Abstract
Objectives
Receptor for advanced glycation end products (RAGE) expressed on adipocytes and immune cells can bind to ligand Nε-(carboxymethyl)-lysine (CML) and trigger dysregulation of adipokines and chronic inflammation. Soluble RAGE (sRAGE) mitigates the detrimental effect of RAGE. We examined the associations between circulating levels of CML-AGE and sRAGE and colorectal cancer (CRC).
Methods
In a case-cohort study of the Women’s Health Initiative Study, blood levels of CML-AGE and sRAGE were measured using ELISA. We used multivariable Cox regression model to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of incident CRC in relation to quartiles (Q) of biomarker levels.
Results
Average follow-up was 7.8 years for 444 cases and 805 subcohort members. In the subcohort, CML-AGE and sRAGE were inversely correlated with BMI (P values < 0.0001). Levels of CML-AGE and sRAGE were not associated with CRC. In BMI-specific analysis, the association between sRAGE and CRC was observed. Among women with BMI ≥ 25 kg/m2, those with highest levels of sRAGE had significantly lower risk for CRC as compared to women with lowest levels of sRAGE (HRQ4 versus Q1: 0.39; 95% CI: 0.17–0.91). This inverse association was not observed among women with BMI <25 kg/m2 (P value for interaction = 0.01).
Conclusions
Among postmenopausal women, the RAGE pathway may be involved in obesity-related CRC.
Keywords: Colorectal cancer, Advanced glycation end-products, Receptor for advanced glycosylation endproducts, Epidemiology, sRAGE, Body weight, Obesity, Nε-(carboxymethyl)-lysine, Pattern recognition receptors
1. Introduction
Colorectal cancer (CRC) is a leading cause of cancer-related deaths in men and women in most industrialized countries. Insulin resistance, excessive adiposity, and chronic inflammation are interrelated mechanisms contributing to CRC development [1]. Receptor for advanced glycation end-products (RAGE) can bind to advanced glycation end-products (AGEs) and trigger dysregulation of adipokines and chronic inflammation.
Advanced Glycation End products (AGEs) are heterogeneous compounds formed irreversibly via non-enzymatic glycation and oxidation reactions between reducing sugars (e.g., glucose and fructose) and free amino acid groups on lipids, proteins and nucleotides endogenously as byproducts of normal metabolism [4,5], or exogenously during high temperature cooking practice and from tobacco smoke [2,3]. High consumption of red meat, which is a rich source of AGEs, has been associated with increased CRC risk [1]. The principal mechanism by which AGEs exert their biological function is through interacting with the full-length multi-ligand RAGE (AGER in HUGO nomenclature), which is present on adipocytes, immune, epithelial, and endothelial cells [6,7].Under normal physiological states, RAGE is expressed at a low level. However, under pathological stress, upregulation of the RAGE gene is observed in companion with the accumulation of its ligands [8,9]. The interaction between RAGE and its ligand results in downstream activation of the pro-inflammatory transcription factor NF-κB [11,12], and has been shown to contribute to insulin resistance and chronic inflammation [9,13–15]. Nε-(carboxymethyl)-lysine (CML)-AGE is one of RAGE ligands and the best characterized AGEs [10]. In addition, accumulation of CML-AGE in the adipose tissue has been shown to contribute to dysregulation of adipokines [16]. A previous study found moderate-to-strong expression of CML-AGE in colon adenocarcinoma tissues [17]. There is also experimental evidence for the involvement of the AGEs/RAGE axis in colon inflammation and tumorigenesis [18–22].
Soluble RAGE (sRAGE) includes a C-truncated endogenous secretory isoform of RAGE (esRAGE); as well as cleavage forms of esRAGE and membrane-bound full-length RAGE [23,24]. sRAGE binds to RAGE ligands without triggering a signaling cascade because it has no transmembrane and intracellular domains. Therefore, sRAGE may be an endogenous factor that protects against AGEs/RAGE-mediated events [25,26].
We proposed that the AGEs/RAGE axis contributes the biological connection between environmental risk factors (such as red meats) and suggestive etiological mechanisms for CRC [12,31]. In this prospective study, we examined the associations between prediagnostic levels of CML-AGE and sRAGE and risk of subsequent development of CRC in the Women’s Health Initiative Observational Study (WHI-OS). We hypothesized that higher baseline levels of CML-AGE and lower levels of sRAGE would be associated with an increased CRC risk, and that the association would differ according to baseline levels of BMI status, insulin and C-reactive protein (CRP).
2. Materials and methods
2.1. Study population
The WHI-OS is a longitudinal cohort of 93,676 postmenopausal women aged 50–79 years who were recruited at 40 different clinical centers across the United States between October 1, 1993 and December 31, 1998 [32]. The present study was based on a case-cohort study population used for previous studies of CRC within the WHI-OS [33,34]. A case of incident CRC was defined and adjudicated according to International Classification of Diseases for Oncology site codes 153.0–153.4, 153.6–153.9, and 154.0–154.1. Study subjects were excluded if they were diagnosed with CRC or censored in the first year of follow-up; had taken diabetes treatment at baseline; or had a history of cancer of colon-rectum, breast, or endometrium. We identified 496 women who developed primary CRC with follow-up through February 29, 2004. A subcohort of 892 women was randomly selected from women who did not develop cancer as of February 29, 2014 [33]. The study protocol was approved by Institutional Review Board of each participating clinical center of the WHI, Baylor College of Medicine, and Michael E. DeBakey VA Medical Center (MED-VAMC).
2.2. Data and bio-specimen collection at baseline
At enrollment, women in the WHI-OS provided informed consent and completed questionnaires regarding demographic and lifestyle/behavioral factors, family history of cancer and medication use. They also completed a validated food frequency questionnaire on food consumption during the prior three months from which daily average nutrient consumption was computed as described elsewhere [35]. Trained WHI staff measured waist and hip circumference, height, and weight. BMI (weight (kg)/height (m)2) and waist to hip ratio (WHR) variables were calculated. At baseline, blood samples were obtained following an overnight fast of at least 12 h and stored at −70°C. Because the plasma/serum samples have been used to test biomarkers previously, all samples have gone through one freeze-thaw cycle, except for one plasma sample that went through two freeze-thaw cycles.
2.3. Measurement of circulating levels of CML-AGE, sRAGE, and other biomarkers
Citrate-plasma was used for measuring CML-AGE using a commercially available ELISA kit (Microcoat Biotechnologie Company, Bernried, Germany). Serum or EDTA-plasma was used for measuring sRAGE using human sRAGE Quantikine ELISA kit (R&D System Inc., Minneapolis, MN). Randomly ordered case and subcohort sample were run together in each batch (96-well plate). A 10% quality control (QC) sample provided by the WHI clinical coordinating center was also randomly placed on each plate. The laboratory personnel who performed the ELISA assay were blinded to QC and case-subcohort status. All samples were assayed in duplicate by the same personnel using kits with the same lot number.
Among 496 cases and 892 subcohort members, we included 442 cases and 800 subcohort members for CML-AGE testing and 318 cases and 638 subcohort members for sRAGE testing due to the use of residual samples from other studies. We included 396 cases and 706 subcohort members in the analysis of CML-AGE after excluding unreliable measurement points as determined by intraplate coefficient of variation (CV) >12%. Data on insulin, adipokines (adiponectin and leptin) and CRP were available for 457 cases and 841 subcohort member as part of earlier studies [33,34].
2.4. Statistical analysis
Untransformed CML-AGE and sRAGE levels were summarized by presenting their median values and interquartile ranges. The reproducibility of the lab assay was evaluated using intra-plate and inter-plate CV, and interclass correlation coefficient (ICC) based on the data of the QC samples. We examined the distribution of multiple factors across quartiles of CML-AGE and sRAGE using the Wilcoxon rank sum test or the ANOVA test for continuous variables and Pearson’s χ2 test for categorical variables. We examined the correlation between CML-AGE and sRAGE and exposure variables using non-parametric Spearman’s rank correlation analysis.
