Accumulation of Mitochondrial DNA Damage in Keratoconus Corneas

Abstract

purpose. To determine whether keratoconus (KC) corneas have more mitochondrial (mt)DNA damage than do normal corneas.

methods. Thirty-three normal corneas and 34 KC corneas were studied. Immunohistochemistry for mitochondria-encoded cytochrome c oxidase (complex IV) subunit 1 (CO-Ι) and porins was performed. Total DNA was isolated and mtDNA genome amplified by either long-extension–polymerase chain reaction (LX-PCR) or short-extension–PCR (SX-PCR). LX-PCR mtDNA was digested with restriction enzymes to confirm full-length mtDNA amplicon. SX-PCR mtDNA was probed by Southern blot analysis. The T414G mutation was analyzed by peptide nucleic acid directed clamping PCR. Real-time PCR measured the ratio of mtDNA to nuclear (n)DNA.

results. KC corneas had decreased CO-Ι in areas of corneal thinning. LX-PCR mtDNA digested with restriction enzymes showed expected size bands except for PstI, which showed two additional bands in some KC corneas (2/18). By both LX-PCR (7.4 ± 3.8 vs. 4.3 ± 2.7, P < 0.04) and SX-PCR (5.5 ± 0.55 vs. 2.4 ± 2.0, P < 0.006), KC corneas had an increased number of smaller-sized bands (representing mtDNA deletions/mutations) compared with normal corneas. Southern blot analysis of SX-PCR products confirmed their mtDNA origin. The T414G mutation was not detected in either KC or normal corneas. KC corneas showed a trend of lower mtDNA-to-nDNA ratio (26%, P < 0.7) than did normal corneas.

conclusions. KC corneas exhibit more mtDNA damage than do normal corneas. The previously reported increased oxidative stress and altered integrity of mtDNA may be related to each other and may be important in KC pathogenesis.

Keratoconus (KC) is a noninflammatory thinning disorder that typically involves the central or inferior cornea. The reported incidence of KC is approximately 1 in 2000 individuals. ^{1 2} KC is a significant clinical problem and a leading indication for corneal transplantation. ³ Since the early 1990s, studies have suggested that KC corneas are oxidatively damaged and have abnormalities in stress-related enzymes, such as extracellular superoxide dismutase, catalase, inducible nitric oxide synthase, aldehyde dehydrogenase 3, and glutathione-S-transferase. ^{4 5 6 7 8} KC corneas also have increased levels of malondialdehyde and peroxynitrite, which are cytotoxic by-products of the lipid peroxidation and nitric oxide pathways. ⁷ Tissues with oxidative damage have increased levels of superoxides, hydrogen peroxide, and hydroxyl radicals collectively known as reactive oxygen species (ROS). The underlying mechanisms of altered antioxidant enzymes and oxidative damage in KC corneas are not clear, but it has been suggested that these components play a part in the pathogenesis of KC. ⁹  

One of the pathways used to metabolize glucose in the human cornea is the tricarboxylic acid (TCA) or Krebs cycle, which occurs within the mitochondria. ¹⁰ In most tissues, mitochondria are a significant endogenous source of ROS, which are generated through oxidative phosphorylation (OXPHOS). ¹¹ In OXPHOS, as excess electrons are donated to oxygen molecules, superoxides are formed. It is estimated that up to 4% to 5% of consumed mitochondrial oxygen is converted to ROS, ^{12 13 14} and these ROS, if not eliminated, can cause damage to DNA, lipids, and proteins. Mitochondria have unique circular DNA that is particularly prone to oxidative damage compared with nuclear (n)DNA because of its proximity to the endogenously generated ROS. ^{15 16 17 18 19 20 21} Furthermore, mitochondrial (mt)DNA is intronless, with a high transcription rate, resulting in a high probability of oxidative modification of the coding region. Finally, mtDNA repair systems appear to be less efficient than those of nDNA. ^{15 16 17} There are serious consequences to having mtDNA rearrangements/deletions, since the mtDNA encode for 13 OXPHOS proteins, 22 transfer (t)RNAs, and 2 ribosomal (r)RNAs. ²² When mitochondria are damaged, OXPHOS is decreased, and ROS production is increased. ^{18 19} The resultant mitochondrial dysfunction can lead to altered gene expression, apoptosis, and loss of cell viability. ^{23 24 25} To date, very little is known about human corneal mitochondria or mtDNA. 

