IFN‐γ‐induced PD‐L1 expression on human melanocytes is impaired in vitiligo

Mounting evidence shows that the PD‐1/PD‐L1 axis is involved in tumor immune evasion. This is demonstrated by anti‐PD‐1 antibodies that can reverse tumor‐associated PD‐L1 to functionally suppress anti‐tumor T‐cell responses. Since type I and II interferons are key regulators of PD‐L1 expression in melanoma cells and IFN‐γ‐producing CD8+ T cells and IFN‐α‐producing dendritic cells are abundant in vitiligo skin, we aimed to study the role of PD‐1/PD‐L1 signalling in melanocyte destruction in vitiligo. Moreover, impaired PD‐1/PD‐L1 function is observed in a variety of autoimmune diseases. It is, therefore, hypothesized that manipulating PD‐1/PD‐L1 signalling might have therapeutic potential in vitiligo. The PD‐1+ T cells were abundantly present in situ in perilesional vitiligo skin, but expression of PD‐L1 was limited and confined exclusively to dermal T cells. More specifically, neither melanocytes nor other epidermal skin cells expressed PD‐L1. Exposure to IFN‐γ, but also type I interferons, increased PD‐L1 expression in primary melanocytes and fibroblasts, derived from healthy donors. Primary human keratinocytes only showed increased PD‐L1 expression upon stimulation with IFN‐γ. More interestingly, melanocytes derived from non‐lesional vitiligo skin showed no PD‐L1 upregulation upon IFN‐γ exposure, while other skin cells displayed significant PD‐L1 expression after exposure. In a vitiligo skin explant model, incubation of non‐lesional vitiligo skin with activated (IFN‐γ‐producing) T cells from vitiligo lesions was previously described to induce melanocyte apoptosis. Although PD‐L1 expression was induced in epidermal cells in these explants, this induction was completely absent in melanocytes. The lack of PD‐L1 upregulation by melanocytes in the presence of IFN‐γ‐producing T cells shows that melanocytes lack protection against T‐cell attack during vitiligo pathogenesis. Manipulating PD‐1/PD‐L1 signalling may, therefore, be a therapeutic option for vitiligo patients.


| INTRODUC TI ON
Vitiligo is a cutaneous pigment disorder characterized by white skin patches due to loss of melanocytes, the pigment-producing cells of the skin. 1 A research study has clarified immune-mediated melanocyte destruction in vitiligo-affected skin, caused by autoreactive melanocyte-specific CD8 + T cells. 2 Moreover, antibody responses against melanocyte antigens, such as tyrosinase and TRP-2, can be found in the sera of patients with vitiligo. 3 Additionally, an imbalance of CD4 + /CD8 + T-cell ratio and regulatory T cells might contribute to T-cell-mediated pathogenesis of vitiligo. 4 Finally, T cells with a resident phenotype (tissue-resident memory T (T RM ) cells) contribute to disease development and flare-up in human vitiligo.
CD8 + T RM cells were enriched in vitiligo-affected skin compared to healthy donor skin, 5,6 indicating immune disturbance in patients with vitiligo. Immune checkpoints, for example programmed cell death 1 (PD-1), modulate immune responses by regulating peripheral tolerance, thereby preventing unwanted inflammatory responses and autoimmunity. 7 Immunotherapy for melanoma shows that anti-PD-1 antibodies can break peripheral tolerance and thereby reactivate anti-tumor responses in patients. 8 Moreover, in 2-43% of patients with melanoma treated with immunotherapy, vitiligo-like depigmentation occurs, which indicates the presence of immune reactivity against both melanoma cells and melanocytes. [9][10][11][12][13] Vitiligo development in patients with melanoma has been associated with a favourable prognosis. [9][10][11][12] As vitiligo can also develop after treatment with anti-PD-1 antibodies, this suggests the involvement of PD-1 and its ligand programmed cell death ligand 1 (PD-L1) in the regulation of tolerance at the level of melanocytes.
Therapeutically targeting PD-1 or PD-L1 seems a promising treatment strategy for human vitiligo. 14 Impaired PD-1/PD-L1 function is involved in a variety of autoimmune diseases, for example type I diabetes and rheumatoid arthritis, 15 indicating the rationale to test the therapeutic potential of increasing PD-1/PD-L1 signalling in autoimmunity. Considering that IFNγ and type I interferons play a crucial role in vitiligo pathogenesis 2,16,17 and both are known to induce immune evasion by PD-L1 expression in melanoma, 18,19 the question arises to what extent PD-L1 upregulation and resulting immune regulation occur in vitiligo.
To our knowledge, in situ PD-1 expression and (cytokine-induced) PD-L1 expression in vitiligo remain incompletely studied. This study, therefore, aimed to study in situ PD-1/PD-L1 expression in human vitiligo specimens. Second, we aimed to investigate the role of PD-1/ PD-L1 checkpoint signalling in melanocyte destruction in vitiligo and how this is influenced by interferons. Our results show lack of (IFNγ-induced) PD-L1 expression by melanocytes from vitiligo patients, in contrast to healthy donor melanocytes, and point to an insufficient ability of melanocytes of vitiligo patients to protect themselves against autoreactive T cells.