In this case-cohort study, the survival time was censored at the time of diagnosis for CRC, date of death, or February 29, 2004, whichever came first. We used Cox proportional hazards regression model to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association of incident CRC with CML-AGE or sRAGE. We calculated biomarker quartile cut-points according to their distributions in the subcohort, and assessed risk estimates for trend using a Wald test by treating the categories as the ordinal variable. All Cox models were computed with robust variance estimation in order to account for the case-cohort design [36,37]. We tested whether all models met the proportional hazard assumption using a published method [38]. Non-linearity using sRAGE as a continuous variable in predicting CRC risk was tested by restricted cubic spline regression [39].
Confounding factors were those associated with both CRC and CML-AGE or sRAGE (P value <0.25 in univariate analysis) or if their inclusion changed the risk estimates by more than 10%, which included age, physical activity (metabolic equivalent of task (MET)-hour/week), energy intake, alcohol use (serving per week), consumptions of saturated fat, available carbohydrate, calcium, total folate, red meat, and waist circumference as continuous variables; and race/ethnicity, smoking status and hormone replacement therapy (categorized as never, former, and current), family history of colorectal cancer, self-reported type 2 diabetes, hypertension, and use of NSAIDs (yes versus no) as categorical variables. All adiposity measures were individually associated with both CRC and sRAGE and CML-AGE. However, given multi-collinearity we were unable to include all three. We therefore conservatively included waist circumference in the multivariate model because this variable attenuated the sRAGE risk estimate the most compared to BMI and WHR. Although we excluded women with self-reported treated type 2 diabetes, there were 20 women with self-reported untreated type 2 diabetes. This variable (yes versus no) was included in the model. Freeze-thaw cycle of the samples was evaluated as an additional confounding factor but did not change the risk estimate. Furthermore, certain medications including statins have been shown to increase sRAGE levels by increasing RAGE shedding in animal models [40] as well as in human [41]. However, in this study, we did not find statin use (yes versus no) confounded the association between sRAGE and CRC. Diet and food variables were energy-adjusted using the density method. Levels of sRAGE (for CML-AGE) or CML-AGE (for sRAGE), insulin, adiponectin, and CRP were included additionally. We draw final conclusions based on the fully adjusted HRs.
We performed stratified analyses by three a priori variables: baseline BMI (<25 versus ≥25 kg/m2), insulin and CRP (median among subcohort as cut-off point). The statistical significance of the interactions was tested using a Wald test on the cross-product interaction term introduced into the multivariable Cox model. To further examine the relation between CML-AGE or sRAGE and insulin, we examined whether adjustment for CML-AGE or sRAGE changed the association between insulin and risk of CRC, and vice versa. Because sRAGE levels were higher when serum samples were used as the testing material compared with when EDTA-plasma samples were used (1630 pg/ml among 410 subcohort members using serum versus 1505 pg/ml among 228 subcohort members using EDTA-plasma, P = 0.02), in a sensitivity analysis, we also generated sample-type specific quartiles to estimate the risk. We also excluded data generated using the EDTA-plasma to examine the association between sRAGE and risk of pancreatic cancer. In addition, we performed sensitivity analyses in those who had been followed up for no less than five year in risk association analysis for sRAGE. Finally, because of the potential biological interaction between CML-AGE and sRAGE, we generated the mutually adjusted CML-AGE or sRAGE value using the residual method [42] and we performed sensitivity analyses by using these mutually adjusted values as the exposure variables for risk estimates.
All statistical analyses were performed using SAS 9.1 (SAS institute Inc., Cary, NC). All tests were two-sided with P-values < 0.05 indicative of statistical significance.
3. Results
3.1. Baseline characteristics and QC data
General characteristics of the cases versus non-cases were previously reported [33,34] (Supplemental Table 1). The HR was 1.49 (95% CI: 1.08–2.08) for BMI ≥ 25 kg/m2 compared with BMI < 25 kg/m2 after adjusting for age, race, smoking status, energy intake, physical activity, alcohol intake and family history of cancer. The baseline level of CML-AGE was significantly lower in cases than in the non-cases (P = 0.04). Table 1 shows general characteristics of the subcohort according to quartile (Q) of CML-AGE or sRAGE. Women with higher CML-AGE or sRAGE levels were more likely to be non-Hispanic white. Both CML-AGE and sRAGE were significantly inversely associated with BMI, waist circumference, WHR, and insulin levels, while positively associated with adiponectin. CML-AGE was positively associated with sRAGE and total fiber intake, but was inversely associated with red meat intake. sRAGE was positively associated with NSAIDs use and calcium intake, but inversely associated with hypertension, alcohol consumption, and CRP level (P values < 0.05). Similarly, among the subcohort members, CML-AGE and sRAGE were inversely correlated with BMI, insulin and CRP, but positively correlated with adiponectin (P values < 0.0001) (Table 2).
Table 1.
Baseline characteristic of study subjects according to levels of CML-AGE or sRAGE in the subcohort.
| Characteristicsa (mean (SD) or%) | CML-AGEb (quartile) n=561 |
P value | sRAGEb (quartile) n=638 |
P value | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 1 | 2 | 3 | 4 | |||
| Range | 185– 463 |
464–585 | 586– 706 |
707– 1339 |
346– 1130 |
1131–1461 | 1465– 1963 |
1964– 5043 |
||
| Age (year) | 62.0 (7.1) |
62.0 (7.4) |
63.3 (7.1) |
63.4 (7.8) |
0.10 | 63.2 (7.0) |
62.6 (7.7) |
62.9 (7.5) |
62.2 (7.4) |
0.70 |
| Race, the other line % non-Hispanic | 81.2 | 80.8 | 88.2 | 90.3 | 0.03 | 76.9 | 86.2 | 88.1 | 93.7 | <0.0001 |
