Intrinsic Abnormalities of Keratinocytes Initiate Skin Inflammation through the IL-23/T17 Axis in a MALT1-Dependent Manner

This article is featured in Top Reads, p.665

Psoriasis is a common, chronic skin disorder histologically characterized by considerable thickening of the epidermis due to the hyperproliferation of keratinocytes accompanied by a dense dermal infiltration of immune cells (1, 2). Although the etiology of psoriasis is complicated, significant progress has been made to understand the pathogenesis of this disease by using imiquimod- (IMQ) (3) or IL-23– (4) induced psoriatic mouse models, which proposed the so-called “IL-23/T17 axis” (5). Currently, it is believed that sustained activation of skin-resident dendritic cells (DCs) promotes the production of type 17 cytokines, such as IL-17 and IL-22, by T cells via secretion of IL-23. IL-17 then activates epidermal keratinocytes to produce CCL20, which can recruit CCR6⁺ cells such as DCs and T cells to shape a feed-forward loop and propagate psoriatic inflammation (6, 7). In other contexts, IL-23 drives the IL-22 production by group 3 innate lymphoid cells (ILC3s) to initiate a local skin inflammatory circuit, in addition to the maintenance of T cells (8). Based on this notion, keratinocytes are considered mainly an executor of the effector cytokines, such as IL-17 and IL-22, not a driver in the pathogenesis of psoriasis.

The debate over which cells trigger the initiation of psoriasis has been a controversial and hot topic (9). Attention turns back to the keratinocytes as a potential site for the origin of psoriasis because genomic-wide association studies have identified many susceptibility genes that are involved in innate immunity, indicating a pivotal role of innate signaling pathways in psoriasis developmental process (10, 11). Therefore, the question, “Which actually comes first in psoriasis: dysregulation of the immune cell biology or keratinocytes?” becomes a lively debate (10, 12). Furthermore, some familial cases imply that intrinsic abnormalities of keratinocytes might be responsible for psoriasis in genetically predisposed individuals (13). The epidermis provides a highly durable barrier between the host and external environment, which is capable of sensing pathogens and producing multiple immunomodulatory factors including cytokines and chemokines to modulate the innate and adaptive immune response (14, 15). However, the distinctive alterations of epidermal keratinocytes enriched in the psoriatic mouse model and the exact sequence of events that lead to the onset of psoriasis are still missing.

Identification of Card14 as a susceptible locus has made a great breakthrough that keratinocyte might play a critical role in the development of psoriasis (11, 13). CARD14 protein is mainly expressed in the skin and some mucosal tissues, such as small intestine, colon, and cecum (16). Both in healthy and lesioned skin, CARD14 expression is restricted to the keratinocytes, not in the immune cells (13, 16, 17). Over 40 missense mutations in the coding sequence of Card14 have been discovered, which are associated with different types of psoriatic skin diseases, including psoriasis vulgaris, generalized pustular psoriasis, pityriasis rubra pilaris, and atopic dermatitis (18). Recent studies have determined how gain-of-function mutations of Card14 induce psoriasis-like skin inflammation in vivo (16, 17, 19–21). Heterozygous and homozygous mice harboring Card14^E138A or Card14^ΔE138 could spontaneously develop extensive skin and systemic inflammation, leading to enhanced activation of the IL-23/T17 axis. In addition to the heterozygous Card14^E138A/+ mice, our group also had accidentally generated a heterozygous Card14^ΔQ136/+ mouse strain that exhibited a relatively weaker psoriasis-like phenotype than Card14^E138A/+ mice (17). These two mouse strains provided bona fide models that more closely resemble psoriatic characteristics for severe and mild human patients. CARD14 is a scaffold protein that assembles B cell lymphoma 10 (BCL10) and MALT lymphoma translocation protein 1 (MALT1), called CBM complex, to activate downstream signaling cascades (22, 23). Nevertheless, the biological functions of CBM complex in psoriasis development are still undetermined.

