KLEIP Deficiency in Mice Causes Progressive Corneal Neovascular Dystrophy

Nicole Hahn,1,2 Christian T. Dietz,1 Sandra Kuhl,¨ 1 Urs Vossmerbaeumer,3 and Jens Kroll1,2

PURPOSE. The BTB-kelch KLEIP/KLHL20 is an linked to corneal dystrophy formation. For example, genomic binding protein that regulates cell-cell contact formation and mutations in the -3 and keratin-12 are linked to an cell migration. The aim of our study was to characterize epithelial dystrophy phenotype, called Meesmann dystrophy,4 KLEIP’s function in ocular health and disease in mice. and mutations in the transforming growth factor-b- induced (TGFBI) gene can cause corneal dystrophies, such as METHODS. KLEIP-/- mice were generated, and were Reis-Bucklers¨ dystrophy, Thiel-Behnke dystrophy, and granular examined histologically and stained for keratin-1, loricrin, type 1-, granular type 2-, and lattice type 1-dystrophy in keratin-12, keratin-14, CD31, LYVE-1, F4/80, E-cadherin, and humans.5 Ki67. Corneal abrasions were performed after opening. Genetic animal models to study onset, progression, and RESULTS. Corneas of KLEIPþ/þ and KLEIP-/- mice were therapeutic intervention of corneal dystrophies currently are indistinguishable at birth. After eyelid opening corneal very limited. For example, keratin-12, a protein that forms epithelial hyperplasia started to manifest in KLEIP-/- mice, intermediate filaments in epithelial cells, is expressed specif- showing a progressive epithelial metaplasia leading to total ically in the corneal epithelium in mice.6 Its gene silencing . In KLEIP-/- mice the initial stratified resulted in a fragile epithelium serving as a mouse model for squamous corneal epithelium was altered to an epidermal Meesmann corneal dystrophy.7 In addition to the , histo-architecture showing several superficial keratinized cells, genetic alterations in different proteoglycans can induce cell infiltrations into the stroma, and several apoptotic cells. corneal malformations in mice. For example deletion of Skin markers and loricrin were positive, and surface keratocan, a cornea-specific keratan sulphate proteoglycan, disease was accompanied by deep stromal vascularization. resulted in a thinner corneal stroma8 and lumican, which gene Expression analysis for E-cadherin in KLEIP-/- corneas showed inactivation in mice showed a cloudy and thin corneal stroma.9 acellular areas in the squamous epithelium, indicating a Another class of molecules that can cause corneal dystrophies progressive fragile corneal integrity. Removal of the virgin in mice are transcription factors. Corneal-specific overexpres- epithelium accelerated strongly development of the epithelial sion of the transcription factor Pax6 induced an abnormal and stromal alterations, identifying mechanical injuries as the cornea with altered epithelial cell morphology and neovascu- major trigger for corneal dystrophy formation and scarification larization;10 inactivation of zinc finger transcription factor in KLEIP-/- mice. Zeb1 in mice correlated with the posterior polymorphous corneal dystrophy,11 and tissue-specific deletion of transcrip- CONCLUSIONS. The data identify KLEIP as an important molecule tion factor Pbx1 in the corneal epithelium resulted in a corneal regulating corneal epithelial integrity. (Invest Ophthalmol Vis dystrophy and clouding.12 Moreover, genetic inactivation of Sci. 2012;53:3260–3268) DOI:10.1167/iovs.12-9676 the transcriptions factors AP-2alpha, Klf4, Klf5, and Cited2 also induced corneal defects in mice.13–16 Together, these mouse models identified few genes that can cause corneal alterations rogredient opacity of the cornea due to stromal reorgani- in mice, but it is unclear mostly whether these genes can cause Pzation, keratinization of the corneal epithelium and 17 1–3 corneal dystrophies in patients as well. Moreover, based on neovascularization, is a major reason for blindness in man. the diversity of corneal dystrophies in patients, several In humans, only few genes have been identified so far that are potential molecular regulators for corneal dystrophy formation still are not yet identified. The BTB-kelch protein KLEIP (Kelch-like ECT2 interacting From the 1Department of Vascular Biology and Tumor Angio- protein), also named KLHL20, has been identified first in genesis, Center for Biomedicine and Medical Technology Mannheim MDCK cells where it co-localizes transiently with F-actin during (CBTM), Medical Faculty Mannheim, Heidelberg University, Man- 18 2 the process of cell-cell contact induction. Recruitment of nheim, Germany; the Division of Vascular Oncology and Metastasis, KLEIP to cell adhesion sites depends on Rac1 activation and German Cancer Research Center (DKFZ-ZMBH Alliance), Heidel- 18 berg, Germany, and the 3Department of , Mainz requires E-cadherin. In addition, KLEIP is induced under 19 University Medical Center, Mainz, Germany. hypoxic conditions that can stimulate RhoA signaling during 20,21 Supported by the Deutsche Forschungsgemeinschaft (KR1887/ endothelial cell migration and in neurite outgrowth. 4-3, KR1887/5-1, and INST 91027/10-1 FUGG). Recent findings also identified KLEIP as a molecule acting Submitted for publication February 10, 2012; revised March 28, together with the wnt/beta- signaling pathway, which 2012; accepted April 9, 2012. may regulate skin thickness.22 In summary, the data highlight Disclosure: N. Hahn,None;C.T. Dietz,None;S. K¨uhl,None; KLEIP as an important molecule in cell-cell contact formation, U. Vossmerbaeumer,None;J. Kroll,None regulation of the cellular architecture, and cellular reorganiza- Corresponding author: Jens Kroll, Center for Biomedicine and tion, and identified KLEIP as a factor regulating cell migration. Medical Technology Mannheim (CBTM), Dept. of Vascular Biology and Tumor Angiogenesis, Medical Faculty Mannheim, Heidelberg Yet, its function in vivo still is unknown. University, Germany, Ludolf-Krehl-Str. 13-17, 68167 Mannheim, To address KLEIP’s function in vivo, we have generated Germany; Telephone þ49-(0)621-383-9965; Fax þ49-(0)621-383- KLEIP knockout mice, and identified KLEIP as an essential 9961; [email protected]. component for the maintenance of corneal integrity. KLEIP-/-

