Genetic deletion of amphiregulin restores the normal skin phenotype in a mouse model of the human skin disease tylosis

ABSTRACT In humans, gain-of-function (GOF) mutations in RHBDF2 cause the skin disease tylosis. We generated a mouse model of human tylosis and show that GOF mutations in RHBDF2 cause tylosis by enhancing the amount of amphiregulin (AREG) secretion. Furthermore, we show that genetic disruption of AREG ameliorates skin pathology in mice carrying the human tylosis disease mutation. Collectively, our data suggest that RHBDF2 plays a critical role in regulating EGFR signaling and its downstream events, including development of tylosis, by facilitating enhanced secretion of AREG. Thus, targeting AREG could have therapeutic benefit in the treatment of tylosis.


INTRODUCTION
The role of RHBDF2 in enhancing amphiregulin (AREG) secretion, and consequently activating the epidermal growth factor receptor (EGFR) pathway, has significance for the skin disease tylosis. In humans, autosomal dominant mutations in RHBDF2 cause tylosis (Blaydon et al., 2012;Saarinen et al., 2012), a form of focal nonepidermolytic palmoplantar keratoderma (PPK) that manifests on the skin of the palms and soles (Stevens et al., 1996). Tylosis patients also present oral leukokeratosis and follicular hyperkeratosis (Field et al., 1997;Hennies et al., 1995). Currently, there is no cure for tylosis. In a study to identify genetic variants underlying tylosis, Blaydon et al. used targeted capture array and next-generation sequencing and identified two heterozygous missense mutations in the RHBDF2 gene that underlie tylosis, p.Ile186Thr and p.Pro189Leu (Blaydon et al., 2012). Further, substantial evidence implicates the involvement of AREG-induced constitutive activation of the EGFR pathway in tylosis (Blaydon et al., 2012;Brooke et al., 2014;Hosur et al., 2014). In addition, skin biopsies of tylosis patients suggest increased EGFR activity (Blaydon et al., 2012). Together, these results, along with the recent findings that mutations in Rhbdf2 in mice result in enhanced EGFR pathway activation (Hosur et al., 2014;Siggs et al., 2014), suggest a possible mechanism for the association between RHBDF2 mutations and tylosis. However, the detailed physiological role of RHBDF2 in facilitating secretion of AREG needs to be established.
Here, we generated mice carrying a human GOF mutation p.P189L ( p.P159L in mice) in RHBDF2 to examine whether GOF mutations in human RHBDF2 lead to enhanced AREG secretion and tylosis. We provide evidence that Rhbdf2 GOF mutations enhance AREG secretion to cause hyperplasia and hyperkeratosis. In addition, we show that genetic ablation of AREG attenuates skin disease in the Rhbdf2 P159L/P159L mouse model of human tylosis. Together, our data strongly suggest that inhibition of AREG could have potential therapeutic value in the treatment of tylosis.

RESULTS AND DISCUSSION The human tylosis disease mutation enhances AREG secretion
Autosomal dominant mutations in the human RHBDF2 gene cause tylosis (Blaydon et al., 2012;Saarinen et al., 2012). Substantial evidence implicates the involvement of AREG-induced constitutive activation of the EGFR pathway (Blaydon et al., 2012;Brooke et al., 2014;Hosur et al., 2014) in human tylosis. We recently showed that similar to GOF mutations in the mouse Rhbdf2 gene (Rhbdf2 cub/cub ), GOF mutations in human RHBDF2 (tylosis mutation p.I186T) promote increased secretion of AREG in vitro (Hosur et al., 2014). To determine in vivo whether the Rhbdf2 cub mutation (Fig. 1A, loss of the cytosolic N-terminal domain of RHBDF2) phenocopies human tylosis, and whether the tylosis p.P189L mutation (Fig. 1B, missense mutation in the cytosolic N-terminal domain of RHBDF2) in RHBDF2 increases AREG secretion, we generated mice carrying the human tylosis mutation p.P189L ( p.P159L in mice) using CRISPR/Cas9-mediated targeting and homology-directed repair in C57BL/6J zygotes (Fig. 1C). We characterized the resulting Rhbdf2 P159L/P159L mice and observed that similarly to Rhbdf2 cub/cub mice (Hosur et al., 2014), mice carrying the human tylosis mutation present with complete hair loss (Fig. 1D). However, unlike Rhbdf2 cub/cub mice (Hosur et al., 2014;Siggs et al., 2014), which present complete hair loss throughout their lives, Rhbdf2 P159L/P159L mice developed a thin layer of curly hair coat that was noticeable in adult mice (Fig. 1E), indicating that the Rhbdf2 cub phenotype is more severe than that of Rhbdf2 P159L . Nevertheless, Rhbdf2 P159L/P159L mice elicited significantly accelerated wound healing, as determined by ear punch closure ( Fig. 2A,B). Additionally, accelerated wound healing was associated with an increase in stimulated secretion of AREG in Rhbdf2 P159L/P159L MEKs compared with Rhbdf2 +/+ MEKs (Fig. 2C).
Next, to examine whether mice carrying the human tylosis mutation phenocopy the Rhbdf2 cub mutation (Hosur et al., 2014), we performed histological analysis of skin sections from Rhbdf2 +/+ and Rhbdf2 P159L/P159L mice. Rhbdf2 P159L/P159L mice presented with follicular dystrophy (F), enlarged sebaceous glands (S), hyperplasia (H), and hyperkeratosis (arrow) of skin (Fig. 2D). To assess enhanced EGFR signaling in the skin of Rhbdf2 P159L/P159L mice, we performed immunohistochemical analyses of skin sections from Rhbdf2 +/+ and Rhbdf2 P159L/P159L mice. We observed a considerable increase in epidermal EGFR signaling pathway downstream effectors phospho-ERK1/2 and phospho-mTOR (Fig. 3A). Similarly, immunoblot analyses of mouse embryonic fibroblasts (MEFs) isolated from Rhbdf2 +/+ and Rhbdf2 P159L/P159L mice also revealed enhanced EGFR signaling as demonstrated by reduced EGFR protein levels (due to downregulation of sustained EGFR receptor activation), and by enhanced protein levels of phospho-AKT, phospho-ERK1/2 and phospho-S6 in Rhbdf2 P159L/P159L MEFs compared with those of Rhbdf2 +/+ mice (Fig. 3B). Also, there was a two-to threefold increase in the levels of the proliferation marker Ki-67 in the skin of Rhbdf2 P159L/P159L mice compared with that of Rhbdf2 +/+ mice (Fig. 3C). Collectively, these data suggest that Rhbdf2 cub/cub and Rhbdf2 P159L/P159L are GOF mutations in the Rhbdf2 gene, and that these mutations enhance secretion of AREG to cause tylosis.