Many recent studies have focused on the role of mitochondria as mediators of oxidative damage in aging and diseases. Mitochondrial dysfunction, ROS formation, and oxidative damage are associated with cardiovascular disease, vasculitis, lupus erythematosus, arthritis, adult respiratory distress syndrome, Alzheimer’s disease, amyotrophic lateral sclerosis, Down syndrome, Parkinson’s disease, smoking-related disease, cataract formation, and retinal degenerations. ^{19 24 25 26 27 28} In some diseases, such as Leber’s hereditary optic neuropathy, mutations of mtDNA are transmitted through maternal inheritance, ^{29 30 31} whereas accumulations of somatic mtDNA damage are present in Alzheimer disease and prostate cancer. ^{32 33 34} In addition, specific mtDNA mutations, such as the T414G mutation in the mtDNA control region, have been identified in diseased and aging tissues by highly sensitive techniques such as peptide nucleic acid (PNA)–directed clamping PCR. ^{34 35}  

KC corneas have abnormal antioxidant enzymes and alterations in the lipid peroxidation and nitric oxide pathways that cause increased levels of ROS and reactive nitrogen species (RS). ^{4 5 6 7} The mtDNA have a higher susceptibility to RS-related damage than nDNA. ^{15 16 17 18 19 20 21} Based on these findings, we hypothesized that RS-related oxidative stress leads to mtDNA damage and altered mitochondrial specific proteins. 

Materials and Methods

Corneal Tissues

In this study, 33 normal corneas and 34 keratoconus corneas were used. The normal corneas were obtained from the National Disease Research Institute (NDRI; Philadelphia, PA) within 24 hours of death. KC corneas were collected from ophthalmologists within 24 hours after surgery. On arrival, corneas were either processed for immunohistochemistry or DNA extraction. The study was approved by the institutional review boards of University of California Irvine and Cedars-Sinai Medical Center. Informed consents were obtained from participants, and the study was performed according to the tenets of the Declaration of Helsinki for research involving human subjects. The corneal transplants were performed by corneal specialists. The diagnosis of KC was based on one or more of the following clinical features: Munson’s sign and Rizzuti phenomenon; slit-lamp findings: stromal thinning, Vogt’s striae, Fleischer ring, and scarring—epithelial or subepithelial; retroillumination signs: scissoring on retinoscopy, oil droplet sign (“Charleux”); photokeratoscopy signs: compression of mires inferotemporally, inferiorly or centrally; or videokeratography signs: localized increased surface power and/or inferior superior dioptric asymmetry. 

Immunohistochemistry

Fifteen normal corneas and 16 KC corneas were examined by immunohistochemistry using the polyclonal antibody to the mitochondrial-encoded cytochrome c oxidase (complex IV) subunit 1 (1:50; complex IV subunit 1; CO-Ι; A-6403, clone 1D6; Molecular Probes, Eugene, OR) or the monoclonal antibody to porin (VDAC; 1:100, clone 20B12; Molecular Probes). The secondary antibody was fluorescein-conjugated goat anti-mouse IgG (1:50, Chemicon, Temecula, CA). 

On arrival in the laboratory, corneas were embedded in optimum cutting temperature (OCT) compound (Tissue-Tek; Sakura Finetech, Torrance, CA) before they were frozen in liquid nitrogen. Five-micrometer-thick tissue sections were cut with a cryostat (Leica, Deerfield, IL), mounted onto microscope slides, and stored in −80°C until further use. Tissue sections were thawed at room temperature, fixed with a 2% paraformaldehyde solution for 10 minutes, and rinsed in phosphate-buffered saline (PBS). Primary antibody was applied to the tissue, incubated 1 hour in a humidified chamber, rinsed with PBS, and incubated 1 hour in secondary antibody. Sections were rinsed with PBS and mounted with glycerol/PBS (1:1). Slides were examined and photographed using a fluorescence microscope (Nikon, Tokyo, Japan) with an attached digital camera. Control experiments included use of the secondary antibody alone. 

Extraction of Total DNA

Normal corneas (n = 18; mean age, 51 years; range, 12–76) and KC corneas (n = 18; mean age, 36 years; range, 12–67) were used (Table 1) . Corneas were snap frozen and stored at −70°C. The frozen corneas were pulverized individually to a fine powder under liquid nitrogen with a mortar and pestle. Whole corneas were then homogenized in 1.5 mL of 1× STE (100 mM NaCl, 25 mM Na₂EDTA, 10 mM Tris-HCl [pH 8.0]). Proteins were digested overnight at 50°C in the presence of 0.5% SDS and 15 μg/mL proteinase K (Invitrogen, Carlsbad, CA). The cellular RNAs were digested for 1 hour at 37°C with 5 μg/mL RNase A (Invitrogen) followed by phenol extraction (Invitrogen). The DNA was precipitated with 0.5 volume 7.5 M ammonium acetate and 2 volumes ethanol. 

