SP/KLF (Specificity protein/Krüppel-like factor) transcription factors comprise an emerging group of proteins that may behave as tumour suppressors. Incidentally, many cancers that display alterations in certain KLF proteins are also associated with a high incidence of KRAS (V-Ki-ras2 Kirsten rat sarcoma viral oncogene homologue) mutations. Therefore in the present paper we investigate whether SP/KLF proteins suppress KRAS-mediated cell growth, and more importantly, the potential mechanisms underlying these effects. Using a comprehensive family-wide screening of the 24 SP/KLF members, we discovered that SP5, SP8, KLF2, KLF3, KLF4, KLF11, KLF13, KLF14, KLF15 and KLF16 inhibit cellular growth and suppress transformation mediated by oncogenic KRAS. Each protein in this subset of SP/KLF members individually inhibits BrdU (5-bromo-2-deoxyuridine) incorporation in KRAS oncogenic-mutant cancer cells. SP5, KLF3, KLF11, KLF13, KLF14 and KLF16 also increase apoptosis in these cells. Using KLF11 as a representative model for mechanistic studies, we demonstrate that this protein inhibits the ability of cancer cells to form both colonies in soft agar and tumour growth in vivo. Molecular studies demonstrate that these effects of KLF11 are mediated, at least in part, through silencing cyclin A via binding to its promoter and leading to cell-cycle arrest in S-phase. Interestingly, similar to KLF11, KLF14 and KLF16 mechanistically share the ability to modulate the expression of cyclin A. Collectively, the present study stringently defines a distinct subset of SP/KLF proteins that impairs KRAS-mediated cell growth, and that mechanistically some members of this subset accomplish this, at least in part, through regulation of the cyclin A promoter.
The SP/KLF (Specificity protein/Krüppel-like factor) family constitutes a group of transcription factors that are present in organisms ranging from yeast to vertebrates . Their structure is defined by the presence of three highly conserved DNA-binding zinc-finger domains and a variant N-terminal domain that contains transcriptional regulatory motifs [2,3]. Identification of an entire repertoire of SP/KLF proteins, along with characterization of their biochemical properties, has been the focus of intensive investigations. These studies have revealed that KLF proteins bind to similar, yet distinct, GC-rich target sequences, and they function either as activators or repressors in cell- and promoter-dependent manners [2–4] by interacting with co-regulator molecules via different types of regulatory domains [2,4–6]. Interestingly, however, expression of KLF proteins is altered in many types of cancer, including breast, head and neck, colon, oesophagus, pancreas, liver, lymphomas and leukaemias. Thus, in spite of extensive biochemical characterizations performed on KLF proteins, their biological functions remain a matter of intense investigation and in particular how they relate to carcinogenesis. Moreover, whether their biochemical properties, such as targeting of specific genes promoters, have a clear impact on their cell biological function remain unclear. Thus investigations that enhance our understanding of the role of SP/KLF proteins at the mechanistic level is of paramount importance for more complete characterization of this important evolutionarily conserved family. Consistent with this goal, the present study focuses on the functional characterization of known members of the SP/KLF family of transcription factors as it relates to their ability to suppress oncogene-mediated cell growth. The results of the present study, for the first time, characterize a novel subset of SP/KLF proteins that antagonize KRAS (V-Ki-ras2 Kirsten rat sarcoma viral oncogene homologue)-mediated cell growth via the inhibition of proliferation and induction of apoptosis. Mechanistically, the growth suppressive function of some members of this subset of SP/KLF proteins is achieved, at least in part, through the down-regulation of the promoter of a key cell-cycle regulator, cyclin A. Since KRAS is mutated in 30% of all tumours , and taking into consideration the increasing evidence supporting the involvement of KLF proteins in different types of cancer, the novel results reported in the present paper are of significant relevance to tumour biology.