Biologics targeting IL-17A contribute high therapeutic effects upon systemic psoriasis treatment, but we previously observed only a partial reduction of psoriatic phenotype in the CARD14-E138A-mutated psoriasis model (17). To investigate the function of IL-17A in the CARD14-mutated psoriasis model, we bred Card14^E138A/+ mice with Il17a^−/− mice and found that depletion of IL-17A could not alleviate the skin inflammation, whereas IL-23 deficiency resulted in a profound response. To address the keratinocyte-intrinsic effects for psoriasis development, we conditionally depleted CARD14-interacting protein MALT1 or BCL10 in keratinocytes in Card14^E138A/+ mice and observed that the skin disease and systemic inflammation were completely dampened, which indicates that the initiation phase of psoriasis indeed occurs in keratinocytes. Interestingly, the alleviated skin inflammation by selective inactivation of MALT proteolytic activity in keratinocytes suggests a promising therapeutic target for early treatment of psoriasis.

All mice used in this study were on the C57BL/6 background. Card14^E138A/+, Card14^ΔQ136/+, Rag1^−/−, Il23^−/−, Bcl10^fl/fl, Malt1^fl/fl, Malt1^C472A/C472A, K14^Cre as well as wild-type mice were housed in specific-pathogen-free facilities at Tsinghua University. Unless otherwise stated, in all experiments, age- and gender-matched mice were used at 4–12 wk of age. The animal experiments were carried out according to the institutional guidelines and were approved by the Institutional Animal Care and Use Committees at Tsinghua University.

HaCaT cells were cultured in DMEM (Life Technologies) supplemented with 10% FBS and 1% FBS in terms of starvation. Primary keratinocytes were isolated from newborn mice and cultured as described in “In vitro keratinocytes isolation and cell culture” below. All cells were cultured at 37°C and 5% CO₂.

Mice were subjected to a daily topical treatment for five consecutive days at a dose of 20 mg of IMQ per ear. Control mice were treated similarly without IMQ. Generally, ear thickness and weight were measured with an engineer’s caliper every day. Tissue samples were taken 1 d later after the last treatment.

To isolate primary keratinocytes from newborn mice, the skin of the whole body was cut into small pieces and incubated in 1 mg/ml Dispase II (Roche) overnight at 4°C. The next day, the epidermis was peeled from skin and digested with 0.05% trypsin-EDTA (Life Technologies) for 10 min at 37°C and filtered with a 100-μM strainer. The cell suspension was neutralized with 1% calcium-free PBS buffer and then centrifuged at 900 rpm for 10 min. The cells were cultured in Medium 154CF (Life Technologies) supplemented with 0.05 mM calcium chloride with HKGS (Life Technologies).

For lymph nodes and spleen sample, organs were mechanically dissociated to obtain single-cell suspensions, and splenic cells were treated with lysis buffer to deplete RBCs. For total ear sample, animals were euthanized and ears were collected and digested with 3 mg/ml Collagenase D (Roche) and 10 mM DNase I (Roche) in serum-free RPMI 1640 medium for 60 min at 37°C, which then was neutralized by adding 10 ml of PBS supplemented with 1% of FBS. Tissues were mechanically disrupted and filtered with a 100-μM strainer. Viability dye (eBioscience) was used to exclude dead cells. For surface staining, cells were stained with indicated surface Ag Abs in FACS buffer (PBS supplemented with 1% FBS) on ice for 30 min. For intracellular cytokine staining, cells were fixed and permeabilized with BD Cytofix/Cytoperm solution buffer followed by staining with indicated cytokines Abs in Perm/Wash buffer (BD Biosciences).

Tissue biopsy specimens were snap-frozen in liquid nitrogen. Frozen tissues were ground in liquid nitrogen and dissolved directly in TRIzol (Invitrogen). Total RNA was extracted and quantified with a NanoDrop. RNA reverse transcription was performed using RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. An ABI 7500 Real-Time PCR system (Applied Biosystems) and Power SYBR Green PCR Master Mix (Genestar) were used for quantitative PCR. Results were normalized to Gapdh and calculated by the change in cycling threshold (2^−ΔΔCt) method. Melting curves were confirmed to ensure amplification of a single product. Primers used in this study are found in Ref. 17.

