Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 May 1.
Published in final edited form as: J Invest Dermatol. 2021 Oct 29;142(5):1360–1371.e15. doi: 10.1016/j.jid.2021.09.031

Transcriptome analysis reveals intrinsic pro-inflammatory signaling in healthy African American skin

Anna Klopot 1, Gleb Baida 1, Alexander Kel 2,3, Lam C Tsoi 4,5,6, Bethany E Perez White 1, Irina Budunova 1,*
PMCID: PMC9038646  NIHMSID: NIHMS1766483  PMID: 34757068

Abstract

Differences in morphology and physiology of darkly pigmented compared to lightly pigmented skin are well recognized. There are also disparities in prevalence and clinical features for many inflammatory skin diseases including atopic dermatitis and psoriasis; however, the underlying mechanisms are largely unknown. We compared the baseline gene expression in full thickness skin biopsies from healthy individuals self-reporting as African American (AA) or White Non-Hispanic (WNH). Extensively validated RNA-Seq analysis identified 570 differentially expressed genes (DEG) in AA skin including immunoglobulins and their receptors such as FCER1G; pro-inflammatory genes such as TNFα, IL-32; EDC (epidermal differentiation cluster) and keratin genes. DEGs were functionally enriched for inflammatory responses, keratinization, cornified envelope formation. RNA-seq analysis of 3D human skin equivalents (HSE) made from AA and WNH primary keratinocytes revealed 360 DEGs (some shared with skin) which were enriched by similar functions. AA HSE appeared more responsive to TNFα pro-inflammatory effects. Finally, AA-specific DEGs in skin and HSE significantly overlapped with molecular signatures of skin in AD and psoriasis patients. Overall, these findings suggest the existence of intrinsic pro-inflammatory circuits in AA keratinocytes/skin that may account for disease disparities and will help to build a foundation for the development of targeted skin disease prevention.

Keywords: transcriptomics, RNA-Seq, African American skin, skin barrier, inflammation, inflammatory skin diseases

Introduction

With the changing demographics of the US population, there is a need to understand the mechanisms, prevention, and treatment approaches of dermatologic diseases that have increased incidence and unique manifestations in non-White skin types.

Differences in morphology and physiology of darkly pigmented skin are recognized (Halder and Nootheti, 2003). There are also disparities in the prevalence, clinical features, and treatment response for skin cancer, keloids (abnormal skin scarring), and many inflammatory skin diseases including atopic dermatitis (AD), psoriasis, acne, hidradenitis suppurativa, lupus erythematosus (Agbai et al., 2014, Alexis and Blackcloud, 2014, Feldman et al., 2013, Halder and Nootheti, 2003, Mei-Yen Yong and Tay, 2017, Price et al., 2019). Subjects with light skin have greater risk of skin cancer and psoriasis, although the overall severity of psoriasis may be higher in non-White patients (Yan et al., 2018). In contrast, AD prevalence is higher in the African American (AA) compared with White non-Hispanic (WNH) population, especially in AA children in the US (Brunner and Guttman-Yassky, 2019), and in African descent children in Europe (Kaufman et al., 2018). The disparity may be even greater than described, given that visualization and scoring of skin erythema is challenging in darker skin types (Kaufman et al., 2018). AA patients with AD have higher serum IgE levels, and different immune polarization (decreased Th1/Th17 activation) (Brunner and Guttman-Yassky, 2019). In addition, AA patients with moderate-to-severe AD may be less responsive to the treatments including IL4/IL13 signaling inhibitor Dupilumab (Alexis et al., 2019).

Although skin pigmentation itself could explain protection against skin cancer (Yin et al., 2014), the increased propensity towards certain inflammatory skin diseases seems to be melanogenesis-independent. Overall, the molecular mechanisms underlying differences in inflammatory skin diseases susceptibility/severity, and the differential response to therapeutic and side effects of anti-inflammatory drugs are poorly understood.

We assumed that the molecular signaling circuits in healthy skin is an important determinant contributing to the risk of skin diseases development. The goal of this work was to define the transcriptome of AA healthy skin. We compared the baseline gene expression in full thickness healthy skin biopsies from individuals self-reporting African American (AA) and White Non-Hispanic (WNH). RNA-seq analysis revealed hundreds of differentially expressed genes in AA skin that were highly enriched for immune responses, formation of the cornified envelope and keratinization. The similarly large differences in global gene expression linked to inflammation and skin barrier function were found in primary human AA compared to WNH keratinocytes growing in 3D human skin equivalent (HSE) cultures. In addition, the AA HSE appeared to be more prone to inflammatory effects of cytokine TNFα. Moreover, DEGs in AA skin and AA HSE significantly overlapped with molecular signatures of AD and psoriasis.