Cell lines, reagents and plasmids
Unless specified, all reagents were from Sigma. Plasmids containing SP1–8 and KLF1–16 cDNAs along with cyclin A2 (CCNA2) promoter (containing a region of the CCNA2 promoter corresponding to −468 to +298 bp upstream of the luciferase gene) and KLF11-deletion constructs were cloned as previously described [8–11]. Oncogenic human KRASV12 plasmid was a gift from Dr Eugenio Santos (National Institutes of Health, Bethesda, MD, U.S.A). To obtain the −89 to −91 CGC>TTT mutant CCNA2 promoter pGL3 reporter construct, site-directed mutagenesis was performed using the QuikChange® site-directed mutagenesis kit as suggested by the manufacturer (Agilent Technologies). The cyclin A2-expression construct was a gift from Dr Robert Sheaff (University of Minnesota, Minneapolis, MN, U.S.A). KLF10-, KLF11- and EV (empty vector)- (Ad5CMV) carrying recombinant adenoviruses were generated in collaboration with the Gene Transfer Vector Core at the University of Iowa. Human pancreatic cancer cell lines and mouse fibroblast (NIH 3T3) cells used in the present study were obtained directly from the American Type Culture Collection and maintained according to their recommendations. MEFs (mouse embryonic fibroblasts) were freshly isolated and cultured from homogenized E13.5 (E is embryonic day) embryos arising from timed mating between wild-type C57BL6 mice using standard methods . Animal experiments were approved by the Animal Facility at Mayo Clinic College of Medicine. The normal human pancreatic-ductal epithelial cell line, HPDE6 (immortalized, but not transformed), was cultured as previously described .
Tumour xenografts and transformation assays
CAPAN2 or L3.6 cancer cells (2.5×106) were injected subcutaneously into the hind leg of athymic nude mice and the resulting tumour size was measured weekly. Nine mice were injected for each experimental condition. We estimated tumour volume (V) from the length (l) and width (w) of the tumour using the formula: 4/3×π×[(l + w)/4]3 . The transformation assays were performed as described previously .
Proliferation and apoptosis assays
Pancreatic cell lines were electroporated using a BTX square-wave electroporator (10 ms, 360 V). Cells were co-transfected with a GFP (green fluorescent protein) vector (5 μg) and the indicated SP/KLF plasmid (25 μg) or control pcDNA3 vector (25 μg). At 48 h post-plating, the cells were pulsed with 100 μM BrdU (5-bromo-2-deoxyuridine) for 1 h, fixed in 3.7% formaldehyde for 10 min at room temperature (25 °C), and permeabilized with 70% ethanol and DNase I at 30 °C for 20 min. After blocking with 5% (v/v) goat serum for 30 min at 37 °C, the coverslips were incubated for 60 min with an anti-BrdU mAb (monoclonal antibody; Roche Applied Science), and then a rhodamine-conjugated anti-mouse secondary antibody (Molecular Probes) and 1 μg/ml DAPI (4′,6-diamidino-2-phenylindole). For quantification, 500 GFP-positive cells in four different 63×fields were counted using a Zeiss confocal microscope. Western blotting was performed as a control of tagged-SP/KLF expression. For cyclin A2-rescue experiments, an MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] assay was performed as described previously . For MEF experiments, the cells were plated at 6000–8000 cells per well in a 96-well plate. At 24 h later, the medium was replaced with serum-medium and the cells were infected with EV- or KLF11-carrying adenovirus at a MOI (multiplicity of infection) of 200. The medium was supplemented with either 5 or 10% FBS (fetal bovine serum) after 3 h. At 48 h later, the medium was aspirated and the wells were washed with PBS. Subsequently, a fluorescent LIVE/DEAD® viability/cytotoxicity assay was performed according to the manufacturer's protocol (Invitrogen). Cells undergoing apoptosis were identified by Hoechst 33860 as described previously . Briefly these cells were identified using light microscopy by their morphological criteria of nuclear fragmentation, nuclear margination, cytoplasmic blebbing and organelle disorganization. The results were validated using annexin V staining. For quantification, 300 GFP-positive cells were counted using a Zeiss confocal microscope.