Stable 6
Abbreviation: SEM, standard error of mean.

| Quantification of immunohistochemical staining analyses of tissue sections
Assessing the presence and multitude of marker-expressing cells Images (magnification ×400) were acquired on a Leica DM microscope using Leica software (Leica Biosystems).

| Tissue preparation
Skin biopsies of discarded breast and abdomen were dissociated into Additionally, primary healthy keratinocytes were kindly provided by the S. Gibbs Lab.

| Cytokine exposure
One day prior to exposure, cells were seeded in a 6-well plate at a density of 2 × 10 5 cells per well in a total volume of 2 mL medium with supplements and culture conditions as stated above. After 24 h, cells were left unexposed or exposed for 48 h with 500 U/mL IFNγ

| Flow cytometry
To detect the effect of cytokine exposure, cells were stained for flow cytometry analysis. The following antibodies were used: APC-

| Significant PD-1 + T cells but limited PD-L1 + cells in perilesional vitiligo skin biopsies
To explore the role of PD-1/PD-L1 in human vitiligo, we studied in situ PD-1 and PD-L1 expression in perilesional human vitiligo skin (see patient characteristics in Table 1). CD3 + cells were observed in all perilesional vitiligo skin biopsies, both dermal and epidermal (  Figure 1A).  Figure 1D). If PD-L1 + cells were observed, even though very few cells did, 78% of PD-L1-expressing cells colocalized with T-cell infiltrates (Table 2, Figure 1E). Most importantly, PD-L1 expression could not be detected on epidermal melanocytes in perilesional skin biopsies (Table 2, Figure 1D), indicating that neither in situ melanocytes nor other epidermal skin cells of vitiligo patients expressed PD-L1, while PD-1-expressing T cells were present.

| Type I and II interferons induce PD-L1 expression in healthy donor skin cells
Since in situ PD-L1 expression was absent in perilesional vitiligo skin and PD-1 + T cells were present, we investigated to what extent PD-L1 expression on human skin cells can be influenced by cytokines present in the skin micro-environment. It is known that CD8 + T cells in (peri)lesional vitiligo skin produce high levels of IFNγ 2,21 and that IFNγ can induce PD-L1 expression in melanoma. 18 Similarly, IFNα-producing plasmacytoid dendritic cells infiltrate active vitiligo skin, 16,17 and although PD-L1 expression is primarily upregulated by IFNγ in melanoma, stimulation with type I interferons also led to increased PD-L1 expression. 19 Exposure to IFNγ, but also type

| IFNγ exposure does not induce PD-L1 expression on primary vitiligo melanocytes
As shown, PD-L1 can be upregulated in various cell types in healthy skin, including melanocytes, upon exposure to inflammatory IFNγ Since PD-L1 is upregulated upon IFNγ exposure in vitiligo skin cells, this impaired response seems to affect melanocytes only and will, therefore, probably not be detected in skin transcriptome analysis of whole skin biopsies but in scRNAseq analyses of melanocytes only.
Together, this suggests a role for PD-L1 in the lack of peripheral tolerance in vitiligo, providing a rationale to target this axis in vitiligo treatment.
In conclusion, this study shows the absence of (IFNγ-induced) PD-L1 expression by melanocytes of vitiligo patients and thereby reveals a potential intrinsic melanocyte defect in vitiligo. This study also highlights the presence of PD-1-expressing T cells in vitiligo skin, which strengthen the notion of manipulating PD-1/PD-L1 signalling to induce peripheral tolerance to melanocytes in vitiligo.

ACK N OWLED G EM ENTS
This study is supported by TRANSCAN JTC 2014 through the Dutch Cancer Society under grant number 7800.

CO N FLI C T O F I NTE R E S T
CJMM was employed by ISA Pharmaceuticals. The authors declare that the research study was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

S U PP O RTI N G I N FO R M ATI O N
Additional supporting information may be found in the online version of the article at the publisher's website.