Myeloperoxidase (MPO) promoter SNPs rs2243828 (−764T>C) and rs2333227 (G-463A) program malignant phenotypes by regulating MPO transcriptional activity. In this study, we enrolled a total of 1,175 controls and 1,078 patients with colorectal cancer with comprehensive clinical and survival information to assess whether these SNPs could affect the susceptibility and development of colorectal cancer. The MPO rs2333227 TT genotype significantly increased the risk of colorectal cancer and decreased the overall survival time of patients. Colorectal cancer cells with the rs2333227 TT genotype exhibited enhanced proliferation, migration, and invasion capacity in vitro and in vivo. Mechanistically, we found that MPO SNP rs2333227 C to T mutation altered the binding affinity of the transcription factors AP-2α to the rs2333227 mutation region, sequentially enhancing expression levels of MPO and activating further IL23A–MMP9 axis–mediated oncogenic signaling. Taken together, our findings indicate that MPO SNP rs2333227 serves as a marker of enhanced risk for development of colorectal cancer.

Significance: MPO polymorphisms are a guide for high risk and poor prognosis in patients colorectal cancer. Cancer Res; 78(10); 2760–9. ©2018 AACR.

Globally, colorectal cancer is the third most common form of cancer and the second leading cause of cancer-related death in the western world, with a lifetime risk in the United States of approximately 7% (1). Several factors are routinely used as prognostic markers for predictive prognosis, including depth of tumor invasion, nodal status, and distant metastasis (2). The 5-year relative survival rate ranges from 90% in colorectal cancer patients with stage І to 10% in colorectal cancer patients with stage IV (3). Other environmental factors thought to influence colorectal cancer prognosis include lifestyle (4) and the tumor immunologic microenvironment (5).

Myeloperoxidase (MPO) is a lysosomal enzyme that is present in neutrophils, monocytes, and tissue macrophages. It protects against microbial infections (6) by generating various free radicals and reactive oxidants. Published data provide evidence of high circulating levels of MPO in patients with heart failure or coronary artery diseases (7–9). The SNPs rs2333227 (G-463A) and rs2243828 (T-764C), located in the 5′ upstream region of the MPO gene, are functional SNPs, as reported in the Carotene and Retinol Efficacy Trial (CARET) cohort (10) and other studies on rectal cancer (11). It is well known that the MPO rs2333227 polymorphism is associated with increased risk for disease, including breast cancer (12), prostate cancer (13, 14), and peripheral arterial disease (15). However, an association between MPO variants and colorectal cancer risk has yet to be reported.

In the current study, we investigated the association between MPO gene polymorphisms and the risk of colorectal cancer in a Chinese population. Experiments were performed to assess the role of SNPs rs2333227 in promoting colorectal cancer.

Study subjects

This case–control study included 1,277 patients with colorectal cancer and 1,175 cancer-free controls. Newly diagnosed and histologically confirmed colorectal cancer patients were recruited between January 2007 and October 2011 the Jiangsu province (China). Controls were randomly selected from patients attending the same hospital for physical examination. The recruited patients had no history of previous cancer in other organs, or preoperative chemotherapy or radiotherapy, and the controls were genetically unrelated to the colorectal cancer patients. The pathologic stage of colorectal cancer was assessed using the Sixth Edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual. The time of death or the last follow-up (in June 2016) of each patient was confirmed by a trained clinical specialist. The median survival time (MST) was 75.0 months, and the maximum follow-up time was 112.7 months. Of the recruited patients, 1,078 patients who completed the follow-up were enrolled in the study, with a response rate of 84.4%. Written informed consent was obtained from each enrolled subject before recruitment. The study was conducted in accordance with Declaration of Helsinki and approved by the Institutional Review Board of Southeast University (Nanjing, China).

DNA extraction and genotype analysis

Because of lack of blood samples, the genomic DNA of the colorectal cancer patients was extracted from paraffin-embedded tissues using the EZNA Tissue DNA Kit (Omega Bio-Tek). For the controls, the genomic DNA was extracted from venous blood using the RelaxGene Blood DNA Kit (Tiangen Biotech). The selected polymorphisms were genotyped using TaqMan allelic discrimination assays with the Quant Studio 6 Flex system (Applied Biosystems). The loading wells without DNA were used as negative controls. Ten percent of the samples were randomly selected for confirmation, and the results were 100% consistent with previous genotyping results.

Tissue microarray construction and IHC

The colorectal cancer tissue microarrays (TMA) were created by the National Engineering Center for Biochip (Shanghai, China). Duplicate 1.0-mm diameter cores of tissue from each sample were punched from paraffin tumor blocks and adjacent colorectal tissues. The IHC assay procedure used has been described previously (16). In brief, the specimens were incubated overnight with rabbit polyclonal anti-MPO antibody (BOSTER) and anti-MMP9 antibody (Abcam) at 4°C. Each specimen was stained with 3, 3′-diaminobenzidine (DAB; Zhongshan Biotech). The evaluation of MPO and MMP9 expression was blindly conducted by two independent pathologists. The specimens were categorized using a semiquantitative immunoreactivity score (17). Briefly, the intensity of immunostaining was deemed as 0 to 3 (0, negative; 1, weak; 2, moderate; 3, strong), and the percentage of immunoreactive cells was regarded as 1 (0%–25%), 2 (26%–50%), 3 (51%–75%), and 4 (76%–100%). The immunoreactivity score (IRS) was calculated by the intensity multiply the percentage. The optimal cut-off value for the MPO and MMP9 IRS was obtained by the receiver operating characteristic (ROC) analysis. Briefly, the area under the curve (AUC) at different cut-off values of MPO IRS for survival time from 1 to 9 years was calculated, respectively. Under this condition, samples with IRS 0–4 and IRS 6–12 were defined as low and high expression for MPO, respectively, whereas samples with IRS 0–6 and IRS 8–12 were defined as low and high expression for MMP9, respectively.

Cell culture

The human colorectal cancer cell lines (SW620 and SW480) were purchased from the Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China) in 2016 and maintained in DMEM (HyClone) at 37°C under 5% CO2. The medium was supplemented with 10% FBS (Sigma), 100 U/mL penicillin, and 100 μg/mL streptomycin (HyClone). The cells were periodically tested and validated to be free of mycoplasma. These two cell lines were tested by short tandem repeat analysis and used within 6 months. The cells were grown for no more than 25 passages for any experiments.