| White | ||||||||||
| % African American | 8.52 | 9.60 | 1.69 | 5.71 | 14.9 | 4.00 | 1.25 | 0.63 | ||
| Body mass index, % of ≥25 | 71.9 | 63.0 | 59.2 | 48.5 | 0.0002 | 67.8 | 62.7 | 54.2 | 51.0 | 0.04 |
| Waist to hip ratio | 0.82 (0.07) |
0.80 (0.08) |
0.79 (0.07) |
0.78 (0.07) |
<0.0001 | 0.82 (0.07) |
0.81 (0.07) |
0.79 (0.07) |
0.78 (0.06) |
<0.0001 |
| Waist circumference (cm) | 89.8 (14.6) |
84.8 (12.8) |
82.3 (10.5) |
80.5 (11.5) |
<0.0001 | 88.2 (14.2) |
86.0 (13.1) |
82.0 (11.4) |
80.4 (11.0) |
<0.0001 |
| Physical activity (MET-hour/wk) | 11.7 (12.3) |
14.1 (14.9) |
15.0 (16.9) |
14.6 (13.8) |
0.14 | 12.8 (15.4) |
13.3 (14.8) |
16.1 (14.2) |
14.2 (15.0) |
0.21 |
| Smoking status | 0.89 | 0.08 | ||||||||
| Never smokers, % | 50.3 | 57.5 | 55.1 | 54.0 | 44.0 | 52.5 | 60.1 | 58.2 | ||
| Former smokers, % | 41.1 | 36.2 | 38.6 | 39.1 | 48.4 | 43.1 | 36.1 | 36.1 | ||
| Current smokers, % | 8.6 | 6.3 | 6.3 | 6.9 | 7.6 | 4.4 | 3.8 | 5.7 | ||
| ≥20 years of smoking, % | 61.9 | 50.7 | 45.4 | 53.2 | 0.15 | 60.5 | 49.3 | 45.0 | 46.7 | 0.12 |
| Colitis, % yes | 1.14 | 2.31 | 0.56 | 0.57 | 0.37 | 0.63 | 1.27 | 0.64 | 2.55 | 0.38 |
| Received colonoscopy, % yes | 53.1 | 52.3 | 56.2 | 58.6 | 0.62 | 56.9 | 53.2 | 56.0 | 59.9 | 0.69 |
| Hypertension, % yes | 30.5 | 34.9 | 25.8 | 28.0 | 0.29 | 36.1 | 34.2 | 27.0 | 22.0 | 0.02 |
| Self-reported diabetes, % yes | 2.3 | 2.3 | 0.6 | 1.1 | 0.47 | 4.4 | 0.6 | 1.3 | 1.3 | 0.06 |
| Family history of CRC, % yes | 17.4 | 16.7 | 14.9 | 18.4 | 0.86 | 18.4 | 11.0 | 13.9 | 21.2 | 0.08 |
| Hormone replacement therapy | 0.72 | 0.16 | ||||||||
| Never users, % | 33.8 | 38.0 | 38.8 | 45.9 | 41.5 | 40.3 | 37.6 | 36.0 | ||
| Former users, % | 15.0 | 17.1 | 15.6 | 14.5 | 16.5 | 14.8 | 18.5 | 17.1 | ||
| Current users, % | 51.2 | 44.9 | 45.6 | 39.6 | 42.0 | 44.9 | 43.8 | 46.9 | ||
| NSAIDs use ≥ once per week | 14.8 | 21.5 | 14.0 | 18.3 | 0.22 | 18.1 | 12.0 | 15.6 | 23.9 | 0.04 |
| Total energy intake (Kcals) | 1630 (663) |
1499 (582) |
1497 (570) |
1519 (563) |
0.11 | 1565 (628) |
1571 (600) |
1617 (613) |
1513 (611) |
0.52 |
| Protein (g) | 30.8 (18.9) |
34.0 (19.6) |
33.7 (20.7) |
32.3 (14.6) |
0.39 | 32.3 (16.2) |
30.8 (17.9) |
30.5 (14.8) |
32.4 (16.0) |
0.63 |
| Total fat (g) | 25.3 (12.8) |
25.9 (15.2) |
26.0 (16.5) |
23.0 (11.1) |
0.16 | 24.7 (14.8) |
24.9 (12.0) |
23.6 (12.3) |
23.9 (12.7) |
0.76 |
| Saturated fat (g) | 8.35 (4.23) |
8.41 (5.19) |
8.54 (5.54) |
7.60 (3.76) |
0.24 | 8.11 (5.08) |
8.12 (3.99) |
7.76 (3.93) |
8.05 (4.67) |
0.88 |
| Red meat (servings) | 0.32 (0.26) |
0.28 (0.26) |
0.29 (0.29) |
0.24 (0.20) |
0.02 | 0.29 (0.26) |
0.28 (0.23) |
0.26 (0.22) |
0.25 (0.22) |
0.34 |
| Available carbohydrate (g) | 84.6 (42.9) |
94.2 (56.9) |
97.5 (57.8) |
96.4 (47.5) |
0.08 | 88.4 (42.1) |
88.3 (43.3) |
89.8 (50.8) |
99.1 (49.7) |
0.12 |
| Total fiber (g) | 7.43 (4.96) |
9.02 (8.14) |
8.91 (6.30) |
9.18 (5.74) |
0.04 | 8.09 (5.27) |
8.25 (5.50) |
8.07 (5.08) |
8.93 (5.64) |
0.44 |
| Folate equivalent (µg) | 225 (149) |
268 (215) |
256 (166) |
260 (155) |
0.11 | 234 (126) |
237 (156) |
245 (195) |
268 (154) |
0.22 |
| Calcium (mg) | 506 (201) |
527 (208) |
532 (227) |
565 (216) |
0.07 | 524 (224) |
499 (193) |
563 (212) |
574 (210) |
0.005 |
| Alcohol (servings per week) | 2.49 (4.84) |
2.80 (6.68) |
2.49 (4.48) |
2.91 (6.39) |
0.86 | 3.42 (7.2) |
3.63 (6.9) |
2.25 (3.5) |
1.82 (3.6) |
0.008 |
| CML-AGE (ng/ml) | 564 (181) |
589 (175) |
627 (187) |
637 (184) |
0.0004 | |||||
| sRAGE (pg/ml) | 1377 (566) |
1602 (595) |
1587 (624) |
1702 (689) |
<0.0001 | |||||
| Insulin (uIU/ml) | 8.28 (5.50) |
6.87 (4.94) |
6.24 (4.18) |
5.84 (4.68) |
<0.0001 | 7.97 (5.7) |
7.54 (5.1) |
5.94 (4.3) |
5.46 (3.9) |
<0.0001 |
| Adiponectin (µg/ml) | 26.7 (12.5) |
29.6 (14.3) |
31.2 (15.4) |
32.1 (14.0) |
0.003 | 27.5 (14.4) |
29.8 (14.9) |
30.2 (12.9) |
33.3 (14.5) |
0.0064 |
| Leptin (ng/ml) | 21.3 (16.4) |
19.4 (29.8) |
17.7 (13.9) |
17.2 (17.7) |
0.24 | 22.2 (32.7) |
19.9 (17.5) |
15.9 (12.5) |
17.6 (17.4) |
0.05 |
| C-reactive protein (mg/l) | 4.24 (6.89) |
3.28 (4.20) |
3.11 (4.08) |
3.13 (4.26) |
0.12 | 4.05 (4.93) |
4.22 (7.42) |
3.37 (4.17) |
2.47 (2.51) |
0.01 |
AIDs, Nonsteroidal Anti-inflammatory Drugs.
Food variables were energy-adjusted.
Table 2.
Spearman correlation coefficient (r) for the selected characteristics with CML-AGE and sRAGE in the subcohort in the WHI-OS Study [1993–2004].
| Characteristics | CML-AGE |
sRAGE |
|---|---|---|
| r | r | |
| Age at randomization | 0.09a | −0.05 |
| BMI, kg/m2 | −0.22b | −0.23b |
| Waist circumference (cm) | −0.23b | −0.23b |
| Waist to hip ratio | −0.22b | −0.22b |
| Years of smoking | 0.06 | −0.13a |
| Total fat (g/day) | −0.12a | −0.0007 |
| Saturated fat (g/day) | −0.11a | 0.003 |
| Available carbohydrate intake (g/day) | 0.08a | 0.09a |
| Total sugar (g/day) | 0.11a | 0.08a |
| Protein (g/day) | 0.02 | 0.03 |
| Red meat (g/day) | −0.15a | −0.04 |
| Processed meat (g/day) | −0.15a | −0.01 |
| Glucose intake (g/day) | 0.12a | 0.07 |
| Fructose intake (g/day) | 0.12a | 0.06 |
| Alcohol consumption | 0.08 | −0.06 |
| Calcium (g/day) | 0.10a | 0.11a |
| Serum insulin | −0.19b | −0.23b |
| Serum glucose | −0.12a | −0.21b |
| Serum IGF-1 | −0.0005 | 0.01 |
| Serum adiponectin | 0.17a | 0.20b |
| Serum leptin | −0.11a | −0.11a |
| CRP | −0.15a | −0.097a |
| sRAGE | 0.16a | 1.0 0 |
P < 0.05.
P < 0.0001.
Using the blinded QC data (28 pairs for sRAGE and 26 pairs for CML-AGE), we calculated the inter-plate CV and ICC for both markers. The interplate CV was 7.85% for sRAGE and was 14.8% for CML-AGE. The ICC was 94.5% for sRAGE and 67.0% for CML-AGE.
3.2. Main effect
In univariate analyses, post-menopausal women with higher levels of CML-AGE were significantly less likely to develop CRC, although the significant association was not found in multivariable analyses (Table 3). For sRAGE, women in the 2nd quartile of baseline levels of sRAGE had significantly reduced CRC risk (HRadjQ2 versus Q1= 0.52, 95% CI: 0.30–0.92) in multivariate models, although a dose-response pattern for the association was not observed (P for trend = 0.90). However, the test of non-linearity of continuous sRAGE in predicting risk of CRC via spline regression did not reject the linearity hypothesis (P = 0.18).
Table 3.
