High expression of TCN1 is a negative prognostic biomarker and can predict neoadjuvant chemosensitivity of colon cancer

Transcobalamin (TCN1) is a vitamin B12 (cobalamin)-binding protein that regulates cobalamin homeostasis. Recent studies and bioinformatic analyses have found that TCN1 is highly expressed in cancer tissues and is associated with tumour aggressiveness and poor prognosis. The present study aimed to detect TCN1 as a novel biomarker for prognosis and chemosensitivity of colon cancer. Next-generation sequencing showed that TCN1 was one of several upregulated mRNAs in colon cancer, which was verified by further bioinformatics analyses. Western blotting (n = 9) and quantitative real time polymerase chain reaction (qRT-PCR, n = 30) revealed that TCN1 was highly expressed in colon cancer tissues at both the protein and mRNA level. A total of 194 cases of colon cancer were examined by immunohistochemistry and revealed that TCN1 expression level was related to advanced stages (P < 0.005). Kaplan–Meier analysis verified that patients with lower TCN1 expression usually had longer overall survival (P = 0.008). In addition, TCN1 was highly expressed in pulmonary metastatic tumour tissues (n = 37, P = 0.025) and exhibited higher levels in right-sided colon cancer than in left-sided colon cancer (P = 0.029). TCN1 expression in specimens that had received neoadjuvant chemotherapy decreased compared with that in colonoscopy biopsy tissues (n = 42, P = 0.009). Further bioinformatics analyses verified that apoptosis pathways might have a role in high TCN1 expression. All the studies revealed that TCN1 expression in colon cancer was significantly associated with malignant biological behaviour. Therefore, TCN1 could be used as a novel biomarker for colon cancer aggressiveness and prognosis and might also be a potential biomarker for predicting neoadjuvant chemosensitivity.

www.nature.com/scientificreports/ metabolism, especially in haematopoiesis and the nervous system 4,5 . Researchers have found that high levels of cobalamin and TCN1 in human serum are associated with leukaemia, hepatocellular carcinoma, and phyllodes of breast tumours [6][7][8] . Overexpression of TCN1 in tumour tissues is associated with tumorigenesis and poor biological behaviour 9 . Chu et al. showed that TCN1 is an oncogene, ranking as one of the top six differentially expressed mRNAs in CRC 10 . Feodorova et al. revealed that NOTUM, TCN1, MACC1, YKL40, GPC3, AXIN2, and IL6 are significantly upregulated in CRC and have roles in tumour invasion and metastasis 11 . Our previous next-generation sequencing (NGS) results verified that TCN1 is one of the top five overexpressed mRNAs in CRC 12 . Taken together, these findings indicate that TCN1 is a CRC-related gene that requires further study. To our knowledge, the present study is the first to investigate the relationship between TCN1 expression and the clinical behaviour of colon cancer. NGS, western blotting, and immunohistochemistry (IHC) revealed that TCN1 was highly expressed in colon cancer tissues compared with adjacent normal mucosal tissues at both the protein and mRNA level. IHC staining of tissue microarray (TMA) slices also verified that TCN1 was highly expressed in colon cancer tissues (73.20%, 142/194) and pulmonary metastatic tumours (83.78%, 31/37). We also found that the expression of TCN1 was higher in right-sided colon cancer than in left-sided. High levels of TCN1 expression were associated with advanced pathological features and poor prognosis. Furthermore, TCN1 expression levels were decreased after neoadjuvant chemotherapy, which provided evidence that TCN1 might have a role in predicting chemosensitivity. Thus, we can conclude that high expression of TCN1 is significantly correlated with advanced clinicopathological features and poor prognosis of colon cancer. It may be novel biomarker for progression and prognosis and might be a potential biomarker to predict neoadjuvant chemosensitivity of colon cancer.
