A pan-cancer analysis of the oncogenic role of KIF13A in human tumors Authors: Shangke Huang, Jizhang Chen, Yongxia Cui Published: 2025 ABSTRACT Despite the increasing evidence supporting the association between KIF13A and cancer, pan-cancer analysis is currently limited. Therefore, we aimed to investigate the potential for KIF13A to contribute to oncogenesis in thirty-three different tumors using publicly accessible databases. Our research findings indicate that KIF13A has lower RNA tissue specificity and exhibits lower levels of expression compared to healthy tissue in most of tumor. However, we discovered distinct associations between KIF13A expression and the outcome of diverse tumor types. Genetic variation analysis revealed that cases of UCEC with genetic alterations in KIF13A exhibited a better prognosis compared to cases without genetic alterations in KIF13A. Analysis of differentially methylated genes (DMGs) in pan-cancer studies revealed that the promoter region of KIF13A exhibited significantly different methylation levels between tumor tissues and adjacent normal tissues across various cancer types. Furthermore, we observed differences in KIF13A (NP_017396.4) phosphorylation levels between normal tissues and primary tumor tissues at different phosphorylation sites across various tumor cases. Specifically, we noted an increased phosphorylation level of KIF13A at the S1698 site in HNSC and HCC, correlating with the early differentiation of human embryonic stem cells. Analysis of immune infiltration revealed a bidirectional regulatory pattern, suggesting that KIF13A may differentially shape the tumor microenvironment across cancer types, potentially influencing immunotherapy responses and stromal interactions. Gene enrichment analysis related to KIF13A indicates that it is involved pathways related to vesicle transport, pigment synthesis, and ubiquitination modification. The synergistic effects of these pathways may collectively contribute to the role of KIF13A in tumor pathogenesis. In conclusion, this pioneering pan-cancer study offers thorough comprehension of the role of KIF13A in various cancers.Supplementary InformationThe online version contains supplementary material available at 10.1007/s12672-025-04327-9. Introduction The development of tumors is intricate, and understanding the relationship among the expression patterns, prognosis, and underlying molecular mechanisms of crucial genes across various cancer types is crucial. With the increasing improvement of tumor genome data in public databases including TCGA (The Cancer Genome Atlas) and GEO (Gene Expression Omnibus), pan-cancer research has become available. Although studies have confirmed the association between KIF13A and tumor development, there are few studies on its pan-cancer implications. KIF13A belongs to the kinesin-3 family and has a distinctive function in organelle dynamics. It was first detected in mouse through cDNA analysis [1] and has been found to be evolutionarily maintained across the evolutionary spectrum from Caenorhabditis elegans to mammals [2, 3]. KIF13A contains seven distinct domains, including the Motor_domain (cl00286) (11–352), FHA (cl00062) (471–534), KIF1B (pfam12423) (748–792), DUF3694 (pfam12473) (1003–1083), DUF3585 (pfam12130) (1171–1284), PEPcase (cl14656), and CAP_GLY (pfam01302) [4–6]. A growing number of studies have demonstrated that KIF13A serves as a key mechanical motor on the endosomal network, facilitating cargo delivery from recycling endosomes (REs) and the Trans-Golgi Network (TGN) to the plasma membrane (PM) [7]. Additionally, studies have revealed that the transport process from endosomes, especially tubular/vesicular endosomes (a subset of total REs), to enlarging melanosomes is contingent upon KIF13A [8]. Moreover, studies have suggested that KIF13A is involved in various cellular processes such as pathogenesis, pathogen survival [9, 10], cell migration [11], and cell division [12]. In addition to the information provided above, a growing number of studies suggest that KIF13A is linked to various diseases. KIF13A was initially discovered at a locus that is suspected to be linked to schizophrenia [3]. Dysfunctions in KIF13A have been associated with the occurrence and progression of tumors [4, 13–19]. However, research on the overall impact of KIF13A on various types of cancer using comprehensive omics and clinical data is currently limited. Here, We present a pan-cancer analysis of KIF13A, utilizing datasets from both the TCGA and GEO databases. This analysis focuses on examining the link between KIF13A expression and factors such as prognosis, DNA methylation, protein phosphorylation, immune infiltration, and related signal pathways. The goal is to uncover potential molecular mechanisms that are involved in the oncogenic functions of KIF13A in 33 different tumor types. Material and methods The investigation of KIF13A expression We logged into TIMER2 (http://timer.cistrome.org/) and entered KIF13A into the “Gene_DE” module. After submitting the data, we obtained box plots that showed the expression of KIF13A across different cancers, their subcategories, and normal tissues. In cases where normal tissues were scarce, We employed the “Expression analysis-Box Plots” function of the GEPIA2 (Gene expression profiling interactive analysis, version 2) platform (http://gepia2.cancer-pku.cn/#analysis) to assess the expression of KIF13A in tumor tissues