Upregulation of ANGPTL6 in mouse keratinocytes enhances susceptibility to psoriasis

Psoriasis is a chronic inflammatory skin disease marked by aberrant tissue repair. Mutant mice modeling psoriasis skin characteristics have provided useful information relevant to molecular mechanisms and could serve to evaluate therapeutic strategies. Here, we found that epidermal ANGPTL6 expression was markedly induced during tissue repair in mice. Analysis of mice overexpressing ANGPTL6 in keratinocytes (K14-Angptl6 Tg mice) revealed that epidermal ANGPTL6 activity promotes aberrant epidermal barrier function due to hyperproliferation of prematurely differentiated keratinocytes. Moreover, skin tissues of K14-Angptl6 Tg mice showed aberrantly activated skin tissue inflammation seen in psoriasis. Levels of the proteins S100A9, recently proposed as therapeutic targets for psoriasis, also increased in skin tissue of K14-Angptl6 Tg mice, but psoriasis-like inflammatory phenotypes in those mice were not rescued by S100A9 deletion. This finding suggests that decreasing S100A9 levels may not ameliorate all cases of psoriasis and that diverse mechanisms underlie the condition. Finally, we observed enhanced levels of epidermal ANGPTL6 in tissue specimens from some psoriasis patients. We conclude that the K14-Angptl6 Tg mouse is useful to investigate psoriasis pathogenesis and for preclinical testing of new therapeutics. Our study also suggests that ANGPTL6 activation in keratinocytes enhances psoriasis susceptibility.


K14-Angptl6
Tg mice develop psoriasis-like epidermal proliferative and inflammatory phenotypes. To assess a potential relationship between ANGPTL6 expression in skin tissue and psoriasis-like conditions, we evaluated phenotypes in K14-Angptl6 Tg mice. We observed no gross difference in skin appearance between K14-Angptl6 Tg and wild-type littermates from birth to 8 weeks of age. However, by 12 weeks skin tissue of K14-Angptl6 Tg mice was reddish and swollen ( Fig. 1a and Supplementary Fig. S1a). Histological analysis revealed a significantly thickened epidermal layer with elongated rete ridges, or thickenings, extending between dermal papillae ( Fig. 1b and Supplementary Fig. S1b), and by 20 weeks K14-Angptl6 Tg mice exhibited skin papillomatosis. Immunohistochemical analysis revealed cells positive for CK14, a marker of proliferating keratinocytes, in suprabasal and basal layers of skin tissue from K14-Angptl6 Tg mice ( Fig. 1c and Supplementary Fig. S1c), suggesting that ANGPTL6 overexpression in keratinocytes promotes proliferation of prematurely differentiated keratinocytes. Moreover, numerous CD68-positive macrophages and CD31-positive endothelial cells were recruited to the dermal layer of K14-Angptl6 Tg skin tissue ( Fig. 1c and Supplementary Fig. S1c), changes not seen in wild-type littermates. Interestingly, it is reported that skin tissues of K14-Angptl6 mice show proliferation of CD31-positive endothelial cells and that Angptl6 enhances blood flow by promoting angiogenesis and arteriogenesis 16,17 .
Appearance of a rash or lesion after application of physical stimulus to healthy skin tissue, known as the Koebner phenomenon, is a common diagnostic for psoriasis 18 . We evaluated skin of K14-Angptl6 Tg mice at 20 weeks of age for a potential Koebner effect in response to tape stripping. Following this stimulus, K14-Angptl6 Tg mice developed a rash with skin exfoliation by 4 day after the procedure, changes not seen in similarly tested wild-type littermate control mice ( Supplementary Fig. S1d). Histological examination of rash lesions revealed keratinocyte hyperplasia and neutrophil accumulation in the dermal layer ( Fig. 1d and Supplementary Fig. S1d), physiological responses associated with human psoriasis.
