Deficiency of the Kruppel-like factor KLF4 correlates with increased cell proliferation and enhanced skin tumorigenesis

Human skin tumor samples and immuohistonchemistry

Skin cancer tissues array from Biomax (SK481, Rockville, MD, US) were deparaffinized and immunohistochemistry (IHC) performed with a rabbit anti-KLF4 antibody [1:50, Zhang et al. (30)] using a standard protocol.

Cell culture and reagents

Human squamous carcinoma cell line SCC13 was cultured in keratinocyte serum-free medium (K-sfm, Invitrogen) supplemented with 25 μg/ml bovine pituitary extract, 100 U/ml penicillin, 100 μg/ml streptomycin, 0.2 ng/ml epidermal growth factor, 0.3 mM CaCl₂ and 2 mM l-glutamine at 37°C in a humidified atmosphere of 5% CO₂. Primary keratinocytes were isolated from 6- to 8-week-old mice and maintained in keratinocyte basal medium KBM-2 (Lonza, Walkersville, MD).

Generation of KLF4-knockdown and -overexpression stable cell lines

Plasmids containing KLF4-specific short hairpin RNA (pRetroSuper-KLF4#2) and the control vector (pRetroSuper) and KLF4 overexpression vector (pBabePuro-KLF4-N) and the control vector (pBabePuro) were kindly provided by Daniel S. Peeper (31). pBabePuro-KLF4-N, expressed N-terminal 100 amino acid fragment of KLF4, can increase endogenous KLF4 expression by an unidentified mechanism. 1.6 μg plasmids were transfected into skin cancer cells using Lipofectamine™ 2000 (Invitrogen). Forty-eight hours after transfection, cells were trypsinized and reseeded onto 10 cm culture plates. Cells were selected with puromycin at a final concentration of 2 μg/ml. Single colonies were selected, confirmed by western blotting and expanded for further experiments.

RNA extraction and quantitative reverse transcription–PCR analysis

Total RNA was prepared using Trizol Reagent (Invitrogen) according to the manufacturer’s instructions. First-strand complementary DNA (cDNA) synthesis was performed using Oligo (dT) and Superscript III reverse transcriptase (Invitrogen). Real-time PCR was performed as described previously (32). Quantitative real-time PCR primers were used as follows: KLF4: 5′-GAAATTCGCCCGCTCCGATGA-3′ (sense) and 5′-CTGTGTGTTTGCGGTAGTGCC-3′ (antisense); glyceraldehyde 3-phosphate dehydrogenase: 5′-TTCGACAGTCAGCCGCATCTTCTT-3′ (sense) and 5′-CAGGCGCCCAAYACGACCAAATC-3′ (antisense).

Protein extraction and western blotting analysis

Total proteins from culture cells and mouse back skin primary cells with indicated treatments were prepared using RIPA buffer, and 50 μg of proteins were separated by 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel followed by western blotting analysis as described (32). The antibodies used were an anti-KLF4 rabbit polyclonal antibody (1:1500, Genespin, Milano, Italy), an anti-p21 antibody (1:1000, Santa Cruz Biotechnology) and an anti-α-Tubulin antibody (1:1000, Santa Cruz Biotechnology).

3-(4,5-dimethylthiazole-2-yl)-2,5-bdiphenyl tetrazolium bromide cell proliferation assay

SCC13 stable cells (2 × 10³) or primary mouse keratinocytes (2 × 10³) isolated from control (KLF4+/+) or KLF4 (−/−) mouse back skin were seeded onto 96-well tissue culture plates and incubated at 37°C in a humidified atmosphere of 5% CO₂ for 12 h, 1, 3 and 5 days, respectively. At the indicated time point, 3-(4,5-dimethylthiazole-2-yl)-2,5-bdiphenyl tetrazolium bromide (MTT) cell proliferation assay kit (ATCC, Manassas, VA) was used to detect absorbance at 570 nm according to the manufacturer’s instructions.

