Kruppel-like factor 8 regulates triple negative breast cancer stem cell-like activity

Abstract

Introduction: Breast tumor development is regulated by a sub-population of breast cancer cells, termed cancer stem-like cells (CSC), which are capable of self-renewing and differentiating, and are involved in promoting breast cancer invasion, metastasis, drug resistance and relapse. CSCs are highly adaptable, capable of reprogramming their own metabolism and signaling activity in response to stimuli within the tumor microenvironment. Recently, the nutrient sensor O-GlcNAc transferase (OGT) and O-GlcNAcylation was shown to be enriched in CSC populations, where it promotes the stemness and tumorigenesis of breast cancer cells in vitro and in vivo. This enrichment was associated with upregulation of the transcription factor Kruppel-like-factor 8 (KLF8) suggesting a potential role of KLF8 in regulating CSCs properties.

Methods: Triple-negative breast cancer cells were genetically modified to generate KLF8 overexpressing or KLF8 knock-down cells. Cancer cells, control or with altered KLF8 expression were analyzed to assess mammosphere formation efficiency, CSCs frequency and expression of CSCs factors. Tumor growth in vivo of control or KLF8 knock-down cells was assessed by fat-pad injection of these cell in immunocompromised mice.

Results: Here, we show that KLF8 is required and sufficient for regulating CSC phenotypes and regulating transcription factors SOX2, NANOG, OCT4 and c-MYC. KLF8 levels are associated with chemoresistance in triple negative breast cancer patients and overexpression in breast cancer cells increased paclitaxel resistance. KLF8 and OGT co-regulate each other to form a feed-forward loop to promote CSCs phenotype and mammosphere formation of breast cancer cells.

Discussion: These results suggest a critical role of KLF8 and OGT in promoting CSCs and cancer progression, that may serve as potential targets for developing strategy to target CSCs specifically.

Keywords: KLF8; O-GlcNAc; OGT; breast cancer; cancer stem cell; self-renewal.

Figure 1

KLF8 inhibition reduced mammosphere formation and CSCs population in TNBC cells. (A) Lysates from MDA-MB-231 cells stably expressing control or KLF8 shRNA was collected for immunoblot analysis using antibodies against indicated proteins (top-left). MDA-MB-231 cells transduced with control or KLF8 shRNA were grown in mammosphere formation assay for 5-7 days. Representative images of mammosphere were taken (top-right) (scale bar 400 μm), and mammosphere larger than 50µm were counted and primary mammosphere formation efficiency was quantified by fractionating counted mammosphere number from total cells plated at the beginning of the assay. Primary mammosphere were then collected and culture again in mammosphere culture condition for 5-7 days to form secondary mammosphere (bottom). (B) Quantified graph of ALDH+ CSCs population detected by flow cytometry from MDA-MB-231, SUM159 and PDX cells HCI-10 stably expressing control or KLF8 shRNA. (C) Quantified graph showing NANOG-GFP+ CSCs population detected by flow-cytometry from MDA-MB-231 and SUM159 cells stably expressing control or KLF8 shRNA. Student t test reported as mean ± SEM, *p<0.05, **p<0.01, and ***p<0.001.

Figure 2

Elevated KLF8 expression increased mammosphere formation and CSCs population in TNBC cells. (A) Lysates from MDA-MB-231 cells stably overexpressing control or KLF8 were collected for immunoblot analysis using antibodies against indicated antibodies (left). MDA-MB-231 cells overexpressing control or KLF8 was grown in mammospheres formation assay for 5-7 days. Representative images of mammosphere were taken (middle) (scale bar 400 μm) and mammosphere larger than 50µm were counted and mammosphere formation efficiency was quantified and graphed (right). (B) Lysates from SUM159 cells overexpressing control or KLF8 were collected for immunoblot analysis using indicated antibodies (left). SUM159 cells overexpressing control or KLF8 was grown in mammospheres formation assay for 5-7 days. Representative images of mammosphere were taken (middle) (scale bar 400 μm) and mammosphere larger than 50µm were counted and mammosphere formation efficiency was quantified and graphed (right). (C) Quantified graph of ALDH+ CSCs population detected by flow cytometry from MDA-MB-231 and SUM159 cells overexpressing control or KLF8. (D) Quantified graph of NANOG-GFP+ CSCs population detected by flow cytometry from MDA-MB-231 and SUM159 cells overexpressing control or KLF8. Student t test reported as mean ± SEM, *p<0.05, **p<0.01, and ***p<0.001.