Hazard ratio (HR) and 95% confidence interval (CI) of colorectal cancer by quartile of CML-AGE or sRAGE levels in the WHI-OS [1993–2004].
| Biomarker |
Quartile |
|||||||
|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | P for trend | ||||
| CML-AGE (ng/ml, range) | 185–463 | 464–585 | 586–706 | 707–1339 | ||||
| Case/non-cases (n/n) | 116/176 | 108/177 | 84/17 | 87/175 | ||||
| HR 95% CIa | 1.0 0 | 0.92 | 0.65–1.31 | 0.67 | 0.46–0.96 | 0.69 | 0.48–1.00 | 0.02 |
| HR 95% CIb | 1.00 | 1.00 | 0.67–1.51 | 0.81 | 0.53–1.24 | 0.81 | 0.52–1.26 | 0.24 |
| HR 95% CIc | 1.00 | 1.07 | 0.67–1.75 | 0.67 | 0.40–1.15 | 0.81 | 0.48–1.38 | 0.22 |
| HR 95% CId | 1.00 | 0.99 | 0.60–1.63 | 0.64 | 0.37–1.10 | 0.81 | 0.48–1.37 | 0.24 |
| HR 95% CIe | 1.00 | 1.11 | 0.68–1.82 | 0.70 | 0.41–1.19 | 0.78 | 0.46–1.34 | 0.18 |
| HR 95% CIf | 1.00 | 0.87 | 0.62–1.76 | 0.70 | 0.40–1.22 | 0.85 | 0.49–1.47 | 0.34 |
| sRAGE (pg/ml, range) | 346–113 0 | 1125–1461 | 1465–1963 | 1964–5043 | ||||
| Case/non-cases (n/n) | 95/160 | 66/159 | 87/160 | 69/159 | ||||
| HR 95% CIa | 1.00 | 0.71 | 0.47–1.05 | 0.93 | 0.64–1.35 | 0.74 | 0.50–1.10 | 0.30 |
| HR 95% CIb | 1.00 | 0.71 | 0.44–1.14 | 1.13 | 0.72–1.78 | 0.83 | 0.50–1.37 | 0.86 |
| HR 95% CIc | 1.00 | 0.59 | 0.35–1.01 | 1.13 | 0.68–1.86 | 0.87 | 0.51–1.51 | 0.88 |
| HR 95% CId | 1.00 | 0.54 | 0.31–0.93 | 1.05 | 0.62–1.78 | 0.83 | 0.47–1.46 | 0.98 |
| HR 95% CIe | 1.00 | 0.55 | 0.31–0.96 | 1.05 | 0.61–1.79 | 0.84 | 0.48–1.48 | 0.98 |
| HR 95% CIf | 1.00 | 0.52 | 0.30–0.92 | 0.97 | 0.56–1.69 | 0.81 | 0.45–1.45 | 0.90 |
Model 1: Adjusted for age.
Model 2: Model 1 + physical activity, waist circumference, alcohol consumption, energy intake, consumption of saturated fat, available carbohydrate, calcium, folate equivalent, and red meat (continuous), and race, smoking, family history of colorectal cancer, NASIDs use, hypertension, diabetes, and hormone replacement therapy (categorical).
Model 3: Model 2 + sRAGE (continuous) for CML-AGE, or CML-AGE (continuous) for sRAGE.
Model 4: Model 3 + adiponectin (continuous).
Model 5: Model 4 + insulin (continuous.
Model 6: Model 5 + CRP (continuous).
3.3. Effect modification and sensitivity analysis
Table 4 shows that there were no significant interactions of CML-AGE by BMI, insulin and CRP levels. Table 5 shows the results for stratified analysis of sRAGE and CRC risk according to BMI, insulin, and CRP. There was a significant inverse association between the highest levels of sRAGE and CRC among women with BMI ≥ 25 kg/m2, but this inverse association was not seen among women with BMI < 25 kg/m2 (P value for interaction 0.01). The interaction with insulin and CRP was not statistically significant.
Table 4.
Association between CML-AGE and risk of incident CRC by BMI, insulin, and CRP in the WHI-OS [1993–2004].
| Factors | Cases/Non-cases | HRa | 95% CIa | HRb | 95% CIb | Cases/Non-cases | HRa | 95% CIa | HRb | 95% CIb |
|---|---|---|---|---|---|---|---|---|---|---|
| BMI <25 kg/m2 | BMI ≥25 kg/m2 | |||||||||
| Q1 | 33/53 | 1.00 | 1.00 | 1.1.1 | 83/123 | 1.00 | 1.1.2 | 1.00 | 1.1.3 | |
| Q2 | 31/68 | 0.68 | 0.25–1.82 | 0.52 | 0.18–1.55 | 77/109 | 1.18 | 0.60–2.34 | 1.25 | 0.63–2.48 |
| Q3 | 39/74 | 0.40 | 0.14–1.10 | 0.30 | 0.10–0.93 | 45/104 | 0.69 | 0.32–1.53 | 0.78 | 0.35–1.76 |
| Q4 | 37/92 | 0.44 | 0.17–1.16 | 0.33 | 0.11–1.00 | 50/83 | 0.89 | 0.43–1.87 | 1.02 | 0.49–2.13 |
| P for trend | 0.06 | 0.04 | 0.50 | 0.77 | ||||||
| P for interaction | 0.49 | |||||||||
| Insulin < median | Insulin ≥ median | |||||||||
| Q1 | 39/64 | 1.00 | 1.1.4 | 1.00 | 77/112 | 1.00 | 1.1.5 | 1.00 | 1.1.6 | |
| Q2 | 42/80 | 0.87 | 0.37–2.04 | 0.77 | 0.30–1.93 | 66/97 | 1.17 | 0.58–2.37 | 1.23 | 0.60–2.50 |
| Q3 | 37/95 | 0.45 | 0.19–1.10 | 0.39 | 0.15–1.01 | 47/83 | 0.79 | 0.38–1.65 | 0.85 | 0.40–1.79 |
| Q4 | 48/110 | 0.69 | 0.28–1.72 | 0.62 | 0.23–1.71 | 39/65 | 0.70 | 0.30–1.61 | 0.77 | 0.33–1.78 |
| P for trend | 0.26 | 0.24 | 0.29 | 0.40 | ||||||
| P for interaction | 0.89 | |||||||||
| CRP < median | CRP ≥ median | |||||||||
| Q1 | 46/82 | 1.00 | 1.00 | 70/94 | 1.00 | 1.00 | ||||
| Q2 | 47/93 | 0.69 | 0.32–1.49 | 0.70 | 0.30–1.65 | 61/84 | 1.48 | 0.73–2.97 | 1.41 | 0.70–2.85 |
| Q3 | 47/95 | 0.66 | 0.30–1.44 | 0.69 | 0.29–1.62 | 37/83 | 0.61 | 0.26–1.41 | 0.58 | 0.24–1.41 |
| Q4 | 41/104 | 0.82 | 0.39–1.72 | 0.97 | 0.43–2.18 | 46/71 | 0.63 | 0.25–1.61 | 0.69 | 0.27–1.77 |
| P for trend | 0.61 | 0.995 | 0.82 | 0.20 | ||||||
| P for interaction | 0.15 |
Model 1: Adjusted for age, physical activity, alcohol consumption, energy intake, consumption of saturated fat, available carbohydrate, calcium, folate equivalent and red meat, sRAGE, adiponectin, and insulin (continuous), and race, waist circumference, smoking, family history of colorectal cancer, NASIDs use, hypertension, diabetes, hormone replacement therapy (categorical).
Model 2: Model 1 + CRP (continuous).
Table 5.
