Involvement of Sp1 Elements in the Promoter Activity of Genes Affected in Keratoconus

Abstract

purpose. Keratoconus is a progressive disease that thins and scars the corneal stroma. In keratoconus corneas, levels of degradative enzymes, including lysosomal acid phosphatase (LAP) and cathepsin B, are elevated, and those of the inhibitors α1-proteinase inhibitor (α1-PI) and α2-macroglobulin (α2-M) are reduced, especially in the epithelial layer. An increased expression of the transcription factor Sp1 was also demonstrated. The role of Sp1 in regulation of the genes affected in keratoconus was examined in this study.

methods. DNA segments, containing 5′-flanking promoter sequences of the α1-PI, LAP, cathepsin B, and α2-M genes were ligated into the secreted alkaline phosphatase (SEAP) reporter gene vector. These constructs, along with the pSVβ-galactosidase control vector, were transfected into cultured human corneal epithelial and stromal cells and skin fibroblasts. Cotransfection with the Sp1 expression vector was performed in parallel. SEAP and β-galactosidase enzyme activities were assayed.

results. In corneal epithelial cells, as in stromal cells, α1-PI promoter activity was suppressed by cotransfection of pPacSp1. The LAP, cathepsin B, and α2-M promoters were functional in corneal cells, whereas activities of these promoters were much lower in skin fibroblasts. Cotransfection experiments indicated that the up- or downregulation of LAP, cathepsin B, and α2-M observed in keratoconus-affected corneas was not mediated by Sp1.

conclusions. These results support the theory that the corneal epithelium, along with the stroma, is involved in keratoconus. An upstream role of Sp1 is indicated and the Sp1-mediated downregulation of the α1-PI gene may be a key event in the disease development.

Keratoconus is a slowly progressive corneal disease. Histopathologic studies have demonstrated that, in early stages, fragmentation of the epithelial basement membrane occurs, with disintegration of Bowman’s layer and fibrillation of the anterior stroma. In later stages, the central cornea is thinned, with destruction of Bowman’s layer and stromal scarring. ^{1 2 3} Keratoconus affects 0.004% to 0.6% ^{1 2 3} of people in the general population and leads to visual handicap in the productive second and third decades of life. No specific treatment of this almost universally bilateral disease exists, except to replace the corneal tissues by surgery when visual acuity is impaired beyond correction with contact lenses. ^{4 5} Most cases reported are isolated, although a positive family history has been described in 6% to 8% of the cases. ^{1 3} Keratoconus has also been associated with eye rubbing, ⁶ contact lens wear, ⁷ Down syndrome, atopy, and connective tissue diseases. ^{1 2 3}  

The exact cause of this disease is still not clear. Previous biochemical studies by our group ⁸ and others ⁹ have shown that the amount of total protein present in keratoconus corneas is lower than that in normal control corneas, whereas protein synthesis proceeds normally in some cases. ^{8 9 10} This led to the formulation of the hypothesis that the abnormality in keratoconus may lie in the degradative pathway of macromolecules. ^{8 10} Subsequent data support the degradation hypothesis, demonstrating increased levels of degradative enzymes ^{11 12} and decreased amounts of protease inhibitors ^{13 14} in keratoconus specimens compared with those of normal and other disease control specimens. The defects were observed not only in the stroma, but also, and even more prominently, in the corneal epithelium, implying a role for this layer in the disease process. 

The degradative enzymes found to be enhanced in corneas with keratoconus include lysosomal acid phosphatase (LAP), acid esterase, acid lipase, ¹¹ and cathepsins B and G. ¹² The inhibitors that are reduced include α1-proteinase inhibitor (α1-PI) and α2-macroglobulin (α2-M). ^{13 14} The up- or downregulation of these genes was noted at both protein and mRNA levels. ¹⁵ In view of the multiple gene involvement and the possibility of a coordinated gene regulation mechanism, several transcription factors were examined. Sp1 was found to be specifically upregulated in keratoconus-affected corneas. ¹⁶  

A putative promoter fragment of the human α1-PI gene was cloned and sequenced in our laboratory. ¹⁷ It is of interest that coexpression of Sp1 in corneal stromal cells suppresses the α1-PI promoter activity, ¹⁷ indicating that Sp1 upregulation may be directly related to α1-PI downregulation in corneas with keratoconus. 

