Effect of Tumor Location on Clinicopathological and Molecular Markers in Colorectal Cancer in Eastern China Patients: An Analysis of 2,356 Cases

Colorectal cancer (CRC) has become a major health concern in China due to its increasing incidence and mortality. This study aimed to clarify the relationship between tumor locations and the clinicopathological molecular marker features in eastern China CRC patients. We continuously collected data on 2,356 CRC patients who underwent surgical resection from January 2017 to April 2019. Right-sided colorectal cancer (RCC), was located from the cecum to the transverse colon and left-side colorectal cancer (LCRC) was located from the splenic flexure to the rectum. The clinicopathological indices (including age, sex, pTNM stage, mucinous production, and distant metastasis) and frequency of molecular markers such as KRAS, NRAS, BRAF, and microsatellite instability (MSI) were statistically analyzed between the RCC and LCRC groups. The associations between clinicopathological characters and molecular markers were also investigated. LCRC and RCC proportions in eastern China CRC patients were 81.75% and 18.25%, respectively. RCC (vs. LCRC) was more frequently observed with higher frequencies of MSI-high (MSI-H) and BRAF mutations in female and younger patients, and was closely associated with metastasis, poor differentiation, and mucinous tumors. Tumor location also showed significant differences in bowel wall infiltration degree and pTNM stage. Mutation rates of KRAS, NRAS, MSI, and BRAF were 40.15%, 3.85%, 6.31%, and 2.30%, respectively. Patients with a KRAS mutation tended to be female, had mucinous, perineural invasive, and polypoid tumor. Those with NRAS mutation tended to develop well-differentiated ulcerative tumors. The BRAF mutation was more relevant with lymph node involvement, deeper infiltration of the bowel wall, mucinous, poorly-differentiated tumor with thrombus, and perineural invasion. Furthermore, MSI-H was more commonly found in younger patients with deeper bowel wall infiltration and a poorly-differentiated polypoid tumor, whereas MSS patients tended to develop lymph node involvement, and a mucinous and perineural invasive tumor. In our study, we found that LCRC and RCC showed different features on the clinicopathological and molecular markers in eastern China CRC patients. Since our data differ from those of Western countries and other regions in China, further studies are required to clarify the regional differences of the clinicopathological and molecular markers in CRC patients.


INTRODUCTION
Colorectal cancer (CRC) is the third most common cancer, with the fourth most common cancer-related mortality worldwide . In China, the incidence, mortality, and burden of CRC are all increasing due to the transition to a westernized lifestyle with obesity and physical inactivity (Center et al., 2009;Gu et al., 2018). Among Chinese patients, CRC shows the fifth highest cancer incidence in men, the fourth in women; and is fifth in cancer-related deaths in men (8.0%) and third in women (9.8%) (Feng et al., 2019).
Both genetic and environmental factors are involved in the tumorigenesis of CRC. For the past decades, studies have demonstrated that genetic molecular markers such as microsatellite instability (MSI), 18q loss of heterozygosity, and CpG island methylator phenotype (CIMP), RAS, and BRAF, among others, are closely associated with the tumorigenesis and prognosis of CRC (Sanz-Garcia et al., 2017;Saeed et al., 2019).
MSI is a genetic hypermutability condition resulting from deficient DNA mismatch repair. It is involved in various types of cancers, including colon cancer, ovarian cancer, skin cancer, and gastric cancer, among others. MSI is most frequently associated with the development of CRC and is found in about 15% of CRC tumors. RAS belongs to the proto-oncogene family that encodes three small GTPase proteins including KRAS, NRAS, and HRAS. RAS gene mutation presents in about 30% of all human cancers, and the mutation proportions of KRAS, NRAS, and HRAS are around 85%, 11-15%, and 1%, respectively (Cox et al., 2014;Chang et al., 2016). KRAS functions downstream in the epidermal growth factor receptor (EGFR) signaling pathway and is involved in cell proliferation, mutation, and cell death. KRAS mutation plays an important role in carcinogenesis, about 30-50% CRC is known to have KRAS gene mutation and the mutation positions are most frequently in codons 12 and 13, in exon 1 (Wagner et al., 2019). NRAS is closely related with KRAS; and its mutation in CRC, which is mainly located at codons 12, 13, or 61 is approximately 1-6% (Vaughn et al., 2011). BRAF is a serine kinase, downstream of KRAS, and in the MAPK signaling pathway. The most frequent mutation point of BRAF is V600E. The incidence of the BRAF V600E mutation is estimated to be about 8-10% in CRC patients (Bahrami et al., 2018;Myte et al., 2019).
CRC is anatomically divided into the right (RCC) and left CRC (LCRC) according to the location of the tumor by the border of the splenic flexure. Past studies have observed the relationship of tumor location with the development, clinicopathological characteristics, treatment, and prognosis of CRC. However, the effect of tumor location on the molecular markers of CRC still remains unclear. Recently, several studies have reported that the clinicopathological and molecular biomarkers of CRC were divergent in different regions (Giraldez et al., 2012;Perea et al., 2015). It is of great value to clarify the correlation between tumor location and genetic markers for patients in different regions. Today, most reports that study differences in molecular markers in RCC and LCRC hail from Western countries, with very few studies from eastern China.
In this study we collected data on 2,356 CRC cases from the affiliated hospital of Qingdao University, a medical center that mainly provides medical care to eastern China patients. We summarized the clinical features and molecular characteristics using RCC and LCRC. Our study provides valuable guidance and reference for the diagnosis and treatment of CRC patients in eastern China.