TCN1 was confirmed to be highly expressed in human colon cancer tissues. Western blot and qRT-PCR assays revealed that TCN1 was highly expressed in most of colon cancer tissues both at the protein ( Fig. 2A, B) and mRNA level (Fig. 2C), but was poorly expressed in adjacent normal colon mucosal tis-  Increased TCN1 expression was associated with clinicopathological features and poor prognosis of colon cancer. The clinicopathological characteristics of the 194 colon cancer patients are presented in Table 1. IHC staining was performed on TMA slides, and immunoreactivity of TCN1 was observed primarily in the cytoplasm (Fig. 3). TCN1 expression was significantly higher in colon cancer tissues than in adjacent normal tissues (73.20% vs. 5.67%, P < 0.01, Fig. 4a). Furthermore, positive TCN1 expression was significantly correlated with T, N, M, TNM stage and age ( Fig. 4b and Table1) and was higher in right-sided colon tumors than in leftsided tumors (Fig. 4e, Table 1). It was much higher in lung metastatic tissues (Fig. 4c) and the expression was consistent with that of the primer colon cancer tissues (Fig. 4d). Furthermore, Kaplan-Meier survival analysis demonstrated that higher TCN1 expression in patients represented significantly poorer OS (Fig. 4f). Univariate analysis showed that age, TNM staging and TCN1 expression were prognostic factors, while multivariate prognostic analysis showed that TCN1 expression level and M staging were independent prognostic factors ( Table 2).

TCN1 expression level decreases after chemotherapy.
After treatment with platinum-containing neoadjuvant chemotherapy, TCN1 expression in tumour tissues was decreased(P = 0.009, Fig. 4g). There was no significant difference in TCN1 expression status between colonoscopy biopsy tissues and specimens removed during surgery (P = 0.3765, Fig. 4h).

GO (Gene Ontology) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis results
of TCN1-related genes in cRc . GO and KEGG analysis was performed, and a total of 5,857 differentially expressed genes (2,683 positively related and 3,174 negatively related, false discovery rate < 0.01) correlated with TCN1 were detected. Spearman's test was conducted to analyse correlations between TCN1 and genes differentially expressed in CRC (red represents positively related genes and blue represents negatively related genes) (Fig. 5A). The top 50 genes that were positively and negatively correlated with TCN1 are shown in heat maps ( Fig. 5B, C). Over-representation analysis (ORA) was conducted in the top 50 genes positively correlated with TCN1. GO analysis showed that the identified TCN1-related gene products were mainly expressed in the cell membrane, Golgi apparatus, endoplasmic reticulum, and vesicles, and are involved in cytokine production, protein glycosylation, apoptosis, exocytosis, and vesicle transport ( Fig. 6A-C). KEGG pathway analysis showed that cancer-related signalling pathways were enriched, including immunity, apoptosis, platinum resistance, and metabolic, hypoxia-inducible factor-1A (HIF-1A) and vascular endothelial growth factor (VEGF) signalling pathways (Fig. 6D). Molecules (marked in red) connected with TCN1 are presented in an apoptotic pathway network in Fig. 7.

Discussion
TCN1 is a vitamin B12-binding protein that is present in the cytoplasm and performs functions with TCN2 and intrinsic factors in the processes of cobalamin transport, metabolism, and homeostasis 13,14 . Studies have shown aberrant serum levels of vitamin B12 and TCN1 in some solid tumours [15][16][17][18][19] . Upregulation of TCN1 is associated with tumourigenesis and progress 20-25 . Chu's meta-analysis of gene expression data based on microarrays demonstrated that TCN1 act as an oncogene and is overexpressed in CRC 10 . We hypothesized that TCN1 might have a role in colon cancer. The NGS results in CRC revealed that TCN1 ranked as the second most upregulated mRNA 12 . Bioinformatic analysis based on the COAD data set verified that TCN1 was indeed overexpressed in CRC. qRT-PCR and western blot assays also verified that TCN1 was highly expressed at both the mRNA and protein level in CRC tissues. All these results provided more evidence that TCN1 has a role in the tumourigenesis of CRC 10,11 , and suggested that TCN1 could be a potential novel biomarker. IHC in TMAs proved that TCN1 was highly expressed in colon cancer tissues and demonstrated that high expression levels were correlated with advanced pathological features. Pulmonary metastatic tumour tissue from  Patients with high expression of TCN1 showed shorter OS and poor prognosis. Previous studies demonstrated that high expression of TCN1 was related to poor prognosis of hypopharyngeal squamous cell and rectal cancer 20,21 . Our findings were concordant with these reports. Univariate analysis proved that TNM stage and TCN1 expression were prognostic factors, while multivariate analysis showed that TCN1 expression and M classification were independent prognostic factors. Further, we compared the expression levels in right-and left-sided colon cancer. Interestingly, TCN1 expression levels were much higher in right-sided than in left-sided colon cancer and the survival time was accordingly shorter, which was paralleled with the literature 27 . This provided further evidence that TCN1 expression was associated with colon cancer behaviour and prognosis.