Investigative Ophthalmology & Visual Science, May 2012, Vol. 53, No. 6 3260 Copyright 2012 The Association for Research in Vision and Ophthalmology, Inc.

Downloaded from iovs.arvojournals.org on 09/30/2021 IOVS, May 2012, Vol. 53, No. 6 KLEIP Deficiency in Corneal Neovascular Dystrophy 3261

mice developed progressively a corneal dystrophy due to epithelial fragility, which was induced mainly by mechanical injuries of the cornea. Therefore, KLEIP-deficient mice represent a unique genetic model to study onset and progression of corneal damages in mice.

METHODS Generation and Genotyping of KLEIP-/- Mice The 129SvEv embryonic stem cell clone XF202 (BayGenomics), carrying the b-geo (beta galactosidases and neomycin resistance) encoding vector pGT2Lxf in the KLEIP locus, was injected into C57/BL6 blastocytes, crossed, and maintained in C57/BL6 Ola mice. Currently, mice are in the F10 generation and named as B6.129-KLEIPtm/Mhm.For genotyping, three primers were used, namely primer S1, which binds in intron 2 of genomic KLEIP; primer A1, which binds in the gene trap vector, and primer A2, which binds in exon 3 of KLEIP. Primer pair S1 and A2 generated the wild type band of 1349 (bp) length, while primer pair S1 and A1 generated the transgenic signal of 591 bp length. Expression of lacZ was analyzed using primer pair L1 and L2 (Supplemental Fig. 1, http://www.iovs.org/lookup/suppl/doi:10.1167/ iovs.12-9676/-/DCSupplemental). For primer sequences see below. Mice were kept under specific pathogen-free conditions according to the animal facility regulations of the Medical Faculty Mannheim, Heidelberg University. All animal experiments were approved by the Regierungspr¨asidium Karlsruhe (protocol numbers 35-9185.83, 35- 9185.81/G-82/11, and I-07/03), and are in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

Antibodies and Histologic Reagents The following antibodies were used for the study: anti-mouse CD31, clone MEC 13.3 (BD Pharmingen, Heidelberg, Germany); anti-mouse Endomucin, clone V.1A7, and anti-mouse keratin-12 (Santa Cruz Biotechnology, Heidelberg, Germany); anti-mouse E-cadherin, clone ECCD-2 (Invitrogen, Darmstadt, Germany); anti-mouse LYVE-1 (RELIA- Tech, Wolfenbuttel,¨ Germany); anti-mouse F4/80, clone CI: A3-1, and donkey-anti-rabbit FITC (Dianova, Hamburg, Germany); anti-mouse keratin-1, anti-mouse keratin-14, and anti-mouse loricrin (Covance, Munchen,¨ Germany); anti-human Ki67 (Novocastra, Wetzlar, Ger- many); goat-anti-rat Alexa 546 and goat-anti-rabbit Alexa 488 (Molecular Probes, Darmstadt, Germany); HRP-conjugated antibodies and ABC-kit (VECTOR Laboratories, Dossenheim, Germany); and DAPI, HE staining solution, Sudan black B solution, Trichrome Stain (Masson) Kit, and anti-human SMA, clone 1A4 (Sigma-Aldrich, Munchen,¨ Germany).