For histopathology analysis, murine ear and tail tissues were fixed with 4% formaldehyde overnight and were then dehydrated and embedded in paraffin. The 5-μm tissue sections were stained with H&E according to standard procedures. Stained tissue sections were scanned using Pannoramic scan (3DHISTECH). The epithelial thickness was quantified in at least five independent regions with the tools of measurement.

For immunohistochemistry and immunofluorescence, 4% formaldehyde-fixed paraffin-embedded dorsal skin, ear, and tail sections were deparaffinized, and Ab was retrieved with sodium citrate, then stained with Ki67 (Abcam), keratin 14 (K14) (Abcam), K10 (Abcam), Loricrin (Abcam), CD45 (Ceville), CD11c (Ceville), and Ly6G (Ceville). The images were captured with a Pannoramic scan (3DHISTECH) or Axio Scan (Zeiss).

For MALT1 cleavage ability, mouse primary keratinocytes were starved overnight without HKGS then stimulated with/without IL-17A (500 ng/ml) or PMA (50 ng/ml)/ionomycin (1 μM) for 2.5 h with or without MG-132 (5 μM) for 15 min. Cells were collected with lysis buffer (150 mM NaCl, 50 mM HEPES, pH 7.4, 1 mM EDTA, 1% Nonidet P-40, and protease inhibitors mixtures). Total cell lysates were subjected to SDS-PAGE and then blotted using the indicated Abs. For human primary keratinocytes, the Dox-induced flag-tagged CARD14 expressed HaCaT stable cell line was constructed. The stable cell line was stimulated by Dox at the indicated time and then was lysed and incubated with anti-flag Abs. The precipitates were then detected with indicated Abs. The Abs used in this assay include RelB (4954; Cell Signaling Technology), CYLD (sc-74435; Santa Cruz) and tubulin (Santa Cruz), second Ab (Easybio), and flag Ab (Abmart).

The ear thickness was measured by vernier caliper. All the statistical analysis was performed using Prism 7 (GraphPad Software). The p values were determined by unpaired two-tailed t tests or two-way ANOVA. Asterisk coding is indicated as *p < 0.05, **p < 0.01, ***p < 0.005, and ****p < 0.001. Statistical parameters, including number of biological replicates and repeat experiments, data dispersion, and precision measures (mean and SEM), are reported in the figure legends.

For data and materials availability, further information and requests for resources and reagents should be directed to and will be fulfilled by corresponding author X.L.

Selectively targeting the IL-17A has resulted in systemically efficient therapies for psoriasis treatment, as a high proportion of patients are getting improvements. Clinicians, however, have experienced groups of patients that do not respond to these biologic therapies (24). To determine the role of IL-17A in the pathogenesis of CARD14-mutation-induced psoriasis, we deleted IL-17A in Card14^E138A/+ heterozygous mice who spontaneously develop severe psoriasis-like skin inflammation at an early age by crossing Il17a^−/− mice to Card14^E138A/+ mice. Surprisingly, Card14^E138A/+Il17a^-/- mice exhibited hardly ameliorated pathology with comparable histological features of epidermal acanthosis as well as similar epidermal thickness in both ears and tails (Fig. 1A), which probably reflects redundant function among various type 17 effector cytokines, such as IL-17F and IL-22.