Chromatin immunoprecipitation assays

Chromatin immunoprecipitation (ChIP) experiments were performed using the ChIP-ITTM magnetic Chromatin Immunoprecipitation Kit (cat. no. 53008, Active Motif) according to the manufacturer's instructions. Briefly, samples were fixed in formaldehyde for 10 minutes at room temperature. Then, samples were neutralized by treated with glycine stop-fix solution for 5 minutes at room temperature. The isolated nuclei and chromatin were incubated with antibody against AP-2α, or control IgG at 4°C overnight. After sequential ChIP, quantitative PCR was performed to determine the precipitated genomic DNA using the Quant Studio 6 Flex system (Applied Biosystems, Life Technologies), followed by human MPO rs2333227 (G-463A) promoter primers identifying the AP-2α–binding sequence: forward: 5′- GCCTCTAGCCACATCATCAA -3′; reverse: 5′- AATTTAGCACTACCAGCCCA-3′.

RNA extraction and real-time RT-PCR analysis

Total RNA was isolated with TRIzol reagent (Invitrogen) from frozen tissues and colorectal cancer cells. Quantitative PCR reactions were performed using SYBR Green Real-Time PCR Kits (Toyobo). Primers were 5′-GACAAATACCGCACCATC-3′ (forward) and 5′-AAGCCGTCCTCATACTCC-3′ (reverse) for MPO; 5′-ATCCGCAAAGACCTGT-3′ and 5′-GGGTGTAACGCAACTAAG-3′ for β-actin.

Western blot analysis

Total proteins were extracted from colorectal cancer tumor tissues with RIPA lysis buffer. Immunoblot assays were used to detect the MPO protein level by using primary antibodies human MPO (1:1,000 dilution, BOSTER Biological Technology) and β-actin (1:10,000 dilution). Semiquantitative analysis of MPO protein levels was performed with Image Lab 3.0 software (BIO-RAD).

CRISPR/Cas9–mediated genome engineering

The CRISPR/Cas9 expression vectors expressing Cas9, puromycin-resistant gene, and sgRNA were constructed (Supplementary Fig. S1A) based on a pX330 vector. The sequences of sgRNA targeting the rs2333227 mutation region are shown in Supplementary Fig. S1A. One day before the transfection, the cells were seeded in 6-well plates (2 × 105 cells/well). Colorectal cancer cells were cotransfected with single-stranded DNA donor construct (0.1 μg/mL) and CRISPR/Cas9 expression plasmids (1 μg/mL) using Lipofectamine 2000 according to the manufacturer's protocol (Invitrogen). After 24 hours, the medium was replaced with complete medium (2 mL) containing puromycin (0.6 μg/mL; VWR Pty Ltd.), and then every 2 days until day 14. Surviving cells were plated in a 96-well plate (1 cells/well) for single cell–derived colonies. Cells were then harvested for DNA sequencing (rs2333227, Supplementary Fig. S1B).

Analysis of colorectal cancer cell colony formation, migration, and invasion

For colony formation ability assays, 500 cells were seeded in 100-mm culture dishes. We added DMEM (HyClone) containing 10% FBS (Sigma) and placed the dishes at 37°C with 5% CO2 in an incubator, allowing the colonies to grow for 10 days. Cell colonies were then stained and imaged according to our previous study (18). Transwell (Corning) assays were performed to detect the effect of different MPO rs2333227 colorectal cancer cell genotypes on colorectal cancer cell migration and invasion. Cells (5 × 104) were added to transwell chamber with a membrane pore size of 8 μm (Corning) in triplicate. The cells (5 × 104) were added to the 8-μm pore insert in serum-free medium. The pore inserts were coated with Matrigel (Cell Biolabs). As a chemoattractant, DMEM containing 10% FBS was added to the lower chambers. After 24 hours, cells were fixed and stained with crystal violet (1%) in PBS/ethanol (85%/15%). Cell migration was determined in an analogous fashion, but in the absence of Matrigel in the transwell chambers.

In vivo nude mice flank tumor experiments

All animal studies were performed with the approval of the Committee on Animal Care and Use of Southeast University. A total of 36 nude mice (6 mice for each group) were injected subcutaneously with 5 × 106 luciferase-expressing stably transformed colorectal cancer cells (SW620 and SW480) with CC, CT, and TT genotype, respectively. Two weeks later, the nude mice were euthanized and tumor, lung, and liver tissues were removed, and luciferase activity was determined with a luminometer (Sirius, Berthold Detection Systems).

ELISA

The MPO and IL23A levels of the colorectal cancer patients’ serum and colorectal cancer cell culture supernatant were detected by ELISA (Abcam) according to the manufacturer's instructions.

Statistical analysis

The genotype frequencies of the controls were analyzed for deviations from the Hardy–Weinberg equilibrium (HWE) using the goodness-of-fit χ2 test. t test (for continuous variables) and χ2 test (for categorical variables) were used to evaluate the differences in demographic characteristics between the colorectal cancer patients and controls and the genotype results of the SNPs. The association between different genotypes and colorectal cancer risk were estimated using the logistic regression models with adjustment for age, gender, smoking, and drinking status. The Kaplan–Meier method and log-rank test were used to analyze the overall survival in different subgroups classified by demographic characteristics, genotype, and clinical information. Univariate or multivariate Cox regression analysis was used to analyze HRs and 95% confidence interval (CI), adjusted for age, sex, drinking or smoking status, tumor site (colon or rectum), grade (low or intermediate/high differentiated), and clinical stage (I to IV). Stepwise Cox regression analysis was conducted to assess predictive factors of colorectal cancer survival, with a significance level of P < 0.05 for entering and P > 0.10 for removing the variables adjusted in Cox regression analysis. The mediation model was used to explore whether the association between MPO rs2333227 C>T polymorphism and colorectal cancer patients’ survival was mediated by MPO protein level. Briefly, the first assumption of the mediation model was applied to evaluate the association between rs2333227 CT/TT genotypes and colorectal cancer patients’ survival. The second assumption of the mediation model was applied to assess the association between rs2333227 CT/TT genotypes and MPO protein levels. For the third assumption, the Cox regression analysis was performed to test the association between levels of MPO protein and colorectal cancer patients’ survival. The statistical significance criterion was set with a two-sided P < 0.05. All the statistical analyses were performed using SAS version 9.1.3 (SAS Institute).