Long-Extension–PCR

The total DNA isolated from normal corneas (n = 10; median age, 37 years; range, 12–56) and KC corneas (n = 16; median age, 33 years; range, 12–48) were used. The LX-PCR was performed according to modified methods of Jessie et al. ³² and Melov et al. ³⁶ Briefly, LX-PCR was performed with 50 ng of extracted DNA, 5 picomoles of primer, 25 μL of PCR mix (Premix D; Epicentre, Madison, WI), and 1 μL of enzyme mix (Failsafe; Epicentre), in a total volume of 50 μL. Samples were denatured for 2 minutes at 90°C, followed by 35 cycles of 10 seconds at 90°C, 16 minutes at 68°C, and a final elongation for 10 minutes at 72°C; forward primer: AAG TAG TAC GCC TCT ACA ACC TACC (15,148–15,174 bp); reverse primer: TTC ATC ATG CGG AGA TGT TGG ATGG (14,841–14,816 bp). These PCR primers were designed specifically to amplify 16,262 bp of the mtDNA, skipping only 334 bp of the cytb gene. To check for the possibility of the spurious amplification of nuclear mtDNA pseudogenes, we applied the primers to cells lacking mtDNA (rho⁻ cells) to ensure that no mtDNA-like sequences could be amplified. LX-PCR products were separated by electrophoresis on an 0.8% agarose gel stained with ethidium bromide. 

To verify by restriction site location the validity of the mtDNA, the samples were treated with five different restriction enzymes—BalI, EcoRI, HindIII, PstI, and ApaI (Invitrogen)—run on an 0.8% agarose gel and then stained with ethidium bromide. The results were compared to those at restriction sites found on www.MITOMAP.org, a human mitochondrial genome database describing known polymorphic mtDNA restriction sites detected by high-resolution screening of human populations. The PstI restriction digestion showed additional bands in two KC corneas. The D loop region of the KC corneal mtDNA was sequenced to identify the nucleotide substitutions. Briefly, the forward primer was 16225F: CAA CTA TCA CAC ATC AACTG. The reverse primer was 16306R: GTA CTG TTA AGG GTG GGTAG. The products were prepared by PCR cleanup, to remove primers and free dNTPs (ExoSAP-IT; USB Corp., Cleveland. OH) and sequenced. Analysis was performed by Chromas/Pro (Technelysium Pty., Ltd., Tewantin, Queensland, Australia) and BioEdit Biological Sequence Alignment Editor (The RNA Society, Madison, WI). 

Short-Extension–PCR

SX-PCR limits the efficiency of full-length mtDNA amplification and selectively amplifies smaller PCR products that are mtDNA in origin. ^{32 36 37} The procedure of Melov et al. ³⁶ was used for selective amplification of the corneal mtDNA PCR products that were <5 kb and >1 kb. ³⁶ Both normal (n = 5; mean age, 31 years; range, 12–39) and KC (n = 6; mean age, 23 years; range, 12–39) corneal DNA samples were used. SX-PCR reactions were performed with the same PCR conditions as LX-PCR, except that 20 ng of DNA was used, and the annealing time was decreased from 16 to 6 minutes at 68°C. Amplified products were separated on 0.8% agarose gels stained with ethidium bromide. All gel images were captured on a fluorescence imaging system (FMBio III; Hitachi, South San Francisco, CA) under identical conditions and were analyzed. 

Southern Blot Analysis

Normal human corneal fibroblast cultures were established according to the methods of Brown et al. ³⁸ The total DNA was isolated and amplified for mitochondrial DNA by using primers specific for the entire mitochondrial genome. The PCR products were separated on a 0.8% agarose gel. Multiple 16.2-kb bands were excised and purified (Nucleotrap Gel Extraction Kit; BD Biosciences, Palo Alto, CA). The eluates were pooled, extracted with phenol-chloroform, and precipitated with ethanol, sodium acetate, and linear acrylamide as a carrier. SX-PCR products were transferred from 0.8% agarose gels to positively charged nylon membranes (Hybond-N+; Amersham Inc., Arlington Heights, IL) using an alkaline blot analysis procedure. The membrane was hybridized with the mitochondrial DNA probe randomly labeled with digoxigenin-11-dUTP (Enzo Life Sciences, Farmingdale, NY) and washed at high stringency. Membrane-hybridized digoxigenin was detected with an anti-digoxigenin horseradish peroxidase–tagged antibody (Enzo Life Sciences). The membrane was developed with a chemiluminescent substrate (SuperSignal West Pico Chemiluminescent Substrate; Pierce Biotechnology, Rockford, IL) and exposed to autoradiograph film (BioMax Light Film; Eastman Kodak, Rochester, NY). 

