E3 ubiquitin ligase NEDD4L negatively regulates keratinocyte hyperplasia by promoting GP130 degradation

Abstract

Psoriasis is mainly characterized by abnormal hyperplasia of keratinocytes and immune cells infiltrating into the dermis and epidermis. Neural precursor cell expressed developmentally downregulated 4‐like (NEDD4L) is a highly conserved HECT type E3 ligase that plays an important role in regulating physiological and pathological processes. Here, we identify NEDD4L as a negative regulator of psoriasis. Nedd4l significantly inhibits imiquimod (IMQ)‐induced skin hyperplasia, and this effect is attributed to the inhibitory effect of NEDD4L on IL‐6/GP130 signaling in keratinocytes. Mechanistically, NEDD4L directly interacts with GP130 and mediates its Lys‐27‐linked ubiquitination and proteasomal degradation. Moreover, the expression of NEDD4L is downregulated in the epidermis from IMQ‐treated mice and psoriasis patients and negatively correlates with the protein levels of GP130 and p‐STAT3 in clinical samples. Collectively, we uncover an inhibitory role of NEDD4L in the pathogenesis of psoriasis and suggest a new therapeutic strategy for the treatment of psoriasis.

NEDD4L functions as a negative regulator of IL‐6 signaling by inhibiting GP130 expression in keratinocytes. As NEDD4L is downregulated in the epidermis of psoriasis patients, increasing the levels of NEDD4L might be a potential therapeutic strategy for psoriasis.

Introduction

Psoriasis is a chronic inflammatory disorder characterized by epidermal and dermal leukocyte infiltration and keratinocyte hyperplasia (Schon & Boehncke, 2005). Several insults, such as drugs and microorganisms from the environment, directly activate skin‐resident dendritic cells (DCs) through pattern recognition receptors (PPRs) (Becher & Pantelyushin, 2012). Stress‐sensitized DCs produce IL‐23, which is defined as the most important cytokine that initiates psoriasis by activating γδ T cells or ILCs to produce IL‐17, TNF‐α, and IL‐22 (Becher & Pantelyushin, 2012). IL‐22 mainly promotes keratinocyte proliferation and antimicrobial peptide production, whereas IL‐17 and TNF‐α upregulate adhesion molecules and induce chemokines, which attract more inflammatory cells to the lesional skin and form a “self‐amplifying‐like inflammation loop” to aggravate the disease (Becher & Pantelyushin, 2012).

IL‐6 is a pleiotropic cytokine that mediates cell growth and differentiation. When IL‐6 binds to its receptor (IL‐6R), the IL‐6 and IL‐6R complex heterodimerizes with two molecules of GP130 (IL‐6 signal transducer, IL‐6ST), leading to the phosphorylation of GP130 and subsequently activating Janus kinase (JAK) and the transcription factor STAT3, which promotes the expression of genes that accelerate proliferation and inhibit apoptosis of the cells (Garbers et al, 2018). Many mucous and skin disorders, such as psoriasis (Galadari & Sheriff, 2005), systemic lupus erythematosus (Wallace et al, 2017), and cutaneous SSC (Du et al, 2010), are accompanied by abnormally elevated IL‐6 expression. Therapeutic agents targeting the IL‐6 axis were effective in treating psoriatic arthritis, rheumatoid arthritis, multiple myeloma, and ongoing Crohn’s disease (Kang et al, 2019). Recombinant soluble GP130Fc protein inhibited colitis with similar efficacy to neutralizing anti‐IL‐6R or anti‐TNF antibodies in mice (Atreya et al, 2000).

Neural precursor cell expressed developmentally downregulated gene 4‐like (NEDD4L) is a highly conserved HECT type E3 ligase regulating the expression of many membrane proteins in epithelial cells (Manning & Kumar, 2018). NEDD4L mediates ubiquitination of the epithelial Na/K/2CI co‐transporter NKCCI/SLCI2A2 and inhibits cell surface expression of epithelial NKCCI/SLCI2A2 in epithelia (Jiang et al, 2017). NEDD4L also limits TGF‐β signal transduction by mediating Smad2/3 polyubiquitination and degradation (Gao et al, 2009). NEDD4L is known to act as a tumor suppressor by ubiquitination of proteins important for cell proliferation and plays an important role in the initiation/progression in the pancreatic, breast, and colorectal cancer (Frampton et al, 2014; Guarnieri et al, 2018; Novellasdemunt et al, 2020). However, the pathological role of NEDD4L on skin diseases, such as psoriasis, remains largely unknown.

In this study, we demonstrated that NEDD4L functions as a negative regulator in epidermal hyperplasia. Nedd4l knockout promoted imiquimod (IMQ)‐induced epidermal proliferation in mice. Ablation or depletion of NEDD4L enhanced the activation of IL‐6/GP130/STAT3 signaling and the resulting proliferation of keratinocytes. Both IL‐6 and GP130 neutralization antibodies abolished the effect of NEDD4L deficiency on the IMQ‐induced epidermal hyperplasia in mice. Mechanistically, NEDD4L catalyzes K27‐linked polyubiquitination of GP130 and promotes GP130 degradation, leading to downregulated IL‐6/STAT3 signaling. Accordingly, decreased NEDD4L expression with enhanced protein expression of GP130 and phosphorylated STAT3 was observed in the epidermis from the clinical sample.

Results

Downregulated NEDD4L expression in the epidermis of IMQ‐treated mice and psoriasis patients

To explore the genes involved in the development of psoriasis, we analyzed the gene expression profiles in the epidermis of IMQ‐treated skin by RNA sequencing. As shown in Fig EV1A, the expression of a HECT ligase subfamily, which has not been reported to be involved in the development of psoriasis, was suppressed after supplementation with IMQ. Among the molecules, Wwp2 and Nedd4l were the most significantly suppressed genes. By consulting the deposited GEO datasets (GEO accession nos. GES13355 and GSE14905), we compared the expression of the genes in normal skin and skin lesions from active psoriasis patients. As shown in Fig EV1B and C, the expression of NEDD4L was significantly decreased in the psoriasis lesions compared with the normal skin and uninvolved skin. However, the expression of Wwp2 was not affected in the psoriasis lesions (Fig EV1D and E). The decrease in NEDD4L mRNA and protein levels in the epidermis from the IMQ‐treated mice was further confirmed by real‐time PCR analysis (Fig EV1F) and immunohistochemistry (IHC) analysis (Fig 1A), respectively. Consistently, the protein expression level of NEDD4L was significantly reduced in the epidermis from the psoriasis patients compared with the healthy controls (Fig 1B). Taken together, these results suggested that the expression of NEDD4L was downregulated in lesions from active psoriasis in humans and in experimental psoriatic mice, which indicated that NEDD4L may be involved in the development of psoriasis.

Figure EV1. Downregulated NEDD4L expression in the epidermis from IMQ‐treated mice and psoriasis patients.

• A
  Six‐ to eight‐week‐old C57BL/6 mice were treated with or without IMQ daily for 6 consecutive days. The epidermis was harvested, and the mRNA were extracted and subjected to RNA sequencing. (control, n = 3, IMQ, n = 3). Red star indicates the gene of interest.
• B–E
  The mRNA expression level of NEDD4L (B, C) and Wwp2 (D, E) in the skins from normal, uninvolved, and psoriasis patients were analyzed in Gene Expression Omnibus database (GEO). FPKM: Fragments per Kilobase Million.
• F
  Real‐time PCR analysis of Nedd4l mRNA expression level in the epidermis of back skins from mice treated with IMQ. (control, n = 5, IMQ, n = 5).
• G–I
  NHEKs were treated with IMQ for 0, 3, 6, and 12 h, and the mRNA expression levels of NEDD4L, EZH2, and EZH1 were detected by real‐time PCR. n = 3 technical replicates.
• J
  The protein expression levels of NEDD4L, EZH2, and EZH1 in IMQ‐treated NHEK were detected by Western blot (WB).
• K
  NHEK cells were transfected with EZH1‐specific small interfering RNA (siRNA) (siEZH1) or negative control siRNA (siNC), and the protein expression levels of NEDD4L, EZH2, and EZH1 were detected by WB after IMQ stimulation.
• L
  The schematic representation of NEDD4L promotor region. The thick truncated lines mark the regions covered by primer sets of interest. TSS: Transcription start site.
• M, N
  Mouse primary keratinocytes and NHEKs were treated with IMQ for 6 h, and ChIP assay was conducted to analyze the local enrichment of histone H3K4me3, H3K9me3, and H3K36me3 in specific regions of the NEDD4L promotor in mouse primary keratinocyte (M) and NHEKs (N). n = 3 technical replicates.

Data information: In heat maps, color scales represent the values of log₂‐tranformed fold changes. Each symbol represents one patient (B–E) or one mouse (F). Data are shown as mean ± s.e.m. and are representative of three independent experiments (F–K) or two independent experiments (M, N). Significant difference was calculated using two‐tailed Student’s t‐test (B–I, M, N). *P < 0.05, **P < 0.05, ***P < 0.05. NS, no significance.

