Expression of a novel CNPY2 isoform in colorectal cancer and its association with oncologic prognosis

CNPY2 isoform expression in CRC cell lines and tissues

CNPY2 isoform1 and isoform2 mRNA was detectable in normal colonic epithelial cells (NCM460) and CRC cells, including SW480, DLD-1, SW620 and HT29, by quantitative real-time polymerase chain reaction (qPCR) analysis. CNPY2 isoform1 was increased in DLD-1, SW620 and HT29 cells, but not in SW480 cells, compared to that in NCM460 cells (Figure 1A, P < 0.01). Meanwhile, expression of CNPY2 isoform2 was significantly higher in all CRC cell lines compared to the NCM460 line (Figure 1B, P < 0.01). We selected clone 2 of the CNPY2 isoform2 antibody for protein detection as it was the best one for immunohistochemistry (IHC) (Supplemental Figure 3). We found that CNPY2 isoform2 protein was highly expressed in the CRC cell lines DLD-1, HT29 and SW620 but was weakly expressed in SW480 and NCM460 cells by western blot (Figure 1C and 1D). In addition, we found that the two isoforms of CNPY2 mRNA were significantly increased in 5 CRC tissues compared with the paired tumor-adjacent normal tissues (Figure 1E, P < 0.01). Moreover, the expression of CNPY2 isoform2 in CRC tissues was also significantly higher than that of CNPY2 isoform1 (Figure 1F, P < 0.05).

Figure 1. CNPY2 mRNA expression evaluated by RT-PCR. (A) CNPY2 isoform1 mRNA was significantly higher in CRC cell lines (DLD-1, SW620, and HT29) than normal colonic epithelial cells (NCM460) **, P < 0.01. (B) CNPY2 isoform2 mRNA was significantly increased in all CRC cell lines (SW480, DLD-1, SW620, and HT29) compared to that in NCM460 cells, **, P < 0.01. (C) Protein expression of CNPY2 isoform2 in CRC cell lines (SW480, DLD-1, SW620, and HT29) and NCM460 cells determined by western blot. (D) The relative expression of CNPY2 isoform2 in NCM460, SW480, DLD-1, SW620, and HT29 was 0.41, 0.36, 0.75, 1.22 and 1.38, respectively. (E) Expression of the two CNPY2 isoforms was significantly higher in tumor tissues than in tumor-adjacent normal tissues (n = 5, **, P < 0.01). (F) Expression of CNPY2 isoform2 was significantly higher compared to that of CNPY2 isoform1 in tumor tissues (n = 5, *, P < 0.05). Relative expression of CNPY2 mRNA was normalized to the internal reference gene GAPDH.

CNPY2 isoform2 expression in CRC tissues

The CNPY2 isoform2 mRNA levels in the series of 57 tumor tissues and tumor-adjacent normal tissues were further examined in this study. Similar to the results shown in Figure 1E, CNPY2 isoform2 mRNA level was significantly elevated in CRC tumor tissues compared to that in tumor-adjacent normal tissues by qPCR analysis (Figure 2A, P < 0.01). In the cohort of 285 patients, the mean IHC score of CNPY2 isoform2 protein expression was substantially higher in tumor tissues than that in tumor-adjacent normal tissues (7.6 ± 0.2 vs. 1.0 ± 0.1, P < 0.001, Figure 2B). Moreover, increased expression of CNPY2 isoform2 in tumor tissues was observed in all patients with stage I-IV disease compared to that of tumor-adjacent normal tissues (P < 0.001, Figure 2C-F), while the expression levels in tumor tissues were comparable among various stages at both the mRNA and protein level (Figure 2G and H). As shown in Figure 3, expression of CNPY2 isoform2 was observed in primary tumors, liver metastatic tumors, tumor-adjacent normal tissues and normal liver tissues. In addition, positive staining of CNPY2 isoform2 protein was predominantly enriched in the cytoplasm of epithelial cells (Figure 3), while some protein was detected in the extracellular matrix (Figure 3A2).