Epidermal water barrier function is also disrupted in the active phase of psoriasis 19,20 ; thus we examined related phenotypes in skin tissue of K14-Angptl6 Tg mice by estimating transepidermal water loss (TEWL). TEWL values significantly increased in K14-Angptl6 Tg mice relative to wild-type littermates, indicative of decreased epidermal water barrier function in transgenics (Fig. 1e). Moreover, STAT3 is reportedly activated in skin cells of psoriasis patients 21 , and transgenic mice expressing constitutively active STAT3 in keratinocytes represent a mouse model of psoriasis 21 . Our western blotting analysis revealed that the ratio of phosphorylated to total STAT3 protein increased in skin tissue of K14-Angptl6 Tg relative to wild-type littermates (Fig. 1f), suggesting STAT3 activation in K14-Angptl6 Tg mice. STAT3 signaling reportedly increases expression of IL17a, IL17f and Ccl20, chemokines/cytokines upregulated in psoriasis skin tissue 22 . RT-PCR analysis revealed that levels of transcripts encoding all three factors significantly increased in skin tissue of K14-Angptl6 Tg relative to wild-type littermates ( Fig. 1g and Supplementary Fig. S1e). We conclude overall that K14-Angptl6 Tg mice exhibit psoriasis-like epidermal characteristics. ANGPTL6 expression is induced by physical injury to skin. Aberrant skin tissue repair is part of the pathogenesis of psoriasis 3 , and psoriasis can be exacerbated by physical stimuli such as sunburn, physical injury, or chemical irritants. Therefore, we asked whether ANGPTL6 expression levels in skin change in response to injury caused by repeated application and removal of adhesive tape in wild-type mice at 12 weeks of age. Levels of ANGPTL6 transcripts significantly increased at the injury site relative to uninjured control mice by 3 day after injury (Supplementary Fig. S1f). Also, ANGPTL6 protein levels significantly increased at day 1 after injury and remained higher that controls but to a lesser extent on day 3 ( Supplementary Fig. S1g), suggesting that physical injury to skin induces ANGPTL6 expression. To identify cell types expressing ANGPTL6, we performed double-immunofluorescence analysis of skin tissues for ANGPTL6 and the macrophage marker Mac-3 following injury. We observed numerous macrophages infiltrating the dermis by 1 day after injury, a period corresponding to the greatest increase in ANGPTL6 protein level, and most infiltrated macrophages were ANGPTL6-positive ( Supplementary Fig. S1h). Some neutrophils and T and B cells were detected in injured skin, and they also were ANGPTL6-positive ( Supplementary Fig. S1i). Keratinocytes observed in these preparations were not ANGPTL6-positive (data not shown). These data indicate that in normal mouse skin, infiltrated macrophages, neutrophils, and T and B cells in dermis produce ANGPTL6 in response to injury.
Increased ANGPTL6 activity in keratinocytes promotes psoriasis-like skin phenotypes independent of S100A9 function. To define molecular mechanisms underlying skin pathologies seen in K14-Angptl6 Tg mice, we undertook gene-expression profiling of whole ear tissue from 20-week-old K14-Angptl6 Tg versus comparably aged wild-type littermates using RNA sequencing technology (Fig. 2a, Tables 1 and 2). Of transcripts that increased in K14-Angptl6 Tg mice, among the top 30 were several genes associated with psoriasis ( Fig. 2a and Table 1) 11,[23][24][25][26][27][28] . High levels of S100a8/S100a9 (also known as calprotectin) are reportedly induced in epidermal keratinocytes in wound healing or in inflammatory processes, including psoriasis, while normal Tg mice at 12 weeks of age (n = 6 per each group). (f) Representative immunoblotting and quantitation of ratio of p-STAT3 to total STAT3 from skin samples of wild-type and K14-Angptl6 Tg mice at 20 weeks of age. Hsc70 served as loading control. Values in wild-type mice were set to 1 (n = 4 per each group). (g) Quantitative real-time RT-PCR analysis of indicated cytokine/chemokine transcripts in ear tissues at 20 weeks of age. Values for respective WT mice were set to 1 (n = 4-5 per each group). Data are means ± SEM. *p < 0.05, **p < 0.01 between genotypes or groups.