Source data are available online for this figure.

Figure 1. Downregulated NEDD4L expression in the epidermis from IMQ‐treated mice and psoriasis patients.

• A, B
  Representative images of the expression of NEDD4L in the skins from IMQ‐treated mice (A) or psoriasis patients (B) as detected by IHC on the left. The NEDD4L expression level was calculated by multiplying the staining intensity score and the extent score and is shown on the right. (A, n = 5, B, normal, n = 15, psoriasis, n = 36).
• C, D
  Representative images of the expression of EZH2 in the skin epidermis from the IMQ‐treated mice (C) and the psoriasis patients (D). The number of EZH2‐positive cells per HPF (high power field) in each mouse and patient is shown on the right. (C, control, n = 3, IMQ, n = 6. D, normal, n = 13, psoriasis, n = 27).
• E
  WB analysis of NEDD4L, EZH2, and EZH1 expression in the EZH2‐specific siRNA (siEZH2)‐ or negative control siRNA (siNC)‐transfected NHEKs with IMQ stimulation.
• F, G
  NHEKs were treated with or without IMQ (1 µg/ml) for 6 h. A ChIP assay was conducted to analyze the local enrichment of histone H3K27me3 in specific regions of the NEDD4L promotor in mouse primary keratinocytes (F) and NHEKs (G). n = 3 technical replicates.

Data information: Data are shown as the mean ± s.e.m. and are representative of three independent experiments (A, C, E) or two independent experiment (F, G). Scale bar, 200 μm (100×) or 50 μm (400×). Significant differences were tested using a two‐tailed Student’s t‐test (A–D, F, G). **P < 0.01, ***P < 0.001. NS, no significance. See also Fig EV1.

Source data are available online for this figure.

Previous studies have reported that histone methyltransferase enhancer of zeste homolog 2 (EZH2) could repress NEDD4L transcription by enhancing histone H3 lysine 27 trimethylation (H3K27me3) at the NEDD4L promotor (Zhao et al, 2015). Interestingly, while significantly decreasing the expression of NEDD4L, IMQ treatment increased the expression of EZH2 but not EZH1 at both the mRNA level (Fig EV1, EV2, EV3, EV4, EV5) and protein level (Fig EV1J) in normal human epidermal keratinocytes (NHEKs). Furthermore, IHC staining showed increased expression level of EZH2 protein in the epidermis from the IMQ‐treated mice and the psoriasis patients (Fig 1C and D). Knockdown of EZH2 (Fig 1E) but not EZH1 (Fig EV1K) reversed the downregulation of NEDD4L protein levels by IMQ stimulation in NHEKs. In conclusion, IMQ‐upregulated EZH2 expression is responsible for the suppression of NEDD4L expression in keratinocytes.

Figure EV2. NEDD4L is critical for inhibiting IMQ‐induced keratinocyte hyperplasia in mice.

• A–K
  Six‐ to eight‐week‐old WT and Nedd4l KO mice were treated with or without IMQ for 3 or 6 consecutive days on the back skin or the ear. (A) Representative images of NEDD4L IHC staining of the back tissues from WT and KO mice. Scale bar, 200 μm. (B) Representative pictures of the WT and Nedd4l KO mice treated with or without IMQ. (C) Scaling, thickness, and erythema of the back skin were scored daily on a scale from 0 to 4, and the scaling, thickness, erythema, and cumulative score (Erythema plus scaling plus thickness) are depicted. (control, n = 3, IMQ, n = 7). (D) Ear thickness of the right ear was measured on the days indicated. (control, n = 3, IMQ, n = 7). (E) Representative images of HE staining of the back tissues are shown on the left. The epidermis thickness measured using the scale in the microscope is shown on the right. (F) Representative images of HE staining of the ear tissues are shown on the left. The epidermis thickness measured using the scale in the microscope is shown on the right. (G) Representative images of Ki67 staining in back tissues are shown on the left. The number of Ki67‐positive cells per HPF in each mouse is shown on the right. (H) Representative images of Ki67 staining in ear tissues are shown on the left. The number of Ki67‐positive cells per HPF in each mouse is shown on the right. (control, n = 3, IMQ, n = 6). (I) Keratinocyte proliferation, (J) differentiation, and (K) inflammation associated genes expression level in the back skins from IMQ‐treated mice are detected using real‐time PCR analysis. (Day 0, n = 3, Day 3 and Day 6, n = 6).
• L, M
  Six‐ to eight‐week‐old WT and Nedd4l KO mice were treated with or without IMQ for 4 consecutive days on the ears, the ears were digested using liberase, and the immune cells were detected by flow cytometry. (L) The percentage of the infiltrating leukocytes, neutrophils, T cells, and IL‐17A‐expressing cells. (M) The cell number of the infiltrating leukocytes, neutrophils, T cells, and IL‐17A‐expressing cells. (For leukocytes, neutrophils, and T cells detection, control, n = 5, IMQ, n = 6; for IL‐17A‐expressing cells measurement, control, n = 3, IMQ, n = 6).
• N
  Representative images of NEDD4L IHC staining of the back tissues from WT and cKO mice are shown on the left. The NEDD4L expression level was calculated by multiplying the staining intensity score and the extent score and is shown on the right (n = 3). Scale bar, 200 μm.
• O
  Scaling, thickness and erythema of the back skin were scored daily on a scale from 0 to 4.

Data information: Each symbol represents one mouse (E–H, L–N). For back tissues, scale bar, 200 μm (100×) or 50 μm (400×). For ear tissues, scale bar, 200 μm (100×) or 200 μm (200×). Data are shown as mean ± s.e.m. and are representative of three independent experiments. Significant difference was tested using a two‐tailed Mann–Whitney U test (C, D, O) or two‐tailed Student’s t‐test (E–N). *P < 0.05, **P < 0.01, ***P < 0.001. NS, no significance.

Source data are available online for this figure.

Figure EV3. NEDD4L inhibits the activation of IL‐6 signaling pathway in keratinocytes.

• A, B
  NEDD4L was knocked out by CRISPER cas9 method in NHEKs. Cell proliferation (A) and cell signaling pathway activation (B) induced by IL‐22 was detected by CCK‐8 and WB, respectively. n = 4 biological replicates.
• C, D
  NHEKs were transfected with NEDD4L‐specific siRNA (siNEDD4L) or negative control siRNA (siNC), the cell proliferation (C) and cell signaling pathway activation (D) induced by IL‐17A, IL‐6 and IL‐22 was detected by CCK‐8 and WB, respectively. n = 4 biological replicates.
• E, F
  Proteins in IL‐6 signaling pathway were detected by WB in WT and NEDD4L KO NHEKs (E) or in WT and NEDD4L knockdown NHEKs (F).
• G
  The relative mRNA expression level of GP130 was detected by real‐time PCR in the WT and NEDD4L KO mouse primary keratinocyte (MPK) or NEDD4L KO NHEKs. n = 3 technical replicates.

Data information: Data are shown as mean ± s.e.m., and are representative of three independent experiments. Significant difference was tested using a two‐tailed Mann–Whitney U test (A, C) or two‐tailed Student’s t‐test (G). **P < 0.01. NS, no significance.

Source data are available online for this figure.

Figure EV4. Nedd4l deficiency promotes IMQ‐induced keratinocyte hyperplasia via IL‐6/GP130 signaling.

• A–F
  Six‐ to eight‐week‐old WT and Nedd4l cKO mice were treated with IMQ for 6 consecutive days, and at the beginning of the experiment, the mice were injected intra‐epidermally with 25 μg anti‐IL‐6 neutralization antibody or with same dosage of isotype on every other day. (A) Representative pictures of the mice treated with IMQ. (B) Erythema, scaling and thickness of the back skin was scored daily on a scale from 0 to 4, and the cumulative score (Erythema plus scaling plus thickness) is depicted. (C) Representative images of HE staining of the back tissues are shown on the left. The epidermis thickness measured using the scale in the microscope is shown on the right. (D) Representative images of Ki67 staining in back tissues are shown on the left. The number of Ki67‐positive cells per HPF in each mouse is shown on the right. (E) Representative images of p‐STAT3 staining in back tissues are shown on the left. The number of p‐STAT3‐positive cells per HPF in each mouse is shown on the right. (F) Keratinocyte proliferation‐associated genes in the skin tissues are detected by real‐time PCR analysis. (A–E, n = 6, F, control, n = 3, IMQ, n = 6).
• G, H
  Six‐ to eight‐week‐old WT and Nedd4l cKO mice were treated with IMQ for 6 consecutive days and at the beginning of the supplementation, the mice were injected intra‐epidermally with 10 μg anti‐GP130 antibody or with same dosage of anti‐isotype control on every other day. (G) Representative images of HE staining of the back tissues are shown on the left. The epidermis thickness measured using the scale in the microscope is shown on the right. (H) Representative images of p‐STAT3 staining in back tissues are shown on the left. The number of p‐STAT3‐positive cells per HPF in each mouse is shown on the right. (control, n = 3, IMQ, n = 6).

