Abstract
Background: Overexpression of the receptor for advanced glycation end-product (RAGE) has been associated with chronic inflammation, which in turn has been associated with increased colorectal cancer risk. Soluble RAGE (sRAGE) competes with RAGE to bind its ligands, thus potentially preventing RAGE-induced inflammation.
Methods: To investigate whether sRAGE and related genetic variants are associated with colorectal cancer risk, we conducted a nested case–control study in the European Prospective Investigation into Cancer and Nutrition (EPIC). Plasma sRAGE concentrations were measured by ELISA in 1,361 colorectal cancer matched case–control sets. Twenty-four SNPs encoded in the genes associated with sRAGE concentrations were available for 1,985 colorectal cancer cases and 2,220 controls. Multivariable adjusted ORs and 95% confidence intervals (CIs) were computed using conditional and unconditional logistic regression for colorectal cancer risk and circulating sRAGE and SNPs, respectively.
Results: Higher sRAGE concentrations were inversely associated with colorectal cancer (ORQ5vs.Q1, 0.77; 95% CI, 0.59–1.00). Sex-specific analyses revealed that the observed inverse risk association was restricted to men (ORQ5vs.Q1, 0.63; 95% CI, 0.42–0.94), whereas no association was observed in women (ORQ5vs.Q1, 1.00; 95% CI, 0.68–1.48; Pheterogeneity for sex = 0.006). Participants carrying minor allele of rs653765 (promoter region of ADAM10) had lower colorectal cancer risk (C vs. T, OR, 0.90; 95% CI, 0.82–0.99).
Conclusions: Prediagnostic sRAGE concentrations were inversely associated with colorectal cancer risk in men, but not in women. An SNP located within ADAM10 gene, pertaining to RAGE shedding, was associated with colorectal cancer risk.
Impact: Further studies are needed to confirm our observed sex difference in the association and better explore the potential involvement of genetic variants of sRAGE in colorectal cancer development.
Introduction
Advanced glycation end-products (AGE) are a heterogeneous group of molecules formed by nonenzymatic reactions between reducing sugars and proteins, lipids, or nucleic acids (1). AGEs are produced endogenously, but diet and lifestyle are likely the largest contributors to the overall AGEs pool particularly from high-temperature processed food products, which contain high amounts of AGEs and/or their precursors (2–4). Glycated proteins tend to become dysfunctional and agglutinate with other reacting molecules to create cross-links and aggregates which can accumulate within diverse tissues in the body (5). The accumulation of AGEs throughout the life course is thought to contribute to intracellular signaling alterations, chronic low-level inflammation, and a decrease in tissue functionality (6).
AGEs are recognized by a multi-ligand cell surface protein receptor, known as the receptor for AGE (RAGE). RAGE consists of an extracellular N-terminal, a transmembrane helix, and an intracellular C-terminal tail (7). RAGE is expressed at low levels in most tissue types, except the lungs in which the expression is generally high (8). Overexpression of RAGE and its high activity have been demonstrated in various cancers including in the colon, breast, brain, prostate, and in the ovaries (9). Binding of AGEs to their receptor triggers a signaling cascade leading to intracellular inflammation with activation of NF-κB, increased secretion of cytokines and chemokines, and elevated production of reactive oxygen and nitrogen species (10).
Soluble RAGE (sRAGE) is a free circulating isoform of RAGE that also binds AGEs and acts as a decoy for RAGE. In contrast to RAGE, binding of AGEs to sRAGE does not induce inflammation and oxidative stress (8). Although the concentration of sRAGE is likely insufficient to bind all circulating AGEs (11), higher sRAGE levels had been associated with low inflammation and lower risk of several chronic diseases, including cancers (12). The variability in sRAGE concentrations is considerably affected by a combination of genetic and environmental factors (13). sRAGE levels have been reported to be elevated in women versus men, younger versus older individuals, and individuals with normal weight versus individuals with overweight and obesity (14–17). Furthermore, genetic determinants of sRAGE expression have also been identified and include SNPs located within advanced glycosylation end-product specific receptor (AGER), a disintegrin and metalloproteinase domain 10 (ADAM10), glyoxalase I (GLO1), and ring finger protein 5 (RNF5) genes (17–21).
