset SCP259 was obtained from the single-cell portal (https://singlecell.broadinstitute.org/single_cell).The K‒M plotter database (https://kmplot.com/) was used for analysis via the LogRank test. A Cox regression model was applied for regression analysis. Oncogenic pathway gene sets were collected from previous research [28].

Single-cell analysis

Single-cell analysis was conducted within the R statistical environment (v4.3.1) deployed on the XiyouCloud platform (https://www.xiyoucloud.net/). The “Seurat” package (v4.3.0) [29] facilitated the processing of raw data and subsequent quality control, which involved the inclusion of genes detected in at least ten cells and the exclusion of cells with more than 20% mitochondrial reads. Cells exhibiting fewer than 500 or more than 7000 genes were also removed. To reduce batch effects, the Harmony package (version 1.2.3) [30] was utilized. Cell type identification was primarily based on the expression patterns of canonical markers specific to major cell populations. Epithelial subset annotations were inherited directly from each of the original collected datasets. The R package “ClusterGVis” was employed to categorize six distinct gene clusters within epithelial cells [31]. Cox regression analysis was conducted to identify poor prognostic genes (HR > 1) within the C1 gene cluster that exhibited progressive upregulation across the four stages. For differential expression analysis, the FindMarkers function was applied with Wilcoxon rank-sum tests, restricted to genes expressed in > 25% of cells in either group of the comparison. Genes with a p value < 0.05 and absolute log2(fold change) > 0.25 were considered differentially expressed genes. For functional enrichment analysis, the R package clusterProfiler (version 4.10.1) was used to identify significantly enriched KEGG pathways [32].

Calculation of gene set activity

“GOBP_GLUTAMINE_TRANSPORT” gene set from MsigDB [33] and the 14 hallmark gene sets were obtained from the CancerSEA database (http://biocc.hrbmu.EdU.cn/CancerSEA/). The AddModuleScore algorithm was used to score hallmark for epithelial cell. Oncogenic pathway activity scores of TCGA and GSE39582 cohorts were calculated by single-sample gene set enrichment analysis (ssGSEA) via GSVA R package.

Prediction of transcription factors and their binding sites

Putative transcription factors regulating NPDC1 were predicted using the predict_TF function from TFTF R package [34]. The 2000-bp promoter region upstream of the NPDC1 transcriptional start site was retrieved from the NCBI database (http://www.ncbi.nlm.nih.gov/; Accession: NC_000009.12). Potential JUND binding sites within NPDC1 promoter region were predicted via the JASPAR database ((http://jaspar.genereg.net/, Matrix ID: MA0492.2) with 85% relative profile score threshold.

Cell lines and cell culture

The murine colon adenocarcinoma cell line MC38 maintained in our laboratory underwent short tandem repeat (STR) profiling for authentication and demonstrated harmful mycoplasma contamination through PCR-based testing. The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; #C3113-0500, VivaCell) supplemented with 10% fetal bovine serum (FBS; #AD00004-500, Azaood) and 1% penicillin‒streptomycin solution (#C3420‒0100, VivaCell) at 37 °C in a humidified 5% CO2 incubator.

DNA construction and siRNA transfection

For stable NPDC1 knockdown, MC38 cells were transfected with lentiviral vectors (ZsGreen-Puro-U6, Corues Biotechnology Co., Ltd., Nanjing, China) expressing short hairpin RNA (shRNA) targeting NPDC1. For stable NPDC1 overexpression, MC38 cells were transfected with lentiviral vectors (ZsGreen-Puro-CMV, Corues Biotechnology Co., Ltd., Nanjing, China) harboring full-length NPDC1 cDNA. SiRNAs targeting NPDC1 and Akt1 were purchased from Beijing Tsingke Biotech Co., Ltd. KRAS^wt and KRAS^G12D mutant plasmids(pCMV3-GFP-Puro-C3×Flag) purchased from Corues Biotechnology Co., Ltd., Nanjing. Cell transfection was performed with Lipofectamine 3000 (#L3000015, Invitrogen) for MC38 cells for 6 h, as described in the manufacturer’s protocol. Transfected cells were selected with 5 µg/mL puromycin medium (#BL528A, Biosharp). RT-qPCR and western blotting were used to verify the efficiency of the sequence.

