Favorable prognosis in colorectal cancer patients with co-expression of c-MYC and ß-catenin

The purpose of our research was to determine the prognostic impact and clinicopathological feature of c-MYC and β-catenin overexpression in colorectal cancer (CRC) patients. Using immunohistochemistry (IHC), we measured the c-MYC and β-catenin expression in 367 consecutive CRC patients retrospectively (cohort 1). Also, c-MYC expression was measured by mRNA in situ hybridization. Moreover, to analyze regional heterogeneity, three sites of CRC including the primary, distant and lymph node metastasis were evaluated in 176 advanced CRC patients (cohort 2). In cohort 1, c-MYC protein and mRNA overexpression and ß-catenin nuclear expression were found in 201 (54.8 %), 241 (65.7 %) and 221 (60.2 %) of 367 patients, respectively, each of which was associated with improved prognosis (P = 0.011, P = 0.012 and P = 0.033, respectively). Moreover, co-expression of c-MYC and ß-catenin was significantly correlated with longer survival by univariate (P = 0.012) and multivariate (P = 0.048) studies. Overexpression of c-MYC protein was associated with mRNA overexpression (ρ, 0.479; P < 0.001) and nuclear ß-catenin expression (ρ, 0.282; P < 0.001). Expression of c-MYC and ß-catenin was heterogeneous depending on location in advanced CRC patients (cohort 2). Nevertheless, both c-MYC and ß-catenin expression in primary cancer were significantly correlated with improved survival in univariate (P = 0.001) and multivariate (P = 0.002) analyses. c-MYC and ß-catenin expression of lymph node or distant metastatic tumor was not significantly correlated with patients’ prognosis (P > 0.05). Co-expression of c-MYC and ß-catenin was independently correlated with favorable prognosis in CRC patient. We concluded that the expression of c-MYC and ß-catenin might be useful predicting indicator of CRC patient’s prognosis.


Background
The c-MYC protein encode by c-MYC gene, acts as transcription factor for variable cellular function including proliferation, differentiation, metabolism, survival, and apoptosis [1,2]. The c-MYC gene can promote tumorigenesis in various malignant tumors [3,4] and mediate the critical role in the colorectal cancer (CRC) progression [5,6]. Deregulation of c-MYC is a consequence of mutations in APC, a central hub in early colorectal carcinogenesis [7].
c-MYC gene amplification, translocation, and alteration of regulatory molecules are major causes of c-MYC protein overexpression [8,9]. Previously, other group indicated that c-MYC amplification and overexpression was showed in approximately 10 and 70 % in CRC, respectively [10]. These studies have deduced that overexpression of c-MYC is controlled by mechanisms other than gene amplification [10]. In recent years, it has been evident that the mechanism of c-MYC overexpression is not restricted to genetic alterations, such as amplification or translocation, but can also occur as a consequence of abnormalities in regulatory molecules [11]; in CRC, ß-catenin is one such regulatory molecule. It is now well established that APC gene mutation, a key driver of adenoma-carcinoma transition, often leads to altered ß-catenin regulation via the wellstudied Wnt signaling pathway [12][13][14]. Regulation of this pathway occurred while changing in nuclear ß-catenin protein levels. A destruction complex maintains a low cytoplasmic concentration of ß-catenin when the Wnt signaling pathway is inactivated. On the contrary, the destruction complex degrades and ß-catenin increases in the cytoplasm, leading to its migration to the nucleus, where it work like a transcriptional factor for c-MYC and cyclin D1 [15,16]. Recent studies reported CRCs with marked WNT and c-MYC signaling activation as a distinct molecular subtype by gene expressionbased CRC classifications, which was associated with relatively better prognosis [17,18]. It suggests that CRCs with activated c-MYC via Wnt signaling pathway have distinct clinicopathologic characteristics, but it has not been confirmed.
Nevertheless, there were a few researches that reported clinicopathological impact of c-MYC and ß-catenin status in CRC. Their prognostic value for CRC patients remains debatable. A recent study reported that c-MYC protein overexpression obtained by immunohistochemistry (IHC) was significantly correlated with better survival of CRC patients [19]. In contrast, other researchers conducted a meta-analysis showing that the accumulation of nuclear ß-catenin could be a biomarker for advanced stage and worse survival of CRC [20]. However, the correlation between immunohistochemical nuclear ß-catenin expression and patient prognosis is quite controversial. Consequently, it is necessary to further evaluate c-MYC and ß-catenin expression to reach a conclusion about their prognostic value.
Recently, the systemic chemotherapy in CRC has made a remarkable development, and targeted therapy has been used to increase survival in advanced CRC patients [21]. However, targeted therapy has no effect in some CRC patients, despite presenting positivity for target-therapy specific molecular examination [22]. Several researchers have demonstrated that CRC shows a regional heterogeneity in KRAS, EGFR, and BRAF mutation, thus tumor heterogeneity may explain this discrepancy between molecular alteration and responses of targeted therapy [23][24][25]. Therefore, molecular alterations between the metastatic and primary lesions need to be discovered to enhance the treatment effect of metastatic CRCs.
The aim of our research was to evaluate the clinical implication of c-MYC and ß-catenin in CRC and evaluate their heterogeneity in primary and distant metastatic tumors. We also analyzed the association between c-MYC and ß-catenin status.