Data information: Data are shown as mean ± s.e.m. and are representative of two independent experiments. Scale bar, 200 μm (100×) or 50 μm (400×). Significant difference was tested using a two‐tailed Mann–Whitney U test (B) or two‐tailed Student’s t‐test (C–H). *P < 0.05, **P < 0.01, ***P < 0.001. NS, no significance.

Source data are available online for this figure.

Figure EV5. NEDD4L interacts with and mediates Lys‐27‐linked polyubiquitination of GP130.

1. Flag‐GP130 was transfected with or without Myc‐NEDD4L into HEK293T cells, and 36 h later, MG‐132 (25 μM), CQ (Chloroquine, CQ, 100 μM), or equal volume DMSO were added into the cultures for 12 h. The cells were harvested for WB using indicated antibodies.
2. WT and KO NHEK cells were stained with anti‐NEDD4L and anti‐GP130 monoclonal antibodies and subjected to immunofluorescence (IF) imaging. Arrows indicate membrane interaction of NEDD4L with GP130 and triangles indicate cytoplasm interaction. Scale bar, 50 μm.
3. Recombinant HA‐NEDD4L‐his (HA‐NEDD4L) proteins were mixed with Flag‐GFP‐his (Flag‐GFP) or Flag‐GP130‐his (Flag‐GP130) proteins, and then, the proteins were pulled down with ANTI‐FLAG M2 magnetic beads and detected using anti‐Flag and anti‐HA antibodies.
4. HEK293T cells were transfected with the combined plasmids of Flag‐GFP, Flag‐GP130, Flag‐STAT3, and Myc‐NEDD4L. 48 h later, whole cell lysates were co‐IP with ANTI‐FLAG magnetic beads, which was followed by WB detection with anti‐Flag or Myc antibodies.
5. A sketch picture of the structure of NEDD4L and GP130 (ECD, extracellular domain, ICD, intracellular domain).
6. Flag‐NEDD4L‐WW was transfected with Myc‐GP130 into HEK293T cells, and whole cell lysates were co‐IP with ANTI‐FLAG M2 magnetic beads, followed by WB with anti‐Flag or anti‐Myc antibodies.
7. Full length Flag‐GP130, GP130‐AKSY, GP130‐aa 765‐918 truncation mutant, or GP130‐aa 765‐918 were transfected with Myc‐NEDD4L into HEK293T cells, and whole cell lysates were co‐IP with ANTI‐FLAG M2 magnetic beads, followed by WB with anti‐Flag or anti‐Myc antibodies.
8. WB analysis of the polyubiquitination of GP130 in HEK293T cells co‐transfected with full length Flag‐GP130, hemagglutinin (HA)‐Ubiquitin (HA‐Ub), and Myc‐NEDD4L or its truncated mutants or enzymatic activity mutant C943S, which followed by denaturation‐IP with ANTI‐FLAG M2 beads and WB analysis with anti‐HA antibodies.
9. WB analysis of the polyubiquitination of GP130 in HEK293T cells co‐transfected with full length Myc‐NEDD4L, HA‐Ub, Flag‐GP130, or Flag‐GP130‐AKSY mutant, which followed by denaturation‐IP with ANTI‐FLAG M2 beads and WB analysis with anti‐HA antibodies.
10. WB analysis of the polyubiquitination of GP130 in HEK293T cells co‐transfected with full length Flag‐GP130, Myc‐NEDD4L, and HA‐Ub or the mutant ubiquitin constructs (K‐R), followed by denaturation‐IP with ANTI‐FLAG M2 magnetic beads and WB analysis with anti‐HA antibodies.
11. WB analysis of the polyubiquitination of GP130 in HEK293T cells co‐transfected with full length Flag‐GP130, Myc‐NEDD4L, and HA‐Ub or its K6R, K27R, K33R, and K6‐27‐33R mutants, followed by denaturation‐IP with ANTI‐FLAG M2 magnetic beads and WB analysis with anti‐HA antibodies.
12. WB analysis of the polyubiquitination of GP130 in HEK293T cells co‐transfected with full length Myc‐NEDD4L, HA‐Ub, and Flag‐GP130 or its lysine mutants (K‐R) in the intracellular domain, followed by denaturation‐IP with ANTI‐FLAG M2 magnetic beads and WB analysis with anti‐HA antibodies.
13. Myc‐NEDD4L or its mutants were co‐transfected with Flag‐GP130, its AKSY domain mutant or K718R mutants into HEK293T cells, and the expression of Flg‐GP130 was detected using WB with anti‐Flag antibodies.

Data information: A representative result from three independent experiments is shown.

Source data are available online for this figure.

Histone modifications are associated with the transcriptional regulation of genes, among which K4 and K36 are methylated in transcriptionally active chromatin, whereas methylation of K9 and K27 is linked with inactive chromatin (Sati et al, 2012). To cover the whole region of the NEDD4L promoter, we designed five pairs of primer sets, and the promoter regions to be amplified are indicated in Fig EV1L. Chromatin immunoprecipitation (ChIP) real‐time PCR experiments showed a significantly increased enrichment of H3K27me3 in P5 regions of NEDD4L promoter in IMQ‐induced mouse primary keratinocytes (Fig 1F) and NHEKs (Fig 1G). However, IMQ stimulation had no effect on H3K4me3, H3K9me3, and H3K36me3 modification in the promoter region of NEDD4L in mouse primary keratinocytes (Fig EV1M) and NHEKs (Fig EV1N). Considering IMQ‐upregulated EZH2 expression is responsible for the suppression of NEDD4L expression in keratinocytes (Fig 1E), the enhanced EZH2 might repress NEDD4L expression via H3K27 trimethylation of the NEDD4L promoter in the progression of psoriasis.

NEDD4L is critical for inhibiting IMQ‐induced keratinocyte hyperplasia in mice

To investigate the regulatory effects of NEDD4L on psoriasis, we generated an IMQ‐induced model of psoriasis skin inflammation in Nedd4l knockout (KO) mice (Fig EV2A). Both the WT and KO mice showed increasing psoriasis‐like disease symptoms, including erythema, scaling, and skin thickness 2 or 3 days after the start of IMQ application; however, the KO mice showed more severe scaling and thickness, but not erythema, than the WT mice (Fig EV2B and C). The independent scores of erythema, scaling, and skin thickness were used to form a cumulative score, and the KO mice exhibited a higher cumulative score than the WT mice (Fig EV2C). Daily application of IMQ to the ears led to a significant increase in ear thickness in the KO mice compared with the WT mice (Fig EV2D). Analysis of the HE‐stained sections from the IMQ‐treated skin showed elevated epidermal thickening both in the back (Fig EV2E) and ear skin (Fig EV2F) in the KO mice compared with the WT mice. This acanthosis was caused by hyperproliferation of keratinocytes, as indicated by the increased numbers of Ki67‐positive keratinocytes in the basal and suprabasal layer of the back (Fig EV2G) and ear (Fig EV2H) skins in the KO mice compared to the WT mice.

Furthermore, Nedd4l knockout significantly upregulated the expression of IMQ‐induced keratinocyte proliferation‐associated genes, including Krt14, S100A7, and CyclinD1 (Fig EV2I), but downregulated the expression of keratinocyte differentiation‐associated genes, including Krt10, Filaggrin, and Loricrin (Fig EV2J). However, Nedd4l knockout had no effect on the expression of psoriasis‐associated inflammatory cytokines, such as IL‐23p19, IL‐22, IL‐17A, and IL‐6 (Fig EV2K). Flow cytometry analysis indicated that there was no significant difference in the percentages (Fig EV2L) and cell numbers (Fig EV2M) of the infiltrating leukocytes, neutrophils, T cells, and IL‐17A‐expressing cells between the WT and KO mice supplemented with IMQ. These results indicate that NEDD4L may function in the epithelial system in the process of IMQ‐induced psoriasis‐like inflammation.

We then crossed Nedd4l^f/f mice with Krt14Cre mice to generate Nedd4l^f/fKrt14Cre mice, in which Nedd4l was specifically knocked out in keratinocytes (conditional knockout, cKO) (Fig EV2N). The Nedd4l cKO mice showed more severe scaling and thickness, but not erythema, than the WT mice after IMQ treatment (Figs 2A and EV2O). Daily IMQ supplementation resulted in an increased cumulative score and ear thickness in the cKO mice (Fig 2B and C). HE‐ and Ki67‐stained sections from the IMQ‐treated skins showed elevated epidermal thickening and increased numbers of Ki67‐positive keratinocytes in the basal and suprabasal layer both in the back (Fig 2D and E) and ear (Fig 2F and G) skins of the cKO mice. In conclusion, these data indicated that epidermal NEDD4L is crucial for inhibiting IMQ‐induced hyperplasia in mice.