Soft agar growth assay
Cells (1×107) were infected with control or KLF11 adenovirus, allowed to recover for 18 h and then mixed with RPMI 1640 medium containing 10% BCS (bovine calf serum) and 0.4% LMP (low-melting point) agarose. This mixture was placed over hardened RPMI 1640 medium containing 10% (v/v) BCS and 1% (w/v) LMP agarose and allowed to harden. The cells were grown for 2 weeks and visible colonies, stained with 0.4% Crystal Violet and containing greater than 50 cells, were counted.
RT (reverse transcription)–PCR analysis, luciferase and ChIP (chromatin immunoprecipitation) assays
Cell-cycle FACS analysis
Adherent cells were harvested, filtered through a 30 μm nylon-mesh filter (Miltenyi Biotec) and fixed for 1 h at 4 °C with 100% ethanol. The fixed cells were incubated with PI (propidium iodide) staining solution (Biosure) and 100 μg/ml RNAse A at room temperature for 40 min. Flow cytometry was performed on a BD™ LSR II instrument, equipped with a Coherent® Sapphire™ 20 mW laser (excitation wavelength, 488 nm; BD Biosciences). PI fluorescence was detected by a photomultiplier tube, a 595 nm long-pass dichroic mirror and 610±20 nm bandpass emission filter. A total of 110000–170000 events were collected using appropriate stop gates set in the PI-fluorescence area compared with width projection to exclude cell aggregates. Acquisition settings were verified by adding internal control cells (chick erythrocyte nuclei; Biosure) to an aliquot of each sample. The results files were analysed with FlowJo software (Treestar) and the Dean—Jett–Fox algorithm.
EMSA (electrophoretic mobility-shift assays)
EMSAs were performed essentially as described in . Annealed double-stranded oligonucleotides corresponding with potential KLF-binding sites in the human CCNA2 promoter were end-labelled with [γ-32P]ATP using T4 polynucleotide kinase: site #1, 5′-GTTTCTCCCTCCTGCCCCGCCCCTGCTCAGTTTCC-3′; site #2, 5′-GCAGGCGTTTTCTCCCGCCCCAGCCAGTTTGTTTC-3′; site #3, 5′-CTAAATCCTACCTCTCCCCGCCCCGCGCAGGCGTTTTC-3′; site #4, 5′-CGGAAGCGTCGGGCCCTAAATCCTAC-3′; and a site #1 mutant probe 5′-GTTTCTCCCTCCTGCCCTTTCCCTGCTCAGTTTCC-3′ as indicated by the manufacturer (Promega). A GST (glutathione transferase)-fusion protein carrying the zinc-finger region of KLF11 was utilized for these studies. GST alone (control) and GST-fusion protein expression was induced in BL21 cells (Agilent Technologies) by the addition of 2 mM IPTG (isopropyl β-D-thiogalactopyranoside) and incubation for 2 h at 37 °C. The cells were lysed and subsequently purified using glutathione–Sepharose 4B affinity chromatography in accordance with the manufacturer's instructions (GE Healthcare Bio-Sciences). Subsequently, EMSAs were performed using 1 μg of purified GST or GST–KLF11 recombinant fusion proteins incubated in a buffer containing 20 mM Hepes (pH 7.5), 50 mM KCl, 5 mM MgCl2, 10 mM ZnCl2, 6% (w/v) glycerol, 200 mg of BSA per ml and 50 mg of poly(dI-dC)·poly(dI-dC) (Sigma) per ml for 7 min at room temperature. Approximately 0.3 ng of the end-labelled probe was then added to each reaction for an additional 20 min, with or without 125×cold competitor anti-GST antibody, or anti-mIgG (monoclonal IgG) antibody as indicated, and loaded immediately on to a 4% non-denaturing polyacrylamide gel. The samples were run for 2 h at 200 V at room temperature, vacuum-dried, and exposed to HyBlot CL™ autoradiography film (Denville Scientific).