In psoriatic plaques, activated DCs drive the polarization and clonal expansion of T17 and T22 cells to produce considerable amounts of IL-17–related cytokines, such as IL-17A, IL-17F, and IL-22 (25). After elimination of T cells by crossing with Rag1^−/− mice, heterozygous Card14^ΔQ136/+ mice, generated as a relatively mild psoriatic model, displayed parallel epidermal thickness compared with the age- and sex-matched Rag1^+/− littermates at 4 wk old (17). But when we examined the dermatological manifestations of Card14^ΔQ136/+Rag1^−/− mice on 12 wk of age, depletion of T cells did not abrogate the skin inflammation characterized by the distinct hyperplasia in comparison with Rag1^+/− littermates (Supplemental Fig. 1A). Of note, the psoriatic phenotypes of Card14^ΔQ136/+ mice gradually increase as they grow up. Until ∼12 wk, their ears begin to appear distinctly red and swollen. To further evaluate the role of T cells in general, we obtained the Card14^E138A/+Rag1^−/− mice by breeding Rag1^−/− mice to Card14^E138A/+ mice. Consistently, the ear and tail thicknesses of Card14^E138A/+Rag1^−/− mice were merely partially reduced even by the age of 4 wk (Fig. 1B, Supplemental Fig. 1B). The expression levels of type 17–related effectors, such as Il17a and Il17f, relatively decreased in Rag1-deficient mice, whereas the mRNA levels of Cxcl1, Il23, and Il22 showed no significant difference and increased Ccl20 expression level with or without T cells (Fig. 1C), which suggests the crucial roles of other immune cells besides T cells in the pathogenesis of psoriasis, like DCs, macrophages, and ILC3 cells. Similar results were also observed in the tails of these mice (Supplemental Fig. 1C). Thus, T cells are involved but not indispensable in severe psoriasis development.

With the discovery of IL-23 as the “master regulator” of T17 cells, several Abs targeting the specific IL-23p19 subunit have been investigated for the clinical treatment of psoriasis (26). Both IL-17A– and/or IL-22–producing T cells and innate lymphoid cells in psoriatic skin and blood are activated in response to IL-23, correlated with disease severity. To parse the function of this cytokine in CARD14-mutation-induced psoriasis in vivo, we depleted the IL-23 in Card14^E138A/+ mice by breeding Il23^−/− mice with Card14^E138A/+ mice. Interestingly, the psoriasis-like lesions featuring redness, scales, and crust formation disappeared in Card14^E138A/+Il23^−/− mice, although the ear thickness as an important indicator of skin inflammation was increased when compared with the healthy controls (Fig. 2A). Histological analysis revealed a reduced cellular infiltrate in the absence of IL-23 (Fig. 2B). We then investigated the skin-infiltrating cells by flow cytometry and found a striking decrease of neutrophils in Card14^E138A/+Il23^−/− mice (Fig. 2C). Meanwhile, IL-23 deficiency dramatically reduced the infiltration of IL-17A–producing cells (Fig. 2D). We next examined the expression level of IL-17–associated cytokines aroused downstream of IL-23. As expected, the expression of Il17a, Il17f, Il22, and S100a8 was totally abolished with the depletion of Il23 (Fig. 2E). Thus, blockade of IL-23 disrupts the IL-23/T17 signaling networks, which results in more sufficient treatment for the CARD14-associated papulosquamous eruption (CAPE). Notably, not only was the expression level of Ccl20 enhanced in the Card14^E138A/+Il23^−/− mice, even much higher than Card14^E138A/+Il23^+/− mice, but Cxcl1 was also remarkably induced in the Card14^E138A/+Il23^−/− mice (Fig. 2E), indicating an exclusive role of keratinocytes under the CARD14 mutation-dependent psoriatic condition.

It was previously reported that inhibition of MALT1 proteolytic activity reduces the CARD14-E138A-induced production of inflammatory effectors in human primary keratinocytes (27). Although it is well documented that CARD14 interacts with BCL10 and MALT1, termed “CBM complex,” for the downstream signal transduction, the biological significance of this complex in vivo is still unclear. For this purpose, we first specifically knocked out MALT1 in keratinocytes in Card14^E138A/+ mice by crossing to Malt1^fl/fl mice and K14-cre mice that express Cre recombinase under the control of the K14 promoter, which referred to Card14^E138A/+K14^CreMalt1^fl/fl mice (Fig. 3A). Keratinocytes in the absence of MALT1 clearly restrained the development of psoriasis-like symptoms with unaffected skin architecture of the ears and tails, also with similar ear thickness to Malt1^fl/fl mice acting as wild-type controls (Fig. 3A). Psoriasis-like aspects of ears and tail skin reflected by the histological analysis were comparable to littermate controls (Fig. 3B, Supplemental Fig. 2A). Thus, MALT1 is indispensable in the pathogenesis of CAPE.