Figure 2. NEDD4L is critical for inhibiting IMQ‐induced keratinocyte hyperplasia in mice.

Six‐ to eight‐week‐old WT and Nedd4l cKO mice were treated with or without IMQ for 6 consecutive days on the back and the ears.

1. Representative pictures of WT and Nedd4l cKO mice treated with or without IMQ.
2. Erythema, scaling, and thickness of the back skin were scored daily on a scale from 0 to 4, and the cumulative score (erythema plus scaling plus thickness) is depicted.
3. Ear thickness of the right ear was measured on the days indicated.
4. Representative images of HE staining of the back tissues are shown on the left. The epidermal thickness measured using the scale in the microscope is shown on the right.
5. Representative images of Ki67 staining in back tissues are shown on the left. The number of Ki67‐positive cells per HPF in each mouse is shown on the right.
6. Representative images of HE staining of the ear tissues are shown on the left. The epidermis thickness measured using the scale in the microscope is shown on the right.
7. Representative images of Ki67 staining in ear tissues (same specimen as in F) are shown on the left. The number of Ki67 positive cells per HPF in each mouse is shown on the right.

Data information: Data are shown as the mean ± s.e.m. and are representative of three independent experiments. For back tissues, scale bar, 200 μm (100×) or 50 μm (400×). For ear tissues, scale bar, 200 μm (100×) or 200 μm (200×). Significant differences were tested using a two‐tailed Mann–Whitney U test (B, C) or two‐tailed Student’s t‐test (D–G). Control, n = 3, IMQ, n = 6. **P < 0.01, ***P < 0.001. NS, no significance. See also Fig EV2.

Source data are available online for this figure.

NEDD4L inhibits the activation of the IL‐6 signaling pathway in keratinocytes

In an imiquimod‐induced psoriasis model, cytokines IL‐22, IL‐17A, and IL‐6 act directly on keratinocytes and promote keratinocyte proliferation (Zheng et al, 2007; Lindroos et al, 2011; Lai et al, 2012). We generated a NEDD4L knockout cell line (KO) in NHEKs to study the effects of NEDD4L on IL‐22‐, IL‐17A‐, and IL‐6‐induced keratinocyte proliferation and cell signaling. NEDD4L knockout had no effect on IL‐22‐induced cell proliferation (Fig EV3A) but significantly promoted IL‐17A and IL‐6‐induced NHEK proliferation (Fig 3A). However, NEDD4L deficiency specifically enhanced IL‐6‐induced STAT3 activation but had no effect on IL‐17A (Fig 3B and C)‐ or IL‐22 (Fig EV3B)‐triggered signaling. These results were confirmed by transient knockdown experiment using RNA interference (RNAi) targeting NEDD4L in NHEKs (Fig EV3C and D). Notably, knockout or knockdown of NEDD4L affected IL‐17A‐induced NHEK proliferation in a delayed manner (after 48 h, Fig 3A). Given that NEDD4L had no effect on IL‐17A signaling, the inhibitory effect of NEDD4L on IL‐17A‐induced NHEK proliferation may be indirect. IL‐6 is an IL‐17A inducible cytokine and possesses potent pro‐proliferative effects (Croxford et al, 2014). Using an anti‐IL‐6 antibody, we further found that the pro‐proliferative effect of NEDD4L deficiency on the IL‐17A‐treated NHEKs was completely abrogated by IL‐6 neutralization (Fig 3D), indicating that NEDD4L regulates IL‐17A‐induced NHEK proliferation by promoting IL‐6 signaling.

Figure 3. NEDD4L inhibits the activation of the IL‐6 signaling pathway in keratinocytes.

• A–F
  NEDD4L was knocked out by the CRISPR Cas9 method in NHEKs. (A) Cell proliferation induced by IL‐17A and IL‐6 was detected by CCK‐8 assays. (B, C) Western blot (WB) analysis of protein expression in NHEKs induced by IL‐17A (B) and IL‐6 (C). (D) WT and NEDD4L KO NHEKs were precoated with anti‐IL‐6 neutralization antibody or an isotype, and the cell proliferation induced by IL‐17A was detected by CCK‐8 assays. (E, F) WT and NEDD4L KO NHEKs were infected with lentivirus overexpressing NEDD4L‐WT or NEDD4L‐C943S mutants. Cell proliferation (E) and cell signaling pathway activation (F) induced by IL‐6 were detected by CCK‐8 and WB analyses, respectively.
• G, H
  Primary keratinocytes were separated from the WT and Nedd4l KO mice. (G) WB analysis of the protein expression in IL‐6‐induced cell signaling pathway activation. (H) WB analysis of the proteins in the IL‐6 signaling pathway.
• I
  Cell proliferation assay of GP130 KO NHEKs transfected with siNEDD4L induced by IL‐17A.

Data information: Data are shown as the mean ± s.e.m. and are representative of three independent experiments (A–I). Significant differences were tested using a two‐tailed Mann–Whitney U test (A, D, E, I). P < 0.01, **P < 0.01, ***P < 0.001. (A, D, E, I), n = 4 biological replicates. See also Fig EV3.

Source data are available online for this figure.

NEDD4L or NEDD4L‐C943S mutant lacking ubiquitin ligase activity was then transfected into NEDD4L KO NHEKs. As shown in Fig 3E and F, only WT NEDD4L, but not the C943S mutant, abolished the increased IL‐6‐induced proliferation and STAT3 phosphorylation resulting from NEDD4L knockout, indicating that the inhibitory effects of NEDD4L on IL‐6 signaling are dependent on its E3 ubiquitin ligase activity. In addition, Nedd4l deficiency in mouse primary keratinocytes notably increased IL‐6‐induced STAT3 phosphorylation (Fig 3G).

IL‐6 interacts with the IL‐6 receptor (IL‐6R) and then heterodimerizes with two molecules of GP130, leading to the activation of GP130 and, subsequently, phosphorylation of the JAK‐STAT3 pathway (Kang et al, 2019). We next evaluated the protein expression levels in the IL‐6 signaling pathway in the WT and NEDD4L KO keratinocytes. Nedd4l knockout strongly promoted GP130 protein expression but had no effect on the expression of other proteins in the IL‐6 signaling pathway in mouse primary keratinocytes (Fig 3H), which was confirmed in the NEDD4L knockout or knockdown NHEKs (Fig EV3E and F). However, NEDD4L deficiency had no effect on the mRNA expression level of GP130 (Fig EV3G). Consistent with the finding that anti‐IL‐6 antibody abolished the promoting effect of NEDD4L deficiency on IL‐17A‐induced NHEK proliferation, GP130 knockout not only decreased IL‐17A‐induced NHEKs proliferation but also abrogated the effect of NEDD4L knockdown on the proliferation of NHEKs (Fig 3I). Collectively, these data indicated that NEDD4L negatively regulates IL‐6 and IL‐17A‐promoted keratinocyte proliferation by inhibiting the IL‐6/GP130/STAT3 signaling pathway.

Nedd4l deficiency promotes IMQ‐induced keratinocyte hyperplasia via IL‐6/GP130 signaling in vivo

We then treated the WT and Nedd4l cKO mice with neutralizing antibodies against either IL‐6 or GP130. As shown in Fig EV4A and B, an intradermal injection of anti‐IL‐6 neutralizing antibody not only resulted in a substantial inhibition of IMQ‐induced skin epidermal thickness and scaling but also abolished the increased skin thickness and scaling in the cKO mice. Furthermore, the epidermal thickness (Fig EV4C), the numbers of Ki67‐positive cells in the basal and suprabasal layer (Fig EV4D), and p‐STAT3‐positive cells in the epidermis (Fig EV4E) of the IMQ‐treated Nedd4l‐deficient mice were decreased to levels comparable with those in the WT mice upon anti‐IL‐6 antibody injection. The effect of Nedd4l cKO on the expression of keratinocyte proliferation‐associated genes (Fig EV4F) was also abolished by treatment with an anti‐IL‐6 neutralizing antibody. Similarly, the increased IMQ‐induced skin thickness, scaling (Fig 4A and B), and skin epidermal proliferation (Fig EV4G) in the cKO mice were completely compromised by intradermal injection of an anti‐GP130 antibody. The enhanced numbers of Ki67 (Fig 4C)‐ and p‐STAT3‐positive cells (Fig EV4H) and the expression of keratinocyte proliferation‐associated genes (Fig 4D) in the IMQ‐treated cKO mice were decreased to levels comparable with those in the WT mice upon anti‐GP130 antibody injection. Collectively, these data suggested that NEDD4L limits IMQ‐induced psoriatic‐like hyperplasia by controlling the IL‐6/GP130 signaling pathway in vivo.

Figure 4. Nedd4l deficiency promotes IMQ‐induced keratinocyte hyperplasia via IL‐6/GP130 signaling.