versus corresponding normal tissues sourced from the GTEx (Genotype-tissue expression) database. We implemented the criteria with a p-value cutoff at 0.01 and log2FC cutoff at 1, ensuring the “Match TCGA normal and GTEX data” option was checked. Additionally, we used the GEPIA2 Pathological Stage Plot tool to generate violin plots that displayed the KIF13A expression patterns across different pathological stages of all TCGA tumors. To enhance precision in box and violin plots, the expression data were converted to log2 [TPM (Transcripts per million) + 1]. Finally, we employed Ualcan (http://ualcan.path.uab.edu/analysis-prot.html), an online platform designed for cancer Omics data analysis, to investigate protein expression pattern using data from the CPTAC (Clinical Proteomic Tumor Analysis Consortium). By submitting “KIF13A” for analysis, We examined the expression profiles of both total and phosphorylated protein variants of KIF13A (NP_071396.4) in tumor versus normal tissues. Survival outcome analysis We accessed GEPIA2 Expression Survival Map tool [20] to retrieve the Overall Survival (OS) and Disease-Free Survival (DFS) significance heatmap for KIF13A across all tumors included in the TCGA. We employed a 50% expression threshold to categorize patients into high and low-expression cohorts. In order to test our hypothesis, we applied the log-rank test. Subsequently, We navigated to the GEPIA2 Expression module to perform Survival Analysis and generate survival plots. Genetic variation analysis We visited the cBioPortal (https://www.cbioportal.org/) [21, 22] and utilized the ‘Quick select’ feature to select the ‘TCGA Pan Cancer Atlas Studies’. By inputting “KIF13A”, we were able to investigate the genetic alteration features of this gene. The results concerning alteration frequency, mutation classifications, and Copy Number Alteration (CNA) in all TCGA tumors are presented in the “Cancer Types Summary”. Furthermore, the mutated residues of KIF13A can be visualized in the schematic representation of the protein structure or the 3D (Three-dimensional) structure within the “Mutations” tool. We employed the “Comparison/Survival” module to gather statistics on Overall Survival, Disease-Free Survival, Progression-Free Survival, and Disease-Specific Survival for TCGA cancer instances, both with and without KIF13A genetic variations. Additionally, Log-rank P-values were calculated and Kaplan–Meier plots were also produced. Profiling immune cell infiltration We utilized the TIMER2 Immune-Gene tool (https://compbio.cn/timer2/) to examine the relationship between KIF13A expression and immune infiltrates in various TCGA Cancer. Various algorithms were employed to estimate immune infiltration level of CD8 + T-cells and cancer-associated fibroblasts. The purity-adjusted Spearman’s rank correlation test was used to calculate P-values and partial correlation values, which were then visualized with a heatmap and scatter plot. KIF13A-related gene enrichment analysis We utilized the STRING platform (https://string-db.org/) to gather information on the KIF13A protein of Homo sapiens. We then configured the parameters: requiring a minimum interaction score of 0.150 (regarded as low confidence), using network edges to represent evidence, limiting the display to 50 interactors in the primary shell, and only considering experimental sources for interactions. As a result, we identified 23 proteins that were experimentally identified and found to bind to KIF13A, referred to as group A. ‌Then, We access the GEPIA2 Similar Gene Detection platform (http://gepia2.cancer-pku.cn/#similar) to retrieve the top 100 target genes that exhibit expression correlation with KIF13A (designated as group B), based on comprehensive TCGA datasets encompassing both tumor and normal tissues. Additionally, pairwise gene Pearson correlation analysis between KIF13A and pre-selected genes (GTF2H1, CLIP1, MIB1, CDYL, KDM1B, ZFP91) will be carried out via the GEPIA2 correlation analysis tool. Scatter plots will be generated using Log2 TPM values to visually represent correlations between KIF13A and target genes, annotated with p-values and Pearson’s R coefficients. Furthermore, we accessed TIMER2.0_Gene_Corr module (http://timer.cistrome.org/) and entered KIF13A, along with the genes we selected (GTF2H1, CLIP1, MIB1, CDYL, KDM1B, ZFP91), into the “Interested Gene” and “Gene Expression” fields. This procedure was implemented to produce heatmap encompassing partial correlation coefficients (cor) and p-values derived from purity-corrected Spearman’s rank correlation analyses for the selected genes. In the end, we merged the two groups into a single list for performing KEGG and GO analysis. The enriched pathways were then visualized using R packages such as ggpubr, clusterProfiler, dplyr, tidyr, and ggplot2. The analysis was carried out using R version 4.4.1. A p-value below 0.05 in a two-tailed test was considered significant. Results Gene expression analysis Here, this study focused on investigating the tumor-promoting function of human KIF13A (mRNA: NM_022113.6; protein: NP_071396.4), as illustrated in Figure S1a. The protein structure of KIF13A, as illustrated in Figure S1b, is conserved among different species and is primarily composed of seven domains, including Motor_domain (cl00286), FHA (cl00062), KIF1B (pfam12423), DUF3694 (pfam12473), DUF3585 (pfam12130), GAP_GLY3694 (pfam01302), and PEPcase (cl4656) domain4. The