Histochemistry Eyes from KLEIPþ/þ and KLEIP-/- mice were enucleated, and corneas were isolated via cutting around the corneoscleral limbal ring. -/- were dissected using a scalpel. Tissues were fixed in 4% formaldehyde FIGURE 1. KLEIP mice developed a corneal dystrophy. (A) Corneal dystrophy phenotype in a KLEIP-/- mouse as indicated by a corneal (PFA) or in Zn-fixative overnight at 48C, or immediately frozen in liquid opacity. A KLEIPþ/þ litter mate served as healthy control. (B) nitrogen, embedded in Tissue TEC (ornithine carbamoyltransferase Incidence of corneal dystrophy in KLEIP-/- mice. KLEIP-/- mice [OCT]-medium) or dehydrated after PFA/Zn-fixation using ethanol- started to manifest a corneal dystrophy three weeks after birth. At 16 series (70%, 80%, 90%, and 100% ethanol, and 100% and 100% xylol, 5 weeks after birth, 90% of KLEIP-/- mice had a severe corneal minutes each), and embedded in paraffin. Both cryo and paraffin dystrophy. KLEIPþ/þ mice did not have a corneal dystrophy (n ¼ 15 sections were cut to a size of 6-8 lm each. To analyze tissue structure per group). (C) Top: Biopsies of KLEIPþ/þ and KLEIP-/- corneal of corneas and glands, mouse tissues were stained using Mayer’s epithelial cells directly after abrasion showed expression of lacZ as an hematoxylin & eosin (HE), trichrome stain (keratin and collagen indicator for KLEIP expression. Bottom: RT-PCR analysis in mouse staining), and Sudan black (lipid staining). Apoptotic corneal cells were corneas (n ¼ 9 pooled samples for each genotype) for KLEIP and lacZ expression. TBP served as a loading control. KLEIP-/- : KLEIP-/- labeled using the TUNEL staining kit from Chemicon (ApopTag Red In dys mice with corneal dystrophy. Scale bar: 50 lm. Situ Apoptosis Detection Kit; Chemicon, Darmstadt, Germany).

Isolation of mRNA, RT-PCR, and Primer Sequences Eyes from KLEIPþ/þ and KLEIP-/- mice were enucleated, and corneas were removed and mechanically homogenized. Total RNA was isolated

Downloaded from iovs.arvojournals.org on 09/30/2021 3262 Hahn et al. IOVS, May 2012, Vol. 53, No. 6

FIGURE 2. Progression of corneal dystrophy formation in KLEIP-/- mice. Progression of corneal dystrophy formation was monitored during the first four months of life in five KLEIPþ/þ and five KLEIP-/- mice. Macroscopic (top) and histologic (bottom) disease progression is shown in one eye of a KLEIP-/- mouse (right) in comparison to a KLEIPþ/þ mouse (left) for each time point. During the whole observation period eyes in KLEIPþ/þ mice remained healthy (left). KLEIP-/- deficient mice were born with normal functional eyelids and corneas, as can be seen after eyelid opening (week 3). During the following weeks KLEIP-/- mice had a corneal plaque (as of week 8), which was vascularized at weeks 10 to 12 (top). HE stainings of corneal sections from KLEIP-/- mice showed morphologic alterations during disease progression (bottom), cell infiltrations into the stroma (weeks 8 to 10), diffuse progressive epithelial metaplasia, and indentations into the corneal stroma (weeks 10 to 12) and corneal neovascularization (weeks 10 to 12). Highlighted areas within section ‘‘12 weeks’’ correspond to the higher magnification (right) showing an epithelial hyperplasia (asterisk), cell infiltrations (point), and neovascularization (arrows). Scale bar: 100 lm.

using the RNeasy Mini Kit, (Qiagen, Hilden, Germany) followed by RT- Hamburg, Germany), 2.1 mg/mL K-ferrOcyanide, and 1.6 mg/mL K- PCR using the Superscript II RT Kit (Promega, Mannheim, Germany). ferricyanide. Then, 16 hours later tissues were washed 3 times in PBS, PCRs were performed under the following conditions: 948Cfor3minutes, and postfixed in 2% PFA/0.1% glutaraldehyde/PBS overnight at 48C. 35· 948C for 45 seconds, 588C for 30 seconds, 728C for 1:30 minutes, and a final extension at 728C for 10 minutes. The following primers were used: Abrasions TBP: forward 50 GGA CCA GAA CAA CAG CCT TCC; reverse 50 Abrasions were performed directly after eyelid opening to monitor CAT GAT GAC TGC AGC AAA TCG epithelial wound healing capacity in KLEIPþ/þ and KLEIP-/- mice. Mice lacZ: forward (L1) 50 TAT CGA TGA GCG TGG TGG TTA TGC C; were anesthetized using isoflurane (Baxter, Unterschleißheim, Ger- reverse (L2) 50 GCG CGT ACA TCG GGC AAA TAA TAT C many) and, for postoperative analgesia, treated mice received 200 mg/ KLEIP: forward 50 GTG ATG GCC TGG GTC AAA TAC; reverse 50 kg body weight Metamizol (Ratiopharm, Ulm, Germany). For wound- GAG GAT CCA TCA TGG CCG CCT AC ing, the corneal epithelium was moistened via short-time incubation Primer for genotyping: forward (S1) 50 CAA GTG CGA TTG AAG using a 2.5% ethanol/water solution. Afterwards, approximately 40% of CAT CC the central corneal epithelium was removed via a hockey spatula. Primer for genotyping: reverse (A1) 50 ACC TGG CTC CTA TGG Cornea epithelial wounding was monitored using Thilorbin (Alcon, GAT AG Freiburg, Germany), and dystrophy development was assessed after 1, Primer for genotyping: reverse (A2) 50 AAA CAT TGC TCG GAA 6, and 12 hours, and 1, 3, 7, and 14 days after abrasion. GTA GG

