Induction of Apoptosis in Human Corneal and HeLa Cells by Mutated BIGH3

Abstract

purpose. To determine the effects of overexpression of mutated BIGH3 in HeLa and human corneal epithelial (HCE) cells.

methods. Six mutations known to be responsible for autosomal dominant corneal dystrophies linked to chromosome 5 were generated in a BIGH3 expression vector and transfected in HeLa and HCE cells. The expression and secretion of the various BIGH3-EGFP fusion proteins were measured by Western blot analysis. Apoptotic cells were identified by Hoechst/propidium iodide and annexin V staining. Lactate dehydrogenase (LDH) activity was measured in the medium of transfected cells. Truncated BIGH3 protein and site-specific mutations were generated to determine the exact region that mediated apoptosis.

result. The overexpressed BIGH3 fusion protein was secreted regardless of its mutation status and was clearly observed in the culture medium. Overexpression of mutated BIGH3 induced apoptosis in both cell lines through activation of caspase-3. Although all the disease-causing mutations tested in this experiment induced apoptosis, the strongest effect was observed with the R124C and R555W mutations. Overexpression of a carboxyl-truncated BIGH3 protein did not induce apoptosis, suggesting that a region located in the C-terminal domain was necessary to mediate cell death. In addition, mutation of the Pro-Asp-Ile (PDI) site at 616-618 was sufficient to prevent induction of apoptosis.

conclusions. Overexpression of mutated BIGH3 induces apoptosis in HeLa and HCE cells through activation of a pathway that uses the PDI domain of the fourth internal Fas domain and activation of caspase-3. Because DI is a known site of interaction with α3β1 integrins, it suggests that integrins play a role in mediating apoptosis in the system used in the current study. This work suggests that apoptosis is a key element in the pathophysiology of BIGH3-related corneal dystrophies.

Corneal dystrophies (CDs) are disorders characterized by the progressive accumulation of deposits in the different layers of the cornea followed by visual impairment. Linkage analysis in large families has localized granular Groenouw type I (CDGG1), lattice type I (CDL1), Avellino (CDA), and Reis-Bückler (CDRB) type CDs on the long arm of chromosome 5 between markers D5S393 and D5S399, in the proximity of IL9. ^{1 2 3 4} Physical mapping and cDNA selection have identified the βIg-h3 (BIGH3)-TGFβ-induced (TGFBI; OMIM 601692; Online Mendelian Inheritance in Man; http://www.ncbi.nlm.nih.gov/Omim/ provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD) ⁵ gene as responsible for each of these CDs linked to chromosome 5. ⁶ BIGH3 is a secreted protein of 683 amino acids and contains four internal repeats. These DNA stretches are called Fas domains because of their homology to the Drosophila fasciclin 1 gene. ⁵ Mutation analysis revealed that R124C was associated with CDLI, R124H with CDA, R555W with CDGGI, and R555Q with CD Thiel-Behnke (CDTB) in most of the cases. ⁷ Recently, new mutations at other positions of BIGH3 have been reported ^{8 9 10 11 12 13 14 15} and other CDs such as CDLIIIa have been linked to mutations of the same gene (Pro501Thr). ¹⁶ Immunohistochemical and biochemical analyses of the nature of these deposits have revealed the presence of the BIGH3 protein (also named kerato-epithelin) as a major component in all types of BIGH3-related CD. ^{17 18 19} The deposits have been precisely localized in the epithelium and Bowman’s layer for CDRB and in the subepithelial layer and stroma for CDGGI, CDLI, CDLIIIA, and CDA. 

Although these deposits are well described, not much is known about their origin. Two-dimensional gel and Western analyses have revealed a general increase in the amount of the BIGH3 protein present in the cornea in three types of CDs with mutation at R124. Different isoforms were seen in R124C and R124H, but not in R124L. ²⁰ Increased amounts of BIGH3 protein have also been observed in CDGG1. ¹⁷ Thus, the generation of abnormal fragments and/or increased amounts of BIGH3 protein seems to represent a common pathogenic mechanism in BIGH3-related CD. 

We therefore investigated the consequences of overexpression of mutated BIGH3 proteins in an in vitro system and observed that in HeLa and human corneal epithelial (HCE) cells, overexpression of BIGH3 induced a strong apoptotic response. 

