Investigation into the effect of skin tone modulators and exogenous stress on skin pigmentation utilizing a novel bioengineered skin equivalent

Abstract Human skin equivalents (HSEs) are a popular technology due to limitations in animal testing, particularly as they recapitulate aspects of structure and function of human skin. Many HSEs contain two basic cell types to model dermal and epidermal compartments, however this limits their application, particularly when investigating the effect of exogenous stressors on skin health. We describe the development of a novel platform technology that accurately replicates skin pigmentation in vitro. Through incorporation of melanocytes, specialized pigment producing cells, into the basal layer of the epidermis we are able to re‐create skin pigmentation in vitro. We observe apical distribution of melanin within keratinocytes and formation of supranuclear caps (SPNCs), only when the epidermal compartment is co‐cultured with a dermal compartment, leading to the conclusion that fibroblast support is essential for correct pigment organization. We also evaluate the commonly observed phenomenon that pigmentation darkens with time in vitro, which we further explore through mechanical exfoliation to remove a build‐up of melanin deposits in the stratum corneum. Finally, we demonstrate the application of a pigmented HSE to investigate drug modulation of skin tone and protection from UV‐induced damage, highlighting the importance of such a model in the wider context of skin biology.


| INTRODUCTION
Human 3D skin equivalents (HSEs) are in vitro bioengineered tissues that recapitulate aspects of skin structure and function. They are extensively used to study areas of skin health including: drug discovery, 1,2 disease modeling, 3,4 exogenous stressors [5][6][7][8] and aging. 9,10 HSEs are advantageous over traditional 2D culture, as recreating the architecture of the tissue is essential to mimicking its function. Although HSEs provide a more physiologically relevant platform, most only contain two distinct cell types, fibroblasts and keratinocytes, 11,12 limiting their application, as appendages and supporting cell types are imperative to the complex, dynamic response of skin to stimuli.
Melanocytes are specialized, pigment producing cells located in the basal layer of the epidermis. 13,14 Melanin pigments are deposited in specialized melanosomes and transferred to adjacent keratinocytes, 15,16 accumulating in the perinuclear area as deposits termed "supranuclear caps" (SPNCs), which protect DNA from UVinduced damage. 17,18 Understanding the role of melanins in health and disease is important not only due to their roles in photoprotection, solar lentigines 19,20 and skin cancers, 21,22 but also due to their role in many pigmentary disorders. 23,24 Although HSEs are now widely used, few models include supporting cell types such as melanocytes. This is disadvantageous as basic HSEs lack innate cellular protection provided by melanins, therefore, their response to external stimuli such as ultraviolet radiation (UVR) is not representative of in vivo tissue. There is need for a more complex, physiologically relevant HSE, which can be used as a robust and reproducible platform technology for a greater range of applications, particularly when investigating exposome impact, as exogenous stressors, including UVR and airborne pollution, are known to induce pigmentary changes. 25,26 Although a limited range of melanocyte-containing HSEs are available, existing systems are limited, by modeling the epidermal compartment only, [27][28][29][30][31] use of exogenous and often animal derived extracellular matrix (ECM) 32,33 or lacking ultrastructural characterization completely. 34 These factors limit their predictive value and physiological relevance through only partially modeling the skin's microenvironment, microanatomy and ultimately function.
While epidermal models offer a reductionist approach, the absence of a dermal compartment can impact experimental outcomes, as complex cellular interactions between the two compartments are lost. This is particularly important for melanocytes, as fibroblast interaction has been reported to impact melanocyte physiology and skin pigmentation. 35,36 Dermo-epidermal cross-talk is thought to affect pigmentation through secretion of soluble factors by fibroblasts 37 such as neurogulin-1, 38 Dikkopf-1 39 and keratinocyte growth factor (KGF). 40 Furthermore, contributions of dermal fibroblasts to basement membrane formation 41 and local ECM proteins 42 have also been reported to affect melanocyte function. For this reason, it is essential to consider both the epidermal and dermal compartments simultaneously.
Previously, we have reported the development of a novel, robust, full-thickness (FT) HSE and its thorough characterization in comparison to human skin. 43 Here, we outline modifications to this method, incorporating human melanocytes into the basal layer of the epidermis to produce a uniformly pigmented HSE. We have conducted a thorough ultrastructural characterization of melanosome distribution and consider the key role of fibroblasts in the regulation and distribution of melanin.
We also investigate the role of desquamation in melanin turnover, which, along with autophagy, has been reported to be integral to skin tone homeostasis. 44,45 Furthermore, we describe the application of this novel pigmented HSE to study the effects of stimuli including drugs and UVR on skin tone. This demonstrates model functionality and its responsiveness to known pigmentation modulators, highlighting potential future applications including the screening of cosmetics and investigations into the effects of UV-damage and photoageing on skin health.