Systematic screening reveals a role for distinct SP/KLF proteins in blocking KRAS-mediated cell growth
To test the effects of SP/KLF in transformation, we initially performed foci-formation assays to assess anchorage-dependent cell growth using the NIH 3T3 transformation model, utilizing well-established oncogenes to induce transformation and, subsequently, observe any reversal of this effect with each of the 24 SP/KLF family members. NIH 3T3 cells were cotransfected with the potent oncogene KRASV12 along with a full-length construct of each member of the SP/KLF family of transcription factors or parental vector (control). Interestingly, only SP5, SP8, KLF2, KLF3, KLF4, KLF11, KLF13, KLF14, KLF15 and KLF16 were able to suppress more than 50% of the transformation mediated by the oncogene KRASV12 (Figure 1). Similar results were obtained using HRAS (v-Ha-ras Harvey rat sarcoma viral oncogene homologue; results not shown). Notably, none of the SP/KLF genes induced foci formation alone or increased the transformation mediated by KRAS (Figure 1). This is congruent with using the NIH 3T3 assay where only one oncogene is needed to achieve maximal transformation, which is different than using MEF-mediated assays, which are better for studying oncogenes due to their requirement of two hits to be fully transformed. Therefore a distinct subset of KLF proteins, namely SP5, SP8, KLF2, KLF3, KLF4, KLF11, KLF13, KLF14, KLF15 and KLF16, are able to significantly suppress KRAS-mediated foci formation.
Inhibition of KRAS-mediated transformation is a defined feature of a subgroup of SP/KLF transcription factors
Suppressor-of-KRAS KLF proteins inhibit anchorage-independent cell growth and in vivo tumorigenesis
To gain an in-depth insight into the molecular mechanisms underlying the suppressive function of this subgroup of SP/KLF proteins, we used a two-tiered experimental strategy. First, we chose one representative member of the subgroup of KLF proteins that suppress KRAS-mediated transformation as a model to search for novel mechanisms that may account for its suppressive function, and, subsequently, investigated whether these mechanisms were utilized by the rest of the subset. For this purpose, we performed these investigations using KLF11, a gene previously identified by our laboratory . To facilitate these studies, we generated adenovirus-expressing KLF11 cDNA to infect several KRAS oncogenic-mutant pancreatic cancer cell lines (AsPC1, Capan2, CFPAC1, L3.6 and MiaPaCa2) to determine the rate of BrdU incorporation at different time points. As shown in Figure 2(a), growth in vitro was significantly inhibited in KLF11-infected cells compared with control-infected cells. In addition, we determined the ability of KLF11 to block colony formation in soft agar by these pancreatic-tumour cells (Figure 2b). KLF11 infection of Capan2, L3.6, MiaPaCa2 and PANC1 cell lines resulted in a substantial reduction in colony formation in soft agar compared with the control-infected populations. As an in vivo validation of this effect of KLF11 on tumour growth in vivo, we injected KLF11- or control-infected Capan2 and L3.6 cells subcutaneously into the flanks of irradiated male athymic nude mice and monitored the increase in tumour volume over a 6-week period. We found that in vivo tumour formation was significantly delayed in the mice with injected tumour cells expressing exogenous KLF11, as compared with control mice (Figures 2c and 2d). We confirmed the expression of KLF11 in the KLF11-infected cells by Western blotting (Figure 2c). These results indicate that KLF11 inhibits the transformed phenotype in vitro and in vivo to support its role as a tumour suppressor in KRAS-mutant tumours.