Cell migration assay

Modified Boyden chamber assay was used to examine cell migration as reported (33). Briefly, SCC13 cells (7 × 10⁴) or primary keratinocytes (7 × 10⁴) isolated from control (KLF4+/+) or KLF4 (−/−) mice were plated onto Transwell chambers (5 μm pore, Corning) coated with Type I collagen. K-sfm or KBM-2 media supplemented with 10% fetal bovine serum was added into the lower chamber. After incubation for 6 h, cells were fixed and stained with 0.1% crystal violet. Non-migrating cells on the upper membrane were removed with a cotton swab. Cells on the bottom surface of each membrane were counted (10–12 fields per well, triplicate for each experimental group) after incubation. Experiments were performed in triplicate.

Cell attachment assay

96-well tissue culture plates were coated with 20 μg/ml fibronectin at 37°C for 1 h and blocked with 5 mg/ml heat-denatured BSA (Sigma) for 1 h. SCC13 cells or primary keratinocytes (2 × 10⁴) isolated from control (KLF4+/+) or KLF4 (−/−) mice were added to each well and cells were allowed to incubate at 37°C for 30 min. After washing twice, attached cells were fixed, washed, and stained with 0.1% (w/v) crystal violet for 10 min. The stained cells were washed thoroughly and dissolved in 2% sodium dodecyl sulfate, and absorbance at 550 nm was measured.

IHC analysis

Following sacrifice, mouse skin tissues were dissected, embedded in paraffin and sectioned at 4 μm. IHC was performed using a standard protocol. Signal was detected using the KPL’s DAB Reagent Set (KPL, Gaithersburg, MD) according to the manufacturer’s instructions, and counterstaining was carried out with Mayer hematoxylin. The following primary antibodies were used: rabbit anti-KLF4 polyclonal antibody (1:500, Genespin, Italy), rabbit anti-Ki67 antibody (1:500, Abcam).

Keratinocyte purification

Collection of mouse dorsal skin keratinocytes was performed as reported previously (34,35). Six- to eight-week-old female mice were killed, dorsal fur was clipped and dorsal skin was removed. All subcutaneous tissues from the skin were scraped, and the skin cut into 0.5 × 1 to 1.5 cm strips and incubated at 32°C in Dulbecco's phosphate-buffered saline with 0.25% trypsin without ethylenediaminetetraacetic acid for 2 h. The epidermis was then scraped into Harvesting Medium (SMEM with 10% fetal bovine serum; Invitrogen). After stirring at 100 r.p.m. on a magnetic stirrer for 20 min at room temperature, the suspension was filtered through a 70 μm cell strainer (BD Biosciences), and keratinocytes were counted using a hemacytometer.

Generation of the tamoxifen-inducible KLF4/CreER mouse model

A construct containing Cre recombinase cDNA fused with a tamoxifen-inducible estrogen receptor (ER) gene (a plasmid kindly provided by Dr Douglas Melton at Harvard Medical School) was used as a PCR template to amplify a fragment containing tamoxifen-inducible CreER cDNA sequences and 5′ and 3′ ends homologous to sequences flanking the mouse tamoxifen-inducible CreER PCR primers: sense = GGTCCCTGTGCGCCCCGCCCGGCCCGCAGGCAGCCACCTGGCGAGTCTGACATGTCCAATTTACTGACCGTACA, antisense = GGCCGGACGCGAACGTGGAGAAGGACGGGAGCAGAGCGTCGCTGACAGCCATTCAGATCGTGTTGGGGAAGCCGTG, the underlined region is homologous to the sequence flanking the KLF4 translation start site. Bacterial artificial chromosome (BAC) clone # RP23-322L22 was obtained from Children's Hospital Oakland Research Institute (CHORI) and the BAC clone was then introduced into a bacterial strain BL250. BL250 bacteria containing RP23-322L22 were made competent and transformed with the amplified tamoxifen-inducible CreER fragment by electroporation according to a previous report (36). Clones in which the amplified fragment was inserted via homologous recombination were selected using resistance to chloramphenicol and kanamycin. Selected colonies were screened for correct recombination by amplification using Cre- and KLF4-specific primers. The FRT-flanked KANA gene was excised from tamoxifen-inducible KLF4/CreER-KANA BAC clones by expressing the FLP recombinase. CHEF gel electrophoresis and DNA sequencing confirmed correct transgene constructions and integrity of the BAC flanking sequence. The tamoxifen-inducible KLF4/CreER BACs were then purified for pronuclear microinjection. Potential founder mice (B6/CBA mixed background) were genotyped by tail DNA amplification using primers specific for the Cre coding sequence. The founder transgenic mice were then backcrossed 6–8 generations to obtain the pure C57BL/6 background. Cre recombinase was induced by daily i. p. injection of tamoxifen (100 mg/kg) (Sigma) for five consecutive days.