• A–D
  Six‐ to eight‐week‐old WT and Nedd4l cKO mice were treated with IMQ for 6 consecutive days, and at the beginning of the treatment, the mice were injected intra‐epidermally with 10 μg anti‐GP130 antibody or with same dosage of isotype on every other day. (A) Representative pictures of the mice treated with IMQ. (B) Erythema, scaling, and thickness of the back skin was scored daily on a scale from 0 to 4, and the cumulative score (Erythema plus scaling plus thickness) is depicted. (C) Representative images of Ki67 staining in back tissues are on the left. The number of Ki67‐positive cells per HPF is shown on the right. (D) Keratinocyte proliferation‐associated genes in the skin tissues are detected using real‐time PCR analysis.

Data information: Data are shown as mean ± s.e.m. and are representative of two independent experiments. Scale bar, 200 μm (100×) or 50 μm (400×). Significant differences were tested using a two‐tailed Mann–Whitney U test (B) or two‐tailed Student’s t‐test (C, D). C, control, n = 5, IMQ, n = 6. D, control, n = 3, IMQ, n = 6. *P < 0.05, **P < 0.01, ***P < 0.001. NS, no significance. See also Fig EV4.

Source data are available online for this figure.

The expression level of NEDD4L negatively correlates with GP130 and p‐STAT3 in the epidermis from clinical samples

To evaluate the correlation of the NEDD4L expression level with GP130 and p‐STAT3 in the clinic, we collected skin tissues from psoriasis patients (n = 36) and protein expression was detected by IHC. As shown in Fig 5A and B, quantitatively standardized IHC analyses revealed that the protein expression level of NEDD4L was significantly decreased, while the protein expression levels of GP130 and p‐STAT3 were enhanced in the skin epidermis from individuals with psoriasis compared with those of the normal controls. Moreover, the expression level of NEDD4L was negatively correlated with the expression levels of GP130 and p‐STAT3 in normal and psoriatic samples (Fig 5C and D). The expression level of active STAT3, which is a downstream target of GP130, was positively correlated with GP130 in normal and psoriatic clinical samples (Fig 5E). Taken together, these data indicated that NEDD4L negatively regulated the progression of psoriasis in humans by inhibiting GP130/STAT3 signaling.

Figure 5. The expression level of NEDD4L negatively correlates with the levels of GP130 and p‐STAT3 in the epidermis from clinical samples.

The skin tissues from psoriasis patients were harvested and subjected to IHC staining. Normal skin sections (Normal) consisted of healthy tissue from the resection edges of cutaneous squamous cell carcinoma (SCC) that appeared to be healthy at the histological level.

1. Representative pictures of the expression of NEDD4L, GP130, and p‐STAT3 in the epidermis of a normal skin section (same specimen) and sections from psoriasis patients. Isotypes are IgG antibodies from same species served as negative control.
2. Semiquantitative scoring of NEDD4L, GP130, and p‐STAT3 in the epidermis detected by IHC.
3. Correlation analysis for NEDD4L and GP130 in the normal and psoriasis groups.
4. Correlation analysis for NEDD4L and p‐STAT3 in the normal and psoriasis groups.
5. Correlation analysis for GP130 and p‐STAT3 in the normal and psoriasis groups.

Data information: Each symbol represents one patient (B–E). Scale bar, 200 μm (100×) or 50 μm (400×). All data are shown as the mean ± s.e.m., and significant differences were analyzed using two‐tailed Student’s t‐test (B) and two‐tailed Pearson’s chi‐square test (C–E) (normal, n = 15, psoriasis, n = 36). *P < 0.05, **P < 0.01, ***P < 0.001.

Source data are available online for this figure.

NEDD4L interacts with and mediates GP130 proteasomal degradation

NEDD4L knockout or knockdown promoted GP130 protein expression level but had no effect on GP130 mRNA expression levels in keratinocytes (Figs 3G and EV3, EV4, EV5). Overexpression of NEDD4L but not its enzymatic activity mutant C943S in HEK293T cells promoted GP130 degradation in a dose‐dependent manner (Fig 6A), and this effect could be blocked by the proteasome inhibitor MG‐132 but not by the lysosome inhibitor chloroquine (CQ) (Figs 6A and EV5A), indicating that NEDD4L promoted GP130 degradation through the proteasomal pathway. Consistently, the half‐life of GP130 protein was strongly reduced by overexpression of NEDD4L (Fig 6B). Endogenous NEDD4L was coimmunoprecipitated with GP130 (Fig 6C) in NHEKs. The immunofluorescence (IF) assay was performed in NHEK cells to detect the co‐localization of NEDD4L and GP130. As shown in Fig EV5B, NEDD4L was localized both in the cytoplasm and on the membrane, whereas GP130 was predominantly localized on the membrane, and to a laser extent, in the cytoplasm. Co‐localization of NEDD4L and GP130 could be observed on the cell membrane (indicated by arrows), and to a lesser extent, in the cytoplasm (indicated by triangles). The GP130 protein expression level in the NEDD4L KO NHEK cells was higher than that in the WT NHEK cells, both on the membrane and in the cytoplasm of the cells. The recombinant HA‐NEDD4L‐his (HA‐NEDD4L), HA/Flag‐GFP‐his (HA/Flag‐GFP), and Flag‐GP130‐his (Flag‐GP130) proteins were purified using a Ni‐NTA‐Sefinose Column to perform pull‐down experiments. As shown in Figs 6D and EV5C, NEDD4L recombinant protein was coprecipitated with GP130 protein, demonstrating that NEDD4L directly interacts with GP130. In the transiently overexpressed HEK293T cells, NEDD4L coprecipitated with GP130 but not STAT3 in the IL‐6 signaling pathway (Fig EV5D). NEDD4L contains one C2 domain, four WW domains and one enzymatic HECT domain (Fig EV5E). To map the domains required for NEDD4L to interact with GP130, we constructed a series of plasmids expressing wild‐type or truncation mutants, in which the C2 (NEDD4L‐∆C2), WW (NEDD4L‐∆WW), or HECT (NEDD4L‐∆HECT) domains were deleted. As shown in Fig 6E, deletion of the WW domain, but not the C2 or HECT domain, disrupted the interaction between NEDD4L and GP130. Furthermore, the binding of GP130 with the WW domain of NEDD4L was detected in transiently overexpressed 293T cells (Fig EV5F), demonstrating that GP130 binds to the WW domain of NEDD4L. GP130 is a membrane‐spanning protein comprising an extracellular domain (aa 1–615), a membrane spanning (aa 616–918), and an intracellular domain (aa 616–918). The GP130 intracellular domain (ICD) carries four subdomains: the JAK1/2‐TYK2‐binding domain (aa 616–734), the Smad7/SHP2/SOCS3‐binding domain (aa 735–764), the STAT3‐biding domain (aa 765–889), and the STAT1/3 binding domain (aa 890–918) (Kamimura et al, 2003; Yu et al, 2017). Therefore, we constructed a series of plasmids expressing the wild‐type or the different domains of GP130 (Fig EV5E) to determine which domain in GP130 is required for its binding to NEDD4L. As shown in Figs 6F and EV5G, aa 765–918 of GP130 was crucial for the interaction of GP130 with NEDD4L. It has been reported that the WW domain of HECT‐type ubiquitin ligase interacts with the PY motif (PPxY, PxY or PxxY, x could be any amino acid) of its substrates (Ding et al, 2013). Consistently, the interaction between GP130 and NEDD4L was abrogated when the conserved residue PKSY (aa 902–905) in GP130 was substituted by AKSY (Fig EV5G). Collectively, these data indicated that NEDD4L directly interacts with GP130 through the WW domain of NEDD4L and the PKSY motif of GP130.

Figure 6. NEDD4L interacts with GP130 and promotes its degradation.

1. NEDD4L promotes GP130 degradation in a dose and enzymatic activity‐dependent manner. HA‐GP130 was transfected with 0.5 μg, 1 μg, or 2 μg of Myc‐NEDD4L or 2 μg of its enzymatic activity mutant into HEK293T cells and 36 h later, MG‐132 (25 μM) was added for further 12 h, and the cell lysates were harvested for WB analysis.
2. Flag‐GP130 was transfected with Myc‐NEDD4L into HEK293T cells, and 36 h later, cyclohexane (CHX, 50 ng/ml) was added into the cells for indicated time. The cell lysates were harvested and analyzed by WB, and the gray values were measured to calculate half‐life of GP130.
3. Whole cell lysates of NHEKs were coimmunoprecipitated (co‐IP) with anti‐IgG or anti‐GP130 antibodies, which was followed by WB detection with anti‐NEDD4L and anti‐GP130 antibodies.
4. Recombinant Flag‐GP130‐his (Flag‐GP130) proteins were mixed with HA‐GFP‐his (HA‐GFP) or HA‐NEDD4L‐his (HA‐NEDD4L) proteins, and then, the proteins were pulled down with ANTI‐HA agarose beads and detected using anti‐Flag and anti‐HA antibodies.
5. Full length Flag‐NEDD4L, NEDD4L‐ΔC2, NEDD4L‐ΔWW, or NEDD4L‐ΔHECT truncation mutants were co‐expressed with Myc‐GP130 in HEK293T cells, and whole cell lysates were co‐IP with ANTI‐FLAG M2 magnetic beads, followed by WB with anti‐Flag or anti‐Myc antibodies.
6. Full length Flag‐GP130, GP130‐aa 616‐918, GP130‐aa 616‐889, GP130‐aa 616‐764, and GP130‐aa 616‐734 were transfected with Myc‐NEDD4L into HEK293T cells, and whole cell lysates were co‐IP with ANTI‐FLAG M2 magnetic beads, followed by WB with anti‐Flag or anti‐Myc antibodies.