RESULTS lacZ Staining of Transgenic Mice Generation of KLEIP-Deficient Mice LacZ staining was performed as described.23 In brief, tissues were fixed in LacZ-fix solution containing 0.2% glutaraldehyde at 48C overnight, To dissect the function of the BTB-kelch protein KLEIP/ washed 3 times in LacZ wash buffer, and incubated overnight at 378C KLHL20 during vertebrate development, the KLEIP gene was in lacZ stain containing 1 mg/mL X-Gal (BIOMOL prod.-No. 02249, disrupted in mice. The genetic modified embryonic stem cell

Downloaded from iovs.arvojournals.org on 09/30/2021 IOVS, May 2012, Vol. 53, No. 6 KLEIP Deficiency in Corneal Neovascular Dystrophy 3263

FIGURE 3. Corneas of KLEIP-/- mice contained several keratinized and proliferating cells, and expressed skin-specific keratins. (A) Masson trichrome staining in corneas of KLEIP-/- and KLEIPþ/þ mice. In comparison to corneas of KLEIPþ/þ mice, corneas of KLEIP-/- mice showed a thickened stroma (collagen green), a thickened and keratinized epithelium (red) and cell infiltrations into the stroma (nuclei dark brown/black). (B) Corneas in KLEIP-/- mice expressed skin-specific markers keratin-1, loricrin, and keratin-14 (brown), whereas corneas of KLEIPþ/þ mice expressed keratin-14 only. (C) Expression of cornea-specific keratin-12 (green) in KLEIP-/- corneas was reduced strongly as compared to KLEIPþ/þ corneas. (D) In contrast to KLEIPþ/þ mice, corneas of KLEIP-/- mice contained proliferating cells within the epithelium and stroma (Ki67 brown). Tissue sections were counterstained with HE. Scale bars: 100 lm.

line XF202, targeting the KLEIP gene, was used that was KLEIP Deficiency in Mice Progressively Resulted in generated by the gene trap technology (Supplemental Fig. 1A, a Corneal Dystrophy, Which Is Associated with an B, C, D, http://www.iovs.org/lookup/suppl/doi:10.1167/iovs. Epithelial Hyperplasia and an Altered Corneal 12-9676/-/DCSupplemental). XF202 cells were injected into blastocysts and transferred into pseudo-pregnant foster moth- Epithelial Cell Differentiation ers. Born chimeras were crossed into C57/Bl6 Ola mice, and a KLEIP-/- mice displayed, starting earliest in the third postnatal genotyping protocol was established (Supplemental Fig. 1E, week, a whitish corneal opacification on both eyes that was http://www.iovs.org/lookup/suppl/doi:10.1167/iovs.12-9676/ macroscopically visible (Fig. 1A). The onset of this pathology -/DCSupplemental). KLEIP-deficient embryos and mice initially varied between weeks 3 and 16, when 90% of the surviving were analyzed for KLEIP expression by RT-PCR analysis, which animals had a corneal opacification (Fig. 1B). Expression showed the expected loss of KLEIP expression (Supplemental analysis in KLEIPþ/þ and KLEIP-/- eyes confirmed expression Fig. 1F, http://www.iovs.org/lookup/suppl/doi:10.1167/iovs. of KLEIP or lacZ as a marker for KLEIP within the corneal 12-9676/-/DCSupplemental). To analyze the effect of KLEIP epithelium, suggesting a direct function of KLEIP in the mouse knock down on viability and fertility in mice, the ratio of cornea (Fig. 1C). In addition, weak expression of lacZ also was KLEIPþ/þ, KLEIPþ/-, and KLEIP-/- mice was determined. observed in the and in the ciliary marginal zone (not Interestingly, we observed a reduced number for KLEIP-/- shown). mice at P28 (wild-type mice versus heterozygous mice versus To characterize in detail onset and progression of corneal homozygous mice: 33% versus 54% versus 13%), showing that dystrophy in KLEIP-/- mice, corneas of KLEIPþ/þ and KLEIP-/- approximately 50% of KLEIP-/- mice die until day 28 of mice were analyzed histologically at different time points. postnatal development. Yet, reasons for this reduced viability While KLEIPþ/þ mice showed a normal corneal anatomical of KLEIP-/- mice presently are unclear. KLEIP-/- mice that structure consisting of the epithelium, stroma, Descemet’s survived beyond P28 were fertile. membrane, and the endothelium, all KLEIP-/- mice had a

Downloaded from iovs.arvojournals.org on 09/30/2021 3264 Hahn et al. IOVS, May 2012, Vol. 53, No. 6

FIGURE 4. KLEIP deficiency in mice induced corneal neovascularization. (A) Brightfield image (top) and whole mount endomucin staining (middle and bottom) of KLEIP-/- corneas showed newly formed blood vessels. Corneas of KLEIPþ/þ mice remained avascular. Scale bars: 500 lm. (B) Corneal sections of KLEIP-/- mice identified blood vessels and lymphatic vessels (CD31 red, LYVE-1 green, DAPI blue) predominantly in the corneal stroma. Corneas of KLEIPþ/þ mice were negative for both markers. (C) Top: Blood vessel maturation in KLEIP-/- corneas as indicated by CD31 staining (green) and alpha SMA (red) staining. Bottom: KLEIP deficiency resulted in corneal infiltrations of macrophages (F4/80 red), while macrophages were not detectable in KLEIPþ/þ corneas. Nuclei were stained with DAPI (blue). Scale bars: 100 lm.

