& 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.

00

Oncogene (2007) 26, 456461

www.nature.com/onc

SHORT COMMUNICATION

KLF8 transcription factor participates in oncogenic transformation

X Wang and J Zhao
Center for Cell Biology and Cancer Research, Albany Medical College, Albany, NY, USA

Kruppel-like factor 8 (KLF8) is a member of the family of KLF transcription factors. Several KLF members have been shown to play a role in oncogenesis. We have previously demonstrated that KLF8 mediates cell cycle progression downstream of focal adhesion kinase (FAK) by upregulating cyclin D1. FAK plays a critical role in transformation and tumorigenesis and is aberrantly upregulated in many types of human cancer. Little is known about the function of KLF8 in these regards. Here we provide evidence suggesting a novel role of KLF8 in oncogenic transformation. We show that KLF8 expression is elevated in several types of human cancer cells and primary tumor tissues. Induced expression of ectopic KLF8 causes serum-independent growth and morphological transformation in NIH3T3 cells and enhances anchorage-independent growth of v-Src-transformed cells. In contrast, expression of a dominant-negative mutant of KLF8 dramatically suppresses the transformed phenotypes induced by v-Src. In addition, the KLF8-enhanced transformation in the v-Src cells was prevented by ablating cyclin D1 expression. Overall, these results indicate that KLF8 is required for v-Src-induced transformation and may play a role in tumor progression of human cancer. Oncogene (2007) 26, 456461. doi:10.1038/sj.onc.1209796; published online 10 July 2006 Keywords: KLF8; transformation; v-Src

Kruppel-like factor 8 (KLF8) belongs to the family of Kruppel-like transcription factors (van Vliet et al., 2000). It shares the well-conserved DNA-binding zincnger domains on its C-terminus. The N-terminal half of KLF8 is thought to determine its functional specicity through recruiting other proteins (Bieker, 2001). Several members of KLF family proteins, including KLF4 (Zhao et al., 2004; Ghaleb et al., 2005), KLF5 (Dong, 2001; Chen et al., 2002; Nandan et al., 2004; Ghaleb et al., 2005), KLF6 (Chen et al., 2003; Kimmelman et al., 2004; Reeves et al., 2004) and

Correspondence: Dr J Zhao, MS338, Center for Cell Biology and Cancer Research, Albany Medical College, 47 New Scotland Avenue, Mail Code 165, Albany, NY 12208, USA. E-mail: zhaojh@mail.amc.edu Received 6 February 2006; revised 23 May 2006; accepted 5 June 2006; published online 10 July 2006

KLF11 (Ellenrieder et al., 2004), have been implicated to be associated with various types of human cancer. We have previously identied KLF8 as an essential effector downstream of focal adhesion kinase (FAK) for cell cycle regulation by activating cyclin D1 gene promoter (Zhao et al., 2003). In addition to physiological role in regulating cell proliferation, survival and migration downstream of adhesion receptors and mitogenic receptors (Schlaepfer et al., 1999), FAK also plays a critical role in oncogenic transformation and tumor progression when deregulated. For example, ectopic expression of constitutively active FAK mutant makes MadinDarby canine kidney (MDCK) cells capable of anchorage-independent growth and forming tumors in nude mice (Frisch et al., 1996). Indeed, FAK is overexpressed or functionally hyperactivated in many types of human tumors (Gabarra-Niecko et al., 2003), including ovarian (Sood et al., 2004; Grisaru-Granovsky et al., 2005) and breast (Cance et al., 2000) carcinomas. Conditional deletion of FAK gene from mouse epidermal keratinocytes prevents carcinogen-induced skin tumor formation (McLean et al., 2004). Inhibition of FAK in rat mammary adenocarcinoma cells suppresses the cancer cell-derived tumor growth as well as lung metastasis (van Nimwegen et al., 2005). These results prompted us to study a potential role of KLF8 in malignant transformation. We rst screened various types of human cancer cell lines to determine whether KLF8 expression is deregulated in the cells. Using quantitative real-time polymerase chain reaction and Western analysis, we found that KLF8 is barely expressed in normal human epithelial cells but highly overexpressed in several types of cancer cell lines established from human patients, including ovarian, breast and renal carcinomas (Figure 1a and b). Consistently, others have demonstrated similar results in the NCI-60 cell lines by cDNA microarray screening (http://smd.stanford.edu/cgi-bin/source/expressionSearch? option cluster&criteria Hs.141120&dataset 2& organism Hs). Immunouorescent staining conrmed the nuclear expression of KLF8 in the ovarian cancer cell line SKOV3ip1 (Figure 1c). Finally, we found that KLF8 is overexpressed in all the human ovarian cancer tissues tested compared to the normal ovarian tissue counterparts in the Human Ovarian Cancer Gene Expression cDNA Panels (OriGene Technologies Inc., Rockville, MD, USA) (Figure 1d). These results strongly suggest a role for KLF8 in the pathogenesis of human cancer.