| Development of pigmented skin equivalents
Initially, we optimized melanocyte inclusion in a simplified epidermal-only skin equivalent (EO-HSE), adapted from a previously described methodology. 43 Pigmented epidermal models appear darker ( Figure 1a) and histological analysis reveals pigmented cells in the stratum basale (Figure 1b: b1, b2). Expression of two melanocyte biomarkers: S100 (Figure 1b Individual typology angle (ITA) is a spectrophotometric measurement of skin tone, dependent upon melanin index (MI), that allows for a more reliable classification than conventional Fitzpatrick classification 46 and is routinely measured in a wide range of dermatology studies. 47 To develop a more complex and physiologically relevant HSE, we applied the optimized parameters to our previously described FT-HSE methodology. 43,49 This HSE is engineered utilizing a porous polystyrene scaffold populated with dermal fibroblasts that secrete endogenous ECM upon which an epidermis is constructed.
Inclusion of melanocytes resulted in a uniformly pigmented HSE Bernerd, 50 the non-pigmented HSE is comparable to "very light" human skin whereas the pigmented HSE would be categorized as "brown" to "dark."

| Desquamation is major regulatory process involved in melanin turnover in vitro
A well-accepted limitation of the use of pigmented HSEs is that they darken with time, thought to be due to impaired desquamation F I G U R E 1 Inclusion of melanocytes produces a pigmented epidermal equivalent. Gross appearance of nonpigmented and pigmented epidermal models (a), two examples shown, diameter 0.6 mm). Histological analysis of epidermal equivalents (b) through H&E staining (b1, b2) reveals the morphology of the epidermis and Fontana-Masson staining (b3, b4) exposes melanin distribution. Immunohistochemical analysis of S100 (b5, b6) and immunofluorescence of gp100 (b7, b8), melanocyte markers identify the location of melanocytes in the stratum basale. Keratin-14 is stained in green highlights basal keratinocytes, gp100 is stained red and DAPI highlights nuclei blue. ITA (c) was significantly reduced thus darker skin tone in pigmented models and similarly a significantly greater melanin index (d) in pigmented models. Data represent mean ± SEM, n = 6, ****p < 0.0001. Scale bars: 50 μm.
in vitro. 53 To better characterize our model, we investigated this in