Maintenance of mouse colonies and the classical DMBA/TPA mouse skin tumor model

Mice were housed in micro-isolator solid-bottomed polycarbonate cages under a pathogen-free condition and experiments were approved by the Institutional Animal Care and Use Committees of University of South Carolina. Skin tumors were generated as described (18). Briefly, the control and tamoxifen-induced KLF4/CreER and KLF4(flox+/+) double transgenic mice were treated with a single dose of DMBA (100 nmol/100 μl acetone; Sigma) followed by repetitive TPA applications (10 nmol/100 μl acetone: Sigma) twice a week for 26 weeks. Acetone-treated control mice (acetone/acetone, DMBA/acetone or acetone/TPA) used as internal controls and did not develop any tumors. Tumors were assessed weekly. Mice were killed 4 weeks after the final treatment. Generation of KLF4(flox+/+) mice has been described previously (37).

Statistical analysis

Quantification of Ki-67-positive cells was performed using Image Pro Plus 5.1 software (Media Cybernetics, Bethesda, MD) on stained skin sections from at least three different samples and the average value was the percentage of Ki-67-positive cells. Two-way analysis of variance and two-tailed t-tests were employed to analyze the data using Prism 5.0 software (GraphPad). P < 0.05 was defined as statistically significant.

Absence of KLF4 expression in patient samples of skin cancer

Recent studies have shown that KLF4 functions either as a tumor suppressor or an oncogene depending on different cellular contexts (17). In order to determine the role of KLF4 in skin tumor development, we analyzed KLF4 expression in 24 formalin-fixed paraffin-embedded human skin tumor samples (12 patients with squamous cell carcinoma and 12 with basal cell carcinoma) and 24 human cancer adjacent normal dermatic tissues. The numbers of female and male patients were the same. As shown in Table I, in squamous cell carcinoma, the lesions included tempus, thigh, face and scalp. Seven cases were malignant grade 1 and the remaining was malignant grade 2 (data not shown). In basal cell carcinoma, lesions were located in eyelids, pars zygomatica, face, nasal ala, corona capitis and frontal region. As expected, KLF4 was expressed in the epithelial cells of 24 normal skin tissue samples but only expressed in 3 of 24 tumor samples detected by immunohistochemical staining. Strong KLF4 staining was observed in the suprabasal layer (arrows in inset) with diffused staining in basal layers in normal tissue (Supplementary Figure 1B is available at Carcinogenesis Online). Typical KLF4 expression patterns of cancer samples (middle and right panels) were shown in Supplementary Figure 1A (available at Carcinogenesis Online). Absence of KLF4 staining in 87.5% (21 in 24 samples) skin tumor samples supports a tumor-suppressive role of KLF4 in skin cancer development.

KLF4 inhibits proliferation and migration of human epithelial carcinoma SCC13 cells

