Krüppel-like Factor 6 (KLF6) Affects the Promoter Activity of the α1-Proteinase Inhibitor Gene

Abstract

purpose. Keratoconus is a progressive disease that thins and scars the cornea. In keratoconus corneas, levels of degradative enzymes, including lysosomal acid phosphatase (LAP) and cathepsin B, are elevated, and those of inhibitors α1-proteinase inhibitor (α1-PI) and α2-macroglobulin (α2-M) are reduced. The present study explored the possible involvement in keratoconus of Krüppel-like factor 6 (KLF6), a transcription factor previously described to be essential for the integrity of the corneal epithelium. The transcript and proteins level of KLF6 and its action in regulating the genes affected in keratoconus were examined in this study.

methods. Semiquantitative RT-PCR, Western blot analysis, immunofluorescence and in situ hybridization were used to investigate the expression of KLF6 mRNA and protein in normal and keratoconus corneas. Modulation by KLF6 of the promoter activity of α1-PI, LAP, cathepsin B, and α2-M genes was studied after transient transfection of KLF6 expression plasmid into corneal epithelial cells using promoter–reporter gene assays. Chromatin immunoprecipitation (ChIP) assays were performed to confirm the interactions between KLF6 and promoters of the genes affected in keratoconus.

results. A global increased expression of the transcription factor KLF6 in terms of mRNAs and proteins was observed in total cornea and/or the epithelium in a substantial number of the keratoconus specimens. The promoter activity of the human α1-PI gene was suppressed by expression of KLF6 in corneal epithelial cells. The ChIP assay confirmed a physical interaction between KLF6 and the α1-PI promoter.

conclusions. Transcription factor KLF6 downregulates the α1-PI gene in corneal epithelial cells and may thereby be involved in keratoconus.

Keratoconus is a noninflammatory disorder that progressively thins and distorts the central portion of the cornea and leads to visual impairment. ¹ The reported incidence varies widely between 60 and 230 per 100,000 of the general population. ² The exact cause of this disease is still not clear; although one notable hypothesis is that the abnormality in keratoconus may lie in the degradative pathway of macromolecular constituents in the cornea. ³ Subsequent data have strongly supported the degradation hypothesis, demonstrating increased levels of degradative enzymes and decreased amounts of protease inhibitors in keratoconus specimens compared with normal and other diseased corneal controls. ^{4 5 6 7}  

The major degradative enzymes found to be enhanced in corneas afflicted with keratoconus are lysosomal acid phosphatase (LAP) and cathepsins B and G. ⁷ The inhibitors that are reduced include α1-proteinase inhibitor (α1-PI) and α2-macroglobulin (α2-M). ^{4 8} The up- and downregulation of these genes are especially seen in the epithelial layers and are described at both mRNA and protein levels. ⁵ In view of the multiple gene involvement and the possibility of coordinated gene regulation, several transcription factors were examined. To date, Sp1 is the unique transcription factor found to be specifically upregulated in keratoconus-affected corneas. ⁹ This factor is further shown to downregulate the promoter activity of α1-PI but has little bearing on promoter activities of other enzyme and inhibitor genes affected in keratoconus. ^{10 11}  

Krüppel-like factor-6 (KLF6, also known as Zf9 or CPBP) is a member of the family of Krüppel-like factors. This family, closely related to the Sp1-like transcription factor family, is composed of 16 members sharing a highly conserved C-terminal DNA-binding domain containing three zinc fingers. ^{12 13} The three carboxyl-terminal C₂H₂ zinc fingers interact directly with the promoter of target genes through a GC box element. Cloned originally from cDNA libraries of placenta, human KLF6 mRNA is ubiquitously expressed, with a high level of expression in the liver, lung, intestine, prostate, and placenta. ^{14 15} KLF6 is shown to play a crucial role in the regulation of genes involved in tissue development, differentiation, angiogenesis, hematopoiesis, cell cycle control, and proliferation. ¹⁶ Target genes of KLF6 reported so far include collagen α1 (I), ¹⁷ keratin 4, ¹⁸ placental glycoprotein PSG5, ¹⁴ urokinase, ¹⁹ transforming growth factor-β (TGFβ) type I and II receptors, ²⁰ a collagen-specific molecular chaperone heat shock protein 47, ²¹ nitric-oxide synthase, ²² leukotriene C4 synthase, ²³ and endoglin. ²⁴ KLF6 is also shown to function as a tumor-suppressor gene linked genetically to prostate cancer or hepatocellular carcinoma. ^{25 26}  