Figure 3

KLF8 regulates expression of stem cell markers in breast cancer cells. (A) Lysates from MDA-MB-231 cells stably overexpressing control or KLF8 were collected for immunoblot analysis of CSCs markers using indicated antibodies (left), and quantified graph of relative level of CSCs detected by immunoblot from MDA-MB-231 cells control or with KLF8 overexpression (right). (B) Quantified graph showing relative mRNA level of CSCs markers as detected by qRT-PCR using probes against genes in MDA-MB-231 cells stably overexpressing control or KLF8. (C) Lysates from MDA-MB-231 cells stably expressing control or KLF8 shRNA were collected for immunoblot analysis of CSCs markers using indicated antibodies (left), and quantified graph of relative level of CSCs levels (right). (D) Quantified graph showing relative mRNA level of CSCs markers as detected by qRT-PCR using probes against indicated targets in MDA-MB-231 cells stably expressing control or KLF8 shRNA. Student t test reported as mean ± SEM, *p<0.05.

Figure 4

KLF8 promotes resistance to paclitaxel in breast cancer cells in vitro. (A) Data-based analysis showing KLF8 expression in breast cancer patients with differential response to taxane treatment. (B) Data-based analysis showing ROC-plot of KLF8 expression in different groups of breast cancer patients treated with taxane. (C) Lysates of MDA-MB-231 cells stably overexpressing control or with KLF8 were collected for immunoblot using indicated antibodies (top-left). MDA-MB-231 cells stably overexpressing control or KLF8, treated with increasing dose of paclitaxel for 48h, were grown in clonogenic assay for 10-14 days. Colonies were stained, counted. Representative images show stained colonies formed in the clonogenic assay after 10-14 days from MDA-MB-231 cells overexpressing control or KLF8, treated with increasing dose of paclitaxel for 48h (top-right). Quantified graph of counted colonies after 10-14 days from MDA-MB-231 cells overexpressing control or KLF8, treated with increasing dose of paclitaxel for 48h, in clonogenic assay (bottom). (D) Quantified graph showing fraction of apoptotic cells from MDA-MB-231 cells overexpressing control or KLF8, treated with increasing dose of paclitaxel for 48h. Two-way ANOVA with Holm-Sidak test reported as mean ± SEM, *p<0.05, and **p<0.01.

Figure 5

KLF8 and OGT/O-GlcNAc formed a potential feed-forward loop in breast cancer cells. (A) Lysates of MDA-MB-231 cells overexpressing control or OGT, transduced with control or KLF8 shRNA, were collected for immunoblot analysis using antibodies against indicated targets (B) Quantified graph showing mammosphere formation efficiency of MDA-MB-231 cells overexpressing control or OGT, transduced with control or KLF8 shRNA, after 5-7 days. (C) Lysates of MDA-MB-231 cells overexpressing control or OGT were collected for immunoblot analysis using indicated antibodies (top). Quantified graph showing relative mRNA level, detected by qRT-PCR, of KLF8 and OGT from MDA-MB-231 cells overexpressing control or OGT in mammosphere formation assay. (D) Lysates of MDA-MB-231 cells stably expressing control or KLF8 shRNA were collected for immunoblot analysis using indicated antibodies (top). Quantified graph showing relative mRNA level, detected by qRT-PCR, of KLF8 and OGT from MDA-MB-231 cells stably expressing control or KLF8 shRNA. (E) Lysates of MDA-MB-231 cells stably overexpressing control or KLF8, transduced with control of OGT shRNA, were collected for immunoblot analysis using indicated antibodies. (B) Quantified graph showing mammosphere formation efficiency of MDA-MB-231 cells overexpressing control or KLF8, transduced with control of OGT shRNA, after 5-7 days in mammosphere formation assay. Student t test reported as mean ± SEM, *p<0.05.

Figure 6

Reduced KLF8 expression impaired breast tumor growth in vivo. (A) Representative images of tumors in immune-compromised mice, formed from MDA-MB-231 cells stably expressing luciferase, transduced with control or KLF8 shRNA 1 day and 6 weeks after injection (left). Quantified graph showing bioluminescence signal from tumors formed from MDA-MB-231 cells stably expressing luciferase, transduced with control (n=4) or KLF8 (n=4) shRNA, 6 weeks after injection (right). (B) Representative images of tumors formed from MDA-MB-231 cells stably expressing luciferase, transduced with control or KLF8 shRNA, taken from mammary fat pad of immune-compromised mice 6 weeks after injection (left). Quantified graph showing weight of tumors formed from MDA-MB-231 cells stably expressing luciferase, transduced with control (n=4) or KLF8 (n=4) shRNA (middle). Quantified graph showing volume of tumors formed from TNBC cells MDA-MB-231 stably expressing luciferase, transduced with control (n=4) or KLF8 (n=4) shRNA (right). Student t test reported as mean ± SEM, *p<0.05.