Ichthyosis molecular fingerprinting shows profound TH17 skewing and a unique barrier genomic signature

Abstract

Background

Ichthyoses are a group of rare skin disorders lacking effective treatments. Although genetic mutations are progressively delineated, comprehensive molecular phenotyping of ichthyotic skin could suggest much-needed pathogenesis-based therapy.

Objective

We sought to profile the molecular fingerprint of the most common orphan ichthyoses.

Methods

Gene, protein, and serum studies were performed on skin and blood samples from 29 patients (congenital ichthyosiform erythroderma, n = 9; lamellar ichthyosis, n = 8; epidermolytic ichthyosis, n = 8; and Netherton syndrome, n = 4), as well as age-matched healthy control subjects (n = 14), patients with psoriasis (n = 30), and patients with atopic dermatitis (AD; n = 16).

Results

Using criteria of a fold change of greater than 2 and a false discovery rate of less than 0.05, 132 differentially expressed genes were shared commonly among all ichthyoses, including many IL-17 and TNF-α–coregulated genes, which are considered hallmarks of psoriasis (defensin beta 4A, kynureninase, and vanin 3). Although striking upregulation of TH17 pathway genes (IL17F and IL36B/G) resembling that seen in patients with psoriasis was common to all patients with ichthyoses in a severity-related manner, patients with Netherton syndrome showed the greatest T-cell activation(inducible costimulator [ICOS]) and a broader immune phenotype with T_H1/IFN-γ, OASL, and T_H2/IL-4 receptor/IL-5 skewing, although less than seen in patients with AD (all P < .05). Ichthyoses lacked the epidermal differentiation and tight junction alterations of patients with AD (loricrin, filaggrin, and claudin 1) but showed characteristic alterations in lipid metabolism genes (ELOVL fatty acid elongase 3 and galanin), with parallel reductions in extracellular lipids and corneocyte compaction in all ichthyoses except epidermolytic ichthyosis, suggesting phenotypic variations. Transepidermal water loss, a functional barrier measure, significantly correlated with IL-17–regulated gene expression (IL17F and IL36A/IL36B/IL36G).

Conclusion

Similar to patients with AD and psoriasis, in whom cytokinedysregulation and barrier impairment orchestrate disease phenotypes, psoriasis-like immune dysregulation and lipid alterations characterize the ichthyoses. These data support the testing of IL-17/IL-36–targeted therapeutics for patients with ichthyosis similar to those used in patients with psoriasis.

The ichthyoses are a heterogeneous group of rare genetic, debilitating cutaneous disorders^1–4 that share defective keratinization, a compromised epidermal barrier, and increased transepidermal water loss (TEWL).^{2, 3, 5} Patients present with xerosis, hyperkeratosis, often severe erythema, and reduced quality of life.^6–10

Current treatments lack specificity and efficacy and are largely limited to topical emollients, keratolytics, and oral retinoids,^8–11 all with potential toxicity.^{3, 12–15} Although corrective gene therapy is evolving, it is still experimental and limited to model systems.^{16, 17} Thus an unmet need exists for pathogenesis-based therapeutic approaches.

Data derived from animal or human models suggest abnormalities in cornification, lipid homeostasis, and keratinocyte adhesion/desquamation among various ichthyoses, implying shared defects in barrier architecture.^18–26 Human skin and blood studies have been limited to a few patients or select ichthyosis subtypes and have shown abnormalities in lipid, cornified envelope (CE), and/or other differentiation measures.^{18, 27–40}

Several observations link ichthyoses, particularly Netherton syndrome (NS), to atopic dermatitis (AD), which is marked by epidermal barrier defects and immune dysregulation.⁴¹ NS shares the eosinophilia and increased T_H2/atopy-related markers (thymic stromal lymphopoietin [TSLP], chemokine like 17 [CCL17]/thymus and activation-regulated chemokine, and IgE) of AD,^{18, 20, 42–49} and polymorphisms of serine protease inhibitor, Kazal-type 5 (mutated in patients with NS) are associated with AD.^50–52 Moreover, clinical improvement in association with reduced T_H2 marker expression has been noted in single cases of NS treated with omalizumab/anti-IgE^{44, 49} and infliximab/anti–TNF-α,⁴⁹ implying potential reversibility of ichthyosis through immune-targeted treatments.

Recently, using a limited RT-PCR panel, our group showed enhanced T_H17/IL-23 responses in patients with ichthyoses with minimal T_H2 inflammation and preserved expression of the terminal differentiation markers filaggrin (FLG), loricrin(LOR), periplakin (PPL).⁵³ These data linked the ichthyoses with psoriasis, a disease with a proved T_H17/IL-23–centered pathogenesis.^54–56 Variable increases in levels of inflammatory cytokines (IL-17, IL-23, IL-8, IL-1β, TNF-α, CCL20, S100As, CCL17/thymus and activation-regulated chemokine, and TSLP) have been reported in patients with ichthyoses^{21, 46, 48, 49, 57–61} but only in limited animal model studies,^{46, 60–62} keratinocyte cultures,^{21, 25} or specific markers of ichthyosis subtypes in human subjects.^{48, 49, 58, 59} Thus a global comprehensive molecular phenotyping across the ichthyoses is warranted.

