Comprehensively analysis exploring the role of kinesin family members in glioma

doi:10.21203/rs.3.rs-1519074/v1

Download PDF

Article

Comprehensively analysis exploring the role of kinesin family members in glioma

https://doi.org/10.21203/rs.3.rs-1519074/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Extensive research indicates that kinesin superfamily (KIFs) regulates tumor progression. Nonetheless, there is little share of literature focused on the potential prognostic and therapeutic role of KIFs in glioma.

Methods

Four independent cohorts were generated into a large combination cohort. Data from The Cancer Genome Atlas (TCGA) database served as the training cohort and data from Chinese Glioma Genome Atlas (CGGA) database served as the validation cohort. The R software were used for statistical analysisa, FigureYa and the online tool of Xiantao were used for mapping.

Results

The signature consist of five KIFs (KIF4A, KIF26A, KIF1A, KIF13A and KIF13B) was successfully identified and showed good performance in prognosis prediction with AUC values corresponding for 1-, 3‐, and 5-years were 0.66, 0.78 and 0.75 in training cohort, and 0.77, 0.82 and 0.83 in validation cohort. We then explored the genomic features differences including immune features and tumor mutation status between high and low risk groups, from which we found that patients in high risk group had a higher level of immune checkpoint modules and IDH1 was identified mutated more frequently in low risk group. Results of GSEA analysis showed that E2F target, mitotic spindle, EMT, G2M checkpoint and TNFa signaling were significantly activated in high risk patients, partially explaining the differential prognosis between two groups. Besides, we also verified the five signature genes in the Human Protein Atlas (HPA) database, which identified of candidates for relevant biomarkers.

Conclusion

The KIFs based signature is a reliable biomarker for the prediction the prognosis of glioma. Glioma patients between two risk groups showed significant discrepancy in immune features, tumor mutation status and prognosis chcracteristics.

glioma

kinesin superfamily

prognosis

signature

bioinformatics analysis

Glioma is the most prevalent and aggressive malignancy in the brain, with about 240,000 new cases diagnosed and 200,000 deaths worldwide every year, renowned for its frequent recurrence, high aggressiveness and mortality ^{1, 2}. Glioma is divided into I-IV grades, in which grade I-II are low-grade gliomas (LGGs) and grade III-IV are high-grade gliomas (HGGs) according to World Health Organization (WHO) standards ³. Notably, glioblastoma (GBM, WHO grade IV) accounts for approximately 60% of all gliomas with the worst prognosis, while LGGs have a low tendency to metastasize and a better prognosis ⁴. Although many advances have been made in the standard treatments for glioma patients including surgical treatment, chemotherapy, radiotherapy, the overall survival (OS) of glioma has not substantially improved (5-year survival rate < 10%) ⁵. Meanwhile, most glioma patients are diagnosed and treated when they have been already in the intermediate and advanced stage, which might be attributed to the high complexity and heterogeneity of glioma and the failure of effective biomarkers entering the bloodstream by the blood-brain barrier (BBB) ^{6, 7}. Therefore, this calls for the identification of more effective targets that roubustly predict the prognosis of glioma, with the overarching goal of improving clinical diagnosis and treatment in patients.