Subject characteristics

The detailed information on the subjects has been shown previously (19). Briefly, the age and sex distribution of colorectal cancer patients and controls were similar. More smokers and drinkers were observed in colorectal cancer patients than the controls (P = 0.0116 and 0.0347 for smoking and drinking status, respectively).

Association between MPO polymorphism and risk of colorectal cancer

The genotype distribution of MPO rs2243828 and rs2333227 among the controls was consistent with HWE (P = 0.1924 for rs2243828, P = 0.5097 for rs2333227). As shown in Table 1, frequencies of MPO rs2243828 genotypes were not significantly different between the colorectal cancer patients and controls (P = 0.3981). However, the frequency distributions of rs2333227 genotypes were significantly different between colorectal cancer patients and controls (P = 0.0066). When the rs2333227 CC genotype was used as a reference, the heterozygous CT and TT genotypes imparted increased risk of colorectal cancer (adjusted OR = 1.27; 95% CI, 1.05–1.53 for the CT genotype; 2.00, 1.12–3.56 for the TT genotype), and the increased risk did not change substantially in a dominant genetic model (adjusted OR = 1.31; 95% CI, 1.09–1.58).

Table 1.

Association between the MPO polymorphisms and risk of colorectal cancer

CasesControls
SNPsGenotypen (%)n (%)PAdjusted OR (95% CI)a
rs2243828 AA 714 (66.2) 787 (67) 0.3981 1.00 (ref) 
 AG 308 (28.6) 341 (29)  0.99 (0.82–1.19) 
 GG 56 (5.2) 47 (4)  1.33 (0.89–1.99) 
 HWE 0.0034 0.1924   
 Ptrend   0.4212  
rs2333227 CC 746 (69.2) 875 (74.5) 0.0066 1.00 (ref) 
 CT 300 (27.8) 281 (23.9)  1.27 (1.05–1.53) 
 TT 32 (3) 19 (1.6)  2.00 (1.12–3.56) 
 CC 746 (69.2) 875 (74.5) 0.0054 1.00 (ref) 
 CT/TT 332 (30.8) 300 (25.5)  1.31 (1.09–1.58) 
 T allele 0.1688 0.1357 0.002  
 HWE 0.7824 0.5097   
 Ptrend   0.002  
CasesControls
SNPsGenotypen (%)n (%)PAdjusted OR (95% CI)a
rs2243828 AA 714 (66.2) 787 (67) 0.3981 1.00 (ref) 
 AG 308 (28.6) 341 (29)  0.99 (0.82–1.19) 
 GG 56 (5.2) 47 (4)  1.33 (0.89–1.99) 
 HWE 0.0034 0.1924   
 Ptrend   0.4212  
rs2333227 CC 746 (69.2) 875 (74.5) 0.0066 1.00 (ref) 
 CT 300 (27.8) 281 (23.9)  1.27 (1.05–1.53) 
 TT 32 (3) 19 (1.6)  2.00 (1.12–3.56) 
 CC 746 (69.2) 875 (74.5) 0.0054 1.00 (ref) 
 CT/TT 332 (30.8) 300 (25.5)  1.31 (1.09–1.58) 
 T allele 0.1688 0.1357 0.002  
 HWE 0.7824 0.5097   
 Ptrend   0.002  

aAdjusted for age, gender, smoking, and drinking status.

The stratification analysis revealed that rs2333227 CT/TT imparted a higher risk for colorectal cancer, especially in subgroups of age (>56 years), female, colon cancer, and tumor–node—metastasis (TNM) stages (III–IV; adjusted OR = 1.42; 95% CI, 1.09–1.85 for age > 56 years; 1.53, 1.15–2.04 for female; 1.64, 1.30–2.08 for patients with colon cancer; 1.44, 1.11–1.87 for patients with TNM stage III tumors; 2.49, 1.70–3.66 for patients with TNM stage IV tumors), compared with colorectal cancer patients with rs2333227 CC genotype (Supplementary Table S1).

Association between rs2333227 and colorectal cancer survival

As shown in Table 2 and Fig. 1A and B, Kaplan–Meier survival curves were used to assess the association between colorectal cancer survival and rs2333227 genotype. In the overall model, the MST of colorectal cancer patients with MPO rs2333227 CC genotype was 79.5 months, and 59.4 and 58.4 months for the CT and TT genotypes, respectively (log-rank P = 0.0026). Compared with colorectal cancer patients with rs2333227 CC genotype, those with rs2333227 CT/TT had a shortened survival time in the dominant model (MST = 79.5 and 59.3 months, respectively, log-rank P = 0.0006). Although previous studies noted that SNP rs2243828 was positively associated with other cancers (8, 16), we found that the SNP rs2243828 does not contribute to the clinical occurrence of colorectal cancer.

Table 2.

Association between MPO rs2333227 polymorphisms and overall survival of the colorectal cancer patients

GenotypePatients (n)Deaths (n)MSTLog-rank PAdjusted HRa (95% CI)
CC 746 427 79.5 0.0026 1.00 (ref) 
CT 300 194 59.4  1.18 (0.99–1.40) 
TT 32 23 58.4  1.28 (0.84–1.96) 
CT/TT 332 217 59.3 0.0006 1.19 (1.01–1.40) 
GenotypePatients (n)Deaths (n)MSTLog-rank PAdjusted HRa (95% CI)
CC 746 427 79.5 0.0026 1.00 (ref) 
CT 300 194 59.4  1.18 (0.99–1.40) 
TT 32 23 58.4  1.28 (0.84–1.96) 
CT/TT 332 217 59.3 0.0006 1.19 (1.01–1.40) 

aAdjusted for age, gender, smoke and drink status, location, grade, and TNM.

Figure 1.

MPO mediated the relationship between rs2333227 C>T polymorphism and colorectal cancer patients’ survival. A, Schematic of the study. B,MPO rs2333227 genotypes and colorectal cancer survival in a dominant genetic model in all patients. C, Protein expression of MPO obtained by IHC assay in human colorectal cancer and adjacent tissues TMA. D, The distribution of MPO protein expression of human colorectal cancer and adjacent tissues TMA in patients carrying three rs2333227 genotypes. P values were calculated with the Wilcoxon test. E,MPO rs2333227 genotypes and protein expression of MPO in human colorectal cancer and adjacent tissues TMA. ***, P < 0.001; NS, no significance. F, Schematic of mediation model for rs2333227 C>T polymorphism, MPO protein level, and patients with colorectal cancer survival. G,MPO rs2333227 genotypes and colorectal cancer survival in a dominant genetic model (log-rank P < 0.001). H, The relationship between MPO protein expression and patients with colorectal cancer survival (log-rank P < 0.001). Kaplan-Meier analysis was calculated with 1,005 patients with colorectal cancer.