Association between sRAGE and incident CRC by BMI, insulin and CRP in the WHI-OS [1993–2004].
| Factors | Cases/Non-cases | HRa | 95% CIa | HRb | 95% CIb | Cases/Non-cases | HRa | 95% CIa | HRb | 95% CIb |
|---|---|---|---|---|---|---|---|---|---|---|
| BMI <25 kg/m2 | BMI ≥25 kg/m2 | |||||||||
| Q1 | 23/57 | 1.00 | 1.00 | 1.1.7 | 72/103 | 1.00 | 1.1.8 | 1.00 | 1.1.9 | |
| Q2 | 17/60 | 0.60 | 0.20–1.77 | 0.56 | 0.17–1.87 | 49/99 | 0.49 | 0.24–1.01 | 0.49 | 0.24–1.01 |
| Q3 | 30/76 | 0.98 | 0.34–2.86 | 1.02 | 0.31–3.31 | 57/84 | 1.08 | 0.55–2.11 | 0.95 | 0.48–1.88 |
| Q4 | 40/79 | 1.88 | 0.65–5.40 | 2.39 | 0.69–8.24 | 29/80 | 0.42 | 0.18–0.97 | 0.39 | 0.17–0.91 |
| P for trend | 0.17 | 0.13 | 0.19 | 0.11 | ||||||
| P for interaction | 0.01 | |||||||||
| Insulin < median | Insulin ≥ median | |||||||||
| Q1 | 26/64 | 1.00 | 1.1.10 | 1.00 | 1.1.11 | 69/96 | 1.00 | 1.1.12 | 1.00 | 1.1.13 |
| Q2 | 29/70 | 0.48 | 0.17–1.32 | 0.39 | 0.13–1.14 | 37/89 | 0.59 | 0.28–1.27 | 0.56 | 0.26–1.22 |
| Q3 | 42/92 | 1.07 | 0.44–2.63 | 0.89 | 0.32–2.45 | 45/68 | 0.97 | 0.46–2.06 | 0.93 | 0.44–1.97 |
| Q4 | 40/95 | 1.15 | 0.48–2.80 | 1.00 | 0.38–2.66 | 29/64 | 0.59 | 0.25–1.44 | 0.58 | 0.24–1.43 |
| P for trend | 0.37 | 0.52 | 0.40 | 0.38 | ||||||
| P for interaction | 0.06 | |||||||||
| CRP < median | CRP ≥ median | |||||||||
| Q1 | 34/70 | 1.00 | 1.00 | 61/90 | 1.00 | 1.00 | ||||
| Q2 | 23/83 | 0.41 | 0.15–1.11 | 0.32 | 0.10–0.95 | 43/76 | 0.63 | 0.27–1.50 | 0.67 | 0.28–1.61 |
| Q3 | 46/82 | 1.24 | 0.56–2.73 | 1.08 | 0.44–2.62 | 41/78 | 0.86 | 0.39–1.94 | 0.88 | 0.39–1.99 |
| Q4 | 39/99 | 0.94 | 0.43–2.07 | 0.87 | 0.36—2.08 | 30/60 | 0.50 | 0.17–1.53 | 0.47 | 0.15–1.46 |
| P for trend | 0.49 | 0.59 | 0.28 | 0.25 | ||||||
| P for interaction | 0.21 |
Model 1: Adjusted for age, physical activity, alcohol consumption, energy intake, consumption of saturated fat, available carbohydrate, calcium, folate equivalent and red meat, CML-AGE, adiponectin, and insulin (continuous), and race, waist circumference, smoking, family history of colorectal cancer, NASIDs use, hypertension, diabetes, hormone replacement therapy (categorical).
Model 2: Model 1 + CRP (continuous).
The sensitivity analyses using the mutually adjusted values of CML-AGE and sRAGE for modeling (Supplemental Table 2) showed similar findings to the main analyses. An interaction effect by BMI was also observed (Supplemental Table 3). Using the sample-type specific cutoff points for sRAGE, we observed the same trend of the association between sRAGE and pancreatic cancer risk. The multivariable HR was 0.64 (95% CI: 0.38–1.09), 0.81 (95% CI: 0.47–1.40), and 0.81 (0.46–1.41 ) for the second, third, and fourth quartile, respectively, compared with the first quartile (P trend = 0.59) (Model 5 in Table 2). After excluding data from 205 cases and 410 subcohort members with EDTA plasma determined sRAGE, the HRs were 0.99 (95% CI 0.52–1.02), 0.85 (95% CI: 0.42–1.70), and 0.85 (95% CI: 0.43–1.71) for the second, third, and fourth quartile compared with the first quartile (P trend = 0.58), respectively. In addition, the interaction effect of sRAGE* BMI in our sensitivity analysis was similar to the main analysis. Among the study subjects with BMI ≥ 25 kg/m2, the HR for CRC was 0.96 (95% CI: 0.40–2.30), 0.90 (95% CI: 0.34–2.38), and 0.43 (0.14–1.32) for the second, third, and fourth quartile, respectively, compared with the first quartile of sRAGE (P trend = 0.19). When we limited the analysis among those who were followed up for ≥5 years (median follow-up 6.2 years), we found a stronger association between sRAGE and CRC. The HR was 0.51 (95% CI 0.18–1.43), 1.36 (95% CI: 0.52–3.59), and 0.50 (95% CI: 0.15–1.58) for the second, third and fourth quartile of sRAGE compared with the lowest quartile.
Finally, we found that higher levels of insulin were associated with significantly increased CRC risk in our cohort (HRadjQ4 versus Q1: 1.64 (95% CI: 1.09–2.48)). Further adjustment for either CML-AGE or sRAGE changed the HRs by less than 10% (e.g., HRadjQ4 versus Q1 : 1.56, 95% CI: 1.02–2.38 for CML-AGE adjustment and HR adjQ4 versus Q1 : 1.53, 95% CI: 0.94–2.50 for sRAGE adjustment, respectively).
4. Discussion
In this prospective case-cohort study of postmenopausal women, we did not observe a significant association between baseline circulating CML-AGE and sRAGE levels and CRC risk. We found an inverse association between pre-diagnostic sRAGE levels and the subsequent risk of incident CRC in women with BMI ≥ 25 kg/m2, independent of insulin levels and CRP. There was no inverse association between sRAGE and CRC in women with normal BMI.
sRAGE is a decoy receptor for RAGE ligands, including CML-AGE, pro-inflammatory high mobility group Box 1 (HMGB1) and S100 family proteins, all of which have been implicated in colorectal carcinogenesis [43,44]. In animal models, one study found that high-fat feeding induced expression of two RAGE ligands, HMGB1 and CML-AGE, in the liver and adipose tissue. Deficiency of RAGE or administration of sRAGE partially blocked the high-fat-diet induced inflammation and weight gain [45]. Circulating levels of sRAGE have been inversely associated with CRP levels and BMI [27,28]. Our previous studies on sRAGE in men found a significant inverse association with the risk of developing CRC [29] and colorectal adenomas [30]. Our present findings were in line with these observations and implied that higher circulating levels of sRAGE may confer a stronger protective effect against CRC in women with BMI ≥25 kg/m2. The crosstalk between RAGE and adiposity has been demonstrated in progression of atherosclerosis in an apoE/RAGE knock-out mouse study [46]. Endogenous secretory RAGE has been associated with platelet activation and oxidative stress among obese women [47]. Platelet count and activation have also been implicated in CRC development [48]. Further understanding of the RAGE-ligand interaction in the tissue microenvironment, as well as the production of sRAGE under inflammatory or obese conditions, may aid interpretation of our study findings.
Our study found circulating CML-AGE was positively correlated with adiponectin, and inversely correlated with indices of excessive adiposity and insulin. This finding was consistent with other studies conducted in children or healthy adults [49,50], but was in contrast to the reported positive association between CML-AGE and insulin resistance and chronic inflammatory diseases [51,52] including incident cardiovascular risk both in type 2 diabetics [53] and in healthy older adults [54]. It is unknown whether this discrepancy was due to different study populations in terms of age, sex, or chronic disease status. Several recent studies examined the involvement of the AGEs/RAGE axis in adiposity [46,55,56]. An animal study showed that CML-AGE accumulation in conjunction with increased expression of RAGE in adipose tissue contributes to the dysregulation of adipokines in obesity and insulin resistance [16]. RAGE-mediated trapping of CML-AGE in the adipose tissue may explain decreased plasma levels of CML in obese subjects [16] and hence our study finding. However, we did not observe a significant interaction between CML-AGE and BMI on CRC risk in our study. Because levels of CML-AGE have also been shown to be modulated by its receptors, including galectin 3, macrophage scavenger receptor [57], AGER-1 and sirtuin 1 [58], its full impact at the tissue level may need to be considered in the context of its receptors.