Collection of samples
A total of 543 CRC cases of this study had been collected in our previous study [26]. To investigate the clinicopathological significance of c-MYC and ß-catenin expression, we collected 367 consecutive CRC patients who underwent surgery between 2005 and 2006 at Seoul National University Bundang Hospital (cohort 1). Additionally, to evaluate the locational heterogeneity of c-MYC and ß-catenin expression, we collected synchronous or metachronous metastatic 176 CRC patients with who had received surgery between 2003 and 2004, as well as between 2007 and 2009 excluding any patient already enrolled in cohort 1 (cohort 2). Pathologists K.S.L and H.S.L reviewed all the cases. Cancer stage was determined from the American Joint Committee on Cancer (AJCC) 7th edition. Clinical and pathologic information was acquired from hospital medical records including patient' outcome and survival.

Tissue array method
Tissue microarray (TMA) was constructed with representative lesions of the donor formalin-fixed paraffinembedded (FFPE) CRC tissues as previously described [27].
Immunohistochemistry c-MYC IHC analysis was performed using an antibody against c-MYC (clone Y69, catalog ab32072, Abcam, Burlingame, CA, USA). ß-catenin IHC used a commercially available antibody against ß-catenin (clone CAT-5H10, Invitrogen, Camarillo, CA, USA). The staining process was performed using an automated immunostainer (BenchMark XT, Ventana Medical Systems), according to the manufacturer's recommendations. Normal colonic mucosa cells were considered as internal negative controls. Normal mucosa was negative for c-MYC nuclear immunostaining. ß-catenin was negative in inflammatory cells, but expressed in colonic epithelium in three patterns: membrane, cytoplasm, and nucleus. We only found ß-catenin nuclear expression in malignant cells. For statistical analysis, c-MYC and ß-catenin immunostaining were regarded as positive when they were expressed in more than 10 % of neoplastic nucleus in any intensity ( Fig. 1) [19,28]. Negative controls were obtained omitting the primary antibody for each immunostaining.

mRNA in situ hybridization
For the detection of c-MYC mRNA transcripts, the RNAscope 2.0 HD detection kit (Advanced Cell Diagnostics, Hayward, CA, USA) was used according to the manufacturer's protocols. The experimental data was interpreted according to the manual in the RNAscope FFPE assay kit: no staining or less than 1 dot/cell at 40× objective view (score of 0); staining in 1-3 dots/cell visible at 20-40× objective view (score of 1); staining in 4-10 dots/cell with no or very few dot clusters visible at 20-40× objective view (score of 2); staining in >10 dots/cell with less than 10 % of positive cells having dot clusters visible at 20× objective view (score of 3); staining in >10 dots/cell with more than 10 % of positive cells having dot clusters visible at 20× objective view (score of 4). A score of 2-4 indicates c-MYC mRNA overexpression (Fig. 1). UBC (ubiquitin C) and dapB (a bacterial gene) were used for positive and negative controls. Tissues were regarded as appropriate when the UBC mRNA signals were visible without difficulty at 10× magnification and the dapB signal was not visible.

