CEND1 and GJB6 co-expression as a prognostic factor in colorectal cancer

Background: Perineural invasion is an important mechanism of cancer progression that is not well understood at present. The present study explored the relationship between GJB6, CEND1, and cell-cell communication as regulators of colorectal cancer patient survival and clinicopathological ndings. Method: Immunohistochemical staining was performed to assess CEND1 and GJB6 expression levels in CRC patient samples, while survival outcomes were assessed using Kaplan-Meier curves and log-rank tests. Results: Elevated CEND1 expression was associated with tumor location, poor differentiation, and perineural invasion, while GJB6 expression was positively correlated with TNM stage, distant metastasis, and perineural invasion. In addition, GJB6 and CEND1 protein levels were correlated with one another in CRC patient tissues, and high expression of both of these proteins was associated with a higher risk of perineural invasion. CEND1+/GJB6+ status was also associated with poorer patient survival, highlighting both of these proteins as prognostic biomarkers in CRC patients. Conclusion: Elevated levels of CEND1 and GJB6 are independent predictors of poorer CRC patient prognosis.


Introduction
Colorectal cancer (CRC) is among the leading causes of cancer-related mortality globally [1]. Aberrant neuronal distribution can facilitate tumor progression, as tumors can spread via perineural invasion.
However, the speci c role of neurons in the context of cancer progression is not well understood, particularly in the context of CRC. While local CRC patients have a 5-year survival rate of almost 90%, this rate falls to just 64% in those exhibiting perineural invasion [2]. CRC is currently treated via a combination of radical surgery and adjuvant chemotherapy or related approaches [3,4]. Some recent evidence suggests that the innervation of tumors can facilitate their growth. As such, it is vital that the molecular mechanisms governing CRC perineural invasion be clari ed [5][6][7][8]. To that end, we have focused on identifying novel biomarkers of nerve in ltration associated with this cancer type.
Many different neuron-related proteins have been identi ed in cancer tissues. For example, CEND1 (Cell Cycle Exit and Neuronal Differentiation 1) is a gene encoding the BM88 protein that is speci c to neuronal lineage cells wherein it regulates cell cycle exit and differentiation [9,10]. As such, this gene may play a role in the development of perineural invasion in certain cancer types [11,12]. The development of such invasion is inextricably linked to cell-cell contact. Gap junctions form as membrane channels which facilitate small molecule transfer between neighboring cells, regulating processes such as cell cycle progression, signal transduction, and invasion. These junctions are formed by different connexin proteins, of which the isotypes GJA1 (connexin 43) [13,14]and GJB2 (connexin 26) [15,16]are the best studied.
Both GJA1 and GJB2 are frequently upregulated in tumors, wherein they are associated with invasion, migration, and poor prognostic outcomes. GJB6 is a gene encoding connexin 30 that has also recently been shown to regulate metastasis and chemoresistance in lung cancer [17], gastric cancer [18], renal cancer [19], glioma [20], and breast cancer [21]. The functional relevance of GJB6 in CRC, however, has yet to be de ned.
CEND1 is involved in regulating CCND1 [10], which is a gene speci cally associated with the G1/S phase of the cell cycle, while GJB6 has been linked to the regulation of cellular proliferation via multiple pathways [22], including pathways related to CCND1. As such, we herein explored the potential prognostic relevance of CEND1 (BM88) and GJB6 (connexin30) and the correlations between these proteins in CRC patient samples.

Patients and samples
In total, samples were collected from 87 CRC (51 male, 36 female; median age: 69 years, range: 31-81) patients that underwent surgical tumor resection without having undergone previous radiotherapy or chemotherapy at the rst a liated hospital of Xi'an Jiaotong university between January 2012 and December 2014. Pairs of tumor and paracancerous tissues were obtained from each patient, with all histological diagnoses having been con rmed independently by thee experienced pathologists. The Ethics Committee of the rst a liated hospital of Xi'an Jiaotong university approved the present study, which was consistent with the Declaration of Helsinki. All patients provided informed consent. Table I, with TNM staging having been conducted as per the American Joint Committee on Cancer (AJCC) criteria. Tissues were snap-frozen immediately following collection and stored at -80°C. Tissues used for immunohistochemical (IHC) staining were xed for 24 h with 4% paraformaldehyde, dehydrated, and para n-embedded.