We observed an overall trend of an inverse association between sRAGE and CRC which was also observed for CRC in the ATBC Finnish male smoker cohorts [29]. However, some discrepancies were also noted. For example, we did not find an association between CML-AGE and CRC in the present study in contrast to the ATBC Study. The source of AGEs was different in these two cohorts in that the study subjects in the ATBC Study might have been exposed to both tobacco AGEs and dietary AGEs. In the WHI, only 6% of the women were current smokers and diet may have been the major source of AGEs. However, compared with the ATBC Study, women in the WHI cohort had slightly higher levels of CML-AGE (561 ng/ml in ATBC versus 586 ng/ml in WHI, median), and much higher levels of sRAGE (572 pg/ml in ATBC versus 1463 pg/ml in WHI, median) although the same type of ELISA assay kits were used in two studies. Due to the observed higher levels of sRAGE in women than men, it is unknown whether the harmful effect of CML-AGE was offset by higher sRAGE in the WHI Study. In addition, unlike the present study, there was no interaction between sRAGE and BMI in the ATBC Study. While nearly 70% had a BMI greater than 25 kg/m2 in the WHI Study, the Finnish male smokers were generally lean. The narrow and lean range of BMI in the ATBC Study might explain the absent interaction between CML-AGE/sRAGE and BMI. Collectively, our observations indicated that the effect of CML-AGE and sRAGE on CRC risk may depend on exposure characteristics such as BMI and smoking. However, additional research is needed to further elucidate potential gender differences (such as sex hormones) in these associations.
Our study had several limitations. First, we used previously thawed biological samples from prior case-cohort studies. Nevertheless, the study samples went through the same freeze-thaw cycles. Moreover, the adjustment for freeze-thaw cycle did not materially change risk estimates. For CML-AGE, the steps of sample dilution and proteinase K digestion introduced variability resulting in the need to exclude data from 140 participants. However, the characteristics of these 140 participants were not different from the rest of the study subjects used in data analysis. We used both serum and EDTA plasma in measuring sRAGE. Measurement errors due to the sample resource could have compromised the accuracy of the data. However, the sensitivity analysis showed the same trend of association as the main finding we reported. Second, the association between sRAGE levels and CRC risk was not linear, which complicates the potential dose-response based support of causality or effect mediation. Research in additional cohorts is necessary to help further refine understanding of biological interaction between sRAGE and the host system. For example, sRAGE levels could have been affected by polymorphisms of RAGE (AGER) [59,60]. Genetic background should be considered in future studies. Finally, we performed multiple stratified analyses to explore the plausible association between CML-AGE/sRAGE, adiposity, insulin and CRP. We did not exclude the possibility of a false positive finding based on the limited sample size. The observed interaction effect between sRAGE and BMI needs to be interpreted with caution especially in the absence of dose-response seen for sRAGE among women with higher BMI.
5. Conclusions
We report a novel finding suggesting the involvement of sRAGE in obesity-related CRC in post-menopausal women. The balance between the synthesis of anti-inflammatory sRAGE and pro-inflammatory RAGE ligands may be an important determinant of RAGE mediated reactions. Further studies with larger sample size are needed to elucidate the inter-related mechanisms among adiposity, insulin resistance, chronic inflammation, and RAGE-ligand axis in diverse populations. Broadly, our research may have implications concerning the role of pattern recognition receptors such as RAGE in fostering an inflammatory tumor microenvironment in the development of CRC.
Supplementary Material
Acknowledgments
This research was supported in part by the National Cancer Institute (5R03CA156626, PI: Jiao), the Dan L Duncan Scholar Award, Gillson Longenbaugh Foundation, and Golfers Against Cancer organization (To L Jiao); Texas Medical Center Digestive Disease Center (P30 DK56338); and Houston Veterans Affairs Health Services Research Center of Innovations (CIN13–413). The views expressed in this article are those of the authors and do not necessarily represent the views of the NCI and the Department of Veterans Affair or other funders.
We acknowledge the dedicated efforts of investigators and staff at the Women’s Health Initiative (WHI) clinical centers, the WHI Clinical Coordinating Center, and the National Heart, Lung and Blood program office (listing available at http://www.whi.org). We also recognize the WHI participants for their extraordinary commitment to the WHI program. For a list of all the investigators who have contributed to WHI science, please visit: http://www.whiscience.org/publications/WHI_investigators_longlist.pdf. The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, US Department of Health and Human Services through contracts N01WH22110, 24152, 32100-2, 32105-6, 32108-9, 32111-13, 32115, 32118–32119, 32122, 42107-26, 42129-32, and 44221, and the Cancer Center Support Grant NIH:NCI P30CA022453. We thank Dr. Xiaonan Xue for statistical advice.
Footnotes
Conflict of interest
None.
Authorship contribution
Liang Chen, Li Jiao—Conception and design, acquisition of data, experiment, data analysis and interpretation, drafting and revising article, final approval.
Lesley Tinker, Howard Strickler, Gloria Y.F. Ho, Marc Gunter— Conception and design, acquisition of data, interpretation of data, revising article, final approval.
Haleh Sangi-Haghpeykar, Hashem B. El-Serag—Acquisition of data, revising article, final approval.
Zhigang Duan, Kathryn Royse—Data analysis and interpretation, revising article, final approval.
Thomas Rohan, Craig Logsdon, Donna L White—Data interpretation, revising article, final approval.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.canep.2016.04.004
References
- 1.WCRF/AICR Expert Report, Food, Nutrition, Physical Activity and the Prevention of Cancer: a Global Perspective. 2007 doi: 10.1017/S002966510800712X. [DOI] [PubMed] [Google Scholar]
- 2.Cerami C, Founds H, Nicholl I, Mitsuhashi T, Giordano D, Vanpatten S, et al. Tobacco smoke is a source of toxic reactive glycation products. Proc. Natl. Acad. Sci. U. S. A. 1997;94:13915–13920. doi: 10.1073/pnas.94.25.13915. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Uribarri J, Woodruff S, Goodman S, Cai W, Chen X, Pyzik R, et al. Advanced glycation end products in foods and a practical guide to their reduction in the diet. J. Am. Diet. Assoc. 2010;110:911–916. doi: 10.1016/j.jada.2010.03.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ott C, Jacobs K, Haucke E, Navarrete SA, Grune T, Simm A. Role of advanced glycation end products in cellular signaling. Redox Biol. 2014;2:411–429. doi: 10.1016/j.redox.2013.12.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Semba RD, Nicklett EJ, Ferrucci L. Does accumulation of advanced glycation end products contribute to the aging phenotype? J. Gerontol. A Biol. Sci. Med. Sci. 2010;65:963–975. doi: 10.1093/gerona/glq074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ott C, Jacobs K, Haucke E, Navarrete SA, Grune T, Simm A. Role of advanced glycation end products in cellular signaling. Redox Biol. 2014;2:411–429. doi: 10.1016/j.redox.2013.12.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tang D, Kang R, Coyne CB, Zeh HJ, Lotze MT. PAMPs and DAMPs: signal 0s that spur autophagy and immunity. Immunol. Rev. 2012;249:158–175. doi: 10.1111/j.1600-065X.2012.01146.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Brett J, Schmidt AM, Yan SD, Zou YS, Weidman E, Pinsky D, et al. Survey of the distribution of a newly characterized receptor for advanced glycation end products in tissues. Am. J. Pathol. 1993;143:1699–1712. [PMC free article] [PubMed] [Google Scholar]