The kinesin superfamily (KIFs) is a class of microtubule-based motor proteins that play a key role in the transport of intracellular various substances, including RNA, protein complexes, organelles and other substances ⁸. Up to date, a total of 45 KIFs have been discovered and identified in humans, among which several family members were closely involved in tumorigenesis and progression of various cancers ⁹. For instance, several mitotic KIFs including KIF4A, KIF11, KIF14, KIF15, KIF18A, KIF18B, KIF23, KIFC1, and KIF2C) were upregulated and activated in ovarian cancer tissues that act as oncogenes, recognized to be one of most important markers of poor prognosis ¹⁰. Another study revealed that overexpression of KIF18A was independently associated with poor OS of lung adenocarcinoma patients, and demonstrated that KIF18A knockdown significantly inhibited proliferation of cells both in vitro and in vivo ¹¹. Besides, KIF15 was reported to be notably upregulated in several cancer tissues compared with adjacent tissues such as breast cancer and glioma, implying that KIF15 could serve as a potential prognostic biomarker and therapeutic target ^{12, 13}. Notably, KIF4A, KIF9, KIF18A, and KIF23 were verified to have significant clinical implications in both LGG and GBM by Sang et al ¹⁴. In addition, KIF11 was identified as a target controlled cell growth and proliferation of glioma, and administration of a highly specific KIF11 inhibitor resulted in a markedly reduced of glioma tumor initiating cells ¹⁵. Given the prominent values of KIFs in various tumors, there is an urgent need to screen KIFs closely associated with the prognosis in glioma, which will be beneficial in improving prognosis evaluations and providing insights into compelling therapeutic targets.

In our study, to understand the potential role of KIFs in glioma, we comprehensively retrieved TCGA and CGGA databases and finally generated four independent cohorts into a large combination cohort (1680 samples). Next, after differential expression analysis of all KIFs, 40 KIFs were screened for further analysis. Then a prognostic signature was established based on the classic statistic (Cox Proportional Hazards) method that the signature could effectively predict prognosis of glioma patients. Kaplan–Meier survival analysis, immune features analysis and tumor mutation analysis were performed to identify the clinical feature and genomics feature between two groups. Notably, expression of signature KIFs found in HPA database showed considerable differences. These results will improve our understanding of the role of KIFs in glioma and provide a basis for personalized treatment.

Data acquisition and processing

All methods were performed in accordance with approved guidelines. TCGA-GBM and LGG data including RNA sequencing and clinical information (5 normal samples and 698 tumor samples) were downloaded were retrieved from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/), served as the training cohort. Besides, partial data from normal tissue (1152 samples) were derived from Genotype Tissue Expression (GTEx) database (https://gtexportal.org/). Two datasets (325 samples and 693 samples) including mRNA and matched clinical data of glioma were retrieved from Chinese Glioma Genome Atlas (CGGA) database (https://cgga.org.cn/), served as the validation cohort. TCGA data were preprocessed with the following steps: data merging, normalized as TPM, probe annotation and exclude patients with missing clinical information and survival time less than 30 days. Then, the “ComBat” algorithm in the “sva” R package (version 4.0.2) was applied to batch normalize the data and reduce the likelihood of batch effects between different datasets.

Differential expression analysis

45 KIFs were obtained from the known literature ⁹. KIFs RNA sequencing expression data of glioma from TCGA database and GTEx database were downloaded using UCSC XENA data hubs (https://tcga.xenahubs.net), and were processed by the Toil process ¹⁶. Differential expression analysis of all 45 KIFs was performed between 1157 normal and 689 tumor samples, and results were visualized using the “ggplot2” R package (version 4.0.2). Then the heatmap with clinical parameters was drawn based on the “ComplexHeatmap” R package (version 4.0.2).

Construction of KIFs based prognostic signature

The detailed clinicopathological parameters of glioma patients, including age, gender, grade, survival status were shown in Supplementary Table 1. The TCGA cohort was used to construct a prognostic signature, followed by validation in the CGGA cohort. KIFs with p < 0.001 after differential expression analysis were selected for univariate survival Cox analysis, among which KIFs related to prognosis with p < 0.01 were identified for further analysis. The least absolute shrinkage and selection operator (LASSO) is an acknowledged algorithm for feature selection with high-dimensional data¹⁷. Lasso regression analysis was then performed with the optimal value of penalty parameter (λ) determined according to the tenfold cross-validations to select significant KIFs based on the “glmnet” R package (version 4.0.2). Finally, a prognostic signature was constructed after multivariate survival Cox analysis and the KIFs with independent prognostic predictive value were enrolled for signature construction in accordance with the following formula: riskscore = coef1*gene1 + coef2* gene2 + coef3* gene3 + … + coefN* geneN.