phylogenetic relationships of KIF13A protein across species ‌are presented in Figure S2‌, illustrating evolutionary divergence patterns. To investigate the tissue specificity of KIF13A expression, we accessed the HPA (Human Protein Atlas) website to analyze its expression levels in non-tumor tissues using the Consensus dataset of HPA / GTEx / FANTOM5. The findings revealed that KIF13A exhibited lower RNA tissue specificity, suggesting that it has a ubiquitous expression profile. The highest expression levels were observed in skeletal muscle, followed by the tongue, heart muscle, and urinary bladder (Fig. S3a). Additionally, a low level of RNA-based discrimination among blood cell types does not indicate specificity when examining KIF13A expression in various blood cells from the HPA/Monaco/Schmiedel datasets (Fig. S3b), with Neutrophils displaying the highest levels of KIF13A expression. However, our analysis using mass spectrometry and immunoassays in Peptide Atlas did not detect KIF13A in plasma, indicating a lack of physiological outward leakage of the intracellular KIF13A protein (Fig. S3c and d). Next, we accessed the TIMER2_Gene_DE module to examine KIF13A expression across different cancer types in TCGA database. As illustrated in Fig. 1a, KIF13A expression is generally suppressed in tumor tissues compared to their normal counterparts. This trend is observed in various cancers such as BLCA, BRCA, KICH, KIRP, LUAD, READ, THCA, UCEC, COAD, KICR, PRAD and LUSC (Wilcoxon test, p < 0.05). In contrast, the expression of KIF13A is significantly higher in tumor tissues than in normal tissues for CHOL, HNSC, and LIHC. (Wilcoxon test, p < 0.05) (Fig. 1a). As the TIMER2 database does not contain normal tissue data for all tumors, we utilized the GEPIA2_boxplot tool, which includes normal tissue samples from the GTEx dataset serving as controls, to reanalyze the expression status of KIF13A across different cancer types. The findings indicate that the expression status of KIF13A is generally lower in most tumors, such as ACC and UCS, compared to normal tissue. Nevertheless, the expression status of KIF13A is elevated in tumor tissues of LGG and PAAD (Fig. 1b, Wilcoxon test, p < 0.05). For the remaining tumors, the difference in expression level between normal and tumor tissue is not significant (Figure S4a).Fig. 1Illustrates the expression levels of the KIF13A gene in various tumors and pathological stages. a The expression status of KIF13A in different cancers or specific cancer subtypes was analyzed using TIMER2. Statistical significance is indicated by * P < 0.05; ** P < 0.01; *** P < 0.001. b For the types CHOL, LGG, PAAD, ACC, BLCA, UCEC, and UCS in the TCGA project, corresponding normal tissues from the GTEx database were utilized as controls. Box plot data is provided with significance denoted by ** P < 0.01, * P < 0.05. c Utilizing the CPTAC dataset, the protein expression levels of KIF13A were analyzed between normal tissue (N) and primary tissue (P) in various cancers including Ovarian Cancer, Colon Cancer, Clear cell RCC, Hepatocellular carcinoma, UCEC, Glioblastoma multiforme, Pancreatic adenocarcinoma, Head and neck squamous carcinoma, Lung squamous cell carcinoma, and LUAD. d Analysis of the TCGA data revealed the expression levels of the KIF13A gene across different pathological stages (stage I, stage II, stage III, and stage IV) in OV, COAD, KIRC, and SKCM. Log2 (TPM + 1) was used for log-scale representation Illustrates the expression levels of the KIF13A gene in various tumors and pathological stages. a The expression status of KIF13A in different cancers or specific cancer subtypes was analyzed using TIMER2. Statistical significance is indicated by * P < 0.05; ** P < 0.01; *** P < 0.001. b For the types CHOL, LGG, PAAD, ACC, BLCA, UCEC, and UCS in the TCGA project, corresponding normal tissues from the GTEx database were utilized as controls. Box plot data is provided with significance denoted by ** P < 0.01, * P < 0.05. c Utilizing the CPTAC dataset, the protein expression levels of KIF13A were analyzed between normal tissue (N) and primary tissue (P) in various cancers including Ovarian Cancer, Colon Cancer, Clear cell RCC, Hepatocellular carcinoma, UCEC, Glioblastoma multiforme, Pancreatic adenocarcinoma, Head and neck squamous carcinoma, Lung squamous cell carcinoma, and LUAD. d Analysis of the TCGA data revealed the expression levels of the KIF13A gene across different pathological stages (stage I, stage II, stage III, and stage IV) in OV, COAD, KIRC, and SKCM. Log2 (TPM + 1) was used for log-scale representation Additionally, the Ualcan_CPTAC dataset also showed decreased levels of total KIF13A protein in cancer tissues of ccRCC, UCEC, OV, COAD, LIHC, LUAD, and LUSC (Fig. 1c, Wilcoxon test), but a elevated levels of total KIF13A protein in cancer tissues of HNSC, PAAD, and GBM (Fig. 1c, Wilcoxon test). Results from all four databases indicate that the expression pattern of KIF13A in most tumor tissues is lower than in normal tissues. Therefore, we speculate that KIF13A acts as a tumor suppressor in most tumors, while it may function as a tumor promoter in others. However, further experimental verification is required. Currently, some studies suggest that, in certain tumors, KIF13A promotes cancer cell invasion by participating in vesicle transport [14], regulating key signaling pathways [15], and influencing the tumor microenvironment [23]. However, other