Prediagnostic Leptin, Adiponectin, C-Reactive Protein, and the Risk of Postmenopausal Breast Cancer

Breast cancer is the most frequently diagnosed malignancy among women in the United States., with an estimated 230,480 new cases and 39,520 deaths occurring annually (1, 2). For postmenopausal breast cancer, obesity has been associated with an increased risk of disease (3, 4). Several possible mechanisms for this association have been hypothesized, including estrogen production in non-ovarian tissues, changes in circulating adipokine concentrations, chronic inflammation, hyperinsulinemia, and increases in insulin-like growth factors (5–7).

Leptin and adiponectin are adipokine hormones produced by metabolically active white adipose tissue (8). Leptin increases in concert with adiposity and has been shown to have mitogenic effects on epithelial cells and to promote cellular proliferation, migration, and invasion in breast cancer cell lines (6, 7); properties potentially increasing breast cancer risk and progression. Adiponectin is inversely associated with adiposity and may act to reduce the risk and the progression of breast cancer via its antiproliferative and possibly pro-apoptotic effects on breast cancer cells (5). Chronic inflammation is hypothesized to increase the risk of cancer by creating a tissue microenvironment high in reactive oxygen and nitrogen species leading to the potential for malignant DNA alterations and by elevating the levels of cancer promoting inflammatory cytokines (9).

Prospective epidemiologic studies examining the association of leptin and adiponectin with the risk of postmenopausal breast cancer are limited. In a cohort of women in Sweden, prediagnostic levels of leptin were not associated with disease risk overall, nor among women older than 54 years when examined as a proxy for postmenopausal status; however, a positive association of leptin with breast cancer risk was suggested for advanced-stage disease (10). While the results were not presented separately for postmenopausal women, no association between prediagnostic leptin concentrations and breast cancer risk was detected among women at high risk for breast cancer participating in the National Surgical Adjuvant Breast and Bowel project Protocol (NSABP) P1 (11). In the combined Nurses' Health Study (NHS) and the NHSII cohorts, prediagnostic adiponectin was inversely associated with postmenopausal breast cancer risk (12), although 2 later studies did not confirm this association (10, 13). To the best of our knowledge, prior prospective epidemiologic studies examining the association of C-reactive protein (CRP), a marker of acute-phase inflammatory response, with the risk of breast cancer have not reported findings separately for postmenopausal women. For all women, studies have primarily reported null findings (14–18), although a positive association was reported in at least one prior prospective study (19).

For the present nested case–control study, we examined whether prediagnostic serum concentrations of leptin, adiponectin, the leptin:adiponectin ratio, and CRP were associated with the risk of postmenopausal breast cancer among women participating in the biospecimen subcohort of the Multiethnic Cohort (MEC) Study. In addition, we examined whether the associations were independent of body mass index (BMI) and established risk factors for postmenopausal breast cancer.

We conducted a nested case–control study of breast cancer among postmenopausal women participating in the biospecimen subcohort of the MEC. The MEC is a longitudinal study designed to investigate the associations of dietary and lifestyle factors with the incidence of cancer, and has been described previously in detail (20). Briefly, from 1993 to 1996, more than 215,000 men and women who were residing in Hawaii and California and were between 45 and 75 years of age at recruitment entered the cohort. Potential participants were identified through drivers' license files, voter registration lists, and Medicare files to obtain a multiethnic sample of African Americans, Japanese Americans, Latinos, Native Hawaiians, and whites. At cohort entry, participants completed a self-administered, 26-page baseline questionnaire that included queries on demographic characteristics, anthropometric measures, medical history, family history of cancer, reproductive and menstrual history, cancer risk factors, and detailed questions on diet.

The prospective MEC biospecimen subcohort was established from 2001 to 2006 by asking surviving cohort members to provide specimens of blood and urine (21). Blood samples were drawn and processed within 4 hours of collection by centrifugation and the components (serum, plasma, buffy coat, red cells) were aliquoted by automation into 0.5-mL cryotubes and stored in the vapor phase of liquid nitrogen (−186°C). For approximately 95% of the participants contributing to the biorepository, fasting blood samples (≥8 hours) were obtained. In total, 67,594 cohort members contributed to the biorepository from which the cases and controls were selected for the present study. The study protocol was approved by the institutional review boards of the University of Hawaii (Honolulu, HI) and the University of Southern California (Los Angeles, CA).