The present study was undertaken to determine whether in corneal epithelial cells, as in stromal cells, the α1-PI promoter is functional and Sp1 sensitivity exists. In addition, experiments were performed to examine whether altered expression of genes including LAP, cathepsin B, and α2-M in keratoconus corneas are related to the upregulated Sp1 level in both normal human corneal stromal and epithelial cells. Promoter fragments of these genes, all of which contain Sp1 sites, ^{18 19 20 21} were cloned into reporter vectors. Their functional activities in corneal cells and skin fibroblasts were measured and the effects of Sp1 coexpression were evaluated. 

Materials and Methods

Tissue Culture

Primary human corneal epithelial cells (donor age, 19 years) were purchased from Cascade Biologics, Inc. (Portland, OR). These cells were grown and maintained in supplemented medium (EpiLife HCGS; Cascade Biologics, Inc.), containing bovine pituitary extract, bovine insulin, hydrocortisone, bovine transferrin, and mouse epidermal growth factor. First- or second-passage cells were used for transfection experiments. 

Corneal stromal fibroblasts were cultured from normal human corneas (donor ages: 10, 19, 35, and 44 years) obtained from the Illinois Eye Bank, Chicago. Normal human skin fibroblasts were purchased from American Type Culture Collection (Manassas, VA; donor age, 28 years) as a nonocular control cell type. These cells were grown and maintained in Dulbecco’s modified Eagle’s medium supplemented with glutamine, 10% (vol/vol) fetal calf serum, nonessential and essential amino acids, and antibiotics, as previously described. ²²  

The tenets of the Declaration of Helsinki for research involving human subjects were followed, and the research was approved by the University of Illinois Institutional Review Board. 

Construction of Promoter Fragments

DNA fragments containing the 2.7- and 1.4-kb 5′-flanking sequence of the α1-PI gene were ligated into one of the secreted alkaline phosphatase (SEAP) series vectors (pSEAP2-Basic; Clontech, Palo Alto, CA), as previously described, ¹⁷ yielding pα1PI2.7-SEAP+ and pα1PI1.4-SEAP+ constructs. Promoter fragments for LAP, cathepsin B, and α2-M genes were made by long PCR, using gene-specific primers selected through the computer software (Vector NTI; InforMax, Ltd., Oxford, UK) based on the known promoter sequences. ^{18 19 20 21} The primers for LAP (exon 1) were upstream (US), TTGTGCAGGGCAGGAACGGTA, and downstream (DS), GCGGCATCACCACCAGGTT; for cathepsin B, US, GATCCCAGGCGCGGGTTCTG, and DS, TTGGCGTTGCCGGAGCGGTT; and for α2-M, US, TCTGTAGCAAACATAGGATC, and DS, TCTGGTCCCAAACACTTCCC. All primers were synthesized by Genemed Biotechnologies, Inc. (South San Francisco, CA). The PCR products were analyzed on a 1.0% agarose gel and were cloned into a vector (pGEM-T Easy; Promega, Madison, WI). The expected sizes of the PCR products were 0.67 (−586 to +79), 0.44 (−361 to +74), and 5.77 (−5761 to +12) kb, containing two, nine, and two putative Sp1 binding sites, respectively. A shorter cathepsin B promoter region fragment (0.26 kb;− 190 to +74, containing seven putative Sp1 sites) was made from the 0.44-kb cathepsin B DNA fragment by BssHII digestion. 

The LAP, cathepsin B, and α2-M promoter fragments subcloned into the pGEM-T Easy vector (Promega) were ligated into the EcoRI multiple cloning sites of the pSEAP2-Basic vector (Clontech), yielding the pLAP-SEAP+, pCatB0.44-SEAP+, and pα2M-SEAP+ vectors. The shorter cathepsin B promoter fragment was ligated with HindIII linker at both ends and subcloned into multiple cloning sites of the pSEAP2-Basic vector, yielding pCatB0.26-SEAP+. Constructs were purified, partially sequenced, and restriction digested to confirm the identity and orientation of the inserts. The genes examined and constructs made in this study are summarized in Table 1 . 