studies have shown that KIF13A, through its weak dimer state regulated by Rab22A, plays a crucial role in maintaining intracellular transport balance. Disruption of this fine regulation can lead to abnormal membrane transport and dysregulation of growth factor receptor recycling, thereby promoting tumor progression. Therefore, normally functioning KIF13A may exert a tumor-suppressive effect by maintaining cellular homeostasis [24]. Ultimately, we also utilized the “Pathological Stage Plot” tool of GEPIA2 to examine the association between KIF13A expression and the pathological stages of various cancers, including OV, COAD, SKCM, and KIRC. The significant difference among the pathological stages of tumor is evident (Fig. 1d, Wilcoxon test, P < 0.001). As the tumor progresses, the expression level of KIF13A gradually decreases (Fig. 1d). Based on these research findings, we hypothesize that with tumor progression, the tumor-suppressive function of KIF13A gradually weakens, and it may serve as a biomarker for the dynamic monitoring of tumors [14, 25]. It is not significant for the remaining cancers (Figure S4b). To further validate the research results, we logged into the GENT2 database to re-analyze the expression status of KIF13A. The research results also indicated lower expression of KIF13A in the majority of cancer tissues comparing with normal tissue (two-sample t-test, p < 0.05) (Figure S5). Survival analysis We stratified cancer patients into high and low KIF13A expression groups and analyzed the link between KIF13A expression pattern and patient outcomes across multiple tumor types in the TCGA dataset. As shown in Fig. 2a, high levels of KIF13A expression were associated with poor overall survival (OS) for KICH (p = 0.047) and LIHC (p = 0.0017), while they were linked to good OS for KIRC (p = 0.00019), LGG (p = 3.5e-06), and READ (p = 0.031). Additionally, high levels of KIF13A expression were associated with good Disease-free survival (DFS) for COAD (p = 0.028) and LGG (p = 0.00045), whereas they were linked to poor Disease-free survival (DFS) for KICH (p = 0.071) (Fig. 2b, Table S8). The correlation was not significant for the rest of the cancers. It is worth noting that some cancers exhibit discrepancies between overall survival (OS) and disease-free survival (DFS) (Fig. 2). These differences may arise from both biological and methodological factors. Biologically, there is considerable heterogeneity in tumor aggressiveness, metastatic patterns, and sensitivity to treatment among different tumors [26, 27]. For example, variation in tumor progression rates following recurrence [28] or differences in responses to second-line therapies [29–31] can lead to divergence between OS and DFS. Methodologically, DFS relies on the timely detection of postoperative recurrence [32, 33]; if the follow-up period is too short or the frequency of surveillance is inadequate, early recurrences may go undetected [34]. In contrast, OS is less influenced by non-tumor-related deaths and may better reflect long-term survival trends [35]. In summary, discrepancies between OS and DFS result from the interplay of tumor biology, follow-up strategies, and therapeutic interventions, and warrant further analysis in the context of the clinicopathological features and molecular mechanisms of specific cancer subtypes.Fig. 2The correlation between KIF13A expression and the survival prognosis of cancers in TCGA. We utilized the GEPIA2 tool to conduct overall survival (a) and disease-free survival (b) analyses of various tumors in TCGA based on KIF13A expression levels. The survival map and Kaplan–Meier curves with significant results are provided The correlation between KIF13A expression and the survival prognosis of cancers in TCGA. We utilized the GEPIA2 tool to conduct overall survival (a) and disease-free survival (b) analyses of various tumors in TCGA based on KIF13A expression levels. The survival map and Kaplan–Meier curves with significant results are provided To further investigate the association between KIF13A expression pattern and prognosis, we accessed the gene chip data from the Kaplan–Meier Plotter website [36, 37]. Consistent with the findings shown in Fig. 2, the results revealed that high levels of KIF13A expression are correlated with favorable prognosis in patient with colon cancer (RFS, p = 8.2e − 06; PPS, p = 0.00015) (Figs. S6f, 2b). Then, we accessed the RNA-seq_Pan-cancer database of Kaplan–Meier tool to assess the association between OS/RFS and KIF13A expression. Again consistent with the results in Fig. 2, the findings revealed that high levels of KIF13A expression are associated with favorable overall survival (OS) in READ (p = 0.0035), KIRC (p = 5.6e-07), but associated with unfavorable prognosis in LICH (RFS, p = 0.04) (Fig. S7b, Fig. 2 and Table S7). Overall, we identified distinct associations between KIF13A expression and clinical outcomes across various tumor types. This may suggest that KIF13A exerts dual roles as either a tumor promoter or suppressor within the tumor microenvironment—potentially through mechanisms such as the regulation of microtubule dynamics—in cancers such as COAD, READ, KIRC, and LIHC. Therefore, its expression pattern could serve as a potential prognostic biomarker for cancer [6, 14, 24]. Interestingly, analysis of KIF13A expression reveals that KIF13A is expressed at low levels in most tumors. Survival outcome analysis further indicates