Regular linkages of the cohort to the Surveillance, Epidemiology, and End Results (SEER) cancer registries for Hawaii and California were conducted to identify incident cases of breast cancer diagnosed over the follow-up period. For this analysis, cases were defined as postmenopausal women who contributed a blood sample to the MEC biorepository before receiving a diagnosis of primary invasive breast cancer according to the most recent tumor registry linkage (December 31, 2009). Breast cancer diagnoses were classified using the International Classification of Diseases for Oncology, Third Edition codes C50.0-C50.9 and were restricted to invasive malignancies. One control for each case was randomly selected from the eligible pool of postmenopausal women who were alive and free of a diagnosis of breast cancer at the age of the case's diagnosis and who matched the case on year of birth (±1 year), location (Hawaii or California), ethnicity, date of blood draw (±6 months), time of blood draw (±2 hours), hours fasting before blood draw (<6, 6–<8, 8–<10, ≥10), and hormone replacement therapy use at blood draw (current versus not current). A total of 706 eligible cases were identified during the follow-up period and matched to 706 controls for the present analysis.

All assays were conducted at the Analytical Biochemistry Shared Resource of the University of Hawaii Cancer Center. Frozen serum samples were retrieved from the MEC biorepository for matched case–control sets. The matched samples were thawed and analyzed together within batches by laboratory personnel blinded to case–control status. ELISAs (R & D Systems) were used to obtain measurements of serum adiponectin and leptin. Quantification of CRP was conducted by turbidimetric measurement using a Latex Particle Enhanced–Based kit (Pointe Scientific) and the Cobas MiraPlus clinical chemistry analyzer (Roche Diagnostics). Blinded replicate samples of pooled serum were included in each analysis batch to assess quality control. On the basis of a total of 69 duplicate samples, the intrabatch coefficients of variation for leptin, adiponectin, and CRP were 6.4%, 9.4%, and 5.0%, respectively.

Characteristics of the cases and the controls were compared using the χ² test for categorical variables, the t test for normally distributed continuous variables, and the Wilcoxon rank-sum test for nonnormally distributed continuous variables. To assess the interrelations between leptin, adiponectin, CRP, and BMI among controls, Spearman partial correlation coefficients, adjusting for age at blood draw, hours fasting before blood draw, time of blood draw, and ethnicity were computed. Conditional logistic regression with matched sets as strata was used to calculate the ORs and the 95% confidence intervals (CI) for postmenopausal breast cancer. Quartiles of serum leptin, adiponectin, the leptin:adiponectin ratio, and CRP were examined in all models using cutoff points based on the exposure distribution among controls. The lowest exposure group served as the referent in all models. Linear trends were assessed by Wald tests of the parameter estimates for the natural log-transformed continuous variables. Established risk factors for breast cancer, including positive first-degree family history of breast cancer, age at menarche (≤12, 13–14, ≥15 years), age at menopause (<45, 45–49, 50–54, ≥55 years), age at first live birth (15–17, 18–20, 21–25, >25 years), parity (nulliparous, 1, 2–3, ≥4 children), oral contraceptive use (<1, 1–5, >5 years), alcohol consumption (grams of ethanol/d), physical activity (hours of moderate and vigorous physical activity/d), and pack-years of cigarette smoking were examined as potential confounders but were not included in the final models as they were not found alone, or in combination, to change the risk estimates by more than 10% (22). Sensitivity analyses of the main effects for all biomarkers were conducted by excluding women diagnosed with breast cancer within 1, 2, 3, 4, and 5 years from the date of blood draw. Heterogeneity in the ORs by the length of follow-up time was assessed by Wald tests of the cross-product terms for the natural log-transformed biomarker concentrations and follow-up time.