Methods

Constructs and Antibody

pGEM-BigWT, a BIGH3 cDNA clone, was obtained by subcloning the amplicon generated by RT-PCR from RNA isolated from a normal human fibroblast cell line with primers BIGH3-1 (5′-CGG TCG CTA GCT CGC TCG GT-3′) and BIGH3-2 (5′-TTC CTC CTG TAG TGC TTC AA-3′) in the pGEM-T plasmid, as described by the manufacturer (Promega, Madison, WI) and standard PCR conditions (GeneAmp; Applied Biosystems, Foster City, CA). The lattice mutation (R124C) was introduced in pGEM-BigWT to form pGEM-BigR124C by replacing the BamHI-ClaI DNA fragment with a BamHI-ClaI fragment obtained by PCR amplification using a modified primer BIGH3-3 (5′- gttggatccACC ACC ACT CAG CTG TAC ACG GAC TGC AC -3′) and BIGH3-4 (5′-ccttatcgATG AGG TGC ACC ACC CCG TT-3′), respectively. The PCR product was digested with BamHI and ClaI and subcloned in the prepared pGEM-BigWT plasmid. 

The pGEM-BigR555W plasmid containing the Groenouw mutation (R555W) was generated by a similar approach. The BglII-ClaI fragment of pGEM-BigWT was replaced with a fragment obtained from the BglII-ClaI double digestion of an amplicon obtained by RT-PCR from the RNA of a patient with CDGGI, by using random hexamers and primers BIGH3-5 (5′-TTG AGC TGC TCA ATG CCC TC) and BIGH3-2. pGEM-BigWT, pGEM-BigR124C and pGEM-BigR555W were digested with NheI and BglII. The respective fragments coding for BIGH3 were isolated and subcloned into the pEGFP-N2 vector digested with the same endonucleases. The BIGH3 stop codon was removed by an additional cycle of PCR and subcloning using primers BIGH3-6 (5′-CTG TCG ACC ACT ACA TTG ATG AGC TA-3′) and BIGH3-7 (5′-CTG TCG ACA ATG CTT CAT CCT CTC TAA TAA C-3′). The different pEGFP-N2 plasmids were digested with BglII and SalI and the BglII-SalI fragments were replaced by the BglII-SalI double-digested PCR products just described to form pBigWT, pBigR124Cm, and pBigR555W, respectively. 

All the mutations (N622K, G623D, H626R, V631D, sat, and grl) were introduced by using the oligonucleotide-directed mutagenesis technique. pBigWT was used as template for two sequential PCR amplifications. In the first PCR, a forward primer (BIGH3-6) and a reverse primer (BIGH3-xR) containing the desired mutation were used. In the second PCR, the reverse primer EGFPas (5′-TGA ACA GCT CCT CGC CCT TG-3′) was used with a forward mutation-specific primer (BIGH3-xF). The two amplicons were mixed together, treated with Klenow, and used as the template for a new PCR amplification with BIGH3-6 and EGFPas. pBigN622K, pBigG623D, pBigH626R, and pBigV631D were generated by replacing the BbsI-BglII DNA fragment with the BbsI-BglII fragment from the respective mutation-containing these amplicons. Similarly, the amplicons containing the mutations sat (Ser-Ala-Thr at 616-618) and grl (Gly-Arg-Leu at 64-644) were introduced into pBigWT and pBigR124C to obtain pBigWT-sat, pBigR124C-sat, pBigWT-grl, and pBigR124C-grl. Mutation-specific primers for the different mutations were N622K: BIGH3-8reverse(R) (5′-CAT GGA CCA CGC CTT TTG TGG CCA TG-3′) and BIGH3-9forward(F) (5′-CAT GGC CAC AAA AGG CGT GGT CCA TG-3′); G623D: BIGH3-10R (5′-GAC ATG GAC CAC GTC ATT TGT GGC CAT G-3′) and BIGH3-11F (5′-CAT GGC CAC AAA TGA CGT GGT CCA TGT C-3′); H626R: BIGH3-12R (5′-GGT GAT GAC ACG GAC CAC GCC ATT TGT G-3′) and BIGH3-13F (5′-CAC AAA TGG CGT GGT CCG TGT CAT CAC C-3′); V631D: BIGH3-14R (5′-GAG GCT GCA GAT CAT TGG TGA TGA CAT G-3′) and BIGH3-15F (5′-CAT GTC ATC ACC AAT GAT CTG CAG CCT C-3′); sDI: BIGH3-16R (5′-GGC CAT GAT GTC AGA CTC GGC AAC AGG CTC C-3′) and BIGH3-17F (5′-GGA GCC TGT TGC CGA GTC TGA CAT CAT GGC C-3′); sat: BIGH3-18R (5′-CGC CAT TTG TGG CCA TGG TGG CAG ACT CGG CAA CAG GCT C-3′) and BIGH3-19F (5′-GGA GCC TGT TGC CGA GTC TGC CAC CAT GGC CAC AAA TGG C-3′); and grl: BIGH3-20R (5′-AGA GTC TGC AAG TTC AAG CCG TCC TTC CTG AGG TCT GTT-3′) and BIGH3-21F (5′-AAC AGA CCT CAG GAA GGA CGG CTT GAA CTT GCA GAC TCT-3′). 