Validation of the prognostic signature

Based on prognostic risk scores calculated by the above formula, glioma patients were stratified into into high-risk score group and low‐risk score group. Kaplan–Meier (KM) survival analysis was carried out to compare the OS outcomes of the two groups. Distribution curves, scatter dot plot, and heatmap were drawn to visualize riskscore distribution, number of censored patients and expression of prognosis‐related KIFs. Furthermore, the time-dependent receiver operating characteristics (ROC) and the area under the ROC curve (AUC) were performed to estimate the predictive accuracy of the signature at 1‐, 3‐, and 5-years. Moreover, univariate and multivariate Cox regression analyses were also performed to investigate the independence of riskscore as a predictive factor from different clinical variables (including age, gender and grade).

Functional enrichment analysis

Gene set enrichment analysis (GSEA) was performed using R “ClusterProfiler” package (version 4.0.2). Hallmark gene set from the Molecular Signatures Database (MsigDB, https://software.broadinstitute.org/gsea/msigdb/) was chosen as a reference gene set to explore the difference in the oncogenetic pathways between different groups. The relevant enrichment pathways with normalized p < 0.001 and FDR < 0.25 were considered as significant.

Exploration of immune features

Immune score, stromal score, and estimate score between two risk groups were calculated using R “estimate” package (version 4.0.2). 52 genes were selected to be immune-checkpoint–relevant transcripts and the expression values of these 52 genes were extracted to perform differential expression analysis between two risk groups. Furthermore, correlation analysis between riskscore and expression of CTLA4, PD1, PD-L1 and PD-L2 was performed using Pearson’s correlation.

Mutation signature analysis

Somatic mutations of glioma samples was downloaded from TCGA database. Mutation types including Frame_Shift_Del, Frame_Shift_Ins, In_Frame_Del, In_Frame_Ins, Missense, Nonsense, Nonstop, Splice_Site and Translation_Start_Site were defined as non-synonymous mutation variants, while other types of mutation were defined as synonymous mutation variants. Genes with the number of times the mutation occurred more than 20 times in at least one group and p < 0.05 were considered significant mutation genes, and the interactions of these mutation genes were analyzed by mafetools.

Immunohistochemistry

Immunohistochemistry images of KIFs identified from multivariate survival Cox analysis in normal tissue and tumor tissues were obtained from The Human Protein Atlas (HPA) (https://www.proteinatlas.org)¹⁸. Pictures of normal samples were selected from patient 3643, 3740 and 3731, while pictures of glioma samples were selected from patient 3064, 2790 and 3226.

Data combination and KIFs quantification

After comprehensively retrieving the publicly available database, four BLCA cohorts met the our criteria, including TCGA-LGG, TCGA-GBM and two detasets from CGGA database (Fig. 1). Meanwhile, apparent inter-assay variances were found in these datasets (Fig. 1A; Comp 1: 29.9% variance, Comp 2: 11.9% variance). Therefore, R “sva” package was used to eliminate potential inter-assay differences as much as possible. As shown in Fig. 1B, likelihood of batch effects was significantly reduced and four datasets were generated into a combined glioma cohort (Fig. 1B; Comp 1: 18.1% variance, Comp 2: 9.6% variance). Furthermore, we used ssGSEA package to quantify the scores of KIFs expression profiles in each patient (Fig. 1C; 1680 samples). The detailed clinical characteristics of glioma patients were shown in Supplementary Table 1.