progressive epithelial metaplasia leading to a corneal opacity of these glands finally can lead to corneal damages24 as observed (Fig. 2). Histologic analysis of the altered corneas in KLEIP-/- in KLEIP-/- mice. However, examination of the meibomian mice performed at different time points revealed a reproducible glands, goblet cells, and lacrimal glands did not reveal major sequence. Initially, we observed epithelial hyperplasia. This was structural differences in KLEIPþ/þ and KLEIP-/- mice (Supple- followed by metaplasia with a reorganization of the original mental Fig. 2, http://www.iovs.org/lookup/suppl/doi:10.1167/ non-keratinized stratified squamous epithelium towards a iovs.12-9676/-/DCSupplemental) suggesting functional tear keratinized cell population on the surface (Figs. 2, 3A–C). This layer-producing tissues. Furthermore, we addressed the ques- was accompanied by focal epithelial indentations into the tion whether KLEIP deletion in mice may affect embryonic eye corneal stroma, and a massive thickening of the epithelial layer development and whether other structures in the eye in (Fig. 2). The final stage of the disease resulted in a situation in KLEIP-/- mice are altered by the KLEIP deletion as well. To this which the tissue displayed an epidermal organization pattern end we assessed macroscopically and histologically KLEIP-/- with a keratinized superficial layer (Figs. 2, 3A–C). Immuno- embryos and newborns ranging from age E11, E14, E18.5, and staining of the corneal epithelium in KLEIP-/- mice revealed P0 for eye development. However, we could not observe ob- Ki67-positive, proliferating cells within the epithelium and vious histologic and anatomical differences between KLEIP-/- stroma (Fig. 3D), and altered expression of skin- and cornea- and KLEIPþ/þ eyes (Supplemental Fig. 3, http://www.iovs.org/ specific keratins.24 While KLEIPþ/þ corneas expressed only lookup/suppl/doi:10.1167/iovs.12-9676/-/DCSupplemental). keratin-12 and keratin-14, KLEIP-deficient corneas expressed In addition, besides corneal opacity in juvenile and adult strongly skin-specific keratin-1, loricrin, and keratin-14, but only KLEIP-/- mice, eyes in KLEIP-/- did not have obvious weakly keratin-12, indicating an altered differentiation in KLEIP- pathologic alterations in other structures of the eye: retina, deficient corneas from a cornea-like into a skin-like phenotype , , , and were not affected by the KLEIP (Figs. 3B, C). deletion (data not shown). Together, the data indicate that KLEIP deletion in mice leads to a progressive corneal dystrophy. Periocular Lacrimal Glands and Embryonic Development of the Eye Remained Unaltered in Corneas in KLEIP-/- Mice Were Highly KLEIP-/- Mice Vascularized Next, we analyzed histologically the meibomian glands, goblet Healthy corneas are optically transparent and avascular due to cells, and lacrimal glands in KLEIP-/- mice, because dysfunction soluble VEGF receptor sFlt-1.25 In contrast to KLEIPþ/þ mice,

Downloaded from iovs.arvojournals.org on 09/30/2021 IOVS, May 2012, Vol. 53, No. 6 KLEIP Deficiency in Corneal Neovascular Dystrophy 3265

which did not show any blood vessels in the corneas, we observed in KLEIP-/- whole mount cornea samples macro- scopically a substantial formation of blood vessels that orig- inated typically in the limbal circumference, and could be visualized further by whole mount endothelial specific endomucin staining (Figs. 2, 4A). Cross-sections of KLEIP-/- corneas identified location of blood (CD31 staining) and lymphatic vessels (LYVE-1 staining) mostly in the corneal stroma (Fig. 4B). These blood vessels were covered by pericytes as indicated by alpha SMA staining (Fig. 4C). In addition to newly formed blood vessels in the corneal stroma of KLEIP-/- mice, we also observed several F4/80-positive cells, indicating infiltrating macrophages (Fig. 4C). The control group of KLEIPþ/þ mice, however, was completely negative for F4/80-positive cells. In conclusion, the data showed a strong induction of angiogenesis and lymphangiogenesis, and infiltra- tions of macrophages in KLEIP-/- corneas.