Figure 2. CNPY2 isoform2 expression in CRC tissues detected by RT-PCR and TMA-IHC. (A) Preferential expression of CNPY2 isoform2 mRNA was observed in CRC tissues compared matched tumor-adjacent normal tissues, (n = 57, **, P < 0.01). (B) Increased expression of CNPY2 isoform2 protein was detected in tumor tissues compared to matched tumor-adjacent normal tissues among all patients (n = 285, 7.6 ± 0.2 vs. 1.0 ± 0.1, **, P < 0.01). (C-F) Increased expression of CNPY2 isoform2 protein was detected in tumor tissues compared to matched tumor-adjacent normal tissues among patients with stage I-IV disease (C: Stage I,7.7 ± 0.6 vs. 0.5 ± 0.1; D: Stage II,7.4 ± 0.3 vs. 1.0 ± 0.2; E: Stage III,8.6 ± 0.3 vs. 1.3 ± 0.2; F: Stage IV,7.5 ± 0.4 vs. 1.0 ± 0.2; **, all P < 0.01). (G) Expression level of CNPY2 isoform2 mRNA was slightly higher in stage II CRC but did not reach statistical significance compared to that in other stages (n = 57, P > 0.05). (H) CNPY2 isoform2 protein expression was comparable among patients with disease stage I-IV (n =2 85, P > 0.05).

Figure 3. Representative TMA-IHC staining of CNPY2 isoform2 protein. (A1, A2) Colon adenocarcinoma with high CNPY2 isoform2 expression in the cytoplasm of CRC cells and some expression in the extracellular matrix; (B1, B2) Colon adenocarcinoma with low CNPY2 isoform2 staining in the cytoplasm of CRC cells; (C1, C2) Tumor-adjacent normal tissues with positive staining of CNPY2 isoform2; (D1, D2) Tumor-adjacent normal tissues with negative staining of CNPY2 isoform2. (E1, E2) High CNPY2 isoform2 expression in liver metastatic tumor tissues; (F1, F2) High CNPY2 isoform2 expression in normal liver tissues. Original magnification was 40×with 100 μm scale per in A1, B1, C1, D1, E1, and F1; 400× with 100 μm scale per bar in A2, B2, C2, D2, E2, and F2.

Cutoff point for CNPY2 isoform2 expression and clinicopathological characteristics

The 285 cases were randomly divided into a training cohort (n = 142) and an independent validation cohort (n = 143). We selected the optimal cutoff CNPY2 isoform2 score as 3.7 at the highest chi-square value of 7.337 in the training cohort (Figure 4). Patients were further divided into 2 groups: 247 (86.7%) patients in the high CNPY2 isoform2 expression group (IHC score ≥ 3.7) and 38 (13.3%) patients in the low CNPY2 isoform2 expression group (IHC score < 3.7). The associations between CNPY2 isoform2 expression and clinicopathological parameters, including gender, age, tumor location, tumor size, histological type, clinical stage, preoperative serum carcinoembryonic antigen (CEA) and CA199, were assessed. No significant association was found between CNPY2 isoform2 expression with all the above clinicopathological characteristics in total cohorts, the training cohort and the validation cohort (Table 1).

Figure 4. X-tile analysis for OS based on CNPY2 isoform2 expression among patients in the training cohort. The optimal cutoff value of the CNPY2 isoform2 expression score was 3.7 at the maximum chi-square value of 7.337.

Association between the CNPY2 isoform2 expression and survival outcome

The median follow-up period for all the patients was 58 months (range, 6 – 123 months). During the follow-up period, 71 (24.9%) patients died of disease progression. In the training cohort, low expression of CNPY2 isoform2 was associated with a lower 5-year overall survival (OS) rate compared to that with high expression of CNPY2 isoform2 (41.7% vs. 77.7%, P = 0.007, Figure 5A). Low expression of CNPY2 isoform2 also identified patients with a lower 5-year OS rate in the validation cohort (47.1% vs. 78.8%, P = 0.002, Figure 5B). As shown in Table 2, univariate analysis revealed that low CNPY2 isoform2 expression was a predictor of worse 5-year OS in both the training cohort [hazard ratio (HR) = 2.809; 95% confidence interval (CI), 1.288–6.125; P = 0.009) and validation cohort (HR = 2.885; 95% CI, 1.442–5.654; P = 0.003). In the multivariate Cox model, low CNPY2 isoform2 expression was identified as a predictor for shorter 5-year OS in the training cohort (HR = 5.001; 95% CI, 2.156–11.598; P < 0.001) and validated in the validation cohort (HR= 2.443; 95% CI 1.197- 4.983, P = 0.014). In addition, metastatic disease (stage IV) was a predictor of worse 5-year OS for CRC patients in both the training cohort and validation cohort.

Figure 5. Kaplan–Meier curves of OS among patients with colorectal cancer. (A) The training cohort (n = 142). (B) The validation cohort (n = 143).