Wild-type and lattice BIGH3 constructs lacking part of Fas-4 domain and C-terminal region of BIGH3 from amino acid 541-683 were constructed as follows. pBigWT and pBigR124C were double-digested with BbsI and SmaI. The 541-683 fragment was discarded, and the plasmids were blunted and self-ligated. This generated pBigWTΔ^COOH and pBigR124CΔ^COOH. These constructs do not have the PDI and RGD domains at positions 616-618 and 642-644, respectively. All constructs were sequenced using standard techniques to confirm sequence integrity (Fig. 1) . 

The rabbit polyclonal anti-serum anti-KE426-682 against the protein BIGH3 has been described. ¹⁷ Human monoclonal antibody to caspase-3 was obtained from Alexis Biochemicals (San Diego, CA) and was used as recommended by the manufacturer. 

Transient Transfection

Cells (10⁵) were transfected in 35-mm dishes with transfection reagent, according to the manufacturer’s instructions (DOTAP; Roche Molecular Biochemicals, Mannheim, Germany). One microgram and 1 to 1.5 μg of plasmid were transfected into HeLa and HCE cells, respectively. 

Coloration of Cells and Counting

Apoptotic cells were discriminated from normal cells by the Hoechst/propidium iodide (PI) technique. ²¹ Briefly, cells were incubated with Hoechst 33342 (12.5 μg/mL) and PI (5 μg/mL) for 5 minutes before visualization under an inverted fluorescence microscope (model Axiovert 25; Carl Zeiss Meditech, Oberkochen, Germany). HeLa and HCE cells were colored and counted 24 and 48 hours after transfection, respectively. The experiments were performed with different DNA preparations and repeated a minimum of three times. At least 1000 cells were counted in each experiment. Apoptotic cells were also identified based on translocation of the phospholipid phosphatidylserine to the cell surface, as observed by annexin V labeling. The annexin V-Cy3 apoptosis detection kit was used according to the manufacturer’s protocol (BioVision Research Products, Mountain View, CA), and cells were observed by fluorescence microscopy. 

Detection of Lactate Dehydrogenase Activity

Cell cultures were harvested 24 hours (HeLa) or 48 hours (HCE) after transfection, and the medium was removed to determine the extracellular lactate dehydrogenase (LDH) content. Intracellular LDH content was measured by lysing the cells with Triton X-100 (1% in PBS). Medium and intracellular LDH contents were analyzed by the cytotoxicity detection kit (LDH; Roche Molecular Biochemicals) according to instructions provided by the manufacturer. Absorbance was read with an immunoassay system (Dynatech, Cambridge, MA) at 490 nm. LDH efflux was expressed as a percentage of the total LDH release (extracellular LDH/extracellular+intracellular LDH). HeLa and HCE cells were transfected with 1 μg of pEGFP-flag, pBigWT, pBigR124C, or pBigR555W. Cell culture medium was collected after 24 hours for HeLa and after 48 hours for HCE cells. The remaining cells were then lysed and recovered, and the LDH activity was measured in these samples. 