Recently, KLF6 was reported to be expressed in human corneal epithelial cells ^{27 28} and during mouse eye development. ²⁹ It was also demonstrated that KLF6 upregulates the transcription of the keratin 12 (K12) gene by binding directly to a CTCCACCCA sequence at the promoter region of the K12 gene in human corneal epithelial cells. ²⁷ Because KLFs and Sp1-like factors are highly related zinc-finger proteins, ¹³ and in light of the Sp1 findings, the present study was undertaken to determine whether KLF6 mRNA and protein levels are altered in keratoconus corneas. Experiments were also performed to examine whether KLF6 regulates the expression of keratoconus-related genes including LAP, cathepsin B, α1-PI, and α2-M in corneal epithelial cells. 

Materials and Methods

Tissue and Cell Collection

All tissue and cell collections were in compliance with the Declaration of Helsinki on research involving human subjects. Informed consents were obtained from subjects after the nature and possible consequences of the study had been explained to them. To obtain the number of samples necessary for all the experimental assays, the collection was undertaken by both teams in Clermont-Ferrand and Chicago. The research was approved by the institutional human experimentation committee (CCPPRB) of Clermont-Ferrand and by the University of Illinois at Chicago Institutional Review Board. 

At the Clermont-Ferrand site, 15 human corneas not used for transplantation were collected from donors aged 18 to 45 years (from the regional eye bank). Fifteen keratoconus corneas were also collected from 15 patients aged 24 to 37 years who underwent corneal transplantation in the Department of Ophthalmology at University Hospital at Clermont-Ferrand. The human epithelial cells were taken from 20 eyes of 20 patients with astigmatism or myopia (as the normal group) who underwent epithelial ablation for photorefractive keratectomy with an excimer laser. Just before keratectomy, an area 8 mm in diameter encompassing the whole corneal epithelial layer was scraped mechanically, immediately frozen in liquid nitrogen, and stored at −80°C until use. Twenty epithelia of keratoconus corneas were also collected. In addition, stromal tissues from 20 normal and 20 keratoconus corneas were isolated. The clinical diagnosis of corneal diseases was confirmed by standard histopathologic examinations. The corneal cells and tissues were used for semiquantitative RT-PCR and Western blot assays. 

In parallel, at the Chicago site, a total of 28 normal human donor eyes (age range, 6–74 years) were obtained within 48 hours of death from the Illinois Eye Bank (Chicago, IL) or the National Disease Research Interchange (Philadelphia, PA). Half corneal buttons from 35 patients with keratoconus (age range, 22–70 years) and 37 patients with other corneal diseases (22 aphakic/pseudophakic bullous keratopathy [ABK/PBK], 14 Fuchs’corneal dystrophy, and 1 corneal scar) were obtained from either the Corneal Service at the University of Illinois at Chicago or the Corneal Associates of New Jersey (Theodore Perl) through the National Keratoconus Foundation. The diagnosis of corneal diseases was again confirmed by clinical and histopathologic examinations. Corneas collected were used for in situ hybridization, immunofluorescence staining, and/or Western blot analysis assays. 

RNA Extraction and Semiquantitative Reverse Transcription-PCR

Total RNA was extracted from human total cornea, stromal, and corneal epithelial cells with a purification kit (Quickprep MicroRNA), according to the instructions of the manufacturer (GE Healthcare, Les Ulis, France). The RNA quantity was determined by spectrophotometer measurement at 260 and 280 nm (ratio with proteins). The RNA quality was studied by the RNA/protein ratio (260:280 nm) and by gel electrophoresis (2% agarose), to observe the presence of intact 28S and 18S RNA bands. In a first step, cDNA was generated (from 2.5 μg of total RNA) with random primers (Superscript First-Strand Synthesis System for RT-PCR; Invitrogen-Gibco, Cergy-Pontoise, France). In a second step, the specific oligonucleotide primers used for the PCR reaction were originally generated using the web program “Primer3” (http://www-genome.wi.mit.edu/cgi-bin/ primer/primer3_www.cgi) based on the published full-length human mRNA sequences of each specific gene: KLF6, sense5′-GGCAACAGACCTGCCTAGAG-3′, antisense 5′-AGGATTCGCTGCTGACATCT-3′; and β-actin, sense 5′-TCCCTGGAGAAGAGCTACGA-3′ and antisense 5′-AGCACTGTGTTGGCGTAC AG-3′. The expected molecular sizes for KLF6 and β-actin are 165 and 180 base pairs, respectively. The sequences of the PCR products were confirmed using the DNA dye terminator cycle sequencing kit (Applied Biosystems, Courtaboeuf, France) with DNA sequencer (model 377; Applied Biosystems). 