Transfection for Promoter Activity Analysis and Cotransfection with the Sp1 Expression Vector

The activity of the putative α1-PI, LAP, cathepsin B, andα 2-M promoters in normal human corneal epithelial cells, corneal stromal cells, or skin fibroblasts was investigated in transient transfection assays. The promoter plasmids and pSEAP2-Basic (promoterless reporter gene vector), pSEAP2-Control (positive control, driven by the simian virus 40 early promoter), along with the pSV-β-galactosidase control vector (Promega) used to normalize the transfection efficiency, were used in cell transfections. 

The cells were plated at 40,000 cells/well on 24-well plates 24 hours before DNA transfection. They received fresh medium and were transfected 2 hours later using FuGENE 6 transfection reagent (Roche Molecular Biochemicals, Indianapolis, IN). In brief, 0.75 μg of the test plasmid (pα1PI2.7-SEAP+, pα1PI1.4-SEAP+, pLAP-SEAP+, pCatB0.44-SEAP+, pCatB0.26-SEAP+, pα2M-SEAP+, pSEAP2-Basic, or pSEAP2-Control) and 0.2 μg pSV-β-galactosidase vector were mixed with 2.85 μl of FuGENE 6 in 100 μl serum-free medium, as recommended by the manufacturer. Another series of cells also received 0.1875 μg of the Sp1 expression vector pPacSp1 (a generous gift of Robert Tjian, University of California Berkeley). None of the test plasmids was added to cells serving as negative control cultures. The medium was collected from each well 48 hours later for SEAP assays. 

For SEAP assay, 100 μl of the medium was mixed with 300 μl dilution buffer. After a 30-minute incubation at 65°C, 100-μl aliquots were mixed with 100 μl each of assay buffer and reaction buffer supplied from the kit, according to the manufacturer’s protocol (Tropix, Inc., Bedford, MA). The enzyme activity, represented by the light emission, was read for 5 seconds on a luminometer (MGM Instruments, Inc., Hamden, CT). For β-galactosidase assays, cells harvested were lysed with 70μ l Galacto-Lysis Solution (Tropix, Inc.) and centrifuged at 14000 rpm to pellet the debris. The extract was used for measurements ofβ -galactosidase activity and protein content (BCA method; Pierce, Rockford, IL). Assays were performed in triplicate, and each experiment was repeated at least three times. The β-galactosidase activity was used to normalize the SEAP enzyme activity. The significance of the data was analyzed by two-tailed Student’s t-tests. 

Results

Functional Analysis of the α1-PI 5′-Flanking Element in Human Corneal Epithelial Cells and Their Responsiveness to Sp1 Expression

In cultured human corneal epithelial cells, DNA fragments containing 2.7 kb (pα1PI2.7-SEAP+) and 1.4 kb (pα1PI1.4-SEAP+) of the 5′-flanking sequence of the α1-PI gene were approximately 19 and 15 times more active at driving the SEAP reporter gene expression than the promoterless pSEAP2-Basic vector used to detect the basal activity (Fig. 1) . This result indicated that the α1-PI promoter sequences conferred functional activity in corneal epithelial cells. The 1.4-kb fragment was sufficient for promoter activity. 

When either the pα1PI2.7-SEAP+ or pα1PI1.4-SEAP+ construct was cotransfected with Sp1 expression vector pPacSp1 into corneal epithelial cells, the level of SEAP expression was significantly reduced (P < 0.05; Fig. 1 ). 

