Metabolic and transcriptional effects of bazedoxifene/conjugated estrogens in a model of obesity-associated breast cancer risk

Abstract

Many risk-eligible women refuse tamoxifen for primary prevention of breast cancer due to concerns about common side effects such as vasomotor symptoms. Tamoxifen may also induce or worsen insulin resistance and hypertriglyceridemia, especially in women with obesity. The combination of bazedoxifene and conjugated estrogens (BZA/CE) reduces vasomotor symptoms and is currently undergoing evaluation for breast cancer risk reduction. However, the impact of BZA/CE on insulin resistance and metabolic health, particularly in those with excess adiposity, is understudied. Here, we examined the effects of obesity on response to BZA/CE in a rat model of breast cancer risk using older ovary-intact rats. Female Wistar rats received carcinogen to increase mammary cancer risk and were fed a high-fat diet to promote obesity. Lean and obese rats were selected based on adiposity, and then randomized to BZA/CE or vehicle for 8 weeks. BZA/CE reduced adiposity, enriched small (insulin-sensitive) mammary adipocytes, increased the abundance of beneficial metabolic gut microbes (Faecalbaculum rodentium and Odoribacter laneus), and reversed obesity-associated changes in lipids and adipokines. BZA/CE also reversed obesity-induced mammary enrichment of cell proliferation pathways, consistent with risk-reducing effects. Together, these data support the use of BZA/CE to improve metabolic health and reduce breast cancer risk in individuals with obesity.

Keywords: Adipose tissue; Breast cancer; Metabolism; Obesity; Oncology.

Figure 1. BZA/CE decreases body weight, adiposity, and food intake in lean and obese rats.

(A) Study design diagram depicting times at which rats were given high-fat diet (HF), carcinogen (N-methyl-N-nitrosourea, MNU), and randomized to BZA/CE treatment or vehicle control. (B) Cumulative weight gain across the 8-week treatment period. (C) Percentage body fat across the treatment period, measured by qMR. (D–F) Mass of mammary, gonadal, and visceral adipose depots at the end of study. (G) Total lean mass, as measured by qMR, at the end of study. (H and I) Energy intake (kcal) and feed efficiency (calculated as mg of weight gained per kcal consumed) across the 8-week treatment period. Data in line graphs represent mean ± SEM; data in box-and-whisker plots represent mean ± interquartile range. Data were analyzed by 2-way ANOVA, with main effects of adiposity, treatment, and their interaction. For all graphs, n = 6 lean control, 8 lean BZA/CE, 8 obese control, and 9 obese BZA/CE rats.

Figure 2. BZA/CE improves markers of insulin resistance and metabolic health.

(A) Glucose and (B) insulin were measured in end-of-study plasma samples and used to calculate (C) HOMA-IR. Plasma levels of (D) cholesterol, (E) non-esterified free fatty acids (NEFAs), and (F) triglycerides (TG) were measured at end-of-study plasma using colorimetric assays. (G) Liver mass and (H) percentage fat in liver (by qMR) at end of study. Plasma levels of (I) estradiol and (J) BZA measured in end-of-study plasma using ELISA and mass spectrometry, respectively. All data are presented as mean ± interquartile range and were analyzed by 2-way ANOVA, with main effects of adiposity, treatment, and their interaction. In A–H, n = 5–6 lean control, 8 lean BZA/CE, 8 obese control, and 8–9 obese BZA/CE rats. In J, n = 3 lean control, 7 lean BZA/CE, 3 obese control, and 9 obese BZA/CE.

Figure 3. BZA/CE increases adiponectin/leptin ratio and reduces the size of mammary adipocytes.

(A) Plasma leptin, (B) adiponectin, and (C) the ratio of adiponectin to leptin in end-of-study plasma (n = 6 lean control, 8 lean BZA/CE, 7–8 obese control, and 8 obese BZA/CE rats). (D) Cell size distribution of subcutaneous/mammary adipocytes in lean and obese BZA/CE or control rats at the end of study (n ≥ 1200 cells per tissue, with n = 6 lean control, 6 lean BZA/CE, 8 obese control, and 8 obese BZA/CE rats per group). Differences in adipocyte cell frequency distribution were examined by χ² analysis. (E) Mean adipocyte diameter of cells in subcutaneous/mammary depots at the end of study (n = 6 lean control, 6 lean BZA/CE, 8 obese control, and 8 obese BZA/CE rats per group). Box-and-whisker plots indicate mean ± interquartile range and were analyzed by 2-way ANOVA, with main effects of adiposity, treatment, and their interaction. For C, Bonferroni-adjusted pairwise comparisons are also shown.

Figure 4. Impact of BZA/CE on the gut microbiome.

(A) Principal coordinate analysis (PCoA) of β-diversity (Bray-Curtis dissimilarity index) showing no difference between BZA/CE-treated versus control rats. (B) Proportional abundance of each microbiome phylum in control and BZA/CE-treated rats. Bar graph indicates mean ± SEM. (C) Relative proportional abundance of bacterial species in control and BZA/CE-treated rats. Each colored box in the bar graph represents a bacterial taxon and the height of the box represents the relative abundance of that bacteria within the sample. “Other” represents lower abundance taxa. (D) Proportional abundance of key bacterial species known to be linked to obesity/metabolic health that were also altered with BZA/CE treatment. Box-and-whisker plots indicate mean ± interquartile range, and were analyzed by Mann-Whitney test. *P < 0.05, **P < 0.01. (E) Correlation between the proportional abundance of Faecalbaculum rodentium and end-of-study plasma leptin and percentage body fat; Spearman’s correlation coefficients are shown. All analyses represent n = 13–14 control and 16 BZA/CE-treated rats.

Figure 5. Impact of BZA/CE on mammary gland gene expression.

Mammary glands from lean and obese, untreated, and BZA/CE-treated rats were subjected to gene expression microarray analysis. (A) GSEA MSigDB Hallmark gene sets significantly enriched/decreased in obese untreated rats compared with lean untreated. (B) GSEA MSigDB Hallmark gene sets significantly enriched/decreased in BZA/CE-treated rats compared with untreated controls, with lean shown in blue, obese shown in red. The size of each symbol indicates the level of significance. NES, normalized enrichment score; significance was defined as an NES > 1.5 with FDR q < 0.05. Open circles are provided for comparison purposes, indicating the direction and fold change of genes that did not meet our predefined levels of significance. n = 5 samples per group (lean control, lean + BZA/CE, obese control, and obese + BZA/CE).

Figure 6. Differentially expressed genes in mammary glands of lean and obese rats with and without BZA/CE treatment.

(A) Heatmap of expression of leading-edge genes from proliferation-related pathways from Figure 5B. (B) Left: Venn diagram showing the number of differentially expressed genes in the mammary gland of obese versus obese BZA/CE compared with lean versus lean BZA/CE. Right: Heatmap showing expression of 13 genes that were differentially expressed in both obese versus obese BZA/CE and lean versus lean BZA/CE. n = 5 samples per group (lean control, lean BZA/CE, obese control, and obese BZA/CE).