KLF11 inhibits proliferation, anchorage-independent growth and tumour formation in vivo in pancreatic-cancer cells
Suppressor-of-KRAS KLF proteins inhibit cell proliferation and induce apoptosis
Subsequently, we investigated the cellular mechanisms underlying the effect of the additional KRAS-suppressor KLF proteins by examining if the suppression of neoplastic transformation by SP/KLFs was, similar to KLF11, due at least in part to a change in cell proliferation. For this purpose, we tested the ability of these proteins to regulate BrdU incorporation at various time points. This was performed first in PANC1 cells, which are cancer cells containing the endogenous oncogenic mutant KRAS. As shown in Figure 3(a), transfection with SP5, SP8, KLF2, KLF3, KLF4, KLF11, KLF13, KLF14, KLF15 and KLF16, the same subset that suppressed KRAS-mediated transformation, resulted in decreased proliferation in PANC1. Subsequently, we repeated this experiment in two additional cell lines, BxPC3 and NIH 3T3. Both of these cell lines contain wild-type endogenous KRAS, which were then transfected with the oncogenic variant of KRAS. Analysis of BxPC3 (Figure 3b) and NIH 3T3 cells (Figure 3c) confirm that the same subset, with the exception of KLF13, decrease proliferation of these cells in the presence of oncogenic KRAS, as determined by the reduction in the number of BrdU-positive cells. Next we examined the effect of these SP/KLFs on cell death via morphological criteria of nuclear fragmentation, nuclear margination, cytoplasmic blebbing and organelle disorganization, as visualized by Hoechst DNA staining. Interestingly only SP5, KLF3, KLF11, KLF13, KLF14 and KLF16 increased the number of apoptotic cells by more than 2-fold in PANC1 cells (Figure 3d). Thus, taken together, these results identify a subgroup of SP/KLF transcription factors that suppress oncogenesis and suggest that suppression of KRAS neoplastic transformation is explained, in part, by the ability of these molecules to inhibit cell proliferation and for some of this subset, to also induce apoptosis. Conceptually, these results identify two important cellular mechanisms underlying the suppression of neoplastic transformation by these proteins, namely an increase in apoptosis and concurrent decrease in cellular proliferation. More importantly, these results demonstrate that KLF11 is a reliable suppressor protein to utilize as a robust model in gathering more mechanistic data.
Suppressor-of-KRAS KLF proteins inhibit cell proliferation and induce apoptosis
Distinct suppressor-of-KRAS KLF proteins share a common mechanism of silencing cyclin A
Since inhibition of cellular proliferation was common among all members of the SP/KLF family that suppressed KRAS-mediated transformation, we investigated the molecular mechanisms underlying this effect of the suppressor-of-KRAS KLF proteins. Consequently, we performed RT–PCR for genes involved in regulating the cell cycle, including cyclin D1, cyclin A2 and cyclin B1. Notably, the mRNA levels of cyclin A2 were significantly decreased in three SP/KLF transcription factors with suppressor-of-KRAS growth functions, specifically KLF11, KLF14 and KLF16 (60.8±15%, 47.0±7.4% and 70.2±0.1% of control levels respectively) (Figure 4a). These KLF proteins were also able to repress the CCNA2 promoter activity analysed via a luciferase-based reporter assay (KLF11, 39.4±12.1%; KLF14, 11.4±3.0%; and KLF16, 33.4±6.1% compared with control) (Figure 4b). Again, using KLF11 as a model, we analysed the cell-cycle profile via flow cytometry in KLF11-overexpressing cells, which resulted in a lower percentage of the cell population in G0/G1-phase with a mean change of 12.025±1.4% compared with control and a concordant increase in cells in S-phase with a mean difference of 11.26±2.1% compared with the control (Figure 4c). There was no change in cells in G2/M-phase between the parental vector (EV) control and KLF11 groups (mean difference 0.475%±2.95%). These results, along with decreased BrdU incorporation mediated by KLF11 (Figures 2a and 3a–3c), suggest that this increased population of cells in S-phase represent an arrest at this stage of the cell cycle. We find that this effect of KLF11 on the CCNA2 promoter is not only functional in cancer cells, as similar results on cell growth analysed by fluorescent LIVE/DEAD® viability/cytotoxicity assay (approximately a 30% reduction in cell survival compared with control-infected cells, P<0.05) and CCNA2 promoter activity (30.9±18.4% relative to control, P<0.05) were obtained in normal fibroblasts and normal pancreatic-ductal cells respectively (Figures 4d–4e).