Western Blot Analysis

The medium of transfected cells was collected 24, 48, and 72 hours after transfection. Samples for SDS-PAGE analysis were prepared by addition of Laemmli buffer and boiled for 10 minutes. Equal volumes of medium were subjected to electrophoresis in 10% SDS-PAGE. Proteins were transferred onto polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore, Bedford, MA) by semidry electroblot. Membranes were blocked for 2 hours in PBST buffer (PBS and 0.1% Tween-20) containing 5% nonfat dry-milk. Diluted anti-KE426-682 (1:4000) was applied for 2 hours at room temperature. Peroxidase labeled anti-rabbit antibody (Amersham Life Science, Amersham, UK) was incubated for 1 hour at room temperature. Immunostaining was visualized by enhanced chemiluminescence (ECL; Amersham Life Science). For the caspase-3 assay, proteins were extracted from HeLa cells 48 hours after transfection. Membranes were incubated separately with caspase-3 human monoclonal antibody and β-tubulin (Santa Cruz Biotechnology, Santa Cruz, CA) as recommended by the manufacturer. Densitometry analysis was performed on computer (LabWorks; UVP, Upland, CA). Signals were normalized with β-tubulin and expressed as percentages of pBigWT signal. 

Statistics

Data are expressed as the mean ± SEM. Statistical significance was assessed by unpaired t-test for each mutation compared with the wild-type construct. Differences were considered significant at P < 0.05. 

Results

Transient Expression of Mutated BIGH3

The wild-type BIGH3 protein is a secreted protein found in the extracellular matrix compartment. However, it was not known whether the mutated BIGH3 protein is secreted with the same kinetic. To investigate, we analyzed GFP fusion protein of the wild-type, R124C, and R555W forms of BIGH3 in transfected HeLa and HCE cells. Western blot analysis of medium obtained at 24, 48, and 72 hours established clearly that recombinant BIGH3 (rBIGH3) protein accumulated in the medium independent of its mutation status (Fig. 2A) . Endogenous and BIGH3 protein from fetal bovine serum (FBS) were also observed in the medium. No rBIGH3 protein was observed in cells transfected with the control plasmid pEGFP-flag. 

The efficiency of transfection and expression of the different constructs were evaluated in HCE and HeLa cells. HCE cells were transfected with identical amounts of pBigWT, pBigR124C, and pBigR555W. An equal volume of medium recovered 48 hours after transfection was analyzed by Western blot. The blot showed smaller amounts of secreted protein in HCE cells than in HeLa cells (Fig. 2B) . 

Induction of Apoptosis by Overexpression of Mutated BIGH3

On examination of transfected cells, we observed that mutated rBIGH3 induced a higher rate of cell death than rBIGH3-WT. Hoechst/PI staining showed that a great proportion of these dead cells presented condensation and fragmentation of the nucleus—two characteristics of cellular apoptosis (Figs. 3A 3B 3C) . In addition, translocation of phosphatidylserine to the cell surface of apoptotic cells was detected by annexin V labeling (Fig. 3D) . 

To quantify the apoptosis induced by the mutated BIGH3 constructs, we transfected HeLa cells with 1 μg of pEGFP-flag, pBigWT, pBigR124C, or pBigR555W. Twenty-four hours later, cells were stained with Hoechst 33342/PI and scored. Transfection with mutated constructs induced an increased apoptotic response from 3.0% ± 0.9% (pBigWT) to 13.1% ± 2.1% (P < 0.001) and 14.6% ± 4.4% (P < 0.001) for pBigR124C and pBigR555W, respectively (Fig. 4) . 

These experiments were repeated in HCE cells to investigate whether this phenomenon was cell-type specific or represented a common process. HCE cells were transfected with 1.5 μg of pEGFP-flag, pBigWT, pBigR124C, or pBigR555W obtained from different DNA preparations. After 48 hours, a similar induction of apoptosis was observed: the apoptosis rates increased from 3.4% ± 0.6% for pBigWT to 11.8% ± 4.6% (P < 0.01) and 14.6% ± 3.7% (P < 0.01) for pBigR124C and pBigR555W, respectively (Fig. 4) . 

