Expression of cyclin-dependent kinases and their clinical significance with immune infiltrates could predict prognosis in colorectal cancer

Highlights • The expression and prognostic values of AURKA and RB1 may also be significant to CRC diagnosis than previously studies.• The association of CDKs with immune infiltrates may serve as target molecules for immunotherapy in CRC.• The expression of CDK is significant among CRC subtypes and therefore, it can be inferred as a potential biomarker in the cancer subtype.• An increase in tumor purity was positively correlated with the expression of CDK-1 in COAD due to CD4+ cells and CDK-4 in COAD and READ resulting from a fraction of immune cells.


Introduction
Colorectal carcinogenesis is a sequential process that results from genetic alterations in either oncogenes or tumor-suppressor genes. Molecular alterations of genes have also been implicated in the development and prognosis of CRC. The survival rate of CRC patients relies on the stage during diagnosis and decreases as the tumor advances [1]. To date, the mechanisms that drive the pathogenesis of CRC are poorly understood and the development of resistance to cytotoxic and targeted therapies remains the primary reason for treatment failure. Therefore, the development of new therapeutic targets that can translate into better clinical outcomes is urgently required for individuals with CRC. Additionally, the identification of biomarkers for early diagnosis and prognosis could improve the efficiency of current treatments for CRC and represent molecular markers for therapeutic targets.
Cyclin-dependent kinases (CDKs) are such potential biomarkers for the diagnosis of CRC due to their crucial role in the elongation step of the global transcription process [2]. CDKs and their specific functions in cancer have been comprehensively investigated; they are involved in regulation of the cell cycle by interacting with specific cell-cycle-regulatory cyclins [3]. The CDK family consists of 20 members (CDK-1 to . This class of serine/threonine kinases has been classified into two groups based on their involvement in tumor formation through the regulation of cellcycle or transcription [4][5][6]. CDK-1, CDK-4, and CDK-5 fall within the cell-cycle-related subfamilies while CDK-7, CDK-8, CDK-9, CDK-11, and CDK-20 are classified as transcriptional subfamilies. Malumbres and Barbacid [7] reported that mammalian cells require a step-wise activation of at least four different CDKs to drive cells through interphase, and for CDK-1 to proceed through mitosis. CDK-1 promotes the G 2 -M transition and regulates G 1 progression and G 1 -S transition [8,9], suggesting that CDK-1 is one of the most important CDKs for cell-cycle regulation [10]. Uncontrollable cell proliferation is one of the hallmarks of malignant tumors and is often driven by mutations in CDK activity. Altered CDK expression and/or activity is observed in many cancer subtypes [11,12]. In addition to cell-cycle regulation, CDK-1 may regulate apoptosis by inhibiting phosphorylation of caspase-8 [13,14]. CDK-1 functions include apoptosis, senescence, angiogenesis, cell cycle progression, and adhesion/migration [15][16][17], all of which are believed to contribute to its role in tumorigenesis. CDK-1 has been implicated in several cancers, including ovarian [18,19], thyroid [20], bladder [21], colorectum [22], breast [23,24], and liver cancer [25]. CDK-4 is also a key player in cell cycle and mediates cell progression through the G1 phase [26]. Its dysregulation has been found to play a significant role as regulator of signaling in cellular pathways in several diseases. Since the frequent deregulation of CDK-4 in cancer often leads to addiction to its activity, it is emerging as a bright therapeutic target [27,28]. CDK inhibitors were recently investigated in preclinical and clinical trials. These inhibitors are believed to act as anticancer drugs by blocking CDKs to prevent uncontrolled cellular proliferation that is a hallmark of cancers specifically in CRC [29]. In earlier studies, the relationship between genetic mutations and CDK overexpression were investigated [30]. The overexpression of CDK4 was observed in colorectal tumors that carry the type II transforming growth factor-β receptor mutation [31]. This mutation resulted in microsatellite instability in 15 % CRC patients [31]. Zhao et al. [32] reported the expression of CDK-4 in 74 specimens through grading, and concluded that stronger immunostaining for CDK-4 was predictive of a worse prognosis (P < 0.001). Tumor immune cells play essential roles in cancer development and progression. However, the association with prognosis in CRC remains elusive [33]. A recent study showed that the expression of CDK-9 was negatively associated with the infiltration of CD8 + T cells at the tumor site in CRC patients [34]. Thus, CDKs may have roles in tumorigenesis, progression, and metastasis, and may serve as an underlying prognostic factor for tumors. Since the TCGA transcriptomic datasets have proven the relevance of CDKs in CRC [35], targeting CDKs could prove an effective strategy to enhance the efficacy of treatment in CRC. With clear emphasis on CDK-1 and CDK-4, the association between CDKs expression and clinical pathological parameters, including prognostic value, was investigated, and the biological function and immune infiltrate activities of CDKs in patients with CRC was examined using an in silico approach (Fig. 1). This study suggests that the involvement of CDKs in CRC has a pivotal clinical significance in CRC.