- 9.Kalea AZ, Schmidt AM, Hudson BI. RAGE: a novel biological and genetic marker for vascular disease. Clin. Sci. (Lond.) 2009;116:621–637. doi: 10.1042/CS20080494. [DOI] [PubMed] [Google Scholar]
- 10.Singh R, Barden A, Mori T, Beilin L. Advanced glycation end-products: a review. Diabetologia. 2001;44:129–146. doi: 10.1007/s001250051591. [DOI] [PubMed] [Google Scholar]
- 11.Schmidt AM, Hori O, Cao R, Yan SD, Brett J, Wautier JL, et al. RAGE: a novel cellular receptor for advanced glycation end products. Diabetes. 1996;45(Suppl. 3):S77–S80. doi: 10.2337/diab.45.3.s77. [DOI] [PubMed] [Google Scholar]
- 12.Yan SF, Ramasamy R, Naka Y, Schmidt AM. Glycation, inflammation, and RAGE: a scaffold for the macrovascular complications of diabetes and beyond. Circ. Res. 2003;93:1159–1169. doi: 10.1161/01.RES.0000103862.26506.3D. [DOI] [PubMed] [Google Scholar]
- 13.Cai W, He JC, Zhu L, Chen X, Zheng F, Striker GE, et al. Oral glycotoxins determine the effects of calorie restriction on oxidant stress, age-related diseases, and lifespan. Am. J. Pathol. 2008;173:327–336. doi: 10.2353/ajpath.2008.080152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Uribarri J, Cai W, Peppa M, Goodman S, Ferrucci L, Striker G, et al. Circulating glycotoxins and dietary advanced glycation endproducts: two links to inflammatory response, oxidative stress, and aging. J. Gerontol. A Biol. Sci. Med. Sci. 2007;62:427–433. doi: 10.1093/gerona/62.4.427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Yan SF, Ramasamy R, Schmidt AM. The RAGE axis: a fundamental mechanism signaling danger to the vulnerable vasculature. Circ. Res. 2010;106:842–853. doi: 10.1161/CIRCRESAHA.109.212217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Gaens KH, Goossens GH, Niessen PM, van Greevenbroek MM, van der Kallen CJ, Niessen HW, et al. Nepsilon-(carboxymethyl)lysine-receptor for advanced glycation end product axis is a key modulator of obesity-induced dysregulation of adipokine expression and insulin resistance. Arterioscler. Thromb. Vasc. Biol. 2014;34:1199–1208. doi: 10.1161/ATVBAHA.113.302281. [DOI] [PubMed] [Google Scholar]
- 17.van Heijst JW, Niessen HW, Hoekman K, Schalkwijk CG. Advanced glycation end products in human cancer tissues: detection of Nepsilon-(carboxymethyl) lysine and argpyrimidine. Ann. N. Y. Acad. Sci. 2005;1043:725–733. doi: 10.1196/annals.1333.084. [DOI] [PubMed] [Google Scholar]
- 18.Heijmans J, Buller NV, Hoff E, Dihal AA, van der PT, van Zoelen MA, et al. Rage signalling promotes intestinal tumourigenesis. Oncogene. 2012;9 doi: 10.1038/onc.2012.119. doi: http://dx.doi.org/10.1038/onc.2012.119. [DOI] [PubMed] [Google Scholar]
- 19.Hu X, Cheng Y. Possible participation of receptor for advanced glycation end products (RAGE) in the origin of cancer stem cells in diabetic patients with colon cancer. Med. Hypotheses. 2013;80:620–623. doi: 10.1016/j.mehy.2013.02.004. [DOI] [PubMed] [Google Scholar]
- 20.Zill H, Gunther R, Erbersdobler HF, Folsch UR, Faist V. RAGE expression and AGE-induced MAP kinase activation in Caco-2 cells. Biochem. Biophys. Res. Commun. 2001;288:1108–1111. doi: 10.1006/bbrc.2001.5901. [DOI] [PubMed] [Google Scholar]
- 21.Dahlmann M, Okhrimenko A, Marcinkowski P, Osterland M, Herrmann P, Smith J, et al. RAGE mediates S100A4-induced cell motility via MAPK/ERK and hypoxia signaling and is a prognostic biomarker for human colorectal cancer metastasis. Oncotarget. 2014;5:3220–3233. doi: 10.18632/oncotarget.1908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Liang H, Zhong Y, Zhou S, Peng L. Knockdown of RAGE expression inhibits colorectal cancer cell invasion and suppresses angiogenesis in vitro and in vivo. Cancer Lett. 2011;313:91–98. doi: 10.1016/j.canlet.2011.08.028. [DOI] [PubMed] [Google Scholar]
- 23.Hanford LE, Enghild JJ, Valnickova Z, Petersen SV, Schaefer LM, Schaefer TM, et al. Purification and characterization of mouse soluble receptor for advanced glycation end products (sRAGE) J. Biol. Chem. 2004;279:50019–50024. doi: 10.1074/jbc.M409782200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hudson BI, Carter AM, Harja E, Kalea AZ, Arriero M, Yang H, et al. Identification, classification, and expression of RAGE gene splice variants. FASEB J. 2008;22:1572–1580. doi: 10.1096/fj.07-9909com. [DOI] [PubMed] [Google Scholar]
- 25.Vazzana N, Santilli F, Cuccurullo C, Davi G. Soluble forms of RAGE in internal medicine. Intern. Emerg. Med. 2009;4:389–401. doi: 10.1007/s11739-009-0300-1. [DOI] [PubMed] [Google Scholar]
- 26.Yamagishi S, Matsui T. Soluble form of a receptor for advanced glycation end products (sRAGE) as a biomarker. Front. Biosci. (Elite Ed.) 2010;2:1184–1195. doi: 10.2741/e178. [DOI] [PubMed] [Google Scholar]
- 27.McNair ED, Wells CR, Mabood QA, Basran R, Pearce C, Orvold J, et al. Modulation of high sensitivity C-reactive protein by soluble receptor for advanced glycation end products. Mol. Cell. Biochem. 2010;341:135–138. doi: 10.1007/s11010-010-0444-3. [DOI] [PubMed] [Google Scholar]
- 28.Norata GD, Garlaschelli K, Grigore L, Tibolla G, Raselli S, Redaelli L, et al. Circulating soluble receptor for advanced glycation end products is inversely associated with body mass index and waist/hip ratio in the general population. Nutr. Metab. Cardiovasc. Dis. 2009;19:129–134. doi: 10.1016/j.numecd.2008.03.004. [DOI] [PubMed] [Google Scholar]
- 29.Jiao L, Taylor PR, Weinstein SJ, Graubard BI, Virtamo J, Albanes D, et al. Advanced glycation end products, soluble receptor for advanced glycation end products, and risk of colorectal cancer. Cancer Epidemiol. Biomark. Prev. 2011;20:1430–1438. doi: 10.1158/1055-9965.EPI-11-0066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Jiao L, Chen L, Alsarraj A, Ramsey D, Duan Z, El-Serag HB. Plasma soluble receptor for advanced glycation end-products and risk of colorectal adenoma. Int. J. Mol. Epidemiol. Genet. 2012;3:294–304. [PMC free article] [PubMed] [Google Scholar]
- 31.Prasad C, Imrhan V, Marotta F, Juma S, Vijayagopal P. Lifestyle and advanced glycation end products (AGEs) burden: its relevance to healthy aging. Aging Dis. 2014;5:212–217. doi: 10.14336/AD.2014.0500212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Design of the women’s health initiative clinical trial and observational study. The women’s health initiative study group. Control. Clin. Trials. 1998;19:61–109. doi: 10.1016/s0197-2456(97)00078-0. [DOI] [PubMed] [Google Scholar]
