Search tips
Search criteria 


Logo of brjopthalBritish Journal of OphthalmologyVisit this articleSubmit a manuscriptReceive email alertsContact usBMJ
Br J Ophthalmol. 2007 June; 91(6): 722–727.
Published online 2007 January 10. doi:  10.1136/bjo.2006.111070
PMCID: PMC1955591

Development of a DNA chip for the diagnosis of the most common corneal dystrophies caused by mutations in the βigh3 gene



To develop a diagnostic DNA chip to detect mutations in the βigh3 gene causing the most common corneal dystrophies (CDs).


Samples from 98 people, including patients with βigh3‐associated CDs (β‐aCDs), were examined. Specific primer and probe sets were designed to examine exons 4 and 12 of the βigh3 gene, in order to identify mutant and wild‐type alleles. Mutations were then identified by hybridisation signals of sequence‐specific probes immobilised on the slide glass.


Direct sequencing of exons 4 and 12 of the βigh3 gene in the patients' genome showed that β‐aCDs could be mainly classified into five types: homozygotic Avellino corneal dystrophy (ACD), heterozygotic ACD, heterozygotic lattice CD I, heterozygotic Reis–Bucklers CD and heterozygotic granular CD. Blind tests were performed by applying the target DNA amplified from the genomic DNA isolated from the peripheral blood of the participants onto a DNA chip. The results obtained by DNA chip hybridisation matched well with the direct DNA sequencing results.


The DNA chip developed in this study allowed successful detection of β‐aCDs with a sensitivity of 100%. Mutational analysis of exons 4 and 12 of the βigh3 gene, which are the mutational hot spots causing β‐aCDs, can be successfully performed with the DNA chip. Thus, this DNA chip‐based method should allow a convenient, yet highly accurate, diagnosis of β‐aCDs, and can be further applied to diagnose other types of CDs.

Genetic analyses of most common βigh3‐associated corneal dystrophies (β‐aCDs) such as Avellino corneal dystrophy (ACD), Reis–Bucklers CD (RBCD), lattice CD type I (LCD I) and granular CD (GCD) have shown that most cases are caused by amino acid substitutions within the βigh3 gene. In the case of ACD, the 124th amino acid arginine of the βigh3 protein is substituted by histidine. Similarly, LCD I is caused by R124C, whereas RBCD is caused by R124L. GCD (also called GCD I) is caused by R555W.1,2,3 Thus, molecular genetic analysis of the mutations in the βigh3 gene is most adequate to diagnose patients having different types of β‐aCDs, in particular those with an atypical or ambiguous corneal appearance.4,5,6

The molecular genetic analysis of the mutations in the βigh3 gene has mainly relied on direct sequencing. However, DNA microarray‐based detection of point mutations can be a much more efficient alternative. To analyse the point mutation in a gene by direct sequencing, the DNA fragment containing a mutated site is amplified and extracted with high purity before sequencing. Even when using the automatic DNA sequencer, each signal peak should be checked directly from the raw data to find just a single gene mutation. This sequential work is rather cumbersome and takes 24–48 h. DNA microarray‐based assay can be a simple but powerful alternative, as it allows simultaneous handling of multiple samples and sequences7 and requires only a small amount of the amplified product that can be directly applied onto a DNA chip without any purification step. It takes no more than 6–8 h, and analysis can be performed simply by checking the signal on the expected probe spot regions. In this paper, we report the development of a DNA chip for detection of the mutations in the βigh3 gene, which allows rapid, convenient and accurate diagnosis of β‐aCDs.

Materials and methods

Patients and collection of samples

This study was performed in accordance with the Helsinki Agreement, and all participants gave their informed consent before the test, which was approved by Severance Hospital IRB Committee (No 4‐2006‐0075).