Construction of KIFs based prognostic signature

Further differential expression analysis of all 45 KIFs were performed between TCGA tumor samples and matched TCGA and GTEx normal samples. A total of 40 diferentially expressed KIFs with p < 0.001 were screened, among which 10 KIFs including KIF17, KIF1C, KIF12, KIF5C, KIFC2, KIF19, KIF25, KIF1A, KIF5A and KIF2B were found decreased expression in tumor samples (Fig. 2A), while other 30 KIFs were signifcantly upregulated in glioma (Fig. 2B-D). Patients with complete survival information, including survival status and survival time > 30 days, were selected for further analysis. TCGA cohort served as the training cohort and CGGA cohort served as the validation cohort. Firstly, we performed univariate Cox analysis to identify prognosis-related KIFs and the result showed that 6 KIFs were protective factors, and 16 KIFs were risk factors for glioma patients (Fig. 3A). Next, 11 OS-related KIFs were extracted after 1,000 iterations using LASSO Cox regression analysis (Fig. 3B-C). Finally, a prognosis model based on 5 KIFs including KIF41, KIF26A, KIF11, KIF13A and KIF13B, was established with the formula: riskscore = 0.3336 * KIF4A − 0.0988 * KIF26A − 0.08192 * KIF1A − 0.4628 * KIF13A + 0.3890 * KIF13B (Fig. 3D). Then, KM survival analysis was performed to compare the OS time between the high expression group and low expression group for each prognostic KIFs (Fig. 3E).

Evaluation and verification of the prognostic signature

The best-cutoff of riskscore were calculated using R “maxstat” package, with the standard that the minimum sample size > 25% and the maximum sample size < 75%, and we best cut-off value was 1.183056 in training cohort and − 0.61779624 in validation. Glioma patients were classified into high-risk group and low‐risk group according to the best cut-off value. KM survival analysis revealed that the OS time of the high-risk score group was significantly shorter than the low-risk score group both in both training (Fig. 4A; HR = 3.54, p = 2.8e-23) and validation cohorts (Fig. 4B; HR = 4.92, p = 1.2e-71). Consistently, with the increase of riskscore, the expression of KIF4A and KIF13B as well as the mortality rate of glioma patients increased remarkably (Fig. 4C-D). Furthermore, time-dependent curve analysis was performed using R “pROC” package, and AUC values corresponding for 1‐, 3‐, and 5-years of signature were 0.66, 0.78 and 0.75 in training cohort, and 0.77, 0.82 and 0.83 in validation (Fig. 4E-F). In addition, univariate and multivariate Cox regression analyses were performed to determine whether the riskscore of signature could be a prognostic factor for glioma independent of conventional clinical characteristics including age, gender, and grade, from which we can see that riskscore were related to the prognosis with the lowest p value not only in training cohort but also in validation cohort (Fig. 4G-H).

Gene set enrichment analysis

With the Hallmark gene set as a reference set, we performed GSEA analysis to explore the underlying oncogenetic difference between two groups. The result showed that the E2F target, mitotic spindle, epithelial-mesenchymal translational (EMT), G2M checkpoint and TNFa signaling via NFkB pathways were significantly activated in high risk patients (Fig. 5A). Furthermore, patients were calssified into high expression group and low expression group on the basis of median expression of 5 KIFs screened from multivariate Cox regression analysis. GSEA analysis was also performed between these groups, and results showed that complement, EMT, TNFa signaling via NFkB, estrogen response late and hypoxia were activated in low KIF26A patients (Fig. 5B), E2F target, G2M checkpoint, MYC targets, phosphorylation and spermatogenesis were activated in low KIF13B patients (Fig. 5C). In another hand, allograft rejection, complement, EMT, estrogen response late and hypoxia were found enriched in low KIF13A patients (Fig. 5D), EMT, TNFa signaling via NFkB, E2F target, inflammatory response and interferon gamma response were found enriched in high KIF4A patients (Fig. 5E), and complement, EMT, TNFa signaling via NFkB, hypoxia and inflammatory response were activated in low KIF1A patients (Fig. 5F).