Patients
We continuously collected data on 2,566 CRC patients who underwent surgical resection at the Affiliated Hospital of Qingdao University from January 2017 to April 2019. RCC was defined as CRC from the cecum to the transverse colon, while LCRC was located from the splenic flexure to the rectum. As summarized in Figure 1, patients who were treated with radiotherapy or chemotherapy before surgery and who lacked accurate tumor location were excluded. Finally, 2,356 cases were selected for this study analysis. For patients who developed two or more colorectal tumors, the more advanced one was selected for this study. Informed consent was obtained from all patients, and the Ethics Committee of Qingdao University approved this study.

Genomic DNA Extraction
Fresh CRC and corresponding normal tissues were fixed by 4% paraformaldehyde at 4°C overnight, paraffin embedded, and sectioned at 5 mm for hematoxylin and eosin (HE) staining for later use. The HE section was observed under the microscope; areas that were rich in tumor cells (the proportion of tumor cells was over 20%) were selected, while non-tumor areas and necrotic areas were avoided as much as possible. Tumor tissues were scraped in a clean Eppendorf tube according to the HE section, and the genomic DNA in paraffin tissues was extracted using a Tiangen paraffin embedded tissue DNA extraction kit (Tiangen Biotech, Beijing, China), according to the manufacturer's instructions.

Microsatellite Instability
Microsatellite status was defined by six microsatellite markers (NR21, NR24, NR27, BAT25, BAT26, and MONO-27) and detected by an MSI detection kit (Microread, Beijing, China). Polymerase chain reaction (PCR) amplification was performed as follows: a total volume of 10 µl reaction mixture contained 20 ng of genomic DNA, 1 X PCR buffer, 1 X MSI Primer Mix, and Taq DNA Polymerase I. The running protocol was set up as follows: 5 min at 95°C once; 30 s at 94°C, 1 min at 60°C, and 1 min at 70°C for 30 cycles; 30 min at 60°C and forever at 15°C. The PCR production was diluted 1:6 and mixed with 9 µl ROX500 and HI-DI; after denaturation for 3 min at 96°C, the reaction complex was analyzed by DNA fragmentation assays (Applied Biosystems 3500DX, Massachusetts, USA). Allelic sizes were evaluated by GeneMapper software ver 4.1 (Thermo Fisher, Massachusetts, USA). MSI status was classified as Microsatellite stable (MSS), MSI-low (MSI-L, 1 marker unstable) and MSI-high (MSI-H, over 2 markers unstable).

Mismatch Repair (MMR) Protein Immunohistochemistry
Immunohistochemistry for mismatch repair (MMR) proteins was performed as previously described . Briefly, tissues were paraffin embedded and sectioned at 3 mm. Immunohistochemistry staining for MLH1, PMS2, MLH2, and MSH6 (Ventana Medical Systems Inc, Tucson, AZ, USA) was accomplished on a BenchMark XT automated staining system (Ventana Medical Systems, Inc., Tucson, AZ, USA). Two pathologists evaluated the staining results and the standards for diagnosis were as follows: expressions of all proteins were considered proficient MMR, loss of expression of one or more of the MLH1, PMS2, MLH2, and MSH6 proteins indicating DNA mismatch repair.