DNA sequence analysis

DNA was extracted from the peripheral blood of participants with the JETQUICK Blood & Cell Culture DNA Spin Kit (GENOMED GmbH, Löhne, Germany). Exons 4 and 12 of the βigh3 gene were amplified by PCR with origin of replication (Ori) 4 primer sets and Ori 12 primer sets, respectively (table 11 and fig 11).). The PCR was conducted with a PCR thermal cycler T1 (BioMetra GmbH, Gottingen, Germany). The amplified products were purified with a PCR purification kit (Qiagen, Hilden, Germany). For the verification of mutations and identification of the mutation types, the DNA isolated from all the 98 participants were also sequenced with an automatic DNA sequencer (ABI Prism 377, Applied Biosystems, Foster, California, USA).

figure bj111070.f1
Figure 1 Gene maps and haplotypes of the βigh3 gene. Coding exons are marked by black bars. PCR primers used for the amplification of the targets and the amplified products are shown. Arrows indicate the mutational hot spots in exons 4 ...
Table thumbnail
Table 1 Primers used in this study

Microarray fabrication

The 3′ termini of designed probes were modified with amine residues using an amino‐linker column (Cruachem, Glasgow, Scotland) for their immobilisation on the aldehyde‐coated slide glasses (CEL Associates, Huston, Texas, USA) by forming amine–aldehyde covalent bonds. The 15mer‐poly A oligonucleotide labelled with fluorescent dyes such as Cy3, Cy5 or fluorescein isothiocyanate (FITC) was used as position marker. The 11mer, 13mer, 15mer and 17mer probes were designed and tested for their efficiencies in detecting the mutations. The probes were resuspended in 3× saline sodium citrate spotting solution (0.45 M NaCl, 15 mM C6H5Na3O7, pH 7.0) and spotted on the slide glass using MicroGridII (BioRobotics, Cambridge, UK). After spotting, the slide glasses were air dried at room temperature for 12 h or baked at 80°C for 1 h.

Preparation of target DNA

To prepare the target DNA, asymmetric PCR was carried out in a 50 μl reaction mixture containing 1× buffer, 0.2 mM deoxynucleotides triphosphates, 1 U Taq polymerase (Takara Shuzo, Shiga, Japan), 100 ng chromosomal DNA, 3 µM forward primer and 30 μM reverse primer. The 5′ terminus of each reverse primer used in asymmetric PCR was modified with Cy5, Cy3 or FITC. The ACD and Ori 12 primer sets were used for amplifying exons 4 and 12, respectively (table 11).). By using a 10 times higher concentration of fluorescent‐labelled reverse primer, mainly single‐stranded DNAs which are antisense to the probes could be prepared. The ACD and Ori 12 primer sets were used for amplifying exons 4 and 12, respectively (table 11).). The sizes of the amplified products from exons 4 and 12 were 147 and 318 bp, respectively(fig 11).

Hybridisation and data analysis

The spotted slides were steamed with boiling water for 1 min and soaked in 70% (v/v) ice‐cold ethanol to remove any non‐covalently bound oligonucleotides. To block the background fluorescence, the unreacted aldehyde groups were treated with sodium borohydride (NaBH4) solution. The PCR‐amplified target products (10 μl each) were directly added to the hybridisation solution containing 6× saline/sodium phosphate/EDTA (SSPE) (0.9 M NaCl, 10 mM NaH2PO4·H2O, 1 mM EDTA, pH 7.4) and 20% (v/v) formamide. Hybridisation was carried out by incubating at 30°C for 4–6 h in Array Chamber X (GenomicTree, Daejeon, Korea). After washing with 3× SSPE for 5 min, 2× SSPE for 5 min and 1× SSPE for 5 min, the slides were scanned using arrayWoRx (Applied Precision LLC, Issaquah, Washington, USA). Data were analysed by ImaGene 6.0 software (BioDiscovery, El Segundo, California, USA). The signal intensities were calculated by global normalisation for all positive spots, after subtracting the background values from the median signal intensities measured at 595 nm for Cy3, 685 nm for Cy5 and 530 nm for FITC.


