SMARCE1 regulates metastatic potential of breast cancer cells through the HIF1A/PTK2 pathway

Abstract

Background: While aberrant activation of the chromatin-remodeling SWI/SNF complexes has been associated with cancer development and progression, the role of each subunit in tumor cells is poorly defined. This study is aimed to characterize the role of SMARCE1/BAF57 in regulating metastasis of breast cancer cells.

Methods: Genetic approaches and chemical inhibitors were used to manipulate the activities of SMARCE1 and its downstream targets in multiple breast cancer cell lines. Xenograft mouse models were used to analyze the role of SMARCE1 in lung metastasis in vivo. Nonadherent culture conditions were used to elucidate the role of SMARCE1 in regulating anoikis. Chromatin immunoprecipitation (ChIP), immunoprecipitation, and immunoblotting assays were designed to dissect the mechanism of action of SMARCE1. Public databases were used to investigate the relationship between SMARCE1 deregulation and breast cancer prognosis.

Results: SMARCE1 knockdown reduced lung metastasis of breast cancer cells and sensitized tumor cells to anoikis. In response to loss of attachment, SMARCE1 interacted with and potentiated transcriptional activity of HIF1A, resulting in rapid PTK2 activation. Both HIF1A and PTK2 were indispensable for SMARCE1-mediated protection against anoikis by promoting activation of ERK and AKT pathways while suppressing the expression of pro-apoptotic BIM protein. Expression data analysis of a large cohort of human breast tumors revealed that high expression of SMARCE1 or PTK2 is associated with poor prognosis and tumor relapse, and PTK2 expression is positively correlated with SMARCE1 expression in basal-like and luminal B subtypes of breast tumors.

Conclusions: SMARCE1 plays an essential role in breast cancer metastasis by protecting cells against anoikis through the HIF1A/PTK2 pathway. SMARCE1-mediated PTK2 activation likely plays a key role in promoting metastasis of basal-like and luminal B subtype of breast tumors.

Keywords: Anoikis; Breast cancer; HIF1A; PTK2/FAK; SMARCE1/BAF57.

Fig. 1

SMARCE1 knockdown reduces lung metastasis of breast cancer in vivo. a Expression levels of SMARCE1 mRNA and protein in LM-EV and LM-SMARCE1-KD cells. b Effect of SMARCE1 knockdown on the growth of xenografts in the fourth inguinal mammary fat pads of female NSG mice. c Spontaneous lung metastasis from orthotopic sites. Metastatic foci of tumor cells expressing red fluorescent protein on the dorsal surface of the left lung lobe were imaged 38 days after tumor cell inoculation (upper panel). The presence of tumor cells in the lungs was visualized by hematoxylin and eosin (H&E) staining of formalin-fixed lung sections (10 μM) and quantified by ImageJ software (lower panel). d Number of circulating tumor cells in blood. Fluorescent tumor cells in mouse blood were isolated and counted 38 days after tumor cell inoculation. EV empty vector, KD knockdown, LM lung metastatic cell line derived from MDA-MB-231

Fig. 2

SMARCE1 knockdown reduces lung colonization of tumor cells inoculated through tail veins. a Number of circulating tumor cells in blood collected at various times after tail vein injection in NSG mice. b Fluorescent tumor cells in lungs of NSG mice 72 h after tail vein injection. Representative images of five lungs for each group were shown. c Fluorescent tumor foci in the left lung lobes of NSG mice 4 weeks after tail vein injection of tumor cells. The area of tumor foci on the dorsal surface of the left lung lobe was quantified by ImageJ

Fig. 3

SMARCE1 knockdown promotes cell migration, but sensitize cells to anoikis. a Effect of SMARCE1 knockdown in LM cells on cell migration and invasion. Boyden chambers, uncoated or coated with Matrigel were used to measure cell migratory and invasive potential, respectively. b Effect of SMARCE1 knockdown in LM cells on viability of cells cultured in dishes coated with polyHEMA to prevent adhesion. c Effect of SMARCE1 knockdown in HCC38 cells. d SMARCE1 overexpression in BT549 cells. e Blocking SMARCE1 expression by siRNA enhanced detachment-induced caspase activation in LM and HCC38 cells. Accell SMARTpool of SMARCE1-siRNA (1 μM) were delivered into cells 48 h prior to detachment. Caspase activities were determined at indicated time points after cell detachment by using the Caspase-Glo 3/7 assay kit. *p < 0.05, Student t test. EV empty vector, KD knockdown, LM lung metastatic cell line derived from MDA-MB-231

Fig. 4

SMARCE1-mediated anoikis resistance requires activation of PTK2, ERK, and AKT pathways. a Viability of LM-EV and LM-SMARCE1-KD cells cultured in polyHEMA-coated dishes in the absence or presence of indicated inhibitors. b Viability of BT549-EV and BT549-SMARCE1 cells cultured in polyHEMA-coated dishes in the absence or presence of indicated inhibitors. For all experiments, viable cells were counted after 7 days in suspension culture. Concentration of inhibitors: 0.5 μM PF562271, 10 μM U0126, 10 μM PD98059, 30 μM ZSTK474, 3 μM dasanitib, 5 μM DAPT, and 10 μM LY2109761. *p < 0.05, Student t test. EV empty vector, KD knockdown, LM lung metastatic cell line derived from MDA-MB-231