The LDH activity of the transfected cells was measured to confirm the presence of cell death. The ratios of released LDH activity over total LDH showed clearly the presence of death in cells. In HeLa-transfected cells, the ratio increased from 0.01 (pEGFP-flag) and 0.005 (pBigWT) to 0.065 (pBigR124C) and 0.04 (pBigR555W; P < 0.01; Fig. 5A ). LDH increase was also observed in HCE transfected by pBigR124C (0.105) and pBigR555W (0.106) compared with pEGFP-flag (0.03) and pBigWT (0.05), P < 0.05. 

Western blot analysis of HeLa protein extracts obtained after transfection with the two mutated constructs showed a reduction in procaspase-3 when compared with the level obtained with overexpression of pBigWT. This indicates activation of caspase-3 (Fig. 5B) . 

Induction of Apoptosis by Overexpression of Other BIGH3 Mutations

Although most mutations arise at amino acids R124 and R555, mutations at other locations have been reported in CDs. We recently identified rare mutations in a large cohort of families with BIGH3-related CD (N622K, G623D, H626R, and V631D) ⁹ and tested several of these specific mutations to see whether apoptosis induced by abnormal rBIGH3 protein is a general mechanism in BIGH3-related CD. Constructs with these mutations were generated (Fig. 1A) and overexpressed in HeLa and HCE cells, as described in the Methods section. Expression of each construct was confirmed by Western blot analysis (data not shown). 

Overexpression of these four BIGH3 mutations induced a strong apoptotic response in both cell lines. In HeLa cells, the rates of apoptosis rose to 6.7% ± 1.0% for pBigN622K (P < 0.05), 8.0% ± 1.9% for pBigG623D (P < 0.05), 5.9% ± 0.7% for pBigH626R (P < 0.05), and 13.5% ± 2.8% for pBigV631D (P < 0.001); and in HCE cells to 7.2% ± 0.5% for pBigN622K (P < 0.05), 7.8% ± 1.3% for pBigG623D (P < 0.05), 7.8% ± 1.1% for pBigH626R (P < 0.05), and 8.5% ± 3.1% for pBigV631D (P < 0.05; Fig. 4 ). The differences between all these responses and basal apoptosis levels were statistically significant. 

Implication of the C-Terminal Part of BIGH3 in Mediating Apoptosis

Analysis of BIGH3 protein identified an RGD motif at amino acids 642-644. This motif represents a potential integrin-binding site. ⁵ Recently, Kim et al. ²² described another integrin-binding site Asp-Ile (DI) at amino acid 617-618 in BIGH3 Fas-4 domain. We therefore investigated whether the C-terminal domain of BIGH3 is involved in the transmission of the apoptotic signal. Two constructs without 142 amino acids from the C-terminal domain (Fig. 1B) , pBigWTΔ^COOH and pBigR124CΔ^COOH, were transfected into HeLa and HCE cells. The rates of apoptosis in HeLa cells transfected with pEGFP-BigWT and pBigWTCΔ^COOH were similar (3.0% ± 0.9% and 3.6% ± 1.0%; P > 0.05) as were rates in HCE cells (3.9% ± 1.2% and 3.3% ± 0.6%; P > 0.05), indicating that the C-terminal truncated proteins were not acting as mutated compounds (Fig. 6) . Transfections with pBigR124CΔ^COOH were accompanied by a reduction in apoptosis from 13.1% ± 2.1% (pBigR124C) to 5.2% ± 0.4% (P < 0.01) in HeLa cells and from 11.8% ± 4.6% to 2.6% ± 0.9% (P < 0.01) in HCE cells (Fig. 6) . 

To determine whether apoptosis was mediated by the RGD motif, mutations were introduced at these three amino acids in wild-type and lattice constructs (Fig. 1C) . Overexpression of these constructs induced a strong apoptotic response in HeLa cells (19.9% ± 8.8% and 20.0% ± 5.8%, respectively; Fig. 7 ) suggesting that the RGD motif was essential for proper function but did not mediate the effects observed previously. 