Figure 1.

MPO mediated the relationship between rs2333227 C>T polymorphism and colorectal cancer patients’ survival. A, Schematic of the study. B,MPO rs2333227 genotypes and colorectal cancer survival in a dominant genetic model in all patients. C, Protein expression of MPO obtained by IHC assay in human colorectal cancer and adjacent tissues TMA. D, The distribution of MPO protein expression of human colorectal cancer and adjacent tissues TMA in patients carrying three rs2333227 genotypes. P values were calculated with the Wilcoxon test. E,MPO rs2333227 genotypes and protein expression of MPO in human colorectal cancer and adjacent tissues TMA. ***, P < 0.001; NS, no significance. F, Schematic of mediation model for rs2333227 C>T polymorphism, MPO protein level, and patients with colorectal cancer survival. G,MPO rs2333227 genotypes and colorectal cancer survival in a dominant genetic model (log-rank P < 0.001). H, The relationship between MPO protein expression and patients with colorectal cancer survival (log-rank P < 0.001). Kaplan-Meier analysis was calculated with 1,005 patients with colorectal cancer.

Close modal

In the stratification analysis (Supplementary Table S2), we found that the CT/TT genotypes were significantly associated with a decreased survival time in the subgroup of age >56 years, female gender, never drink, and smoke (adjusted HR = 1.33; 95% CI, 1.05–1.67; adjusted HR = 1.61; 95% CI, 1.24–2.10, adjusted HR = 1.24; 95% CI, 1.01–1.53; adjusted HR = 1.25; 95% CI, 1.01–1.54, respectively). We also found a similar result in the subgroups of intermediate/high grade and TNM stage III (adjusted HR = 1.25; 95% CI, 1.02–1.54; adjusted HR = 1.29; 95% CI, 1.01–1.65, respectively).

We then performed a Cox stepwise regression analysis to evaluate the effects of selected demographic, clinical information, and the rs2333227 variant on colorectal cancer survival. As a result, four variables (age, drinking status, TNM stage, and rs2333227 dominant model) were included in the model (P < 0.0001 for age, P < 0.0001 for drinking status, P < 0.0001 for TNM stage, and P = 0.013 for rs2333227 dominant model; Supplementary Table S3).

MPO rs2333227 genotype and MPO expression in colorectal cancer

To further determine whether the MPO rs2333227 polymorphism can affect its protein expression, TMAs from each genotype were collected for IHC analysis. IHC assay revealed that MPO protein expression levels were upregulated in the cancer tissue of colorectal cancer patients as well as patients with the MPO rs2333227 CT/TT genotype (Fig. 1C–E).

Next, we used the mediation model to explore whether MPO mediated the relationship between rs2333227 C>T polymorphism and colorectal cancer patients’ survival. We found that the levels of MPO protein were significantly upregulated in the patients carrying the CT/TT genotypes compared with those with the CC genotype (by the logistic regression analysis; adjusted OR = 2.74; 95% CI, 2.00–3.75; Fig. 1E and F), thus supporting the second assumption of the mediation model. Moreover, the first assumption of the mediation model showed the CT/TT genotypes significantly decreased the colorectal cancer patients’ survival time when compared with those with the CC genotype (adjusted HR = 1.21; 95% CI, 1.01–1.44; Fig. 1F and G). Furthermore, Cox regression analysis indicated that high MPO expressions was associated with shorter survival time (adjusted HR = 3.36; 95% CI, 2.77–4.07; Fig. 1F–H), confirming the third assumption of the mediation model. Moreover, the relationship between rs2333227 C>T polymorphism and colorectal cancer patients’ survival became insignificant once the MPO was included in the model (adjusted HR = 0.89; 95% CI, 0.74–1.06; Fig. 1F).

MPO rs2333227 T genotype recruited AP-2α to bind on the -463 region

According to the JASPAR CORE database and Gene Regulation, the MPO rs2333227 SNP site, located at the −463 promoter, is contained in the AP-2α–binding motif. Accordingly, we hypothesized that the MPO rs2333227 variant might affect the binding affinity of AP-2α to the MPO rs2333227 mutation region. Thus, the efficacy binding affinity of AP-2α to the -463 site was assessed by ChIP analysis using colorectal cancer human peripheral white blood. Results showed that the rs2333227 CT/TT genotype promoted AP-2α binding to the MPO promoter, but the CC genotype did not in colorectal cancer human peripheral white blood (Fig. 2A). Similar results were also confirmed in SW620 and SW480 cells with the indicated genotypes (Fig. 2B and C; three separate colonies with rs2333227 CT/TT genotype). To further determine whether the MPO rs2333227 polymorphism can affect MPO transcriptional activity, the MPO mRNA level of colorectal cancer patients was determined. Results showed that individuals with mutant CT/TT genotypes at the rs2333227 had significantly higher MPO levels than those with the CC genotype (Fig. 2D). The MPO protein expression levels of colorectal cancer patients with the CT/TT genotype were significantly higher than those with the CC genotype (Fig. 2E). Moreover, the serum MPO levels of colorectal cancer patients with the CT/TT genotype were statistically significantly higher than those with the CC genotype (Supplementary Table S4). Thus, we posit that higher AP-2α–binding efficiency is driving increased MPO protein expression in the rs2333227 genotypes.

Figure 2.

AP-2α promoted tumor MPO transcription by binding to the rs2333227 T genotype in colorectal cancer patients. A–C, The binding affinity between AP-2α and the indicated MPO rs2333227 genotypes were determined by ChIP assay in colorectal cancer. Human peripheral white blood (A), SW620 cells (B), and SW480 cells (C). D,MPO mRNA levels in colorectal cancer tissues expressing CC or CT/TT genotype were detected by qPCR analysis. E, MPO protein levels in colorectal cancer tissues expressing CC, CT, or TT genotype were detected by immunoblotting. ***, P < 0.001.

Figure 2.