Scientific RepoRts | 6:34690 | DOI: 10.1038/srep34690 epidermis shows very low levels 11,29 . We found that S100a8 and S100a9 transcript (Fig. 2b) and protein ( Fig. 2c) levels increased in skin tissue of K14-Angptl6 Tg relative to wild-type controls. Recent reports indicate that the absence of S100A8 protein was also detected in S100a9 −/− mice due to inefficient translation of S100a8 mRNA or due to instability of S100A8 protein in the absence of its partner, S100A9 11,30 . Moreover, psoriasis-like skin proliferative and inflammatory phenotypes are significantly rescued in psoriasis model mice in which S100A9 is genetically deleted 11 , suggesting that S100A9 could be a useful therapeutic target for psoriasis. Therefore, we asked whether S100A9 loss would alter psoriasis-like skin phenotypes in K14-Angptl6 Tg mice by genetically deleting S100a9 in K14-Angptl6 Tg mice (K14-Angptl6 Tg;S100a9 −/− ) using CRISPR-Cas9-mediated technology ( Supplementary Fig. S2). Overall, the gross appearance of skin in K14-Angptl6 Tg;S100a9 −/− and K14-Angptl6 Tg littermates exhibited indistinguishable inflammatory phenotypes throughout animal's lifespan ( Fig. 3a and Supplementary Figs S3a and S4a). Histological and immunohistochemical analysis of skin tissue from mice of both genotypes with antibodies to CK14, CD68 and CD31 (markers of early keratinocytes, macrophages, and endothelial cells, respectively) revealed no differences in pathology of psoriasis-like lesions ( Fig. 3b and Supplementary Figs S3b and S4b). Moreover, the Koebner phenomenon in skin tissue promoted by tape stripping injury was observed in both K14-Angptl6 Tg;S100a9 −/− and K14-Angptl6 Tg mice ( Supplementary Fig. S3c). Moreover, both genotypes showed comparable STAT3 activation and expression levels of psoriasis-associated chemokine/cytokines (Fig. 3c,d and Supplementary Figs S3d and S4c,d). Thus, S100A9 loss in K14-Angptl6 Tg male or female mice did not rescue psoriasis-like skin phenotypes, suggesting that ANGPTL6-dependent psoriasis emerges independently of S100A9 and that these phenotypes are not sex-dependent.
ANGPTL6 production increases in keratinocytes of some psoriasis patients. Finally, we asked whether ANGPTL6 activation is correlated with skin tissue pathologies seen in psoriasis patients. To do so, we compared ANGPTL6 protein levels between psoriasis patients and non-psoriasis control subjects (Tables 3 and 4). Immunohistochemical analysis of tissue specimens with ANGPTL6 antibody revealed abundant or moderate ANGPTL6 protein in the epidermal layer of 13 of 15 tissue specimens from different patients, whereas no ANGPTL6 protein was detected in all 10 non-psoriasis control specimens (Fig. 4a,c and Supplementary Fig. S5a,c). Two patients did not show elevated ANGPTL6 production in skin tissue ( Fig. 4b and Supplementary Fig. S5b). Immunofluorescence analysis of skin tissue from one psoriasis patient revealed ANGPTL6 production in epidermal keratinocytes, suggesting they are the source of ANGPTL6 in skin tissue of these patients (Fig. 4d). By Tg relative to wild-type littermate mice. Genes shown in red were previously reported to be psoriatic-related genes 11,23-28 . (b) S100a8 and S100a9 transcript levels in psoriatic-like lesions of wild-type or K14-Angptl6 Tg mice at 20 weeks of age. Values for respective WT mice were set to 1 (n = 4-5 per each group). Data are means ± SEM. *p < 0.05, **p < 0.01 between genotypes. (c) Immunohistochemistry staining for S100a8 and S100a9 in neck skin of WT and K14-Angptl6 Tg mice at 20 weeks of age. Scale bar: 100 μ m.
contrast, S100A8 and S100A9 proteins were abundant in skin tissues of all 15 psoriasis patients but not in tissues from any non-psoriasis control subject (Fig. 4, Supplementary Fig. S5, Tables 3 and 4), suggesting that ANGPTL6 production and S100A8/S100A9 expression are not linked in the pathogenesis of psoriasis. Thus, elevated ANGPTL6 levels seen in keratinocytes of psoriasis patients support our observations in mice that ANGPTL6 may induce psoriasis-like phenotypes.