Real-Time PCR Analysis

Real-time PCR was performed with total DNA isolated from normal (n = 10; mean age, 37 years) and KC (n = 16; mean age, 33 years) corneas by using the techniques described by Rodriguez-Santiago et al. ³³ Briefly, DNA was analyzed for either r18s representing nDNA or mtND2 representing the mtDNA. The mtND2 gene is in a relatively stable region near the light-chain origin of replication where mtDNA deletions are not common, and no pseudogenes have been described. ³³ Serial dilutions of DNA control sample were prepared to compare the amplification efficiencies of the target genes by real-time PCR. The difference in threshold cycle (C_T) was plotted against the logarithm of the template amount. The DNA dilution with the slope of the resultant straight line less than 0.1 was chosen for analysis. 

The real-time PCR reactions were performed in a 50-μL final volume containing 50 ng of DNA, 25 μL PCR master mix (QuantiTect SYBR Green PCR Master Mix; Qiagen, Inc., Valencia, CA), and 0.5 μM of each primer, mtND2 and r18s (Sigma Genosys; The Woodlands, TX). The oligonucleotide primer for mtND2 sequences were 5′-GCCCTAGAAATAAACATGCTA-3′ and 5′-GGGCTATTCCTAGTTTTATT-3′. The oligonucleotide sequences for r18s were 5′-ACGGACCAGAGCGAAAGCAT-3′ and 5′-GGACATCTAAGGGCATCAC AGAC-3′. The real-time PCR program consisted of an initial denaturation at 94°C for 15 minutes followed by 35 amplification cycles at 94°C for 15 seconds, 53°C for 30 seconds, and 72°C for 30 seconds. The fluorescent product was detected at the last step of each cycle. After amplification, a melting curve was acquired by cooling the product to 65°C, then heating from 65°C to 98°C with fluorescent readings taken at every 0.2°C. Melting curves were used to determine the specificity of the PCR products. An internal standard was set up to normalize the relative gene expression level, and this standard was run with each different experiment. 

Detection of the T414G Mutation

The aging and disease-associated mtDNA control region mutation on the 414 nucleotide pair (np) was examined in age-matched normal (n = 16) and KC (n = 16) corneal DNA by the PNA-clamping PCR procedure as described in Murdock et al. ³⁵ The presence of theT414G mutation in the PCR product was confirmed by FokI cleavage. The T414G− sample was the PCR product from wild-type plasmid. The T414G+ sample was the PCR product from the T414 mutant plasmid. ³⁴  

Statistical Analysis

The data from the LX-PCR mtDNA and the SX-PCR mtDNA were analyzed by unpaired t-test (two-tailed; GraphPad Software, Inc., San Diego, CA). With the real-time PCR results, relative quantitation was analyzed by the comparative C_T method. ^{39 40}  

Results

In this study, we found that regions of corneal stromal thinning have diminished staining of the mitochondrial-encoded CO-Ι; KC corneas have significantly increased levels of mtDNA damage compared with age-matched normal corneas; and the T414G mutation of the mtDNA control region is not involved in KC. 

Immunohistochemistry

Porins, which are mitochondrial structural proteins that tend to be very stable, were found in the cytoplasm of epithelial, stromal, and endothelial cells in similar patterns in both normal corneas and KC corneas (Fig. 1A) . Normal and KC corneas then were examined for the presence of CO-Ι. In the normal corneas, basal epithelial cells showed positive CO-Ι staining (Fig. 1B , arrows). In contrast, KC corneas had diminished CO-Ι staining in the basal epithelial cells in areas of active disease, which are regions of thinning stroma and absent Bowman’s layer so that the epithelium and stroma are in direct contact with each other. ⁴¹ The CO-Ι staining was normal in regions of KC corneas that had intact Bowman’s layer and relatively normal corneal thickness (data not shown). 