that, in most tumors, patients with low KIF13A expression tend to have a poor prognosis (Figures S6 and S7). However, there are also certain tumor types (ESSC, KIRP, LIHC, BLCA, and THYM) in which low KIF13A expression correlates with a better prognosis (Figures S6 and S7), which appears contradictory. However, previous studies have shown that in tumors with low KIF13A expression, other motor proteins (such as KIFC1) may be compensatory activated within similar pathways [38], thereby diluting the prognostic value of KIF13A. Additionally, epigenetic modifications, such as histone acetylation [39], may enable KIF13A to retain functional activity in some instances. In different tumor subtypes, KIF13A may play opposing roles by regulating processes such as mitotic spindle formation or cell migration [40]. Collectively, these findings may partially explain the differences in prognosis associated with KIF13A expression. Genetic alteration analysis We examined the mutational landscape of KIF13A across diverse TCGA tumor datasets. As illustrated in Fig. 3a, SKCM displayed the most frequent KIF13A alterations (> 9.23%), where mutations predominated, followed by UCEC (7.75%). CNA was the most common type of amplification in OV cases, with an alteration frequency of approximately 7.53% (Fig. 3a), followed by BLCA at 5.84%. As depicted in Fig. 3a, deep deletions constituted the predominant genetic alteration in MESO (1.15%), with STAD ranking second at 1.14%. Additionally, the spectrum of KIF13A-associated genetic variants is comprehensively mapped in Fig. 3b. It was found that missense mutations predominated among genetic alterations. Frameshift mutations in KIF13A were predominantly triggered by the V1116Sfs*23/4 alteration, particularly in STAD (n = 6) compared to EAC (n = 1) and COAD (n = 2) cohorts (Fig. 3b). This results in the translation from Valine (V) to Serine (S) at the 1116 site of the KIF13A, ultimately leading to truncation of the KIF13A protein.Fig. 3Analysis of the mutation feature of KIF13A in various tumors from the TCGA was conducted. By utilizing the cBioPortal tool, we assessed the mutation characteristics of KIF13A across the TCGA tumors. The alteration frequency, along with mutation type (a) and mutation site (b), were examined. The mutation site with the highest alteration frequency (V1116Sfs*23/4) was visualized in the 3D structure of KIF13A (c). Additionally, we explored the potential correlation between mutation status and overall, disease-specific, disease-free, and progression-free survival of UCEC (d) through the cBioPortal tool Analysis of the mutation feature of KIF13A in various tumors from the TCGA was conducted. By utilizing the cBioPortal tool, we assessed the mutation characteristics of KIF13A across the TCGA tumors. The alteration frequency, along with mutation type (a) and mutation site (b), were examined. The mutation site with the highest alteration frequency (V1116Sfs*23/4) was visualized in the 3D structure of KIF13A (c). Additionally, we explored the potential correlation between mutation status and overall, disease-specific, disease-free, and progression-free survival of UCEC (d) through the cBioPortal tool Additionally, we examined the possible link between genetic changes in KIF13A and the survival outcomes of various cancer types. The findings showed that cases of UCEC with altered KIF13A had a more favorable prognosis (OS, p = 0.0278; DSS, p = 0.0231; DFS, p = 0.0129; PFS, p = 2.363e-3) compared to UCEC cases without KIF13A alterations (Fig. 3d). This finding may suggest increased patient sensitivity to treatment or indicate that KIF13A could serve as a new target for combination therapy [41, 42]. However, this correlation was not statistically significant for other type of cancers. Next, we analyzed the correlation between KIF13A mRNA abundance and tumor mutational landscape parameters (TMB/MSI) across 33 TCGA cancer types. The results revealed a positive correlation between KIF13A expression and TMB for COAD (P = 0.026), but negative correlated with TMB for LUAD (P = 0.0026) and THCA (P = 0.019) (Figure S8). In the meantime, a positive relationship exists between KIF13A expression and MSI in LUSC (P = 0.0015), OV (P = 0.011), COAD (P = 6.2e-05), READ (P = 8.4e-10), LMAL (P = 0.046), and STAD (P = 0.0015), but negatively correlated with MSI in PRAD (P = 4.9e-08) and HNSC (P = 0.0042) (Figure S9). This mean that the association between KIF13A expression and tumor mutation burden (TMB) as well as microsatellite instability (MSI) is highly cancer-type specific. For example, in colorectal cancer (COAD), KIF13A expression is positively correlated with genomic instability, whereas in lung adenocarcinoma (LUAD), it shows a negative correlation. These findings imply that KIF13A may influence DNA repair or cell cycle checkpoints through tissue-specific regulatory mechanisms [6, 43], but it needs to be further verified through experiments. KIF13A methylation analysis DNMIVD was employed to evaluate whether KIF13A methylation contributes to pathogenesis in diverse tumors. Analysis of differentially methylated genes (DMG) of KIF13A in pan-cancer studies revealed significant differences in methylation levels between adjacent and tumor tissues in the promoter region of KIF13A for BRCA (P = 0.00107), HNSC (P = 0.029), KIRP (P = 0.00037), and PAAD (P = 0.0004) (Figure S10). Additionally, A significant inverse correlation was observed between