Semiquantitative amplifications were performed per recommendations of the manufacturer (Light Cycler analyzer; Roche Molecular Diagnostics, Meylan, France). After 10 minutes of denaturation at 94°C, 15 μL of cDNA and PCR mix were submitted to 50 cycles of PCR amplifications (1 minute, 94°C; 45 seconds, 53°C; and 45 seconds, 72°C) followed by a final step of elongation (7 minutes, 72°C). The protocol used Sybr Green and based on ΔC_T (delta cycle threshold) determination. The formula used to derive the multiple of change (x-fold) in RNA content was ΔC_T normal/ΔC_T KC. Each determination was realized in triplicate. The KLF6/actin mRNA ratios obtained from normal tissues and cells were normalized to 1.0. All other values were expressed against the normal ones to depict the x-fold differences between normal and pathologic samples. 

In Situ Hybridization

Paraffin sections were prepared from six normal human corneas (donor ages: 27–74 years), and corneas from five patients (ages: 44–54 years) with keratoconus, four patients (ages: 55 to 76 years) with ABK/PBK and two patients (ages: 82 and 82 years) with Fuchs’ corneal dystrophy. In situ hybridization was performed with digoxigenin-labeled KLF6 sense and antisense riboprobes as previously described. ²⁹  

Immunofluorescence Staining

Cryosections (8 μm) of corneas from six normal human donors aged 27 to 64 years, seven patients with keratoconus aged 23 to 65 years, three patients (42–77 years) with ABK/PBK, one 80-year-old patient with Fuchs’ corneal dystrophy, and one 79-year-old patient with corneal scar were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) at 4°C for 30 minutes, rinsed with PBS three times, and incubated in 3% bovine serum albumin (BSA; Sigma-Aldrich, St. Louis, MO) at 25°C for 30 minutes. Immunohistologic staining was performed using epitope-specific polyclonal antibody anti-KLF6 (1:350, SC 7158; Santa Cruz Biotechnology, Santa Cruz, CA) for incubation overnight at 4°C, followed by incubation with FITC-conjugated goat anti rabbit IgG secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) for 3 hours at room temperature. 

For paraffin-embedded sections, corneas from five normal human donors (ages: 42–73 years), five patients with keratoconus (ages: 36–53 years), three patients with ABK/PBK (ages: 63– 91 years), and two patients with Fuchs’ corneal dystrophy (ages: 67 and 76 years) were collected, fixed in formalin, and processed. Immunostaining was performed as previously described. ²⁹ Briefly, with a heat-induced epitope retrieval technique with 10 mM sodium citrate buffer (pH 6.0; NeoMarkers, Fremont, CA), sections were incubated at room temperature sequentially with polyclonal rabbit anti-human KLF6 antibody (1:100) for 90 minutes, biotinylated goat anti-rabbit IgG (1:200; Jackson ImmunoResearch Laboratories) for 45 minutes, and a dichlorotriazinyl aminofluorescein (DTAF) conjugate (Extravidin; Jackson ImmunoResearch Laboratories) for 45 minutes. 

For a negative control, sections were incubated with normal rabbit IgG in place of anti-KLF6. The slides were examined by microscope (Axiophot; Carl Zeiss Meditec, Inc., Dublin, CA) after mounting in an aqueous mounting medium (Vectashield; Vector Laboratories, Burlingame, CA). 