AP-2α promoted tumor MPO transcription by binding to the rs2333227 T genotype in colorectal cancer patients. A–C, The binding affinity between AP-2α and the indicated MPO rs2333227 genotypes were determined by ChIP assay in colorectal cancer. Human peripheral white blood (A), SW620 cells (B), and SW480 cells (C). D,MPO mRNA levels in colorectal cancer tissues expressing CC or CT/TT genotype were detected by qPCR analysis. E, MPO protein levels in colorectal cancer tissues expressing CC, CT, or TT genotype were detected by immunoblotting. ***, P < 0.001.

Close modal

MPO rs2333227 TT genotype promotes MPO expression and colorectal cancer cell proliferation, invasion, and migration

The SNPs rs2333227 (G-463A) and rs2243828 (T-764C), located in the 5′ upstream region of MPO gene, are functional SNPs, as reported in various diseases (10, 11).

We hypothesized that the MPO-mutant rs2333227 genotype might alter colorectal cancer cell invasion and migration. Our results indicated that colony formation in both SW620 and SW480 cells were markedly enhanced by the rs2333227 CT/TT genotype compared with the CC genotype (Fig. 3A and B). The MPO rs2333227 TT genotype significantly promoted the migration and invasion of SW480 and SW620 colorectal cancer cell lines (Fig. 3C and D).

Figure 3.

MPO rs2333227 T genotype promotes in vitro and in vivo cell proliferation, migration, and invasion of human colorectal cancer cell lines. A and B, Representation of colony formation images and colony formation counting are shown [SW620 (A) and SW480 (B)]. C and D, Transwell assay performed to determine the migration and invasion of cell lines SW620 (C) and SW480 (D). E and F, Flank tumor, liver, and lung metastatic burden in nude mice (n = 6 for each group) injected with colorectal cancer cell lines with rs2333227 CC, CT, and TT genotypes, measured by luciferase activity. *, P < 0.05; **, P < 0.01; ***, P < 0.001 .

Figure 3.

MPO rs2333227 T genotype promotes in vitro and in vivo cell proliferation, migration, and invasion of human colorectal cancer cell lines. A and B, Representation of colony formation images and colony formation counting are shown [SW620 (A) and SW480 (B)]. C and D, Transwell assay performed to determine the migration and invasion of cell lines SW620 (C) and SW480 (D). E and F, Flank tumor, liver, and lung metastatic burden in nude mice (n = 6 for each group) injected with colorectal cancer cell lines with rs2333227 CC, CT, and TT genotypes, measured by luciferase activity. *, P < 0.05; **, P < 0.01; ***, P < 0.001 .

Close modal

Next, we injected the colorectal cancer cell lines (SW480 and SW620 with rs2333227 CC, CT, and TT genotype) into the dorsal flanks of nude mice. The results showed that the growth rate of xenograft colorectal cancer cells was significantly higher with the rs2333227 CT/TT genotype than CC genotype (Fig. 3E and F).

IL23A and MMP9 mediated the relationship between MPO and development of colorectal cancer

To further explore the potential mechanism underlying the biological effects of MPO in colorectal cancer development, we constructed a protein interaction network for genes regulated by MPO using STRING version 10.0 (search tool for the retrieval of interacting genes; Fig. 4A). The correlations between mRNA levels of MPO and key components in the predicted network were validated in colorectal cancer tissues (Fig. 4B; Supplementary Fig. S2A–S2I). Among all the components, IL23A transcriptional levels exhibited the strongest correlation with MPO, especially in patients carrying with CT/TT genotype (r = 0.8560 for CT/TT and r = 0.5362 for CC, respectively; Fig. 4B). Given that IL23 upregulates the matrix metalloproteinase 9 (MMP9) expression (20), we next evaluated the mRNA and protein levels of MMP9 in colorectal cancer patients. We found a significantly positive correlation between MPO and MMP9 mRNA expression levels (r = 0.5163, P <0.001, Fig. 4C). Moreover, the correlation in patients with CT/TT genotypes was stronger than with the CC genotype (r = 0.7689 and r = 0.1563, respectively), corroborating the CT/TT genotypes might be a higher risk factors for colorectal cancer metastasis (Fig. 4C). Similar to our finding in MPO, MMP9 were upregulated in colorectal cancer patients’ cancer tissues and patients carrying the MPO rs2333227 CT/TT genotypes (Fig. 4D–F.). Furthermore, colorectal cancer cells carrying the CT/TT genotypes also showed increased mRNA levels of MPO, IL23A, and MMP9 compared with the CC genotype (Supplementary Fig. S3A–S3F). The trend for increased MPO and IL23 protein levels was also consistent with the mRNA levels in the cell culture supernatants assayed by ELISA (Supplementary Fig. S3G–S3J.). Kaplan-Meier analysis showed that both high expressions of those two proteins showed worsening outcome for survival compared with the other groups (one high or both low; Fig. 4G.). Furthermore, multivariate Cox regression analysis indicated that high MPO and MMP9 expressions were independent prognostic markers for colorectal cancer (Table 3). We then performed a time-dependent ROC analysis to evaluate the prognostic efficacy of MPO and MMP9 expressions on colorectal cancer patients. We found that the combination of the clinical variables, MPO, and MMP9 risk score contributed more than either one of these alone (Fig. 4H).

Figure 4.

IL23A and MMP9 mediated the relationship between MPO and development of colorectal cancer. A, Protein interaction network for genes regulated by MPO analyzed by STRING (version 10.0). B, Correlation analysis between mRNA expression levels of MPO and IL23A by qRT-PCR. C, Correlation analysis between mRNA expression levels of MPO and MMP9 by qRT-PCR. D, Protein expression of MMP9 obtained by IHC assay in human colorectal cancer and adjacent tissues TMA. E, MPO protein expressions of human colorectal cancer and adjacent tissues TMA in patients carrying three rs2333227 genotypes were determined. P values were calculated with the Wilcoxon test. F,MPO rs2333227 genotypes and protein expression of MMP9 in human colorectal cancer and adjacent tissues TMA. ***,P < 0.001; NS, no significance. G, The relationship between MPO and MMP9 protein expression and colorectal cancer patients’ survival (log-rank P < 0.001). Kaplan–Meier analysis was calculated in a sample of 915 patients with colorectal cancer. H, Time-dependent ROC analysis for the clinical risk score (TNM, grade), MPO, MMP9, or the combined MPO, MMP9, and the clinical risk score (TNM, grade).

Figure 4.