KLF8 mediates v-Src-induced transformation X Wang and J Zhao

457

Figure 1 KLF8 is overexpressed in human cancer cell lines and primary tumors. (a, b) Total RNA or protein extract prepared from growing cells was used for quantitative real-time PCR (qRT-PCR) or Western blot. Primers used for KLF8 are 50 -TCTGCAGG GACTACAGCAAG-30 (forward) and 50 -TCACATTGGTGAA TCCGTCT-30 (reverse). KLF8 to GADPH ratios were normalized to the value for T80. T80 and MCF-10A are an immortalized normal human ovarian surface (Liu et al., 2004) and breast epithelial cell lines, respectively. Human cancer cell lines are CaVO3, SKOV3, OVCAR5 and SKOV3ip1 (ovarian); 786-0 (renal); and MCF-7, BT549, T47D, MDA-MB-231, and Hs578T (breast). (c) The indicated cells were co-immunostained with KLF8 antibody and Hoechst dye for nuclei. (d) Human ovarian tissues (N1N7, normal; C8C48, cancer) gene expression panels were analysed for KLF8 expression using qRTPCR. The results represent at least two independent experiments.

To test directly if KLF8 is itself sufcient in mediating oncogenic transformation, we took advantage of our Tet-off NIH3T3 cell lines that express inducible ectopic KLF8 (Zhao et al., 2003). These cells do not express the ectopic KLF8 if grown in tetracycline-containing

medium. Induction of expression of the ectopic KLF8 by removal of tetracycline from the medium led to remarkably accelerated cell growth (Figure 2A). The aberrant cell growth was characterized by loss of contact inhibition, overlapping growth (Figure 2B) and the less dependence on the serum in the medium (Figure 2C). These phenotypes of the cells clearly t those of broblasts transformed by oncoproteins such as v-Src. Interestingly, we found that induced expression of KLF8 was not sufcient to induce anchorage-independent growth of the cells in soft agar (data not shown). To determine whether or not KLF8 is required for transformation induced by other oncogenes, we rst established an inhibitory mutant of KLF8. This mutant lacked the C-terminal one-and-a-half of the DNAbinding zinc-nger motifs owing to the deletion of the last 38 amino acids. We named this mutant as KLF8DC38. As KLF8DC38 maintained the entire Nterminus and remained localized in the nucleus (data not shown), we believed that this mutant possibly acted in a dominant-negative manner by competing with wild-type KLF8 for recruiting other regulatory proteins as described above. We have previously shown that KLF8 promotes cell cycle progression through the G1 phase by activating cyclin D1 promoter (Zhao et al., 2003). Co-overexpression of KLF8DC38 signicantly blocked wild-type KLF8-enhanced cell cycle progression and cyclin D1 promoter activation in a dose-dependent manner, as analysed by 5-bromo-20 deoxy-uridine (BrdU) incorporation assays (Figure 3A) and cyclin D1 promoter-driven luciferase reporter assays (Figure 3B), respectively. These results conrmed the dominant-negative function of the KLF8DC38 mutant. We then tested the need for KLF8 in oncogenic transformation by making v-Src-transformed NIH3T3inducible cell lines that expressed the KLF8DC38 mutant. Similar cell lines that expressed the wild-type KLF8 or empty vector were also generated as essential controls. The expression of the KLF8 proteins were under the control of tetracycline (Figure 3C) as described previously (Zhao et al., 2003). Induction of KLF8DC38 expression led to more than twofold inhibition of both anchorage-dependent growth (Figure 3D and F) and anchorage-independent growth that was reected by the decrease in both number and sizes of the colonies formed in the soft agar (Figure 3E and F). Interestingly, induction of the wild-type KLF8 expression in the v-Src-transformed cells resulted in further increase in both anchorage-dependent growth (Figure 3D and F) and anchorage-independent growth (Figure 3E and F). We have recently demonstrated that cyclin D1 is a critical transcriptional target and effector of KLF8 in the cell cycle regulation (Zhao et al., 2003; Wei et al., 2006). Importantly, cyclin D1 has been shown to be essential for v-Src-mediated transformation (Leslie et al., 2006) and we have found that induced expression of the KLF8DC38 mutant in the v-Src cells resulted in a dramatic decrease in the expression of endogenous cyclin D1 (data not shown). We thus tested whether cyclin D1 mediates KLF8-promoted transformation in
Oncogene