KIF13A methylation and its expression in BRCA (P = 0.00048824), HNSC (P = 0.00132566), KIRP (P = 0.000220667), LIHC (P = 0.0167125), SARC (Sarcoma, P = 0.0182044), SKCM (P = 0.0000014301), and STAD (P = 0.0072986), but a positive correlation in COAD (P = 0.0452013) and THCA (P = 0.0300103) (Figure S11). These results may indicated that differential methylation of the KIF13A promoter in different tumors influence tumor development and progression through epigenetic regulation and KIF13A may serve as a potential diagnostic biomarker or therapeutic target [44–46]. Protein phosphorylation analysis Utilizing the CPTAC database, we assessed differential phosphorylation status of KIF13A (NP_017396.4) in primary tumor versus matched normal tissues spanning diverse malignancies. As illustrated in Fig. 4g, the phosphoproteomic analysis revealed significant changes in KIF13A phosphorylation at specific residues. In ccRCC, Primary tumors exhibit attenuated phosphorylation at the S1529 (p = 1.02E-06) residue compared to matched normal renal tissues (Fig. 4a). In LUAD, primary tumors exhibit divergent phosphorylation patterns: elevated at S1529 ((p = 2.58E-08) yet reduced at S363 (p = 4.47E-04) and S1283 (p = 5.71E-29) versus normal tissues (Fig. 4b). For LUSC, primary tumors manifest significant hyperphosphorylation at S1529 (p = 3.38E-07) relative to matched normal bronchial epithelia (Fig. 4c). In HNSC, Primary tumors exhibit significantly elevated phosphorylation at S636 (p = 1.78E-19), T1285 (p = 2.32E-03), T1447 (p = 2.45E-06), S1682 (p = 2.67E-04), S1698 (p = 6.64E-05), and S1700 (p = 2.32E-03) relative to normal counterparts (Fig. 4d). In HCC (hepatocellular carcinoma), primary tumors exhibit hyperphosphorylation at both site S1529 (p = 9.22E-07) and S1698 (p = 3.06E-16) compared to adjacent normal liver tissues (Fig. 4e). For PAAD, primary tumors demonstrate significant hypophosphorylation at S1529 (p = 1.28E-02) compared to normal tissues (Fig. 4f). Overall, KIF13A (NP_017396.4) exhibits tumor-specific phosphorylation heterogeneity‌ across clinical specimens, with divergent patterns between normal and malignant tissues. This cancer-type specific phosphorylation abnormalities of KIF13A suggest that it may participate in tumorigenesis by regulating microtubule dynamics, and could serve as a potential therapeutic target [47, 48].Fig. 4Phosphorylation analysis of the KIF13A protein in various tumors was conducted. Using the CPTAC dataset, we compared the expression levels of the KIF13A phosphoprotein (NP_017396.4) in normal tissue and primary tissue of different selected tumors through UALCAN. Box plots are provided for different cancers, including ccRCC (a), LUAD (b), LUSC (c), HNSC (d), HCC (e), and PAAD (f). Significantly altered phosphoprotein sites are highlighted in the schematic diagram of the KIF13A protein (g). In the diagram, a red arrow pointing upwards indicates an increase in KIF13A expression in tumor tissue at that phosphorylation site compared to normal tissue, while a red arrow pointing downwards indicates a decrease in KIF13A expression in the tumor tissue at that phosphorylation site Phosphorylation analysis of the KIF13A protein in various tumors was conducted. Using the CPTAC dataset, we compared the expression levels of the KIF13A phosphoprotein (NP_017396.4) in normal tissue and primary tissue of different selected tumors through UALCAN. Box plots are provided for different cancers, including ccRCC (a), LUAD (b), LUSC (c), HNSC (d), HCC (e), and PAAD (f). Significantly altered phosphoprotein sites are highlighted in the schematic diagram of the KIF13A protein (g). In the diagram, a red arrow pointing upwards indicates an increase in KIF13A expression in tumor tissue at that phosphorylation site compared to normal tissue, while a red arrow pointing downwards indicates a decrease in KIF13A expression in the tumor tissue at that phosphorylation site Leveraging the PhosphoNET knowledgebase, we performed functional annotation of CPTAC-identified phosphosites on KIF13A (NP_017396.4). This revealed S1698 as a conserved regulatory hub coordinating early pluripotency in human embryonic stem cells (hESCs) [49, 50]. This result may suggest that S1698 is involved in regulating stem cell symmetric division or fate determination via the microtubule-vesicle transport network [51]; however, further experimental validation is required. In a tumorigenic context, further investigation is needed to elucidate the crosstalk between this phosphorylation motif and kinases such as AKT/PKM2 [52], especially given the potential role of S1698 in promoting metastasis during epithelial-mesenchymal transition (EMT) through two mechanisms: first, by acting as a downstream effector node within the kinase network, thereby regulating vesicle transport of mesenchymal markers (such as N-cadherin) [51]; and second, by altering the interaction interface between KIF13A and dynein motors via phosphorylation, which may influence tumor cell polarity and migration [24]. Overall, this discovery provides new avenues for anti-tumor strategies targeting the KIF13A phosphorylation network. Immune infiltration analysis As critical constituents of the tumor microenvironment (TME), tumor-infiltrating immune cells mechanistically contribute to carcinogenesis, malignant progression, and metastatic dissemination. Here, employing seven algorithms (TIMER, CIBERSORT, CIBERSORT-ABS, QUANTISEQ, XCELL, MCPCOUNTER, EPIC), this