Restriction Enzyme Digestion of LX-PCR mtDNA

Figure 2Ashows a schematic of circular mtDNA depicting the locations of the mtDNA-specific LX-PCR primers and the cleavage sites of the restriction enzymes. The LX-PCR mtDNA from normal (n = 16) and KC (n = 16) corneas were digested with five different restriction enzymes (Fig. 2B) . A representative ethidium-bromide–stained gel of normal corneal mtDNA showed that the uncut LX-PCR product was approximately 16.2 kb (Fig. 2B , lane 1), consistent with the known size of the mtDNA. After cleavage with the BalI, EcoRI, HindIII, and ApaI restriction enzymes, the products from normal and KC corneal mtDNA were similar to each other (Fig. 2B) . Normally, the PstI restriction enzyme cuts the mtDNA into 8.3-, 5.8-, and 2.1-kb bands (Fig. 2C , lane 2). The LX-PCR mtDNA from two KC corneas had two additional bands of ∼7.2 kb and 1.1 kb (Fig. 2C , lane 1). None of the normal corneal LX-PCR mtDNA had these PstI patterns. The D loop region was sequenced to identify the nucleotide substitution that would account for the PstI cleavage. Three substitutions in the control region of the D loop were found: 16,217 t→c; 16,247 a→g; and 16,261 c→t. The substitution at nucleotide 16,247 yielded a nucleotide sequence recognized by the PstI enzyme (ctgca/g). 

LX-PCR mtDNA Analysis

The LX-PCR method with the increased extension time allows amplification of the entire mitochondrial genome and any subgenomic-sized product that contains the primer sites. The major LX-PCR product representing full-length mtDNA is ∼16.2 kb. When nucleotide substitutions, rearrangements, or deletions are present in the mtDNA template, LX-PCR yields smaller-sized products. Because in other tissues the mtDNA deletions increase after 60 years of age, ⁴² we examined only corneas <59 years of age to minimize these age-related mtDNA deletions. 

The KC corneas have an increased number of additional smaller-sized LX-PCR mtDNA products than do normal corneas. Smaller-sized mtDNA bands (<16.2 kb and >1 kb) in normal and KC samples were quantified, and the mean values ± SD calculated (Figs. 3A 3B) . The normal versus KC corneal LX-PCR mtDNA had 4.3 ± 2.7 and 7.4 ± 3.8 smaller-sized bands per individual, respectively, (P < 0.04). The smaller products were heterogeneous in size, and no one pattern was dominant. Based on the size of the LX-PCR products, one can infer that the mtDNA deletions eliminated between 6 and 14 kb, which is consistent with reported mtDNA rearrangements in other tissues. 

SX-PCR mtDNA and Southern Blot Analysis

SX-PCR uses a shorter extension time so that smaller-sized mtDNA products are selectively amplified. Figure 4Ashows an ethidium bromide–stained gel of SX-PCR mtDNA from normal and KC corneas, with each lane representing a different individual. Note that although the normal samples were older (mean age, 31 years) than the KC samples (mean age, 23 years), it was the KC samples that had significantly greater numbers of smaller-sized PCR products, representing random nucleotide changes or deletions. When quantified, the KC group had 5.5 ± 0.55 bands per individual versus the normal group with 2.4 ± 2.1 bands per individual (P < 0.006). To verify that the bands were mtDNA specific, the samples were analyzed by Southern blot analysis using a digoxigenin-labeled mtDNA probe (Fig. 4B) . The banding pattern on the Southern blot was nearly identical with the ethidium-bromide–stained gel, indicating that the products were indeed mtDNA. 

PNA-Clamping PCR Analysis

The PNA-clamping PCR procedure of the normal and KC corneas allows identification of a T414G mutation in the control region of mtDNA. In Figure 5 , the procedure was run on a T414G-positive (T414G+) and a T414G-negative (T414G− or wild-type) control. If a sample has the T414G mutation, a PCR product is amplified in the presence and absence of PNA (PNA+ and PNA−). However, in a sample that lacks the T414G mutation, a PCR product is only evident when PNA is absent (PNA−) and no product is amplified when PNA is added (PNA+). When the PNA-clamping PCR procedure was applied to normal and KC corneal DNA, the product was amplified in the absence of PNA, but when PNA was added, no mutant products were found, indicating that the T414G mutation was lacking in all corneas examined. 

Real-Time PCR Analysis

The mtND2 and r18s were analyzed by real-time PCR measuring fivefold dilutions of the total DNA isolated from each cornea (Table 2) . We found a 26% reduction of mtDNA content in KC corneas compared with age-matched normal corneas (P < 0.7). 

Discussion

Our study is the first to examine mtDNA in human corneal tissues. In this study, that KC corneas (median age, 33 years) had increased mtDNA damage compared with normal corneas (median age, 37 years). By immunohistochemistry, regions of active disease in KC corneas exhibited decreased staining for CO-Ι, an important mitochondrial-encoded subunit of complex IV of OXPHOS. In addition, by LX-PCR and SX-PCR, KC corneas had statistically significant increases in the number of smaller-sized bands representing mtDNA deletions and mutations. Furthermore, KC corneas showed a trend for a decreased mtDNA-to-nDNA ratio, compared with age-matched normal corneas. This mtDNA damage could lead to decreased OXPHOS, generation of additional ROS, and further oxidative damage, causing a “vicious cycle” of oxidative stress that contributes to KC pathogenesis. These data support our hypothesis that oxidative stress, including mtDNA damage, is present in KC corneas and may play a central role in the pathologic course of KC. 