KLEIP Regulated Corneal Epithelial Integrity and Development of Corneal Diseases in KLEIP-/- Mice Was Intensified through Experimental Mechanical Injury KLEIP co-localizes transiently with E-cadherin in cultured MDCK cells, suggesting that it regulates formation of cell-cell contacts.18 Therefore, we hypothesized that E-cadherin localization could be altered in the epithelium of KLEIP- deficient corneas, and expression of E-cadherin in KLEIPþ/þ and KLEIP-/- mice was examined. In KLEIPþ/þ corneas, both epithelial cell layers, namely the squamous cell layer and the basal cell layer, were positive for E-cadherin, and both layers displayed the physiologic architecture (Fig. 5A, left). In contrast, squamous and basal cell layers in KLEIP-/- mice largely were disorganized and characterized by acellular areas in the squamous cell layer, suggesting loosened cell-cell contacts and an incomplete regeneration of the corneal epithelium in KLEIP-/- corneas (Fig. 5A, right). Furthermore, corneas of KLEIP-/- mice displayed several apoptotic cells, which were absent in KLEIPþ/þ corneas (Fig. 5B). This observation suggested that high proliferative activity and massive corneal epithelium formation in KLEIP-/- mice (Figs. 2, 3) is a compensatory mechanism due to corneal epithelial fragility, apoptosis, and damage. Next, we hypothesized that the corneal epithelium in KLEIP-/- mice makes corneas more prone to mechanical injury, and an experimental mechanical FIGURE 5. Corneal epithelium of KLEIP-/- mice showed disorganized, injury would induce rapidly a strong corneal dystrophy in acellular and apoptotic areas. (A) Corneas of KLEIPþ/þ mice showed KLEIP-/- mice. To test this hypothesis, abrasions were intact squamous and basal cell layers as indicated by E-cadherin -/- performed by removing mechanically 40% of the cornea in staining. Squamous and basal epithelial cell layers in KLEIP mice þ/þ -/- largely were disorganized, and the squamous cell layer showed several 18-day-old KLEIP and KLEIP mice, and corneal wound acellular areas. Scale bar: 100 lm. (B) Non-dystrophic KLEIP-/- þ/þ closure was analyzed (Fig. 6). In KLEIP mice a rapid corneas of 3-week-old mice showed several apoptotic cells (top, scale corneal regeneration occurred, and after three days injured bars: 500 lm). Apoptotic cells (red) were detected in the squamous corneas were indistinguishable from untreated eyes (Fig. 6A epithelial cell layer (S) but not in the basal layer (B) as indicated by top, B). In contrast, corneal abrasions in KLEIP-/- mice led TUNEL staining (bottom, scale bars: 100 lm). rapidly to corneal opacity within seven days in all treated animals (Fig. 6A bottom, B). Injured KLEIP-/- corneas ISCUSSION showed similar histologic alterations, such as epithelial D hyperplasia, stromal infiltrations, and superficial keratinized Our observations and data reveal that KLEIP-/- mice have a cells (Fig. 6C right) as in 15-week-old non-injured KLEIP-/- distinct phenotype of corneal opacification, epidermal meta- mice (Fig. 2). Likewise, analysis for E-cadherin expression in plasia of the epithelium, and stromal vascularization. This is KLEIP-/- corneas 12 hours after abrasion showed disorga- accompanied by lymphangiogenesis and inflammatory chang- nized and acellular areas in the squamous epithelial cell layer es. This exclusive observation of the described changes in (Fig. 6D), which were similar to the non-injured 15-week-old KLEIP-/- mice would make the phenotype eligible to be KLEIP-/- mice (Fig. 5). Together, the data indicate that KLEIP termed a corneal dystrophy. regulates corneal epithelial integrity, and loss of KLEIP Human corneal dystrophies can be divided into four expression makes corneas more fragile and sensitive to groups according to the affected anatomical structure in the mechanical injury. cornea, namely into epithelial, Bowman Layer, stromal, and

Downloaded from iovs.arvojournals.org on 09/30/2021 3266 Hahn et al. IOVS, May 2012, Vol. 53, No. 6

FIGURE 6. Corneal epithelial abrasion strongly accelerated corneal dystrophy development in KLEIP-/- mice. (A) Left: Brightfield images of mouse eyes directly after abrasions. A 40% removal of the corneal epithelium was visualized using Thilorbin. Middle: Corresponding untreated eyes served as controls for dystrophy development. Pictures were taken 14 days later. Right: Brightfield images of mechanical treated eyes 14 days after corneal abrasion. Injured eyes developed a severe corneal dystrophy in KLEIP-/- mice, but not in KLEIPþ/þ mice. Scale bar: 500 lm. (B) Quantification of data shown in (A), n ¼ 7 per group. All KLEIP-/- mice had a corneal dystrophy on the treated eye within 7 days after abrasion. Of those mice 14% also had a corneal dystrophy on the control eye, whereas KLEIPþ/þ mice had no dystrophy either in the untreated or in the treated eye. (C)HE(top) and Masson trichrome (bottom) staining of corneal sections after corneal abrasion. Thickening of the epithelium and stroma, cell infiltrations in the stroma, and a keratinized epithelium were induced strongly in corneas of KLEIP-/- mice, but not in corneas of KLEIPþ/þ mice. Scale bar: 100 lm. (D) E-cadherin expression in the squamous (S) and basal (B) corneal epithelial cell layers 12 hours after abrasions in 18-day-old KLEIP-/- and KLEIPþ/þ mice. KLEIP-/- corneas showed several acellular areas in the squamous epithelial cell layer 12 hours after abrasions, which were not present in KLEIPþ/þ corneas. Scale bar: 100 lm.

