The incidence and mortality of CRC still go up in many countries[5]. Identification of genes and pathways related to CRC would be helpful for the diagnosis and treatment of the disease [1]. Ga Eun Nam, et al.’ s study [6] suggested that abdominal obesity may be a risk factor for CRC in this East Asian population. Also, a longitudinal analysis study indicated non-alcoholic fatty liver disease (NAFLD) with obesity posed a risk factor for colorectal cancer in apparently healthy Japanese individuals [7]. In addition, Abbinaya Elangovan, et al. conducted a large retrospective, cross-sectional database study which showed the association between obesity and colorectal cancer [8]. Obesity is regarded as one of the key environmental risk factors for the pathogenesis of CRC.
This is the first article to explore the relationship between obesity and colorectal cancer at the gene level of a large sample. We speculated that obesity may be an independent risk factor of CRC. In order to explore the characteristic genes of CRC in obese group and to identified potential molecular markers for early diagnosis, targeted therapy and recognition of prognosis of special CRC population, these characteristic genes were screened from TCGA database. We found that the most significant downregulated genes of non-obese group were mostly similar with those of total group, such as VSTM2A, BMP3, SCARA5, CA1, PKIB and SLC4A4. As for VSTM2A, consistent with our findings, Dong, Y., et al. identified that it was one of the top downregulated secreted protein in CRC and maybe a novel prognostic biomarker for CRC patients, for VSTM2A might suppress colorectal tumorigenesis by directly binding to Wnt signaling co-receptor LDL receptor related protein 6 [9]. In addition, Secco, B., et al. [10] reported VSTM2A as a factor modulating adipogenic commitment, which can maintain and amplify the adipogenic capacity of adipose precursor cells. Cancer tissue in non-obese people has lower VSTM2A expression than normal tissue beside cancer, which seems to contradict energy theory.
In our study, MS4A12, TMIGD1, CA2, GBA3, SLC51B are only sharply downregulated in obese group and SST, PYY, GNG12, CCL13, MCHR2, CCL28, ADCY9, SSTR1, CXCL12, ADRA2A were identified as top ten hub genes in the network in obese group. Among above genes, MS4A12 is related to proliferation and motility of colon cancer cells, participating in cell membrane composition, cell differentiation, proliferation, and cell cycle regulation [11]. Current studies supported that MS4A12 is a colon-specific gene regulated by intestinal differentiation transcription factor CDX2 [12, 13] and in adenomatous polyp, MS4A12 transcription is downregulated while CDX2 gene and protein expression upregulated [13, 14]. Survival analysis of primary colorectal cancer patients showed that patients with low expression of MS4A12 had a worse survival rate than those with high expression of MS4A12, suggesting that MS4A12 was involved in the occurrence and development of primary colorectal cancer and could inhibit the progression of cancer, which may be a potential target for diagnosis and treatment [15]. What’s more, MS4A12 can predict prognosis in early stage colon cancer patient and mainly relate to malignancy of non-metastasis tumor while the prognosis value of metastasis colon tumor is low [13]. However, few evidences confirm the relationship between this gene and obesity.
TMIGD1 is also a downregulated gene in obese group. It not only is a novel adhesion molecule, which is highly conserved in humans and other species, but also a novel tumor suppressor highly expressed in normal coloretal epithelial cells and downregulated in CRC [16]. Kyle Oliver, et al.’ s study showed that TMIGD1 inhibits cell migration and metastasis in colon cancer as a tumor suppressor by arresting cell cycle at G2/M, and down-regulation of TMIGD1 correlates with poor survival [17]. However, few studies have been able to draw on any systematic research into the association between this gene and CRC with obesity.
SLC51B, combined with solute transporter-alpha (SLC51A) is the encode gene of the organic solute transporters alpha and beta (OSTa-OSTb), which is located at the basolateral membrane, involving in the metabolism of bile acids [18]. Some studies indicated a defect about intestinal bile acid and conjugated steroid absorption in SLC51A-deficient mice, causing a decrease in the levels of bile acid, serum triglyceride, cholesterol, and glucose [19, 20]. Wang DF, et al. confirmed that expression of SLC51B which plays a role in bile acid transport, was decreased in the ulcerative colitis. Diminished uptake of bile acids into ileocytes is most likely responsible for the decreased expression of SlC51B in the ileum. In our study, the down-regulation of this gene in the obese group may be associated with metabolic abnormalities in obesity. Notably, many studies have suggested that ulcerative colitis and CRC have the same metabolic pathway, which seems to support abnormal expression of this gene in CRC.
Carbonic anhydrases, a kind of critical enzymes in regulating pH homeostasis in cells, catalyze the interconversion between carbon dioxide and carbonic acid [21, 22]. Norihiro Nakada, et al.’s study showed that CA2, an acytoplasmic enzyme widely expressed in normal organs including colonic mucosa, was downregulated in ulcerative colitis-associated colorectal cancer. However, the mechanism of CA2 downregulation in CRC is not clear [23].