Significant discrepancy of immune features between two risk groups

Based on R “estimate” package, a final immune score, stromal score, and estimate score were calculated, and we found that patients in high risk group had a high level of immune score, stromal score, and estimate score (Fig. 6A-C). Considering the important role of immune checkpoint inhibitor treatment in malignant tumor, we analyzed the correlation between riskscore and multiple checkpoint modules. The result showed a significant difference in many immune checkpoint modules between two risk groups (Fig. 6D). To be noted, patients in high risk group had a higher level of CTLA4, PD-1, PD-L1 and PD-L2 (Fig. 6E-H), indicating that patients with high and low riskscore might have a potential difference in immunotherapy response.

Mutation status in glioma patients between two risk groups

We calculated and visualized the level of TMB of all cancer types, from which a relatively low level of TMB were observed in glioma patients (Fig. 7A). After maftools analysis, no differential mutated genes with p < 0.05 were found in GBM cohort, while EGFR and NF1 mutated more frequently in high risk group and IDH1, CIC, NOTCH1 and FUBP1 were identified mutated more frequently in low risk group in LGG cohort (Fig. 7B). Moreover, significant co-occurrences and mutually exclusives were observed among mutations of these genes (Fig. 7C). Then, somatic mutations data of glioma samples from TCGA database were also analysised, and we found more somatic mutations including non-synonymous, synonymous mutations and all mutations enriched in patients with high risk (Fig. 7D-F). Besides, IDH1 was selected for further study due to its highest number of mutations and its key role in glioma development¹⁹. KM survival analysis was performed to compare the OS time between the IDH1-mutant group and IDH1-wild group, and result showed that patients with IDH1 mutant gliomas had a better prognosis (Fig. 8A). Differential expression analysis of 5 signature KIFs was also conducted between IDH1-mutant group and IDH1-wild group (Fig. 8B).

Immunohistochemistry based on HPA database

Immunocytochemistry showed KIF13A and KIF26A mainly localized in cytoplasm and membrane, and KIF13B and KIF4A mainly concentrated at membrane and nuclear that significantly upregulated in tumor samples. Despise of location, all signature KIFs found in HPA database differentially expressed.

Globally, well known for the high probability of metastasis and recurrence, glioma is still a leading cause of mortality in brain tumors ²⁰. Despite advances in diagnosis of glioma, most glioma patients can only be diagnosed at the intermediate and advanced stage when surgical resection is not an optimal option ²¹. Therefore, accurate patient stratification based on reliable prognostic factors is essential for decision of clinical management and determination of subset of glioma patients suitable for personalized treatment. In this study, we first developed a prognostic signature and successfully validated it based on a large combined cohort. Further analysis revealed that patients in two groups of the signature presented significant discrepancy of mutation status and immunity features, which might provide valuable information for predicting the prognosis and treatment of glioma patients, thereby helping physicians predict prognosis and guide treatment for patients.

The kinesin superfamily (KIFs), known as molecular motors with microtubule (MT) class-binding protein superfamily, play pivotal roles in several biological processes, including axonal transport and microtubule stabilization in our body, especially in the brain, that are important for various highly aggressive malignancies including glioma ²². Although several compounds targeting mitotic kinesins including the taxanes, vinca alkaloids, and epothilones, have been successfully used in treatment of a number of hematologic and solid malignancies, the same efficacy does not appear to glioma patients ²³. However, there is little share of the literature focused on the potential therapeutic and prognostic role of KIFs in glioma.