KLF11 represses the CCNA2 promoter and causes cell-cycle arrest
To determine whether the effect of KLF11 on cyclin A2 is due to a direct interaction with its promoter, we performed ChIP assays. Indeed, KLF11 was found on the endogenous CCNA2 promoter, whereas KLF10, a protein that does not regulate the promoter in the experiments of the present study, is not detected on the promoter in the same ChIP experiments (Figure 5a). KLF10, the SP/KLF family member with highest similarity to KLF11, is part of the same TIEG (transforming growth factor β-inducible early gene) subfamily , but does not alter KRAS-mediated transformation or proliferation in the present study, therefore supporting the specificity of binding to this promoter. Subsequently, we performed genetic-rescue experiments in which cyclin A2 was overexpressed from a heterologous promoter along with KLF11 in PANC1 cells. Under these circumstances, KLF11 arrests cell growth (22±12.9% normalized to EV), however, the expression of exogenous cyclin A2 rescues the effect of KLF11 (88±11.3% normalized to EV), indicating that the cell-cycle regulatory function of this protein depends on its repression of the CCNA2 promoter (Figure 5b). KLF11 contains three distinct repressor domains, R1, R2 and R3, located in the N-terminal domain of the protein [8,11]. Consequently, using deletion mutants, we investigated which of these domains is involved in CCNA2 promoter repression. KLF11 transcriptional repression depends on the specific interaction of co-repressor molecules with each of these domains [9,10]. Specifically, the interaction of the R1 domain with the Sin3A–HDAC (histone deacetylase) co-repressor complex is required for the transcriptional regulatory activity of certain genes. Regarding how these gene regulatory pathways have an impact on cell biology, our laboratory has described previously that many KLF11 biological functions require its gene-silencing activity through this co-repressor complex [9,18]. Interestingly, however, in the context of cyclin A2 expression, further characterization of the mechanism revealed that KLF11 requires the intact full length of the protein in order to repress this promoter activity in cancer cells (Figure 5c). In fact, none of the deletion mutants had any statistically significant difference with the control values. Thus this result suggests that KLF11 uses a novel mechanism to achieve this function, which involves all of its transcriptional repressor domains. Since other co-repressors of KLF11 function currently remain unknown, the Sin3-independent regulation of this promoter by KLF11 raises an interesting area of future investigation, focused on how distinct chromatin-mediated events triggered by this KLF protein results in the repression of the CCNA2 promoter and suppression of KRAS-mediated transformation.
KLF11 inhibits cell proliferation via direct binding and repression of the CCNA2 promoter in a Sin3/HDAC-independent manner
To specifically characterize the KLF11-mediated regulation of cyclin A2, we first analysed the CCNA2 promoter for KLF-binding sites via bioinformatics using pairwise comparison with KLF consensus sites defined according to rules derived from crystal structures [16,19], as well as empirical DNA-binding matrices developed by our laboratory [18,20] to find four potential sites (Figure 6a). To determine which of these candidate sites KLF11 binds, we performed EMSAs. Interestingly, the specific binding of recombinant KLF11 was observed only for site #1, corresponding to −87 through to −92 of the CCNA2 promoter (Figure 6b, lane 1 and 6c, lane 4). The binding of recombinant KLF11 to this sequence was found to be specific by super-shift assays (Figure 6c, lanes 5 and 6). Furthermore, competition was performed with a cold wild-type, but not mutant, probe confirming a high specificity in binding (Figure 6c, lanes 7 and a). Binding was also not observed with the mutant probe to this site (Figure 6c, lanes b–d; nucleotides corresponding to −89 to −91 CGC>TTT). In addition we performed site-directed mutagenesis on the CCNA2 promoter to change the corresponding KLF11-binding site from CCCCGCCCC to CCCTTTCCC. Indeed, upon mutagenesis of this site, KLF11 was no longer able to repress promoter activity and was similar to the EV control (Figure 6d; 41.4±3.7% normalized to EV for wild-type promoter compared with 116±8.8% normalized to EV for the −89 to −91 CGC>TTT mutant promoter), confirming that this is indeed the site of KLF11 binding to the CCNA2 promoter and that this site is necessary for KLF11-mediated repression of this promoter. Taken together, these results indicate that KLF11 can inhibit tumour-cell proliferation and therefore transformation, at least in part, through the modulation of cyclin A2, and that this mechanism requires full-length KLF11 protein, which directly binds to −87 through to −92 of the CCNA2 promoter.
