CRC is one of the most common malignancies, with multiple common types. Despite surgical resection, chemotherapy, and radiotherapy used in CRC treatment, the prognosis of CRC remains poor and CRC has high morbidity and mortality rates [28]. In the present study, we established a prognostic model of COAD and READ according to the OS and tumor staging. The clinical implementation of this developed risk score model is likely to provide more relevant information for clinical diagnoses and may improve the prognosis of CRC patients. ALAD and CHRM2 were considered as hub genes in the PPT network. Furthermore, aminolevulinic acid, levulinic, and loxapine may be potential drugs for CRC treatment. Meanwhile, KAZALD1, HPCAL4, CDH8, SYNPO2, CCND3, and hsa_mir_26b were significantly associated with the OS of CRC patients, and served as potential independent prognostic factors.
PPI analysis showed that ALAD and CHRM2 were connected with the development of CRC, and were also related to the prognosis of CRC patients. ALAD catalyzed the second step in heme biosynthesis and was an inhibitor of the 26S proteasome [29]. Proteasome has been found to be associated with cancer and is a therapeutic target for cancer [30]. Meanwhile, the genetic variants in ALAD are connected with the risk of genitourinary cancers [31]. Neslund et al. indicated the potential interactions between ALAD genetic variants and occupational Pb exposure, suggesting that ALAD may be a current target of prostate cancer [32]. Additionally, a polymorphism in ALAD genes may affect lead toxicokinetics; Van Bemmel et al. revealed that ALAD may alter the risk of renal cell carcinoma related to lead exposure [33]. ALAD has also been connected with prognosis of cancers, such as breast cancer [34]. Drug prediction analysis showed that ALAD was targeted by aminolevulinic acid (ALA) and levulinic acid. ALA is an endogenous metabolite formed by succinyl-CoA and glycine in mitochondria. ALA, as an already approved therapeutic strategy, is considered as a successful prodrug used in cancer treatment. The photosensitizer excited by ALA could contribute to the generation of reactive oxygen species, which induces oxidative damage and consequently leads to cell death [35]. Ferreira et al. indicated that melanoma cancer could be treated by photodynamic therapy using ALA [36]. Prior studies observed that 5-ALA was a promising candidate for CRC diagnosis, and 5-ALA mediated photodynamic therapy could be a therapeutic strategy for CRC [37, 38], which is consistent with our finding that ALA is an effective drug for CRC treatment. Another drug, levulinic acid, was detected in this study. Evidence demonstrated that levulinic acid has antimicrobial activity, and could reduce Salmonella contamination on the surface of various materials. Salmonella infection is the common cause of foodborne diseases [39]; meanwhile, food-borne mutagens and pollution are known risk factors for CRC [40]. Thus, we speculated that levulinic acid may be a potential drug for CRC therapy. CHRM2 belongs to a larger family of G protein-coupled receptors with adenylate cyclase inhibitory effect [41]. Furthermore, the CHRM2 gene is prone to alcohol dependence and drug dependence [42] and may affect the internal nerves of gastrointestinal function [43]. However, there have been no reports on the association of CHRM2 and CRC; thus, further experiments should be performed to explore the effect of CHRM2 on CRC. We also found that CHRM2 was involved in the calcium signaling pathway. Accumulating evidence suggested that intracellular Ca2 + homeostasis is changed in cancer cells, and this alteration participates in tumor initiation, progression, and metastasis. Therefore, targeting calcium signaling is a novel approach for treating malignant tumors [44]. Bush et al. showed that PPARγ-regulated calcium signaling was correlated with CRC [45]. Taken together, CHRM2 serves an important role in the development of CRC by affecting the calcium signaling pathway. In addition, CHRM2 has been targeted by loxapine. Previous studies reported that loxapine was used in the treatment of breast cancer and mental disorder [46]. However, the therapeutic effect of loxapine on CRC remains unclear, and further experiments in preclinical models must be conducted to validate this result.
