Obesity and metabolic syndrome are associated with short-term endocrine therapy resistance in early ER + breast cancer

Abstract

Purpose: Increased body mass index (BMI) and metabolic syndrome (MS) are associated with increased breast cancer recurrence risk. Whether this is due to intrinsic tumor biology or modifiable factors of the obese state remains incompletely understood.

Methods: Oncotype DX Recurrence Scores of 751 patients were stratified by BMI to assess association with tumor-intrinsic recurrence risk. Cellular proliferation by Ki67 after 10-21 days of presurgical letrozole treatment was used to stratify endocrine therapy response (sensitive-ln(Ki67) < 1; intermediate-ln(Ki67)1-2; resistant-ln(Ki67) > = 2). BMI at the time of surgery and MS variables were collected retrospectively for 143 patients to analyze association between therapy response and BMI/MS. Additionally, PI3K pathway signaling was evaluated by immunohistochemistry of phosphorylated Akt and S6.

Results: There was no significant association between BMI and recurrence score (p = 0.99), and risk score distribution was similar across BMI groups. However, BMI was associated with short-term endocrine therapy resistance, with a significant enrichment of intermediate and resistant tumors in patients with obesity (55%, p = 0.0392). Similarly, the relative risk of an endocrine therapy-resistant tumor was 1.4-fold greater for patients with MS (p = 0.0197). In evaluating PI3K pathway mediators, we found patients with 3 or more MS criteria had more tumors with pAkt scores above the median (p = 0.0436). There were no significant differences in S6 activation.

Conclusion: Our findings suggest the association between obesity/metabolic syndrome and breast cancer recurrence is better reflected by response to treatment than tumor-intrinsic properties, suggesting interventions to reverse obesity and/or MS may improve outcomes for breast cancer recurrence.

Keywords: Breast cancer; Endocrine therapy resistance; Metabolic syndrome; Obesity; Recurrence risk.

Fig 1. Number of cases available for analysis.

Cohort 1: Patients with Oncotype tests performed who had clinical data available; Cohort 2: Patients treated on a clinical trial of presurgical endocrine therapy [22] who had endocrine therapy response by Ki67 immunohistochemistry and clinical data available.

Fig 2. BMI is not associated with Oncotype Recurrence Score.

a. Oncotype Recurrence Scores plotted by BMI category; dotted lines show median and quartiles. b. Frequency distribution of Oncotype DX RS for each BMI category.

Fig 3. BMI is associated with resistance to short-term endocrine therapy.

Distribution of endocrine sensitive tumors according to BMI. Endocrine sensitivity was determined from tumor proliferation in the post-treatment sample (Ki67 staining) after 10–21 days of endocrine therapy prior to surgery as previously described [22]. Sensitive tumors were defined by ln(Ki67) < 1, intermediate as ln(Ki67) 1–2, and resistant as ln(Ki67) >= 2. Pre-surgery weights were used to determine BMI for n=147 patients; p-values are displayed for chi-square contingency testing among the 3 groups and between overweight (BMI 25–30) vs obese (BMI >= 30).

Fig 4. Metabolic syndrome is associated with higher risk of resistant disease.

a. Distribution of endocrine sensitive tumors according to the presence of metabolic syndrome (defined as 3 or more criteria). Relative risk is the risk of an intermediate or resistant tumor associated with the presence of metabolic syndrome, p-value by Fisher’s exact t-test. b. Distribution of sensitive and intermediate/resistant tumors by the number of metabolic syndrome criteria present.

Fig 5. Metabolic syndrome is associated with more Akt activation in breast tumors.

Immunohistochemistry with pAkt and pS6 antibodies on a TMA containing post-treatment tumor samples was scored by automated microscopy. a, d. Violin plots display H-scores of pAkt (a) or pS6 (d) from tumors in patients without or with 3 or more metabolic syndrome criteria. b, e. Distribution of tumors with scoring above or below the median pAkt (b) or pS6 (e) scores are shown, p by Fisher’s exact t-test. c, f. Distribution of pAkt (c) and pS6 (f) scores according to endocrine sensitivity and the presence of metabolic syndrome criteria. Median and quartiles are shown as dashed and dotted lines.