Statistical analyses
All statistical analysis was performed with the SPSS version 21 (IBM, Armonk, NY, USA) software. The Chi-square test or Fisher's exact test was used for evaluating the correlation between clinicopathological characteristics and c-MYC and ß-catenin expression. The Pearson correlation coefficient was used for analyzing comparison of detection methods. The Kaplan-Meier method with the log-rank test and multivariate regression were performed to assess survival difference. The survival results were determined with hazard ratio (HR) and its 95 % confidence interval (CI). P < 0.05 was considered statistically significant.

Clinicopathological impacts of c-MYC and ß-catenin expression in consecutive CRC patients
In 367 patients (cohort 1), a c-MYC mRNA in situ hybridization score of 0 was observed in 34 (9.3 %), a score of 1 in 92 (25.1 %), a score of 2 in 123 (33.5 %), a score of 3 in 93 (25.3 %), and a score of 4 in 25 (6.8 %). Consequentially, overexpression of c-MYC mRNA (a score of 2-4) was observed in 241 patients (65.7 %). c-MYC protein overexpression was observed in 201 (54.8 %), and ß-catenin nuclear overexpression was observed in 221 (60.2 %) patients. Table 1 demonstrates the correlations between c-MYC and ß-catenin overexpression and clinicopathological parameters. c-MYC protein overexpression was associated with non-aggressive characteristics, including early pT stage, low-grade differentiation, absence of perineural invasion, and smaller tumor size (P < 0.001, P = 0.007, P = 0.025 and P < 0.001, respectively). In addition, c-MYC protein overexpression was associated with a tumor location in the recto-sigmoid colon. Increased levels of the c-MYC mRNA transcript were associated with microsatellite stable CRC (P = 0.019), located in the sigmoid colon and rectum, and with less aggressive features, similarly to c-MYC protein overexpression. Likewise, ßcatenin nuclear expression was frequently detected in tumors of the recto-sigmoid colon, of low-grade differentiation (P = 0.006), of small size (P = 0.007) and microsatellite stable CRC (P < 0.001).

Locational heterogeneity of c-MYC and ß-catenin status
For analysis the locational heterogeneity of c-MYC and ß-catenin expression, we investigated cancer from three lesion, including the primary, distant and lymph node metastasis (cohort 2). All 176 cases had distant metastatic lesions. Among them, 142 cases had lymph node metastases, even though we dissected more than 20 lymph nodes in all CRC patients respectively. The clinicopathological features of the cohort 2 are indicated in Table 2 as previously reported [31]. Not every cohort 2 patients are stage IV due to metachronous metastasis which develops consequently after treatment of the first primary tumor. The distant metastatic sites were described below: liver in 82 cases (46.6 %), lung in 37 cases (21.0 %), peritoneal seeding in 38 cases (21.6 %), distant lymph nodes in 3 cases (1.7 %), and ovary in 16 cases (9.0 %).
The locational heterogeneity of c-MYC and ß-catenin status is demonstrated in Table 3. Discordance of c-MYC protein overexpression between the primary and distant metastatic cancer was detected in 45.5 % (80 out of 176) of cases, and discordance between the primary and lymph node metastatic cancer was observed in 31.5 % (35 out of 111) of cases. Discordance of c-MYC mRNA overexpression between the primary and distant metastatic cancer was detected in 25.6 % (45 out of 176) of cases, and discordance between the primary and lymph node metastatic cancer was observed in 30.6 % (34 out of 111) of cases. Discordance of nuclear ß-catenin expression between the primary and distant metastatic cancer was detected in 29.0 % (51 out of 176) of cases, while discordance between the primary and lymph node metastatic cancer was observed in 26.1 % (29 out of 111) of cases. Consequently, locational heterogeneity of c-MYC and ß-catenin expression was frequently seen in advanced CRC.