Western Blot Analysis

Individually collected corneal tissues and cells were analyzed as previously described. ²⁷ For analyses of pooled samples, 11 normal human (donor ages: 6–69 years), 18 keratoconus (patient ages: 22–70 years), 12 ABK/PBK (patient ages: 37–85 years), and 9 Fuchs’ dystrophy (patient ages: 47–82 years) corneas were collected. These corneas were incubated at 37°C with 1 mL of 20 mM EDTA-PBS (pH 7.2–7.3) for 45 minutes. The corneal epithelium and the stroma were separated, individually frozen, and stored at −80°C. These samples were divided in groups such that 3 Western blot experiments could be performed. The corneal tissues were pulverized and homogenized in PBS and proteinase inhibitors. The supernatant was collected after centrifugation at 14,000g for 10 minutes. Proteins from samples in the same group were pooled and the protein concentration was measured using the BCA method with BSA as a standard. Total protein (25 μg for each epithelial and 75 μg for each stromal group) was electrophoresed under reducing conditions on 4% to 20% SDS gels. The proteins resolved were transferred onto nitrocellulose membranes. The membranes were blocked and probed for KLF6. For epithelial samples, the membranes were incubated with horseradish peroxidase (HRP)–conjugated rabbit anti-goat antibody (1:10,000 in 1% BSA; Jackson ImmunoResearch Laboratories,) and visualized by enhanced chemiluminescence (ECL; SuperSignal West Pico; Pierce, Rockford, IL). For stromal samples, the membranes were incubated with biotinylated goat anti-rabbit IgG (1:20,000 in 1% BSA; Jackson ImmunoResearch Laboratories) for 1 hour followed by incubation with HRP-streptavidin (1:10,000; Jackson ImmunoResearch Laboratories) for 1 hour and then visualized by ECL. The membranes were stripped at room temperature with IgG elution buffer (ImmunoPure; Pierce) overnight and washed extensively. The blots were treated with the chemiluminescent substrate (SuperSignal; Pierce) and exposed to confirm that all signals had been stripped. The membranes were then reprobed for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) using polyclonal anti-GAPDH (1:6000; Trevigen, Gaithersburg, MD) to calibrate for protein loading. The expected molecular sizes for KLF6 and GAPDH are 34 and 36 kDa, respectively. In some cases, two membranes were prepared; one for KLF6 and one for GAPDH probing to confirm the results. 

Human Corneal Epithelial Cell Culture

A human corneal epithelial cell line transformed with SV40 (ATCC/CRL11135) was cultured under standard conditions (5% CO₂, 95% humidified air, 37°C) in DMEM-F12 supplemented with 5% fetal calf serum, 5 μg/mL of insulin, 0.1 μg/mL of cholera toxin, 50 mg/mL of streptomycin, 50 IU/mL of penicillin, 0.5 mg/mL of epidermal growth factor, and 0.5% DMSO. ²⁷  

Transfection of Cultured Human Corneal Epithelial Cells

The DNA fragments containing the 0.67-, 0.44-, 1.4-, and 5.77-kb 5′-flanking sequences of LAP, cathepsin B, α1-PI, and α2-M genes were ligated into the secreted alkaline phosphatase (SEAP) series vector pSEAP-Basic (BD-Clontech, Palo Alto, CA) yielding pLAP0.67SEAP+, pCatB0.44SEAP+, pα1-PI1.4-SEAP+, and pα2M5.77SEAP+. The constructs were purified, sequenced, and restriction digested to confirm the identity and orientation of the inserts. ¹¹ The activity of the LAP, cathepsin B, α1-PI, and α2-M promoters in human corneal epithelial cells was investigated in transient transfection assays. The promoter plasmids and pSEAP-Basic (promoterless reporter vector, negative control) and pSEAP-Control (positive control driven by the SV40 early promoter) along with the pCMV-β-galactosidase control vector (for normalization of the transfection efficiency; Promega, Madison, WI), were used in cell transfection. Two human KLF6 expression vectors (generous gifts from Scott L. Friedman, Mount Sinai School of Medicine, NY) were used: one for the wild-type human KLF6 gene (pCI-neo-KLF6) and one for a mutated KLF6, unable to bind DNA, called X137 (pCI-neo-X137). 

Human corneal epithelial cells were trypsinized 24 hours before DNA transfection and plated at 40,000 cells/well on 24-well plates. Transient transfection of all DNA constructs was performed by a liposome-mediated method (Lipofectamine and Plus reagent; Invitrogen-Gibco), with 0.75 μg of the test plasmid pLAP0.67SEAP+, pCatB0.44SEAP+, pα1-PI1.4-SEAP+, and pα2M5.77SEAP+ and 0.2 μg pCMV-β-galactosidase. Another series of cells also received 0.5 μg of pCI-neo-KLF6 or the pCI-neo-X137 expression plasmid. None of the test plasmids was added to cells serving as negative control cultures. 

After 2 days, the medium (100 μL) was collected from each well for an SEAP assay according to the manufacturer’s protocol (Roche Molecular Diagnostics). The cells were washed twice with PBS, and treated with 700 μL of cell lysis buffer for 1 hour at 4°C. The lysed cells were centrifuged at 950 g for 5 minutes at 4°C. The determination of β-galactosidase production was performed by an immunoenzymatic assay on 100 μL of supernatants (Roche). Assays were performed in triplicate, and each experiment was repeated at least three times. Results refer to mean ± SEM, and are averages of nine values per experiment. Comparison of means was done by analysis of variance (ANOVA) and Fisher’s t-test (Statview II 1.03 software; Abacus Concepts Inc., Berkeley, CA). The values were considered to be significantly different when P < 0.05. 