Subcutaneous colorectal cancer mouse model

Male C57BL/6J and BALB/c nude mice aged 6–8 weeks were procured from Hunan SJA Laboratory Animal Co., Ltd. MC38 cells (1 × 10⁶) resuspended in 100µL of phosphate-buffered saline (PBS) were subcutaneously inoculated into the left inguinal region of each mouse. On day 14 post inoculation, the tumor-bearing mice were humanely euthanized. The tumors were excised, weighed immediately, and measured via digital calipers. The tumor volume was determined via the following formula: volume=(length × width²)/2.

In vivo tail vein injection model

A lung metastasis mouse model generated via tail vein injection of MC38 cells was used to evaluate cell metastasis. MC38 cells (1 × 10⁶) resuspended in 200µL of PBS were injected into the tail vein of each male C57BL/6J mouse. After the mice were sacrificed, the lungs were paraffin-embedded, and routine HE staining was performed. Intrapulmonary metastases were observed with a Leica imaging system and photographed.

Flow cytometry

The spleens were crushed with 70 μm (#BS-70-CS, Biosharp) filter screens and dissolved in PBS. After centrifugation, red blood cell lysis (#R1010, Solarbio) was used to break the red blood for 5 min. Lysis was terminated by adding an equal volume of PBS. The obtained cells were resuspended in PBS. Single-cell suspensions were subsequently obtained. The tumor tissues were isolated, minced and digested with DMEM containing collagenase type IV (1 mg/ml; #2091GR001, Biofroxx). The cells were placed on a magnetic stirrer and digested for 45 min at 37 °C. After 45 min of digestion, the digested mixtures of cells were filtered through a 70 μm nylon cell strainer. Following red blood cell lysis, a single-cell suspension was obtained after resuspension in PBS. Single-cell suspensions were incubated with surface flow fluorescent antibodies and surface stained at 4 °C for 45 min. Intracellular staining was performed with the Foxp3/Transcription Factor Staining Buffer Set (#00-5523-00, Invitrogen) according to the manufacturer’s protocol. The following antibodies were used: Alexa Fluor 700 anti-mouse CD45 antibody (#103128, BioLegend), phycoerythrin (PE) anti-mouse CD3 antibody (#100206, BioLegend), Peridinin Chlorophyll Protein (PerCP/Cy5.5) anti-mouse CD8a antibody (#100734, BioLegend), APC/Cyanine7 anti-mouse CD4 antibody (#100414, BioLegend), PerCP/Cy5.5 anti-mouse CD45 antibody (#103132, BioLegend), allophycocyanin (APC) F4/80 antibody (#123116, BioLegend), Pacific Blue (PB), anti-mouse CD11c antibody (#117322, BioLegend), PE anti-mouse CD206 antibody (#141706, BioLegend), APC anti-mouse CD11b antibody (#101212, BioLegend), APC/Cyanine7 anti-mouse Gr1 antibody (#108424, BioLegend), PB anti-mouse Ly-6 C antibody (#128014, BioLegend), and PF/Cyanine7 anti-mouse Ly-6G (#127618, BioLegend).

Cell proliferation and colony formation assays

MC38 cells in the logarithmic growth phase were seeded into 96-well plates at a density of 2,000 cells per well. Cell viability was assessed at 0, 12, 24, and 48 h via the Cell Counting Kit-8 assay (#BS350B, Biosharp) according to the manufacturer’s instructions. For the EDU Kit-8 assay for EUD staining (#C0078L, Beyotime), MC38 cells were seeded into 24-well plates containing coverslips and incubated for 24 h. Subsequently, the cells were incubated for 1 h with medium containing 10 μm EdU. The coverslips were then removed, and the cells were fixed with 4% paraformaldehyde. Staining was performed according to the manufacturer’s instructions, and imaging was conducted via an inverted microscope. For the colony formation assay, a total of 1,000 MC38 cells in the logarithmic growth phase were seeded into 6-well plates, and 2.5 mg/ml puromycin culture medium was added every 5 days. After 2 weeks of culture, the cells were fixed with paraformaldehyde and stained with crystal violet. Cell colony analysis via ImageJ (v1.54 g).