We hypothesized that higher circulating sRAGE levels are inversely associated with colorectal cancer development. Previously, only two prospective studies have investigated the association, and showed an inverse association of high sRAGE concentrations with colorectal cancer risk among Finnish male smokers (22) and women with overweight and obesity (23). However, there are sparse data from other prospective studies, and there is a need to carefully investigate possible differences in the association by sex or lifestyle factors. To address these gaps, we studied the association between prediagnostic levels of circulating sRAGE and risk of colorectal cancer in a large, multinational European prospective cohort. We also investigated whether SNPs, reported to be related to sRAGE levels or RAGE function, are associated with colorectal cancer risk.
Materials and Methods
Study population and data collection
We used a case–control design nested within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort. EPIC is an ongoing multicenter prospective cohort with 521,324 participants (70% women) recruited from 23 study centers located in 10 European countries (Denmark, France, Germany, Greece, Italy, the Netherlands, Norway, Spain, Sweden, and the United Kingdom). The rationale and methods of the EPIC study, including information on the recruitment of the participants as well as data collection, have been described previously (24). Participants provided written informed consent before joining the EPIC study. Participant's health history, anthropometry, and socio-demographic and standardized lifestyle variables, including education, smoking, and physical activity, were collected by questionnaire at baseline, prior to disease onset or diagnosis. Physical activity was based on the Cambridge physical activity index: inactive (sedentary job and no recreational activity), moderately inactive (sedentary job with <0.5 hours recreational activity per day/or standing job with no recreational activity), moderately active (sedentary job with 0.5–1 hour recreational activity per day/or standing job with 0.5 hours recreational activity per day/or physical job with no recreational activity), or active (sedentary job with >1 hour recreational activity per day/or standing job with >0.5 hours recreational activity per day/or physical job with at least some recreational activity/or heavy manual job; ref. 25). Dietary intake was assessed at recruitment by validated center-specific questionnaires. At each of the study centers, blood samples were drawn at recruitment (∼80% of participants provided blood samples) and stored in liquid nitrogen (−196°C, liquid nitrogen) at the International Agency for Research on Cancer (Lyon, France) biobank, or in local biobanks (at −150°C in nitrogen vapor in Denmark and −80°C freezers at Malmö and Umeå centers in Sweden; ref. 24).
Follow-up for cancer incidence and vital status
Vital status follow-up (98.4% complete) was collected by record linkage with regional and/or national mortality registries in all countries, except Germany and Greece, and the Italian center of Naples, where data are collected actively. Incident cancer cases were identified through record linkage with regional cancer registries or using a combination of methods, including health insurance records, cancer and pathology registries, and active follow-up through participants and their relatives. Colorectal cancer cases were eligible if they were first incident and histologically confirmed. Cases were defined using the International Classification of Diseases for Oncology. Colon cancers were defined as tumors that occurred in the cecum, appendix, ascending colon, hepatic flexure, transverse colon, splenic flexure, descending and sigmoid colon (C18.0–C18.7), and overlapping and or unspecified origin tumors (C18.8 and C18.9). Rectal cancers were defined as tumors that occurred at the recto-sigmoid junction (C19) or rectum (C20). Cancers of the anal canal were excluded.
Case–control design
From baseline onwards, 1,413 first incident colorectal cancer cases with available blood samples were identified (until June 2003 as endpoint) among all the 2,476 colorectal cancer cases ascertained (Fig. 1). For each identified case, one control was matched by incidence density sampling from all cohort members alive and cancer free at the time of diagnosis of the index case. Cases and controls were matched by age (±1 year), sex, center, and blood collection details, including time (±3 hours) and fasting prevenepuncture (<3, 3–6, and >6 hours); and additionally among women only, by menopausal status (pre-, peri-, and postmenopausal) and hormone replacement therapy use at the time of blood collection (yes/no). After exclusion of participants with incomplete matched case sets (n = 16), those with extreme sRAGE levels (n = 3 controls and one case with sRAGE concentrations unusually high, i.e., >mean + 4 SD), and 32 cases and matched controls from Greece due to unforeseen data restriction issues, 1,361 cases and 1,361 matched controls were included in the sRAGE analysis. Among EPIC participants, 4,487 participants (until December 2012 as endpoint, 2,148 colorectal cancer cases and 2,339 matched controls) have been genotyped previously. After exclusion of 100 colorectal cancer cases and 100 matched controls from Greece and 82 participants with missing lifestyle variable, 1,985 colorectal cancer cases and 2,220 matched controls were included in the genetic analysis. Among the participants who have been genotyped, 972 colorectal cancer cases and 767 noncases overlapped with case–control sets in whom sRAGE measurements were conducted.
sRAGE and genetic data available within EPIC. Two endpoints were used for our data, the first ended in June 2003 and included 1,361 colorectal cancer cases and 1,361 matched controls for the analysis of sRAGE concentrations. December 2012 was considered for the second endpoint, with 1,985 samples of colorectal cancer cases and 2,220 controls analyzed for genetic data. The overlapping between the two samples was used for sensitivity analysis.