| Bioengineered skin equivalents respond to pigment modulating stimuli
Kojic acid (KA) is a well-characterized active commonly used in skin care regimes that lightens skin tone through inhibition of tyrosinase, F I G U R E 2 Incorporation of melanocytes into full-thickness bioengineered skin produces a pigmented construct. Gross appearance of non-pigmented and pigmented full-thickness skin models (a), two examples shown, diameter 12 mm). Histological analysis of skin models in comparison to human skin (b) through H&E staining (b1-b3) and Fontana-Masson staining (b4-b6). Immunological analysis of melanocyte markers S100 (b7-b9) and gp100 (b10-b13) reveal the presence of melanocytes. Keratin-14 is stained in green highlights basal keratinocytes, gp100 is stained red and DAPI highlights nuclei blue. A view of the stratum basale confirms melanocyte distribution (insert, TRP1, red). Quantification of melanin content per model (c) (data represent mean ± SEM, n = 6) demonstrates large quantities of melanin within the pigmented skin equivalent. Similarly ITA (d) (data represent mean ± SEM, n = 18) is significantly reduced in pigmented models indicating darker skin tone. ***p < 0.001, ****p < 0.0001. Scale bars: 50 μm.
applications, investigations into the impact of exogenous stressors on skin health require a more complex system. Herein, we describe the development and in-depth characterization of a pigmented FT-HSE that recapitulates both the structure and distribution of melanin throughout the epidermis. We also demonstrate functionality through its responses to known stimuli including a potent skin lightening agent and UVR.
The formation of melanin SPNCs within keratinocytes is a wellknown phenomenon that protects nuclear DNA from UV-induced damage. 62  F I G U R E 6 Melanin provides innate protection from UVinduced damage in vitro. Gross appearance of non-pigmented and pigmented skin equivalents either non-irradiated (ÀUV) or irradiated 5 times with 3.3 J cm À2 (96.5% UVA, 3.5% UVB) (+UV) and harvested 48 h following exposure (a, diameter 12 mm). H&E (b) reveals epidermal structure (b: b1-b4) and Fontana-Masson (b: b5, b6) exposes melanin content of pigmented models. The epidermis of UV irradiated non-pigmented models is apoptotic with formation of "sunburn cells" (arrow heads), whereas epidermal structure is unaltered in pigmented models following UV irradiation. ITA (c) and melanin index (d) remain unchanged in non-pigmented models, whereas they decrease and increase, respectively, following UV exposure in pigmented models, indicating tanning (data represent mean ± SEM, n = 9). ****p < 0.0001. Scale bars: 50 μm.
We hypothesized that the formation of SPNCs and correct localization of melanocytes to the stratum basale relies on dermo-epidermal interactions. This is due to the observation that SPNCs only form in FT-HSEs as opposed to EO-HSEs. This notion is supported by wellestablished links between dermo-epidermal interactions and gross pigmentation; proposed mechanisms include secretion of fibroblastderived soluble factors, 67 direct physical interaction with the basement membrane, 41 biochemical interactions with ECM components 68 along with direct biophysical influences of the culture substrate. Further investigations are required to elucidate the specific mechanism, or mechanisms, that underpin dermal influence on epidermal pigmentation in this system, however these data do support the requirement of a robust dermal formation to recapitulate accurate pigmentation in vitro.
In this study, we have demonstrated not only the correct localization of melanin, but also functionality. When exposed chronically to UVR, we found that pigmented HSEs remained intact with healthy epidermal structure, whereas non-pigmented HSEs became apoptotic and displayed characteristic sunburn damage. 56 We also observed a tanning effect caused by increased melanogenesis within UVRexposed pigmented HSEs, demonstrating that our engineered HSE reacts to stimuli in-line with the known physiological response. We postulate that this innate protection offered by melanin prevents damage to the epidermis in pigmented HSEs, mediated by the formation of protective SPNCs.
Autophagy has often been considered the driving factor in epidermal melanin homeostasis. 69 However, Joly-Tonetti et al. 45 suggest an alternative process involving asymmetric melanosome distribution, and found no evidence of autophagy in healthy skin. Asymmetric distribution, in combination with desquamation and epidermal turnover, is hypothesized to be key in maintaining skin tone, a theory supported by findings of varying rates of desquamation in different skin tones 70 and skin tone lightening by agents that enhance desquamation. 44 We observed the well-known phenomenon of increasing pigmentation over time in HSEs, 71 and hypothesized that this was due to attenuated desquamation in vitro, 72 as a build-up of stratum corneum concentrates melanin, preventing keratinocyte turnover that occurs in vivo. To test this, we used a common dermatological technique of removing material from the stratum corneum surface by tape stripping. 73 Tape stripping removed layers of the stratum corneum to a point where barrier function was compromised and resulted in significant skin tone lightening, supporting the finding of Joly-Tonetti that desquamation is an important regulatory mechanism. Particularly in darker skin tones as melanin concentrates in the stratum corneum and demonstrated by the "darkly pigmented" melanocytes used to recapitulate pigmentation in this study. Irregular pigmentation is a consequence of many dermatological disorders including: photoageing, 74,75 post-inflammatory hyperpigmentation 76,77 and pigmentary changes such as melasma or senile lentigo. For this reason, many cosmetics contain agents known to modulate skin tone such as thiamidol, 78,79 niacinamide 80 and KA. 80 In vitro pre-clinical assessments have become an important tool to provide insight into the effect of formulations prior to progression to clinical trials. Herein, we describe the ability of KA, a wellcharacterized inhibitor of melanogenesis, to lighten the skin tone of our HSE, establishing its response to known actives and providing an example application. Although there is published literature on the use of 3D technologies to screen hypopigmenting agents, 65,81 some model systems including a commonly cited commercially available technology do not exhibit the expected response to KA. 82 We therefore demonstrate that our HSE not only recapitulates physiologically accurate pigmentation but, crucially, responds to stimuli in a manner akin to native skin.
Through this study we have presented a well-characterized FT-HSE and its thorough ultrastructural analysis. Characterization to this level is lacking in many other pigmented HSEs. We describe the correct localization of melanin, formation of protective SPNCs and the ability of such structures to protect against UVR-induced damage. Furthermore, we also demonstrate the ability of a well-known active to reduce melanin content and lighten skin tone. These data demonstrate the ability of the HSE to  HSEs were generated as previously described. 43,49 For pigmented HSEs, HEMn-DP were trypsinized and seeded simultaneously with HEKn onto a 28-day matured dermal compartment at a ratio of 1:10 (melanocytes:keratinocytes). Cultures were then maintained in submerged culture for 48 h and raised to the ALI to promote keratinocyte differentiation and stratification, for a further 14-days prior to use in experiments.