Migration and invasion assays

For the migration and invasion assays, MC38 cells were seeded into 8-µm-pore Boyden chambers in 24-well Transwell plates (Corning Falcon, 353097) with 8-micron apertures (#354480, Corning). Uncoated inserts were used to assess cell migration, whereas Matrigel-precoated inserts were used for invasion studies. MC38 cells were plated in the upper compartment of the plate in 500µL of serum-free medium, while 700µL of medium supplemented with 10% FBS was added to the lower compartment of the chamber to serve as a chemoattractant. Following a 20-h incubation, the inserts were fixed in 4% paraformaldehyde for 15 min at ambient temperature and stained with crystal violet. The cells remaining on the upper surface of the membrane were gently removed via a cotton swab. The migrated or invaded cells on the lower surface were visualized and photographed via an inverted light microscope.

Annexin V-PI apoptosis assay

MC38 cells (5 × 10⁵) were seeded in a six-well plate and cultured at 37 °C and 5% CO₂ for 24 h with 2 mL of culture medium per well. The cells were digested with trypsin to obtain each well (1 × 10⁶/ml) and processed according to the instructions of the Annexin V/PI staining kit (#640920, #422201, BioLegend), followed by analysis via flow cytometry (BD FACSVerse C6, USA) and FlowJo 10.8.1 software.

Western blot

Cellular proteins were extracted with Western cell lysis buffer (#P0013J, Beyotime) supplemented with protease and phosphatase inhibitors (#P1045, Beyotime). Protein concentrations were measured via a bicinchoninic acid (BCA) assay (#P0010, Beyotime). Thirty micrograms per protein sample was loaded onto 10% SDS‒PAGE gels (#PG112, Epizyme) for electrophoretic separation. Electrophoresis was performed at 90 V for 30 min, followed by 120 V until the samples ran to the bottom of the gel. The proteins were subsequently transferred onto a PVDF membrane (#IPVH00010, Millipore) activated with methanol. The PVDF membrane was blocked with 5% skim milk for 1 h at room temperature and then incubated with the primary antibody overnight at 4 °C. After being washed five times with PBST buffer, the PVDF membrane was incubated with the corresponding secondary antibody for 2 h at room temperature. The visualization of the target protein was achieved via highly enhanced ECL (#BG0015, Bgbiotech). The antibodies used for Western blot analysis were as follows: anti-NPDC1 (1:1000, #DF4225, Affinity), anti-β-Tubulin (1:1000, #AC030, ABclonal), anti-PI3K(p85) (1:1000, #60225-1-IG, Proteintech), anti-KRAS (1:1000, # 12063-1-AP, Proteintech), anti-Flag (1:1000, #66008-4-Ig, Proteintech), anti-rabbit IgG (#SA00001-2, Proteintech) and anti-rabbit IgG (#SA00001-1, Proteintech).

Real-time PCR

Total RNA was harvested from MC38 cells and human CRC tissues via an RNA-Quick Purification Kit (#RN001, ESScience) following the manufacturer’s protocol. Then, the RNA was reverse transcribed via ABScript Neo RT Master Mix with gDNA Remover (#RK20433, ABclonal). Real-time reverse transcription quantitative PCR (RT‒qPCR) was carried out on an ABI 7500 real-time polymerase chain reaction system with BrightCycle Universal SYBR Green qPCR Mix (#RK21219, ABclonal). After normalization to ACTB (for human CRC samples) or Gapdh (for MC38 cells) expression, relative gene levels were quantified by using Gapdh or ACTB as a housekeeping gene (2[Ct Gapdh/ACTB-Ct target gene]).

The primers used in the qPCR analysis were as follows: m-NPDC1: (F) 5’-CTTCAACACCACGAATCTCGC-3’ and (R) 5’-CCTTTGGAGGTTCCTTATGCC-3’. m-Gapdh: (F) 5’-TGTGTCCGTCGTGGATCTGA-3’, (R) 5’-CCTGCTTCACCACCTTCTTGAT-3’.m-JUND: (F) 5’-TTTAGGGCGTTCGACTCCAC-3’, (R) 5’-CCTCATAAACCCAGGCCCTC-3’. h-NPDC1: (F) 5’-CCACTACCAGCACCAACGG-3’, (R) 5’-GCTCTTTATGCCGCTCCAG-3’. h-ACT: (F) 5’- CTCTTCCAGCCTTCCTTCCT − 3’, (R) 5’- AGCACTGTGTTGGCGTACAG-3’.