mRNA expression analysis
The web-based Oncomine database accessed at https://www. oncomine.org/resource/login.html, was used to investigate the transcription levels of CDK-1 and CDK-4 in CRC using the following parameters: p<1 Â 10 -4 , fold-change >2 and gene ranking in the top 10 %. The mRNA expressions were compared between normal and tumor clinical samples.

UALCAN database analysis
UALCAN is a comprehensive, user-friendly, and interactive web resource for analyzing cancer OMICS data [36]. The correlation between CDKs expression and lymph node metastasis in colon and rectum adenocarcinomas was analyzed using UALCAN (http:// ualcan.path.uab.edu/). CDK-1 expression and the clinicopathological parameters relationship was provided in a tabular form.

Predicted protein interaction
Co-expressed proteins were identified according to a method previously described by Fadaka et al. [37]. The database was accessed, downloaded and installed from http://www.cytoscape. org/ (v3.7.2) along with the STRING plug-in (http://string-db.org/ v11). Only genes with a 0.90 confidence level as the minimum required interaction score were considered.

Gene ontology
FunRich is a stand-alone software tool used mainly for functional enrichment and interaction network analysis of genes and proteins [38]. The enrichment analyses of gene ontology such as cellular component (CC), biological processes (BP), molecular function (MF) and biological pathway were analyzed using the functional enrichment (v3.1.3) software accessed at http://funrich. org/index.html.

Mutations analysis for CDK-1 and CDK-4
The Catalogue of Somatic Mutations in Cancer (COSMIC) database coupled with cBioPortal was used to detect genetic mutation frequencies and hotspots for CDK-1 and CDK-4. The COSMIC v90, (released 05-SEP-19) database accessed at https:// cancer.sanger.ac.uk/cosmic is a comprehensive resource for exploring the effect of somatic mutations in all human cancer subtypes [39,40]. This database also provides information pictorially on the distribution and substitutions on the coding strand in cancer subtypes.

Immune Cells Tumor-Infiltration analysis
Tumor Immune Estimation Resource (TIMER v 2.0), a webbased database for comprehensive analysis of tumor-infiltrating immune cells [49,50] (https://cistrome.shinyapps.io/timer/) was used to assessed the interactions between the genes of interest and the immune system for better understanding of CDKs-immune interactions in CRC. Six immune infiltrates namely; B cells, CD4 + T cells, CD8 + T cells, Neutrophils, Macrophages and Dendritic cells were analyzed using this algorithm. Furthermore, the expression report of Oncomine was validated by TIMER v2.0.