investigation systematically profiled tumor-infiltrating immune cell subsets and evaluated their associations with KIF13A expression across TCGA malignancies. This may imply a dual regulatory relationship between KIF13A expression and the tumor microenvironment. For example, In head and neck squamous cell carcinoma (HNSC)—including both HPV-positive and HPV-negative subtypes—as well as in lung squamous cell carcinoma (LUSC), elevated KIF13A transcriptional levels were significantly and negatively correlated with CD8 + T cell infiltration density (Figure S12). This finding suggests that KIF13A may contribute to immune evasion by inhibiting T cell recruitment or function [53, 54]; however, this hypothesis requires further experimental validation. In contrast, in lung adenocarcinoma (LUAD), pancreatic cancer (PAAD), and gastric cancer (STAD), KIF13A expression was positively correlated with cancer-associated fibroblast (CAF) infiltration, indicating a potential role in promoting tumor progression through the activation of pro-tumorigenic stromal signaling [55] (Fig. 5). Recent studies have shown that extracellular vesicles (EVs) secreted by tumor cells can encapsulate various cellular components—such as proteins, lipids, and miRNAs—and transfer these bioactive substances to recipient cells, including CAFs. This process activates resident fibroblasts, inducing a myofibroblast phenotype and thereby promoting tumor progression [56]. As a motor protein, KIF13A may participate in the transport or secretion of EVs. This mechanism may, to some extent, help explain why KIF13A expression is positively correlated with cancer-associated fibroblast infiltration. These cancer-type-specific patterns of microenvironmental interaction not only underscore the potential of KIF13A as a target for immunotherapy but also suggest that its regulation of microtubule dynamics may influence either immune cell migration or CAF activation in different tumor contexts, providing a foundation for the development of targeted intervention strategies [57].Fig. 5Correlation analysis was conducted to examine the relationship between KIF13A expression and immune infiltration of cancer-associated fibroblasts. a. Various algorithms were employed to investigate the potential correlation between KIF13A expression levels and cancer-associated fibroblast infiltration levels in all types of cancer in TCGA. b. The correlation between KIF13A expression levels and cancer-associated fibroblast infiltration levels in LUAD, PAAD, and STAD was analyzed by adjusting for purity Correlation analysis was conducted to examine the relationship between KIF13A expression and immune infiltration of cancer-associated fibroblasts. a. Various algorithms were employed to investigate the potential correlation between KIF13A expression levels and cancer-associated fibroblast infiltration levels in all types of cancer in TCGA. b. The correlation between KIF13A expression levels and cancer-associated fibroblast infiltration levels in LUAD, PAAD, and STAD was analyzed by adjusting for purity Enrichment analysis of KIF13A-related partners To delineate KIF13A’s oncogenic mechanisms, we employed STRING (v12.0) and GEPIA2 databases to define: (i) Group A: 23 high-confidence KIF13A interactors (Fig. 6a), and (ii) Group B: 100 top co-expressed genes. The scatter plot demonstrates that CLIP1, GTF2H1, KDM1B, M1B1, CDYL and ZFP91 show a strong positive correlation (R > 0.45, p < 0.001), and this co-expression pattern is prevalent across multiple cancer types (Fig. 6b, c). The results imply that these genes may participate tumor progression through synergistic regulation of pathways [25, 58–65]. Subsequently, we combined the data from both groups to perform functional enrichment analysis. The results indicated enrichment in pathways related to vesicle transport, pigment synthesis, and ubiquitination modification (Fig. 6d and e). The synergistic effects of these pathways may collectively regulate immune escape mechanisms within the tumor microenvironment. For example, vesicle transport mediates the dynamic localization of immune checkpoint proteins [66–68]; pigment synthesis influences oxidative stress levels, thereby indirectly suppressing immune cell activity [69]; and ubiquitination modification determines the stability of key proteins [70]. These findings provide potential targets for intervention, such as the combined use of immune inhibitors and pathway modulators for targeted therapeutic strategies.Fig. 6 KIF13A-related gene enrichment analysis. a We obtained a list of experimentally determined KIF13A-binding proteins using the STRING tool. b Additionally, we used the GEPIA2 approach to identify the top 100 KIF13A-correlated genes in TCGA projects and analyzed the expression correlation between KIF13A and selected target genes, including CDYL, CLIP1, GTF2H1, KDF1B, ZFP91, and MIB1. c The corresponding heatmap data for detailed cancer types is displayed. d KEGG pathway analysis was performed based on the KIF13A binding and interacted genes. e GO analysis was also conducted based on the KIF13A binding and interacted genes KIF13A-related gene enrichment analysis. a We obtained a list of experimentally determined KIF13A-binding proteins using the STRING tool. b Additionally, we used the GEPIA2 approach to identify the top 100 KIF13A-correlated genes in TCGA projects and analyzed the expression correlation between KIF13A and selected