Next, we stained the tail tissue sections and found not only a largely decreased expression of K14, K10, and Ki-67 but also reduced CD11c⁺ and Ly6G⁺ infiltrates in the mice with the specific absence of MALT1 compared with those of Card14^E138A/+ mice (Fig.3C, Supplemental Fig. 2B). Consistently, Card14^E138A/+K14^CreMalt1^fl/fl mice exhibited profoundly less neutrophil infiltration than Card14^E138A/+ mice (Fig. 3D). The expression of those proinflammatory molecular modulators, such as Tnfα, Il1β, Il6, and Ccl20, as well as the antimicrobial peptides were all completely blocked in conditionally MALT1-deleted animals (Fig. 3E). We then crossed Card14^E138A/+ mice into the mice containing a floxed Bcl10 allele to obtain BCL10 deficiency in keratinocytes. Interestingly, we also observed the normal skin conditions after depletion of BCL10 in the epidermal cells of Card14^E138A/+ mice, including ear swelling and histological examination (Supplemental Fig. 2C, 2D). Altogether, these results provide strong genetic evidence that development of the skin phenotype in Card14^E138A/+ mice fully depends on CBM complex–mediated signaling in keratinocytes.

We next determined whether the MALT1-mediated signaling pathway within keratinocytes was definitely required for initiating the IL-23/T17 axis in the skin. Card14^E138A/+K14^CreMalt1^fl/fl mice displayed the same percentage of infiltrating IL-17A–producing cells in ears as Malt1^fl/fl mice, which was remarkably decreased relative to that in Card14^E138A/+ mice (Fig. 4A). Indeed, the expression of IL-23/T17 axis–associated inflammatory cytokines Il23 and Il17a was completely suppressed among the conditionally MALT1-deleted group (Fig. 4B). To further evaluate the important role of epidermal MALT1 psoriatic inflammation, we then investigated the role of MALT1 in keratinocytes using a well-established model of psoriasiform inflammation in which IMQ is administered to the skin (3). Mice treated with IMQ showed a slight weight loss in the early days but recovered to the original level at the end (Fig. 4C). Repetitive topical application of IMQ caused a severe skin phenotype with marked ear swelling and epidermal hyperplasia in Malt1^fl/fl mice, whereas only weak cutaneous changes were observed in K14^CreMalt1^fl/fl mice (Fig. 4D, 4E). Subsequent analysis of the cellular infiltration revealed dramatically decreased neutrophils in a keratinocyte-specific MALT1-dependent manner (Fig. 4F). Collectively, these observations have presented solid evidence that the CBM complex orchestrates keratinocytes to drive skin inflammation through the IL-23/T17 axis.

Psoriasis is a systemic inflammatory disease with several comorbid conditions. Others reported that the size of spleen and skin draining lymph nodes was enlarged in an IMQ-induced acute psoriasis–like model (3), and we also found the increased size of the axillary, inguinal, cervical, and mesenteric lymph nodes in Card14^E138A/+ mice but not enlargement of spleen and thymus. However, the size of these immune organs was restricted in Card14^E138A/+K14^CreMalt1^fl/fl mice (Supplemental Fig. 3A). Moreover, the proportion of neutrophils in the skin draining lymph nodes and spleen was also reduced (Supplemental Fig. 3B, 3C), indicating that systemic inflammation was the secondary response of psoriasis and once the skin disease was cured, the comorbidities would disappear simultaneously.