Associations for the circulating biomarkers with the risk of postmenopausal breast cancer were also examined in analyses stratified by ethnicity, BMI (<25, 25–29.9, ≥ 30), and age at blood draw (<67.1, ≥67.1 years, median). Heterogeneity in the ORs by ethnicity, BMI, and age at blood draw was assessed by Wald tests of the cross-product terms for continuous variables. Unconditional polytomous logistic regression adjusting for the matching factors was used to examine the associations for serum biomarkers by clinical tumor characteristics including stage at diagnosis (localized, regional, and distant metastasis) tumor size (<2.0, 2.0–5.0, >5.0 cm), axillary node status (N0, N1), grade (I, II, III), and hormone receptor status (model 1 = ER⁺ vs. ER⁻, model 2 = PR⁺ vs. PR⁻, model 3 = ER⁺ or PR⁺ vs. ER⁻ PR⁻). Heterogeneity in the ORs for postmenopausal breast cancer in polytomous models was assessed by global Wald tests contrasting the parameter estimates comparing case subgroups with all controls. All tests were 2-sided; P < 0.05 was considered statistically significant. All data analyses were conducted using SAS 9.2 statistical software (SAS Institute Inc.).

The mean age at blood draw was 67.8 years of age, and the mean BMI at cohort entry was 26.7 and 26.1 kg/m² for cases and for controls, respectively (Table 1). Japanese American women comprised the largest ethnic group (35%), followed by whites (22%), Latinas (19%), African Americans (15%), and Native Hawaiians (10%). Cases were more likely to report a higher BMI than were controls (P = 0.05) but were similar with respect to the other breast cancer risk factors examined. The levels of serum leptin, the leptin:adiponectin ratio, and CRP were higher for cases than for controls (P < 0.05 for the rank-sum test). Prediagnostic levels of leptin (r = 0.62, P ≤ 0.001), adiponectin (r = −0.36, P ≤ 0.001), and CRP (r = 0.30, P = < 0.001) were correlated with BMI among controls (Table 2). Statistically significant correlations were also observed among the analytes (P ≤ 0.001).

In the basic models accounting for the matching factors, higher prediagnostic serum leptin [OR_{Q4 vs. Q1}, 1.94 (1.37–2.75); P_trend ≤ 0.001), the leptin:adiponectin ratio [OR, 1.91 (1.36–2.68); P_trend = 0.005], and CRP [OR, 1.41 (1.01–1.96); P_trend = 0.014], were associated with an increased risk of postmenopausal breast cancer (Table 3). The associations for leptin [OR, 1.90 (1.28–2.80); P_trend = 0.001] and the leptin:adiponectin ratio [OR, 1.82 (1.25–2.66); P_trend = 0.035] remained statistically significant in models adjusting for BMI, as well as in models additionally adjusting for CRP, history of diabetes, and history of hypertension as proxies for the metabolic syndrome. The association for CRP was modestly attenuated when adjusted for BMI; however, there remained a statistically significant positive linear trend in the odds for disease (P_trend = 0.045). Adiponectin was not associated with the risk of postmenopausal breast cancer in any model examined. Mutual adjustment for adiponectin and leptin, or all 3 biomarkers combined, was not found to influence the respective ORs for postmenopausal breast cancer (data not shown). Compared with women with a BMI < 25 kg/m², overweight [OR, 1.63 (1.27–2.09)] and obese [OR, 1.44 (1.07–1.94)] women were at an increased risk for postmenopausal breast cancer. This remained unchanged upon adjustment for leptin, adiponectin, or CRP when examined alone or in combination (data not shown).

No heterogeneity in the ORs for leptin, adiponectin, the leptin:adiponectin ratio, or CRP on the risk of postmenopausal breast cancer was detected in the sensitivity analyses examining the consistency of the associations for samples obtained between 1 and 5 years before the date of diagnosis (P_{heterogeneity} > 0.05; data not shown). In addition, the ORs for serum biomarkers were similar across strata in analyses stratified by ethnicity and age at blood draw (data not shown), as well as in models examining the associations by clinical tumor characteristics including stage at diagnosis, tumor size, axillary node status, grade, and hormone receptor status (Supplementary Tables S1 and S2). Heterogeneity in the ORs for continuous variables was detected for the cross-products terms of BMI with leptin (P = 0.014), the leptin:adiponectin ratio (P = 0.011), and CRP (P = 0.008). In models stratified by BMI (<25, 25–29.9, ≥30 kg/m²), no clear association with the risk of postmenopausal breast cancer was detected for the biomarkers examined; however, the results were based on a limited number cases (Table 4).