Using the LX-PCR procedure that detects the qualitative changes in mtDNA, ^{32 36 43} we found that KC corneas had significantly increased numbers of mtDNA deletions compared with normal corneas, similar to findings in prostate cancer, Down syndrome, Alzheimer’s disease, Parkinson’s disease, and diabetes. ^{19 32 33 44 45 46 47 48} Acquired mtDNA mutations, such as those found in aging tissues or diseases, are often heterogeneous and can accumulate to over 50% before developing clinically recognized pathologic states. ³⁰ We speculate that KC corneas acquire mtDNA deletions and that they contribute to its pathogenesis. 

KC corneas had a significant increase of these smaller-sized mtDNA bands compared with normal corneas, which is consistent with other diseased tissues. ^{32 36} Numerous studies suggest that mtDNA deletions and mutations increase with age. ^{37 42} In the SX-PCR experiments, the two youngest normal corneas (12 and 29 years) had virtually no bands, despite selective amplification for smaller-sized products (Fig. 4 , lanes 1, 2). However in the KC group, all the corneas <30 years of age, including the youngest (12 years), had multiple SX-PCR mtDNA bands (Fig. 4 , lanes 6–10). As suggested by other studies, we selectively amplified the products between 1 and 5 kb and then demonstrated by Southern blot analysis that these smaller-sized PCR products were mtDNA, not PCR artifact. ^{32 36 37} It should be noted that the median age for KC corneas was generally younger than that of the normal corneas. If the mtDNA mutations were a consequence of aging, then one would expect to find more mutations in the normal group than in the KC group. This was contrary to our findings and therefore, supports our hypothesis that KC mtDNA damage is disease related and not age related. 

With respect to the LX- and SX-PCR mtDNA deletions, the younger KC corneas (age range, 12–29 years; Fig. 3 ) had a pattern of smaller-sized bands that were similar to those in older normal corneas (age range, 38–48 years). This may be related to some type of premature aging, such as is described in patients with Down syndrome. ^{49 50 51} The incidence of KC in patients with Down syndrome is 50 to 300 times higher than in the general population, ^{52 53 54 55} and, as in KC, oxidative stress with mitochondrial damage has been implicated in Down syndrome pathogenesis. ^{44 45 51} Mitochondrial dysfunction and subsequent increase of ROS formation may play a role in both KC and Down syndrome. 

PNA-clamping PCR is a highly sensitive technique to identify low levels of point mutations in mtDNA that occur with aging or disease. ^{34 35 56} The somatic T414G mutation is in the mtDNA control region, adjacent to the mtDNA transcriptional start sites. ³⁵ The T414G mutation is found in 65% of Alzheimer’s brains but is absent in age-matched normal brains. ³⁴ In this study, the T414G mutation was not found in either normal (n = 16; age range, 12–76 years) or KC (n = 16; age range, 12–67 years) corneas. The T414G mutation was unexpectedly absent, even in older corneas, which is different from the accumulation of the T414G mutation found in muscle from individuals >35 years of age. ³⁵ This suggests that the T414G mutation may be tissue specific and at least in the cornea does not occur. It may be that other control region mutations are involved in KC corneas, and their involvement will be the subject of future studies. 

Our real-time PCR study showed the KC corneas had a 26% reduction of mtDNA compared with age-matched normal corneas, similar to the 28% mtDNA reduction reported for Alzheimer’s brains. ³³ These values represent a trend toward a lower mtDNA to nDNA but are not statistically significant, which is surprising considering the extensive mtDNA damage found in the KC corneas. One explanation may be that the mtND2 gene is in a stable region of the mtDNA genome not susceptible to damage. ³⁴ Moreover, clinically, KC corneas are not homogeneous throughout. For example, the stromal thinning and topographical steepening are typically found in the inferior half of the cornea but the superior region retains relatively normal characteristics. ² In the present study, we collected the DNA from the entire cornea, perhaps masking the mtDNA changes found in the affected, lower portion of the cornea. In addition, as the KC cornea thins, apoptosis occurs, ⁵⁷ leading to cellular loss in the anterior region of the cornea, again decreasing the amount of mutated mtDNA in the pool of total mtDNA. 