Data information: A representative result from three independent experiments is shown. See also Fig EV5.

Source data are available online for this figure.

NEDD4L mediates Lys‐27‐linked polyubiquitination of GP130

As NEDD4L is a HECT E3 ubiquitin ligase, we then examined whether NEDD4L directly ubiquitinates GP130. Indeed, wild‐type but not the C943S mutant of NEDD4L efficiently promoted GP130 ubiquitination (Fig EV5H). The WW domain or ubiquitin ligase activity mutant with a HECT domain deletion of NEDD4L completely lost the ability to promote GP130 ubiquitination (Fig EV5H), and the GP130 mutant with the PKSY motif mutated to AKSY was resistant to NEDD4L‐mediated ubiquitination (Fig EV5I). An in vitro ubiquitination assay confirmed that GP130 is a direct target of NEDD4L, whereas the ubiquitination effect was abrogated by the deletion of the WW domain, HECT domain, or the mutation of cysteine 943 to serine (Fig 7A). Furthermore, NEDD4L deficiency abrogated the ubiquitination of GP130 in NHEKs (Fig 7B), and reconstitution of NEDD4L KO NHEKs with wild‐type NEDD4L, but not NEDD4L‐ΔC2, ΔWW, or the C943S mutant, restored the ubiquitination of GP130 in the NEDD4L‐deficient NHEKs (Fig 7C). To determine the nature of NEDD4L‐mediated GP130 ubiquitination, we transfected HEK293T cells with NEDD4L‐ and GP130‐expressing vectors in the presence of constructs expressing wild‐type ubiquitin (HA‐Ub‐WT) or its mutants. As shown in Figs 7D and EV5J, NEDD4L mainly promoted Lys‐27‐ and, to a lesser extent, Lys‐6‐ and Lys‐33‐linked polyubiquitination of GP130. Mutation of Lys‐6, Lys‐27, and Lys‐33 completely abrogated the ubiquitination effect of NEDD4L on GP130 (Fig EV5K).

Figure 7. NEDD4L mediates Lys‐27‐linked polyubiquitination of GP130.

1. Recombinant Flag‐GP130, Myc‐NEDD4L, Myc‐NEDD4L‐ΔC2, Myc‐NEDD4L‐ΔWW, Myc‐NEDD4L‐ΔHECT, or Myc‐NEDD4L‐C943S, and ubiquitination‐associated proteins were mixed in adenosine 5′‐triphoshpate buffer and incubated at 30°C for 2 h. Then, the reaction aliquots were quenched and measured by WB with anti‐Flag antibodies.
2. Whole cell lysates of WT and NEDD4L KO NHEKs were immunoprecipitated with anti‐GP130 antibody, and the immune‐precipitates were analyzed by WB with anti‐Ub antibodies.
3. WT or NEDD4L KO NHEK cells was infected with lentivirus that overexpress NEDD4L‐WT, NEDD4L WW domain‐truncated mutant (NEDD4L‐ΔWW) or its enzymatic activity mutant (NEDD4L‐C943S), the cells were followed by denaturation‐IP with anti‐GP130 antibody and WB analysis with anti‐Ub antibodies.
4. WB analysis of the polyubiquitination of GP130 in HEK293T cells co‐transfected with Flag‐GP130, Myc‐NEDD4L, HA‐Ub, or its mutants expressing only one lysine (K‐only), respectively, which followed by denaturation‐IP with ANTI‐FLAG M2 beads and WB analysis with anti‐HA antibodies.
5. WB analysis of the polyubiquitination of GP130 in HEK293T cells co‐transfected with full length Myc‐NEDD4L, HA‐Ub, and Flag‐GP130 or its lysine mutants (K‐R) in the intracellular domain, followed by denaturation‐IP with ANTI‐FLAG M2 magnetic beads and WB analysis with anti‐HA antibodies.

Data information: A representative result from three independent experiments is shown. See also Fig EV5.

Source data are available online for this figure.

E3 promotes the ubiquitin chain linked to the lysine residues of the substrates (Ding et al, 2013). A series of GP130 mutants possessing a single lysine mutant (K‐R) in the intracellular domain of GP130 were constructed to explore which lysine in GP130 was responsible for the linkage of the ubiquitin chain mediated by NEDD4L. As shown in Figs 7E and EV5L, the mutation of K718 to arginine completely abrogated NEDD4L‐mediated ubiquitination of GP130, which indicated that the ubiquitin chain mediated by NEDD4L is on Lys‐718 of GP130. Consistent with the above results, only the wild‐type NEDD4L and ΔC2 mutant, but not the ΔWW, ΔHECT truncation mutants or C943S mutant, promoted GP130 degradation (Fig EV5M). The GP130 AKSY mutant or K718R mutant, which lost the interaction motif or the ubiquitin chain linkage site, respectively, was resistant to NEDD4L‐mediated GP130 degradation (Fig EV5M). Taken together, these data indicate that NEDD4L directly mediates Lys‐27‐linked ubiquitination at Lys‐718 in GP130, which leads to the proteasomal degradation of GP130.

Discussion

NEDD4L is a highly conserved HECT‐type E3 ligase that is known to bind and regulate a number of membrane proteins, aiding their internalization and turnover. NEDD4L has been demonstrated to be a tumor suppressor and reduced NEDD4L expression contributes to the initiation/progression of a number of cancers, such as breast cancer (Guarnieri et al, 2018), colorectal cancer (Novellasdemunt et al, 2020), and pancreatic cancer (Frampton et al, 2014). In this study, we demonstrated that NEDD4L protects against the development of psoriasis. NEDD4L expression in the epidermis was strongly repressed in patients with psoriasis.

Psoriasis is a common skin disease that affects nearly two percent of people worldwide, but the pathogenic mechanisms remain unclear. With the development of psoriatic animal models, the IL‐23/IL‐17 axis has been identified as a crucial mediator of psoriasis (van der Fits et al, 2009). Furthermore, monoclonal antibodies that target IL‐23 (both IL‐23p19 and p40 subunits) and IL‐17 have been used in the clinical treatment of psoriasis. The proinflammatory properties of IL‐17 have been mainly attributed to its capability to induce the expression of IL‐6, which activates the oncogenic transcription factor STAT3 (Du et al, 2010). Strikingly, STAT3 was found to be activated in lesional keratinocytes from virtually all psoriatic patients, which suggested that STAT3 activation in keratinocytes might be necessary for the development of psoriasis (Sano et al, 2005). IL‐6 is a potent cytokine that promotes keratinocyte proliferation, whereas it inhibits keratinocyte differentiation (Du et al, 2010). In animal studies, IL‐6 functions downstream of IL‐17A to exacerbate neutrophil microabscess development in IL‐17A‐driven psoriasiform lesions through the induction of IL‐6R‐expressing neutrophils into the skin (Croxford et al, 2014). IL‐17C⁺IL‐6 KO mice showed a delayed onset in psoriasis‐like disease (Fritz et al, 2017). Psoriasis‐like inflammation and skin lesions induced by intradermal injection of IL‐23 or IL‐17F were also dependent on the induction of IL‐6 (Fujishima et al, 2010; Lindroos et al, 2011). The expression of IL‐6 was increased in the blood and skin tissues of psoriasis patients (Galadari & Sheriff, 2005). The IL‐6 promoter gene single‐nucleotide polymorphism (SNP) at position ‐174 (IL‐6‐174G/C, rs1900795), which causes elevated IL‐6 expression, is a well‐documented risk factor of psoriasis (Nie et al, 2016). It would be also interesting to investigate whether the SNP rs7529229 in the IL‐6R gene has a protective role in psoriasis (Swerdlow et al, 2012), since it increases the expression of soluble IL‐6R levels in the blood, which could interact with soluble GP130 to form IL‐6 buffer in the blood. In clinical case studies, tocilizumab, an anti‐IL‐6 receptor antibody, has been reported to successfully and completely resolve moderate to severe psoriasis in some cases induced by TNF‐α inhibitor rituximab (Jayasekera et al, 2014). Most evidence to data, however, showed that tocilizumab failed to ameliorate psoriasis (Wendling et al, 2012; Blauvelt, 2017). We here found that anti‐IL‐6 or anti‐GP130 treatment significantly inhibits IMQ‐induced psoriasis in mouse, indicating that the IMQ‐induced psoriasis skin inflammation covers several but not all aspects of human psoriasis. IL‐17A significantly enhanced NEDD4L deficiency‐promoted keratinocyte proliferation, and this effect could be attributed to the induction of IL‐6 expression both in vitro and in vivo. Interleukin‐6 can either signal via the membrane‐bound receptor (Classic signaling) or in complex with the soluble IL‐6 receptor (sIL‐6R, trans‐signaling), and in both cases, homodimerization of GP130 is needed for the activation of intracellular signaling (Garbers et al, 2015). Aberrant activation of GP130 signaling is associated with the initiation/progression of many diseases, including autoimmunity, inflammation, and even cancer in clinic, making GP130 a promising drug target for therapy (Jones et al, 2011; Xu & Neamati, 2013). In this work, we demonstrated that NEDD4L regulated IL‐6/GP130/STAT3 signaling via mediating GP130 degradation, which is crucial for controlling keratinocyte proliferation in psoriasis patients and in IMQ‐induced mice.

