A pivotal role of Krüppel-like factor 5 in regulation of cancer stem-like cells in hepatocellular carcinoma

Abstract

In hepatocellular carcinoma (HCC), there exists a highly tumorigenic subset of cells defined by high expression of CD44 and CD133 that has been reported to contain cancer stem-like cells (CSCs). Krüppel-like factor 5 (KLF5) regulates many factors involved in cell cycle, migration, inflammation, angiogenesis and stemness and has cancer-promoting effects in some cancers. While some reports have indicated that KLF5 may have important roles in regulation of CSCs, what role, if any, KLF5 plays in regulation of CSCs in HCC remains to be elucidated. Flow cytometric analysis of CD44 and CD133 in HCC cell lines revealed subpopulations of CD44(High)/CD133(High) and CD44(Low)/CD133(Low) cells. We subsequently sorted these subpopulations and identified KLF5 as a gene that is significantly upregulated in CD44(High)/CD44(High) cells via RNA sequencing using next generation sequencing technology. Moreover, KLF5 overexpression enriched the CD44(High)/CD133(High) subpopulation and, consistent with the up-regulation of CD44(High)/CD133(High) cells, KLF5 overexpressing cells were more resistant to anti-cancer drugs and displayed enhanced colony-formation capacity. By contrast, knock-down of KLF5 by siRNA diminished the CD44(High)/CD133(High) subpopulation. When KLF5 was acetylated by TGF-β1, the KLF5-mediated CD44(High)/CD133(High) subpopulation enrichment was abrogated. Oppositely, ectopic expression of an acetylation-deficient KLF5 mutant further increased CD44(High)/CD133(High) subpopulations as compared to cell expressing wild-type KLF5. These findings provide novel mechanistic insight into a pivotal role for KLF5 in the regulation of CSCs in HCC.

Keywords: CD133; CD44; KLF5; acetylation; cancer stem cell; hepatocellular carcinoma.

Figure 1.

CD44^High/CD133^High cells in hepatoma cell lines. (A) Hepatoma cell lines, Huh7 and HepG2, were stained with anti-CD44 and anti-CD133 and analyzed by FACS. Cells expressing high CD44 and CD133 (CD44+/CD133+ cells) and those expressing low CD44 and CD133 (CD44−/CD133− cells) were sorted by FACS Aria III as gated. (B) Anchorage independent growth was examined by soft-agar colony formation assay with the sorted CD44+/CD133+ cells and CD44−/CD133− cells. (*P < 0.05 vs. CD44−/CD133− cells) (C) Sensitivity to anti-cancer drugs, CDDP and 5FU, was analyzed by MTS assay (* P < 0.05 vs. CD44−/CD133− cells).

Figure 2.

KLF5 is significantly upregulated in CD44+/CD133+ cells. RNA sequence was performed using sorted Huh7 cells. Results of principal component analysis (A) and volcano plot (B). (C) Expression levels of CD44, CD133 and KLF5 were validated by real-time RT-PCR in Huh7 and HepG2 cells (* P < 0.05 vs. CD44−/CD133− cells).

Figure 3.

Knockdown of KLF5 decreases CD44+/CD133+ population in hepatoma cell lines. (A) Two independent sequences of siRNA against KLF5 were used to knock-down KLF5 at the concentration of 5nM. Expression levels of KLF5 and CD44 were determined by real-time RT-PCR 48 hours after siRNA transfection. (* P <0.05 vs. scramble siRNA) (B) KLF5 was knocked-down by KLF5-A siRNA at a concentration of 5nM. 48 hours after siRNA transfection, the CD44+/CD133+ subpopulation was analyzed by FACS. A histogram shows relative fold change of CD44+/CD133+ cells in Huh7 and HepG2 cells. (* P<0.05 vs. scramble siRNA) (C) KLF5 was knocked-down by KLF5-A siRNA at a concentration of 5nM. 48 hours after siRNA transfection, sensitivity to 5FU was analyzed by MTS assay (* P<0.05 vs. siKLF5).

Figure 4.

Overexpression of KLF5 increases the CD44+/CD133+ subpopulation and renders cells more resistant to anti-cancer drugs in hepatoma cell lines. KLF5 was transduced by retrovirus-mediated gene transfer into Huh7 cells. Expression level of KLF5 was examined by real-time PCR (A) and western blot (B). (* P < 0.05 vs. empty) (C) The CD44+/CD133+ subpopulation was determined by FACS. A histogram shows relative fold change of CD44+/CD133+ cells in Huh7 cells. (* P < 0.05 vs. empty) (D) KLF5 overexpressing cells and empty control cells were subjected to soft-agar colony formation assay. Chemotherapeutic reagents, CDDP (10 μM) and 5FU (5 μM), were added to the overlaid medium. Histogram shows relative colony numbers (* P < 0.05 vs. empty control).

Figure 5.

TGF-β1 decreases the CD44+/CD133+ subpopulation independent of altered KLF5 expression. Control (Empty) and KLF5 overexpressing (KLF5) Huh7 cells were treated with TGF-β1 (5ng/ml) for 48 hours. (A) Expression of KLF5 was determined by WB. (B) Expression of CD44 was determined by real-time PCR (* P < 0.05 vs. Empty/TGF-β1 (−), # P < 0.05 vs. Empty/TGF-β1 (−), ** P < 0.05 vs. KLF5/TGF-β1 (−)). (C) Expression of CD44/CD133 and percentage of CD44+/CD133+ cells were determined by FACS.

Figure 6.

Acetylation status of KLF5 is critical for regulation of the CD44+/CD133+ subpopulation. Wild type (KLF5-WT) and acetylation-deficient (KLF5-K369R) KLF5 were transduced by retrovirus-mediated gene transfer into Huh7 cells. (A) Expression of KLF5 was determined by WB. (B) Expression of CD44 was determined by real-time PCR (* P < 0.05 vs. Empty, # P < 0.05 vs. KLF5-WT). (C) Expression levels of the known KLF5 target genes, Oct4, Pim1 and p15 were determined by real-time PCR (* P < 0.05 vs. Empty, # P < 0.05 vs. KLF5-WT). (D) Expression of CD44/CD133 and percentage of CD44+/CD133+ cells were determined by FACS.