Type-1 cytokines regulate MMP-9 production and E-cadherin disruption to promote melanocyte loss in vitiligo

Abstract

Loss of melanocytes is the pathological hallmark of vitiligo, a chronic inflammatory skin depigmenting disorder induced by exaggerated immune response, including autoreactive CD8 T cells producing high levels of type 1 cytokines. However, the interplay between this inflammatory response and melanocyte disappearance remains to be fully characterized. Here, we demonstrate that vitiligo skin contains a significant proportion of suprabasal melanocytes, associated with disruption of E-cadherin expression, a major protein involved in melanocyte adhesion. This phenomenon is also observed in lesional psoriatic skin. Importantly, apoptotic melanocytes were mainly observed once cells were detached from the basal layer of the epidermis, suggesting that additional mechanism(s) could be involved in melanocyte loss. The type 1 cytokines IFN-γ and TNF-α induce melanocyte detachment through E-cadherin disruption and the release of its soluble form, partly due to MMP-9. The levels of MMP-9 are increased in the skin and sera of patients with vitiligo, and MMP-9 is produced by keratinocytes in response to IFN-γ and TNF-α. Inhibition of MMP-9 or the JAK/STAT signaling pathway prevents melanocyte detachment in vitro and in vivo. Therefore, stabilization of melanocytes in the basal layer of the epidermis by preventing E-cadherin disruption appears promising for the prevention of depigmentation occurring in vitiligo and during chronic skin inflammation.

Keywords: Autoimmune diseases; Cytokines; Dermatology; Inflammation; Skin.

Conflict of interest statement

Conflict of interest: NB, CJ, FXB, JS, and KB are inventors on the patent application “New therapies in the prevention or treatment of a depigmenting disorder” (EP16306719.2; filing date: December 16, 2016).

Figures

Figure 1. Melanocyte loss results from defective adhesion of melanocytes and not from apoptosis.

(A) Representative immunofluorescence analysis showing melanocytes (red, Melan-A) and basal layer of epidermis stained with an anti–type VII collagen antibody (green) in control healthy skin, vitiligo perilesional skin, skin obtained from patients with concomitant vitiligo and psoriasis, and lesional psoriatic skin. Dashed lines represent dermoepidermal layer. Arrows show suprabasal melanocytes in different conditions. Scale bars: 50 μm (top), 20 μm (bottom). (B) Proportion of suprabasal melanocytes in control healthy skin (n = 5), stable or active vitiligo perilesional skin (n = 18; black squares: stable vitiligo, red squares: active vitiligo), perilesional skin of patients with association of vitiligo and psoriasis (n = 4), and lesional psoriatic skin (n = 4). (C) Representative analysis of epidermal cell death using cleaved caspase-3 antibody (green). Melanocytes were stained with anti-MITF antibody (red) in control healthy skin, stable and active vitiligo perilesional skin, or lesional skin from psoriasis, cutaneous lupus erythematous, and toxic epidermal necrolysis. Dashed lines represent dermoepidermal layer. Scale bar: 20 μm. (D) Proportion of cleaved caspase-3+ MITF+ basal (circles) or suprabasal (triangles) melanocytes in control healthy skin (n = 3), stable (n = 4) or active vitiligo perilesional skin (n = 6), lesional skin of psoriasis (n = 6), cutaneous lupus erythematous (n = 6), and toxic epidermal necrolysis (n = 3). (E) Representative immunofluorescence analysis of expression of E-cadherin (green) and melanocytes (red, Melan-A staining) in control healthy skin, perilesional stable or active vitiligo skin, and lesional psoriatic skin. Staining is representative of 10 independent patients. Dashed lines represent dermoepidermal layer. Arrows identify suprabasal melanocytes. Scale bars: 50 μm (right) and 10 μm (left). (F) Assessment by ELISA of soluble E-cadherin levels in the sera of healthy controls (n = 18) or patients with stable (n = 37), progressive (n = 38) vitiligo, or psoriasis (n = 20). Data show mean ± SEM. *P < 0.05, **P < 0.01; calculated with a Kruskal-Wallis test.

Figure 2. Combined activity of IFN-γ and TNF-α reproduces human vitiligo features in a 3D model of reconstructed pigmented epidermis in vitro.