KRAS, NRAS, and BRAF
Patients were tested for KRAS (codons 12 and 13), NRAS (codons 12, 13, and 61), and BRAF (V600E) to detect gene mutation; detailed detection variants are shown in Table 1. Genomic DNA was extracted from paraffin embedded tissues using a Tiangen kit (Tiangen Biotech, Beijing, China). RAS and BRAF mutations were detected by the Human RAS and BRAF mutation detection kit with PCR fluorescence probe, according to the manufacturer's instructions (AmoyDx, Xiamen, China). Briefly, for each gene analysis, 25 µl total volume of complex mixture containing 0.3 mm primers and Taqman probes, 200 mm dNTPs, 200 mm Taq polymerase, and 90 ng of DNA. PCR amplification were set up using ABI 7500 as follows: 42°C, 5 min; 94°C, 3 min (94°C, 15 s; 60°C, 60 s); 40 cycles. The running data were analyzed by 7500 software ver 2.3 (Applied Biosystems, Massachusetts, USA).

Statistical Analysis
All data were analyzed by SPSS 19.0.0 statistical analysis software (IBM Corp., Armonk, NY, USA). The relationship between the two groups (RCC and LCRC) was evaluated using a standard chisquare test. Data that were not qualified for the chi-square test, were merged into the groups to reach the standard. p < 0.05 was defined as statistically significant.

Clinicopathological Characteristics by RCC and LCRC
The summary of the basic clinicopathological indices with respect to the tumor location is shown in RCC was more frequently associated with younger female patients (p = 0.000), a higher risk of metastasis (p = 0.003), poorlydifferentiated carcinoma (p = 0.000), and mucin production (p=0.000). Tumor location was also involved in the infiltration degree of the bowel wall and pTNM stage. In contrast, LCRC and RCC showed no significant differences in lymph node involvement, polypoid gross type, tumor, and perineural invasion.

Molecular Differences in RCC and LCRC
We checked the MSI, KRAS, NRAS, and BRAF mutation status in this study. For MSI detection, among the 2,329 cases, 757 cases were analyzed by immunohistochemistry staining for MMR proteins and 1,572 cases were analyzed by PCR; both DNA mismatch repair and MSI-H results were considered. Since MSI-L CRC showed no difference with the MSS tumor (Pawlik et al., 2004), we merged the MSI-L (n = 3) and MSS (n = 2179) tumor as MSS/L, which differed from MSI-H (n = 147). Among the 2,356 patients with CRC, the detection rates of molecular markers were 80.56% for KRAS and NRAS, 42.57% for BRAF, and 98.85% for MSI. The mutation rates for these molecular markers were 40.15%, 3.85%, 2.30%, and 6.31% for KRAS, NRAS, BRAF, and MSI, respectively. RCC was significantly associated with higher incidence of MSI-H (p = 0.000) and BRAF mutation (p = 0.001) compared with LCRC. However, there was no measurable difference between RCC and LCRC on the KRAS and NRAS mutation. Furthermore, we checked the correlation between the genetic markers and found that the KRAS mutation was accompanied with a lower mutation rate of NRAS, while the mutation status of KRAS and BRAF was incompatible, no association was found between the KRAS mutations and the BRAF mutation.

BRAF Mutation was Specifically Low in Eastern China CRC Patients
Among the 1,003 CRC patients with BRAF test, 23 patients with a positive BRAF mutation was found with a mutation rate of 2.29% -much lower than that in the published data of 8-10% (Bahrami et al., 2018;Myte et al., 2019). This result indicated that BRAF mutation rate was specifically low in eastern China patients. Also, with respect to the tumor location, the mutation rates for LCRC and RCC were 1.49% and 6.36%, respectively. We analyzed the mutation rate of BRAF in each pTNM stage and found that it was mostly detected in stage III (n = 18/23, 78.26%).