Patient clinicopathological information is compiled in
Immunohistochemistry IHC staining was used to assess protein expression. Brie y, samples were depara nized, rehydrated, blocked for 10 min with 3% H 2 O 2 to suppress endogenous peroxidase activity, and then probed overnight with antibodies speci c for BM88 (ab113087, 1:200, Abcam) or connexin 30 (ab199178, 1:200, Abcam) at 4°C. Sections were then probed for 30 min at room temperature with biotinylated secondary antibodies, and DAB was then used to detect substrates of interest, while hematoxylin was used for nuclear counterstaining.
Numbers of positive cells in 10 random elds of view were counted via microscopy, with staining intensity and percentage positive cells both being scores. Samples were then categorized into the following groups: negative (-), ≤ 5% stained cells, 0 points; weak expression (+), 1-3 points; moderate expression (++), 4-5 points; strong expression (+++), 6-7 points. Samples in the negative and weak staining groups (-and +) were considered to be negative for the purposes of statistical analyses, whereas moderate and strong staining groups (++ and +++) were considered positive.

Statistical analysis
Data were analyzed via Student's t-tests and chi-squared tests as appropriate, with Spearman correlation analyses being used to gauge the relationship between the expression of CEND1 and GJB6. SPSS v22.0 (IL, USA) was used for all statistical testing, with P < 0.05 as the signi cance threshold.

Results
Evaluation of the relationship between CEND1 expression and CRC patient characteristics Pathological and demographic data corresponding to patients in the present study is compiled in Table I. An immunohistochemical staining approach was rst used to evaluate CEND1 protein levels in the samples from these CRC patients. Relative to paired paracancerous tissue samples (18/87, 20.69%), signi cantly more tumor samples were CEND1+ (37/87, 42.53%)(P<0.01, Figure 1).
When the relationship between CEND1 protein levels and CRC patient clinicopathological data was evaluated, we found that CEND1 expression was signi cantly related to location (P=0.012), poor differentiation (P=0.024), and perineural invasion (P=0.036) ( Table I).
The relationship between CEND1 and GJB6 expression in CRC As discussed above, CEND1 expression can promote the migration and metastasis of tumor cells through a mechanism potentially linked to the activation of GJB6. As such, we next evaluated the expression of these two proteins simultaneously in 87 CRC patient tumor samples, separating patients into four staining categories: CEND1+/GJB6-, CEND1+/GJB6+, CEND1-/GJB6-, and CEND1-/GJB6+. This analysis revealed CEND1 and GJB6 to be strongly co-expressed, with co-expression of these markers being signi cant (r=0.534, P<0.05; Table II). In contrast, no relationship between CEND1 and GJB6 expression was detected in paracancerous samples (r=0.124, P>0.05; Table II). Overall these ndings suggested GJB6 and CEND1 protein levels to be correlated with one another speci cally in CRC tumor tissues.
The relationship between GJB6 and CEND1 levels and CRC tumor perineural invasion As discussed above and shown in Table I, there was a signi cant relationship between perineural invasion and high expression of CEND1 (15/25, 60%; P = 0.036) or GJB6 (13/25, 52%; P=0.043) in CRC patients. When evaluating the relationship between combined CEND1 and GJB6 expression and perineural invasion status, we found that such invasion was observed for 40% of tumors expressing both markers (10/25). Spearman correlation analyses further con rmed that CEND1 and GJB6 co-expression was signi cantly correlated with perineural invasion in this patient cohort (P<0.05; Table III).
The relationship between GJB6 and CEND1 levels and CRC patient overall survival We next assessed the association between CRC patient survival and the expression of CEND1 and GJB6 for all 87 patients for whom follow-up data through 2020 were available. These patients were strati ed according to their CEND1 and GJB6 co-expression status (CEND1+/ GJB6-, CEND1+/ GJB6+, CEND1-/ GJB6-, and CEND1-/ GJB6+), revealing that patients co-expressing both of these proteins had sign icantly reduced overall survival (OS) relative to the other three groups, thus suggesting that GJB2 and CEND1 are potentially valuable biomarkers of CRC patient outcomes (P<0.05, Figure 2).
To explore these ndings in greater depth, univariate analyses of prognostic factors in these CRC patients were conducted via log-rank tests (Table IV), revealing histological grade, TNM stage, perineural invasion, CEND1 expression, GJB6 expression, and CEND1/GJB6 co-expression to all be signi cant predictors of CRC patient OS (P<0.05, Table III). Subsequent multivariate analysis con rmed that TNM stage (P<0.01), perineural invasion, and CEND1+/GJB6+ expression (P<0.05) were all independent predictors of CRC patient outcomes.