It is well accepted that MALT1 not only acts as a scaffold protein, providing an assembly platform for integrating signals from cell surface, but also possesses proteolytic activity to cleave a limited repertoire of proteins resulting in fine-tuned immune responses (28). Therapeutic targeting of MALT1 proteolytic activity by small molecule inhibitors has shown promising effects for inflammatory diseases and cancer (28). Based on the prominent function of MALT1, we hypothesized that inactivation of its protease activity might dampen the skin diseases. Therefore, we crossed Card14^E138A/+ mice with Malt1^C472A/C472A mice that ubiquitously inhibited the proteolytic activity of MALT1. Histological analysis represented reduced dermal infiltrate and epidermal thickening in Card14^E138A/+Malt1^C472A/C472A mice (Supplemental Fig. 4A, 4B). In line with this, mutation of MALT1 protease activity attenuated the IMQ-induced ear swelling (Fig. 5A). H&E staining of the ear sections showed that the epidermis was less thickened in Malt1^C472A/C472A mice (Fig. 5B). Thus, MALT1 protease activity makes a great contribution to the psoriasis development.

The important role of MALT1 protease in psoriasis development was currently reported (21), but considering the early lethality of MALT1 protease-dead mutant mice due to multiorgan inflammation and autoimmunity (29), systemic inhibition of MALT1 may lead to some undesired side effects for the treatment of disease. We assumed a particular inhibition of MALT1 protease activity confined to the epidermis. Upon different stimulation, we detected the strong cleavage of RelB and CYLD, two MALT1 substrates, in the lysates of primary keratinocytes from wild-type animals (Fig. 5C). Because the cleaved fragments of RelB and CYLD may be degraded by proteasome complex, we treated these cells with proteasome inhibitor MG-132 to detect the MALT1 activation in these cells. In addition, we also noticed the cleaved RelB and CYLD in the cultured primary keratinocytes derived from Card14^E138A/+ newborn pups without treatment (Fig. 5D). Furthermore, once expressed by doxycycline (Dox), the human HaCaT cells with CARD14-E138A protein indeed triggered the enhanced cleavage of RelB and CYLD (Fig. 5E). Accordingly, MALT1 protease activity is greatly induced in the psoriatic keratinocytes.

Based on the same disease phenotype of Malt1^C472A/C472A mice and Malt1^C472A/- mice, but the healthy phenotype of Malt1^C472A/+ mice (29), we intercrossed Card14^E138A/+ mice with K14^Cre mice, Malt1^fl/fl mice, and Malt1^C472A/C472A mice to generate Card14^E138A/+K14^CreMalt1^fl/C472A mice, in which only protease-dead MALT1 could be expressed in keratinocytes, allowing us to study the biological role of MALT1 protease activity within keratinocytes in vivo along with intact scaffold function. Selective mutation of MALT1 protease activity in keratinocytes abated ear swelling, epidermal hyperplasia, and inflammation in ears and tails at an early age (Fig. 6A, 6B). We analyzed the total ear cell suspensions by flow cytometry and found a significantly reduced infiltration of neutrophils (Fig. 6C). Mutated MALT1 gave rise to abate the expression levels of not only the proinflammatory cytokines but also the IL-23/T17–related genes (Fig. 6D). Systemic inflammation characterized by the enlargement of lymph nodes as well as the infiltration of neutrophils into spleen and skin draining lymph nodes showed a decrease in a MALT1 protease-dependent manner (Supplemental Fig. 4C–E). It should be mentioned that specific inhibition of MALT1 in the epidermal cells demonstrated a slight improvement in the older Card14^E138A/+ mice with extremely severe disease (Supplemental Fig. 4F).

Considering the extraordinary severity of Card14^E138A/+ mice, we further selectively inactivated MALT1 protease activity in the Card14^ΔQ136/+ mice that developed a relatively weaker psoriasis-like phenotype than Card14^E138A/+ mice (17). Interestingly, the Card14^ΔQ136/+K14^CreMalt1^fl/C472A mice showed similar ear thickness to K14^CreMalt1^fl/+ control animals, which was thinner than Card14^ΔQ136/+K14^CreMalt1^fl/+ psoriatic mice (Fig. 6E). Furthermore, daily topical application of IMQ cream to the skin of the ears for 5 d induced considerable scaling and plumping of the skin in control animals; however, K14^CreMalt1^fl/C472A mice exhibited two weaker psoriatic manifestations (Fig. 6F). The percentage of neutrophils was significantly reduced in IMQ-treated K14^CreMalt1^fl/C472A mice compared with Malt1^fl/C472A control (Fig. 6G). Taken together, these data demonstrated that MALT1 proteolytic activity in keratinocytes provided a critical contribution to the development of psoriatic skin inflammation and could serve as a promising therapeutic target for topical treatment of psoriasis.