Results

1. Differences in healthy African American and White Non-Hispanic skin transcriptome

We recently noticed significant differences in AA and WNH skin transcriptome (Lili et al., 2019). We have now advanced this observation using deep bioinformatics re-analysis and validation of differential gene expression in whole thickness skin biopsies from 17 healthy volunteers: 9 AA and 8 WNH subjects of both sexes (grouping based on Fitzpatrick score combined with selfidentification, see Materials and Methods). There were five females and three males in WNH group; five females and four males in AA group (the demographics is presented in Supplemental Table S1).

The analysis of RNA-seq results identified 570 differentially expressed genes (DEGs, p-value<0.05 and FC/fold change>1.5) in AA versus WNH skin (Fig. 1, Supplemental Table S2). Interestingly, ~40% DEGs in AA skin belong to different classes of non-coding RNAs, including long non-coding RNAs (lincRNAs, anti-sense RNAs, sense overlapping RNAs, sense intronic RNAs), short non-coding RNAs (miRNAs, rRNAs, Mt RNAs, small nuclear RNAs) and pseudogenes (Fig. 1a, Supplemental Table S1, Supplemental Fig. S1). The abundance of non-coding RNAs suggests that the differential gene expression in AA compared to WNH skin is regulated at both transcriptional (via effect of transcription factors, RNA splicing, subcellular localization and stability (Hombach and Kretz, 2013, Hu et al., 2018) and post-transcriptional levels.

Figure 1. Transcriptome differences in African American (AA) and White Non-Hispanic (WNH) healthy skin.

Figure 1.

Open in a new tab

Total RNA was extracted from 17 full thickness skin biopsies (9 AA and 8 WNH samples) and used for RNA-sequencing. (a). Heatmap of 50 most differentially expressed genes (P<0.05, FC >1.5) (rows) in individual RNA samples (columns) from AA versus WNH skin. (b). qRT-PCR validation of selected DEGs in individual AA (red circles) versus WNH (blue squares) skin RNA samples. RPL27 was used as a cDNA normalization control. Statistical analysis was performed using unpaired two-tailed t test. (c). Results of GSEA pathway enrichment analysis of normal skin DEGs with Hallmark and Reactome molecular signatures.

Importantly, AA skin DEG set included only four known pigment-related genes (KRT17, KRT75, LMX1A, SLC7A11) compared with 170 cloned pigment genes (www.espcr.org/micemut/), suggesting that overall differential gene expression in AA skin did not correlate with melanogenesis/skin color.

DEG families in skin, their functional enrichment by GSEA

DEGs upregulated in AA skin included numerous genes involved in inflammation and immune response/innate immunity: TNFα, interleukins (IL32, LIF), cytokines and their receptors (CCL2, CCL13, CCL18, CCL21, CCR1, CCR5), CD (cluster of differentiation) antigens; interferon-induced proteins (IFI6, IFI27, IFI44L). Several immunoglobulin E receptors were also overexpressed in AA skin: FCER2 essential for IgE production and FCER1G, the major receptor for IgE, a key immunoglobulin involved in atopic diseases, and elevated in serum of AD patients. In addition, ~ 60 DEGs encode light and heavy immunoglobulin chains (Fig. 1, Supplemental Table S2).

The other large group of upregulated DEGs was from the epidermal differentiation cluster (EDC) (de Guzman Strong et al., 2010), encoding late cornified envelope (LCE) proteins 1D, 1E, 1F, 2A, 3E, 5A and small proline rich proteins (SPRR1B and 5) important for skin barrier function (Fig. 1, Supplemental Table S2). Genes encoding some collagens were also differentially upregulated.

Among down-regulated DEGs were multiple keratins and keratin-associated proteins (Supplemental Table S2), trichohyalin (THCC) and its analog TCHHL1, all of which are important for hair morphogenesis and skin barrier organization (Harland and Plowman, 2018). The peptidyl arginine deiminase genes, PADI1, 3 and 4, involved in deamination of filaggrin (FLG), keratin 1, and trichohyalin were also down-regulated in AA skin. Finally, several olfactory receptor (OR) family members were among down-regulated DEGs in AA skin. ORs can stimulate hair follicle growth and keratinocyte proliferation (Cheret et al., 2018). We extensively validated RNA-seq results by qRT-PCR of representative DEGs from gene groups/families described above (Fig. 1b).

Sex could play an important role in gene expression in different human tissues including skin. Thus, in current work to identify DEGs, we implemented a linear model (limma R package) adjusted for sex and age as described previously (Lili et al., 2019). To provide the additional proof that in the volunteer cohort under study, sex has not significantly affected gene expression, we re-analyzed Q-PCR data used for AA DEG validation. We found that the expression of all validated genes differentially expressed in AA skin (Fig. 1b) was similar in male and female skin (Supplemental Fig. S2). Thus, our findings on differential gene expression in AA skin are not confounded by sex of participants

To evaluate systemic effects of dimorphic gene expression in skin, we performed GSEA pathway enrichment analysis via comparison with Hallmark gene sets (consisting of genes with coherent expression) and Reactome molecular signatures (gene sets derived from online pathway databases and PubMed publications). The major functions of upregulated DEGs were inflammation, interferon-alpha and -gamma responses, and allograft rejection signaling along with epithelial-mesenchymal transition (EMT). The down-regulated DEGs were enriched by keratinization and formation of cornified envelope processes (Fig. 1c).