Statistical analysis
Based on the Bioinformatics approaches employed, the Kaplan-Meier plot was generated by PrognoScan, GEPIA, and UCSC-Xena. PrognoScan and GEPIA analyses were displayed along with Hazzard ratio (HR) and p-values of the log-rank test. Expression analysis results were presented as fold changes, p-values, and ranks. The difference in expression was verified using Student's t-test. Differences between and within groups were carried out using ANOVA for multiple tissues. The correlation of gene expression was assessed by statistical significance, Pearson's correlation, and spearman's correlation. The scatter/ boxplot was created using TIMER computational tool and was calculated using spearman's correlation between two groups of genes of interest. The significance level was set as p < 0.05.

Expression of CDKs
To determine the expression of CDKs among subtypes of CRC, it is important to explore their expression among different cancer types (Figs. 2a, b and 3 ). The expression of these genes was investigated in both solid tumors as well as hematological malignancies (Fig. 2). The expression of CDK-1 was upregulated in 17 cancers including breast cancer, CRC, lung cancer and sarcoma in total unique analyses of 455 different datasets. Underexpression of this gene was also identified in brain and central nervous system (CNS) cancer, breast cancer and leukemia. Although the expression of CDK-4 was determined in 463 different analyses, its expression was up-regulated in 53 cancer datasets and downregulated in two cancer studies (leukemia). The pathogenesis of CRC is unique among human tumors. This common malignant tumor usually develops in a benign precursor lesion, the adenoma, which is visible on the mucosal surface of the large bowel, constituting an adenoma-adenocarcinoma sequence. The expression of CDK-1 and CDK-4 were significantly elevated in multiple CRC types when compared to normal colorectum tissues (p = 1.0 Â 10 À4 ). These CRC types include Colon Adenocarcinoma, Cecum Adenocarcinoma, Cecum Adenocarcinoma, Rectosigmoid Adenocarcinoma, Colon Mucinous Adenocarcinoma, Rectal Mucinous Adenocarcinoma, Rectal Adenocarcinoma, Colon Adenocarcinoma, Colon Adenoma, Rectum Adenoma, Colon Adenocarcinoma, Rectal Adenocarcinoma, Colorectal Carcinoma and Colorectal Adenocarcinoma with different p-values and fold changes (Table 1 and Fig. 5).
These expressions were further validated using the computational tool (TIMER v 2.0) (Fig. 2b). Clear emphasis were placed on CRC as the cancer type of interest and their expressions were evaluated using the GEPIA database (Fig. 3b).

Genetic alterations of CDKs and clinicopathological parameters in CRC patients
To understand the importance and mechanism underlying the expression of CDKs in CRC, the expression profiles of CDKs were examined across CRC subtypes using OMICS data (obtained from the UALCAN database) based on different clinical-pathological parameters ( Table 2). This analysis differs from Oncomine expression analysis in Table 1 which evaluated the expression of CRC subtypes compared to normal counterpart. Table 2 specifically investigated normal expression of CDK 1 and 4 in selected clinicpathological parameters. Based on gender, an increase in CDK expression was observed in males and females with CRC when compared to individuals without the disease. The age criterion demonstrated significantly increased expression of CDK genes. mRNA levels in tumors of patients of all ages >20 -100 years, compared with normal individuals of all ages. In addition, the CDKs mRNA expression was significantly increased in sample types (p < 1.0 Â 10 À12 ), race, weight, tumor stages, histological subtypes, and nodal metastasis status compared with the corresponding normal tissues ( Table 2). The nodal metastatic status of these genes was represented with a box plot using the CRC subtypes Colon adenocarcinoma (COAD) and Rectum adenocarcinoma (READ) (Fig. 4).

Survival analysis
The prognostic values of CDK-1 and CDK-4 were assessed using the PrognoScan-based Kaplan-Meier analyses. The overall survival (OS) and disease-free survival (DFS) of CDK-1 were determined in five different datasets (p < 0.05) in CRC. The results obtained were consistent across the five datasets and confirmed that the expression of CDK-1 positively correlated with poor OS rates in CRC patients (Fig. 6). In contrast, when the prognostic value of  CDK-4 was assessed in three different databases using the same bioinformatics tools (Fig. 7B-D), the expression of CDK-4 in the GSE14333 dataset inversely correlated to poor prognosis in disease-free survival (DFS) in CRC (Fig. 7). Although CDK-1 and CDK-4 were positively correlated, their prognostic values with respect to mRNA expression are different ( Fig. 7A and Supplementary Fig. 1).