Fig. 5

SMARCE1-dependent activation of PTK2, ERK, and AKT pathways by cell detachment. a Effect of SMARCE1 knockdown in LM cells on detachment-induced activation of pro-survival (FAK/PTK2, ERK, and AKT) and pro-apoptotic (BIM) signaling pathways. b Effect of SMARCE1 knockdown in HCC38 cells. c Effect of SMARCE1 overexpression in BT549 cells. The levels of protein expression and phosphorylation were determined by immunoblotting assays. GAPDH was included as a loading control. EV empty vector, KD knockdown, LM lung metastatic cell line derived from MDA-MB-231

Fig. 6

SMARCE1 collaborates with HIF1A to activate PTK2 transcription in detached cells. a Detachment-induced PTK2 mRNA expression is abolished by SMARCE1 knockdown in LM and HCC38 cells, whereas enhanced by SMARCE1 overexpression in BT549 cells. b Detachment-induced recruitment of HIF1A and SMARCA4 to the promoter region of PTK2 is diminished by SMARCE1 knockdown in LM cells. Chromatin binding of HIF1A, SMARCE1, and SMARCA4 was measured by ChIP-qPCR assay. Average fold enrichment of PTK2 promoter DNA by indicated antibodies (vs. control IgG) from three independent experiments were presented. c Detachment induces interaction between SMARCE1 and HIF1A in LM-EV and LM-SMARCE1-KD cells. The protein interaction was examined by immunoprecipitation/immunoblotting assays. *p < 0.05, Student t test. EV empty vector, KD knockdown, LM lung metastatic cell line derived from MDA-MB-231

Fig. 7

SMARCE1 orchestrates chromatin remodeling of PTK2 promoter in detached cells. a Overlapping primer designed for the nucleosome scanning assay and their amplicon sizes. The location of each primer set is given relative to the transcription start site (TSS). b Significant nucleosomal displacement is observed on PTK2 promoter (from −802 to −492 bp) in LM-EV cells after 0.5, 1, and 2 hours under suspension culture (vs. adherent cells), but not in LM-SMARCE1-KD cells. Results from three independent experiments were presented as means ± SE. EV empty vector, KD knockdown, LM lung metastatic cell line derived from MDA-MB-231

Fig. 8

HIF1A activates PTK2 transcription and downstream survival pathways in detached cells. a Detachment induced HIF1A protein accumulation is eliminated by HIF1A knockdown in MDA-MB-231 cells. b HIF1A knockdown abolishes detachment induced PTK2 mRNA expression in MDA-MB-231 cells. c HIF1A knockdown sensitizes MDA-MB-231 cells to anoikis. Viable cells in suspension culture for 10 days were counted. d Effect of HIF1A knockdown in MDA-MB-231 cells on detachment-induced activation of pro-survival (PTK2, ERK, and AKT) and pro-apoptotic (BIM) signaling pathways. e Effect of PTK2 inhibition in MDA-MB-231 cells on detachment-induced activation of pro-survival and pro-apoptotic signaling pathways. The mRNA levels were measured by qPCR. The levels of protein expression and phosphorylation were determined by immunoblotting assays with GAPDH included as a loading control. *p < 0.05, Student t test. EV empty vector, KD knockdown, LM lung metastatic cell line derived from MDA-MB-231

Fig. 9

PTK2 activation is essential for SMARCE1-mediated anoikis resistance in breast cancer cells. a Expression levels of PTK2 mRNA in LM derivatives. b PTK2 protein expression and phosphorylation in LM derivatives cultured in polyHEMA-coated dishes. c Effect on PTK2 overexpression on viability of LM derivatives cultured in polyHEMA-coated dishes. d Effect on PTK2 overexpression on caspase activation in LM derivatives cultured in polyHEMA-coated dishes. Caspase activity was normalized to cell numbers and results from three independent experiments were presented as means ± SE. e Expression levels of PTK2 mRNA in BT549 derivative. f Effect of PTK2 knockdown on viability of BT549 derivative cultured in polyHEMA-coated dishes. *p < 0.05, Student t test. EV empty vector, KD knockdown, LM lung metastatic cell line derived from MDA-MB-231

Fig. 10

Blocking SMARCE1 expression sensitizes non-tumorigenic mammary epithelial cells to anoikis. a Effect of SMARCE1 knockdown by stable shRNA expression on viability of 184B5 cells cultured in polyHEMA-coated dishes. b Effect of stable shRNA expression (SMARCE1-KD) and SMARCE1-siRNA delivery (SMARCE1-i) on caspase activation in 184B5 cells cultured in polyHEMA-coated dishes. Caspase activity was measured by using the Caspase-Glo 3/7 assay kit and normalized to cell numbers. The results from three independent experiments were presented as means ± SE. c Detachment-induced PTK2 mRNA expression is abolished by SMARCE1-KD or SMARCE1-i in 184B5 cells. mRNA levels were measured by qPCR. EV empty vector, KD knockdown

Fig. 11

Higher SMARCE1 and PTK2 expression is associated with poor prognosis of breast cancer patients. a Kaplan-Meier analysis of recurrence-free survival (RFS) of breast cancer patients based on SMARCE1 and PTK2 expression using a dataset of 3554 breast tumor samples (http://kmplot.com/analysis/). To construct Kaplan-Meier curves all percentiles between the lower and upper quartiles were computed and the best performing threshold was used as cutoff to divide the patient cohort. The log-rank test p value was calculated by using RFS up to 15 years as end point. b Correlated expression of SMARCE1 and PTK2 in basal-like and luminal B subtypes of breast cancer according to the TCGA database. Tumors with PTK2 amplification were excluded from the correlation analysis. c Frequency of genetic alterations of SWI/SNF complex subunits in breast tumors according to the TCGA database