The present study is the first detailed genomic and cellular profiling of the most common orphan forms of ichthyosis(NS, autosomal recessive congenital ichthyosis [ARCI] subtypes of lamellar ichthyosis [LI] and congenital ichthyosiform erythroderma [CIE], and epidermolytic ichthyosis [EI]) compared with control subjects. To identify whether ichthyosis bears similarity to either AD (T_H2 skewed) or psoriasis(T_H17/IL-23 polarized), we included these common diseases as comparators. Our fingerprinting showed ichthyotic skin to harbor robust T_H17/IL-23 skewing and activation of genes considered to be molecular classifiers of psoriasis (kynureninase [KYNU], vanin 1 [VNN1], and vanin 3 [VNN3]). Patients with ichthyoses showed less T_H2 expression than patients with AD (and in many cases less than control subjects) and lacked its characteristic differentiation abnormalities. Significant correlation between disease severity and IL-17activation in skin and blood suggests systemic T_H17 activation and further supports the use of systemically administered IL-17 antagonists as therapy.

Methods

Patients’ characteristics and samples

Institutional review board–approved consent was provided by subjects (≥12 years) and parents (<18 years), and demographics, medical history, physical examination, clinical severity scores (Ichthyosis Area and Severity Index [IASI], including Ichthyosis Area and Severity Index–erythema [IASI-E], and Ichthyosis Area and Severity Index–scaling [IASI-S]), pruritus scores (5-D itch and Numeric Rating Scale), and TEWL were obtained, as reported.⁵³ The IASI score consists of erythema (IASI-E) and scaling (IASI-S) subscores and quantifies score intensity within body regions.⁵³ Four-millimeter biopsy specimens and sera were obtained (see details in the Methods section in this article’s Online Repository at www.jacionline.org). Because patients with ichthyosis were significantly younger than those in the psoriasis and AD groups, a sensitivity analysis was performed, confirming no significant differences in results when adjusting for age (data not shown; Table I and see the Methods section in this article’s Online Repository).

Table I.

Patients’ demographics

	Parameter	Control subjects (n = 21)	Patients with ichthyosis (n = 29)	P value, control subjects vs patients with ichthyosis	Patients with NS (n = 4)	Patients with LI (n = 8)	Patients with CIE (n = 9)	Patients with EI (n = 8)	P value, subtypes
Age (y)	Mean ± SD	26.1 ± 20.4	21.2 ± 15.4	.362	15.2 ± 7.4	27.7 ± 15.6	19.5 ± 15.6	19.7 ± 18.3	.551
	(range)	(6.4-57)	(2.2-56.9)		(6.3-23)	(9.2-56.9)	(4.0-44)	(2.2-55)
Sex	F/M	10/11	16/13	.810	2/2	6/2	3/6	5/3	.358
Race	EA/AA/As/H	16/1/0/4	20/3/3/3	.443	2/1/0/1	6/1/0/1	6/1/2/0	6/0/1/1	.590
IASI score	Mean (±SD)	NA	27.9 ± 8.4	NA	31.0 ± 10.8	29.2 ± 7.5	23.5 ± 9.2	30.4 ± 6.8	.284

AA, African American; As, Asian; EA, European American; F, female; H, Hispanic; M, male; NA, not applicable.

Gene-expression analyses

HGU133Plus2.0 microarrays (GSE#108640; Affymetrix, Santa Clara, Calif) were used for gene arrays, as previously described.⁶³ RNA was extracted for RT-PCR with the Qiagen miRNeasy Mini Kit (Qiagen, Valencia, Calif), as previously described.^{64, 65} Expression values were normalized to human acidic ribosomal protein. Primers and probes are listed in Table E1 in this article’s Online Repository at www.jacionline.org. See the Methods section in this article’s Online Repository for details.

Serum immunoassays

Electrochemiluminescence immunoassays (Singulex, Alameda, Calif) were used to quantify IL-17A in sera from patients with ichthyosis, as previously described.⁶⁶

Immunostaining and immunofluorescence

Immunohistochemistry was performed on frozen sections by using the listed antibodies (see Table E2 in this article’s Online Repository at www.jacionline.org), as previously described.⁶⁷Immunofluorescence staining for neutral lipids with Nile red was performed, as previously eported.^{68, 69} See the Methods section in this article’s Online Repository for details.