Protein partners of CDKs
Using CDK-1 and CDK-4 as a queries, associated proteins were generated using STRING v11 plugin Cytoscape 3.7.2with a 0.90 confidence level as the minimum required interaction. Each of these genes generated 10 other genes through curated database entries, experimental validation, text mining, co-expression and/or protein homology (Fig. 8).

Gene ontology of associated genes
To identify functional categories and characteristic biological attributes of CDKs (Fig. 8), Gene Oncology (GO) enrichment analysis was performed using the FunRich software.
Analysis of GO terms were filtered for cellular component (CC), biological process (BP), molecular function (MF) and biological pathway (BP) (Fig. 9). The top 10 GO terms of the associated genes in BP, CC, MF and BP were considered significant at p-value of < 0.05 (Fig. 9A-D). These genes were involved in the different GO terms such as cell communication, ontology; BP, kinase binding, Note: Adenoma is a benign for of adenocarcinoma. Sometimes this counterpart transforms into adenocarcinomas. ontology; MF, cyclin-dependent protein kinase holoenzyme complex, ontology; CC and cell cycle, ontology; BP. The biological pathway analysis showed that cell cycle was highly enriched amongst the CDKs associated genes, which are closely related to FOXM1 transcription factor network, regulation of retinoblastoma protein and the E2F transcription factor network (Fig. 9D).

Genomic correlation of CDK genes in clinical data
To lower the barriers of access to the complex data sets and accelerate the translation of genomic data into new biological insights, therapies, and clinical trials, the cBioPortal database was used to profile the CDK genes in ten CRC studies. This software tool accessed at https://www.cbioportal.org/ reduces molecular profiling data from cancer tissues and cell lines into readily understandable genetic, epigenetic, gene expression and proteomic events. The patterns of gene alterations were visualized in specific CRC studies and all the relevant genomic alterations in these studies were reported (Figs. 10 and 11, Tables 3 and 4). The results are represented as a bar graph plot of alteration frequency (%) against mutation data, copy number alteration and study data (Fig. 10). The mutation observed after the query of the CDKs include mutation, amplification, deep deletion and multiple alteration. Table 3 shows the alteration frequency of CDKs in different CRC data sources. This table reported the percentages of occurrence in selected case number, in different types of alteration. In COAD (CPTAC-2, 2019) with a sample number of 105, the frequency of aleration observed were 17.14 %, 15.25 %, and 10.48 % for mutation, deep deletion and amplification respectively. Overall the highest form of alteration was observed for mutation in all the CRC cases except in Colorectal (TCGA) data where 14.62 % for amplification was observed as opposed 13.21 % for mutation. Fig. 11 shows the percentages of alterations in CDK-4, CCNE1, CDKN2A, CCND2, RB1, CDK-1, BUB1, NCAPG, AURKA and KIF11 genes in the CRC. Grey bars along a vertical line represent the same sample interrogated for genetic alteration. With two combined studies containing 704 samples, 25 % of the queried genes were altered (178). The mutations observed in CRC were deep deletion, missenses and truncation mutations. Also, CRC may be thought to frequently amplify CDKs. With respect to the 10 querried genes, AURKA was frequently altered in CRC with 8% frequency of alteration with amplification as the major alteration followed by RB1 with 5% alterations (amplification, missense, deep deletion and truncating mutation). However, among there genes, CDKN2A, CDK1 and CDK4 were least altered with 1.9 %, 1.7 % and 1.4 % frequencies of alteration respectively. In addition, the percentages of alteration frequencies of five selected CRC samples (TCGA, Firehose; TCGA, Nature; Pan Cancer; MSKCC and CPTAC-2 were further investigated. The result shows that AURKA has the highest alteration frequency across all five samples which sugegest that the AURKA gene may be a good diagnostic biomarker for CRC (Table 4).

