Colorectal mucinous adenocarcinoma indicates a meaningful subtype: A whole genome sequencing study.
Yunhua Xu,Xiguang Chen,Yuqiao Chen,Xiaofeng Wu,Qing Fang,Xiangwen Tan,Shuxiang Li,Qiulin Huang,Xuyu Zu,Kai Fu,Shuai Xiao
DOI: https://doi.org/10.1002/ctm2.1246
IF: 8.554
2023-01-01
Clinical and Translational Medicine
Abstract:Dear Editor, Colorectal cancer (CRC) is a common lethal gastrointestinal tumour. Mucinous adenocarcinoma (MAC) is a special histological subtype of CRC, which characterized by abundant extracellular mucin.1 MAC has distinct characteristics compared with the commonest subtype, non-specific adenocarcinoma (AC), including clinicopathologic factors, molecular features, therapy response, and prognosis.2-5 However, since the molecular mechanisms differences between the MAC and AC are still unclear, all of the current treatment guidelines rarely notice their distinction.6, 7 To illustrate these issues, we firstly performed RNA-sequencing on 40 samples comprising 15 MACs, 15 ACs, and 10 normal colorectal tissues, which were collected from our central (hereinafter referred as USC set). The baseline feature of CRC cases is appeared in Table S1. Through the clinicopathological features analysis, we found MAC had a more proximal colon location (p < .05) and larger tumour size (p < .05). In addition, MAC showed a higher pT stage and more frequent dMMR tendency (Table 1). Then, the TCGA-COAD transcriptome and clinical data (referred as TCGA set) were also downloaded and analyzed. Through analyzing TCGA clinical data, we got a similar result: no significant difference in other items except dMMR (Table S2). To elucidate the genome characteristic between MAC and AC, we analyzed the TCGA-COAD mutation data. Results showed that the mutation atlas of them was notably different. In the AC group, APC (74.6%) had the highest mutation frequency followed by TP53 (59.0%), TTN (49.4%) (Figure 1A). While in the MAC group, TTN (72.4%), APC (65.5%), and KRAS (46.6%) are the most mutated genes (Figure 1B). Additionally, we observed a high mutation frequency of BRAF, whereas TP53 was rare in MAC (Figure 1A, B). Additionally, comparing the common 10 mutant genes, we found only the mutation rate of TTN, RYR2, and OBSCN were distinctly different (Figure 1C). In addition, we observed that the integral genome alteration frequency of MAC is higher than AC (Figure 1D). Finally, we compared the fraction genome alteration of the above three genes. Although the mutation rate of these genes was higher in MAC, the fraction genome alteration was lower in MAC, and each gene had a unique mutational feature (Figure 1E). To study the transcriptomics features of MAC and AC, we performed differentially expressed genes (DEGs) analysis in TCGA and USC sets, respectively. In the TCGA set, we found thousands of DEGs between tumour and normal tissue (Figure 2A,B), and hundreds of DEGs between MAC and AC (Figure 2C). In the USC set, we got similar results (Figure 2D–F). The distribution of DEGs was shown as the ternary plot. The variation was apparent among transcriptional properties of different subtypes (Figure 2G,H). Finally, differences and similarities of DEGs for the TCGA and USC datasets are summarized in Figure 2I. Furthermore, we synthesize intersection DEGs of MAC versus AC in USC and TCGA set and performed functional enrichment analysis to uncover the potential gene functions. Through GO, KEGG, cancer hallmarks, and Reactome analysis, we found the gene functions of these DEGs were correlated with symporter activity, and metabolism-related molecules and pathways (SM1, Figure S1). In 2011, Weinberg et al. summarized 10 cancer hallmarks, which become the cornerstone of tumour characteristic study.8 Thus, we analyzed the 10 hallmarks differences between AC and MAC. Results showed that evading growth suppressors and inducing angiogenesis appeared a marked weak activity. Nevertheless, reprogramming energy metabolism and tumour-promoting inflammation were highly expressed. Unfortunately, we didn't observe a hallmark that was significantly different between MAC and AC (SM1, Figure S2A,C). In addition, we found some genes played an important role in multiple hallmarks, while others played the exclusive roles in single hallmark (SM1, Figure S2B,D). According to Sadanandam et al. reported that CRCs could be classified into five cell phenotypes, which had distinct therapy response.9 We analyzed the cell subtype differences of MAC and AC both in TCGA and USC sets. Result showed that the samples of two datasets were well classified into five cell subtypes (Figure S3A,C), and the proportion of each subtype was different (Figure S3B,D). The Goblet-like group accounted for the largest proportion in the MAC, followed by Stem-like, and Ta was the least. On the contrary, the predominant cell subtype of AC was the Ta group, and the other four subtypes are relatively evenly distributed (Figure S3). The consensus molecular subtype (CMS) is a new and important classification system for CRC.10 We also explored the association between CMS and MAC, and results showed that there was good consistency between subtype prediction and template features of TCGA and USC set (Figure 3A, D). The signal feature of each subtype was also studied. Results showed that the CMS1 group has obvious MSI characteristics, and the CMS2 group has distinct MSS, MYC, and cell cycle features, and differentiation signal was visible in CMS3, while TGF-Beta and EMT signal was prominent in CMS4 (Figure 3B, E). Our results showed that AC group was enriched in CMS2 and CMS4 subtypes, and MAC group was enriched in CMS3 and CMS4 subtypes, especially in CMS3 (Figure 3C, F). Previous studies indicated MAC was correlated with poor chemotherapy response and prognosis. To better understand the underlying molecular mechanisms, we analyzed GEO dataset of CRC drug resistance (GSE83129). GSEA was carried out to acquire the enriched pathway by our sequencing results and drug resistance dataset. Six shared GOBP pathways were enriched in MAC and chemotherapy non-responder group (Figure 4A,B). The common genes interaction among these six GOBP pathways are shown in Figure 4C. We further found six intersection genes in GOBP①/GOBP③(Figure 4D), and 15 intersection genes in GOBP④/GOBP⑤ (Figure 4E). These 21 evolved drug-resistant gene clusters might confer chemotherapy resistance to MAC. In conclusion, this whole genome sequencing study preliminarily revealed the molecular and functional characteristics of MAC, as well as potential clinical value, which indicated that MAC tend to metabolic and mesenchymal phenotypes, contributing to worse prognosis and chemotherapy resistance. This work was supported by the Clinical Medical Technology Innovation Project of Hunan Province (grant number: 2020SK51815), the Scientific Research Fund Project of Hunan Provincial Health Commission (grant numbers: 20201919 and 202104010105), and Natural Science Foundation of Hunan Province (grant number: 2022JJ30538). The authors declare no conflicts of interest. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.