Although GP130 was initially characterized as the signal transducer subunit of the IL‐6 receptor, GP130 also functions as the β‐receptor for IL‐11, oncostatin‐M (OSM), leukemia inhibitory factor (LIF), ciliary neurotrophic factor (CNTF), cardiotrophin‐1 (CT‐1), erythropoietin (EPO), IL‐12, IL‐10, and leptin (Hunter & Jones, 2015). Among the cytokines, LIF, OSM (Bonifati et al, 1998), and IL‐11 (Ameglio et al, 1997) were all reported to be upregulated in the lesional skin of psoriasis. Since NEDD4L constitutively interacts with GP130 and inhibits GP130 expression in resting cells, NEDD4L might also negatively regulate signal transduction of these factors and function in a broad range of biological processes. Interestingly, it has been reported that IL‐6 stimulation‐induced lysosome‐dependent degradation of GP130, which is mediated by the E3 ligase c‐CBI, is critical for cessation of IL‐6‐mediated signaling in HeLa and MEF cells (Tanaka et al, 2008). As a methyltransferase enzyme, EZH2 regulates multiple genes expression (Lin et al, 2017). DNA damage‐binding protein 2 (DDB2) can bind to the promotor region of NEDD4L and recruit EZH2 histone methyltransferase to repress NEDD4L transcription by enhancing histone H3 lysine 27 trimethylation at the NEDD4L promoter (Zhao et al, 2015). In this study, we broadened this view by demonstrating that in keratinocytes, IMQ dramatically suppressed NEDD4L expression by promoting EZH2‐mediated H3K27me3 at the promoter of NEDD4L. These results indicate that the mechanisms underlying the regulation of NEDD4L expression might be conserved in keratinocytes during the initiation of psoriasis.

Taken together, the present study identifies NEDD4L as a negative regulator of IL‐6 signaling by inhibiting GP130 expression in keratinocytes, bringing novel insight into the pathogenesis of psoriasis. Considering that NEDD4L could inhibit the progression of psoriasis and was downregulated in the epidermis of the lesional skin from psoriasis patients, overexpression of NEDD4L, for example by using an adeno‐associated virus (AAV) in vivo gene delivery system (Wang et al, 2019), might be a potential therapeutic strategy for psoriasis treatment.

Materials and Methods

Ethics

The experiment license to use human paraffin‐embedded psoriasis was approved by the Medical Research Ethics Committee of Zhejiang University. In addition, informed consent was obtained from all of the subjects involved, and the experiments were conducted according to the principles expressed in the Declaration of Helsinki.

Animals

Heterozygous Nedd4l mice were purchased from JAX^® Mice, America. Nedd4l^f/f mice were purchased from Cyagen Bioscience. KO mice and the WT littermate control were generated by crossing of Nedd4l heterozygous. Nedd4l Keratinocyte‐specific knockout mice (cKO) were generated by crossing of Nedd4l^f/f and Nedd4l^f/f Kertin14Cre mice. All the mice were maintained under the specific‐pathogen‐free (SPF) condition in the Laboratory Animal Center of Zhejiang University. All animal experiments were conducted in accordance with Institutional Animal Care and Use Committee guidelines at the School of Medicine, Zhejiang University.

Plasmids

HA‐tagged GP130 and Flag‐tagged‐STAT3 plasmids were gifts of professor XinHua Feng (Life Science Institute, Zhejiang University). Flag‐tagged GP130 WT or its variants, including GP130‐Intracellular domain (Flag‐GP130‐ICD, aa 616–918), aa 616–889, aa 616–764, aa 616–734, aa 764–918, Δaa 764–918, and AKSY mutants were generated by PCR and subcloned into pcDNA3.1‐EGFP‐Flag‐His vector using Not1 and Sfa1 restriction enzymatic sites. Myc‐tagged GP130 WT was also subcloned into pcDNA3.1‐Myc‐His vector using Not1 and Sfa1 restriction enzymatic sites for immunoprecipitation and protein purification. Flag‐tagged and Myc‐tagged NEDD4L WT or its variants, including NEDD4L‐ΔC2, NEDD4L‐ΔWW, NEDD4L‐ΔHECT, and NEDD4L‐C943S, were generated by PCR and subcloned into pcDNA3.1‐EGFP‐Flag‐His or pcDNA3.1‐EGFP‐Myc‐His vectors using Not1 and Sfa1 restriction enzymatic sites. HA‐Ub‐K6R, K11R, K27R, K29R, K33R, K48R, and K63R or HA‐Ub‐K6‐only, K11‐only, K27‐only, K29‐only, K33‐only, K48‐only, and K63‐only were generated by point mutation method. NEDD4L, WW domain, HECT domain truncation mutants, and its enzymatic activity mutant C943S were cloned to pHAGE‐IRES‐ZsGreen lentivirus vector for overexpression. Primers for lentivirus construction were list in Appendix Table S1.

IL‐6 and GP130 neutralization experiment

For IL‐6 and GP130 neutralization experiment, 25 μg of anti‐IL‐6 (Biolegend, Clone number, MP5‐20F3) or 10 μg anti‐GP130 neutralizing antibody (R&D System, Clone number, 125605) or anti‐IgG isotype control per mouse were injected intradermally 4 h before IMQ application, and the injection continued every other day. Mice were sacrificed at day 6, and skin tissues were collected for further analysis. Erythema, scaling, and thickness were scored daily on a scale from 0 to 4. The cumulative score was calculated by plus the scores of erythema, scaling, and thickness.

Recombinant protein purification

Recombinant Flag/HA‐GFP‐His protein (Flag/HA‐GFP), Flag‐GP130‐his protein (Flag‐GP130), and HA‐NEDD4L‐his protein (HA‐NEDD4L) were purified using Ni‐NTA‐Sefinose Column (Sangon biotec (ShangHai) Co., Ltd.) according to the manufacture’s protocol.

Co‐IP and WB

For endogenous co‐IP assay, WT or NEDD4L Knockout NHEKs were harvest and lysed in lysis buffer (5% glycerol, 1% Nonidet P‐40, 5 mM EDTA, 100 mM NaCl, 20 mM Tris–HCl [pH, 8.0], 1 mM phenylmethylsulfonyl fluoride). After centrifugation for 10 min under 12,000 g at 4°C, the supernatants were collected and subjected to overnight incubation with anti‐GP130 antibody (Abcam, USA) and protein G magnetic beads (Bio‐Rad). The beads were washed with TBS (50 mM Tris–HCl, 150 mM NaCl, PH 7.4) for three times, and the proteins were eluted by loading buffer for immunoblot assay. For exogenous co‐IP, GFP control, NEDD4L (or its domain mutants) and GP130 (or its domain mutants) plasmids were co‐transfected into HEK293T cells, and the cell lysates were harvest and processed as described in endogenous co‐IP. The Protein G magnetic beats were substitute by ANTI‐FLAG M2 beads. For WB analysis, the cells were harvested and were lysed in lysis buffer (5% glycerol, 1% Nonidet P‐40, 0.1% SDS, 5 mM EDTA, 100 mM NaCl, 20 mM Tris–HCl [PH,8.0], 25 mM β‐glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 μg/ml aprotinin, and 1 μg/ml leupeptin). After centrifugation, the cell lysis was subjected to SDS–PAGE for separation and subsequently, transferred onto nitrocellulose membranes and then for antibody detection. The antibodies and their dilutions used were listed in Appendix Table S2.

Immunofluorescence (IF) staining

WT and NEDD4L KO NHEK cells were fixed with 4% paraformaldehyde for 10 min, and the membranes were permeabilized in 0.3% Triton X‐100 for 20 min. After blocking with 5% bovine serum albumin and 3% goat serum for 1 h. The cell were then incubated with anti‐mouse GP130 antibody (SANTA CRUZ BIOTECHNOLOGY, E‐8, SC‐376280, 1:100) and anti‐rabbit NEDD4L antibody (Cell Signaling Technology, 4013S, 1:200) overnight at 4°C. After wash, the cells were incubated with Dylight 594‐goat anti‐rabbit IgG and Dylight 488‐goat anti‐mouse IgG and the images were harvested using fluorescence confocal microscope (Olympus IX81‐FV1000).