KLF11-mediated repression requires an SP/KLF site at −87 through to −92 of the CCNA2 promoter
KLF proteins are eliciting significant attention because of their misexpression in different cancers, as well as their role in cell proliferation, apoptosis, senescence, migration, adhesion and angiogenesis [2,21]. Many of these tumours and the cell processes mediated by KLF proteins share the activation of the KRAS pathway in common, either by signalling or mutations; however, the role of KLF proteins in KRAS-mediated transformation has remained poorly understood. The present study was designed to fill this existing gap in knowledge and outlines novel mechanistic information on how KLF-mediated transcriptional pathways can regulate cancer-associated processes. Using well-established assays, we have performed the first family-wide functional screening for KLF proteins, identifying a subset that inhibits the ability of KRAS to mediate neoplastic transformation. Our functional screening yielded SP5, SP8, KLF2, KLF3, KLF11, KLF13, KLF14, KLF15 and KLF16 as novel SP/KLF molecules with growth-suppressor properties in the context of the an oncogenic KRAS pathway. This SP/KLF subgroup has suppressive activity on KRAS transformation similar to our positive control KLF4 . KLF5, which exhibits well-characterized effects opposite to those of KLF4 , did not suppress KRAS transformation, thereby serving as an internal negative control for the assay. Thus the present study reveals that mechanistically at the cellular level, the suppression activities of these KLF proteins shared at least two functional mechanisms (Figure 3), namely decreased cell proliferation and enhanced apoptosis. Collectively, these experiments identify a subset of KLF proteins that antagonize KRAS-mediated transformation and outline two cellular mechanisms that primarily account for this effect.
Insights into the molecular mechanisms shared by suppressor KLF genes was derived from a more focused functional characterization of KLF11, chosen as a model for this group of proteins. We selected this gene for further study because of its growing medical relevance, owing to recent reports describing its methylation-dependent inactivation and down-regulation in several malignancies including leukaemia, myeloproliferative disorders, oesophageal adenocarcinoma, pancreatic cancer and germ cell tumours, as well as head and neck cancer, supporting its candidacy as an actual tumour suppressor in humans [17,24–28]. Congruent with this idea, our investigations at the cellular level show that KLF11 not only inhibits the KRAS-mediated transformed phenotype in foci formation and agar colony formation in vitro, but also the in vivo growth of KRAS-mutant xenografts (Figure 2). At the molecular level, we find that KLF11 and two other KLF suppressor proteins, KLF14 and KLF16, share the ability to down-regulate cyclin A2 at the transcriptional level. This down-regulation of cyclin A2 is congruent with the existence of multiple potential cis-regulatory SP/KLF sites in the promoter of this gene as defined by bioinformatics and based upon DNA-binding matrices with stringent theoretical and empirical support [18,20]. As demonstrated for KLF11, down-regulation of cyclin A2 occurs via binding of this SP/KLF transcription factor to the CCNA2 promoter to silence its expression. Similar to what happens with other tumour-suppressor proteins, down-regulation of cyclin A2 triggers cell-cycle arrest in S-phase . A rescue experiment, performed by the concomitant overexpression of cyclin A2 abolishes this effect, further supporting the veracity of this phenomenon (Figure 5b). Therefore, together with our cell biological data, this molecular information demonstrates that some of these KLF proteins have the ability to antagonize KRAS-mediated transformation, at least in part, by inhibiting cell growth at S-phase via a shared biochemical mechanism, namely the down-regulation of cyclin A.