The specific probes were designed on the basis of the sequences of the normal and mutant βigh3 genes (fig 11 and table 22).). First, the probe length and the hybridisation time were optimised for the diagnosis of CDs. It was concluded that the use of 15mer probes and hybridisation at 30°C for 4–6 h resulted in the highest sensitivity and specificity for detecting the mutations in exon 4 (fig 22).). The wild type‐to‐mutant signal ratio, defined as the ratio of the spot signal of the wild‐type probe for normal sequence (denoted as W) to that of each type of CD probe for each mutated sequence (denoted as M), was used for the probe optimisation. Each probe signal for the mutated sequence was set to one, and the ratios of other probe signals to each mutated sequence were calculated according to the probe length and the hybridisation time. Figure 2A2A shows that single‐base mismatch could be discriminated successfully using a 13mer probe in the homozygotic ACD, but the 15mer probes showed the highest discrimination efficiency in different types of CDs. The discrimination can be further enhanced by changing the probe concentration. Hybridisation time was optimised with the heterozygotic ACD and the homozygotic ACD samples. As shown in fig 2B2B,, hybridisation for 4–6 h was found to be optimal for the highest sensitivity and specificity. The probes finally selected are shown in bold letters in table 22.

figure bj111070.f2
Figure 2 Hybridisation results according to the probe length (A) and the hybridisation time (B). The results are normalised for the signal obtained with the wild‐type sequence versus that obtained with the mutant sequence (W/M). ACD, Avellino ...
Table thumbnail
Table 2 Probes designed for detecting the mutations in the βigh3 gene

The DNA chip‐based diagnosis of CDs was performed using blind tests for a total of 98 blood samples. The genotypes of these samples were also determined by direct sequencing for the verification of DNA chip results. DNA sequence analysis showed that the 98 participants could be classified as 27 normal, 10 homozygotic ACD, 57 heterozygotic ACD, 1 heterozygotic LCD I, 1 heterozygotic RBCD and 2 GCD (table 33).

Table thumbnail
Table 3 βigh3 gene mutations in 98 participants

All microarray experiments were carried out in duplicate. The results of DNA chip‐based diagnosis matched 100% with those obtained with direct DNA sequencing for all 98 samples. Four types of CDs—namely, homozygotic ACD, heterozygotic ACD, heterozygotic LCD I and heterozygotic RBCD—were detected in 96 samples. Figure 33 illustrates the DNA chip design, the expected results and the representative DNA chip hybridisation results. We also designed the probes for detecting GCD (two samples) in an attempt to analyse the possibility of detecting other types of CDs with DNA chip technology. Figure 44 shows the applied DNA chip design and the representative results. The DNA chip developed in this study could successfully diagnose different types of CDs with 100% sensitivity, 100% specificity and 100% positive predictive value. Table 44 summarises the DNA chip hybridisation results of blind samples. It should be noted that ACD is the most common type of CD detected in this study. Even though we cannot truly claim the high specificity, sensitivity and the predictive value for other types of CDs as the number of samples was too small, the results clearly showed that the DNA microarray developed in this study can successfully distinguish between different types of CDs.

figure bj111070.f3
Figure 3 The first version of DNA chip designed for the detection of mutations in the βigh3 gene. (A) The DNA chip design and probe information. (B) Expected hybridisation results according to the types of the corneal dystrophies (CDs). ...
figure bj111070.f4
Figure 4 The final version of DNA chip designed for the detection of mutations in the βigh3 gene. (A) The DNA chip design using the probes listed in table 22.. (B) Expected hybridisation results according to the types of corneal ...
Table thumbnail
Table 4 Summary of DNA microarray‐based detection of the βigh3 gene in this study