Gene hotspot mutation for CDKs
The mutations of CDK-1 and CDK-4 were carried out using the COSMIC database for CRC coupled with the cBioPortal database for hotspot mutation in human cancers. These databases were queried for CDK-1 and CDK-4 gene mutations. The pie charts illustrate that the mutant types of CRC were primarily missense and synonymous substitutions (Figs. 12A and 13 A). The CDK-1 and CDK-4 CRC data contained G > A and C > T mutations, each accounting for over 20 % of the genes coding strand (Figs. 12A and 13 A). Using cBioPortal, 85,583 combined studies involving 82,346 patients from 277 studies of human cancers were investigated. CDK mutations across these studies reported a total of 169 and 302 mutations for CDK-1 and CDK-4 respectively. The mutation sites for CDK-1 was detected  (Figs. 12B and 13 B).

Immune infiltration analysis
TIMER was used to investigate the molecular characterization of tumor-immune cell interaction. The genes of interest were used  as queries against the six immune infiltrates using two cancer types (COAD and READ). The output of each query generated scatter plots for each of the infiltrates in COAD and READ (Fig. 14). In COAD, CDK-1 and CDK-4 clearly shows significant positive correlations with tumor purity while in READ, only CDK-4 was positively correlated with tumor purity. The expression levels of each gene were displayed against tumor purity (left). This is because genes that are highly expressed in the microenvironment are expected to have negative association with tumor purity. For each immune cell, the partial Spearman's correlation was adjusted by tumor purity (Fig. 14).