Chromatin Immunoprecipitation Assays

Human corneal epithelial cells (10 × 10⁶) were transfected with the KLF6 expression vector, the mutated KLF6 (X137), or the empty plasmid (control). The chromatin immunoprecipitation (ChIP) assay was performed as previously described. ³⁰ Briefly, KLF6 was cross-linked to DNA in a 1% formaldehyde solution (1% formaldehyde, 0.5 mM EGTA, 50 mM HEPES [pH 8], 100 mM NaCl, and 1 mM EDTA) for 15 minutes, and washed twice with PBS and 0.125 M glycine solution. The samples were sonicated (10 seconds, kept on ice 30 seconds and repeated for five times) and then centrifuged at 1500g for 10 minutes at 4°C. The supernatants were diluted in five volumes of an immunoprecipitation (IP) buffer (0.1M KCl, 20 mM Tris-HCl [pH 8], 5 mM MgCl₂, 1 mM phenylmethylsulfonyl fluoride (PMSF), 10% glycerol, and 0.1% Tween 20) and subjected to immunoprecipitation by incubation with the anti-KLF6 antibody (SC 7158; Santa Cruz Biotechnology) overnight at 4°C, treatment with salmon sperm DNA-BSA-Sepharose beads, followed by treatment with 10 mM phenyl phosphate for 15 minutes. The resultant immunoprecipitates were centrifuged at 1000 rpm for 5 minutes at 4°C and washed five times with IP buffer. The final precipitated DNA was dissolved in 500 μL of elution buffer (1 M NaCO_3, 1 M NaHCO₃, and 10% SDS). The DNA released from the cross-linked material by heating at 65°C for 4 hours and by extensive digestion with RNase A (50 μg/mL at 37°C for 30 minutes) and proteinase K (0.5 mg/mL at 65°C for 3 hours) was purified using a DNA purification kit (Qiagen Inc., Valencia, CA). PCR was performed with primers flanking the two potential KLF6-binding sites in the α1-PI promoter (see Fig. 4B ) using the following PCR protocol: primary denaturation at 94°C for 3 minutes; 94°C for 1 minute, 60°C for 1 minute, 72°C for 1 minute, 34 times; and final elongation at 72°C for 7 minutes. Amplified fragments were analyzed on a 1.5% agarose gel. The sequences of the PCR products were confirmed using the DNA dye terminator cycle sequencing kit (Applied Biosystems) with the DNA sequencer (model 377; Applied Biosystems). The noantibody or nonspecific antibody (anti-digoxigenin) immunoprecipitation samples served as the negative control. 

Results

KLF6 mRNA Levels and Localization in Keratoconus Corneas

Semiquantitative RT-PCR assays demonstrated that the KLF6 mRNA levels were significantly increased in 15 keratoconus versus 15 normal corneas (2.8 fold increases, P < 0.0001, Fig. 1A ). The increase in corneas affected with other diseases such as Fuchs’ corneal dystrophy or edema, although statistically significant, was only 1.2-fold (P < 0.035, Fig. 1A ). Further analysis of cellular types constituting the cornea revealed that epithelial cells extracted from keratoconus corneas had a KLF6 mRNA level 8.7-fold higher (P < 0.0001) than normal corneal epithelial cells (Fig. 1B) . Levels of KLF6 mRNAs in keratocytes were modestly increased, respectively, by 1.4- and 1.5-fold (P < 0.035) in the stromas of keratoconus and other diseased corneas (Fig. 1C) . 

In situ hybridization experiments showed that KLF6 transcripts were detected in the endothelium, epithelium, and stroma of all normal and pathologic (keratoconus, Fuchs’ corneal dystrophy, and edema) corneas (data not shown). 

KLF6 Protein Localization and Levels in Keratoconus Corneas

Immunofluorescence with either cryo- or paraffin sections showed nuclear staining for KLF6 mainly in epithelial cells (Fig. 2) , keratocytes, and endothelial cells (data not shown) of human corneas. Cytoplasmic staining was also noted. The KLF6 intensity in the epithelial layer of keratoconus or other diseased corneas was either similar or stronger than that of normal human control subjects (Fig. 2) . Among the keratoconus specimens examined (n = 12), one third of the corneas were noted to have enhanced KLF6 staining. In the other disease group (n = 10), 20% of the corneas had intensified KLF6 staining in the corneal epithelium. The staining intensity in the stromal layer was variable, but essentially comparable in all corneas (Fig. 2) . 