CDs have most often been categorised by their clinical appearance examined by light and electron microscopes. However, their appearance is variable, and thus is often misdiagnosed. Recently, various types of mutations causing different CDs have been reported,8,9,10,11 and the molecular basis for the common and rare CDs is becoming apparent. However, the disease phenotypes of CDs seem to be more diverse than expected previously. Thus, it is desirable to diagnose them accurately by direct DNA sequencing and related genetic analyses. In this paper, we reported the development of a DNA chip that allows simple, yet highly specific and accurate, detection of mutations in the βigh3 gene causing the CDs. The discrimination of the single‐base mismatch from the perfect match could be optimised by examining the different lengths and concentrations of probes and the hybridisation time. Through the blind tests, we were able to conclude that this DNA chip‐based method allows successful detection of various mutation types of the CDs in 98 samples. Thus, the DNA chip developed in this study allows specific and accurate diagnosis of multiple types of β‐aCDs, which should be useful in the present clinical diagnosis and classification of β‐aCDs.


This work was supported by Medigenes. Further support by the LG Chem Chair Professorship (SYL) is appreciated.


ACD - Avellino corneal dystrophy

β‐aCD - βigh3‐associated corneal dystrophy

CD - corneal dystrophy

FITC - fluorescein isothiocyanate

GCD - granular CD

LCD I - lattice CD type I

Ori - origin of replication

RBCD - Reis–Bucklers CD

SSPE - saline/sodium phosphate/EDTA


Competing interests: None.


1. Afshari N A, Mullally J E, Afshari M A. et al Survey of patients with granular, lattice, avellino, and Reis‐Bucklers corneal dystrophies for mutations in the BIGH3 and gelsolin genes. Arch Ophthalmol 2001. 11916–22.22 [PubMed]
2. Chau H M, Ha N T, Cung L X. et al H626R and R124C mutations of the TGFBI (BIGH3) gene caused lattice corneal dystrophy in Vietnmese people. Br J Ophthalmol 2003. 87686–689.689 [PMC free article] [PubMed]
3. Ellies P, Renard G, Valleix S. et al Clinical outcome of eight BIGH3‐linked corneal dystrophies. Ophthalmology 2002. 109793–797.797 [PubMed]
4. El‐Ashry M F, Abd El‐Aziz M M, Ficker L A. et al BIGH3 mutation in a Bangladeshi family with a variable phenotype of LCDI. Eye 2004. 18723–728.728 [PubMed]
5. Konishi M, Yamada M, Nakamura Y. et al Varied appearance of cornea of patients with corneal dystrophy associated with R124H mutation in the BIGH3 gene. Cornea 1999. 18424–429.429 [PubMed]
6. Yoshida S, Yoshida A, Nakao S. et al Lattice corneal dystrophy type I without typical lattice lines: role of mutational analysis. Am J Ophthalmol 2004. 137586–588.588 [PubMed]
7. Mitterer G, Huber M, Leidinger E. et al Microarray‐based identification of bacteria in clinical samples by solid‐phase PCR amplification of 23S ribosomal DNA sequences. J Clin Microbiol 2004. 421048–1057.1057 [PMC free article] [PubMed]
8. Aldave A F, Gutmark J G, Yellore V S. et al Lattice corneal dystrophy associated with the Ala546Asp and Pro551Gln missense changes in the TGFBI gene. Am J Ophthalmol 2004. 138772–781.781 [PubMed]
9. Hirano K, Hotta Y, Fujiki K. et al Corneal amyloidosis caused by Leu518Pro mutation of betaig‐h3 gene. Br J Ophthalmol 2000. 84583–585.585 [PMC free article] [PubMed]
10. Hirano K, Hotta Y, Nakamura M. et al Late‐onset form of lattice corneal dystrophy caused by leu527Arg mutation of the TGFBI gene. Cornea 2001. 20525–529.529 [PubMed]
11. Warren J F, Abbott R L, Yoon M K. et al A new mutation (Leu569Arg) within exon 13 of the TGFBI (BIGH3) gene causes lattice corneal dystrophy type I. Am J Ophthalmol 2003. 136872–878.878 [PubMed]

Articles from The British Journal of Ophthalmology are provided here courtesy of BMJ Publishing Group