target genes, including CDYL, CLIP1, GTF2H1, KDF1B, ZFP91, and MIB1. c The corresponding heatmap data for detailed cancer types is displayed. d KEGG pathway analysis was performed based on the KIF13A binding and interacted genes. e GO analysis was also conducted based on the KIF13A binding and interacted genes Discussion Kinesins, as microtubule-dependent molecular motors, play key roles in tumor cell proliferation, migration, invasion, and metastasis by regulating cytoskeletal dynamics, cell cycle progression, and signal transduction pathways [71]. Numerous studies have investigated the roles of kinesins in cancer progression. For example, kinesin subtypes such as KIF11 and KIF15 enhance tumor cell division efficiency by promoting mitotic spindle formation and chromosome segregation [72], while KIF23 and KIF20A are implicated in tumor cell invasion and metastasis by regulating extracellular matrix degradation and microtubule network remodeling [73, 74]. Meanwhile, vesicle transport, a core mechanism of intracellular material trafficking, also plays a crucial role in immune evasion and epithelial-mesenchymal transition (EMT) [68, 75–77]. Vesicle transport influences the activation and function of immune cells in the tumor microenvironment by regulating antigen presentation, expression of immune checkpoint molecules, and intercellular communication, thereby promoting tumor immune escape68. During EMT, vesicle transport facilitates the transition of epithelial cells to mesenchymal cells by transporting EMT-related transcription factors, cell adhesion molecules, and signaling molecules, thus promoting tumor cell migration and invasion [78]. In addition, kinesins, as molecular motors, mediate the directional transport of vesicles along microtubules. In recent years, increasing attention has been paid to the interplay between kinesins and vesicle transport; for example, certain kinesin family members can directly participate in the directional transport of vesicles, while abnormalities in vesicle transport pathways may, in turn, affect the subcellular localization and function of kinesins [71, 79, 80]. This bidirectional interaction provides a new perspective for understanding tumor heterogeneity and drug resistance [81, 82]. However, the synergistic regulatory network between these two processes in cancer progression still requires further investigation. This perspective may help guide the design of combination strategies to overcome relapse driven by tumor plasticity. Among kinesin family members, KIF13A plays a critical role in cancer progression due to its unique biological characteristics and holds promise as a potential biomarker for tumor progression [6, 14]. Its distinctiveness is reflected in three aspects: functional specificity, subcellular localization, and involvement in signaling pathways. Functionally, KIF13A primarily participates in membrane trafficking and cytoskeletal remodeling, influencing tumor cell migration and invasion by regulating vesicle transport and cytoskeletal dynamics6—unlike subtypes such as KIF11 and KIF15, which are mainly involved in cell division [72]. Regarding subcellular localization, KIF13A is found at the cell membrane and on vesicles, participating in intercellular communication and signal transduction [6, 24], whereas KIF23 and KIF20A are localized in the cytoplasm and microtubules [73, 83], where they are involved in cytoskeletal remodeling and extracellular matrix degradation. In terms of signaling, KIF13A affects tumor cell fate and plasticity by modulating pathways such as Wnt/β-catenin, Notch, and the mTORC1-KIF13A-M6PR axis [23, 84], while other subtypes mainly contribute to cancer progression by regulating the cell cycle and microtubule dynamics. As a potential biomarker, KIF13A can aid in the early diagnosis and prognostic assessment of cancer by enabling detection of its expression levels or genetic variations [85, 86], thus providing a basis for individualized treatment strategies. In drug development, KIF13A serves as both a key target for therapy and an indicator for companion diagnostics [24]; inhibitors targeting KIF13A—such as small molecules[87–89] or monoclonal antibodies [90]—can disrupt tumor microtubule dynamics and inhibit cell division. At the same time, KIF13A-based genetic testing can be used to identify sensitive patient populations, improve drug efficacy, and reduce the risk of toxicity [85, 86]. This association exemplifies the synergistic strategy of “biomarker–target–drug” in precision medicine and is expected to open new avenues for cancer therapy. Supplementary Information Supplementary material 1. Supplementary material 2. Supplementary material 3. Supplementary material 4. Supplementary material 5. Supplementary material 6. Supplementary material 7. Supplementary material 8. Supplementary material 9. Supplementary material 10. Supplementary material 11. Supplementary material 12. Supplementary material 13. Supplementary material 14. Supplementary material 15. Supplementary material 16. Supplementary material 1. Supplementary material 2. Supplementary material 3. Supplementary material 4. Supplementary material 5. Supplementary material 6. Supplementary material 7. Supplementary material 8. Supplementary material 9. Supplementary material 10. Supplementary material 11. Supplementary material 12. Supplementary material 13. Supplementary material 14. Supplementary material 15. Supplementary material 16.