We found that there was no significant difference in TCN1 expression between colonoscopy biopsy tissues and surgical specimens, which suggested that biopsy tissues obtained by colonoscopy might represent surgical specimens for the detection of TCN1 expression, and thus would make detection easier. Nevertheless, only a small cohort from a single centre was examined, prospective experiments with a larger sample size should performed to confirm the results. We also discovered that the expression level of TCN1 decreased after administration of platinum-containing chemotherapy. It could be inferred that the chemoradiotherapy will affect TCN1 expression through some mechanism. Zhang et al. screened differentially expressed genes from the Gene  21 . Researchers also proved that TCN1 expression levels were associated with neoadjuvant chemotherapy sensitivity in hypopharyngeal squamous cell cancer and chemoradiotherapy sensitivity in rectal cancer 20,21 , but the underlying mechanism remains unclear.
To reveal the potential downstream mechanism of TCN1 in CRC, genes related to TCN1 were selected to determine the way it functions. Because TCN1 might have a role as an oncogene 10 , genes positively correlated  www.nature.com/scientificreports/ with TCN1 were finally chosen for GO and KEGG analysis. As expected, the result proved that TCN1 is associated with many cancer-related genes and pathways, which provided evidence that TCN1 might affect tumourigenesis and progress by affecting transcription. Apoptotic pathways were enriched in the KEGG analysis and verified that lysosomes and apoptotic proteins of caspase 3 were located in a prominent location in the pathway network, suggesting that TCN1 may partially function through the apoptotic pathway. In the meantime, we found that TCN1 was significantly associated with immune response pathways, which provided a new avenue for our future study. Nevertheless, an experiment has not undertaken to verify this inference, which is the limitation of our study. We are focusing on this area and will explore the underlying mechanism in our future work. To our knowledge, this is the first report to discuss the relationship between expression of TCN1 and colon cancer behaviour. The research proves that TCN1 is associated with prognosis and presents clinic evidence for previous research 10,11 . It also provides a new area for future study, especially the underlying mechanism of tumour invasion and the mechanism of predicting chemosensitivity.
Taken together, we conclude that TCN1 expression is significantly overexpressed in colon cancer and is correlated with advanced pathological features. TCN1 may be a predictor of prognosis and a potential biomarker for predicting chemosensitivity. However, further in vivo and in vitro experiments should be performed to explore the underlying mechanism.   www.nature.com/scientificreports/ mucosal tissues (n = 39) were snap-frozen in liquid nitrogen and stored at − 80 °C for qRT-PCR assay and western blotting. None of the patients were treated by radiotherapy or chemotherapy before surgery.