Western blot analyses on individual corneas showed a 2.3-fold increase of KLF6 level in keratoconus compared to normal corneas (n = 15). KLF6 level in epithelial cells derived from keratoconus corneas was 3.8-fold (P < 0.001) higher than that from normal corneas (Fig. 3A) . Those derived from other diseased corneas by comparison had only a 1.4-fold (P < 0.025) higher KLF6 level than normal control subjects. When analyses were performed on pooled samples for increased sensitivity, heterogeneity concerning increases (or lack thereof) of KLF6 protein levels could be observed in the epithelial layer in keratoconus and other disease groups (Fig. 3B , see panels 1, 2, and 3). No increase of KLF6 protein was observed in any of the stromal samples (Figs. 3A 3B , panel 4). These results were in general consistent with the immunofluorescence data. 

hKLF 6 Represses α1-PI Promoter in Human Corneal Epithelial Cells

The role of human KLF6 (hKLF6) was investigated in the transcriptional regulation of keratoconus-related genes, including LAP, cathepsin B, α1-PI, and α2-M. When transfected into cultured corneal epithelial cells, the pLAP0.67SEAP+, pCatB0.44SEAP+, pα1-PI1.4-SEAP+, and pα2M5.77SEAP+, vectors were 20.1, 18.7, 22.3, and 28.3 times more active, respectively, at driving SEAP reporter gene expression than the SEAP-basic vector (Fig. 4) . When cotransfected with the hKLF6 expression vector, the level of SEAP expressions remain unaltered with all the constructs except pα1-PI1.4-SEAP+, with which a significant reduction (P < 0.05) was observed. Mutated KLF6 (expression plasmid containing X137 stop mutant that lacks the DNA-binding domain and that should not bind DNA promoter) failed to repress the α1-PI promoter–reporter activity (Fig. 4) . Likewise, a mutation of the potential KLF6-binding site abolished the α1-PI promoter–reporter construct (data not shown). These results indicated that hKLF6 could repress the α1-PI promoter activity in a region present in the pα1-PI1.4 construct. 

hKLF6 Interactions with DNA Sequence of the α1-PI Promoter

The 1400-bp sequence of the α1-PI promoter fragment ¹⁰ was examined for hKLF6-binding sites. Using a genomic database MatInspector Release Professional; Genomatix, Munich, Germany) we found two putative binding motifs in the pα1-PI1.4 construct: 5′-GTCCAAGCTCCCGCC CTC-3′ and 5′-CAGGTTGGAGGGGCGGCAACCT-3′, with, respectively 92% and 90%, similarity (Fig. 5A)with the consensus sequence of hKLF6-binding motif already described in other target genes. ¹⁴ To check the ability of these sequences to interact physically with hKLF6, ChIP assays were performed using human corneal epithelial cells transiently transfected with the hKLF6 expression vector. As shown in Figure 5B(lane 4), hKLF6 was bound to one amplified sequence containing 5′-GTCCAAGCTCCCGCCCTC-3′, demonstrating a physical interaction of hKLF6 and α1-PI promoter. This specific interaction was verified by the absence of sequence amplifications when empty vector (Fig. 5B , lanes 1 and 2), nonspecific antibody (lanes 1 and 3), or mutated KLF6 (expression plasmid containing an X137 stop mutant that lacks the DNA-binding domain and should not bind the DNA promoter, lane 5) were used. 

Discussion

In the present study, we demonstrated for the first time that the level of transcription factor KLF6 was increased in the cornea of a number of keratoconus cases. It is of note that while the transcript of KLF6 was relatively consistently upregulated in the epithelium and the stroma (Fig. 1)of keratoconus corneas, the transcript level was not always directly applicable to the protein level. Our results showed that an increase in KLF6 protein was only found in the epithelial layer, not in the stroma of keratoconus corneas. The increase of protein in the epithelium was also somewhat variable, not in all but in some keratoconus corneas. The heterogeneity of clinical context (evolution step, treatment, and/or other concomitant corneal alterations) present at the time of biological collection of keratoconus cells or tissues could to some degree explain the variation of translation rate for KLF6 transcripts. The turnover of the KLF6 protein may also be another factor. By Western blot analysis, a similarly variable increase in KLF6 protein was also detected in some of the other diseased epithelia. The induction in disease conditions is not surprising, as KLF6 has been reported to be induced in cells in response to injury such as occurs during hepatic fibrosis ¹⁷ or vascular surgery. ²⁴ KLF6 is also known to be a labile protein in vitro that disappears quickly after withdraw of stimuli. ^{19 24}  