IL23A and MMP9 mediated the relationship between MPO and development of colorectal cancer. A, Protein interaction network for genes regulated by MPO analyzed by STRING (version 10.0). B, Correlation analysis between mRNA expression levels of MPO and IL23A by qRT-PCR. C, Correlation analysis between mRNA expression levels of MPO and MMP9 by qRT-PCR. D, Protein expression of MMP9 obtained by IHC assay in human colorectal cancer and adjacent tissues TMA. E, MPO protein expressions of human colorectal cancer and adjacent tissues TMA in patients carrying three rs2333227 genotypes were determined. P values were calculated with the Wilcoxon test. F,MPO rs2333227 genotypes and protein expression of MMP9 in human colorectal cancer and adjacent tissues TMA. ***,P < 0.001; NS, no significance. G, The relationship between MPO and MMP9 protein expression and colorectal cancer patients’ survival (log-rank P < 0.001). Kaplan–Meier analysis was calculated in a sample of 915 patients with colorectal cancer. H, Time-dependent ROC analysis for the clinical risk score (TNM, grade), MPO, MMP9, or the combined MPO, MMP9, and the clinical risk score (TNM, grade).

Close modal
Table 3.

Multivariate Cox regression analysis of MPO and MMP9 expression and selected clinical pathologic variables in colorectal cancer survival

MPOaMMP9aCombineda,bCombineda,c
VariablesAdjusted HR (95% CI)PAdjusted HR (95% CI)PAdjusted HR (95% CI)PAdjusted HR(95%CI)P
Age (>56 vs. ≤56) 1.41 (1.18–1.67) 0.0001 1.47 (1.24–1.75) <0.0001 1.39 (1.17–1.65) 0.0002 1.52 (1.24–1.85) <0.0001 
Gender (females vs. males) 0.86 (0.71–1.03) 0.0975 0.82 (0.68–0.98) 0.0284 0.80 (0.67–0.96) 0.019 0.89 (0.72–1.10) 0.2861 
Smoke (ever vs. never) 1.04 (0.83–1.30) 0.7538 0.94 (0.76–1.18) 0.5959 0.97 (0.78–1.21) 0.7728 1.11 (0.85–1.45) 0.4503 
Drink (ever vs. never) 1.16 (0.91–1.46) 0.2312 1.19 (0.94–1.50) 0.1410 1.17 (0.93–1.47) 0.1819 1.05 (0.79–1.39) 0.7327 
Location (rectum vs. colon) 0.89 (0.75–1.06) 0.2060 0.90 (0.76–1.07) 0.2307 0.86 (0.72–1.03) 0.0935 0.95 (0.78–1.17) 0.6402 
Grade (intermediate/high vs. low) 0.95 (0.80–1.14) 0.5813 0.93 (0.78–1.12) 0.4515 0.98 (0.82–1.17) 0.832 0.94 (0.76–1.16) 0.5652 
TNM(3/4 vs. 1/2) 5.17 (4.24–6.29) <0.0001 6.10 (5.04–7.39) <0.0001 5.63 (4.64–6.84) <0.0001 5.71 (4.55–7.18) <0.0001 
Protein (high vs. low) 3.48 (2.89–4.21) <0.0001 2.23 (1.88–2.64) <0.0001 2.91 (2.40–3.53) <0.0001 4.04 (3.25–5.03) <0.0001 
MPOaMMP9aCombineda,bCombineda,c
VariablesAdjusted HR (95% CI)PAdjusted HR (95% CI)PAdjusted HR (95% CI)PAdjusted HR(95%CI)P
Age (>56 vs. ≤56) 1.41 (1.18–1.67) 0.0001 1.47 (1.24–1.75) <0.0001 1.39 (1.17–1.65) 0.0002 1.52 (1.24–1.85) <0.0001 
Gender (females vs. males) 0.86 (0.71–1.03) 0.0975 0.82 (0.68–0.98) 0.0284 0.80 (0.67–0.96) 0.019 0.89 (0.72–1.10) 0.2861 
Smoke (ever vs. never) 1.04 (0.83–1.30) 0.7538 0.94 (0.76–1.18) 0.5959 0.97 (0.78–1.21) 0.7728 1.11 (0.85–1.45) 0.4503 
Drink (ever vs. never) 1.16 (0.91–1.46) 0.2312 1.19 (0.94–1.50) 0.1410 1.17 (0.93–1.47) 0.1819 1.05 (0.79–1.39) 0.7327 
Location (rectum vs. colon) 0.89 (0.75–1.06) 0.2060 0.90 (0.76–1.07) 0.2307 0.86 (0.72–1.03) 0.0935 0.95 (0.78–1.17) 0.6402 
Grade (intermediate/high vs. low) 0.95 (0.80–1.14) 0.5813 0.93 (0.78–1.12) 0.4515 0.98 (0.82–1.17) 0.832 0.94 (0.76–1.16) 0.5652 
TNM(3/4 vs. 1/2) 5.17 (4.24–6.29) <0.0001 6.10 (5.04–7.39) <0.0001 5.63 (4.64–6.84) <0.0001 5.71 (4.55–7.18) <0.0001 
Protein (high vs. low) 3.48 (2.89–4.21) <0.0001 2.23 (1.88–2.64) <0.0001 2.91 (2.40–3.53) <0.0001 4.04 (3.25–5.03) <0.0001 

aMultivariate Cox regression analyses were performed with 915 subjects because of loss of sample during MPO and MMP9 antigen retrieval.

bMultivariate Cox regression analyses were calculated with following proteins expression status (one high vs. both low).

cMultivariate Cox regression analyses were calculated with following proteins expression status (both high vs. both low).

In the current study, we explored the role of promoter MPO SNPs in susceptibility and development of colorectal cancer. By conducting a hospital-based case–control study, we found that subjects carrying rs2333227 CT/TT genotypes had a significantly higher risk for colorectal cancer and poorer survival outcomes when compared with the CC genotype. With in vivo and in vitro assays, we demonstrated that transcription factor AP-2α presented stronger binding affinity with rs2333227 CT/TT genotype than CC genotype, thus upregulating MPO expressions, and ultimately promoting colorectal cancer proliferation and metastasis. Therefore, our results verified the critical role of MPO functional variant (rs2333227) in colorectal cancer.