Discussion
CDKs participate in the regulation of cell cycle progression and transcription, and aberrant expression of these genes is often observed in various human tumors. Dysregulation in cell cycle progression has been associated with bowel (small and large intestine) diseases particularly colon and rectum cancers representing one-fifth of all cancers globally [57]. Unlike normal cells, cancer cells grow uncontrollably without regulation by cell-cycle due to a loss of function of checkpoint integrity. This could be due to CDK inhibitor inactivation leading to overexpression of either cyclin or CDKs. Therefore, these kinases are potential targets for cancer theranostics. Using bioinformatics approaches, this study investigated the roles of CDKs in relation to their expressions and immune infilterates in CRC. Most adenocarcinomas of the colorectum arise in a visible benign precursor lesion, the adenoma, which is a monoclonal proliferation of dysplastic nonmalignant epithelial cells [58]. Therefore, both adenoma and adenocarcinoma of the colorectum were considered in this study.
The expression of CDKs was explored in 20 different cancer subtypes with emphasis on different sections of the colorectum, using different computational tools. Oncomine, TIMER and GEPIA webbased tools were used to evaluate and validate the mRNA expression of CDK-1 and CDK-4 in different cancer types (solid tumor and hematological malignancies), different CRC sections as well as clinico-pathological parameters of CRC in relation to normal tissues. Relative to normal colorectum tissue, gene expression of CDK-1 and CDK-4 were up-regulated in all major forms of cancer with high mRNA expressions observed in CRC. Xue et al. [59] revealed that the expression of CDK-1 was exceptionally high in CRC in contrast with normal tissues, and it predicts distance metastasis risk in CRC; CDK-1 can also alter the progression of CRC through phosphorylation of JAK1 to initiate the JAK/STAT3 signaling pathway. In addition, high expression of CDK-1 reportedly stimulate the proliferation and migration of CRC while its inhibition restricted CRC proliferation [60]. Furthermore, frequent CDK-1 expression was correlated with therapeutic target, and the therapeutic resistance of BRAF mutant human CRC can be suppressed by targeting CDK-1 [3]. Similarly, poor prognosis was observed in patients with high expression of CDK-4 (P < 0.001) [61]. Therefore, inhibition of these kinases may be useful in  cancer therapy and their expression patterns could become a novel prognostic marker in CRC. The protein expression level was also explored for reliable prognostic values for CRC patients using in silico web-based servers. Bioinformatics analyses have provided useful results based on large experimentally validated data accesible in various databases. Gene expression prognostic signature for high-grade serous ovarian cancer was investigated by resampling/crossvalidation method with Cox regression analysis followed by Kaplan-Meier analysis and log-rank testing [62]. Similarly, He et al. [63] probed the survival analysis of the expression level of MANF in HCC using the Kaplan-Meier method. In addition, Cox Moreover, high mRNA expression of CDK-1 is linked to poor OS in CRC, hepatocellular carcinoma and lung cancer [65][66][67]. This result revealed that CDK-1 expression levels in CRC correlated with the poor OS of patients with the disease. In addition, high expression of CDK-1 is a prognostic factor for CRC. Protein associations have been previously reported to be pivotal in biological processes and cellular function in both prokaryotes and eukaryotes including humans (normal and disease states) [68]. The discovery of associated proteins with specific gene(s) of interest are important to understand the mode of action of diseases such as cancer to improve the development of therapeutic drugs and treatmentsstrategies. Numerous experimental approaches have been deployed for the identification of interacting genes but the processes are demanding with limited outputs. However, bioinformatics approach has been increasingly used to validate and predict elusive protein partners. Therefore, this study identified twenty associated proteins with CDK-1 and CDK-4 using STRING v11 plugin Cytoscape 3.7.2 since proteins are constantly regulated and rarely function in isolation. These genes are crucial to the roles and specific molecular networks of CDK-1 and CDK-4 in diseases, most importantly CRC.
The ten associated proteins identified for CDK-1 include       The study identified a panel of genes which could predict RB1 status in both cultured cancer cell lines and clinical tumor samples. AURKA genes was studied to provide cancer therapeutic interventions due to their significant expression in cancer patient samples [70]. Similarly, Korobeynikov et al. [71] suggested that the dual targeting of AURKA and PAK1 may be a promising therapeutic strategy for cancer treatment from a combinatory study of alisertib and FRAX1036. Another study by Ding et al. [72] shows that MAD2L1 may be a novel oncogene in an in vitro 20 clinical CRC samples. Since previous studies have predicted the association of these genes as promising targets in cancer therapy [73,74], CDKs together with their protein partners could serve as a confirmation of the strength of the methodology used in this study. The discovery of pathways and specific processes that are significant with factors regulating activities to genes of interest is important in cancer research. FunRich software was used to evaluate the enrichment of the protein partners in BP, MF and CC. These genes were involved in: a) cell communication and signal transduction in BP, b) kinase binding and kinase regulator activity in MF, c) cyclin-dependent protein kinase holoenzyme complex and nucleoplasm in CC and d) cell cycle, mitotic and FOXM1 transcription factor network in BP. These processes can be further probed for their involvement in CRC in order to tailor specific strategies for CRC therapeutics.
In a combined study of 3,953 samples, the associated genes were altered in 21 % (817) of queried samples reporting 46 % altered cases as the alteration frequency (cancer types summary). From the bar chart, the alterations include mutation, amplification, deep deletion and multiple alterations. The mutation data and copy number alterations were also represented. The ratio of alteration ranged between 1.45 and 45.71 % with the dominance hierarchy as Colon Cancer (CPTAC-2 Prospective, Cell 2019), Colorectal Adenocarcinoma (TCGA, Nature 2012) and Colorectal Adenocarcinoma (TCGA Firehose Legacy). Furthermore, alterations in the associated genes (CDK-1, CDK-4, CCNE1, CDKN2A, RBI, NCAPG, AURKA, and CCND2) were determined in five CRC studies (CRC, TCGA, Firehose, CRC, TCGA, Nature, CRC, Pan-Cancer, MSKCC and CPTAC-2). The percentages of alterations in these genes ranged from 0 to 11% for individual genes. AURKA and RB1 genes were amplified predominantly in the combined studies, suggesting that their expression could be significant to CRC prognosis. Since these genes are positively correlated with CDK genes, their prognostic values deserve to be assessed for theranostic purposes.
The alterations frequencies and hotspot mutations of CDK-1 and CDK-4 were analyzed with COSMIC and cBioPortal databases. The results revealed that the types of mutations of CDK-1 and CDK-4 in the CRC data were missense mutations among others. Furthermore, the alteration frequency of these genes in CRC is notably low. CDK genes were altered in 2,537 of the queried samples (3%), with a somatic mutation frequency of 0.2 % for both CDK-1 and CDK-4. For CDK-1, 169 mutations including 130 missenses, 32 truncating, 4 inframe insertions, and 3 others with E163* hotspot were observed. In a total of 302 mutations for CDK-4, including 264 missenses, 17 truncating, 7 in-frame insertions, and 14 others its hotspot was identified as a R24A/C/H/L hotspot. The clinical implications of these mutation hotspots are likely to be oncogenic with gain of function as the biological effect. CDK-1 and CDK-4 are significant kinases in M-phase andG1 stages of cell-cycle respectively. Unlike the aberrant expression of CDK-1 in stage-2 colon and rectum adenocarcinoma being associated with a higher risk of metastasis [75], mutations in the catalytic subunit of CDK-4 gene has also been reported to be common and its abnormal expression has been uncovered in over 90 % of all bowel cancers [76]. The identified hotspots for these kinases may be a promising drug target.
During tumor development and progression, cancer and immune cells interact by multiple genes and pathways. Therefore, it is worthy to explore the correlation between gene expression and immune infiltration level. Tumor purity plays an important role in understanding the pathogenic mechanism of tumors [77]. This is a crucial characteristic that cannot be ignored in cancer genomics or epigenomics data analysis [78]. It is noteworthy that the increase in tumor purity was inversely correlated with the expression of CDK-1 in READ. Specifically, the immune infiltration analysis may have resulted from an increase in the numbers of infiltrating B cells, CD8+cells, neutrophils, and dendritic cells. Conversely, an increase in tumor purity was positively correlated with the expression of CDK-1 in COAD due to CD4+cells and CDK-4 in COAD and READ resulting from a fraction of immune cells.