Discussion
Over the past decades, various mouse models of skin pathologies resembling psoriasis have provided important insight into molecular mechanisms undelying psoriasis 3 . Nonetheless, differences in immune-associated gene expression between mouse models and psoriatic patients 31 require development of additional models that reflect unique skin phenotypes and gene expression patterns seen in patients. In this study, we report that K14-Angptl6 Tg mice develop psoriasis-like epidermal proliferation consisting of prematurely differentiated keratinocytes, which contributes to epidermal barrier dysfunction, and also exhibit psoriasis-like skin tissue inflammation marked by infiltration of neutrophils and vascular endothelial cells into the dermal layer. We also observed enhanced STAT3 activation in skin tissue of transgenics and upregulated expression of psoriasis-associated chemokines/cytokines. Finally, we report that ANGPTL6 production is elevated in keratinocytes from some psoriasis patients. Our results strongly suggest that increased ANGPTL6 production from keratinocytes contributes to psoriasis pathogenesis and that K14-Angptl6 Tg mice could serve as alternate mouse models of the disease. They also suggest that K14-Angptl6 Tg mice could serve as a murine psoriatic model for preclinical testing of biologics useful to attenuate the condition.
Macrophages reportedly play an important role in the pathophysiological development of psoriasis, and their deletion reportedly can rescue psoriasis skin phenotypes 32,33 . In this study, we observed that infiltrating macrophages express ANGPTL6 in response to injury of normal mouse skin. Thus, deletion of macrophages via clodronate liposome treatment may rescue ANGPTL6-dependent psoriasis when skin tissues repeatedly suffer injurious stimuli. In regard to skin tissue repair after injury, however, ANGPTL6 may function in keratinocyte proliferation. Further studies are needed to investigate mechanisms by which ANGPTL6 promotes not only psoriasis development but also skin tissue repair in physiological conditions in response to normal skin injury. Psoriasis starts with angiogenesis in the superficial dermal microvasculature 34 . Psoriatic skin is characterized by increased expansion and permeability of the endothelium of the superficial microvasculature 35 . These phenotypes allow leukocyte transmigration into areas of inflammation via enhanced expression of cell adhesion molecules associated with the onset of psoriasis 36,37 . We previously reported that ANGPTL6 induces not only neovascularization but also enhanced permeability of the local vasculature 16 . These data suggest that ANGPTL6 modulates permeability of the microvasculature and/or immature neomicrovasculature and accelerates leukocyte transmigration into dermal areas at the onset of inflammation.
Although human psoriasis has a multifactorial etiology, it is strongly influenced by genetics 2 . A search for genes responsible for familial psoriasis via genome-wide association study (GWAS) identified several putative psoriasis susceptibility loci 38 . Interestingly, human ANGPTL6 is located at 19p13.2, located in the psoriasis susceptibility region PSORS6 14,15 . It would be of interest to investigate whether ANGPTL6 levels are elevated in skin tissues of individuals with mutations in PSORS6. Also, a recent report suggests significant association of ANGPTL6 gene polymorphisms (rs6511435) tends to be associated with a 20% higher risk of metabolic syndrome 39 , while another found that serum ANGPTL6 levels are significantly higher in subjects with metabolic syndrome than in healthy individuals 40 . Interestingly, psoriasis patients have a higher prevalence of metabolic syndrome 41,42 ; thus it would also be of interest to investigate whether there are ANGPTL6 gene polymorphisms associated with psoriasis.