MPO catalyzes oxidants generation, thus contributing to the risk of cancer (21). It has been reported that MPO rs2333227 T allele alters the susceptibility of breast (12) and prostate cancer (14). However, few studies were paid attention to the role of MPO rs2333227 in colorectal cancer. In our cohort, we demonstrated that rs2333227 TT genotype was a risk factor for colorectal cancer and associated with worse clinical features, suggesting that different clinical features of colorectal cancer might be regulated by different molecular mechanisms. Interestingly, we found that age, drinking status, TNM stage, and rs2333227 dominant model were independent factors affecting colorectal cancer in our stepwise Cox regression analysis. Taken together, our results have verified the oncogenic role of rs2333227 TT genotype in colorectal cancer.

Transcription factor AP-2α promotes the expressions of target genes (such as DEK; ref. 22) through directly binding to the promoter region. By conducting the ChIP assay, we demonstrated that AP-2α preferred to binding with MPO rs2333227 CT/TT genotype. As colorectal cancer tissues, serum and cells carrying CT/TT genotype exhibited much higher MPO protein levels, we speculated that AP-2α was responsible for rs2333227 CT/TT genotype–induced higher transcriptional activity. As few reports were focused on AP-2α/rs2333227 T allele axis, further researches are still needed to confirm our findings.

It has been reported that MPO plays an important role in cellular processes, such as DNA damage, oxidative stress (23), and innate immune response (24). Its major function is to mediate hypochlorous acid (HOCl) generation (25). However, HOCl and other oxidants derived by MPO are considered important mediators of oxidative damage to cellular biomolecules (26) in that they may damage the structure of host tissues (27). Moreover, HOCl can decrease the adhesiveness of extracellular matrix proteins and affect endothelial function (11), destabilizing the tissue environment surrounding endothelial cells (28). This suggests that overexpressed MPO in human tumors may be associated with enhanced cell invasion and migration. Moreover, we found that MPO protein expression levels were upregulated in the cancer tissue of colorectal cancer patients, which corroborates that MPO could be an oncogene in colorectal cancer.

To further ascertain the downregulators of MPO that drive aggressive phenotype, protein interaction network were constructed (29, 30). Among the 10 candidates, IL23A showed the strongest correlation with MPO levels, especially in patients carrying TT genotype. Interestingly, Lan and colleagues showed that IL23A was upregulated in the colorectal cancer tissues (31). Previous studies have showed that IL23A promotes tumor incidence and growth by targeting MMP9 (20). In our study, we again demonstrated that MMP9 was not only the downregulator of IL23A, but also synergistically predicted colorectal cancer patients’ prognosis with MPO. Taken together, our results indicated that MPO/IL23A/MMP9 axis might promote colorectal cancer progression and metastasis.

In summary, our results revealed that functional rs2333227 TT genotype upregulated MPO expressions in colorectal cancer, thus activating IL23A/MMP9 signaling–medicated tumor growth and development. However, further studies are still required to confirm our findings.

No potential conflicts of interest were disclosed.

Conception and design: M. Aschner, R. Chen

Development of methodology: R. Chen

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Wang, H. Sun, R. Lu, X. Li

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Q. Meng, S. Wu, J. Xu, N. Gao, M. Aschner

Writing, review, and/or revision of the manuscript: Q. Meng, S. Wu, X. Li, M. Aschner, R. Chen

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): H. Yang, B. Tang, R. Chen