endothelial dystrophies.5 Corneal epithelial dystrophies, in KLEIP has been identified originally as a molecule regulating particular Meesmann corneal dystrophy, are caused partially transiently cell-cell contact sites via E-cadherin localization in in human by mutations in the keratin-3 and keratin-12 genes.5 MDCK cells,18 which implied that KLEIP also may regulate In mice, homozygous deletion of the keratin-12 gene led to a corneal integrity in epithelial cells via cell-cell contact -/- fragile and thinner corneal epithelium.7 Keratin-12 is a protein formation. Although corneas in newborn KLEIP mice belonging to the intermediate filaments and is expressed initially are normal, after eyelid opening an epithelial specifically in the corneal epithelium.6,26 In our study, mice hyperplasia developed progressively, leading to a corneal that harbored a homozygous deletion for the BTB-kelch KLEIP opacity. Experimental mechanical injury of the cornea strongly and rapidly induced formation of a corneal dystrophy within gene also showed a corneal dystrophy in the epithelium, but -/- few days, which clearly showed that mechanical injury is the corneal alterations were different from the keratin-12 -/- -/- 7 major trigger for corneal dystrophy development in KLEIP mice. While corneas in keratin-12 mice are thinner, mice. This is supported strongly by the observation that the -/- corneas in KLEIP mice experienced an epithelial hyper- corneal epithelium in 4-month-old KLEIP-/- mice showed a plasia, including keratinized cells, an altered expression of similar epithelial metaplasia and acellular epithelial areas as keratins, and they showed several stromal infiltrations. This compared to KLEIP-/- corneas 12 hours after abrasions. phenotype suggests a different function for KLEIP in Therefore, the data show that loss of KLEIP expression makes maintenance of the corneal integrity as compared to corneas more fragile and sensitive to mechanical injury. As a keratin-12. compensatory mechanism to the injuries, corneal epithelial

Downloaded from iovs.arvojournals.org on 09/30/2021 IOVS, May 2012, Vol. 53, No. 6 KLEIP Deficiency in Corneal Neovascular Dystrophy 3267