Careful molecular dissection of CCNA2 promoter repression revealed that the full-length KLF11 protein was required for its repressive effect and specific binding of KLF11 occurs at site #1, corresponding to −87 through to −92 of the CCNA2 promoter (Figure 6). In particular, the requirement for the full-length KLF11 protein, unlike previously characterized gene targets for KLF11-mediated repression [17,27], indicates that cyclin A2 repression is Sin3a–HDAC-independent, a complex that interacts with a defined 20 amino acid peptide located within the R1/SID (Sin3-interacting domain) region of KLF11. Notably, since other co-repressors of KLF11 function currently remain unknown, the Sin3-independent regulation of this promoter by KLF11 suggests a yet-undiscovered mechanism of repression among the three KLF proteins observed to have a significant effect on cyclin A2 expression, namely KLF11, KLF14 and KLF16. Rationally, the best shared characteristic of these proteins, outside of Sin3 binding, is that all possess a similar SP1-like DNA-binding domain. Consequently, the ability of KLF proteins to compete with each other through their zinc-finger domains for binding to the same promoter target may explain some measure of redundancy currently assumed in this field of research [3,5]. In this theoretical context, our results suggest that different KLF proteins might down-regulate cyclin A2 in different cells or their specificity, in this regard, might come from their relative abundance in distinct cell types. Moreover, since many KLF proteins are regulated at the transcriptional and post-translational level, some of these proteins may not be functional for regulating cyclin A2, in particular in cells where these mechanisms are operational in a context-dependent manner. Nevertheless, the biochemical ability of these proteins to silence cyclin A2 expression, as described in the present paper, should fuel future studies that will define the cell-specific and the biological context, based on expression and signalling, in which each suppressor KLF protein performs this function.
In conclusion, the present study identifies a subgroup of SP/KLF proteins that blocks KRAS-mediated neoplastic transformation, thus significantly expanding our understanding of the repertoire of proteins that can antagonize the function of the KRAS oncogene. Regarding SP/KLF proteins, the present study identifies not only new modulators of the neoplastic phenotype, but also provides novel mechanisms underlying the function of these proteins. At the cellular level, there appear to be a network of SP/KLF proteins with the ability to negatively regulate growth by the inhibition of cell proliferation and induction of apoptosis. At the molecular level, the results described in the present paper advance the field of SP/KLF proteins and their role in the control of cancer-associated processes, since a regulatory role for any SP/KLF on cyclin A2 expression has not previously been determined. Therefore, collectively, both the phenomenological and mechanistic discoveries of the present study further our understanding of the role of KLF proteins in cell-cycle control, tumour suppression and KRAS antagonism.
We thank Dr Eugenio Santos (National Institutes of Health, Bethesda, MD, U.S.A.) for the KRASV12 plasmid and Dr Robert Sheaff (University of Minnesota, Minneapolis, MN, U.S.A.) for the cyclin A2-expression construct.
green fluorescent protein
V-Ki-ras2 Kirsten rat sarcoma viral oncogene homologue
mouse embryonic fibroblast
transforming growth factor β-inducible early gene
Martin Fernandez-Zapico, Gwen Lomberk and Raul Urrutia designed all of the experimental work. Martin Fernandez-Zapico, Gwen Lomberk, Shoichiro Tsuji, Cathrine DeMars, Michael Bardsley, Yi-Hui Lin, Luciana Almada, Jing-Jing Han and Navtej Buttar participated in the experiments. Martin Fernandez-Zapico, Gwen Lomberk and Raul Urrutia analysed and interpreted the results, as well as directed and supervised the research. Michael Bardsley and Tamas Ordog performed and interpreted results of cell-cycle analysis. The paper was written by Gwen Lomberk and Raul Urrutia and revised before submission by Martin Fernandez-Zapico, Navtej Buttar, Debabrata Mukhopadhyay and Tamas Ordog.
This work was supported by the National Institutes of Health [grant number DK 52913] (to R.U.) and [grant number 76845] (to N.B.), and the Mayo Clinic Pancreatic SPORE (Specialized Program of Research Excellence) P50 [grant number CA102701] (to M.F.-Z.), as well as the Mayo Clinic Center for Cell Signaling in Gastroenterology [grant number NIDDK P30DK084567].
These authors contributed equally to this work.