Prognostic impact of c-MYC and ß-catenin expression in CRC
All CRC patients of our study were successfully included survival analysis ( Fig. 2 and Additional file 1: Table S1). In the consecutive cohort (cohort 1), the median followup was 55 months (1-85 months) as previous reported [26]. c-MYC protein overexpression, mRNA overexpression and nuclear ß-catenin expression were significantly correlated with an improved survival in Kaplan-Meier analysis (P = 0.011, P = 0.012 and P = 0.033, respectively). When prognostic analysis was performed using the combined status of c-MYC and ß-catenin expression, positivity for both proteins (c-MYC/ß-catenin: +/+) was observed 84/367 (22.9 %) cases and was significantly correlated with an improved survival (P = 0.012). We additionally investigated the c-MYC protein overexpression by density of staining -scoring each tumor as low (0-1) to high (2)(3) in cohort 1. The percentage of positive neoplastic cells was correlated with density of staining of neoplastic cells (ρ, 0.789; P < 0.001), which was categorized as strong correlation [30]. However, the staining density of c-MYC protein was not significantly correlated with patients' prognosis (P = 0.070, Additional file 2: Figure S1).
In the cohort with metastases (cohort 2), the median follow-up was 43 months (1-105 months), as previous reported [31]. In the primary cancer, Kaplan-Meier analysis showed that c-MYC protein overexpression and nuclear ß-catenin expression were significantly correlated with improved prognosis (P = 0.005 and P = 0.007, respectively), but mRNA overexpression was not (P = 0.258). Co-expression of c-MYC and ß-catenin (c-MYC/    Figure S1). The multivariate Cox's proportional hazards regression model of c-MYC and ß-catenin expression was described in Table 4, and indicated that co-expression of c-MYC and ß-catenin was an independent prognostic factor for better survival in both cohort 1 and primary tumor of cohort 2 (P = 0.048 and P = 0.002, respectively). However, individual analysis of c-MYC protein overexpression, mRNA overexpression, and nuclear ß-catenin expression did not independently predict better prognosis.