Conclusion
Dysregulation of kinase genes is commonly encountered in bowel cancers. In this study, CDK-1 and CDK-4 genes were overexpressed in all the CRC subtypes with respect to normal tissues and the expression of CDK-1 was associated with poor prognosis as opposed to CDK-4. Based on the associated proteins, RB1 and AURKA were positively correlated and showed significant relationships with CDKs and their expressions were significantly altered in CRC. The most frequent alteration was found to be a mutation with hotspots E163* and R24A located on the Pkinase domain for both CDK-1 and CDK-4 respectively. In addition, one of the crucial findings of the current study was the identification of a positive correlation between CDK-1 and increased numbers of infiltrating B cells, CD8+ cells, neutrophils and dendritic cells and a contrasting correlation between CDK-4 and the immune infiltrates in both CRC subtypes. This is a potential novel feature of CRC in the context of the immune response, driven by CDK-1 and CDK-4. In concluson, the results of this study justified exploring CDKs as prognostic markers as well as immunotherapeutic target for CRC as well as other cancers. The role of these genes in CRC and the potential molecular mechanisms deserve further molecular investigation.

Data availability
The datasets generated during and/or analysed during the current study are presented in the manuscript. All other databases used for data generation are duly referenced and their links are also provided.

Ethical approval and consent to participate
Not applicable.

Consent for publication
Not applicable.

Funding
Not applicable.

Declaration of Competing Interest
The authors report no declarations of interest.