We then mutated the putative α3β1 integrin-binding site identified in BIGH3 Fas-4 domain by Kim et al. ²² The pBigWT and pBigR124C constructs containing a modified Pro-Asp-Ile (PDI) motif at amino acids 616-618 into Ser-Ala-Thr (sat) were generated as described in the Methods section and transfected in HeLa and HCE cells. In HeLa cells, the rate of apoptosis was 4.8% ± 0.3% for pBigWT-sat. This is not significantly different from the basal rate of 3.0% ± 0.9% obtained with pBigWT (P > 0.05; Fig. 7 ). This indicated that, in contrast to mutation of the RGD domain, the modification of the PDI motif did not render the protein apoptogenic. When the same mutations were engineered in a lattice background and overexpressed in HeLa cells, the rate of apoptosis was reduced from 13.1% ± 2.1% for pBigR124C to 3.7% ± 0.9% for pBigR124C-sat (P < 0.01). Similar results were obtained with transfections in HCE cells (Fig. 7) . 

Discussion

Recent immunohistology analyses in BIGH3-related CD showed that increased amounts of the BIGH3 protein is a major, if not the sole, component of all types of deposits, whether amyloid or nonamyloid. ^{17 18 20} We therefore developed a cellular model in which different mutated BIGH3 transcripts were overexpressed and investigated their role in the pathogenesis of BIGH3-related CD. 

When overexpressed in HeLa and HCE cells, mutated and wild-type rBIGH3 proteins were normally secreted into the medium as shown by Western blot analysis. The EGFP portion of the fusion protein was not processed. In addition, no fusion protein was observed in the different cellular compartments by fluorescence microscopy, which indicated that full-length rBIGH3 did not accumulate in the cell. It cannot be excluded that processed recombinant proteins that do not have the EGFP reporter are still present in the cells. However, it is unlikely that great amounts were present, because we have not observed degradation products in Western blot analysis of proteins extracted from transfected HeLa and HCE cells. 

Overexpression of the mutated rBIGH3 induced a strong apoptotic response in the two analyzed cell lines. This response was present and reproducible for all the clinically relevant mutations—that is, mutations that have been linked to BIGH3-related CD, as well as for mutations engineered in regions not yet linked to CD (i.e., Arg-Gly-Asp at amino acids 642-644, a putative integrin-binding site. The strongest effect was obtained with the R124C and R555W constructs. 

Apoptosis is a complex mechanism that can be mediated by numerous pathways involving intracellular and/or extracellular signals. Western blot analysis indicated that the recombinant BIGH3-EGFP fusion protein was secreted together with the endogenous BIGH3. As no fusion protein was observed in the different compartments and the success rate of transfection rarely exceeded 5%, mutated rBIGH3 protein may induce apoptosis by interacting with surrounding cells, either directly through binding to some cellular receptors or through other members of the extracellular matrix system. Recently, it was shown that purified wild-type BIGH3 protein binds collagen I, fibronectin, fibrillin and lamin and that binding to collagen II and VI is moderate and minimal compared with collagen IV. ^{23 24} An abnormal interaction between mutated BIGH3 protein and one of these extracellular proteins could be responsible for the induction of apoptosis. The mechanism by which this would occur is not self-explanatory, but ultimately it activates caspase-3, a strong effector of apoptosis. 

Alternatively, direct interaction with membrane receptors such as integrins can be postulated. Computer analysis of the C-terminal region of BIGH3 ⁵ and work by Kim et al. ²² have identified two integrin recognition sites, an Arg-Gly-Asp (RGD) sequence at amino acids 642-644 and Asp-Ile (DI) at 617-618, although no experimental data have yet shown integrin binding to the former site, despite several attempts. ²⁵  

To identify the precise region of rBIGH3 that mediated apoptosis, constructs that lacked the C-terminal part of BIGH3 where the DI and RGD motifs are located were generated. Although overexpression of the BIGH3 constructs without the C-terminal had no effect on apoptosis, as expected, overexpression of the C-terminal partially deleted lattice constructs was associated with basic levels of apoptosis, in both HeLa and HCE cells. This suggests that interaction with the C-terminal part is needed for mutated BIGH3 to initiate apoptosis. 