Here we identified 22 KIFs closely associated with glioma patients prognosis (6 protective and 16 risk factors) using Cox analysis. Further analysis resulted in the selection of five KIFs (KIF4A, KIF26A, KIF1A, KIF13A and KIF13B) with prognostic values. Among them, KIF4A and KIF13B were risk factors upregulated in the highrisk group, whereas KIF26A, KIF1A and KIF13A were protective factors downregulated in the highrisk group. KIF4A was found to be remarkably upregulated in primary colorectal carcinoma and contributes to proliferation of colorectal carcinoma through modulation of p21-mediated cell cycle progression ²⁴. A recent study reported that low expression of KIF26A had a positive correlation with distal metastasis and poor survival in patients with gastric cancer via regulating the focal-adhesion pathway and repressing the occurrence of epithelial-to-mesenchymal transition ²⁵. De S et al. revealed that overexpression of KIF1A was closely related to the cell resistance to docetaxel in breast cancer ²⁶. In addition, a target of miR-1290, KIF13B is typically involved in activation Wnt pathway and increased the expression of reprogramming-related transcript factors c-Myc and Nanog in colon cancer tissues ²⁷. Given both biological and tumorigenic functions of the five KIFs that were highly consistent with the results from bioinformatics analyses in our study, the prognostic value of five KIFs was demonstrated of valid predictions of clinical outcomes in glioma patients.

We further explored the underlying biological difference between patients in not only high and low risk group, but also in groups of different expression of five KIFs through GSEA and burden analysis. With the Hallmark gene set as a reference set, results of GSEA analysis showed that the E2F target, mitotic spindle, epithelial-mesenchymal translational (EMT), G2M checkpoint and TNFa signaling via NFkB pathways were significantly activated in high risk patients. EMT is the process that epithelial cells transform into motile mesenchymal cells, which is closely related to malignant biological behavior of cancer ²⁸. Recently, EMT has attracted much attention in tumorigenesis and cancer progression of glioma, and a previous study have been reported that the inhibition the process of EMT increased invasion and migration in glioma ²⁹. The mitotic spindle is a fundamental physiological process in cells responsible for the generation of a mitotic spindle with two spindle poles and daughter cells. The abnormal mitotic spindle in cancer cells may be related to the increase of cellular heterogeneity and metastasis as well as facilitation of transformation of cancerous cells ³⁰. G2/M checkpoint, as the DNA-damage in the cell cycle, is a limitation step of cell which is closely correlated with cancer cell growth and migration ³¹. All these results in our study indicated that the patients in high risk group may aberrantly activate the above pathways, leading to worse genomic and prognosis characteristics.

In addition, we found that patients with high riskscore was significantly positively correlated with immune score, stromal score, and estimate score and revealed a higher level of all main immune checkpoints, implying that patients in high risk group could be considered had a poor response to immunotherapy. Furthermore, mutation anslysis revealed that more somatic mutations including non-synonymous, synonymous mutations and all mutations enriched in patients with high risk, and IDH1 was the most significant hub gene associated with mutation. Then differential expression analysis of five signature KIFs was also conducted between IDH1-mutant group and IDH1-wild group. Finally, we also verified the five signature genes in the Human Protein Atlas (HPA) database, which identified of candidates for relevant biomarkers.

Briefly, we successfully developed a five KIFs based prognostic signature which could be used to stratify patients to further predict clinical outcomes and guide treatment. However, this study had some limitations need to be noted. First, the signature should be further validated in other prospective cohorts. Second, the five signature KIFs need to be further studied, including functional experiments and potential molecular mechanisms.

In summary, our study comprehensively evaluated the KIFs of glioma patients based on a combined cohort with a large population. The prognostic signature constructed by 5 KIFs was proved to be a powerful tool for predicting prognosis of glioma patients. We also explored the difference of biological pathway, immune features and mutation status between two risk groups. Finally, KIF4A, KIF13A, KIF13B and KIF26A were identified from HPA database, and all these KIFs differentially expressed between normal and tumore samples. In the future, these 5 KIFs may become more effective targets for the treatment of glioma, thereby providing a further personalized and accurate prognostic surveying tool.

Acknowledgements

All authors would like to thank the specimen donors and research groups for TCGA and CGGA.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Availability of data and materials

Supplementary material for this article is available online.

Authors' contributions

PD and FX collected the data and performed all analysis. PD and LJ wrote the manuscript. All the

authors participated in the data analysis and approved the final version of the manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The author reports no conflicts of interest in this work.