Statistical analysis

Statistical analysis was performed with R software (www.R-project.org) and packages available through Bioconductor(www.bioconductor.org). Quality control was performed, as previously described.⁶³ Expression values were obtained by using the GC Robust Multiarray Average algorithm. Batch effects were removed by identifying and adjusting for surrogate variables with the sva package.⁷⁰ Probe sets with 15 or more samples and expression values of greater than 3 were kept for analysis. Data were log₂-transformed and fitted with a linear model. Fold changes (FCHs) between comparisons of interest were estimated, and hypothesis testing was conducted by using contrasts under the general framework for linear models in the R software limma package. P values from the moderated (paired) t test were adjusted for multiple hypotheses by using the Benjamini-Hochberg procedure. Genes with FCHs of greater than 2 and a false discovery rate (FDR) of less than 0.05 were considered differentially expressed. Hierarchical clustering of samples/conditions used a McQuitty agglomeration algorithm. Gene-set variance analysis was performed by using unsupervised sample-wise enrichment.⁷¹ Gene-set overrepresentation analysis was performed with XGR software⁷² and crowd-extracted expression of differential signature disease signatures.⁷³

All 29 samples from patients with ichthyoses, 21 control subjects, and 10 patients with psoriasis and AD were available for RT-PCR analyses. mRNA expression RT-PCR data were log₂-transformed and fitted to a linear model as above. Values of less than the limit of detection were imputed as 20% of the minimal value over the limit of detection, as previously eported.⁷⁴ See the Methods section in this article’s Online Repository for details.

Results

Twenty-nine patients (age, 2.2-56.9 years; mean age, 21.2 years; 16 female and 13 male patients) with ichthyosis and known mutations, including 4 with NS, 8 with ARCI-LI, 9 with ARCI-CIE, and 8 with EI, were enrolled (Table I and see Table E3 in this article’s Online Repository at www.jacionline.org). Healthy control subjects (n = 21) were between 6.4 and 57 years old (mean age, 26.1 years). Data from previous studies on 16 patients with moderate-to-severe AD (age, 33-69 years; mean age, 49.8 years; mean SCORAD score, 61.2)⁶³ and 30 patients with psoriasis (age, 33–67 years; mean age, 48.5 years; mean Psoriasis Area and Severity Index score, 23.6)⁷⁵ served as comparators.

Genomic profiling identifies the common and individual ichthyosis phenotypes

Gene arrays were performed on 21 of the 29 patients with ichthyoses (6 patients with ARCI-CIE, 7 patients with ARCI-LI, 5 patients with EI, and 3 patients with NS) and 14 healthy age-matched control subjects.⁵³ Although a principal component analysis of microarray data demonstrated considerable overlap among the ichthyosis subtypes, it also distinguished between individual subtypes and healthy control subjects, which clustered separately (see Fig E1 in this article’s Online Repository at www.jacionline.org). Using criteria of an FCH of greater than 2 and an FDR of less than 0.05, we identified the common ichthyosis molecular phenotype, consisting of a core set of 132 differentially expressed genes (DEGs; 113 upregulated and 19 downregulated genes), as well as unique genomic signatures of each subtype (Fig 1, A). Among upregulated genes in the common ichthyosis transcriptome were many psoriasis hallmark genes, including Defensin beta4A (DEFB4A), kynureninase (KYNU), VNN3, kallikrein 8 (KLK8), late cornified envelope 3D (LCE3D), inducible nitric oxide synthase regulator gene arginase 1, the NOD-like receptor gene caspase recruitment domain family member 6 (CARD6), and the novel psoriasis autoantigen phospholipase A2, group 4D (PLA2G4D) (Fig 1, A, and see Table E4 in this article’s Online Repository at www.jacionline.org).^{76, 77}

Fig 1.

A, Venn diagram showing DEGs comparing ichthyosis subtypes: NS, LI, CIE, and EI. The box represents the common ichthyosis molecular phenotype. B, Heat map representing DEGs upregulated and downregulated in the common ichthyosis phenotype, with a table of fold changes of these genes in patients with psoriasis (PSO), AD, and ichthyosis subtypes. C, Venn diagram depicting how gene alterations of the common ichthyosis phenotype shown to be shared in Fig 1, A, overlap with the molecular signatures of AD and psoriasis. Red, Upregulated; blue, downregulated. ⁺P < .1, *P < .05, **P < .01, and ***P < .001.

However, many DEGs differed among the unique subtypes. To assess these differences, we evaluated uniquely expressed DEGs (Fig 1, A). Of all ichthyoses, patients with EI expressed the greatest number of unique DEGs (223 upregulated and 150 downregulated genes), with upregulation in KLKs (KLK7 and KLK10), keratins (KRT4, KRT72, and KRT9), cell cycle/proliferation (cell division cycle 6/20/25A), and cell transport molecules (dynein axonemal assembly factor 3, clathrin interactor 1, and cilia and flagella associated protein 97). Two hundred thirty-four DEGs were unique to patients with LI, 61 of which were upregulated, including the psoriasis-associated cytokine IL-20 and IL-17–related genes (IL36A and lipocalin 2 [LCN2]). Downregulated DEGs in patients with LI consisted of lipid metabolism (arachidonate 15-lipoxygenase, type B, Acyl-CoA wax alcohol acyltransferase 1, and fatty acid binding protein 7 [FABP7]), T_H1 (CXCL9/CXCL11), and T_H2 (CCL8/CCL17 and IL31RA) markers, as well keratins (KRT35/KRT40/KRT75/KRT81/KRT83/KRT85/KRT86) and keratin-associated protein 1. One hundred fifty-eight DEGs were unique to patients with CIE, including upregulation of the junctional protein desmocollin 1 and downregulation of metalloproteases (ADAM metallopeptidase domain 33 and matrix metallopeptidase [MMP] 9), aquaporin 1(AQP1), collagens (COL14A1), and integrins (ITGA11 and ITGA5). Patients with NS expressed the fewest unique DEGs (63 upregulated and 33 downregulated DEGs), including upregulation of microtubule-associated proteins (MAP4/7), the interferon-induced molecule 2′-5′-Oligoadenylate synthetase 3, and the TNF receptor superfamily member 11a (see Table E4).