Laboratory analyses
Circulating sRAGE concentrations were measured in citrated plasma samples by ELISA (Quantikine, R&D Systems), following the manufacturer's instructions. Previous studies have reported that sRAGE is stable in plasma over a long period of time (26). Analyses were run with case–control sets randomized across batches (n = 40 batches, with an average of 35 case–control pairs analyzed per batch). Intra- and interbatch coefficients of variation (CV) were assessed by measuring three different samples used as quality controls in duplicate in each. Mean intra- and interbatch CVs were 1.25% and 6.0%, respectively. C-reactive protein (CRP) concentrations were determined using a high-sensitivity assay (Beckman-Coulter).
DNA genotyping and genetic variants selection
DNA was extracted from buffy coats from citrated blood samples at the Center for Inherited Disease Research (Johns Hopkins University, Baltimore, MD) using the HumanOmniExpressExome-8v1–2 array as described elsewhere (27). All SNPs met criteria for quality control for genotyping call rate (above 95%). Candidate SNPs selected for our study were those previously associated with sRAGE levels. Most of these SNPs appear to be located within the AGER gene, with rs2070600 being the most important and explaining 22% of the variability in sRAGE concentrations in Caucasians (17). In addition to AGER, three additional genes that contain SNPs associated with sRAGE were identified: RNF5, a neighboring gene which encodes for RAGE (28), ADAM10, which encodes for metalloproteinases involved in the shedding of RAGE ectodomain to form sRAGE (29), and GLO1, which encodes for glyoxalase enzyme responsible to metabolize methylglyoxal and prevent aberrant AGEs formation (30). The main SNPs were from AGER (rs2070600, rs1800625, rs1800624, rs184003, and rs2854050), ADAM10 (rs653765), and RNF5 (rs9469089; refs. 17–21, 31–38). We additionally considered less-studied SNPs located within AGER (rs1035798, rs1800684, rs3131300, rs3134940, rs2269422, rs2853807, rs9391855, and rs17846798), ADAM10 (rs514049), RNF5 (rs57409105, rs41268928, and rs17493811), and GLO1 (rs4746, rs1130534, rs1049346, rs6932648, and rs10484854). The choice of this supplementary group of SNPs was based on the potential influence and interactions they may have in modulating sRAGE levels directly, or through AGEs (13, 17, 21, 31, 39–41).
Genotype distributions were in Hardy–Weinberg equilibrium (cutoff P = 1 × 10−3) for all the SNPs considered, with the exception of rs6932648, which was consequently excluded from the analysis. The selected SNPs and their characteristics are detailed in Supplementary Table S1. To select the independent variants, linkage disequilibrium (LD) pruning (LD ≤ 1%) was performed using NCI LDlink tools (https://ldlink.nci.nih.gov). We found the following independent variants (highly correlated variants are in brackets): rs2070600 (rs41268928, rs9391855, and rs2854050), rs1800625 (rs3131300 and rs3134940), rs1800624 (rs17846798), rs4746 (rs1130534 and rs10484854), rs17846798 (rs57409105), rs9469089, rs1800684, rs2269422, rs2853807, rs1049346, rs17493811, and rs653765 (rs514049). A flowchart outlining the selection of the independent SNPs is detailed in Supplementary Fig. S1.
Among the 767 control subjects who had both sRAGE and genetic data available, we assessed the association between the independent genetic variants and log-transformed sRAGE levels using linear regression models (Supplementary Table S2). The SNPs in the following genes were significantly associated with sRAGE levels: AGER (rs2070600 and rs1800625), RNF5 (rs9469089), and GLO1 (rs4746). Although rs653765 (ADAM10) was not associated with sRAGE levels, we decided to conserve it in our analysis for two main reasons: first, as a major variant of metalloproteinases, which are involved in the shedding of the ectodomain of RAGE to produce sRAGE; and second, this variant was previously associated with sRAGE levels in other populations (21). Overall, five SNPs (rs2070600, rs1800625, rs9469089, rs4746, and rs653765) were examined for the association with colorectal cancer risk.