Discussion
There were several studies on c-MYC status in various malignancies. Some malignant tumors with c-MYC overexpression including gastric carcinoma, esophageal squamous cell carcinoma, and soft tissue leiomyosarcoma are associated with poor survival [32][33][34]. Likewise, several cancers with c-MYC gene amplification tend to be correlated with poor survival [35][36][37]. Interestingly, c-MYC mRNA overexpression in CRC was reported to be correlated with improved survival [5], but this was opposite result to previous other study [38]. Nevertheless, recent research indicated that immunohistochemical c-MYC overexpression was significantly associated with better prognosis of CRC patients in univariate model, but not in multivariate model [19]. In addition, many studies have shown that ß-catenin is crucial part of the Wnt signaling pathway in CRC development [39]. Recently, a meta-analysis study showed that nuclear overexpression of ß-catenin appeared to be associated with progressive disease for CRC patients [20]. However, prognostic value of nuclear overexpression of ß-catenin in CRC patients remains controversial [40,41].
In our study, overexpression of c-MYC protein in the consecutive cohort was significantly correlated with an improved prognosis in univariate model, but not in multivariate model. The prognostic significance of nuclear ß-catenin expression is similar to that of c-MYC protein overexpression. We performed a combined analysis of c-MYC and ß-catenin expression because these proteins are closely related. Astonishingly, co-expression of c-MYC and ß-catenin correlated with an improved prognosis by univariate and multivariate analysis. Although the advanced cohort was mainly consisted of stage IV CRC patients (111 cases; 63.1 %), coexpression of c-MYC and ß-catenin was independently predicted favorable prognosis. Furthermore, overexpression of c-MYC and ß-catenin-except c-MYC mRNA-in the advanced cohort was significantly correlated with better prognosis using a univariate model. Consequently, co-expression of c-MYC and ß-catenin determined by IHC might be of use in the assessment of CRC patients.
We also demonstrated that ß-catenin nuclear expression significantly associated with its target molecule c-MYC in CRC patients (ρ, 0.282; P < 0.001). Overexpression of c-MYC can be caused by complex regulatory pathways and multiple communications with other factors, rather than just c-MYC gene alterations [42]. An example of such a mechanism is signaling via ß-catenin, a c-MYC regulator whose nuclear accumulation is correlated with c-MYC overexpression [7,43]. ß-catenin increases in the cytoplasm and undergoes translocation to the nucleus, where it plays as a transcription factor for target genes such as c-MYC [16]. These processes explain that nuclear expression of ß-catenin is partly responsible for c-MYC overexpression. As a result, coexpression of c-MYC and ß-catenin can be considered as c-MYC overexpression via ß-catenin in CRC. The APC gene mutation is the initial step of CRC oncogenesis [44] and often lead to deregulation of ß-catenin [45]. Thus, ß-catenin-dependent c-MYC overexpression can be suggested in early colorectal carcinogenesis. In addition, recent studies suggested that high-level nuclear β-catenin in CRC was significantly correlated with high Ki67 expression [46], and indicated that tumor proliferative activity was inversely related to CRC aggressiveness and metastases [47,48]. For this reason, c-MYC overexpression via ß-catenin might have an influence on improved prognosis of CRC patients. Moreover, E Melo et al. reported that CpG island methylation interrupts several Wnt target genes, including ASCL2 and LGR5 during CRC progression and promoter methylation of Wnt target genes is a powerful predictive factor for CRC relapse [16]. Therefore, silencing of ß-catenin/ Wnt pathway by methylation generates CRC progression and worse prognosis. It is noteworthy that our   result adds clinical evidence to support these previous studies [16][17][18][46][47][48]. The lack of ß-catenin expression in CRC patients with presenting c-MYC overexpression shows a rather worse survival, presumably because, in these patients, the c-MYC is controlled by other regulatory factors. Future studies will be required to dissect the different mechanisms of ß-catenin-mediated c-MYC overexpression.
In the advanced cohort, regional heterogeneity of c-MYC and ß-catenin expression was frequently observed in advanced CRC. Nonetheless, c-MYC and ß-catenin expression was associated with better prognosis in the primary, not distant and lymph node metastatic cancer. Consequently, when we evaluate prognosis with c-MYC and ß-catenin, tissue from primary CRC should be used. In daily practice for pathologists, metastatic cancer tissue can only occasionally be obtained. Some researchers suggest that regional heterogeneity should be considered as a potential limitation to the evaluation of prognostic and therapeutic value in tissue from metastatic lesions [49,50]. Therefore, the importance of regional heterogeneity must be assessed in biomarker research.

Conclusions
Our study comprehensively evaluated the c-MYC and ßcatenin status of CRC patients. Overexpression of c-MYC protein, mRNA, and ß-catenin nuclear expression were observed in 54.8, 65.7, and 60.2 % of consecutive CRC patients, respectively. c-MYC protein overexpression was significantly correlated with mRNA overexpression and with ß-catenin nuclear expression. Interestingly, co-expression of c-MYC and ß-catenin-in other words, c-MYC overexpression via ß-catenin-was an independent improved prognostic factor in both the consecutive and advanced cohort. These findings indicate that c-MYC and ß-catenin IHC can be used as prognostic marker of CRC patients. However, further investigations on the detailed mechanism of connection between c-MYC and ß-catenin and its impact on patients' outcome in CRC are needed.