KLF8 mediates v-Src-induced transformation X Wang and J Zhao

458

the v-Src cells. When cyclin D1 expression was knocked down from the v-Src cells expressing inducible KLF8 (Figure 4A, upper panel), both anchorage-dependent (Figure 4B and D) and anchorage-independent (Figure 4C and D) growths were signicantly reduced regardless of the equal induction of KLF8 expression (Figure 4A, bottom panel). Taken together, our results show that KLF8 promotes serum-independent growth, loss of contact

inhibition, morphological change into spindle-like phenotypes and overlapping cell growth, the hallmark features of oncogenic transformation. Our results also suggest that KLF8 is required for malignant transformation induced by the v-Src oncoprotein and cyclin D1 plays a critical mediating role. Our failure to induce anchorage-independent growth by overexpressing KLF8 alone in HIN3T3 cells could be owing to the use of monoclonal cell lines. We have noticed that for some

Figure 2 KLF8 induces transformed growth and morphology. (A) KLF8 promotes anchorage-dependent growth. Our Tet-off NIH3T3 cells expressing inducible KLF8 or the mock cells (Zhao et al., 2003) were grown in Dulbeccos modied Eagles medium plus 10% serum under uninduced (U) or induced (I) conditions at the density of 2 104 cells in 35-mm wells. The cells were re-fed and counted every other day. (B) KLF8 transforms the cell phenotypes. Representative phase-contrast images of the KLF8 cells showing transformed morphology under induced (KLF8-I) condition (a) compared to those grown under uninduced (KLF8-U) condition (b) and the mock cells (c, d). The images were taken after 90 h. Similar results were obtained for six independent cell lines. (C) KLF8 reduces serum need for the cell growth. Experiments were performed as in panel a except that the media contained 0.5% serum.

Figure 3 KLF8 is required for v-Src-induced transformation. (A and B) KLF8DC38 functions in a dominant-negative manner. (A) Overexpression of KLF8DC38 blocks KLF8-promoted cell cycle progression. The pHan-KLF8DC38 plasmid encoding Myc-tagged KLF8DC38 protein was transfected into the NIH3T3 cells expressing inducible KLF8 for 16 h under uninduced conditions. After 24 h serum deprivation, the cells were stimulated with serum with (I) or without (U) KLF8 induction for 12 h before being prepared for BrdU incorporation assays as described previously (Zhao et al., 2000). (B) Co-expression of KLF8DC38 prevents activation of cyclin D1 promoter by wild-type KLF8. Increased amounts of pHan-KLF8DC38 were co-transfected into NIH3T3 cells with 1 mg of either pKH3 vector or pKH3-KLF8 along with cyclin D1 promoter reporter. Reporter assays were performed as described previously (Zhao et al., 2001). The protein expression was conrmed by Western blot (data not shown). (C) Induced expression of the HA-tagged KLF8 proteins. V-Src-transformed NIH3T3 cell lines that express inducible HA-KLF8 (KLF8), HA-KLF8DC38 (DC38) or control vector only (mock) were generated as described previously (Zhao et al., 2003). Western blotting was conducted using anti-HA and whole lysates of cells grown for 72 h under induced (I) and uninduced (U) conditions. (D) KLF8 enhances but its DC38 mutant inhibits growth of v-Src transformed cells. Serum-stimulated growth of the indicated cells was examined as in Figure 2a. (E) KLF8 enhances but its DC38 mutant inhibits anchorage-independent growth of the v-Src transformed cells. Three thousand cells were grown in soft agar for 2 weeks under uninduced (U) and induced (I) conditions. Top media were refreshed twice a week. After 2 weeks, colonies were counted from 10 independent elds and normalized to the uninduced mock cells. (F) The representative phase-contrast pictures of the cells grown on culture dishes for 2 days (a, b, e, f, I, j) and 4 days (c, d, g, h, k, l) or grown in soft agar for 2 weeks (a0 f0 ). The results are representatives for three triplicate experiments.
Oncogene