We further assessed how different subtypes vary in their inflammatory profiles through analysis of a previously published immune gene subset^{63, 67, 78} represented in a heat map (see Fig E2 in this article’s Online Repository at www.jacionline.org). Relative to other ichthyoses, patients with NS demonstrated the greatest upregulation of T-cell activation markers (MMP12, inducible costimulator [ICOS], and C-C motif chemokine receptor 7), interferon/T_H1 products (IFNG and 2′−5′-oligoadenylate synthetase-like protein [OASL]), and T_H2 (IL4R, IL5, IL13, and CCL18/CCL26) markers, although increases in the latter group were much smaller than in patients with AD. T_H17 (PI3, S100A9/12, and IL17A/F) measures were upregulated in all patients with ichthyoses, with the greatest increases in patients with NS; CCL20 and LCN2 were exceptions and greatest in patients with LI (see Fig E2).

The common upregulated ichthyosis profile overlaps with psoriasis

We compared the global ichthyosis profile with that of psoriasis and AD, which can clinically resemble ichthyoses.^{42, 47, 48, 49, 51, 52} As shown in an unsupervised hierarchical clustering heat map of the top 100 upregulated ichthyosis DEGs (Fig 1, B, and by individual samples in Fig E3 and Table E4 in this article’s Online Repository at www.jacionline.org), there is a striking similarity to psoriasis and little similarity to AD. Upregulated genes in patients with ichthyosis included IL-17/TNF-α–associated molecules (IL36B/IL36G, peptidase inhibitor 3 [PI3], and S100A7A/S100A9/S100A12), with many showing psoriasis-level expression (Fig 1, B). The intersection of the common ichthyosis lesional phenotype with AD and psoriasis signatures is represented in Fig 1, C, as a Venn diagram. When compared with normal skin, 37 DEGs were shared among all diseases, including upregulation of CARD6, DEFB4A, IL36G, LCE3D, PI3, S100A9/12, and VNN3. Notably, the common ichthyosis molecular phenotype demonstrated larger overlap with psoriasis than AD, with psoriasis sharing 92 (69%) of 132 DEGs and AD sharing only 38 (28%) of 132 DEGs. Only ichthyosis and psoriasis showed unique upregulation in KLK8, KYNU, and PLA2G4D (see Table E5 in this article’s Online Repository at www.jacionline.org).

An IL-17/TNF-α gene signature is shared among ichthyosis subtypes

Because an IL-17–skewed pattern was shared across ichthyoses and many IL-17–related genes are synergistically/additively induced by IL-17/TNF-α, we evaluated individual ichthyoses for a subset of IL-17– and TNF-α–regulated genes (Fig 2). All ichthyosis subtypes demonstrated highly significant increases in IL-17/TNF-α hallmark psoriasis genes (KYNU, VNN3, Wiskott-Aldrich syndrome like, dedicator of cytokinesis 9, and kinesin family member 1B; Fig 2); these were not as highly upregulated or were even downregulated in patients with AD.

FIG 2.

Heat map of additive and synergistic IL-17/TNF-a–induced genes across ichthyosis subtypes (NS, LI, CIE, and EI), with fold changes also depicted for psoriasis (PSO) and AD lesions versus normal skin. Red, Upregulated; blue, downregulated. 1P < .1, *P < .05, **P < .01, and ***P < .001

To validate the array data, we also measured mRNA expression of representative IL-17/TNF-α–regulated genes using quantitative PCR/RT-PCR in the 29 patients with ichthyoses. As seen in the box plots in Fig 3, A, all subtypes showed similar highly significant increases in expression of IL-17/TNF-α–induced gene mRNA (IL17F, IL36A/IL36B/IL36G, KYNU, epiregulin [EREG], and VNN1/3) to psoriasis (P < .05 for all). A heat map depicting FCHs of mRNA expressions in skin of patients with lesional ichthyosis versus normal skin (see Fig E4in this article’s Online Repository at www.jacionline.org) also shows overall greater similarity to psoriasis but not AD.

Fig 3.

A, Real-time/RT-PCR–derived mRNA expression of T_H17/TNF-α–induced cytokines and markers in normal (N), psoriasis (PSO), AD, NS, LI, CIE, and EI lesions versus normal skin represented as box plots. Red dots indicate means, black dots indicate individual samples, horizontal black lines indicate medians, and vertical black lines indicate values within 1.5 times the interquartile range of the median; the top and bottom of the boxes indicate 75th and 25th percentiles, respectively. P values above horizontal lines denote significance among groups. B-H,Immunohistochemistry staining of IL-36R in normal (NRML), AD, and psoriasis (PSO) lesions compared with CIE, LI, EI, and NS tissues (×10 magnification). Scale bars = 100 μm.