Clinical specimen collection

Paraffin-embedded specimens were collected from 299 patients diagnosed with colorectal adenocarcinoma who underwent primary tumor resection and recruited at Sun Yat-sen University Cancer Center between February 2006 and January 2014. Patients with preoperative anti-cancer treatment, multiple primary CRCs, or a history of other active malignancies (except for basal cell carcinoma of the skin) and those who died or were lost to follow-up within 6 months after surgery were excluded. A total of 285 eligible cases were finally included in current study. Additionally, fresh CRC cancer tissues and the matched tumor-adjacent normal tissues were harvested from an independent cohort of 57 preoperative treatment naive CRC patients and stored at 80°C for biochemical studies. Basic characteristics of the 57 patients are shown in Supplemental Table 1. The detailed clinicopathologic information of the eligible patients was reviewed through the electronic medical records system, and follow-up data were collected from the tracking system. This study was undertaken in accordance with the ethical standards of the World Medical Association Declaration of Helsinki. The study and consent procedure were approved by the Institutional Research Ethics Committee of Sun Yat-sen University Cancer Center (Approval number: B2017-042-01), and informed consents for using tissue samples, before the initial treatment, were obtained from the patients.

Cell lines and culture

Human CRC cell lines (SW480, SW620, HT29 and DLD-1) were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA), while the human normal colon epithelial cell line NCM460 was acquired from INCELL (San Antonio, TX, USA). Before use, all cells were negatively tested for mycoplasma contamination and authenticated based on STR fingerprinting at Medicine Lab of Forensic Medicine Department of Sun Yat-sen University. HT29 was cultured in McCoy's 5A medium, while the remaining cell lines were cultured with RPMI-1640 medium (Gibco, Grand Island, NY, USA); both were supplemented with 10% fetal bovine serum (FBS, Gibco Invitrogen Co., Grand Island, NY, USA), 100 units/ml penicillin and 100 mg/ml streptomycin. All cell lines were maintained at 37°C in a humidified atmosphere of 5% CO₂.

Generation of CNPY2 isoform2 monoclonal antibodies

The open reading frame of CNPY2 isoform2 (GenBank accession. no. NM_001190991) encoding amino acids (AAs) 21–84 (excluding the first 20 AAs as the predicted signal peptide) was polymerase chain reaction (PCR)-amplified from the total cDNA of human smooth muscle cells and inserted into the pET30a vector between HindIII and XhoI restriction sites, resulting in recombinant CNPY2 fused to a 6×His tag at its C-terminus (CNPY2-His). The E. coli BL21 (DE3) strain was transformed with recombinant plasmids. A single colony was inoculated into medium containing the appropriate antibiotic (kanamycin). Isopropyl beta-D-1-thiogalactopyranoside (IPTG) was introduced for induction, and CNPY2 isoform2-His recombinant protein was purified from inclusion bodies. Proteins were analyzed by 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blot using standard protocols for molecular weight and purity measurements. The theoretical isoelectric point of this recombinant CNPY2 isoform2-His protein is ~7.27, and the theoretical molecular weight is ~7.96 kDa (DNA and protein sequences are shown in Supplemental Table 2). This recombinant CNPY2 isoform2-His protein was used to immunize mice for four rounds of injection, including booster immunization, and eventually, the splenocytes were fused with myeloma cells to generate hybridoma clones. Then, 5 positive clones were selected by enzyme-linked immunosorbent assay (ELISA) using the short peptide as the mapping epitope. We compared the efficacy of these 5 clones using IHC, and finally, we selected the best clone for IHC staining. The selected clone antibody binds to the epitope that was unique to the CNPY2 isoform2 C-terminus region and does not cross react with isoform1 or any other known proteins. Monoclonal antibody was purified from the supernatant of the selected clone using a standard protein A/G protocol.

Detection of CNPY2 mRNA

The total cellular RNA was extracted from frozen samples and cultured cells using TRIzol reagent (Invitrogen, CA, USA), and DNase I was applied to exclude potential genomic DNA contamination. Reverse transcription (RT) was performed using a Revert Aid First Strand cDNA Synthesis Kit (Thermo Scientific, Waltham, MA, USA) following the manufacturer’s instruction. The RT reaction was incubated at 25°C for 5 min, then at 42°C for 60 min and terminated at 70°C for 5 min. qPCR primers were designed and are presented in Supplemental Table 3. The primers designed for the CNPY2 isoform2 only amplified CNPY2 isoform2 and not isoform1 or any other known transcripts, which was confirmed in human smooth muscle cells (Supplemental Figure 2). Real-time PCR amplification was performed with DreamTaq DNA Polymerase (Thermo Scientific, Waltham, MA, USA) using the ROCHE 480 system. First, the reaction was performed at 95°C for 5 min with an S1000 Thermal Cycler (Bio-Rad) for initial denaturation, and then, amplification was performed at 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s for 40 cycles, followed by a final extension at 72°C for 7 min. Relative quantification of CNPY2 mRNA was performed using the 2^−∆∆Ct method.