Immunohistochemistry (IHC)

Paraffin-embedded CRC patient tissue slides were incubated at 60 °C overnight, deparaffinized in xylene, and rehydrated in alcohol. Antigen retrieval was performed in citrate buffer (pH = 6.0) at 95 °C for 5 min, followed by natural cooling to room temperature. Permeabilization was then performed with 0.2% Triton X-100 at room temperature for 20 min, after which endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 20 min. Next, the sections were blocked with blocking solution (#10018001060) for 1 h and incubated at 4 °C overnight with the indicated NPDC1 antibody (1:400). The following day, the tissue sections were incubated with an HRP-conjugated secondary antibody for 30 min and detected using DAB substrates for 30 s, followed by Mayer’s hematoxylin solution staining (#AR005, BOSTER). Histochemistry score (H-Score) = (percentage of weak intensity cells*1)⁺ (percentage of moderate intensity cells*2) ⁺ (percentage of strong intensity cells*3).

Immunofluorescence

The cells or tissues were directly fixed with 4% paraformaldehyde for 15 min, then permeabilized with 0.3% Triton X-100. After blocked 3% BSA for 1 h, the cells or tissues were incubated with the primary antibodies overnight at 4℃. The cells or tissues were then incubated with the secondary antibodies Alexa Fluor 488-labeled donkey anti-mouse IgG (1:800; # 715-545-151, Jackson), Cy3-labeled goat anti-rat IgG (1:800; # 711-165-152, Jackson) for 1 h at room temperature. Then, the cells or tissues were incubated for 5 min with DAPI for nucleus staining. Finally, images were obtained with fluorescence and laser confocal microscopes (Leica). The following primary antibodies were used: anti-NPDC1 (1:400; #DF4225, Affinity), anti-Ki67 (1:400; #ab16667, Abcam), anti-PI3K(p85α) (1:400; #60225-1-IG, Proteintech).

Statistical analysis

All experiments, except for those involving mice, were performed in at least three independent biological replicates, with technical replicates for each experiment. The data are expressed as the mean ± SD. For normally distributed data, unpaired or paired two-tailed Student’s t tests were used to compare the significance of differences between two groups of independent samples. For the in vivo experiments related to animals, the number of biological replicates was 4 or 5. The animals were allocated to the control experimental groups via a blinding and randomization method. The survival rate was determined via the Kaplan‒Meier method. Statistical analysis was performed with GraphPad Prism 8.0. A p value less than 0.05 was considered statistically significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: Rongchen Shi, Houjie Liang, Lili Zhang. Methodology: Qing Qin, Dapeng Zhang, Yusai Xie, Fan Zhang, Yulan Huang, Renchao Deng. Investigation: Qing Qin, Dapeng Zhang, Yusai Xie, Fan Zhang, Yulan Huang, Renchao Deng. Visualization: Rongchen Shi, Houjie Liang, Lili Zhang. Supervision: Rongchen Shi, Houjie Liang, Lili Zhang. Writing: original draft: Qing Qin, Rongchen Shi. Writing: review & editing: Rongchen Shi. All the authors have read and approved the final manuscript.

Funding

This work was supported in part by the National Natural Science Foundation of China (82303245) and the Scientific and Technological Research Program of Chongqing Municipal Education Commission (KJQN202412809).

Data availability

All the data are available from the corresponding authors upon request.

Declarations

Ethics approval and consent to participate

Mice: All applicable international, national, and institutional guidelines for the care and use of animals were followed. The Animal Care and Use Committee of Army Medical University (AMU) reviewed and approved all animal experiments in this study. No: AMUWEC20230037. Human specimens: Human sample collection for this study was conducted in accordance with the Declaration of Helsinki, and was approved by the Bishan Hospital of Chongqing Medical University Ethics Committee. Written informed consent was obtained from all the patients. Cqbykyll-20250110-36.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no potential conflicts of interest.