DISCUSSION
RCC and LCRC were first proposed as two distinct tumors by Bufill (1990), from the perspective of molecular genetics. The clinical performance, prognosis, and sensitivity to targeted therapy differs significantly depending on the tumor location. For the past decade, the role of molecular markers in the diagnosis and prognosis of cancer has been increasingly prominent. Since the indices of cancer vary in different regions and people, clarifying the effect of colorectal tumor location on clinicopathological features and molecular markers in regional areas is of great value for clinicians. In this study, for the eastern China patients, the frequency of RCC was 18.25%, lower than in previous reports from other regions in China such as Shanghai (24.4% and 25.3%), Shantou (36.9%), Tianjin (50.6%), and Guangdong (17.5%), as well as Japan (26.3% and 29.3%) (Xu et al., 2010;Watanabe et al., 2012;Liu et al., 2017;Peng et al., 2017;Qin et al., 2017;Natsume et al., 2018;Guo et al., 2019), whereas the RCC frequency in the United States was as high as 42% (Siegel et al., 2014). This finding indicates that the distribution of RCC in eastern China patients might be different compared to those of other regions in China and Western countries.
For the RCC and LCRC clinicopathological characteristics, our findings were similar to those in published data. RCC was more frequently associated with female patients, metastasis, mucinous, poorly-differentiated carcinoma, and a higher correspondence with the BRAF mutation and MSI-H status (Wangefjord et al., 2013;Li et al., 2015;Fu et al., 2019). Furthermore, our results showed that RCC was more common in younger patients who were under 50 years old, providing a diagnostic reference for clinicians.
For the molecular markers in CRC, the mutation rates of MSI-H, KRAS, and NRAS were reported to be 6-15% (Horvat and Stabuc, 2011;Gelsomino et al., 2016;Chang et al., 2017;  Samstein and Chan, 2017), 38.5-40% (Natsume et al., 2018), and 1-6% respectively (Downward, 2003). In our study, the frequencies of the MSI-H (6.31%), KRAS (40.15%), and NRAS (3.85%) mutation were in accordance with those in the Western countries and other regions in China, whereas the BRAF (2.29%) mutation was specifically lower. The BRAF mutation rate was reported to be around 2.5%-6.15% in different regions of China (Peng et al., 2017;Wang et al., 2018), and 3.7-11.3% in other Asian countries (Lee et al., 2015;Fujiyoshi et al., 2017;Natsume et al., 2018), for the eastern China patients, this incidence was 2.23%, much lower than those of Western countries, reported as 17-19.4% (Hutchins et al., 2011;Loupakis et al., 2015). We analyzed the pTNM stage and found that the BRAF mutation was mostly detected at stage III which was in line with previous data (Sayagues et al., 2018). The low mutation rate in eastern China patients might be due to the regional difference and varying genetic predispositions. Further studies need to be completed to investigate the effect of environmental and genetic factors on the molecular markers. Furthermore, several reports showed that the frequency of the BRAF mutation in RCC was around 4-10% in Asia, and in our data at around 6.36%, in agreement with the published data. RCC more frequently corresponded with a higher mutation rate of MSI-H and BRAF but had no association with the KRAS and NRAS mutation. Also, as Table 4 shows, we checked the mutation status of RAS and BRAF and found them to be irreconcilable, this was in accordance with previous reports (De Roock et al., 2010). We additionally analyzed the correlation between the molecular markers and basic pathological features. The KRAS mutated tumor was reported to be more frequent in female patients with mucinous differentiation and polypoid growth (Wangefjord et al., 2013;Li et al., 2015). In our study, we found it was more frequently observed in female patients, accompanied with perineural invasion, mucin production, and polypoid gross type tumor, which is mainly consistent with the published data from Western countries. As another member of the RAS family, NRAS mutation is rare and shows no significant relevance with histologic features (Irahara et al., 2010); nevertheless, we discovered that it was associated with a well/ moderate differentiation grade and ulcerative carcinoma. The BRAF-mutated tumor was more often related with pathologic characteristics such as a location on the right-side, lymph-node metastases, mucin component, tumor thrombus, perineural invasion, and low differentiation grade. These findings were in agreement with existing studies (Li et al., 2015). The MSI-H phenotype has been reported to be involved in poor differentiation, mucinous histology, and right-sided colon location (Battaglin et al., 2018). Moreover, we found that it was regularly observed in younger patients (<50 years old, p = 0.000), a deeper infiltration of bowel wall, and polypoid gross type tumor. The distribution of the correlations between molecular markers and histologic features in our study were not completely in accordance with the published data of Western countries and other regions in China, and we assume that this difference is a consequence of multiple sample sizes and regional diversity.