Western blot analysis

Western blot analysis was conducted following our previously described method [24]. Briefly, equal amounts of protein were loaded and separated using 15% SDS-PAGE. The proteins were then electrophoretically transferred onto a polyvinylidene fluoride (PVDF) membrane. The membrane was incubated with the CNPY2 isoform2 and β-actin antibody (1:5000 dilution, 4970S, Cell Signaling Technologies, Danvers, MA, USA) at 4°C overnight. The membrane was washed and incubated with a secondary antibody at room temperature for 1 hour. After the membrane was washed 3 times with Tween (TBS-T) buffer, the proteins were detected using an enhanced chemiluminescence (ECL) reagent kit.

Tissue microarrays (TMAs) and IHC

TMA slides contained 285 CRC and tumor-adjacent normal tissues, including 48 cases with liver metastasis specimens. TMAs were constructed using a tissue array instrument (Beecher Instruments, Sun Prairie, WI, USA). Briefly, each tissue core with a diameter of 0.6 mm was punched from the marked areas in selected FFPE tissues and organized in the TMA blocks. The blocks were sectioned into 4 μm slices and then mounted onto glass slides. After dewaxing and treatment with 0.3% hydrogen peroxide, the slides were incubated with a primary CNPY2 isoform2 antibody (1:800 dilution) overnight in a moist chamber at 4°C. Subsequently, the slides were incubated with anti-mouse/rabbit secondary antibody from REAL En Vision (Dako, Glostrup, Denmark) for 30 min at 37.5°C. The IHC staining was followed by 3', 3-diaminobenzidine tetrahydrochloride (DAB, Dako, Glostrup, Denmark) staining.

IHC scoring was determined by the percentage and intensity of positively stained cell. The positive staining was scored as follows: “0” (less than 5% positively stained cells), “1” (5–24% positively stained cells), “2” (25–49% positively stained cells), “3” (50–74% positively stained cells), and “4” (75–100% positively stained cells). The intensity was scored according to the following standard: “0” (negative staining); “1” (weak staining); “2” (moderate staining), and “3” (strong staining). The final score was generated by multiplying the percentage score with the staining intensity score. Two trained pathologists blindly evaluated all the specimens.

Definition of survival endpoint

Follow-up data of all included patients were available for analysis. Within the follow-up period, OS was defined as the length of time (in months) from the date of surgery until death from any cause or last follow-up. The last follow-up visit was in March 2017.

Statistical analysis

An optimal cutoff value for the IHC score that separates patients into low and high CNPY2 isoform2 expression groups with respect to OS was determined using X-tile software (version 3.6.1; Yale University, New Haven, CT, USA) at the largest chi-square value in the training cohort, as described previously [25,26]. Categorical variables are presented as percentages, and they were compared using the chi-square test or Fisher’s exact test. Continuous variables are presented as the median (range) or mean (standard deviation). Student’s t-test was used for two-group comparison and one-way ANOVA for multiple-group comparison. Survival outcomes were summarized by the Kaplan–Meier method. Potential effects of clinical variables on OS were examined using univariate Cox’s proportional hazards model; variables that were statistically significant in univariate Cox models were further assessed with multivariate Cox models using a forward stepwise method. HRs and CIs were subsequently calculated. All analyses were performed using IBM SPSS statistics software, version 21.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism version 6.0 (GraphPad Software, La Jolla, CA, USA). P-values less than 0.05 were considered significant based on two-sided statistical tests.

Ethics approval and consent to participate

Study was approved by the Institutional Research Ethics Committee of Sun Yat-sen University Cancer Center (Approval number: B2017-042-01). The informed consents for using tissue samples, before the initial treatment, were obtained from the patients.

Consent for publication

Not applicable.

Availability of data and materials

The datasets analysed during the current study were available from the corresponding author on reasonable request. Anyone who is interested in the information should contact fangyj@sysucc.org.cn and wands@sysucc.org.cn.