To test the function of KLF4 in cell proliferation in vitro, we generated KLF4-knockdown stable cells (designated as shKLF4) by transduction with an established KLF4-specific short hairpin RNA vector (31) and KLF4-overexpressing stable cells (designated as KLF4-N) by transduction with an N-terminal version of KLF4 in human epidermal squamous carcinoma cell line SCC13 and in epidermoid carcinoma cell line A431. Note that the full-length KLF4 was not used because of its strong antiproliferative effect and very few stable colonies obtained in the establishment of stable cells. Quantitative real-time PCR analysis and western blot analysis (Figure 1A) confirmed the expected levels of KLF4 messenger RNA and protein in KLF4-knockdown and -overexpressing stable SCC13 cells. To test whether KLF4 has an effect on cell proliferation in SCC13 cells, we performed the MTT assay in KLF4 knockdown and overexpression cells. As expected (38), cell proliferation was significantly accelerated after KLF4 knockdown (Figure 1B). Consistently decreased proliferation was observed upon KLF4 overexpression (Figure 1C). The same results were obtained in A431 cells after KLF4 knockdown and overexpression (Supplementary Figure 2B is available at Carcinogenesis Online and data not shown). Mechanistically, previous work determined that KLF4-mediated G₁ arrest was via transcriptional induction of p21 (38). In our study, p21 protein levels were indeed decreased in shKLF4 cells when compared with that in control cells as shown in Figure 1A. Consistently, increased p21 protein levels were observed in KLF4-overexpressing cells (Figure 1A). Our data thus support that KLF4 inhibits tumor cell proliferation at least in part through a p21-dependent manner.

One of the important properties of tumor cells is their increased mobility. To evaluate whether KLF4 regulates cell migration, we performed a transwell assay using KLF4 knockdown and overexpression cells. As shown in Figure 1D, the number of migrated shKLF4 cells had a 1.41-fold increase when compared with that of the shCon cells. Consistently, the number of migrated KLF4-overexpressing cells was 55% less than that of the control cells. In addition, it is well established that cell attachment or detachment with the matrix where the tumor grows in the microenvironment is the ‘hallmark’ of cell migration and invasion during metastatic processes (39). Notably, the shKLF4 cells had increased capacity of attachment with fibronectin when compared with the shCon cells, and consistently KLF4-overexpressing cells showed decreased adhesion when compared with the control cells (Figure 1E), indicating that KLF4 has an inhibitory effect on cell adhesion.

Generation of an inducible KLF4 knockout mouse model

Full KLF4 knockout mice die shortly after birth due to dehydration which partially prevents us from studying the role of KLF4 in adult (15). To circumvent this situation and to study the in vivo function of KLF4, we first generated tamoxifen-inducible KLF4/CreER transgenic mouse model using a Cre recombinase as the transgene derived from a mouse KLF4 gene-containing BAC (RP23-322L22) (36). This BAC clone contains ∼119kb 5′ and 82kb 3′ fragments of KLF4 gene. In this transgene, a Cre recombinase cDNA fused with a tamoxifen-inducible (TM) ER (CreER™) gene was precisely inserted into the KLF4 locus at the initiating methionine without deleting any endogenous KLF4 sequences, so that CreER™ gene expression was under the control of a big 5′ sequence of the KLF4 gene. A diagram of the construction strategy to generate a KLF4/CreER™ transgene was shown in Figure 2A. Western blot confirmed the expression of Cre at a protein level and its expression did not influence that of the endogenous KLF4 (Figure 2B). Fidelity of the model was also confirmed by matched expression patterns of Cre with that of endogenous KLF4 by IHC in mouse small intestine and colon (data not shown). Then, these mice were crossed with KLF4(flox+/+) mice (37) to generate tamoxifen-inducible KLF4/CreER (+/−) and KLF4(flox+/+) double transgenic mice and KLF4 was knocked out by daily tamoxifen i. p. injection at a dose of 100 mg/kg for five consecutive days. As shown in Figure 2C (high magnification in Supplementary Figure 3C is available at Carcinogenesis Online), while KLF4 staining indicated by red arrows (left) was abundant in hair follicles of the mice induced with vehicle control, the staining was almost absent in hair follicles of mice with tamoxifen induction (right), suggesting a successful KLF4 knockout in our KLF4-inducible knockout model. Western blotting analysis also confirmed the drastically reduced protein levels of KLF4 in skin keratinocytes after tamoxifen induction when compared with that after the control induction (Figure 2D).