Our gene expression profile analysis of K14-Angptl6 Tg mice revealed several genes whose expression levels were previously reported to increase in psoriasis skin tissues 11,23-28 . One of those was S100A9, which was recently suggested to be a potential therapeutic target for psoriasis 11 . Interestingly, psoriasis-like skin proliferative and inflammatory phenotypes were not rescued in K14-Angptl6 Tg mice by S100A9 deletion. This observation is consistent with our analysis of patient specimens: we observed elevated ANGPTL6 levels in keratinocytes of some psoriasis patients, while S100A8 and S100A9 were elevated in keratinocytes of all patients. We suggest that ANGPTL6-dependent mechanisms associated with psoriasis may be independent of those governed by S100A9. Our gene profiling analysis also revealed A230050P20Rik to be one of the most significantly upregulated transcripts in K14-Angptl6 Tg mice. Interestingly, a recent paper showed that chromosome 19 open reading frame 66 (C19orf66), the human ortholog of A230050P20Rik, functions as an interferon-stimulated cellular inhibitor of dengue virus replication 43 Table 3. ANGPTL6, S100A8 and S100A9 expression in skin biopsies from psoriasis patients. In summary, ours is the first study to report that in mice, epidermal ANGPTL6 activity promotes aberrant epidermal barrier function associated with keratinocyte hyperproliferation and hyperactive chemokine/cytokine signaling, and is accompanied by neutrophil infiltration and concomitant release of proinflammatory factors in skin lesions. We also found that infiltrated macrophage-derived ANGPTL6 is upregulated by physical stimuli, such as skin injury. We conclude that K14-Angptl6 Tg mice represent a novel symptomatic, histological, and molecular model of the condition and could be useful in preclinical testing of new therapeutics against psoriasis.

Methods
Animal study. For this study, 20-week-old male and female wild-type mice on a C57BL/6N genetic background (Kyudo Co., Ltd., Saga, Japan), transgenic mice expressing Angptl6 in keratinocytes (K14-Angptl6 Tg) 13 , and S100a9-knockout (S100a9 −/− ) mice, plus respective wild-type littermates on a C57BL/6N background were used. All animals were fed a normal diet (CE-2, CLEA, Tokyo, Japan), bred in a mouse house under specific pathogen-free (SPF) conditions with automatically controlled lighting (12 h on, 12 h off), and maintained at a stable temperature of 23 °C in the Animal Resource Facility at Kumamoto University. Tape stripping analysis was performed using 20 strokes with transparent tape on dorsal skin of 10-to 12-week-old mice following hair removal, under sodium pentobarbital anesthesia. TEWL of mouse skin was determined using a Tewameter TM300 (Courage and Khazaka, Cologne, Germany) as described 44 . All the experimental procedures were approved by the Committee on Animal Research at Kumamoto University and conducted in accordance with guidelines of Institutional Animal Committee of Kumamoto University. Immunohistochemical staining. Human and mouse skin tissues were embedded in OCT compound (Sakura Finetechnical, Tokyo, Japan) and frozen in liquid nitrogen. Mouse skin tissues were fixed in 4% paraformaldehyde for 24 h and embedded in paraffin. Hematoxylin and eosin (HE) was used to stain 6 μ m-thick frozen sections and 4-μ m-thick paraffinized sections. For immunohistochemistry, a rabbit polyclonal anti-human ANGPTL6 antibodies (1:100) was produced by immunizing rabbits with a synthetic peptide corresponding to amino acids 392-408 (NDKPESTVDRDRDSYSG) of ANGPTL6, and a rabbit polyclonal anti-mouse Angptl6 (1:100) was used previously reported 13 . Supplementary Table S2 shows information relevant to other primary antibodies. Fixed sections from back, neck and ear skin of K14-Angptl6 Tg mice and their littermate wild-type controls were stained with 1:500 diluted anti-mouse Angptl6 antibody. Sections were pretreated with periodic acid (Nichirei, Tokyo, Japan) to inhibit endogenous peroxidases. Subsequently, specimens were incubated overnight with each antibody dilution. After washing sections with phosphate-buffered saline (PBS), immunostaining was performed using Histofine rat stain kit, Histofine goat stain kit (Nichirei Biosciences Inc), or EnVision/horseradish peroxidase (Dako) according to the manufacturers' instruction, and specimens were then counterstained with hematoxylin. As negative controls, the same procedures were performed using isotype control IgG rather than primary antibodies. Peroxidase activity was visualized by incubation with a 3,3-diaminobenzidine solution and analysis using a BIOREVO BZ-9000 microscope (Keyence, Osaka, Japan). For double immunofluorescence, a rabbit polyclonal anti-human ANGPTL6 antibody (1:100) was used with other primary antibodies and Alexa Fluor-conjugated secondary antibodies, as indicated in Supplementary Table S2. After washing sections with PBS, fluorescent images were captured by confocal laser microscopy (FluoView FV1200, Olympus Corporation, Japan).