Statistical analysis
Case–control differences in baseline characteristics were evaluated using Student paired t test and Wilcoxon signed-rank test for continuous variables and Kruskal–Wallis test for categorical variables. Spearman rank correlation was used to correlate sRAGE levels to anthropometry, dietary intakes, and other biomarkers. We divided sRAGE concentrations into quintiles based on the distribution in the control group. Conditional logistic regression was used to compute ORs and 95% confidence intervals (CI) for the associations between circulating levels of sRAGE and colorectal cancer risk. We ran two different models by including for each successive model additional adjustment variables incrementally. Model 1 (crude) was conditioned on the matching factors. Model 2 was additionally adjusted for body mass index (BMI), height, education (none, primary, technical and professional, secondary, and higher), physical activity (inactive, moderately inactive, moderately active, and active), smoking status, duration, and intensity (never; cigarettes/day 1–≤15, 16–≤25, and >26; and former smokers ≤10, 11–≤20, >20 years, and occasional), dietary energy, and intakes of alcohol, red and processed meat, dietary fiber, and dairy products. Dietary factors included as adjustment factors have been previously associated with colorectal cancer and/or sRAGE levels (42). P values for the linear trend (Ptrend) were obtained by including the median value of each quintile as a continuous variable in the model. We also examined sRAGE levels as a continuous variable, per SD increment.
Stratified analyses were performed by anatomic subsites (colon vs. rectal cancers and proximal colon vs. distal colon cancers), sex (men vs. women), age groups (<50, ≥50–<55, ≥55–<60, ≥60–<65, and ≥65), smoking status (never, former, and ever), alcohol intake (tertiles), physical activity (inactive, moderately inactive, moderately active, and active), and BMI (<25, ≥25–<30, and ≥30 kg/m²); and below or above sex-specific recommended cutoffs for waist circumference (WC; men, 94 cm and women, 80 cm) and waist-to-hip ratio (WHR; men, 0.90 and women, 0.85), and in women by menopausal status (pre-, post-, and perimenopause). The cutoffs for WC and WHR were based on the World Health Organization's definitions of central adiposity in European men and women (43). Additional stratified analyses were conducted for CRP (tertiles) as a marker of inflammation. Pheterogeneity was calculated using the Wald test. For subgroup analyses by anthropometric measures, individual models were run for BMI, WC, and WHR in men and women separately (model 2 without BMI). In sensitivity analyses, we excluded cases diagnosed during the first 2 years of follow-up and rerun the analyses.
We assessed the association between the genetic variants and colorectal cancer risk using data of all participants genotyped in EPIC study to increase the statistical power of the analysis. The associations between the five independent genetic variants and colorectal cancer risk were assessed by unconditional logistic regression models. Two models were run, an unadjusted model and a multivariable adjusted model, adjusted for sex, age, BMI, smoking status, alcohol, and country. Additive (major allele = 0, heterozygous = 1, and minor allele = 2), dominant (major allele = 0 and heterozygous + minor allele = 1), and recessive models (major allele + heterozygous = 0 and minor allele = 1) were run for the genetic variants. In sensitivity analyses, we analyzed the participants with overlapping genetic and sRAGE concentrations data. All the statistical analyses were performed using Stata 14.0 (StataCorp). P < 0.05 was considered statistically significant.
Data sharing statement
For information on how to submit an application for gaining access to EPIC data and/or biospecimens, please follow the instructions at http://epic.iarc.fr/access/index.php.
Results
Baseline characteristics and sRAGE levels in cases and controls are presented in Table 1. Compared with controls, colorectal cancer cases had higher BMI, WC, WHR, and CRP concentrations, and consumed more alcohol and less dairy products and fruit and vegetables. sRAGE concentrations were slightly lower in colorectal cancer cases than controls (1,086 vs. 1,130 pg/mL), but this was mainly observed among men (982 vs. 1,066 pg/mL in male cases vs. controls, respectively); whereas among women, sRAGE was 1,185 pg/mL in cases and 1,191 pg/mL in controls. BMI, WC, WHR, and alcohol intake were all negatively correlated with sRAGE levels, whereas sugar and confectionaries, fruit and vegetable, and cereals intake showed positive correlations (Supplementary Table S3). Women with higher sRAGE levels had lower CRP concentrations (Spearman rho = −0.156; P = 0.004).