Functional Analysis and Sp1 Responsiveness of the LAP, Cathepsin B, and α2-M 5′-Flanking Elements

In normal human corneal stromal cells, the vector that contained 0.67-kb 5′-flanking sequence of the human LAP gene (pLAP-SEAP+) had 53 times greater activity than the pSEAP2-Basic vector (Fig. 2) . Likewise, the SEAP activity in stromal cells driven by constructs containing the 0.44-kb promoter fragment of the cathepsin B gene (pCatB0.44-SEAP+) and the 5.77-kb α2-M promoter sequence (pα2M-SEAP+) was significantly higher (approximately 64- and 21-fold, respectively) than the basal expression (Figs. 3 and 4) . In contrast, pCatB0.26-SEAP+ showed little promoter activity (Fig. 3) . None of the promoter activity was modified by pPacSp1 cotransfection. 

Compared with corneal cells, the promoter activity of the LAP, cathepsin B, and α2-M genes was much lower in skin fibroblasts (Figs. 2 3 and 4) , suggesting that there may be tissue- or cell type–specific expression of these genes. 

When transfected into cultured corneal epithelial cells, the pLAP-SEAP+, pCatB0.44-SEAP+, and pα2M-SEAP+ vectors were approximately 168, 127, and 7.6 times more active, respectively, at driving SEAP reporter gene expression than the pSEAP2-Basic vector (Fig. 5) . The pCatB0.26-SEAP+ was again, as in corneal stromal cells, not functional in the epithelial cells. When cotransfected with pPacSp1, the level of SEAP expressions remained unaltered with all constructs except pCatB0.44-SEAP+, for which a significant reduction (P < 0.05) was observed (Fig. 5) . 

Discussion

The thinning in corneas affected by keratoconus is believed to be due to extensive degradation of connective tissue elements contained therein. ¹² Studies from our laboratory have provided evidence that an increase in the net activities of serine and cysteine proteinases, such as cathepsins B and G, that have been shown to be released to the extracellular milieu ^{23 24 25 26} and are capable of digesting gelatin, casein, and extracellular matrix elements, ^{27 28} may be responsible for the abnormal degradative processes in keratoconus. ¹² The levels and activities of cathepsins B and G were elevated ¹² and the levels of inhibitors α1-PI and α2-M were reduced ^{13 14} in corneas with keratoconus, presumably resulting in the increased net (a balance of enzyme and inhibitor) degradative activity detected. The biochemical abnormalities were observed both in the corneal stroma, where thinning occurs in keratoconus, and the corneal epithelium. ^{11 12 13 14 15} An involvement of the epithelial layer in the disease has thus been surmised. 

By immunostaining, Western blot analysis, and electrophoretic mobility shift assay, an upregulation of transcription factor Sp1 was observed in the stroma and the epithelium of keratoconus-affected corneas. ¹⁶ It was further noted that the basal level of Sp1 expression in stromal cells cultured from eyes with keratoconus was greater than that in normal human cells, ²⁹ suggesting that the diseased corneal cells carry and retain the Sp1 abnormality, even after they are maintained in tissue culture through two passages. The Sp1 defect therefore appears to be an inborn error in keratoconus, not merely occurring secondarily as a result of other factors. 

In an earlier study, we showed that an augmented level of Sp1 directly repressed the promoter activity of the α1-PI gene in corneal stromal cells. ¹⁷ The present study further demonstrates a similarα 1-PI downregulation conferred by Sp1 in corneal epithelial cells. These results may argue for both an epithelial involvement and an upstream role of Sp1 in the development of keratoconus. The Sp1-mediated α1-PI downregulation appears to be a key event. It is possible that in the corneal epithelium, the Sp1/α1-PI abnormality contributes to the degradation and breaks in the epithelial basement membrane zone, resulting in the earliest pathologic features ^{30 31} seen in keratoconus. In the stroma as well, the anomaly may lead to increased degradative activity and eventual manifestation of thinning and scarring. 