Ubiquitination assay

For endogenous GP130 ubiquitination assay, WT or NEDD4L Knockout NHEK cells were lysed and boiled for 5 min in lysis buffer supplemented with 1% SDS. After 5 min cooling in ice, the cell lysate was diluted for ten times with lysis buffer, and then centrifugation for 10 min under 12,000 g at 4°C, the supernatants were collected and subjected for overnight incubation with anti‐GP130 antibody (Abcam, USA) and protein G magnetic beads (Bio‐Rad). After incubation, the beads were washed three times with TBS (50 mM Tris–HCl, 150 mM NaCl, pH 7.4) and then eluted by loading buffer and subsequently, for SDS–PAGE separation. For exogenous GP130 ubiquitination assay, HEK293T cell were transfected with Flag‐tagged GP130, Myc‐tagged NEDD4L and HA‐tagged Ubiquitin or its mutants, 36 h later, 25 μM (Final concentration) of MG‐132 were added to inhibit the function of proteasome and 12 h later, cells were harvested and processed as described in endogenous ubiquitination assay. The Protein G magnetic beats were substitute by ANTI‐FLAG M2 beads. In vitro ubiquitination assay: In brief, 150 ng E1, 300 ng E2 (UbcH5α), 500 μg/ml Ubiquitin, 1 μg Flag‐GP130, and 750 ng Myc‐NEDD4L were reacted in ubiquitination buffer which contains 25 mM Tris–HCl, PH 7.6, 5 mM MgCl₂, 100 mM NaCl, 0.2 μΜ DTT, and 2 mM ATP for 2 h under 30°C. Loading buffer was added to stop the reaction and subsequently subjected to SDS–PAGE separation and WB detection. Proteins E1, E2 (UbcH5α), and ubiquitin were kindly provided by professor ZongPing Xia (Life Science Institute, Zhejiang University) for in vitro ubiquitination assay.

IHC staining

The skin tissues were fixed in 4% formalin, and all the immunohistochemistry assay were accomplished by the technicians of core facilities of Zhejiang University school of medicine. The staining extent was on a scale of 0–3, corresponding to the percentage of positive cells (0–10%, 11–25%, 26–75%, and 76–100%, respectively) and the staining intensity (negative, score = 0; weak, score = 1; strong, score = 2; very strong, score = 3). A staining value from 0 to 9 was calculated by multiplying the intensity score with the staining extent score for each specimen. The number of EZH2‐, Ki67‐, or p‐STAT3‐positive cells per high power filed (HPF) in each mouse was counted (at least 8 HPFs per mouse were counted). For correlation analysis, semiquantitative scoring for p‐STAT3 was performed on a scale from 0 to 4 (0, negative; 1, < 10 positive cells/HPF; 2, 10–50 positive cells/HPF; 3, 51–100; 4, > 100 positive cells/HPF).

Real‐time PCR

mRNA from the skin tissues and NHEK cells were extracted using TRIzol (Vazyme, China)/trichloromethane/isopropanol method. cDNA was synthesized from mRNA using SuperScript II reverse transcriptase (TAKARA, Beijing), according to the manufacture’s protocol. Real‐time PCR assays were designed to determine transcript levels of Keratin‐14, Keratin‐10, CyclinD1, C‐Myc, S100A7, IL‐23p19, IL‐22, IL‐17, IL‐6, Filaggrin, Loricrin, and Actin. Expression levels were measured using the LightCycler 480 II machine (Roche). Primers were list in Appendix Table S3.

ChIP‐Quantitative PCR assay

ChIP assays were performed according to the manufacturer’s instructions with a SimpleChIP Plus Enzymatic Chromatin IP Kit (Magnetic Beats) (Cell Signaling Technology, CST, #38191). Briefly, the primary mouse keratinocyte and NHEK cells were fixed by 1% formaldehyde. The cross‐linked chromatin was sonicated in a water bath at 4°C using Bioruptor UCD‐200 sonicator to obtain DNA fragments sized between 150 and 500 base pairs. Chromatin from 1 × 10⁶ cells was used for each ChIP experiment. Antibodies against H3K4me3, H3K9me3, H3K27me3, or H3K36me3 were used. The DNA from cross‐linking was purified for quantitative PCR analysis. Primers used for ChIP‐qPCR assay were list in Appendix Table S4.

NEDD4L and GP130 KO NHEK cell line generation

Small guide RNAs (gRNA) target NEDD4L or GP130 for knockout were designed and subcloned into PEP‐KO (PEP‐330X) vector. After transfection, NHEK cells were further screened using 1 μg/ml puromycin. The surviving cell was further transferred into 96‐well plate to form monoclonal cell lines. The KO cells were detected by real‐time PCR and WB analysis. The sequences for NEDD4L gRNAs were list in Appendix Table S5.

RNA interfering

For transient knockdown, small interfering RNA was transfected into NHEK cells using INTERFERin@ according to the manufacture’s protocol (Polyplus). siRNA oligonucleotide sequences were used: NEDD4L siRNA (5′‐GAGUCCUAUCGGAGAAUUATT‐3′), EZH2 siRNA (5′‐AAGACUCUGAAUGCAGUUGCUUU‐3′), EZH1 siRNA (5′‐AAUUGUCAUAGACCUUUCCGCTT‐3′), “Non‐sense” siRNA (Negative control) (5′‐UUCUCCGAACGUGUCACGUTT‐3′).

Cell culture and reagents

HEK293T, NHEK cells (Normal human epidermal keratinocytes, NHEK) were obtained from American Type Culture Collection. Cells were cultured at 37°C under 5% CO₂ in Dulbecco Modified Eagle Medium (DMEM) containing 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin (Ma et al, 2017). NHEKs were stimulated with IMQ (100 ng/ml), IL‐17A (100 ng/ml), IL‐22 (10 ng/ml) and IL‐6 (10 ng/ml) stimulation for the indicated times. Mouse primary keratinocyte was prepared as described previously (Lichti et al, 2008). Once the mouse primary keratinocytes reached confluence, they were stimulated by IL‐6 for indicated time. ANTI‐FLAG M2 beads were purchased from Sigma‐Aldrich.

IMQ‐induced skin hyperplasia

IMQ‐induced mouse skin hyperplasia was performed according to a previous work (van der Fits et al, 2009). Six‐ to eight‐week‐old WT/KO or WT/cKO mice received a daily topical application of IMQ cream (5%) on the shaved back and the ears for indicated days. To score the severity of the skin inflammation, an objective scoring system was developed baseing on the clinical Psoriasis Area and Severity Index (PASI). Erythema was scored using a scoring table with red taints and scaling was scored independently on a scale from 0 to 4 (0, none; 1, slight; 2, moderate; 3, marked; 4, very marked.). The thickening was measured using a caliper on a scale from 0 to 4 (0, none; 1, thickening 0–25%; 2, thickening 26–50%; 3, thickening 51–75%; 4, thickening ≥ 76%). The cumulative score served as the severity of inflammation was calculated by adding the score of erythema, scaling, and thickness (scale 0–12) (van der Fits et al, 2009).

Human samples

All the paraffin‐embedded skin sections form psoriasis and cutaneous squamous cell carcinoma (SCC) patients were obtained from Department of Dermatology and Venereology, Sir Run Run Shaw hospital affiliated to Zhejiang university medical college. Normal control skin sections consisted of healthy tissue from the resection edges of cutaneous SCC biopsies that appeared to be normal at the histological level. The expression of NEDD4L, GP130, and p‐STAT3 was detected by IHC staining and was scored 0–4 according to the staining intensity.

Statistical analysis

All data are expressed as the mean ± SEM. and are representative of at least two or three independent experiments. Statistical significance between two experimental groups was calculated using unpaired two‐tailed Student’s t‐test. For NHEKs proliferation, significant difference was tested by calculating the area under the curve using a two‐tailed Mann–Whitney U test. Correlation studies of immunohistochemically stained psoriasis clinical specimens were analyzed using the Pearson’s chi‐square test with a 95% confidence interval (CI) (Lin et al, 2017). P < 0.05 was considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001, NS, Non‐significant, P > 0.05).

Author contributions

Experiment design and data analysis: HL, XW, and WL; Biochemical and cellular studies: HL and ZL; Animal studies: HL and WL; Clinical samples: HC; Providing and analyzing RNA‐sequencing data: YS; Manuscript writing and revision: HL, XW and HA; Manuscript editing: All authors.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Expanded View Figures PDF

Source Data for Expanded View

Review Process File

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

Source Data for Figure 6

Source Data for Figure 7

EMBO reports (2021) 22: e52063.

Data availability

The datasets produced in this study are available in the following databases: RNA‐seq data: NCBI SRA: PRJNA687994 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA687994).