Selected baseline demographic and lifestyle characteristics of study participants by colorectal cancer status, EPIC study 1992–2012.
sRAGE and colorectal cancer risk
sRAGE concentrations were inversely associated with colorectal cancer risk in multivariable adjusted analyses (OR comparing the highest with the lowest quintile, ORQ5vs.Q1, 0.75; 95% CI, 0.58–0.98; Ptrend = 0.035; Table 2). Subgroup analyses by sex showed an inverse risk association for men (ORQ5vs.Q1, 0.63; 95% CI, 0.42–0.94; Ptrend = 0.001), but not in women (ORQ5vs.Q1, 0.94; 95% CI, 0.63–1.38; Ptrend = 0.754; Pheterogeneity = 0.006). In men, sRAGE concentration was associated with a lower risk of both colon cancer (OR per SD increment, OR, 0.84; 95% CI, 0.70–0.99) and rectal cancer (OR, 0.80; 95% CI, 0.64–0.99), with no heterogeneity across anatomic subsites (Pheterogeneity = 0.607; Table 3). The magnitude of the inverse association appeared stronger for distal colon cancer (OR, 0.61; 95% CI, 0.44–0.84) compared with proximal cancer (OR, 0.94; 95% CI, 0.69–1.29), but no heterogeneity was observed (Pheterogeneity = 0.671). In women, no association was found between sRAGE concentration and colon cancer (OR, 0.99; 95% CI, 0.85–1.15) or rectal cancer (OR, 1.06; 95% CI, 0.86–1.32). Stratified analyses by age groups, BMI categories, WC and WHR cutoffs, and smoking status showed no significant differences across strata (Fig. 2). Women in higher CRP tertiles tended to have higher colorectal cancer risk associated with sRAGE (Pheterogeneity across = 0.011; Fig. 2).
ORs and 95% CIs for colorectal cancer risk associated with circulating sRAGE (quintiles and continuous), EPIC study 1992–2012.
ORs and 95% CIs for risk of colorectal cancer anatomic subsites associated with circulating sRAGE (continuous, per SD), EPIC study 1992–2012.
Multivariable adjusted OR and 95% CI of the associations between RAGE level and colorectal cancer, stratified by lifestyle, obesity, CRP, and menopause status. Multivariable adjusted OR and 95% CI were computed for the stratified analysis. All the analyses were conditional logistic regression models conditioned on matching factors and adjusted for BMI, education, physical activity, smoking status, dietary energy, and intakes of alcohol, red and processed meat, dietary fiber, and dairy products. The analyses stratified by BMI, physical activity, smoking, and alcohol were not adjusted for their respective variables.
Analyses of genetic variants
Table 4 presents the association of the genetic variants with colorectal cancer risk. While comparing minor allele versus major allele, rs1800625 (AGER, G vs. A, OR, 1.15; 95% CI, 1.02–1.29) was associated with an increased risk of colorectal cancer, whereas rs653765 (ADAM10, C vs. T, OR, 0.88; 95% CI, 0.80–0.97) was associated with a lower colorectal cancer risk, in univariate models. After multivariate adjustments, the association remained statistically significant for rs653765 (ADAM10, C vs. T, OR, 0.90; 95% CI, 0.82–0.99), but not for rs1800625 (AGER, G vs. A, OR, 1.11; 95% CI, 0.99–1.25).
ORs and 95% CIs for colorectal cancer risk associated with SNPs associated with sRAGE levels, EPIC study 1992–2012.
Sensitivity analysis
Exclusion of the cases that occurred within the first 2 years of follow-up did not change the associations between sRAGE concentrations and colorectal cancer (Table 1). The associations between SNPs and colorectal cancer in participants with overlapping genetic and sRAGE data showed similar, but no statistically significant associations for rs653765 (ADAM10, OR, 0.90; 95% CI, 0.78–1.05) or rs1800625 (AGER, G vs. A, OR, 1.00; 95% CI, 0.83–1.19; Supplementary Table S4).
Discussion
In this large, case–control study nested within a European prospective cohort, we found that prediagnostic circulating sRAGE levels were inversely associated with colorectal cancer risk in men, but not in women. The associations observed between sRAGE concentration and colorectal cancer did not vary by age or by lifestyle factors, including obesity and smoking status, suggesting that sex is the main effect modifier in the association between sRAGE levels and colorectal cancer. With respect to the SNP analyses, we found that the minor allele of rs653765 (ADAM10) was inversely associated with risk of colorectal cancer, whereas an increased risk was suggested for rs1800625 (AGER). However, we did not observe the association between rs653765 and levels of sRAGE.