For the treatment of potentially resectable colon cancer with RAS and BRAF wild-type status, LCRC with FOLFOXIRI ± cetuximab is recommended to patients, while FOLFOXIRI ± FIGURE 2 | The correlation between clinicopathological characteristics and tumor locations that shows statistically significant difference. Compared with LCRC, RCC is more relevant with higher frequencies of MSI-high (MSI-H) and BRAF mutation, and higher incidence in female and younger patients, and was closely associated with bowel wall invasion, metastasis, poor differentiation, and mucinous tumors. (LCRC, left-sided colorectal cancer; RCC, right-sided colorectal cancer; MSI, microsatellite instability; MSS/L, microsatellite stable plus microsatellite instability-low; MSI-H, microsatellite instability-high; Old, ≥50 years old; Young, <50 years old; T, bowel wall invasion; M, metastasis).   FIGURE 3 | The correlation between clinicopathological characteristics and molecular markers that shows statistically significant difference. KRAS mutation more commonly occurred in female patients with a mucinous, perineural invasive, and polypoid tumor while NRAS mutation is more associated with well-differentiated ulcerative tumors. BRAF mutation was more relevant to lymph node involvement, deeper infiltration of the bowel wall, mucinous, a poorly-differentiated tumor with thrombus, and perineural invasion. MSI-H was more commonly found in younger patients with deeper bowel wall infiltration and a poorly-differentiated polypoid tumor. (MSI, microsatellite instability; MSS/L, microsatellite stable plus microsatellite instability-low; MSI-H, microsatellite instability-high; Old, ≥50 years old; Young, <50 years old; T, bowel wall invasion; N, lymph node involvement; M, metastasis). bevacizumab is recommended for RCC patients. For patients with RAS or BRAF mutation, regardless of tumor location, FOLFOXIRI ± bevacizumab is recommended (Falcone et al., 2007). For the palliative treatment of colon cancer, further classification has been made with respect to the tumor location. In first line treatment, patients with both KRAS, NRAS, and BRAF wild-type are suitable for intense medical treatment, doublet chemotherapy plus cetuximab is reco mmen de d to the LCRC patie nts and d oub let chemotherapy plus bevacizumab is preferred for RCC patients (Tejpar et al., 2017). For those who are not suitable to undergo intense care but have an MSH status, immune checkpoint inhibitors are recommended. In second line treatment, despite the status of the RAS/BRAF gene, immune checkpoint inhibitors are recommended for patients with an MSH status (Diaz and Le, 2015;Overman et al., 2018). Patients who are RAS wild-type and BRAF V600E mutated are recommended a VIC regimen. In third line treatment, despite the status of the RAS and BRAF gene, fruquinitinib is recommended Yuan et al., 2019).
There were several limitations in our study: (1) a single medical center study; (2) lack of clinical treatment data such as applications of target drugs, chemotherapy, or radiotherapy; (3) no track of survival data; and (4) other molecular markers like CIMP, 18q loss of heterozygosity, CIMP were not included in this research.
In summary, our study collected data on 2,356 cases of CRC and analyzed the relationship between tumor location, clinicopathological, and molecular features. We found that RCC (vs LCRC) was significantly associated with a higher incidence of the MSI-H and BRAF mutation but showed no measurable difference on the KRAS and NRAS mutation. Considering the limited reports published on the correlation between tumor location and molecular markers for eastern China patients, our study provides a valuable reference for physicians and researchers to study CRC.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
Informed consent of all patients was obtained and the Ethic Committee of Qingdao University approved this study.

AUTHOR CONTRIBUTIONS
YS and XX conceived and designed the study. XX contributed reagents, protocols, and materials. WR, GL, YX, and XW conducted the experiments and analyzed the raw experimental data. LZ provided pathological diagnosis. YS and LW collected and statistically analyzed the data. YS wrote the paper. XX modified the manuscript.