The current data indicate that Sp1 cotransfection confers little impact on the transcriptional regulation of the LAP, cathepsin B, and α2-M genes in corneal stromal cells. The constructs that contain the promoter fragments of these genes (Table 1) were made from sequences available in the literature. ^{18 19 20 21} LAP is one of the lysosomal hydrolases that are involved mostly in lipid metabolism, not in turnover of protein or connective tissues. The 0.67-kb LAP promoter has no TATA and CAAT box sequences, has high GC content, two Sp1 binding sites, and a region complying with the properties of a CpG island. ¹⁸ Similar to the α1-PI gene, ¹⁷ the characteristics of the LAP promoter identify it as a housekeeping gene. Unlike the α1-PI gene, however, the LAP gene promoter does not respond to the Sp1 overexpression. This demonstration is in line with our previous finding that LAP activity is enhanced in both the cornea and the conjunctiva ³² of eyes with keratoconus, whereas Sp1 upregulation is seen only in the former. ^{16 33} Taking all evidence together, we suggest that the LAP abnormality in keratoconus may represent merely a microenvironmental change on the ocular surface. Perhaps environmental factors that have long been speculated to be associated with keratoconus, such as eye rubbing and contact lens wear, modify the hydrolase levels in cells both in the cornea and the conjunctiva. 

The human cathepsin B promoter region also possesses high GC content and has no canonical TATA and CAAT boxes. ²¹ Such TATA-less promoters are often activated by Sp1 and are dependent on the presence of clusters of Sp1 binding sites in the proximity of the transcription start site. Cathepsin B thus appears to be typical of an Sp1-regulating promoter. ^{34 35} A total of nine Sp1 binding sites were identified in the 0.44-kb fragment used in our study. Electrophoretic mobility shift assays and site-directed mutagenesis studies have confirmed the involvement of multiple Sp1 sites in regulation of the cathepsin B proximal promoter. ²¹ In this vein, it is somewhat surprising that the results indicate that cathepsin B expression in corneal stromal cells is insensitive to Sp1 coexpression and hence may not be regulated directly by Sp1. Either additional factors are required, or the regulation is through an Sp1-independent mechanism. A shorter 0.23-kb promoter fragment with seven Sp1 sites has been shown to be transactivatable by Sp1 cotransfection in Schneider’s Drosophila line 2. ²¹ We therefore made and tested an additional construct containing a 0.26-kb promoter sequence and the same Sp1 binding sites. The functionality of this construct, however, was minimal in corneal cells. 

The human α2-M gene promoter contains a TATA box. There are two potential Sp1 binding sites, although their functionality has yet to be established. ²⁰ The absence of Sp1 responsiveness of theα 2-M promoter was not entirely unexpected. 

In corneal epithelial cells, the 0.67-, 0.44- and 5.77-kb promoter fragments for LAP, cathepsin B, and α2-M genes were all functional. Similar to that found in stromal cells, the LAP and α2-M promoters were not Sp1 responsive. The cathepsin B promoter activity, on the other hand, was suppressed by the Sp1 cotransfection. The experiments were repeated three times, and repression was observed in every experiment. The significance of this result is unclear at present. 

In conclusion, the present study demonstrates, on the basis of transient transfection assays, that among the known enzyme and inhibitor genes affected in keratoconus, only α1-PI is regulated by Sp1. The Sp1 regulation is operative in both corneal epithelial and stromal cells. These results support the epithelial involvement theory and suggest that the Sp1-mediated downregulation of the α1-PI gene may be a key event in the development of keratoconus. 

²  YM, XW, and YL contributed equally to the study.

 Supported by Grants EY03890 and EY05628 and Core Grant EY01792 from the National Institutes of Health. BYJTY is a recipient of a Senior Scientific Investigator Award from Research to Prevent Blindness, Inc.

 Submitted for publication December 8, 2000; revised March 30, 2001; accepted April 19, 2001.

 Commercial relationships policy: N.

 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.

 

Table 1.

 

Summary of the Genes Studied and Constructs Made

Genes	Alteration Observed in Keratoconus	Constructs for Transfection	Size and Region of 5′-Flanking Fragments*	Number of Putative Sp1 Sites
α1-PI	Downregulated	pα1PI2.7-SEAP+	2.7 (−2703 to+9)	12
		pα1PI1.4-SEAP+	1.4 (−1397 to +9)	10
LAP	Upregulated	pLAP-SEAP+	0.67 (−586 to+79)	2
Cathepsin B	Upregulated	pCatB0.44-SEAP+	0.44 (−361 to+74)	9
		pCatB0.26-SEAP+	0.26 (−190 to+74)	7
α2-M	Downregulated	pα2M-SEAP+	5.77 (−5761 to +12)	2

*  Size is expressed in kilobases with region shown in parentheses.