IL-36α, IL-36β, and IL-36γ subunits exert their proinflammatory effects through binding

IL-36 receptor (IL-36R), which activates intracellular signaling pathways.⁷⁹ To extend our findings, we analyzed IL-36R protein expression through immunohistochemistry. IL-36R is known to form feedback loops driving IL-17 cytokine production in patients with psoriasis.^{80, 81} Although in normal (Fig 3, B) and AD lesions (Fig 3, C) IL-36R showed minimal epidermal expression, widespread expression of IL-36R was noted in psoriatic epidermis (Fig 3, D) and patients with all types of ichthyoses (Fig 3, E–H).

Different ichthyosis subtypes show unique epidermal differentiation and lipid genomic alterations

Ichthyoses were historically considered disorders of lipid transport and epidermal barrier alterations. Hence we assessed genes relating to the epidermal differentiation complex, CE, and lipids across each subtype, as reported in a heat map (Fig 4, A). Although patients with AD showed significant downregulation of differentiation products versus control subjects (LOR, FLG, PPL, envoplakin, corneodesmin, late cornified envelope 1B/E, and annexin A9 [ANXA9]),⁶⁴ ichthyosis showed minimal downregulation of differentiation products, with some measures even showing upregulation (EREG, ANXA9, psoriasis susceptibility 1 candidate 2, LOR, and FLG) versus normal skin (Fig 4, A, yellow box). Significant decreases in claudins (CLDNs; CLDN1/8/23) were seen in patients with AD and those with psoriasis, with variable and less significant reductions across the ichthyoses (Fig 4, A). Lipids were most prominently decreased in patients with AD but were also downregulated in those with psoriasis. Among ichthyoses, patients with LI demonstrated the greatest and most significant lipid downregulation (ELOVL fatty acid elongase 3, galanin [GAL], hydroxyacid oxidase 2, and FABP7), whereas patients with EI generally showed preserved expression of lipid-associated genes (Fig 4, A, green box).

Fig 4.

A, Heat map of epidermal differentiation complex (EDC), CE, and lipid genes expressed across normal skin (N)and skin of patients with psoriasis (PSO), AD, and ichthyoses(NS, LI, CIE, and EI), showing FCHs versus normal skin. Greenand yellow boxes indicate gene sets displaying differences in patients with ichthyoses from those with the other conditions. Red, Upregulated; blue, downregulated. ⁺P < .1, *P < .05, **P < .01, ***P < .001. B-H, Lipid staining using Nile red immunofluorescence in normal (NRML), AD, and psoriasis (PSO) lesions compared with tissue from patients with CIE, LI, EI, and NS, with insets showing decreased corneocyte compaction and reduced extracellular lipids (white arrows; ×20 magnification).

To better visualize stratum corneum (SC) intercellular lipids and validate array findings, we performed Nile red immunofluorescence staining in skin from patients with ichthyoses, patients with AD, patients with psoriasis, and healthy subjects (Fig 4, B–H). Although control skin (Fig 4, B) was characterized by extracellular lipids organized in lamellar sheets between compacted corneocytes, patients with ichthyoses (Fig 4, E–H) showed variable patterns in extracellular lipid layers. Patients with CIE and those with NS showed similar lipid staining as control subjects with compact corneocytes and prominent lipid lamellae, whereas patients with EI demonstrated less distinct lamellar layering and foci of greater intensity (“lacunae-like” defects). Consistent with its greatest downregulation of lipid genes among ichthyoses (Fig 4, A), patients with LI showed diminished corneocyte compaction, visibly reduced lipid lamellae, and intensely staining intracellular, globular lipid-like structures. Qualitatively, patients with AD demonstrated the most profound decreases in lipid layers and SC compaction (Fig 4, C), whereas patients with psoriasis (Fig 4, D) revealed more compact SC and less profound lipid defects.

Pathway analysis highlights similarities to psoriasis

We also performed gene-set variation analysis using previously defined gene lists for IL-17–regulated immune axes.⁶⁷ This is represented as bar plots depicting increases in IL-17–upregulated and synergistic/additive IL-17/TNF-α genes in patients with all types of ichthyoses, most profoundly in patients with NS and psoriasis compared with patients with AD and control subjects, which showed smaller activation of these gene sets (P < .05 for all; Fig 5, A and B). Expression of CE genes was decreased significantly in patients with AD but relatively unchanged or even increased in patients with psoriasis and ichthyoses (Fig 5, C). Lipid genes (Fig 5, D) were reduced significantly in patients with AD (P < .001) and, to a lesser extent, in patients with ichthyoses, except EI, which resembled a lipid profile of normal skin.

Fig 5.