Mutation of the three amino acids RGD into GRL was accompanied by a strong apoptotic response even in the wild-type form of BIGH3. The strong apoptotic response triggered by all the constructs containing the GRL mutations prevented us from studying the role of RGD in apoptosis. Mutation of the PDI domain itself at amino acids 616-618 completely abolished the proapoptotic capacity of a lattice construct. It seems therefore that mutated BIGH3 induces apoptosis through an integrin-related pathway. 

Cellular models are useful tools for dissecting complicated physiological and pathologic mechanisms, but they should be validated by in vivo experiments. Apoptosis is difficult to assess, because only a very small proportion of cells may show signs of apoptosis at any given time. This is true even more so when the disease under study is a slow-progressing disease such as BIGH3-related CD. Kim et al. ²⁶ recently reported a study of apoptosis in keratoconus. Using the terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) assay, they were able to identify apoptosis in two of four lattice CDs and in one of three corneas with granular stromal dystrophy, whereas apoptosis was never observed in six normal corneas. Although the main purpose of their work was not to study apoptosis in CD, their finding provide strong support to our observation that apoptosis is a major event in CD. 

Although BIGH3 is expressed in various tissues, it seems that deposits are observed only in the cornea of affected individuals. There is no simple answer to explain this. In our model, apoptosis was also seen in HeLa, a cell line not originating from the cornea. One could postulate that the mutated BIGH3 protein and/or degradation products induce apoptosis and that the cornea lacks a general clearing mechanism, allowing for the progressive accumulation and deposition. A thorough analysis of corneas obtained from patients and the development of animal models are now needed to investigate and evaluate in vivo the role of apoptosis in the pathogenesis BIGH3-related CD. 

In summary, in our study mutated rBIGH3 protein triggered apoptosis in HCE and HeLa cells. The results suggest that aberrant activation of the α3β1 integrin-related pathway by mutated BIGH3 protein is part of the pathophysiological process that leads to BIGH3-related CD. 

 Supported Grant 31-52940 from the Swiss National Science Foundation.

 Submitted for publication July 3, 2002; revised January 2, 2003; accepted January 27, 2003.

 Disclosure: S. Morand, None; V. Buchillier, None; F. Maurer, None; C. Bonny, None; Y. Arsenijevic, None; F.L. Munier, None; D.F. Schorderet, 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 1.

 

Schematic view of the constructs used. SS: signal sequence; Fas-1 to Fas-4: internal domains homologous to the Drosophila melanogaster fasciclin gene. The position of the different mutations is indicated. (A) Point mutations to generate constructs that reproduce known human mutations of BIGH3. (B) Constructs with deletion of a C-terminal region extending from amino acid 541 to 683. (C) Constructs with mutations in the RGD domain. (D) Constructs with mutations in the PDI domain.

Figure 2.

 

BIGH3 expression analysis by Western blot after transfection of the different constructs as indicated. (A) Culture medium was collected at the indicated time and a human antibody against BIGH3 was used. *Recombinant BIGH3; **human endogenous BIGH3 and bovine BIGH3 from the serum used for tissue culture. (B) Evaluation of the expression of the indicated constructs. One microgram of each construct was transfected, and an equal volume of cell culture medium was recovered 48 hours after transfection and analyzed as in (A). Asterisks are as in (A).

Figure 3.

 

Photographs of HCE cells in culture stained with Hoechst/PI, 24 hours after transfection with (A) pBigWT, (B) pBigR124C, and (C) pBigR555W. (D) Annexin V labeling of HCE cells.

Figure 4.

 

Effect of overexpression of the indicated BIGH3 constructs. HeLa and HCE cells were transfected. The percentage of apoptotic cells is given.

Figure 5.

 

(A) LDH activity in HeLa and HCE cells. LDH efflux is given as the percentage of the total LDH released. (B) Western blot analysis of caspase-3 activation. Reduction of the β-tubulin-normalized procaspase-3 is indicated as a percentage of the activity observed in pBigWT.

Figure 6.

 

Apoptotic responses of HCE and HeLa cells transfected with constructs harboring a C-terminal deletion. The percentage of apoptotic cells is shown.

Figure 7.

 

Apoptotic responses of HeLa and HCE cells transfected with constructs harboring various mutations in the PDI and RGD domains of BIGH3. The percentage of apoptotic cells is shown.