(A–C) Reconstructed human pigmented epidermis (RHPE) containing melanocytes were stimulated for 24 hours in the presence or absence of 10 ng/mL of TNF-α and IFN-γ, alone or in combination. (A) Representative immunofluorescence analysis of Melan-A (red) and E-cadherin (green) expression. Alteration of E-cadherin expression around melanocytes. Dashed lines represent dermoepidermal layer. Scale bars: 20 μm (left), 10 μm (right). (B) Proportion of suprabasal melanocytes in the different conditions (n = 4). (C) Representative analysis of epidermal cell death using a TUNEL assay (green); melanocytes were stained with anti-Melan-A antibody (red). Dashed lines represent the dermoepidermal layer. Scale bar: 20 μm. Staining is representative of 3 independent experiments. (D) Primary cultures of normal human epidermal keratinocytes (NHEK, left, n = 10) or melanocytes (NHEM, right, n = 11) were treated for 24 hours with 20 ng/mL of IFN-γ and/or TNF-α. CDH1 gene expression in epidermal cells was analyzed by real-time PCR. Results are shown as the percentage of change compared with the control culture. (E) Dose-response study of CDH1 gene expression in NHEM after 24 hours stimulation with TNF-α and/or IFN-γ. (F) Kinetic analysis of CDH1 gene expression in NHEM in response to 20 ng/mL of TNF-α and/or IFN-γ. Results from 1 experiment are shown in E and F and are representative of 4 independent experiments with 4 independent donors. GAPDH was used as a housekeeping gene. (G) Confocal microscopy analysis of RHPE treated in the presence or absence of the combination of 10 ng/mL of TNF-α and IFN-γ for 24 hours. Sections were stained for E-cadherin (green) and LAMP-1 (a marker for lysosomes and late endosomes, red). Merge shows the presence of E-cadherin molecule into LAMP-1 vesicle structures. Scale bars: 20 μm (top); 10 μm (bottom). Stainings are representative of 3 independent experiments. (H) Assessment by ELISA of soluble E-cadherin levels in cell-free supernatants of RHPE treated for 24 hours in the presence or absence of 10 ng/mL of TNF-α and IFN-γ. Data in B, D, and H show mean ± SEM. *P < 0.05, **P < 0.01; calculated with 2-tailed Mann-Whitney (B) or Wilcoxon (H) tests.

Figure 3. MMP-9 levels are increased in patients with vitiligo and correlate with soluble E-cadherin levels and surface of depigmentation.

(A) ELISA levels of MMP-9 in the sera of healthy controls (n = 18), patients with stable (n = 37) or active (n = 37) vitiligo, and patients with psoriasis (n = 20). (B) Active MMP-9 levels in the sera of healthy controls (n = 22), patients with stable (n = 30) or active (n = 39) vitiligo, and patients with psoriasis (n = 19). (C) Spearman’s rho correlation (2-tailed) between MMP-9 and soluble E-cadherin levels in the sera of patients with vitiligo (n = 73). (D) Spearman’s rho correlation (2-tailed) between serum active MMP-9 and body surface area (BSA) involved in patients with vitiligo (n = 59). (E) Representative IHC staining of MMP-9 expression in healthy control skin, perilesional skin of stable and active vitiligo, and lesional psoriatic skin. Scale bar: 100μm. (F) Semiquantitative analysis of MMP-9 expression in skin from healthy controls (n = 5), perilesional skin of vitiligo patients with stable (n = 11) or active (n = 10) disease, and lesional psoriatic skin (n = 8). (G and H) Inflammatory transcriptomic profile of perilesional skin of patients with stable (n = 3) and active (n = 6) vitiligo was assessed using NanoString technology. (G) The most upregulated genes are shown. Results show the change in gene expression between the 2 groups. (H) Predicted protein-protein interaction networks for upregulated genes using STRING online tool. The thickness of edges represents the strength of data support. The thicker the edge between 2 proteins, the more these proteins are linked based on the enrichment evidenced by STRING. (I and J) Reconstructed human pigmented epidermis (RHPE), NHEK, and NHEM were stimulated for 24 hours in the absence or presence of TNF-α and IFN-γ. (I) Real-time PCR analysis of MMP9 gene expression in RHPE (n = 7), NHEK (n = 9), and NHEM (n = 7). Data are shown as fold increase above the control culture. GAPDH was used as a housekeeping gene. (J) Levels of MMP-9 in cell-free culture supernatants of RHPE (n = 13, left) and NHEK (n = 9, right). (K) Western blot analysis of MMP-9 expression in NHEK treated for 24 hours in the presence or absence of 20 ng/mL of TNF-α and IFN-γ. Data in A, B, F, I, and J show mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001; calculated with Kruskal-Wallis (A, B, and F) or Wilcoxon tests (I and J).