We also provided evidence that an augmented level of KLF6 directly repressed the promoter activity of the α1-PI gene in corneal epithelial cells. We thus established a new target gene of KLF6, adding to those already described for this transcription factor: collagen α1(I), ¹⁷ keratin-4, ¹⁸ glycoprotein PSG5, ¹⁴ urokinase type plasminogen activator, ¹⁹ human immunodeficiency virus placental long terminal repeat, ³¹ TGFβ type I and II receptors, ²⁰ K-12, ²⁷ insulin like-growth factor 1 receptor, ³² nitric oxide synthase, ²² and endoglin. ²⁴ α1-PI, a major protease inhibitor in human serum, is a member of the serine proteinase inhibitors superfamily (serpins), critical in preventing and controlling proteolysis. ³³ One of its primary physiologic roles is to protect the elastic fibers in lung alveoli from excessive digestion by neutrophil elastase. The importance of this protein was evidenced by the observations that genetically α1-PI-deficient patients have early-onset degenerative lung disease or a liver disease. ³⁴ The liver is the predominant site of α1-PI synthesis, but the protein is also synthesized in blood monocytes and macrophages, alveolar macrophages, intestinal epithelial cells, human breast carcinoma, ³⁴ and human corneal cells. ³⁵ In keratoconus corneas, the level of α1-PI is reduced. ⁴ Such a reduction, together with increased expression of degradative enzymes (LAP, cathepsins, ³⁶ matrix metalloproteinase, ³⁷ ) and the decreased level of another protease inhibitor, α2-M, is believed to contribute to accelerated degradation of macromolecules, leading to thinning in the keratoconus corneas. ⁷  

Promoter studies showed that the promoter activity of α1-PI in corneal epithelial cells and corneal fibroblasts is downregulated by overexpression of transcription factor Sp1. ^{10 11} As Sp1 is demonstrated to be upregulated in keratoconus, the Sp1-mediated downregulation of α1-PI is speculated to be a key event in the development of keratoconus. ^{10 11} The most proximal Sp1 site (−100/−87) in the promoter fragment is mapped to be an essential element involved in the negative regulation of the α1-PI promoter activity by Sp1. ³⁸  

Our KLF6 results suggest that this transcription factor may contribute to the pathophysiology in at least some keratoconus corneas. The increase of KLF6 protein in the epithelial layer argues for and supports the previous notion of epithelial involvement in keratoconus. ^{4 39 40} The increased KLF6 may independently or in conjunction with Sp1 suppress the expression of the α1-PI gene, escalating pathologic conditions. Direct physical interaction between KLF6 and Sp1 has been documented by coimmunoprecipitation, pulldown experiments, and the GAL4 one hybrid system. Transfected KLF6 and Sp1 have also been shown to regulate cooperatively the expression of genes including endoglin and collagen α1 (I). ²⁴ Our preliminary result suggested that KLF6 could act synergically with Sp1 on the α1-PI promoter activity in corneal epithelial cells. The identified binding site of KLF6 is located at the most proximal and the key Sp1 binding site previously described in the α1-PI promoter, ¹⁰ highlighting again the possibility of KLF6 and Sp1 molecular interactions for the transcriptional regulation of this gene. The decreased α1-PI level detected in keratoconus could thus be a combinatory result of negative regulation by KLF6 and Sp1. 

Furthermore, KLF6 has been described as a regulator for the transcription of TGFβ, ²⁴ platelet-derived growth factor (PDGF), ⁴¹ and inducible nitric oxide synthase. ²² Of interest, cellular levels of both TGFβ and PDGF were found to be increased in keratoconus corneas. ^{42 43} Although testing the oxidative stress hypothesis, expression of inducible nitric oxide synthase was noted by Buddi et al. ⁴⁴ in diseased corneas but not in normal human control subjects. Together, the findings suggest that KLF6 could have an upstream role for regulation of these genes in keratoconus. 

In conclusion, we demonstrate for the first time that KLF6 (an Sp1-like member of the zinc finger Krüppel-like transcription factor family) is capable of downregulating the α1-PI gene and that an increased KLF6 protein level may be implied in the pathophysiology of keratoconus. The current data, along with the previously described importance of KLF4, -5, and -13 in corneal physiology ²⁸ strongly imply a role for KLF factors in the cornea. This, plus the implication that KLF15 is a participant in retinal physiology ⁴⁵ also underscores a possibility that KLF family members may be key players in the global eye physiology and pathology. 