RnA extraction, quantitative real-time pcR, and next-generation sequencing. Total RNA was extracted from tissues using the Trizol (Life Technologies, Massachusetts, USA) reagent 12,29 . cDNA was synthesized from the total RNA using the GoScript Reverse Transcription System (Promega, Wisconsin, USA) according to the manufacturer's instruction. A SYBR Green PCR kit (Promega, Wisconsin, USA) was used in the amplification process with a 7,500 Real-time PCR system (Applied Biosystems, Massachusetts, USA). The primers (HQP017978, Eockville, MD, USA) were purchased from GeneCopoeia. GAPDH was used as an internal control. The qRT-PCR results were analyzed with the 2 −ΔΔCt method. All the reactions were conducted in triplicate and the values are shown as means ± SD. mRNA data were selected from our earlier NGS results 11 , which had been deposited in GEO database (GSE104836, https ://www.ncbi.nlm.nih.gov/ geo/query/acc. cgi?acc = GSE104836). The differential expression of the genes was assessed by limma package 30 (https ://bioco nduct or.org/ packages/release/bioc/html/limma.html) in R Software using RNA-seq read counts. |log2Fold-Change|> 2 and adjusted P < 0.05 as cut-off criteria for screening the differentially expressed genes (DEGs). A heatmap and volcano plot were drawn using pheatmap and ggplot2 packages in R software to indicate the DEGs.  Haematoxylin-eosin (HE) staining and immunohistochemistry (IHC). HE staining was performed for tumour diagnosis, and TCN1 expression was detected using IHC in tumour and adjacent normal tissues. Continuous sections with a thickness of 4 μm were performed for HE and IHC staining on Envision. Anti-TCN1 monoclonal antibody was purchased from AB-clonal (1:60, a6614, ABclonal, Wuhan, China). All slides were interpreted by two pathologists independently who were blinded to the clinical information. The IHC staining score included the proportion of positively stained tumour cells and the staining intensity. The www.nature.com/scientificreports/ proportion of positively stained tumour cells was scored as follows: 0 (no tumour cells stained), 1 (< 25% tumour cells stained), 2 (25-50% tumour cells stained), 3 (50-75% tumour cells stained), and 4 (75-100% tumour cells stained). The grading of staining intensity was evaluated by the following criteria: 3 (brown, strong staining), 2 (yellow brown, moderate staining), 1(light yellow, weak staining), and 0 (no staining). The final total staining score was calculated by multiplying the proportion of stained tumour cells and the staining intensity score. TCN1 expression was scored and a score of ≤ 3 indicated negative TCN1 expression, while a score of > 3 was considered as positive TCN1 expression. A score of ≤ 6 was designated as low expression and a score of > 6 as high expression. The pathological diagnosis was made in accordance with the histological classification of tumours developed by the World Health Organization.
GO and KEGG enrichment analysis of TCN1-related genes in cRc based on the Linkedomics database. The LinkedOmics database (https ://linke domic s.org), which has been used to analyse TCGA cancer-associated datasets 27 , was selected to identify the differentially expressed genes related to TCN1. The database included 379 CRC cohorts. Results from LinkedOmics were assigned and ranked. GO functional enrichment analysis and KEGG pathway analysis were performed by gene set enrichment analysis (GSEA). Spearman's test was conduct to perform statistical analyses. All the results were graphically showed in heat maps and volcano plots. Linkfinder-related modules were selected to show the results of differentially expressed genes and the network analyses.
Statistical analysis. Statistical analysis was performed using the software package SPSS 21.0 (Chicago, IL, USA) and figures were constructed with Prism5.0 (GraphPad, Lnc., La Jolla, CA, USA) and R (version: 3.6.1). Correlations between TCN1 expression and clinicopathological features were determined using chi-squared tests or Fisher's exact tests, as appropriate. The endpoints were defined as patient death or the final follow-up date. Overall survival (OS) was defined as the time between surgery time and death or the last follow-up visit. Kaplan-Meier survival analysis was conducted to plot survival curves and log-rank tests were used to evaluate the prognostic differences between subgroups. Cox multivariate regression analysis was conducted after prognostic significance at the univariate level was determined. A P value of < 0.05 was considered statistically significant.
ethics approval. We confirm that all the methods had been carried out in accordance with the relevant guidelines and regulations of the Declaration of Helsinki.