RAGE is a pattern recognition receptor that recognizes multiple ligands such as S100, high mobility group box 1 protein (HMGB1), and amyloid-β peptide, in addition to the AGEs (44). RAGE is overexpressed in several diseases of the colon, including inflammatory bowel diseases (45). RAGE action in colon tissues may participate in colorectal cancer tumor initiation, progression, and invasion (46–48). sRAGE, by acting as a decoy of RAGE, binds to AGEs in the circulation and clears them by decreasing interaction with full-length cell surface RAGE. The evidence from mouse studies shows that injection of sRAGE is associated with a reduction in the expression of inflammatory mediators, such as TNFα (49). Evidence from case–control studies also shows that elevated sRAGE levels are associated with a lower risk of several cancers, including liver (50) and pancreatic cancer (51). This suggests that higher concentrations of sRAGE are protective against AGE-induced inflammation, which is involved in the etiology of various chronic diseases, such as diabetes and cancers, but the mechanisms need further exploration.
The underlying reasons for the observed difference between men and women in the association between sRAGE concentration and colorectal cancer risk are unclear. Several previously published studies that compared sRAGE levels between men and women suggest higher circulating levels in women (14, 15, 17), which we also observed in our study. One explanation of the sex difference in sRAGE levels may be that estrogens stimulate sRAGE expression and production (52). Estrogens have also been reported to reduce AGEs production and AGE-related inflammation (53). In our study, women with higher sRAGE levels had lower CRP concentrations (Spearman rho = −0.156; P = 0.004) and lower colorectal cancer risk, suggesting that sRAGE may possibly reduce colorectal cancer risk in women, by mitigating overall inflammation. However, analysis by menopausal status showed no differences across strata in our study population. Our findings suggest that additional studies are needed to understand the physiologic sex differences in sRAGE levels and how they may translate into the differential colorectal cancer risk associations that we have observed in this study.
Interestingly, the two previous publications on sRAGE and colorectal cancer in prospective cohorts have been conducted in men (22) and in women (23) only. The Alpha-Tocopherol, Beta-Carotene Cancer Prevention study reported high serum sRAGE levels to be associated with low colorectal cancer risk in Finnish male smokers (22). We expanded this observation by showing that such an inverse association was also observed in male never smokers. We expected to observe a greater reduction in colorectal cancer risk in nonsmokers compared with smokers, but our findings did not differ by smoking status. Smoking may be a source of AGEs exposure (2), but the magnitude of the contribution of smoking to overall AGEs exposures remains to be explored. sRAGE levels have been reported to be higher, lower, or unchanged in smokers compared with nonsmokers (54–56). It is still unknown whether smoking could induce an adaptive mechanism of sRAGE synthesis to cope with sustained formation of AGEs from glycotoxins contained in cigarettes. In a previous nested case–control study on a subsample of 1,249 postmenopausal women in the Women's Health Initiative study, higher sRAGE levels were observed to be associated with lower colorectal cancer risk in individuals with overweight and obesity, but not among normal weight postmenopausal women (23). Overall, our findings showed that sRAGE levels were associated with an inverse risk of colorectal cancer only in men, with no difference in magnitude across smoking status or any other lifestyle factor was observed.
We found that rs653765 located within ADAM10 (C vs. T) was associated with lower risk for colorectal cancer. However, rs653765 (ADAM10) was not associated with sRAGE levels in our study, in contrast to previous studies in which the minor allele of rs653765 was associated with lower sRAGE levels (21). Another SNP, rs1800625, located in the promoter region of AGER is involved in the initiation of the production of the RAGE or its isomers (39). Xu and colleagues (57) reported in a meta-analysis of 18 case–control genetic studies that the recessive model of rs1800625 was associated with an increase of overall cancer risk while analyzing case–controls studies of 6,246 cases of renal, lung, breast, cervical, liver, oral, breast, and colorectal cancers. Although our findings with genetic variants are intriguing, they may be attributed to the diversity of functions associated with AGER and ADAM10 genes. The production of sRAGE through the shedding of RAGE is dependent on ADAM10 levels. Thus, the overexpression of AGER coupled with lower ADAM10 activity will result in higher transmembrane RAGE and lower circulating sRAGE levels. This suggests that the interactions between AGER and ADAM10 may provide a better understanding of the genetic implications of RAGE and sRAGE in colorectal cancer development. In addition, the associations observed with the genetic data could be explained by other functions of the SNPs examined, particularly in the case of ADAM10 when considering its multiple actions, such as the formation of amyloid inclusions and the cleavage of a range of proteins (58). We did not observe a significant association between rs2070600 (AGER) and colorectal cancer, albeit our study showed that the major allele (C allele) of this SNP associates with higher sRAGE levels. A meta-analysis of 15 case–control studies showed that homozygous minor allele of this SNP was associated with an increased risk of all cancers (59). The absence of association of this SNP with colorectal cancer may be due to low statistical power, particularly as carriers of the minor allele are rare. Additional studies, using genetic data from larger research consortia, are needed to explore the link between the expression of AGER, ADAM10, and RNF5 genes, and levels of sRAGE and colorectal cancer initiation and development.