KLF4 deficiency is accompanied by increased cell proliferation in vivo

Eight weeks after induction of tamoxifen-inducible KLF4/CreER (+/−) and KLF4(flox+/+) double transgenic mice with tamoxifen, no obvious phenotypic changes of mouse hairs were observed. However, number of the hair follicles was prominently increased upon KLF4 knockout as indicated by the arrows in Supplementary Figure 4A is available at Carcinogenesis Online when compared to that in the control by hematoxylin and eosin staining, which was consistent with an inhibitory function of KLF4 in proliferation in vitro. In addition, changes of suprabasal cells from a single layer to multiple layers in KLF4 knockout mice was also observed (insets in Figure 3A and Supplementary Figure 4B), suggesting a high proliferative status of these cells upon KLF4 knockout. The total number of keratinocytes purified from 3 × 3 cm² back skin was increased from 1.81 × 10⁷ ± 1.15 × 10⁶ in control mice to 2.77 × 10⁷ ± 7.02 × 10⁵ in KLF4 knockout mice (Figure 3B). Consistently, IHC staining with the proliferation marker Ki67 in skin tissue also confirmed the increased proliferation of hair follicles upon KLF4 knockout. As shown in Figure 3C and D and Supplementary Figure 4D (available at Carcinogenesis Online), the proportion of Ki67-positive cells increased from 8% to 43% after KLF4 knockout. These results indicate that deficiency of KLF4 is accompanied by increased cell proliferation of mouse skin keratinocytes.

KLF4 inhibits migration and adhesion of mouse primary keratinocytes ex vivo

Our previous study showed increased proliferation status of keratinocytes from the dorsal skin of KLF4 knockout mice. Using purified keratinocytes from the control and KLF4 knockout mouse back skin, we performed the MTT assay to further examine the cell proliferation ex vivo. As expected, the cell proliferation was significantly accelerated upon KLF4 knockout (Figure 4A). Next, we evaluated whether KLF4 regulates cell migration and adhesion ex vivo by the transwell assay. As shown in Figure 4B, the number of migrated cells purified from KLF4 knockout mice was increased 1.53-fold when compared with that from the control mice. Similar results were obtained when the transwell chambers were coated with Type I collagen (Figure 4C), an assay for cell invasion, indicating that KLF4 inhibits both migration and invasion of primary mouse skin keratinocytes. Furthermore, in an attachment assay, KLF4 knockout keratinocytes exhibited an 80% increase of cell attachment to fibronectin when compared with the control keratinocytes (Figure 4D). Taken together, these results suggest that KLF4 inhibits cell migration and adhesion of mouse primary keratinocytes.

KLF4 deficiency is accompanied by the enhanced skin tumorigenesis in a DMBA/TPA mouse skin cancer model

We have shown that KLF4 knockdown was associated with increased proliferation of cells of skin squamous cell carcinoma (Figure 1) and knockout of KLF4 increased proliferation of primary mouse skin keratinocytes (Figures 3 and 4), suggesting a tumor-suppressive role of KLF4 in skin cancer development. To test this hypothesis, we utilized a classical two-step DMBA/TPA mouse skin cancer model. Five pairs of tamoxifen-inducible KLF4/CreER(+/−)/KLF4(flox+/+) double transgenic female mice at 4-week-old were i. p. injected with vehicle (KLF4+/+ mice) or tamoxifen (KLF4−/− mice) to induce KLF4 knockout. They were then treated with a single topical application with the carcinogen DMBA (100 nmol/100 μl acetone), followed by promotion with the phorbol ester TPA (10 nmol/100 μl acetone) for 26 weeks (18). Tumors were found on KLF4−/− back skin after 12 weeks of promotion and by 30 weeks 80% percent KLF4−/− mice had developed tumors. In contrast, control mice developed tumors after 15 weeks and the tumor incidence was 20% (Figure 5A, left). Tumor multiplicity in KLF4−/− mice was two tumors per mouse but only 0.6 tumors per mouse in control mice (Figure 5A, right). Moreover, tumor size in control mice was smaller (Figure 5B). Highly differentiated and hyperkeratotic morphology was shown in KLF4−/− tumor by hematoxylin and eosin staining (Figure 5C). The accelerated incidence and increased multiplicity of mouse skin tumors in KLF4 knockout mice strongly support a tumor-suppressive role of KLF4 in the DMBA/TPA mouse skin tumor model.