- 33.Gunter MJ, Hoover DR, Yu H, Wassertheil-Smoller S, Rohan TE, Manson JE, et al. Insulin, insulin-like growth factor-I, endogenous estradiol, and risk of colorectal cancer in postmenopausal women. Cancer Res. 2008;68:329–337. doi: 10.1158/0008-5472.CAN-07-2946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Ho GY, Wang T, Gunter MJ, Strickler HD, Cushman M, Kaplan RC, et al. Adipokines linking obesity with colorectal cancer risk in postmenopausal women. Cancer Res. 2012;72:3029–3037. doi: 10.1158/0008-5472.CAN-11-2771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Patterson RE, Kristal AR, Tinker LF, Carter RA, Bolton MP, gurs-Collins T. Measurement characteristics of the women’s health initiative food frequency questionnaire. Ann. Epidemiol. 1999;9:178–187. doi: 10.1016/s1047-2797(98)00055-6. [DOI] [PubMed] [Google Scholar]
- 36.Self SG, Prentice RL. Asymptomic distribution theory and efficiency results for case-cohort studies. Ann. Stat. 1988;16:64–81. [Google Scholar]
- 37.Barlow WE, Ichikawa L, Rosner D, Izumi S. Analysis of case-cohort designs. J. Clin. Epidemiol. 1999;52:1165–1172. doi: 10.1016/s0895-4356(99)00102-x. [DOI] [PubMed] [Google Scholar]
- 38.Xue X, Xie X, Gunter M, Rohan TE, Wassertheil-Smoller S, Ho GY, et al. Testing the proportional hazards assumption in case-cohort analysis. BMC Med. Res. Methodol. 2013;13:88. doi: 10.1186/1471-2288-13-88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Govindarajulu US, Spiegelman D, Thurston SW, Ganguli B, Eisen EA. Comparing smoothing techniques in Cox models for exposure-response relationships. Stat. Med. 2007;26:3735–3752. doi: 10.1002/sim.2848. [DOI] [PubMed] [Google Scholar]
- 40.Quade-Lyssy P, Kanarek AM, Baiersdorfer M, Postina R, Kojro E. Statins stimulate the production of a soluble form of the receptor for advanced glycation end products. J. Lipid Res. 2013;54:3052–3061. doi: 10.1194/jlr.M038968. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Tam HL, Shiu SW, Wong Y, Chow WS, Betteridge DJ, Tan KC. Effects of atorvastatin on serum soluble receptors for advanced glycation end-products in type 2 diabetes. Atherosclerosis. 2010;209:173–177. doi: 10.1016/j.atherosclerosis.2009.08.031. [DOI] [PubMed] [Google Scholar]
- 42.Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epidemiologic studies. Am. J. Clin. Nutr. 1997;65:1220S–1228S. doi: 10.1093/ajcn/65.4.1220S. [DOI] [PubMed] [Google Scholar]
- 43.Ichikawa M, Williams R, Wang L, Vogl T, Srikrishna G. S100A8/A9 activate key genes and pathways in colon tumor progression. Mol. Cancer Res. 2011;9:133–148. doi: 10.1158/1541-7786.MCR-10-0394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Luo Y, Chihara Y, Fujimoto K, Sasahira T, Kuwada M, Fujiwara R, et al. High mobility group Box 1 released from necrotic cells enhances regrowth and metastasis of cancer cells that have survived chemotherapy. Eur. J. Cancer. 2013;49:741–751. doi: 10.1016/j.ejca.2012.09.016. [DOI] [PubMed] [Google Scholar]
- 45.Song F, Hurtado del PC, Rosario R, Zou YS, Ananthakrishnan R, Xu X, et al. RAGE regulates the metabolic and inflammatory response to high-fat feeding in mice. Diabetes. 2014;63:1948–1965. doi: 10.2337/db13-1636. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Ueno H, Koyama H, Shoji T, Monden M, Fukumoto S, Tanaka S, et al. Receptor for advanced glycation end-products (RAGE) regulation of adiposity and adiponectin is associated with atherogenesis in apoE-deficient mouse. Atherosclerosis. 2010;211:431–436. doi: 10.1016/j.atherosclerosis.2010.04.006. [DOI] [PubMed] [Google Scholar]
- 47.Vazzana N, Guagnano MT, Cuccurullo C, Ferrante E, Lattanzio S, Liani R, et al. Endogenous secretory RAGE in obese women: association with platelet activation and oxidative stress. J. Clin. Endocrinol. Metab. 2012;97:E1726–E1730. doi: 10.1210/jc.2012-1473. [DOI] [PubMed] [Google Scholar]
- 48.Kawai T, Akira S. Signaling to NF-kappaB by toll-like receptors. Trends Mol. Med. 2007;13:460–469. doi: 10.1016/j.molmed.2007.09.002. [DOI] [PubMed] [Google Scholar]
- 49.Accacha S, Rosenfeld W, Jacobson A, Michel L, Schnurr FJ, Shelov S, et al. Plasma advanced glycation end products (AGEs), receptors for AGEs and their correlation with inflammatory markers in middle school-age children. Horm Res. Paediatr. 2013;80:318–327. doi: 10.1159/000354831. [DOI] [PubMed] [Google Scholar]
- 50.Sebekova K, Somoza V, Jarcuskova M, Heidland A, Podracka L. Plasma advanced glycation end products are decreased in obese children compared with lean controls. Int. J. Pediatr. Obes. 2009;4:112–118. doi: 10.1080/17477160802248039. [DOI] [PubMed] [Google Scholar]
- 51.Mark AB, Poulsen MW, Andersen S, Andersen JM, Bak MJ, Ritz C, et al. Consumption of a diet low in advanced glycation end products for 4 weeks improves insulin sensitivity in overweight women. Diabetes Care. 2014;37:88–95. doi: 10.2337/dc13-0842. [DOI] [PubMed] [Google Scholar]
- 52.Yamagishi S, Ueda S, Okuda S. Food-derived advanced glycation end products (AGEs): a novel therapeutic target for various disorders. Curr. Pharm. Des. 2007;13:2832–2836. doi: 10.2174/138161207781757051. [DOI] [PubMed] [Google Scholar]
- 53.Hanssen NM, Beulens JW, van DS, Scheijen JL, van der AD, Spijkerman AM, et al. Plasma advanced glycation endproducts are associated with incident cardiovascular events in individuals with type 2 diabetes: a case-cohort study with a median follow-up of 10 years (EPIC-NL) Diabetes. 2015;64:257–265. doi: 10.2337/db13-1864. [DOI] [PubMed] [Google Scholar]
- 54.Kizer JR, Benkeser D, Arnold AM, Ix JH, Mukamal KJ, Djousse L, et al. Advanced glycation/glycoxidation endproduct carboxymethyl-lysine and incidence of coronary heart disease and stroke in older adults. Atherosclerosis. 2014;235:116–121. doi: 10.1016/j.atherosclerosis.2014.04.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Chen CY, Abell AM, Moon YS, Kim KH. An advanced glycation end product (AGE)-receptor for AGEs (RAGE) axis restores adipogenic potential of senescent preadipocytes through modulation of p53 protein function. J. Biol. Chem. 2012;287:44498–44507. doi: 10.1074/jbc.M112.399790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Monden M, Koyama H, Otsuka Y, Morioka T, Mori K, Shoji T, et al. Receptor for advanced glycation end products regulates adipocyte hypertrophy and insulin sensitivity in mice: involvement of toll-like receptor 2. Diabetes. 2013;62:478–489. doi: 10.2337/db11-1116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Chatzigeorgiou A, Kandaraki E, Piperi C, Livadas S, Papavassiliou AG, Koutsilieris M, et al. Dietary glycotoxins affect scavenger receptor expression and the hormonal profile of female rats. J. Endocrinol. 2013;218:331–337. doi: 10.1530/JOE-13-0175. [DOI] [PubMed] [Google Scholar]
- 58.Cai W, Ramdas M, Zhu L, Chen X, Striker GE, Vlassara H. Oral advanced glycation endproducts (AGEs) promote insulin resistance and diabetes by depleting the antioxidant defenses AGE receptor-1 and sirtuin 1. Proc. Natl. Acad. Sci. U. S. A. 2012;109:15888–15893. doi: 10.1073/pnas.1205847109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Gaens KH, Ferreira I, van der Kallen CJ, van Greevenbroek MM, Blaak EE, Feskens EJ, et al. Association of polymorphism in the receptor for advanced glycation end products (RAGE) gene with circulating RAGE levels. J. Clin. Endocrinol. Metab. 2009;94:5174–5180. doi: 10.1210/jc.2009-1067. [DOI] [PubMed] [Google Scholar]
- 60.Han Z, Liu Q, Sun C, Li Y. The interaction between obesity and RAGE polymorphisms on the risk of knee osteoarthritis in Chinese population. Cell. Physiol. Biochem. 2012;30:898–904. doi: 10.1159/000341467. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.