In this case–control study nested within the MEC, we found an increased risk of postmenopausal breast cancer among women with the highest prediagnostic levels of leptin, the leptin:adiponectin ratio, and CRP. The associations for leptin and the leptin:adiponectin ratio remained after adjustment for BMI, as well as after adjustment for proxies for the metabolic syndrome including CRP concentration and history of diabetes and hypertension.

In an analysis conducted in the Northern Sweden Health and Disease Cohort (10), the association between prediagnostic leptin and the risk of breast cancer differed by the stage at diagnosis with an inverse association reported for early stage and a suggestive positive association reported for late stage disease when examined among pre- and postmenopausal women combined. In our sample of postmenopausal women, the associations were similar for early- and late-stage disease and in agreement with several previous retrospective case–control studies reporting positive findings (23, 24). In contrast to our findings, no association between prediagnostic leptin and breast cancer risk was detected among women at high risk for breast cancer participating in the NSABP-P1 trial; however, the results were not presented separately for postmenopausal women (11). It is also noteworthy, that in our study, the associations for leptin and the leptin:adiponectin ratio, while based on a limited number of cases, reached statistical significance only among normal weight women. A recent study examining the relationship of leptin in the circulation and in the breast tissue of healthy women (25), reported correlations for leptin of r = 0.62 (P = 0.009) for normal weight women, r = 0.36 (P = 0.08) for overweight women, and r = 0.03 (P = 0.86) for obese women. Thus, if circulating and breast tissue leptin levels are more highly correlated among women at the lower end of the BMI range, then stronger associations for serum leptin and breast cancer risk may also be expected in this group. Furthermore, the correlations between leptin concentrations measured in plasma and in breast tissue were also reported by Llanos and colleagues (25) to differ by race, potentially contributing to the discrepant findings reported in epidemiologic studies.

Our null results for adiponectin are in accord with 2 previous prospective studies that failed to detect an association with breast cancer risk among pre- or postmenopausal women (10, 13); however, in an analysis of the combined NHS and the NHSII cohorts, a statistically significant inverse association for adiponectin was reported after adjustment for BMI and breast cancer risk factors (12). The limited number of prospective analyses conducted to date, coupled with the inconsistent findings, warrants additional studies aimed at further clarifying the role of circulating adipokines in the etiology and the progression of breast cancer.

Elevated leptin levels have been shown in breast cancer cell lines to stimulate downstream signaling pathways similar to that of estrogen receptor-α, as well as to enhance aromatase activity and interfere with anti-estrogen pathways (26), potentially increasing the growth and development of breast tumors in an estrogen-dependent manner. Adiponectin receptors are also expressed in breast cancer cell lines with the administration of adiponectin shown to retard cellular growth rates and to increase apoptosis (27–30). This model of an opposing role for leptin and adiponectin in the etiology and the progression of breast cancer highlights the importance of examining the potential interrelation of these adipokine hormones. While in our sample, there is no evidence that the ratio of leptin:adiponectin or further adjustment for BMI improves prediction of risk over leptin alone (c-statistic for all models = 0.57), future prospective studies should include the ratio measure, as it remains plausible that it may exhibit a stronger association with disease risk than either marker alone (31) and has been correlated with increased tumor aggressiveness in patients with breast cancer (23).

In contrast to our findings, previous prospective studies examining the association between circulating levels of CRP and the risk of breast cancer, including a recent meta-analysis (16), have primarily reported null results (13–17). While our results could reflect chance findings and require additional replication, the null results obtained previously could also, in part, reflect the inherent challenges of using CRP as a marker of long-term inflammation. CRP, a marker of acute-phase inflammatory response, is influenced by age, lifestyle factors, and underlying disease states. Thus, there is the potential for misclassification error and the attenuation of risk estimates should a single measure not well-represent long-term exposure. In addition, circulating CRP reflects systemic rather than local levels of inflammation. Should levels of inflammation directly in the breast tissue have more relevance in the etiology of breast cancer and be poorly correlated with a systemic measure of CRP, circulating CRP levels may be an inadequate proxy of the exposure of interest. Although the examination of local tissue biomarkers is challenging, it may be reasonable to use breast tissue from women undergoing clinical procedures, such as breast reduction surgery, for comparison. It also remains plausible that circulating CRP levels may partially reflect subclinical disease, and therefore discrepancies may arise from differences in the length of time between biospecimen collection and the date of diagnosis; however in our sample, no heterogeneity in the risk of postmenopausal breast cancer was detected for samples obtained between 1 and 5 years before the date of diagnosis. Future studies focusing on inflammatory markers measured in breast tissue may provide for additional insight into the role of the inflammatory process in the etiology and the progression of postmenopausal breast cancer.