| UV exposure
Bio-Sun UV (Vilber Lourmat, Collégien, France) was used to expose HSEs to UVR. This system has a defined spectrum including a range of wavelengths that peak at 365 nm (UVA) or 312 nm (UVB). During irradiation the sample is exposed to a mixed spectra as a result of energy levels selected for both UVA and UVB. Delivery of both spectra begins in unison, however the time of delivery completion varies according to the dose selected for each type of radiation.
A physiologically representative UVR exposure was calculated using previously reported parameters to represent northern hemisphere countries including USA and UK. A UVA:UVB ratio of 27 was selected based on the calculated average daily UV spectrum published by Christiaens et al., 2005. 83 A terrestrial dose of 110-160 J cm À2 was selected representative of latitudes 30 to 60 83 and an average time of 10% daylight time spent outdoors assumed, which signifies average exposure of children and indoor workers. [84][85][86] An outdoor condition factor of 30% was also incorporated to account for environmental features. [85][86][87] This published data was collated and a representative UVR dose of 3.3-4.8 J cm À2 was calculated.
In this study, HSEs were irradiated every 24 h for 5 consecutive

| ITA & melanin index readings
Individual typological angle (ITA) is a colorimetric measure of skin tone phenotype, through which skin tone can be classified from "very light" to "dark" tones. This offers a more reliable greater range skin tone measurements compared with well accepted Fitzpatrick scaling. 50 ITA and MI readings were obtained using the SkinColourCatch (Delfin Technologies, Surrey, UK).

| Transepidermal water loss measurements
Transepidermal water loss (TEWL) was measured using the Vap-oMeter (Delfin Technologies). HSEs were allowed to equilibrate at room temperature for 20 mins prior to measurement.

| Immunofluorescence
Sections were deparaffinized in Histoclear and rehydrated through a series of ethanols. Antigen retrieval was performed in citrate buffer pH 6 (Sigma-Aldrich) at 95 C for 20 min, followed by blocking and permeabilization for 1 h with 20% neonatal calf serum (NCS, Sigma-Aldrich) and 0.4% Triton X-100 (Sigma-Aldrich) in phosphate buffered saline (PBS). Samples were then incubated overnight at 4 C in primary antibody diluted in blocking buffer (gp100, Abcam, ab137078,

| Epidermal whole mount staining
In order to visualize the distribution of melanocytes across the epidermis, the epidermal layer was separated from the underlying dermis through enzymatic digestion. The HSE was placed stratum corneum down in a solution of 500 U ml À1 Dispase (Sigma-Aldrich) and incubated at room temperature for 20 min. Following this the epidermis was peeled using forceps, from the dermis and washed twice in PBS. The epidermal sample was then blocked and permeabilized for 1 h with 20% NCS and 0.4% Triton X-100 in PBS. Samples were then incubated at room temperature for 2 h in primary antibody diluted in blocking buffer (TRP1, Abcam, ab6190709, 1:100). Samples were washed three times in PBS and incubated with the appropriate secondary antibody diluted in blocking buffer for 1 h at room temperature (donkey anti-mouse Alexa Fluor 594) and washed three times in PBS. Epidermises were then mounted on microscope slides using Vectashield.

| Light microscopy
Histology, Fontana-Masson and immunohistochemistry images were captured using Leica ICC50 high-definition camera and Brightfield microscope. Immunofluorescence images were taken using the Zeiss 880 confocal microscope with Airyscan and Zen software.

| Electron microscopy
Samples were prepared for TEM as previously described 43 and images captured on a Hitachi H7600 TEM.

| Melanin quantification assay
Quantification of melanin content was achieved using a previously published methodology. 88 4.14 | Statistical analysis GraphPad Prism software was used to measure the statistical significance by use of a Student's t-test or one-way ANOVA with Tukey's post-hoc as appropriate. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

| CONCLUSIONS
In this study, we have described the development of a complex, pigmented HSE that is able to appropriately modulate skin pigmentation in response to external stimuli such as UVR and drugs. We

CONFLICT OF INTEREST
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
No datasets were generated or analyzed during the current study.