Data Availability Statement

All data reported are contained in the manuscript, including RNA sequencing and clinical information retrieved from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/), Genotype Tissue Expression (GTEx) database (https://gtexportal.org/) and Chinese Glioma Genome Atlas (CGGA) database (https://cgga.org.cn/).

Ostrom Q, Bauchet L, Davis F, et al. The epidemiology of glioma in adults: a "state of the science" review[J]. Neuro-oncology, 2014, 16(7): 896–913.
Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries[J]. CA: a cancer journal for clinicians, 2021, 71(3): 209–249.
Louis D, Perry A, Wesseling P, et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary[J]. Neuro-oncology, 2021, 23(8): 1231–1251.
Jiang G, Mu J, Liu X, et al. Prognostic value of miR-21 in gliomas: comprehensive study based on meta-analysis and TCGA dataset validation[J]. Sci Rep, 2020, 10(1): 4220–4220.
Ostrom Q, Cioffi G, Waite K, et al. CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2014–2018[J]. Neuro-oncology, 2021, 23(iii1-iii105.
Arvanitis C, Ferraro G, Jain R. The blood-brain barrier and blood-tumour barrier in brain tumours and metastases[J]. Nature reviews Cancer, 2020, 20(1): 26–41.
Lapointe S, Perry A, Butowski NA. Primary brain tumours in adults[J]. Lancet (London, England), 2018, 392(10145): 432–446.
Hirokawa N, Noda Y, Tanaka Y, et al. Kinesin superfamily motor proteins and intracellular transport[J]. Nature reviews Molecular cell biology, 2009, 10(10): 682–696.
Miki H, Setou M, Kaneshiro K, et al. All kinesin superfamily protein, KIF, genes in mouse and human[J]. Proceedings of the National Academy of Sciences of the United States of America, 2001, 98(13): 7004–7011.
Liu H, Chen C, Fehm T, et al. Identifying Mitotic Kinesins as Potential Prognostic Biomarkers in Ovarian Cancer Using Bioinformatic Analyses[J]. Diagnostics (Basel, Switzerland), 2022, 12(2):
Zhong Y, Jiang L, Lin H, et al. Overexpression of KIF18A promotes cell proliferation, inhibits apoptosis, and independently predicts unfavorable prognosis in lung adenocarcinoma[J]. IUBMB life, 2019, 71(7): 942–955.
Sheng J, Xue X, Jiang K. Knockdown of Kinase Family 15 Inhibits Cancer Cell Proliferation In vitro and its Clinical Relevance in Triple-Negative Breast Cancer[J]. Current molecular medicine, 2019, 19(2): 147–155.
Takahashi S, Fusaki N, Ohta S, et al. Downregulation of KIF23 suppresses glioma proliferation[J]. Journal of neuro-oncology, 2012, 106(3): 519–529.
Cho SY, Kim S, Kim G, et al. Integrative analysis of KIF4A, 9, 18A, and 23 and their clinical significance in low-grade glioma and glioblastoma[J]. Sci Rep, 2019, 9(1): 4599.
Venere M, Horbinski C, Crish JF, et al. The mitotic kinesin KIF11 is a driver of invasion, proliferation, and self-renewal in glioblastoma[J]. Science translational medicine, 2015, 7(304): 304ra143.
Vivian J, Rao AA, Nothaft FA, et al. Toil enables reproducible, open source, big biomedical data analyses[J]. Nat Biotechnol, 2017, 35(4): 314–316.
Tibshirani R. The lasso method for variable selection in the Cox model[J]. Statistics in medicine, 1997, 16(4): 385–395.
Uhlen M, Zhang C, Lee S, et al. A pathology atlas of the human cancer transcriptome[J]. Science (New York, NY), 2017, 357(6352):
Yu D, Boekholder R, VanBrocklin M, et al. TMOD-19. MUTANT IDH1 PROMOTES GLIOMA FORMATION IN VIVO[J]. Neuro-Oncology, 2018, 20(Suppl 6): vi272-vi272.