KLF8 mediates v-Src-induced transformation X Wang and J Zhao

459

unknown reasons, heterogeneous polyclonal cell lines, but not individually cloned cell lines, are more efcient for inducing anchorage-independent growth and tumor formation by oncogenes such as Ras (Lim and Counter, 2004). Alternatively, KLF8 may be able to induce celltype-specic transformation like FAK, constitutive activation of which is sufcient in the induction of complete malignant transformation of MDCK epithelial cells (Frisch et al., 1996), but not of broblasts (Renshaw et al., 1999). Our experiments are currently in progress to test this interesting hypothesis as to whether or not KLF8 can fully transform epithelial cells. Another possibility is that KLF8 acts in cooperation with another weak oncogene(s) to regulate transformation. For instance, to transform cells, v-Src needs not only to promote cell proliferation but also to promote adhesion-independent cell survival and to regulate the dynamics of cytoskeleton and cellcell and cellmatrix adhesions (Frame, 2004). These multiple requirements

by v-Src involve various intermediate signaling proteins with oncogenic potential such as Myc, Akt, STAT and Rho family proteins. It will be intriguing to investigate how KLF8 cooperates with these proteins in oncogenic transformation. As FAK signaling has also been demonstrated to play a key role in cell invasion (Schlaepfer and Mitra, 2004) and tumor metastasis (Benlimame et al., 2005; van Nimwegen et al., 2005), it will be interesting to test whether KLF8 also plays such roles. In summary, our studies described here clearly establish a novel role for KLF8 in regulating oncogenic transformation. Our nding that KLF8 is overexpressed in human cancer (Figure 1), combined with considerable evidence that link elevated expression and/or activity of Src, FAK and cyclin D1 with the development of human cancer (Diehl, 2002; Gabarra-Niecko et al., 2003; Summy and Gallick, 2003), strongly suggests a potentially signicant role for KLF8 in tumor progression of human cancer.

Oncogene

KLF8 mediates v-Src-induced transformation X Wang and J Zhao

460

Figure 4 Cyclin D1 is a critical mediator for KLF8-promoted transformation in the v-Src cells. (A) Efcient knockdown of cyclin D1 in the v-Src cell line with induced expression of HA-KLF8. Western blot with cyclin D1 and HA antibodies conrmed the knockdown of cyclin D1 expression by its specic siRNA (Dharmacon, Chicago, IL, USA) but not by a scramble control siRNA for KLF8 (Zhao et al., 2003) and the equal induction of the HA-KLF8. (B, C) Knockdown of cyclin D1 prevents KLF8 from promoting the transformation of the v-Src cells. Serum-stimulated growth (B) and soft agar colony formation (C) of the v-Src cells expressing induced HA-KLF8 with or without cyclin D1 knockdown were examined similarly as described in Figure 3d and e. (D) The representative phase-contrast pictures of the cells grown on culture dishes for 2 days (a, b) and 5 days (c,d), or grown in soft agar for 2 weeks (e, f). The results represent for three triplicate experiments.