cells start to proliferate. Insufficient wound healing based on 8. Liu CY, Birk DE, Hassell JR, Kane B, Kao WW. Keratocan- loosened cell-cell contacts promotes a continuous epithelial deficient mice display alterations in corneal structure. J Biol cell proliferation and epithelial metaplasia, and finally a total Chem. 2003;278:21672–21677. corneal opacity in KLEIP-/- mice. Interestingly, this observa- 9. Saika S, Shiraishi A, Liu CY, et al. Role of lumican in the corneal tion resembles closely a spontaneous scarification with epithelium during wound healing. J Biol Chem. 2000;275: neovascularization as seen in connection with external triggers 2607–2612. in patients. For instance, chemical and thermal injury27 to a 10. Davis J, Piatigorsky J. Overexpression of Pax6 in mouse cornea healthy cornea leads to a similar loss of epithelial and stromal directly alters corneal epithelial cells: changes in immune clarity, and eventually neovascularization. function, vascularization, and differentiation. Invest Ophthal- Healthy corneas usually are avascular as a major prerequi- mol Vis Sci. 2011;52:4158–4168. site for corneal transparency, which is regulated mainly by the 11. Liu Y, Peng X, Tan J, Darling DS, Kaplan HJ, Dean DC. Zeb1 soluble VEGF receptor sFlt-1.25 sFlt-1 binds VEGF with high mutant mice as a model of posterior corneal dystrophy. Invest affinity, acts as a trap for VEGF and, therefore, prevents Ophthalmol Vis Sci. 2008;49:1843–1849. formation of new blood vessels driven by VEGF.28 A recent 12. Murphy MJ, Polok BK, Schorderet DF, Cleary ML. Essential role study also suggested an additional mechanism showing that for Pbx1 in corneal morphogenesis. Invest Ophthalmol Vis transcription factor FoxC1 regulates bioavailability of VEGF by Sci. 2010;51:795–803. modifying expression of metalloproteinases.29 In this study 13. Dwivedi DJ, Pontoriero GF, Ashery-Padan R, Sullivan S, KLEIP deficiency in mice resulted in a strong formation of new Williams T, West-Mays JA. Targeted deletion of AP-2alpha blood and lymphatic vessels in the corneal stroma. Although leads to disruption in corneal epithelial cell integrity and KLEIP is known to be involved in endothelial cell function and defects in the corneal stroma. Invest Ophthalmol Vis Sci. angiogenesis20, the data suggest that vascularization of corneas 2005;46:3623–3630. in KLEIP-/- mice is a secondary phenomenon. It could be due 14. Young RD, Swamynathan SK, Boote C, et al. Stromal edema in to high proliferative activity and scarification of the corneal klf4 conditional null mouse cornea is associated with altered epithelium. collagen fibril organization and reduced proteoglycans. Invest In summary, our data suggest a new important function for Ophthalmol Vis Sci. 2009;50:4155–4161. KLEIP as an essential regulator for corneal integrity in mice. 15. Kenchegowda D, Swamynathan S, Gupta D, Wan H, Whitsett J, The data establish KLEIP-/- mice as a unique mouse model to Swamynathan SK. Conditional disruption of mouse Klf5 study corneal dystrophy formation and scarification, and it is results in defective eyelids with malformed meibomian glands, tempting to speculate that patients suffering from corneal abnormal cornea and loss of conjunctival goblet cells. Dev epithelial dystrophies may show genetic alterations in the Biol. 2011;356:5–18. human KLEIP gene. Further research will be directed now at 16. Chen Y, Carlson EC, Chen ZY, et al. Conditional deletion of elucidating the role of KLEIP in physiologic regeneration of the Cited2 results in defective corneal epithelial morphogenesis corneal epithelium, and in the formation of cell-cell contacts. and maintenance. Dev Biol. 2009;334:243–252. Also, pharmacologic ways to support the apparently fragile 17. Lisch W, Seitz B. [New international classification of corneal equilibrium of the cornea in KLEIP-deficiency merits deeper dystrophies (CD)]. Ophthalmologe. 2011;108:883–896; quiz consideration. Clinically, it would be desirable if the sponta- 897. neous pathology in KLEIP knockout animals could be used to 18. Hara T, Ishida H, Raziuddin R, Dorkhom S, Kamijo K, Miki T. study mechanisms of progressive corneal scarring. Novel kelch-like protein, KLEIP, is involved in actin assembly at cell-cell contact sites of Madin-Darby canine kidney cells. Acknowledgments Mol Biol Cell. 2004;15:1172–1184. 19. Yuan WC, Lee YR, Huang SF, et al. A Cullin3-KLHL20 ubiquitin The authors would like to thank the Nikon Imaging Center of ligase-dependent pathway targets PML to potentiate HIF-1 Heidelberg University. signaling and prostate cancer progression. Cancer Cell. 2011; 20:214–228. References 20. Nacak TG, Alajati A, Leptien K, et al. The BTB-Kelch protein KLEIP controls endothelial migration and sprouting angiogen- 1. Weiss JS, Mller HU, Lisch W, et al. The IC3D classification esis. Circ Res. 2007;100:1155–1163. of the corneal dystrophies. Cornea. 2008;27(suppl 2):S1– 21. Lin MY, Lin YM, Kao TC, Chuang HH, Chen RH. PDZ-RhoGEF S83. ubiquitination by Cullin3-KLHL20 controls neurotrophin- 2. Klintworth GK. Corneal dystrophies. Orphanet J Rare Dis. induced neurite outgrowth. J Cell Biol. 2011;193:985–994. 2009;4:7. 22. Yamaguchi Y, Passeron T, Hoashi T, et al. Dickkopf 1 (DKK1) 3. Qazi Y, Wong G, Monson B, Stringham J, Ambati BK. Corneal regulates skin pigmentation and thickness by affecting Wnt/ transparency: genesis, maintenance and dysfunction. Brain beta-catenin signaling in keratinocytes. FASEB J. 2008;22: Res Bull. 2010;81:198–210. 1009–1020. 4. Irvine AD, Corden LD, Swensson O, et al. Mutations in cornea- 23. Lobe CG, Koop KE, Kreppner W, Lomeli H, Gertsenstein M, specific keratin K3 or K12 genes cause Meesmann’s corneal Nagy A. Z/AP, a double reporter for cre-mediated recombina- dystrophy. Nat Genet. 1997;16:184–187. tion. Dev Biol. 1999;208:281–292. 5. Aldave AJ, Sonmez B. Elucidating the molecular genetic basis 24. Vauclair S, Majo F, Durham AD, Ghyselinck NB, Barrandon Y, of the corneal dystrophies: are we there yet? Arch Ophthal- Radtke F. Corneal epithelial cell fate is maintained during mol. 2007;125:177–186. repair by Notch1 signaling via the regulation of vitamin A 6. Liu CY, Zhu G, Westerhausen-Larson A, et al. Cornea-specific metabolism. Dev Cell. 2007;13:242–253. expression of K12 keratin during mouse development. Curr 25. Ambati BK, Nozaki M, Singh N, et al. Corneal avascularity Eye Res. 1993;12:963–974. is due to soluble VEGF receptor-1. Nature. 2006;443:993– 7. Kao WW, Liu CY, Converse RL, et al. -deficient mice 997. have fragile corneal epithelia. Invest Ophthalmol Vis Sci. 26. Liu JJ, Kao WW, Wilson SE. Corneal epithelium-specific mouse 1996;37:2572–2584. keratin K12 promoter. Exp Eye Res. 1999;68:295–301.

Downloaded from iovs.arvojournals.org on 09/30/2021 3268 Hahn et al. IOVS, May 2012, Vol. 53, No. 6

27. Tsai RJ, Li LM, Chen JK. Reconstruction of damaged corneas receptor, FLT-1, and its heterodimerization with KDR. Bio- by transplantation of autologous limbal epithelial cells. N Engl chem Biophys Res Commun. 1996;226:324–328. J Med. 2000;343:86–93. 29. Seo S, Singh HP, Lacal PM, et al. Forkhead box transcription factor FoxC1 preserves corneal transparency by regulating 28. Kendall RL, Wang G, Thomas KA. Identification of a natural vascular growth. Proc Natl Acad Sci U S A. 2012;109:2015– soluble form of the vascular endothelial growth factor 2020.

Downloaded from iovs.arvojournals.org on 09/30/2021