Zhang X, Wei X, Liu J, et al. Up-regulation of long non-coding RNA BLACAT1 predicts aggressive clinicopathologic characteristics and poor prognosis of glioma[J]. Medicine (Baltimore), 2021, 100(11): e20722-e20722.
Zhang Q, Xu B, Hu F, et al. Tenascin C Promotes Glioma Cell Malignant Behavior and Inhibits Chemosensitivity to Paclitaxel via Activation of the PI3K/AKT Signaling Pathway[J]. J Mol Neurosci, 2021, 71(8): 1636–1647.
Yang Y, Gao L, Weng N-N, et al. Identification of Novel Molecular Therapeutic Targets and Their Potential Prognostic Biomarkers Among Kinesin Superfamily of Proteins in Pancreatic Ductal Adenocarcinoma[J]. Front Oncol, 2021, 11(708900–708900.
Jackson JR, Patrick DR, Dar MM, et al. Targeted anti-mitotic therapies: can we improve on tubulin agents?[J]. Nat Rev Cancer, 2007, 7(2): 107–117.
Hou P, Jiang T, Chen F, et al. KIF4A facilitates cell proliferation via induction of p21-mediated cell cycle progression and promotes metastasis in colorectal cancer[J]. Cell death & disease, 2018, 9(5): 477.
Ma R, Zhang H, Chen H, et al. MiR-19a/miR-96-mediated low expression of KIF26A suppresses metastasis by regulating FAK pathway in gastric cancer[J]. Oncogene, 2021, 40(14): 2524–2538.
De S, Cipriano R, Jackson M, et al. Overexpression of kinesins mediates docetaxel resistance in breast cancer cells[J]. Cancer research, 2009, 69(20): 8035–8042.
Wu J, Ji X, Zhu L, et al. Up-regulation of microRNA-1290 impairs cytokinesis and affects the reprogramming of colon cancer cells[J]. Cancer letters, 2013, 329(2): 155–163.
Nisticò P, Bissell MJ, Radisky DC. Epithelial-mesenchymal transition: general principles and pathological relevance with special emphasis on the role of matrix metalloproteinases[J]. Cold Spring Harbor perspectives in biology, 2012, 4(2):
Wang B, Zhang X, Wang W, et al. Forkhead box K2 inhibits the proliferation, migration, and invasion of human glioma cells and predicts a favorable prognosis[J]. Onco Targets Ther, 2018, 11(1067–1075.
Chou J, Lin Y-C, Kim J, et al. Nasopharyngeal carcinoma–review of the molecular mechanisms of tumorigenesis[J]. Head Neck, 2008, 30(7): 946–963.
Mongre RK, Mishra CB, Prakash A, et al. Novel Carbazole-Piperazine Hybrid Small Molecule Induces Apoptosis by Targeting BCL-2 and Inhibits Tumor Progression in Lung Adenocarcinoma in Vitro and Xenograft Mice Model[J]. Cancers (Basel), 2019, 11(9): 1245.

No competing interests reported.

SupplementaryTable1.xls

Download PDF

Version 1

posted

You are reading this latest preprint version

Comprehensively analysis exploring the role of kinesin family members in glioma

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

Introduction

Methods

Data acquisition and processing

Differential expression analysis

Construction of KIFs based prognostic signature

Validation of the prognostic signature

Functional enrichment analysis

Exploration of immune features

Mutation signature analysis

Immunohistochemistry

Results

Data combination and KIFs quantification

Construction of KIFs based prognostic signature

Evaluation and verification of the prognostic signature

Gene set enrichment analysis

Significant discrepancy of immune features between two risk groups

Mutation status in glioma patients between two risk groups

Immunohistochemistry based on HPA database

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1