GBA3 is an enzyme belonging to the glycosid e hydrolase family 1 that is able to hydrolyse a wide variety of substrates with a b-D-glucose moiety linked to a hydrophobic group, mostly present in the kidney, liver, spleen, intestine and lymphocytes of mammals [24]. Very little is known about the role of this enzyme in cancer. In our study, GBA3 was downregulated in CRC tissue in obese group. Gene expression profiling of intestinal metaplasic lesions from patients, which are precursors of gastric cancers, identified GBA3 among the top 25 significantly upregulated genes [25]. GBA3 might play an important role in the detoxifcation and/or biotransformation of dietary xenobiotic plant β-glycosides in roots and tubers, which are the main food in African populations but not in Asian and European populations. And it can be observed an overlap in Eurasia between the highest frequency of GBA3 loss and a main dietary source of meat or fish, while in Africa the lowest frequency of GBA3 inactive alleles meets the general trend in diets higher in glycoside-rich foods and poorer in meat or fish [26].
In the enrichment analysis of the three groups, DEGs in MF term are mainly focused on ion channel activity, and on inflammation in BP term, but in the obesity group, DEGs also focused on metabolism. In terms of CC, cell membrane covered the most in the obese group, while the non-obese group and total group preferred cell connection structure. The KGGG enrichment analysis of the three groups mainly focused on the common energy metabolism signal pathway.
Though KM survival analysis, PDCD11 may be potential molecular markers for non-obese patients with CRC. PDCD11 is a NF-κB-binding protein that colocalizes with U3 RNA in the nucleolus and is required for rRNA maturation and generation of 18S rRNA. PDCD11 is necessary for Fas ligand (FasL) expression, and PDCD11 overexpression is known to induce transcription of FasL, leading to the induction of apoptosis through Fas/FasL/caspase death pathway [27–30]. Current studies on PDCD11 mainly focus on the nervous system, such as transient ischemic attack and schizophrenia [31, 32]. Importantly, our study argued that PDCD11 upregulated in CRC tissue, belonged to low expression category in non-obese group with longer survival time (Supplement txt. 2; Fig. 6a). There might be other pathways affecting CRC tumorogenesis, which is not clear now.
It is found that SRSF5, SRSF11 and HNRNPA1 had proven to be associated with the prognosis of CRC in total group of our study. SRSF5, previously called SRp40, is a member of the SR protein family that has early been identified as a splicing regulator [33, 34]. Currently, research about SRSF5 is mainly focused on lung cancer, and it is hyperacetylated and upregulated in human lung cancer. Little is known about the relationship between SRSF5 and colorectal cancer. In our analysis, SRSF5 was downregulated in colorectal tissue and belonged to low expression category in total group (Supplement txt. 2; Fig. 6c). The survival time of low expression SRSF5 group was less than that of the high expression group in total group (P = 0.025). Yuhan Chen, et al.’ s study [35] indicated that upon glucose intake, the splicing factor SRSF5 is specifically induced and promotes the alternative splicing of cell cycle and apoptosis regulator 1 to produce special proteins, which promote tumor growth by enhancing glucose consumption and acetyl-CoA production. SRSF5 seems to be a tumor suppressor gene in colorectal cancer and responds to high glucose to promote cancer development.
HNRNPA1 is also a well-known splicing regulator with effects antagonistic to SR proteins [36]. Upregulated expression and aberrant cytoplasmic localization of HNRNPA1, as determined by immunohistochemical staining, were noted in CRC [37]. HNRNPA1 not only contributed to the promotion of cell growth through the regulation of energy metabolism to use glycolysis efficiently, but upregulated the promoter activity of TRA2B which was considered as an oncogene [38, 39]. Sun Y, et.al argued that the phosphorylation of Ser6 of hnRNPA1 was a predictor of poor prognosis for patients with CRC [40]. HNRNPA1 was upregulated in CRC tissue in total group of our study, too. However, it belonged to low expression category with shorter survival time (Supplement txt. 2; Fig. 6d). Interestingly, an animal study showed that Hepatic HNRNPA1 overexpression activated the CaM/Akt pathway and repressed the mTOR/SREBP-1C pathway to ameliorate hyperglycemia and steatosis in obese mice [41]. Therefore, we surmise that this gene might upregulated in obese people with CRC and lead to a poor prognosis of CRC.
At present, there are few researches about SRSF11 on CRC. Ji Hoon Lee, et.al suggested that SRSF11 is a novel the telomerase RNA component binding protein that localizes to nuclear speckles with subnuclear structures that are enriched in premessenger RNA splicing factors. It acts as a nuclear speckle-targeting factor that is essential for telomerase association with telomeres and provides a potential target for modulating telomerase activity in cancer [42]. We found that SRSF11 still belonged to low expression category in total group with shorter survival time (Supplement txt. 2; Fig. 6b). It seemed to confirm that SRSF11 is a tumor suppressor gene in CRC.
There are some limitations in this study. Firstly, patients were categorized by BMI. However, some studies suggested that measures of abdominal adiposity, such as waist circumference (WC) and waist-to-hip ratio (WHR), might be more strongly associated with colorectal neoplasia risk than BMI [6, 43]. Secondly, our study was performed only by pure bioinformatics analysis. Finally, our article only found the relevant colorectal cancer genes in the non-obese population, but not in obese population. Therefore, further experiments are needed to validate the results based on tumor samples and clinical data.