A-D, Gene-set variance analyses (GSVA) for subsets of genes upregulated with IL-17 (Fig 5, A), genes synergistically and additively induced by IL-17/TNF-α (Fig 5, B), CE genes (Fig 5, C), and lipid metabolism genes (Fig 5, D). Expression is represented as a GSVA score for normal (N) skin and skin from patients with psoriasis (PSO), AD, NS, LI, CIE, and EI. Means ± SEMs are shown. P values above horizontal lines denote significance among groups. ⁺P < .1, *P < .05, **P < .01, and ***P < .001. E, Disease signature enrichment analyses of upregulated genes in ichthyosis lesions versus control lesions performed against curated disease signature sets from the Broad Institute. Top representative hits of polarized inflammatory/immune diseases are represented as a bar plot, and all outputs of the enrichment analyses are presented in Table E9. The green line indicates a false discovery rate cutoff of 0.05.

Pathway enrichment analyses using function-based pathway databases (KEGG, REACTOME, Biocarta, and the Pathway Interaction Database; see Tables E6–E10 in this article’s Online Repository at www.jacionline.org) revealed significantly upregulated pathways in the common and individual ichthyosis signatures (FDR < 0.05). These included focal adhesion and mitogen-activated protein kinase signaling for patients with CIE; cyclins, cell-cycle regulation, and lipoprotein metabolism for patients with EI; chemokine signaling and extracellular matrix for patients with LI; and extracellular matrix for patients with NS. We further evaluated gene enrichment of the ichthyosis phenotype with available signatures of other polar inflammatory diseases (see Table E11 in this article’s Online Repository at www.jacionline.org).⁷³ Ichthyosis showed the most significant overlap with the psoriasis vulgaris fingerprint but also shared its profile with other T_H17-mediated dermatoses (acne) and T_H2 conditions (AD; Fig 5, E).⁸²

IL-17– and TNF-α–regulated genes correlate with ichthyosis severity and epidermal barrier disruption

To determine how molecular and cellular biomarkers in the skin of patients with ichthyosis correlated with clinical severity scores (IASI and its subscores) and epidermal barrier disruption (TEWL), we performed Spearman correlation coefficient analysis presented as correlation plots in Fig 6, A and B, and Fig E4. The highest correlations with the inflammatory component of IASI (IASI-E) included IL-17/TNF-α–regulated genes, such as IL17F (r = 0.71, P = 9.54 × 10⁻⁶), IL36G (r = 0.62, P = .0002), KYNU (r = 0.57, P = .001), and KLK13 (r = 0.57, P = .001; Fig 6, A, and see Fig E5 and Table E11 in this article’s Online Repository at www.jacionline.org). Highly significant correlations were also noted between select markers and total IASI scores (IL17F and KLK13, P < .05) and TEWL levels (IL17F, EREG, IL36A, IL36B, IL36G, KLK13, and VNN3; all P < .01; Fig 6, A, and see Fig E5).

Fig 6.

A, Representative Spearman correlations between real-time/RT-PCR–derived mRNA expression of skin markers with IASI-E and TEWL. Dots represent ichthyosis subtypes: orange dots, CIE; green dots, EI; blue dots, LI; and purple dots, NS. B,Serum IL-17A levels in normal (N) and ichthyosis (ICH) serum represented as box plots. The mean of each group is shown as a large red dot, black dots are individual samples, horizontal black lines indicate medians, vertical black lines indicate values within 1.5 times the interquartile range of the median, and the top and bottom boundaries of the boxes indicate 75th and 25th percentiles, respectively. P values above horizontal lines denote significance among groups. C, Spearman correlation heat map of clinical (IASI-E, IASI total, IASI-S, pruritus, and TEWL), transcriptional, and serum immune markers. The black boxdemonstrates groups of tightly clustered markers/parameters. +P < .1, *P < .05, **P < .01, and ***P < .001.

We also assessed IL-17A levels in serum. IL-17A activation extended to the blood compartment, with highly significant increases in serum from patients with ichthyosis versus control subjects (P < .001; Fig 6, B, and see Fig E6 in this article’s Online Repository at www.jacionline.org) and the highest increases in patients with NS (P < .01), highlighting its high-grade immune activation.

A Spearman rank correlation coefficient heat map of all clinical parameters (IASI, IASI-E, IASI-S, pruritus, and TEWL) in patients with ichthyosis compared with skin mRNA and serum expression is presented in Fig 6, C(red, positive correlations; blue, negative correlations; color intensity reflects correlative strength). A black box demonstrates IL-17−associated markers in ichthyotic skin (KYNU, VNN1/3, IL36A/G, and IL17F) and serum IL-17A clustering tightly with disease severity (IASI-E), whereas epidermal dysfunction (TEWL) correlated most with EREG and IL36B mRNA. Individual correlations are shown in Table E12 in this article’s Online Repository at www.jacionline.org.