Restriction digestion was used extensively to identify mtDNA patterns. With four of the restriction enzymes (BalI, EcoRI, HindIII, and ApaI), the patterns between normal and KC were similar to each other. In our study, the LX-PCR mtDNA from two KC corneas had two additional bands of ∼7.2 kb and 1.1 kb after digestion with PstI, whereas none of the normal corneas had this pattern. This mtDNA pattern is heteroplasmic (mixed pattern of multiple mtDNA forms), since some of the 8.3-kb band is still present. The expected cleavage pattern for PstI is at two sites (6914 and 9024) yielding products of 8.3, 5.8, and 2.1 kb (Fig. 2C) . The only reported mtDNA polymorphism for PstI is a loss of the 6914 site (www.MITOMAP.org). Since the cleavage pattern found in the two KC corneas had not been reported as a polymorphism (www.MITOMAP.org), we sequenced the D loop region of the mtDNA and found three nucleotide substitutions, one of which (16,247 a→g) accounted for the gain of a previously unreported PstI site. These substitutions have been reported in populations with Polynesian mtDNAs, ^{58 59} and it is unlikely that they play a role in KC. 

The cause of the mtDNA damage in KC corneas is unknown. However, KC corneas have abnormal antioxidant enzymes, including a ∼2-fold increase in RNA and activity levels of catalase ⁸ which is upregulated in response to hydrogen peroxide. ^{60 61 62} In addition, increased iNOS levels are present ⁷ which suggests elevated nitric oxide, an RS. Moreover, KC corneas also have an accumulation of malondialdehyde, a cytotoxic aldehyde, and peroxynitrite, a result of superoxides and nitric oxide, both of which can damage DNA. The presence of these reactive species may cause acquired mtDNA damage within the KC corneas. Furthermore, we speculate that the loss of mtDNA integrity may decrease OXPHOS and further increase ROS formation, leading to additional oxidative stress. This cycle of events may play a significant role in KC pathogenesis. 

 Supported by National Eye Institute Grant EY06807, the Schoellerman Charitable Foundation, the Discovery Fund for Eye Research, the Guenther Foundation, Research to Prevent Blindness, the Skirball Molecular Ophthalmology Program, and the National Keratoconus Foundation.

 Submitted for publication November 29, 2004; revised January 4, 2005; accepted January 5, 2005.

 Disclosure: S.R. Atilano, None; P. Coskun, None; M. Chwa, None; N. Jordan, None; V. Reddy, None; K. Le, None; D.C. Wallace, None; M.C. Kenney, 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.

 

Table 1.

 

Characteristics of Normal and Keratoconus Corneas Used for the mtDNA Analysis

Normal	Age	Gender	Race	Cause of Death	Keratoconus	Age	Gender	Race
1	12	M	B	MVA	1	12	M	N/A
2	29	F	W	Seizure	2	17	F	B
3	33	M	W	CVA	3	20	M	W
4	38	F	W	Asphyxia-suicide	4	22	M	N/A
5	39	M	N/A	N/A	5	28	F	N/A
6	39	F	W	Pneumonia	6	29	M	H
7	42	M	B	Stomach cancer	7	30	F	W
8	44	F	B	Heart failure	8	30	M	H
9	46	M	W	Respiratory failure	9	35	F	W
10	48	M	W	CVA/CMV	10	35	M	B
11	60	NA	N/A	GSW chest	11	38	M	W
12	60	F	W	Fall-brain death	12	39	M	W
13	64	M	W	Aortic dissection	13	39	M	N/A
14	70	M	W	Ventricular fibril.	14	41	M	N/A
15	73	F	W	MI/diabetes	15	54	F	W
16	74	F	W	Respiratory failure	16	56	M	W
17	75	M	W	MI	17	62	M	N/A
18	76	F	W	Brain hemorrhage	18	67	M	H

 N/A, not available; W, White; B, African-American; H, Hispanic; MI, myocardial infarction; CVA, cerebral vascular accident; GSW, gun shot wound; MVA, motor vehicle accident; CMV, cytomegalovirus.

Figure 1.

 

Immunohistochemistry showed decreased CO-Ι staining in KC corneas. (A) In normal and KC corneas, the porin antibody staining of epithelial, stromal, and endothelial cells was similar. (B) In the normal corneas, anti-CO-Ι antibody stained the basal epithelial cells, stromal, and endothelial cells. In KC corneas, the regions of active disease showed decreased CO-Ι staining in epithelial cells and anterior stromal cells. No staining was observed when only secondary antibody (IgG only) was used on the tissue sections. CO-Ι, cytochrome c oxidase (IV) subunit 1; NL, normal; KC, keratoconus; E, epithelium; S, stroma; Endo, endothelium. Scale bar, 30 μm.