CARD14 mutations have been independently reported to be associated with psoriasis, pityriasis rubra pilaris, and psoriasis arthritis, which refers to CAPE (13, 30, 31). By making full use of the CARD14-E138A mutant mouse model, we found that genetic deletion of Il23 resulted in much better limitation of the psoriatic skin inflammation than elimination of IL-17A, which completely blocked the IL-23/T17 axis, indicating that other IL-23–regulating cytokines, such as IL-17F or IL-22, might play redundant roles in the pathogenesis of psoriasis. Interestingly, specifically knocking out MALT1 in keratinocytes in both Card14^E138A/+ mice and IMQ-treated mice pronouncedly abrogated the development of psoriasis-like symptoms. Therefore, our data indicate that psoriasis indeed originates from intrinsic abnormalities of keratinocytes that secrete proinflammatory cytokines and chemokines to give rise to intricate communication networks in innate and adaptive immune responses.

Several IL-23 and Il-17 antagonists have been approved for the treatment of psoriatic disease based on their high efficacy and tolerability (6). However, it has been noticed that subsets of psoriasis patients with recalcitrant disease do not respond well to these biologic therapies, causing a long-term cycle of mediation error for patients and clinicians. We found that targeting IL-17A was insufficient to restrain the skin inflammation in CARD14-mutated mice, suggesting that when primary mutations occur in keratinocytes, these types of psoriatic inflammation may not respond well to biologic therapies using IL-17 antagonists. However, polymorphisms in CARD14 increased responsiveness to anti–TNF-α treatment, including adalimumab, etanercept, and infliximab (32). Consistent with this, CARD14-E138A–induced skin inflammation was ameliorated by TNF blocking in vivo (16). Thus, although CARD14 mutations are found in a small minority of psoriasis patients, and a much smaller proportion than the number of IL-17 nonresponders, it remains an important factor for predicting response to biologic therapies in the clinic, which needs more studies in the future.

Animal models provide a convenient survey of the extensive and complex mechanisms in the pathogenesis of diseases. It is well accepted that psoriasis is a T cell–driven inflammation (33, 34). However, when we eliminated T cells from Card14^E138A/+ mice, these mice still had severe skin disease, and the psoriatic phenotypes were barely alleviated. Similar results were observed in the mice that conditionally expressed of the human CARD14-E138A mutation following tamoxifen injection (16). In addition, another study has shown that IL-17– and IL-22–producing ILC3s may directly contribute to the pathogenesis of psoriasis-like disease (8). Therefore, T cell–independent mechanisms may also play important roles to induce psoriatic inflammation. Investigating these mechanisms may reveal novel therapeutic targets for the treatment of chronic skin disorders.

Small molecule drugs may provide the alternatives to the use of biologics with less costly long-term management. Our studies showed that Card14^E138A/+ mice with selectively inactivated MALT1 protease activity in keratinocytes exhibited abated ear swelling, epidermal hyperplasia, and inflammation in ears, backs, and tails. Particularly, when we abolished the MALT1-catalytic function in keratinocytes of Card14^ΔQ136/+ mice, which developed a mild manifestation of skin disease, the skin and systemic inflammation thoroughly disappeared. These results highlight the importance of MALT protease-mediated epithelial signaling pathways in psoriatic inflammation, suggesting that MALT1 protease activity may serve as a potential target for topical treatment of psoriasis.

We thank Dr. Chen Dong for kindly providing us reagents and Il23^−/− and Il17a^−/− mice. We thank Dr. Gang Ma from Shanghai Jiao Tong University for kindly sharing K14-cre mice.

The authors have no financial conflicts of interest.