Study supervision: R. Chen

This work was financially supported by National Natural Science Foundation of China (81472938 and 91643109 to R. Chen; 81703261 to J. Xu), the National Key Research and Development Program of China (2017YFC0211600 and 2017YFC0211603 to R. Chen), the Natural Science Foundation of Jiangsu Province (BK20151418 to X. Li; BK20171060 to J. Xu), the Fund of the Distinguished Professor of Jiangsu Province to R. Chen, the Fund of the Distinguished Talents of Jiangsu Province to R. Chen (BK20150021), the Six talent peaks project in Jiangsu Province to R. Chen(2016-WSN-002), the Natural Science Foundation of Jiangsu Higher Education Institutions to J. Xu (17KJB330002), the Fund of the Post-graduate Innovative Talents to Q. Meng (KYZZ16_0137), the Fundamental Research Funds for the Central Universities to R. Chen and X. Li, and the National Institute of Environmental health Sciences (NIEHS; R01 ES10563, R01 ES07331, and R01 ES020852 to M. Aschner).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Quan
B
,
Qi
X
,
Yu
Z
,
Jiang
Y
,
Liao
M
,
Wang
G
, et al
Pathway analysis of genome-wide association study and transcriptome data highlights new biological pathways in colorectal cancer
.
Mol Genet Genomics
2015
;
290
:
603
10
.
2.
Walther
A
,
Johnstone
E
,
Swanton
C
,
Midgley
R
,
Tomlinson
I
,
Kerr
D
. 
Genetic prognostic and predictive markers in colorectal cancer
.
Nat Rev Cancer
2009
;
9
:
489
99
.
3.
Brenner
H
,
Kloor
M
,
Pox
CP
. 
Colorectal cancer
.
Lancet
2014
;
383
:
1490
502
.
4.
Haydon
AM
,
Macinnis
RJ
,
English
DR
,
Giles
GG
. 
Effect of physical activity and body size on survival after diagnosis with colorectal cancer
.
Gut
2006
;
55
:
62
7
.
5.
Galon
J
,
Costes
A
,
Sanchez-Cabo
F
,
Kirilovsky
A
,
Mlecnik
B
,
Lagorce-Pages
C
, et al
Type, density, and location of immune cells within human colorectal tumors predict clinical outcome
.
Science
2006
;
313
:
1960
4
.
6.
Arnhold
J
,
Flemmig
J
. 
Human myeloperoxidase in innate and acquired immunity
.
Arch Biochem Biophys
2010
;
500
:
92
106
.
7.
Baldus
S
,
Heeschen
C
,
Meinertz
T
,
Zeiher
AM
,
Eiserich
JP
,
Munzel
T
, et al
Myeloperoxidase serum levels predict risk in patients with acute coronary syndromes
.
Circulation
2003
;
108
:
1440
5
.
8.
Karakas
M
,
Koenig
W
. 
Myeloperoxidase production by macrophage and risk of atherosclerosis
.
Curr Atheroscler Rep
2012
;
14
:
277
83
.
9.
Reichlin
T
,
Socrates
T
,
Egli
P
,
Potocki
M
,
Breidthardt
T
,
Arenja
N
, et al
Use of myeloperoxidase for risk stratification in acute heart failure
.
Clin Chem
2010
;
56
:
944
51
.
10.
Choi
JY
,
Neuhouser
ML
,
Barnett
MJ
,
Hong
CC
,
Kristal
AR
,
Thornquist
MD
, et al
Iron intake, oxidative stress-related genes (MnSOD and MPO) and prostate cancer risk in CARET cohort
.
Carcinogenesis
2008
;
29
:
964
70
.
11.
Slattery
ML
,
Lundgreen
A
,
Welbourn
B
,
Wolff
RK
,
Corcoran
C
. 
Oxidative balance and colon and rectal cancer: interaction of lifestyle factors and genes
.
Mutat Res
2012
;
734
:
30
40
.
12.
He
C
,
Tamimi
RM
,
Hankinson
SE
,
Hunter
DJ
,
Han
J
. 
A prospective study of genetic polymorphism in MPO, antioxidant status, and breast cancer risk
.
Breast Cancer Res Treat
2009
;
113
:
585
94
.
13.
Cheng
TY
,
King
IB
,
Barnett
MJ
,
Ambrosone
CB
,
Thornquist
MD
,
Goodman
GE
, et al
Serum phospholipid fatty acids, genetic variation in myeloperoxidase, and prostate cancer risk in heavy smokers: a gene-nutrient interaction in the carotene and retinol efficacy trial
.
Am J Epidemiol
2013
;
177
:
1106
17
.
14.
Cheng
TY
,
Barnett
MJ
,
Kristal
AR
,
Ambrosone
CB
,
King
IB
,
Thornquist
MD
, et al
Genetic variation in myeloperoxidase modifies the association of serum alpha-tocopherol with aggressive prostate cancer among current smokers
.
J Nutr
2011
;
141
:
1731
7
.
15.
Haslacher
H
,
Perkmann
T
,
Gruenewald
J
,
Exner
M
,
Endler
G
,
Scheichenberger
V
, et al
Plasma myeloperoxidase level and peripheral arterial disease
.
Eur J Clin Invest
2012
;
42
:
463
9
.
16.
Wu
DM
,
Zhu
HX
,
Zhao
QH
,
Zhang
ZZ
,
Wang
SZ
,
Wang
ML
, et al
Genetic variations in the SMAD4 gene and gastric cancer susceptibility
.
World J Gastroenterol
2010
;
16
:
5635
41
.
17.
Wang
S
,
Wu
X
,
Chen
Y
,
Zhang
J
,
Ding
J
,
Zhou
Y
, et al
Prognostic and predictive role of JWA and XRCC1 expressions in gastric cancer
.
Clin Cancer Res
2012
;
18
:
2987
96
.
18.
Li
X
,
Lv
Y
,
Gao
N
,
Sun
H
,
Lu
R
,
Yang
H
, et al
microRNA-802/Rnd3 pathway imposes on carcinogenesis and metastasis of fine particulate matter exposure.
Oncotarget
2016
;
7
:
35026
43
.
19.
Wu
S
,
Meng
Q
,
Zhang
C
,
Sun
H
,
Lu
R
,
Gao
N
, et al
DR4 mediates the progression, invasion, metastasis and survival of colorectal cancer through the Sp1/NF1 switch axis on genomic locus
.
Int J Cancer
2018
Feb 15.
doi: 10.1002/ijc.31318.
20.
Langowski
JL
,
Zhang
X
,
Wu
L
,
Mattson
JD
,
Chen
T
,
Smith
K
, et al
IL-23 promotes tumour incidence and growth.
Nature
2006
;
442
:
461
5
.
21.
Winterbourn
CC
,
Kettle
AJ
. 
Biomarkers of myeloperoxidase-derived hypochlorous acid
.
Free Radic Biol Med
2000
;
29
:
403
9
.
22.
Qiao
MX
,
Li
C
,
Zhang
AQ
,
Hou
LL
,
Yang
J
,
Hu
HG
. 
Regulation of DEK expression by AP-2alpha and methylation level of DEK promoter in hepatocellular carcinoma
.
Oncol Rep
2016
;
36
:
2382
90
.
23.
Klebanoff
SJ
. 
Myeloperoxidase: friend and foe.
J Leukoc Biol
2005
;
77
:
598
625
.
24.
Mayadas
TN
,
Cullere
X
,
Lowell
CA
. 
The multifaceted functions of neutrophils
.
Annu Rev Pathol
2014
;
9
:
181
218
.
25.
Hickey
MJ
. 
MPO and neutrophils: a magnetic attraction
.
Blood
2011
;
117
:
1103
4
.
26.
Himmelfarb
J
,
McMenamin
ME
,
Loseto
G
,
Heinecke
JW
. 
Myeloperoxidase-catalyzed 3-chlorotyrosine formation in dialysis patients
.
Free Radic Biol Med
2001
;
31
:
1163
9
.
27.
Hansson
M
,
Olsson
I
,
Nauseef
WM
. 
Biosynthesis, processing, and sorting of human myeloperoxidase
.
Arch Biochem Biophys
2006
;
445
:
214
24
.
28.
Fu
X
,
Kassim
SY
,
Parks
WC
,
Heinecke
JW
. 
Hypochlorous acid generated by myeloperoxidase modifies adjacent tryptophan and glycine residues in the catalytic domain of matrix metalloproteinase-7 (matrilysin): an oxidative mechanism for restraining proteolytic activity during inflammation
.
J Biol Chem
2003
;
278
:
28403
9
.
29.
Szklarczyk
D
,
Morris
JH
,
Cook
H
,
Kuhn
M
,
Wyder
S
,
Simonovic
M
, et al
The STRING database in 2017: quality-controlled protein-protein association networks, made broadly accessible.
Nucleic Acids Res
2017
;
45
:
D362
D8
.
30.
Huang
DS
,
Zhang
L
,
Han
K
,
Deng
S
,
Yang
K
,
Zhang
H
. 
Prediction of protein-protein interactions based on protein-protein correlation using least squares regression
.
Curr Protein Peptide Sci
2014
;
15
:
553
60
.
31.
Lan
F
,
Zhang
L
,
Wu
J
,
Zhang
J
,
Zhang
S
,
Li
K
, et al
IL-23/IL-23R: potential mediator of intestinal tumor progression from adenomatous polyps to colorectal carcinoma.
Int J Colorectal Dis
2011
;
26
:
1511
8
.