 Supported by INSERM grant (FC, GM, VS) and EY03890 (BYJTY), EY05628 (BYJTY), and core Grant EY01792 from the National Institutes of Health, Bethesda, Maryland.

 Submitted for publication May 4, 2005; revised September 22 and October 14, 2005; accepted December 2, 2005.

 Disclosure: F. Chiambaretta, None; H. Nakamura, None; F. De Graeve, None; H. Sakai, None; G. Marceau, None; Y. Maruyama, None; D. Rigal, None; B. Dastugue, None; J. Sugar, None; B.Y.J.T. Yue, None; V. Sapin, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 4.

 

Repression of α1-PI promoter activity in human corneal epithelial cells by KLF6. The human wild-type KLF6 (pCI-neo-KLF6) or mutated KLF6 (pCI-neo-X137) expression vector was cotransfected into 50% confluent human corneal epithelia by using lipophilic substrate, together with the β-galactosidase expression plasmid and the promoter SEAP reporter plasmids: LAP (pSEAP-LAP), cathepsin B (pSEAP-CatB), α2-M (pSEAP-α2M), and α1-PI (pSEAP-α1-PI). SEAP activity was normalized by β-galactosidase activity. Experiments were repeated six times. Error bars, SD. For each reporter plasmid, *significant difference (P < 0.05) between the different conditions.

Figure 1.

 

Relative quantitative RT-PCR for KLF6 transcripts in normal, keratoconus, and other diseased corneas. For comparison of mRNA levels of KLF6 in normal and keratoconus corneas, semiquantitative RT-PCR assays were performed on extracts of 15 whole corneas (A), 10 corneal epithelia (B), and 10 corneal stromas (C). The KLF6 mRNA quantities against those of the β-actin mRNA are shown as mRNA ratios. The normal values were normalized to 1.0, and other values were expressed relative to the normal ones. Fuchs’ dystrophy and edema corneas were used as the other diseased group. Experiments for determination of KLF6 transcript in each sample were repeated three times. Error bars, SD. Statistically significant differences between normal and keratoconus tissues or cells were noted with *P < 0.035 or **P < 0.0001.

Figure 2.

 

Immunofluorescence staining of KLF6 protein in normal, keratoconus, and other diseased corneas. KLF6 protein expression was examined in normal human (NH) corneas and in those affected by keratoconus (KC), pseudophakic bullous keratopathy (PBK), or Fuchs’ corneal dystrophy (Fuchs’) by immunofluorescence using primary antibody against human KLF6 or normal rabbit IgG (negative control). ²⁹ Paraffin-embedded sections were used. Epi, corneal epithelium; S, corneal stroma. Bar, 50 μm.

Figure 3.

 

Western blot analyses of KLF6 protein in normal, keratoconus, and other diseased corneas. (A) To compare proteins levels of KLF6 in corneal samples, Western blot assays were performed on extracts of 15 individual total corneas (part 1), 10 corneal epithelia (part 2), and 10 corneal stromas (part 3). The KLF6 protein quantity was normalized against that of the GAPDH protein and shown as protein ratios. Values from pathologic samples were expressed against those of normal subjects. Experiments for determination of KLF6 protein in each sample were repeated three times. Error bars, SD. Significant differences between normal and keratoconus tissues or cells were noted with *P < 0.025 or **P < 0.0001. (B) Western blot analysis assays illustrating the heterogeneity of the KLF6 protein level variations, in pooled epithelia (parts 1, 2, and 3) and stromas (part 4) from normal, keratoconus, and other diseased (Fuchs’ corneal dystrophy and edema) corneas. The membranes were probed with anti-KLF6 and anti-GAPDH. Protein ratios are shown.

Figure 5.

 

Binding of KLF6 with the α1-PI gene promoter. (A) Partial DNA sequence of the 3-kb fragment of the α1-PI gene promoter. Nucleotide numbers are related to the published sequence (HSU88408). Putative KLF6 binding sites are in italic. Underscored and bold sequences: specific sense and antisense primers used to amplify DNA regions containing the two putative KLF6 binding sites. One KLF6 binding site contains the most proximal Sp1 binding site (Sp1-1) previously described ¹⁰ to be an essential element in the negative regulation of α1-PI promoter activity. (B) ChIP assays are known to determine the KLF6 ability to bind a specific sequence of the α1-PI gene promoter. Two regions were examined: KLF6 binding sites 1 and 2 (respectively, lanes 1–5, and lane 6) using specific primers described in (A). Lanes 1 and 2: negative controls using empty vector; lanes 1 and 3: nonspecific antibody; and lane 5: mutated KLF6.