Discussion
Herein, we explored the expression of CEND1 and GJB6 in human CRC tumor tissue samples in order to understand the relationships between these proteins and both clinicopathological ndings and patient prognosis. Overall we observed elevated levels of both CEND1 and GJB6 in tumor tissues relative to paired paracancerous samples. Increased CEND1 expression was associated with tumor location, poor differentiation, and perineural invasion, while GJB6 expression was associated with TNM stage, distant metastasis, and perineural invasion. Notably, levels of CEND1 and GJB6 were correlated with one another in CRC tumor tissues, and high levels of both of these proteins (CEND1+/GJB6+) were associated with poorer patient prognosis and a higher risk of perineural invasion. Multivariate analysis con rmed that both TNM stage and CEND1/GJB6 co-expression were independent risk factors associated with CRC patient survival.
Appropriately managing early-stage CRC patients can be challenging, as tumor resection and adjuvant therapies are often associated with disappointing outcomes. [1] There may therefore be a need to reevaluate patients without any evidence of metastasis prior to chemotherapy in order to consider other relevant pathological ndings such as the presence of perineural invasion [8]. While lymph node metastasis and distant metastasis are well known to be associated with a poor prognosis, perineural invasion is often overlooked as a prognostic indicator, as it is not taken under consideration in the context of TNM staging. Future research is essential to understand how perineural in ltration in uences overall CRC disease progression.
Perineural invasion is an important event that occurs in the context of cancer progression wherein neonatal neural cells and axons can penetrate the tumor microenvironment in response to certain cytokines in a mechanism that may mimic normal neurogenesis. CEND1 is a neuronal lineage-speci c regulator of cell cycle progression and differentiation that controls the division of neuronal precursor cells and associated neuronal maturation [10]. Its relevance in the context of oncogenesis is not well understood. However, given that it regulates the cell cycle via a p53-dependent/Cyclin D1/pRb signaling pathway, it has the clear potential to in uence tumor cell malignancy [11]. We herein con rmed this hypothesis, highlighting CEND1 as a promising therapeutic target in CRC that warrants further study.
Communication between cells is an essential facet of the tumor microenvironment, regulating key processes including perineural invasion-associated tumor progression. Gap junctions and the connexins that compose them are essential mediators of such cell-cell communication, facilitating metabolic and signaling exchanges between cells that can in uence the progression of diseases including sensorineural hearing loss, reversible blindness, Charcot-Marie-Tooth X-linked disease, demyelinating diseases, skin diseases, arrhythmias, and cancer [23][24][25]. While connexin 26 and connexin 43 are often studies in the context of cancer, the role of connexin 30 (encoded by GJB6) in CRC remains to be clari ed. As gap junctions can form from heteromeric or homomeric combinations of connexins, relative proportions of these proteins in cells are critical to effective cell-cell communication. Connexin 43 has previously been identi ed as a biomarker of positive breast cancer patient prognosis [26], whereas connexin 30 was suggested to be associated with a negative prognosis. In this study, we determined that the expression of GJB6 was associated with poor CRC patient outcomes and that this was correlated with the coexpression of CEND1. CEND1 may interact with one another through both the CCND1 pathway and other mechanisms, given that CEND1 is sensitive to intracellular calcium concentrations, which are in turn modulated by gap junction channels such as those formed by connexin 30. However, more research is required to elucidate these mechanisms.

Page 7/15
Overall, the results of the present study suggested that CEND1 and GJB6 may regulate CRC tumor cell-cell communication in the context of perineural invasion. Future studies will help to clarify the roles of connexins and BM88 in this oncogenic context, aiding in the development ofnovel e cacious treatments for this deadly disease.

Declarations
Ethical Approval and Consent to participate This project was approved by the ethics committee of the First A liated Hospital, Medical College, Xi'an Jiaotong University for clinical research, and all specimens were obtained from patients with informed consent.

Human and Animal Rights
There is no experiments on human or animals.

Consent for publication
All authors have reviewed the nal version of the manuscript and approved it for publication.

Availability of data and material
We declared that materials described in the manuscript, including all relevant raw data, will be freely available to any scientist wishing to use them for non-commercial purposes, without breaching participant con dentiality.

Con ict of Interest
There is no competing interest in this work.  Figure 1 Immunohistochemistry of CEND1 and GJB6 in colorectal cancer (CRC). A: High expression of CEND1 in tumor tissues (×200); B: High expression of CEND1 in Tumor tissues (×200); C: Low expression of GJB6 in CRC tissues (×200); D: Low expression of GJB6 in adjacent non-tumors tissues (×200).