Figure 1.

 

Relative SEAP enzyme activities in normal human corneal epithelial cells transfected with pSEAP2-Basic, pα1PI2.7-SEAP+, and pα1PI1.4-SEAP+ vectors along with the pSV-β-galactosidase control vector in the absence or presence of Sp1 expression vector pPacSp1. Note that the activity of pα1PI2.7-SEAP+ and pα1PI1.4-SEAP+ vectors was approximately 19 and 15 times higher, respectively, than that of the pSEAP2-Basic vector. *Significantly different from data obtained from cells transfected without Sp1 coexpression (P < 0.05). The positive control, using the pSEAP2-Control vector, yielded more than an order of magnitude higher SEAP activity. The SEAP activity was normalized by β-galactosidase enzyme activity. Data are presented as mean ± SD (n = 3). Two-tailed Student’s t-tests were used to analyze the significance of the data. Experiments were repeated four times with similar results.

Figure 2.

 

Relative SEAP enzyme activities in normal human corneal stromal cells and skin fibroblasts. Cells were transfected with pSEAP2-Basic and pLAP-SEAP+ vectors along with the pSV-β-galactosidase control vector (used to normalize transfection efficiency) in the absence or presence of the Sp1 expression vector pPacSp1. In corneal stromal cells and skin fibroblasts, respectively, the activity of the pLAP-SEAP+ vector was approximately 53 and 3.3 times higher than that of the pSEAP2-Basic vector. Coexpression of Sp1 did not significantly alter the SEAP expression driven by the pLAP-SEAP+ vector. Data are presented as mean ± SD (n = 4). Experiments were repeated five times with similar results.

Figure 3.

 

Relative SEAP enzyme activities in normal human corneal stromal cells and skin fibroblasts. Cells were transfected with pSEAP2-Basic and pCatB0.44-SEAP+ and/or CatB0.26-SEAP+ vectors along with pSV-β-galactosidase control vector without or with pPacSp1. The activity of the pCatB0.44-SEAP+ vector for corneal and skin cells, respectively, were approximately 64 and 13 times higher than that of the pSEAP2-Basic vector. The pCatB0.26-SEAP+ vector was not functional in corneal stromal cells. Cotransfection with pPacSp1 did not modify the SEAP expression level by either the pCatB0.44-SEAP+ or pCatB0.26-SEAP+ vector. Data are presented as mean ± SD (n= 3). Experiments were repeated four times.

Figure 4.

 

Relative SEAP enzyme activities in normal human corneal stromal cells and skin fibroblasts. Cells were transfected with pSEAP2-Basic and pα2M-SEAP+ vectors along with the pSV-β-galactosidase control vector without or with pPacSp1. The activity of the pα2M-SEAP+ vector, for corneal and skin cells, respectively, was approximately 21 and 1.4 times higher than that of the pSEAP2-Basic vector. The activity was not significantly varied by cotransfection of pPacSp1. Data are presented as mean ± SD (n = 4). Experiments were repeated four times.

Figure 5.

 

Relative SEAP enzyme activities in normal human corneal epithelial cells transfected with pSEAP2-Basic, pLAP-SEAP+, pCatB0.44-SEAP+, pCatB0.26-SEAP+, and pα2M-SEAP+ vectors along with pSV-β-galactosidase control vector in the absence or presence of pPacSp1. The activity of pLAP-SEAP+, pCatB0.44-SEAP+, pCatB0.26-SEAP+, and pα2M-SEAP+ vectors was approximately 168, 127, 3.4, and 7.6 times higher, respectively, than that of the pSEAP2-Basic vector.* Significantly different from data obtained from cells transfected without Sp1 coexpression (P < 0.05). The SEAP activity was normalized by β-galactosidase enzyme activity. Data are presented as mean ± SD (n = 3). Two-tailed Student’s t-tests were used to analyze the significance of the data. Experiments were repeated three times with similar results.

The authors thank Chan Boriboun for technical assistance.