The results of the prognostic model showed that KAZALD1 and HPCAL4 may be considered as significant independent prognostic factors for CRC. KAZALD1 was over-expressed in cancer samples and was associated with shorter survival of patients with high-grade glioma [47]. Moreover, Alvi et al. indicated that KAZALD1 hypomethylation was related to poor survival of small bowel adenocarcinoma, and could be used as a prognostic biomarker [48]. The proteins encoded by HPCAL4 may regulate intracellular calcium homeostasis, which is connected to tumorigenesis and cancer progression [49]. Bavarva et al. observed that HPCAL4 was relevant to various cancer types, such as breast cancer, lung cancer, and CRC [50]. Preoperative staging of CRC is the basis of surgical strategy, and incomplete staging can lead to poor prognosis [51]. Meanwhile, tumor staging can provide valuable prognostic information and contribute to the design of treatment strategies [52]. In the present study, we observed that CDH8, SYNPO2, CCND3, and hsa_mir_26b may be independent prognostic factors of CRC staging. The proteins encoded by CDH8 were type II classical cadherin from the cadherin superfamily [53]. A study comparing the gene expression profiles of cancer and health samples showed that the abnormalities of genes such as CDH8 appeared only in cancer tissues [54]. Lee et al. revealed that CDH8 was a prognosis factor of non-small cell lung cancer (NSCLC) [55]. Furthermore, Lu et al. demonstrated that CDH8 may be used for predicting the survival of patients with stage Ⅰ NSCLC [56]. SYNPO2 is a member of the podin family and is a repressor of tumor cell invasion [57]. Kai et al. indicated that SYNPO2 could induce the formation of the complex stress fiber network in the cell body, and promote the migration of PC3 prostate cancer cells under serum stimulation [58]. Additionally, Kandimalla et al. revealed that SYNPO2 methylation was significantly correlated with the tumor stage of bladder cancer [59]. The proteins encoded by CCND3 belong to the highly conserved cyclin family, and these proteins are involved in the phosphorylation of tumor suppressor protein Rb [60]. Several studies reported that miR-592 could inhibit cell proliferation and tumor growth of CRC cells by suppressing CCND3 expression [61]. Edelman et al found that the expression level of CCND3 was significantly altered in patients with stage Ⅳ squamous cell lung cancer [62]. Another gene, hsa_mir_26b, was significantly related to the OS of CRC. Lin et al. found that miR-26b could induce pig granulosa cell apoptosis [63]. Meanwhile, Ma et al. observed that miR-26b was markedly connected with the invasiveness and metastasis of CRC cells, and it might serve as a prognostic factor for CRC [64]. However, the effect of hsa-mir-26b on the prognosis of tumor staging had not been reported. Thus, the relationship between hsa-mir-26b and cancer staging needs further study. Notably, the AUC values of the ROC curve for the COAD OS, COAD stage, READ OS, and READ stage prognostic models were 0.677, 0.917, 0.814, and 0.859, respectively, indicating that these genes had a good performance in survival prediction. Taken together, KAZALD1, HPCAL4, CDH8, SYNPO2, CCND3, and hsa_mir_26b were considered as potential prognostic markers for CRC. However, the specific action mechanism of these genes needs to be elucidated in further studies.
Overall, we systematically investigated the molecular mechanism underlying the development and prognosis of CRC. First, the DE-RNAs related to the prognosis of CRC were screened. Even though the action mechanisms of these genes are still not fully understood, our study has provided a theoretical basis for further research on CRC prognostic. Second, a number of drugs were predicted by drug-target analysis. Although these drugs have not been clinically proven, our work has provided guidance for clinical drug use. Some anticancer drugs may improve disease-free survival or relapse-free survival of CRC [65], but there is currently no effective drugs for improving OS. In the present study, OS was regarded as the endpoint for developing molecular markers; thus, these drugs may be beneficial for improving the OS of CRC. Third, the prognostic models may assist clinicians in providing individualized treatment in CRC patients.