Figure 4. MMP-9 inhibition allows melanocyte stabilization both in vitro and in vivo.

(A–C) Reconstructed human pigmented epidermis (RHPE) were treated for 24 hours in the presence or absence of increasing concentrations of active MMP-9. (A) Representative immunofluorescence staining of Melan-A (red) and E-cadherin (green). Dashed lines represent the dermoepidermal layer. Arrows show suprabasal melanocytes in the different conditions. Scale bar: 20 μm. (B) Proportion of suprabasal melanocytes in the different conditions. (C) Assessment by ELISA of soluble E-cadherin levels in cell-free supernatants. (D–F) RHPE were treated for 24 hours in the presence or absence of 10 ng/mL of TNF-α and IFN-γ and/or 1, 10 or 100 μM of MMP-9 inhibitors Ab142180 or SB-3CT. (D) Representative immunofluorescence staining of Melan-A (red) and E-cadherin (green). Dashed lines represent the dermoepidermal layer. Arrows show suprabasal melanocytes in the different conditions. Scale bar: 20 μm. (E) Proportion of suprabasal melanocytes in the different culture conditions. (F) Assessment by ELISA of soluble E-cadherin levels in cell-free supernatants. (G–I) The base of C57BL/6 mouse tail was treated daily for 6 days with intradermal injections of saline buffer (control), or the combination of 1 μg of TNF-α and IFN-γ and/or 1.25 mg/mL of SB-3CT. (G) In vivo schema of C57BL/6 mice treatment. (H) Representative immunofluorescence analysis of Melan-A (red) and E-cadherin (green) staining in the different groups. Dashed lines represent the dermoepidermal layer. Arrows show suprabasal melanocytes. Scale bars: 20 μm. (I) Proportion of suprabasal melanocytes was assessed in the different groups (n = 7–9). Data in B, C, E, F, and I show mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001; ****P < 0.0001, calculated with 2-tailed Mann-Whitney test.

Figure 5. Inhibition of JAK prevents type 1 cytokine-mediated melanocyte detachment both in vitro and in vivo.

(A–D) Reconstructed human pigmented epidermis (RHPE) were treated for 24 hours in the presence or absence of 10 ng/mL of TNF-α and IFN-γ and/or 0.1 or 1 μM of tofacitinib or ruxolitinib. (A) Representative immunofluorescence staining of Melan-A (red) and E-cadherin (green). Dashed lines represent the dermoepidermal layer. Arrows show suprabasal melanocytes in the different conditions. Scale bar: 20 μm. (B) Proportion of suprabasal melanocytes in the different culture conditions. (C and D) Assessment by ELISA of (C) soluble E-cadherin or (D) active MMP-9 levels in cell-free supernatants. (E) Vitiligo perilesional epidermis were treated for 24 hours in the presence or absence of 1μM of tofacitinib. MMP-9 levels were assessed by ELISA in cell-free supernatants (n = 3). (F and G) C57BL/6 mouse tails were treated by intradermal injection of saline solution (control), 1 μg TNF-α and IFN-γ, and/or 1 mM of tofacitinib or ruxolitinib, according to the same protocol described in Figure 4F. (F) Representative immunofluorescence analysis of Melan-A (red) and E-cadherin (green) staining in the different groups. Dashed lines represent the dermoepidermal layer. Arrows show suprabasal melanocytes. Scale bars: 20 μm. (G) Proportion of suprabasal melanocytes was assessed in the different groups (n = 6–10). Data in B–D and G show mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001; ****P < 0.0001, calculated with 2-tailed Mann-Whitney test.

Figure 6. Putative model of primary event leading to loss of melanocytes in depigmenting disorders.

Type 1 cytokines TNF-α and IFN-γ produced by activated Trm cells induce an E-cadherin defect in melanocytes. TNF-α and IFN-γ induce the production of MMP-9 by epidermal cells, especially keratinocytes, that cleave E-cadherin (Ecad) to release its soluble form. E-cadherin cleavage leads to melanocyte destabilization. This effect is inhibited in the presence of MMP-9 or JAK inhibitors.