Discussion

Disrupted terminal differentiation, lipid homeostasis, increased TEWL, and associations with atopy have historically linked ichthyoses to AD.^{2, 3, 18, 27–41, 43, 50–52, 83–86} However, our recent profiling of ichthyotic skin using a limited gene-expression approach demonstrated enhanced T_H17/IL-23 response with marginal T_H2 skewing and lack of abnormalities in 3 differentiation proteins (FLG, LOR, and PPL), aligning it more closely to the psoriasis profile.⁵³

The present study is the first comprehensive genomic skin fingerprinting of the most common orphanichthyosis subtypes to date, with comparisons to tissue from healthy control subjects, patients with AD, and patients with psoriasis. Our data further establish a predominant T_H17/IL-23 and IL-36 signature underlying the common and individual ichthyosis phenotypes, expanding on our prior report⁵³ and profiling of single patients with erythrokeratodermia-cardiomyopathy (mutation in desmoplakin [DSP])⁸⁷ and ichthyosis Curth-Macklin (keratin 1 [KRT1]).⁸⁸ TNF-αantagonism is beneficial in patients with psoriasis, likely because of its effects on IL-17/TNF-α synergistically/additively induced genes rather than on TNF-α–regulated genes alone.^{89, 90} Here we show a striking increase in expression of IL-17/TNF-α synergistic genes, including those representing molecular hallmarks of psoriasis (KYNU, VNN3, IL36B/G, and EREG) across ichthyoses and particularly NS. Furthermore, the expression of these IL-17/TNF-α–regulated genes significantly correlated with sores on IASI-E, the inflammatory component of ichthyosis severity. Among these psoriasis classifiers are (1) KYNU, which is involved in tryptophan metabolism⁹¹; (2) VNN3, expression of which is increased by T_H17/T_H1 but not by T_H2 cytokines,^{92, 93} as suggested by its increased expression in patients with ichthyosis and psoriasis but not patients with AD; and (3) IL36A/IL36B/IL36G, proinflammatory cytokinescoregulated in keratinocytes by IL-17/TNF-α⁹⁴ that amplify IL-17 cytokine production^{80, 81} and are overexpressed in psoriatic skin,^{95, 96} discriminating it from AD skin.^{53, 80, 81, 97} Our study shows increased mRNA and protein expression of IL-36 family members and IL-36R and highly significant correlations with ichthyosis severity and TEWL, further suggesting IL-36/IL-36R involvement in the T_H17 axis in the ichthyoses, similar to psoriasis.^{44, 48, 49, 58, 59, 96} Our data also extend the prominent severity-associated T_H17 profile in ichthyosis to the blood compartment, supporting an accentuated systemic inflammatory phenotype, particularly in patients with NS.

Beyond the common ichthyosis phenotype, the present study is the first to highlight the unique genomic signatures of each ichthyosis variant. NS showed the greatest increases in inflammatory pathways and particularly those related to T_H17 versus other subtypes. In mouse NS models a few T_H17 markers showed increases,20, 60 and in 1 patient with NS, T_H17 markers were reduced with TNF-α antagonism and clinical improvement.⁴⁹ Patients with NS carry serine protease inhibitor, Kazal-type 5 mutations, leading to unopposed KLK5–protease-activated receptor 2–TSLP activation and induction of T_H2 and TNF-α, an identified gene target of KLK5–protease-activated receptor 2 signaling.^{46, 47, 98} Contrary to patients with NS, which has been linked classically to the atopic diathesis^{18, 46, 48, 50–52, 60, 86} and harbors T_H2 and pronounced T_H17 activation, T_H2 skewing was largely absent in other ichthyoses. This might be attributable to reciprocal regulation of the T_H17 and T_H2 axes^99–101 or might be secondary to KYNU metabolites, which were implicated in T_H2 downregulation.⁹¹ Hence T_H17 upregulation in patients with ichthyoses might be the outcome of complex interactions between different arms of adaptive immunity that extend to the systemic level, as evidenced in our report.

Ichthyoses have barrier defects,^{2, 5, 53, 102} as reflected by increased TEWL,^{2, 3, 5, 18, 28, 30, 31, 53, 102} and many forms have defective lamellar body transport and lipid homeostasis. Hence we comprehensively analyzed epidermal barrier genes, including those contributing to terminal differentiation, tight junctions (TJs), and lipid metabolism to dissect contributions of each component to impaired barrier function. A preserved differentiation profile was seen across ichthyoses, supporting our limited prior data and other past studies and possibly reflecting the relative lack of T_H2-mediated suppression of epidermal differentiation.^{36, 60, 103–105} CLDN1/CLDN8/CLDN23, which are crucial for TJ formation and highly suppressed in skin of patients with AD,^{64, 106} were variably downregulated in patients with ichthyoses, although less significantly than in those with AD and even psoriasis. IL-17 has been suggested to suppress TJs and adhesion molecules (including CLDN1/CLDN4 and occludin), perhaps explaining their reduced expression in patients with ichthyosis.¹⁰⁷

Our gene profiling, corroborated by Nile red staining, also demonstrates unique abnormalities in SC lipid structure among the ichthyoses. Although overall lipid measures (FA2H, FABP7, and ELOVL3) were suppressed in all patients with ichthyoses (except those with EI) compared with control subjects, the greatest downregulation was found in patients with LI, which was even more pronounced than in patients with AD and psoriasis. Our patients with LI all harbored mutations in both alleles of TGM1, which encodes transglutaminase 1.^{108, 109} Transglutaminase 1 cross-links cutaneous ω-hydroxyceramides to terminal differentiation proteins, playing an important role in CE assembly.^{110, 111} Thus TGM1 inactivity in patients with LI might drive the LI barrier abnormalities in addition to the underlying inflammation/immune activation, particularly of T_H17. Notably, decreased lipid levels were shown to upregulate psoriasis-like inflammation in mice (IL-17, IL-22, IL-23, S100As, and β-defensins),¹¹² and preadipocytes treated with IL-17 have reduced lipid-related gene expression,^{113, 114} highlighting the need to further investigate the role of IL-17 in epidermal lipid defects in patients with ichthyosis.