The strengths of our study include the large number of cases and controls, the prospective design, and the availability of dietary and lifestyle factors and genetic variants. Our study was, however, limited by the fact that we did not differentiate between endogenous secretory RAGE (esRAGE) and proteolytically cleaved RAGE (cRAGE), the two components of sRAGE. esRAGE is formed by alternative splicing of RAGE mRNA and cRAGE is produced by the shedding of the ectodomain of RAGE by metalloproteinases located at the surface of the cells. esRAGE is stable throughout the life course, whereas cRAGE levels vary with age and environmental factors (60). Because we have measured the total pool of plasma sRAGE we, therefore, cannot discern whether the different variants of sRAGE have specific and potentially opposite associations with study outcomes. Although the variability of cRAGE makes it a poor biomarker for a prospective study, cRAGE levels data would have permitted us to explore the association between SNPs from the ADAM10 gene, levels of cRAGE, and colorectal cancer risk. Our study was also limited by the fact that lifestyle factors and blood samples were collected at the recruitment, and may not necessarily reflect changes over years. Moreover, we cannot rule out residual confounding or unmeasured confounders, such as lifetime history of anti-inflammatory medication use.
In conclusion, we observed that prediagnostic circulating sRAGE levels were inversely associated with colorectal cancer risk in men, but not among women. We also found that the minor allele of rs653765 (ADAM10) was inversely associated with colorectal cancer risk. Additional studies are, however, required to further investigate how genetic variation and sex may affect sRAGE levels or modify its association with colorectal cancer risk.
Authors' Disclosures
No disclosures were reported.
Disclaimer
Where authors are identified as personnel of the International Agency for Research on Cancer/World Health Organization, the authors alone are responsible for the views expressed in this article and they do not necessarily represent the decisions, policy, or views of the International Agency for Research on Cancer/World Health Organization. The funders had no role in study design, data collection, analysis, and decision to publish or in preparation of the article.
Authors' Contributions
E.K. Aglago: Formal analysis, writing–original draft, writing–review and editing. S. Rinaldi: Formal analysis, writing–review and editing. H. Freisling: Conceptualization, methodology, writing–review and editing. L. Jiao: Conceptualization, methodology, writing–review and editing. D.J. Hughes: Conceptualization, methodology, writing–review and editing. V. Fedirko: Supervision, methodology, writing–review and editing. C.G. Schalkwijk: Conceptualization, validation, writing–review and editing. E. Weiderpass: Resources, supervision, writing–review and editing. C.C. Dahm: Resources, writing–review and editing. K. Overvad: Resources, writing–review and editing. A.K. Eriksen: Resources, writing–review and editing. C. Kyrø: Resources, writing–review and editing. M.-C. Boutron-Ruault: Resources, writing–review and editing. J.A. Rothwell: Resources, writing–review and editing. G. Severi: Resources, writing–review and editing. V. Katzke: Resources, writing–review and editing. T. Kühn: Resources, writing–review and editing. M.B. Schulze: Resources, writing–review and editing. K. Aleksandrova: Resources, writing–review and editing. G. Masala: Resources, writing–review and editing. V. Krogh: Resources, writing–review and editing. S. Panico: Resources, writing–review and editing. R. Tumino: Resources, writing–review and editing. A. Naccarati: Resources, writing–review and editing. B. Bueno-de-Mesquita: Resources, writing–review and editing. C.H. van Gils: Resources, writing–review and editing. T.M. Sandanger: Resources, writing–review and editing. I.T. Gram: Resources, writing–review and editing. G. Skeie: Resources, writing–review and editing. J.R. Quirós: Resources, writing–review and editing. P. Jakszyn: Resources, writing–review and editing. M.-J. Sánchez: Resources, writing–review and editing. P. Amiano: Resources, writing–review and editing. J.M. Huerta: Resources, writing–review and editing. E. Ardanaz: Resources, writing–review and editing. I. Johansson: Resources, writing–review and editing. S. Harlid: Resources, writing–review and editing. A. Perez-Cornago: Resources, writing–review and editing. A.-L. Mayén: Writing-review and editing. R. Cordova: Writing–review and editing. M.J. Gunter: Resources, supervision, writing–review and editing. P. Vineis: Resources, supervision, writing–review and editing. A.J. Cross: Resources, supervision, writing–review and editing. E. Riboli: Resources, supervision, writing–review and editing. M. Jenab: Conceptualization, supervision, funding acquisition, validation, investigation, methodology, project administration, writing–review and editing.
