Unusual Phenotype of an Individual With the R124C Mutation in the TGFBI Gene

We describe here an unusual phenotype associated with the arginine-124→cysteine(R124C) mutation of the TGFBI gene. The proband wasa 44-year-old woman who exhibited translucent lattice lines and circular centralepithelial opacity in both corneas and who had a typical history of recurrentcorneal erosions. She and members of her family harbored the R124C mutationof TGFBI, which has been implicated as a geneticchange responsible for lattice corneal dystrophy (LCD) type I. We performedkeratoplasty on her left eye and examined the removed tissue histopathologically.Light microscopy revealed deposition of mucopolysaccharide and disruptionof the Bowman layer, but no Congo red–positive component was detected.Furthermore, apple-green dichroism was not apparent with polarized microscopy.Electron microscopy revealed numerous vacuoles within keratocytes as wellas separated collagen fibrils, but no amyloid fibrils, in the stroma. TheR124C mutation of TGFBI is thus not necessarily associatedwith amyloid deposition.

Point mutations of the TGFBI gene (also knownas βig-h3), which encodes keratoepithelin,¹ have been shown to be responsible for corneal dystrophies.Mutation of codon 124 of TGFBI from arginine to cysteine(R124C), histidine (R124H),² or leucine(R124L)³ has thus been associated with typeI, Avellino dystrophy, and superficial granular corneal dystrophy, respectively.Although the mutation site is the same, it is possible to differentiate theR124C, R124H, and R124L mutations both clinically with a slitlamp microscopeand pathologically.

Keratoepithelin is present in several human tissues, including the cornea.It forms fibrils and interacts with various extracellular matrix proteins,including fibronectin and collagen;⁴ theseproperties are thought to be related to the ability of the mutant proteinto form amyloid deposits in corneal dystrophies. Keratoepithelin also playsa role in the adhesion and migration of dermal fibroblasts.⁵

As far as we are aware, all reported cases of the R124C mutation of TGFBI have been diagnosed clinically as LCD type I, withdeposition of amyloid in the cornea confirmed pathologically. We now presentan unusual case of an individual with the R124C mutation of TGFBI without histopathologic evidence of amyloid deposition.

A 44-year-old woman was referred to our hospital in October 1998 withcomplaints of pain in her left eye. In childhood, she had been diagnosed byher local ophthalmologist as having corneal dystrophy, and she had developedrecurrent corneal erosions several times in both eyes that had also been treatedby her local specialist. Medical information from her local ophthalmologistrevealed that she had taken steroid eyedrops because the recurrent erosionwas severe and inflammation persisted. At 20 years of age, she underwent trabeculotomyin her left eye as a result of steroid-induced glaucoma. She was referredto our cornea service because of an unusual pattern of healing of epithelialerosion. At the first visit, her visual acuity was 20/40 OD and 20/100 OS,and her intraocular pressure was 10 mm Hg OD and 15 mm Hg OS. Slitlamp examinationrevealed corneal epithelial and subepithelial opacity as well as fine translucentopacity in the stroma characterized by lattice lines in both eyes (Figure 1). On the basis of the slitlamp examinationand her past history, we diagnosed the patient as having LCD. In January 1999,we performed penetrating keratoplasty on her left eye because of a decreasein left visual acuity to 20/600. No episodes of rejection or recurrent erosionhave been apparent since the surgery, and the graft has remained clear (Figure 1 The visual acuity of her left eyeis currently 20/15 with correction.

We obtained written informed consent for genetic analysis of TGFBI from the proband and 3 members of her family: her father, firstson, and husband. This analysis was approved by the ethical committee of YamaguchiUniversity Hospital. We detected a heterozygous R124C mutation of TGFBI in the proband, her father, and her first son (Figure 2). We did not detect a mutation of TGFBI in the husband. We were not able to perform genetic analysis on othermembers of her family. Slitlamp examinations of her father, 1 of her 2 brothers,and her 2 sons were performed. The father, who died several years ago, hadbeen diagnosed with LCD and underwent corneal transplantation in both eyes2 decades ago at another hospital; no medical records of his surgeries wereavailable to us. The slitlamp photographs reveal, however, that the graftsin both his eyes were clear and not edematous (Figure 3A). No translucent lines or other findings suggestive ofcorneal dystrophy were apparent in his own peripheral cornea tissue. The corneasof the proband's brother, however, exhibited typical translucent lines andcentral superficial opacity (Figure 3B).He had also experienced several episodes of recurrent erosion. We were notable to obtain informed consent for genetic analysis of the brother to confirmthe presence of the mutation. The corneas of both the proband's sons exhibitedtranslucent lines, but central superficial opacity was not apparent (Figure 3C and D). On the basis of the clinicalfindings for this family, we conclude that the corneal abnormality is inheritedin an autosomal dominant manner. On the basis of the clinical findings andgenetic analysis, we diagnosed the proband as having LCD type I.