Acknowledgements We thank Drs Jinsong Liu (T80) and Dihua Yu (SKOV3ip1) of the University of Texas, MD Anderson Cancer Center, Jin Q Cheng (SKOV3 and OVCAR5) of University of South Florida College of References Benlimame N, He Q, Jie S, Xiao D, Xu YJ, Loignon M et al. (2005). J Cell Biol 171: 505516. Bieker JJ. (2001). J Biol Chem 276: 3435534358. Cance WG, Harris JE, Iacocca MV, Roche E, Yang X, Chang J et al. (2000). Clin Cancer Res 6: 24172423. Chen C, Bhalala HV, Qiao H, Dong JT. (2002). Oncogene 21: 65676572. Chen C, Hyytinen ER, Sun X, Helin HJ, Koivisto PA, Frierson HF et al. (2003). Am J Pathol 162: 13491354. Diehl JA. (2002). Cancer Biol Ther 1: 226231. Dong JT. (2001). Cancer Metast Rev 20: 173193. Ellenrieder V, Buck A, Harth A, Jungert K, Buchholz M, Adler G et al. (2004). Gastroenteroloy 127: 607620. Frame MC. (2004). J Cell Sci 117: 989998.
Oncogene

Medicine, C Michael DiPersio (MCF-7 and MDA-MB-231) and Gang Liu (MCF-10A) of our Center for kindly providing the cell lines. This work was supported by Albany Medical College Institutional Grant, Wendy Will Case Cancer Fund and American Cancer Society Grant (RSG CCG-111381) to JZ.

Frisch SM, Vuori K, Ruoslahti E, Chan-Hui PY. (1996). J Cell Biol 134: 793799. Gabarra-Niecko V, Schaller MD, Dunty JM. (2003). Cancer Metast Rev 22: 359374. Ghaleb AM, Nandan MO, Chanchevalap S, Dalton WB, Hisamuddin IM, Yang VW. (2005). Cell Res 15: 9296. Grisaru-Granovsky S, Salah Z, Maoz M, Pruss D, Beller U, Bar-Shavit R. (2005). Int J Cancer 113: 372378. Kimmelman AC, Qiao RF, Narla G, Banno A, Lau N, Bos PD et al. (2004). Oncogene 23: 50775083. Leslie K, Lang C, Devgan G, Azare J, Berishaj M, Gerald W et al. (2006). Cancer Res 66: 25442552. Lim KH, Counter CM. (2004). Mol Cell 15: 491492.

KLF8 mediates v-Src-induced transformation X Wang and J Zhao

461 Liu J, Yang G, Thompson-Lanza JA, Glassman A, Hayes K, Patterson A et al. (2004). Cancer Res 64: 16551663. McLean GW, Komiyama NH, Serrels B, Asano H, Reynolds L, Conti F et al. (2004). Genes Dev 18: 29983003. Nandan MO, Yoon HS, Zhao W, Ouko LA, Chanchevalap S, Yang VW. (2004). Oncogene 23: 34043413. Reeves HL, Narla G, Ogunbiyi O, Haq AI, Katz A, Benzeno S et al. (2004). Gastroenterology 126: 10901103. Renshaw MW, Price LS, Schwartz MA. (1999). J Cell Biol 147: 611618. Schlaepfer DD, Hauck CR, Sieg DJ. (1999). Prog Biophys Mol Biol 71: 435478. Schlaepfer DD, Mitra SK. (2004). Curr Opin Genet Dev 14: 92101. Sood AK, Cofn JE, Schneider GB, Fletcher MS, DeYoung BR, Gruman LM et al. (2004). Am J Pathol 165: 10871095. Summy JM, Gallick GE. (2003). Cancer Metast Rev 22: 337358. van Nimwegen MJ, Verkoeijen S, van Buren L, Burg D, van de Water B. (2005). Cancer Res 65: 46984706. van Vliet J, Turner J, Crossley M. (2000). Nucleic Acids Res 28: 19551962. Wei H, Wang X, Gan B, Urvalek AM, Melkoumian ZK, Guan JL et al. (2006). J Biol Chem 281: 1666416671. Zhao J, Bian ZC, Yee K, Chen BP, Chien S, Guan JL. (2003). Mol Cell 11: 15031515. Zhao J, Pestell R, Guan JL. (2001). Mol Biol Cell 12: 40664077. Zhao J, Zheng C, Guan J. (2000). J Cell Sci 113(Part 17): 30633072. Zhao W, Hisamuddin IM, Nandan MO, Babbin BA, Lamb NE, Yang VW. (2004). Oncogene 23: 395402.

Oncogene