Acknowledgments
The funding for this work (WCRF 2015/1391, to principal investigator, M. Jenab) was obtained from Wereld Kanker Onderzoek Fonds, as part of the World Cancer Research Fund International grant program. The authors thank the EPIC study participants and staff for their valuable contribution to this research. The authors especially thank Mr. Bertrand Hemon and Dr. Aurelie Moskal for their support in preparing the databases and providing technical support pertaining to the data analysis, along with Ms. Audrey Brunat-Manquat for her assistance with the laboratory analyses for sRAGE. The coordination of the EPIC study was financially supported by the European Commission (DG-SANCO) and the International Agency for Research on Cancer. The national cohorts were supported by Danish Cancer Society (Denmark); Ligue Contre le Cancer, Institut Gustave Roussy, Mutuelle Générale de l'Education Nationale, and Institut National de la Santé et de la Recherche Médicale (INSERM, France); German Cancer Aid, German Cancer Research Center (DKFZ), and Federal Ministry of Education and Research (BMBF, Germany); Italian Association for Research on Cancer (AIRC), National Research Council, Associazione Iblea per la Ricerca Epidemiologica (AIRE-ONLUS) Ragusa, Associazione Volontari Italiani Sangu (AVIS) Ragusa, and the Sicilian Government (Italy); Dutch Ministry of Public Health, Welfare and Sports (VWS), Netherlands Cancer Registry (NKR), LK Research Funds, Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF), and Statistics Netherlands (the Netherlands); Health Research Fund (FIS); Regional Governments of Andalucía, Asturias, Basque Country, Murcia (No. 6236) and Navarra; the Centro de Investigación Biomédica en Red en Epidemiología y Salud Pública and Instituto de Salud Carlos II (ISCIII RETIC, RD06/0020, Spain); Health Research Fund (FIS) - Instituto de Salud Carlos III (ISCIII), Regional Governments of Andalucía, Asturias, Basque Country, Murcia and Navarra, and the Catalan Institute of Oncology - ICO (Spain); Swedish Cancer Society, Swedish Scientific Council, and Regional Government of Skåne and Västerbotten (Sweden); Cancer Research UK, Medical Research Council, Stroke Association, British Heart Foundation, Department of Health, Food Standards Agency, and the Wellcome Trust (United Kingdom). This work was also supported by grants from Cancer Research UK (14136 to EPIC-Norfolk; and C570/A16491, C8221/A19170, and C8221/A29017 to EPIC-Oxford) and Medical Research Council (1000143 to EPIC-Norfolk and MR/M012190/1 to EPIC-Oxford, United Kingdom). The EPIC-Norfolk study (DOI 10.22025/2019.10.105.00004) has received funding from the Medical Research Council (MR/N003284/1 and MC-UU_12015/1) and Cancer Research UK (C864/A14136). The authors are grateful to all the participants who have been part of the project and to the many members of the study teams at the University of Cambridge (Cambridge, England) who helped enable this research. This work was partially financially supported by the Fondation de France (FDF, grant no. 00081166 to H. Freisling and R. Cordova and FDF grant no. 00089811 to A.-L. Mayén).
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.
Footnotes
Note: Supplementary data for this article are available at Cancer Epidemiology, Biomarkers & Prevention Online (http://cebp.aacrjournals.org/).
Cancer Epidemiol Biomarkers Prev 2021;30:182–92
- Received June 8, 2020.
- Revision received July 31, 2020.
- Accepted October 9, 2020.
- Published first October 20, 2020.
- ©2020 American Association for Cancer Research.
References
Volume 30, Issue 1