There are several strengths to the current study, including the prospective design allowing for prediagnostic biomarker and covariate assessment, as well as the ethnic diversity and wide range of BMI values of the study sample, and the population-based sampling frame allowing for the generalizability of results. The final approximate response rates for cohort members by initially assigned ethnicity were 49% for Japanese Americans, 43% for whites, 39% for Native Hawaiians, 23% for African Americans, and 20% for Latinos (20). There are also limitations. First, similar to other studies in this area, only a single measurement for each biomarker was available for analysis. This may have resulted in the potential for misclassification error and the attenuation of risk estimates. However, we did have sufficient power to observe statistically significant main effects for the total study population and previous reports have shown CRP and adipokine levels to be relatively stable within healthy individuals over time (32–34). Second, BMI was self-reported and used as proxy for adiposity and may not adequately reflect levels of adipose tissue for specific subpopulations, including women in the upper age range of our study; however, no differences in the risk of disease were detected in analyses stratified by age at diagnosis. In addition, information on relevant anthropometric measures such as waist–hip circumference was not available. Third, we were unable to control for additional physiologic parameters in our analyses, such as circulating estrogen and insulin, possibly associated with both BMI and the risk of postmenopausal breast cancer. Fourth, no distinction between type and duration of hormone replacement therapy use was made when matching cases, information was not available on hysterectomy or oophorectomy status, and baseline data were collected several years before blood collection. Fifth, the statistical power was limited to detect associations when testing for interactions and for subgroup analyses.

In conclusion, our findings suggest that adipokines and systemic inflammation, as reflected by circulating CRP levels, are associated with the risk of postmenopausal breast cancer independently of BMI. Further prospective epidemiologic and laboratory-based experimental studies are needed to clarify the role of adipokines and CRP in the etiology and the progression of postmenopausal breast cancer, especially those considering heterogeneity in the associations by BMI and race or ethnicity.

No potential conflicts of interest were disclosed.

Conception and design: L.R. Wilkens, B.E. Henderson, L. Le Marchand, L.N. Kolonel, M.T. Goodman

Development of methodology: R.V. Cooney, L. Le Marchand, M.T. Goodman

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.A. Franke, R.V. Cooney, L. Le Marchand, L.N. Kolonel, M.T. Goodman

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N.J. Ollberding, Y. Kim, Y.B. Shvetsov, L.R. Wilkens, R.V. Cooney, G. Maskarinec, M.T. Goodman

Writing, review, and/or revision of the manuscript: N.J. Ollberding, Y. Kim, L.R. Wilkens, R.V. Cooney, G. Maskarinec, B.Y Hernandez, L. Le Marchand, L.N. Kolonel, M.T. Goodman

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Kim, A.A. Franke, B.Y Hernandez, L. Le Marchand, L.N. Kolonel, M.T. Goodman

Study supervision: M.T. Goodman

The authors thank all participants in the Multiethnic Cohort Study and Laurie Custer at the Analytical Biochemistry Share Resource of the University of Hawaii Cancer Center for carrying out the biochemical assays.

The Multiethnic Cohort Study has been supported by grants R37 CA 54281 and R01 CA 63464 from the National Cancer Institute. The SEER tumor registries in Hawaii and Los Angeles are supported by the NIH, Department of Health and Human Services (contracts N01-PC-35137 and N01-PC-35139, respectively). N.J. Ollberding was supported by a postdoctoral fellowship on grant R25 CA 90956. The Analytical Biochemistry Shared Resource of the University of Hawaii Cancer Center in supported, in part, by grant P30-CA71789 from the National Cancer Institute.

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.