Figure 2.

 

Restriction enzyme digestion of corneal LX-PCR mtDNA. (A) Schematic of double-stranded circular mtDNA showing locations of PCR primers and restriction enzyme sites. Arrows: The 3′ end of the forward LX-PCR primer (15,148 bp) and reverse (14,841 bp) primer sites. The recognition sites for the restriction enzymes ApaI, BalI, EcoRI, PstI, and HindIII are shown on the mtDNA. (B) Representative ethidium bromide gel of normal (n = 18) and KC (n = 18) corneal LX-PCR mtDNA after restriction enzyme digestion showing that products were the expected sizes for mtDNA. (C) A distinct restriction digestion pattern was generated by PstI in two KC corneas but not in normal corneas. Lane 1: additional bands of ∼7 and ∼1.3 kb that were found in two KC corneal LX-PCR mtDNA after PstI digestion; lane 2: the expected PstI pattern for mtDNA of 8.3, 5.6, and 2.1 kb; lane M: molecular mass marker.

Figure 3.

 

Increased numbers of mtDNA deletions/mutations in KC corneal LX-PCR mtDNA. (A) Ethidium-bromide–stained gel showing LX-PCR mtDNA from normal and KC corneas. Each lane represents a different individual, with ages given above the lanes. The 16.2-kb product corresponds to the full-length mtDNA. Smaller-sized LX-PCR products (<16.2 and >1.0 kb) represent mtDNA deletions. (B) Quantitation of smaller-sized bands per individual in normal and KC corneal LX-PCR mtDNA. Normal corneas had 4.3 ± 2.7 mtDNA bands per individual, and KC corneas had 7.4 ± 3.76. P < 0.04. Lane M: molecular mass marker.

Figure 4.

 

Increased number of mtDNA deletions and mutations in KC corneal SX-PCR mtDNA. (A) Ethidium-bromide–stained gel showing SX-PCR mtDNA products from normal and KC corneas. The KC corneas had an increased number of smaller-sized SX-PCR bands compared with normal corneas (5.5 ± 0.55 vs. 2.4 ± 2.1 bands per individual, respectively; P < 0.006). (B) Southern blot using a digoxigenin-labeled mtDNA probe showed bands similar to those found in the ethidium bromide gel, indicating that these smaller-sized SX-PCR products were mtDNA.

Figure 5.

 

The PNA-clamping PCR procedure shows that the normal (n = 16) and KC (n = 16) corneal mtDNA lacked the T414G mutation. The PCR products were amplified with and without PNA and then digested with FokI. As seen in the positive T414G mutation control sample (T414G+), PCR products were amplified in the presence (lane PNA+) and absence (lane PNA−) of PNA. In the wild-type (T414G−) sample, the PCR product was present when PNA was absent, but the product was not amplified when PNA was added. In all corneal samples, when PNA was absent (PNA−), the PCR products appeared but when PNA was added (PNA+) no PCR products were found, indicating that the corneal samples did not have the T414G mutation.

Table 2.

 

Relative Quantification of mtND2 and r18s by Real-Time PCR

	CT ND2 (Mean ± SD)	CT 18s (Mean ± SD)	ΔCT * (Mean ± SD)	ΔΔCT , † (Mean ± SD)	Rel. to ND2	Range
												Low	High
Internal Standard	13.771 ± 0.393	16.289 ± 0.587	−2.518 ± 0.856	0.000 ± 0.856	1.000	0.553	1.809
NL	14.566 ± 1.067	16.579 ± 0.733	−2.013 ± 0.632	0.505 ± 0.632	0.705	0.455	1.092
KC	14.867 ± 1.752	16.438 ± 0.863	−1.571 ± 1.529	0.946 ± 1.529	0.519	0.180	1.498

 The values represent the mean threshold cycle (C_T) for target amplification ± standard deviation. Each sample was performed in triplicate and with fivefold serial dilutions. Shown are the results for a single dilution representing 10 ng of DNA. NL, normal corneas; KC, keratoconus corneas; C_T, threshold cycles; Rel., Relative.

*  ΔC_T was the difference between C_T of mtND2 and the C_T of 18s.

†  ΔΔC_T was calculated by subtracting the ΔC_T of internal standard from the ΔC_T of NL and KC.

‡  The range for mtND2 relative to NL and KC were determined by the expression: 2^−(ΔΔC_T ± SD).

The authors thank Ezra Maguen, Anthony Nesburn, Frank Price, Yaron Rabinowitz, and Theodore Perl for providing diseased human corneas and the National Disease Research Interchange for supplying normal human corneas.