Liver X receptors induce antiproliferative effects in basal-like breast cancer

Abstract

Liver X receptors (LXRs) are nuclear transcription factors important in the regulation of cholesterol transport, and glucose and fatty acid metabolism. The antiproliferative role of LXRs has been studied in a variety of malignancies and may represent a therapeutic opportunity in cancers lacking targeted therapies, such as triple-negative breast cancer. In this study, we investigated the impact of LXR agonists alone and in combination with carboplatin in preclinical models of breast cancer. In vitro experiments revealed a dose-dependent decrease in tumor cell proliferation in estrogen receptor-positive breast cancer cells, whereas LXR activation in vivo resulted in an increased growth inhibitory effect in a basal-like breast cancer model (in combination with carboplatin). Functional proteomic analysis identified differences in protein expression between responding and nonresponding models related to Akt activity, cell-cycle progression, and DNA repair. Furthermore, pathway analysis suggested that the LXR agonist in combination with carboplatin inhibits the activity of targets of E2F transcription factors and affects cholesterol homeostasis in basal-like breast cancer.

Keywords: LXR; PDX; RPPA; basal-like breast cancer.

Fig. 1

Effects of LXR activation in luminal and basal‐like breast cancer cell lines. A colorimetric cell proliferation assay (MTS) was used to detect the inhibitory effect of increasing concentrations of the LXR agonist GW3965 after 7 days in (A) T47D (n = 3) (luminal cell line), (B) MDA‐MB‐231 (n = 3) and (C) MDA‐MB‐468 (n = 3) (both basal‐like cell lines). All proliferation assays are presented as percentage of vehicle (DMSO) and error bars represent SEM. *Student's t‐test P < 0.05. (D–F) Relative mRNA expression of target genes (ABCG1 and SREBP1c) on treatment with 10 μm GW3965 (n = 2) compared with DMSO (n = 1) in T47D, MDA‐MB‐231 and MDA‐MB‐468, respectively. The relative expression of DMSO treatment was set to 1.0 and data presented as mean.

Fig. 2

LXR activation in patient‐derived xenograft models of basal‐like breast cancer. Growth curves showing the effect of the LXR agonist GW3965 (GW) in combination with carboplatin, presented as a median relative tumor volume in (A) MAS98.12 and (B) HBCx39. The addition of GW3965 to carboplatin resulted in a significant reduction in tumor growth in MAS98.12 *P < 0.05 (Wilcoxon's P = 0.022), but not in HBCx39 (Wilcoxon's P = 0.089). Number of tumors in each group MAS98.12: CTRL (n = 19), GW (n = 16), Carboplatin (n = 12), Combination (n = 13) HBCx39: CTRL (n = 10), GW (n = 7), Carboplatin (n = 11), Combination (n = 5). (C, D) LXR activation by GW3965 induced the expression of target genes in both tumor models. Number of tumors in each group MAS98.12: CTRL (n = 7), GW (n = 8), Carboplatin (n = 4), Combination (n = 8) HBCx39: CTRL (n = 10), GW (n = 6), Carboplatin (n = 12), Combination (n = 7). (E) Gene expression analysis detected a higher expression of LXRα encoded by NR1H3 and LXRβ encoded by NR1H2 in MAS98.12 compared with HBCx39 (data derived from RNA expression array analysis of one tumor from each tumor model). CTRL, control; Cbx, carboplatin; GW, GW3965; Combination/Cbx + GW, carboplatin and GW3965.

Fig. 3

Protein expression by RPPAs in basal‐like patient‐derived xenograft (PDX) models HBCx39 compared with MAS98.12 before treatment. (A) Volcano plot demonstrating magnitude of fold‐change protein expression (log₂(FC); x‐axis) and statistical significance of this change (Student's t‐test P‐value; −log₁₀, y‐axis) in HBCx39 (n = 2) versus MAS98.12 (n = 2) untreated controls. Colored points represent differently expressed proteins in which green points have a significant P‐value (< 0.05) and |log₂(FC)| > 0.5, red points have a significant P‐value and |log₂(FC)| < 0.5, whereas the yellow and gray points have a nonsignificant (NS) P‐value and a |log₂(FC)| > 0.5 or < 0.5, respectively. (B) Boxplot demonstrating selected differently expressed proteins between the MAS98.12 and HBCx39 PDX models from same samples. *Student's t‐test unadjusted P < 0.05.

Fig. 4

Protein expression by RPPAs comparing carboplatin and GW3965 combination versus carboplatin alone‐treated MAS98.12. (A) Volcano plot demonstrating magnitude of fold‐change protein expression (log₂(FC); x‐axis) and statistical significance of this change (Student's t‐test P‐value −log₁₀; y‐axis) between samples treated with GW3965 and carboplatin combination (n = 3) versus carboplatin alone (n = 3) in MAS98.12. Colored points represent differently expressed proteins in which green points have a significant P‐value (< 0.05) and |log₂(FC)| > 0.5, red points have a significant P‐value and |log₂(FC)| < 0.5, whereas the yellow and gray points have a nonsignificant (NS) P‐value and |log₂(FC)| > 0.5 or < 0.5, respectively. (B) Boxplot demonstrating four of the differently expressed proteins between the same combination‐ and carboplatin‐treated samples *Student's t‐test unadjusted P < 0.05.

Fig. 5

Pathway activation in basal‐like breast cancer xenografts models before and after treatment with LXR agonist GW3965. Gene Set Variation Analysis (GSVA) was used to generate single sample scores for gene signatures in the Hallmark database base on the RPPA proteomic data. Differential expression analysis was then performed with the r package limma, and changes in pathway enrichment between the groups of comparison are displayed as volcano plot with magnitude of fold‐change pathway enrichment score (log₂FC; x‐axis) and statistical significance of this change (pairwise t‐test unadjusted P‐value −log₁₀; y‐axis). P‐values < 0.05 was considered significant (green). (A) Pathway enrichment comparing the two basal‐like models MAS98.12 (n = 2) and HBCx39 (n = 2) at baseline. (B) Pathway enrichment comparing the carboplatin and GW3965 combination (n = 3) versus carboplatin alone (n = 3) treated MAS98.12 tumor samples. Several Hallmark signatures showed significant changes in enrichment between combination and carboplatin, and single sample scores were used to assess distribution across all treatment groups in MAS98.12. (C) A distinct drop in enrichment of E2F genes, associated with proliferating cancer cells, was observed in the combination group (n = 3), but not in the other treatment groups (control; n = 2, GW and carboplatin; n = 3). (D) Cholesterol homeostasis‐associated genes showed increased enrichment in both groups treated with the LXR agonist GW3965 (n = 3), but not with carboplatin (n = 3) alone or in the control group (n = 2). *Pairwise t‐test adjusted (Bonferroni) P < 0.05. CTR, control; GW, GW3965; Combination, carboplatin and GW3965.

Fig. 6

Schematic illustration of downstream effects of LXR activation. GW3965 primarily activates LXR dimerization (with the corresponding Retinoid X receptor), dissociation from corepressor, and recruitment with coactivators. Activated LXR induces the expression of downstream target genes like ABCG1, SREBP1c, and FASN. Studies have also shown LXR‐mediated decrease in PI3K‐subunit phosphorylation and corresponding increase of PTEN activation, as well as modulation of AKT signaling by LXR agonist activation. Altogether, this results in reduced proliferation, growth, and survival of cells (Created with BioRender.com).