Histopathologic examination by hematoxylin-eosin staining and lightmicroscopy of the corneal specimen of the proband obtained during surgeryrevealed that the layered structure of the corneal epithelium was irregular,with prominent disarrangement of the basal cell layer (Figure 4A). The Bowman layer was defective in the central regionof the cornea. In the same area, small numbers of infiltrated cells were presentin the anterior portion of the stroma. Collagen fibers in the anterior stromaof the central cornea were separated, possibly as an artifact of fixation.We did not detect eosin-positive amorphous components. The Descemet membraneand endothelial cells appeared intact without associated pathologic features.Congo red staining did not reveal any positive components in the corneal stroma(Figure 4B), and apple-green dichroismin the cornea was not detected by polarized light microscopy (Figure 4C). Mucopolysaccharide deposition was apparent in the anteriorportion of the stroma but was restricted to the site of the defect in theBowman layer, as revealed by staining with toluidine blue (pH 7.1) (Figure 4D). Electron microscopy revealedthe presence of separated collagen fibrils in the stroma as well as numerousvacuoles indicative of lipid deposition in keratocytes; again, these vacuoleswere apparent only at the site of the defect in the Bowman layer (Figure 4E and F). However, 8-nm to 10-nmnonbranching fibrils were not observed by electron microscopy. These pathologicalfindings were thus not indicative of amyloid deposition. Indeed, the diagnosisprovided by our pathologist (M.T.) was not amyloidosis of the cornea but healedcorneal ulcer.

We have presented a case of an unusual phenotype associated with theR124C mutation of TGFBI. Although the proband exhibitedclinical characteristics similar to those of LCD, amyloid deposition in thecorneal stroma was not detected histopathologically. The definition of LCDtype I is based on the presence of translucent lines and of amyloid, as revealedby Congo red staining or apple-green dichroism, in the anterior corneal stroma.The R124C mutation of TGFBI has previously been associatedwith LCD type I.² The genetic, clinical,and pathological findings for the patient described here indicate that thephenotype associated with the R124C mutation may be varied.

The proband harbored a heterozygous R124C mutation in TGFBI. Her father and eldest son also harbored the same mutation. Furthermore,1 of her brothers showed translucent lines and central epithelial opacityon slitlamp examination, findings typical of LCD type I. On the basis of theslitlamp examination of the proband and the pattern of inheritance of thiscondition in her family, we initially diagnosed her as having LCD type I.However, analysis of the corneal specimen obtained during surgery failed todetect amyloid deposition, precluding a diagnosis of LCD type I. The probandthus manifested an unusual phenotype associated with the R124C mutation of TGFBI.

The clinical characteristics of the patient included autosomal dominantinheritance, glasslike linear stromal opacity, epithelial-subepithelial opacityat the center of the cornea, and a past history of recurrent corneal erosions.On the basis of our clinical observations, we attempted to differentiate thiscase from gelatinous droplike dystrophy, Avellino dystrophy, and corneal leukomaassociated with infectious disease. In the proband, the surface of the cornealepithelium was relatively smooth, and lattice lines were apparent in the stroma.In contrast, in individuals with gelatinous droplike dystrophy, the surfaceof the corneal epithelium is irregular and bumpy and exhibits protrusions.In patients with Avellino dystrophy, the corneal opacities are isolated, eachbeing circular and 1 to 2 mm in diameter. In contrast, in the present case,the corneal opacity was diffuse. The proband had no memory of a previous seriouscorneal infection such as herpes, measles, or treponemiasis. Our clinicalfindings thus excluded these 3 alternative diagnoses.

It is not known why the R124C mutation of TGFBI inthe proband did not cause LCD type I. It is possible that an inhibitory mechanismprevented amyloid accumulation in the corneal stroma or that a mechanism foramyloid degradation was operative. Although it is thought that amyloid depositsin tissues do not disappear after they have formed,⁶ mouseastrocytes were recently shown to degrade amyloid-β.⁷

In conclusion, we have described a case of an unusual phenotype associatedwith the R124C mutation of TGFBI. The proband exhibitedclinical characteristics of LCD type I but showed no amyloid deposition. Theproband also did not appear to suffer from a different type of corneal dystrophy.Although molecular genetics has demonstrated the relation between specificgene mutations and hereditary diseases, extensive clinical and pathologicalvalidation is still required to verify that a particular mutation is reallyresponsible for disease pathogenesis.

The authors have no relevant financial interest in this article.

Correspondence: Dr Morishige, Department of Biomolecular Recognitionand Ophthalmology, Yamaguchi University School of Medicine, 1-1-1 Minami-Kogushi,Ube, Yamaguchi 755-8505, Japan (morishig@yamaguchi-u.ac.jp).