Although ichthyosis and psoriasis have different primary causes, our data highlight the similarities between these 2 diseases. In patients with psoriasis, T_H17 T cells become self-reactive to endogenous proteins,^{54, 55, 115–126} and the IL-17 produced by T cells, along with IL-26 and IL-29, trigger major genomic changes in epidermal keratinocytes, including induction of cytokines (ie, IL-36) that drive the typical epidermal hyperplasia.^{93, 95, 117, 118, 125, 127–129} Many other IL-17–induced products create feed-forward inflammatory responses, leading to chronic T-cell activation in the skin and sustained reactive responses of keratinocytes and other cells creating the psoriasis phenotype. IL-17 blockade with therapeutic antibodies resolves most of the clinical or molecular psoriasis phenotype.^{55, 120, 130, 131}

In contrast, the ichthyoses are caused by genetic defects primarily in keratinocyte structural genes or enzymes, leading to barrier abnormalities that are most often present at birth. In patients with ichthyosis, initial triggering of IL-17–dominated inflammatory responses stems presumably from exposure to microbes or other antigens through a disrupted barrier, which one predicts sustains immune activation by constant antigenic exposure. However, skin responses to IL-17 have the same potential to alter keratinocyte transcriptional/differentiation programs, and these can create feed-forward inflammatory loops, as in patients with psoriasis. We propose that the shared phenotypic features among ichthyosis subtypes result from primary barrier abnormalities, which in turn lead to cutaneous molecular and structural responses to IL-17–centered inflammation with hyperplasia and TJ/lipid changes. Given that so many patients with known genetic defects in the proteins/lipids that are critical to normal barrier function show shared IL-17/IL-23 skewing, it would be unlikely that the explanation is an additional primary underlying genetic abnormality of innate immunity. The extent to which IL-17–centered inflammation contributes to overall disease pathogenesis and clinical disease can only be determined by means of direct antagonism of IL-17 or upstream regulators. Recently, we reported 2 cases of syndromic ichthyoses from a primary defect in DSP(erythrokeratodermia-cardiomyopathy and severe dermatitis, multiple allergies, and metabolic wasting syndromes) treated successfully with IL-12/23p40 antagonism⁸⁷ with improved ichthyosis severity (both erythroderma and scaling) and TEWL, suggesting that T_H17/IL-23 activation could contribute to disease pathogenesis. Our data form the basis for an ongoing clinical trial ( NCT03041038) to test the therapeutic efficacy of IL-17A antagonism in patients with ichthyosis, which can better dissect the complex relationship of IL-17 and the underlying barrier defects.

Our study has several limitations. First, although this is the largest characterization of ichthyosis subtypes, we acknowledge the limited sample size because of disease rarity. Second, we could only screen for IL-17A in sera from patients with ichthyosis, suggesting the need for a more comprehensive investigation of systemic T_H17 activation in patients with ichthyosis.

In summary, ichthyotic skin and serum have strong T_H17 skewing, with an increase in markers coregulated by IL-17/TNF-α, preserved epidermal differentiation, and downregulation of lipid metabolism and TJ genes best resembling psoriasis. Our study offers new insights into the possible role of the IL-17 pathway in perpetuating the barrier defects of ichthyosis and advocates for testing of IL-17, IL-36,⁹⁶ and IL-23 antagonism in patients with these diseases. Such studies will not only help dissect disease pathogenesis but might also provide hope for better treatment for these devastating disorders.

Clinical implications.

Our report offers novel insights into the role of IL-17/IL-36 responses and unique barrier alterations in ichthyosis pathogenesis, advocating for directed immune antagonism in these patients. This comprehensive molecular phenotyping could direct much-needed targeted therapeutics for these diseases.

Abbreviations used

AD

Atopic dermatitis

ARCI

Autosomal recessive congenital ichthyosis

CCL17

Chemokine like 17

CE

Cornified envelope

CIE

Congenital ichthyosiform erythroderma

CLDN

Claudin

DEG

Differentially expressed gene

EI

Epidermolytic ichthyosis

EREG

Epiregulin

FCH

Fold change

FDR

False discovery rate

FLG

Filaggrin

IASI

Ichthyosis Area and Severity Index

IASI-E

Ichthyosis Area and Severity Index–erythema

IASI-S

Ichthyosis Area and Severity Index–scaling

IL-36R

IL-36 receptor

KLK

Kallikrein

KYNU

Kynureninase

LI

Lamellar ichthyosis

LOR

Loricrin

NS

Netherton syndrome

OASL

29-59-oligoadenylate synthetase-like protein

PPL

Periplakin

SC

Stratum corneum

TEWL

Transepidermal water loss

TJ

Tight junction

TSLP

Thymic stromal lymphopoietin

VNN3

Vanin 3