Animals
Mice were bred and maintained under modified barrier conditions at The Jackson Laboratory (JAX). The Animal Care and Use Committee at JAX approved the experimental procedures. For the tylosis mutation, we targeted the Rhbdf2 locus in C57BL/6J zygotes by pronuclear microinjection of Cas9 mRNA (60 ng/μl), single guide RNA (sgRNA) (30 ng/μl), and singlestranded oligonucleotide DNA (ssDNA) (1 ng/μl). A total of 10 pups were born, and five of the 10 pups exhibited partial to complete hair loss. Tail DNA was amplified using PCR and then sequenced, and subsequently founder mice carrying the desired modified alleles were backcrossed to C57BL/6J mice to validate germline transmission. The resulting offspring heterozygous for the Rhbdf2 P159L allele were intercrossed to generate homozygous C57BL/6J-Rhbdf2 P159L/P159L mice. Similar to the Rhbdf2 cub/+ heterozygous mice (Hosur et al., 2014;Johnson et al., 2003;Siggs et al., 2014), the Rhbdf2 P159L/+ heterozygous mice present with a normal hair coat, and moreover there is no evidence of epidermal hyperplasia or hyperkeratosis. We recently showed that B6.Cg-Areg Mcub Rhbdf2 cub /J mice carry a T-to-G point mutation that disrupts the coding frame and introduces a premature stop codon in the Areg gene (Hosur et al., 2014). B6. Cg-Areg Mcub Rhbdf2 cub /J do not produce a functional AREG protein; hence, Areg Mcub/Mcub mice are referred to as Areg −/− in this manuscript.

Ear hole closure and histology
Ear punch closure, histology and histological analyses of hematoxylin and eosin (H&E)-stained sections were performed as previously described (Hosur et al., 2014). Slides were scanned using a Nanozoomer digital slide scanner (2.0-HT, Hamamatsu Photonics, Hamamatsu, Japan) and analyzed with the help of Dr Rosalinda Doty, a board-certified pathologist at JAX.

Isolation of fibroblasts and keratinocytes
MEFs were isolated from 13.5 days postcoitum mouse embryos using Liberase DL (Sigma-Aldrich). Embryos were dissected in 15 ml sterile serum-free DMEM and minced with a razor blade in a 100 mm petri dish. The DMEM medium was replaced with fresh serum-free DMEM containing Liberase DL (0.5 mg/ml), and the cell suspension was allowed to incubate at 37°C for 75 min. Following incubation, cells were centrifuged, washed three times with sterile PBS, re-suspended in DMEM containing 10% fetal bovine serum and antibiotic/antimycotic (Thermo Fisher Scientific), and plated in collagen-coated T-75 flasks. Mouse embryonic keratinocytes were isolated from the skin of embryonic day 18 mouse embryos using Neutral Protease (Dispase) (Worthington Biochemical Corporation, Lakewood, USA). Briefly, skin was peeled off, washed twice with sterile PBS, and incubated in Keratinocyte Growth Medium-2 (Lonza, Allendale, USA) containing antibiotic/antimycotic (Thermo Fisher Scientific). Following overnight incubation at 4°C, epidermis was carefully separated from dermis and placed in a sterile petri dish containing 2 ml trypsin (TrypLE Select Enzyme, Thermo Fisher Scientific) for 30 min at room temperature. Subsequently, trypsin was neutralized with Defined Trypsin Inhibitor (Thermo Fisher Scientific), the cell suspension was centrifuged, and cells were seeded in collagen-coated six-well plates.

AREG ELISA
MEKs seeded in collagen-coated six-well dishes were stimulated with 100 nM PMA (R&D Systems, Minneapolis, USA) for the indicated times. AREG protein levels in the conditioned medium were measured using a Mouse Amphiregulin DuoSet ELISA Developmental Kit (#DY989, R&D Systems). A spectrophotometer (SpectraMax 190, Molecular Devices, Sunnyvale, USA) was used to determine the optical density.

Statistical analysis
GraphPad Prism v6 was used to analyze data. Student's t-test was used to assess the statistical difference between two groups. P<0.05 was considered statistically significant.