2. Differential gene expression in 3D human skin equivalents (HSE) made from AA and WNH keratinocytes closely resembles skin transcriptome dimorphism

Skin consists of multiple cell types: keratinocytes, melanocytes, dermal fibroblasts, adipocytes, immune and endothelial cells. To address keratinocyte contribution to skin transcriptome, we employed HSE made from 4 individual cultures of AA and 5 individual cultures of WNH primary neonatal human epidermal keratinocytes, NHEK. All HSE were maintained at liquid/air interface for 11 days and had completely developed multilayer epidermis without apparent morphological differences between AA and WNH HSEs.

The analysis of RNA-seq results revealed 363 DEGs (P≤0.05, FC ≥1.5): 169 upregulated and 194 down-regulated in AA HSEs (Fig. 2a, Supplemental Table S3). As in skin, the significant amount of DEGs (~ 20%) was presented by non-coding RNAs including long non-coding (linc) RNAs, short non-coding and pseudogenes (Supplemental Fig. S1).

Figure 2. Differential gene expression in control 3D human skin equivalents (HSE) made from AA and WNH NHEK.

Figure 2.

Open in a new tab

HSE made from AA (4 individual cultures) and WNH (5 individual cultures) normal human epidermal keratinocytes (NHEK) were maintained at air-liquid interface for 11 days to allow for epidermal maturation. Experiment was repeated three times. Extracted RNA was used for RNA-sequencing.

(a). Heatmap of 50 top most differentially expressed genes (P<0.05, FC >1.5) (rows) in AA and WNH HSEs (columns). (b). qRT-PCR validation of DEGs in individual AA HSEs (red circles) versus WNH HSEs (blue squares). RPL27 was used as normalization control. Statistical analysis was performed using unpaired two-tailed t test. (c). Results of GSEA pathway enrichment analysis of AA HSE DEGs with Hallmark and Reactome. (d). Venn diagram showing common DEGs for AA human skin (Versus WNH skin) and AA HSE (versus WNH HSE) transcriptome.

DEGs upregulated in AA HSEs include multiple inflammation-related genes: chemokines (CCL5, CCL20); interleukin receptors (IL27RA, ILR7); interferon-induced proteins (IFIT1, IFI44, IFITM2, SLFN11), antigen-presenting major histocompatibility complex genes (HLA-C, HLA-DRB1, CIITA) along with ERAP1, an aminopeptidase involved in trimming HLA class I-binding precursors. In addition, several structural genes were upregulated in AA HSEs: collagens (COL1A1, COL22A1, COL4A1, COL5A3), collagen cross-linking lysyl oxidase (LOXL2); keratinocyte late differentiation markers (LCE1D, LCE1E, SPRR3) (Fig. 2a,b, Supplemental Table S3).

At the same time, the expression of some keratins (KRT4, 23, 78), potassium channels (KCN), transmembrane serine proteases (TMPRSS); carboxylesterases (CES1, 2, 3) involved in fatty acid/cholesterol/xenobiotics metabolism was lower in AA HSEs (Fig. 2, Supplemental Table S3). qRT-PCR experiments confirmed dimorphic expression of selected DEGs in HSEs (Fig. 2b).

More detailed GSEA analysis of AA DEGs by GO identified a large set of ~ 20 down-regulated DEGs related to lipid metabolism including PPARG coactivator PPARGC1A, carboxylesterases CES1, 2, and 3, ceramide synthase CERS2, choline phosphotransferase CHPT1, phospholipid phosphatase PLPPR4, dehydrocholesterol reductase DHCR24, perilipin 2 (PLIN2), apolipoprotein APOL4 and some others. This is an important finding as lipid metabolism which plays an essential role in skin barrier function, is known to be significantly altered in skin of patients with AD and some other inflammatory skin diseases (Berdyshev et al., 2018, Ewald et al., 2015).

Further, we revealed 10 common DEGs between AA skin and AA HSE (Fig. 2d) including COL1A1, LCE1D, LCE1E; FAP (membrane gelatinase involved in EMT); MRAP2 (modulator of melanocortin receptor signaling), and several pseudogenes. Moreover, GSEA pathway enrichment analysis with Hallmark and Reactome revealed remarkably similar enrichment in AA HSE and in skin: by processes related to epithelial-mesenchymal transition (EMT), inflammation and immune system-related Interferon-alpha and -gamma responses; Allograft rejection signaling along with TNF-α signaling via NF-kB and IL2-STA5. The down-regulated DEGs in AA HSE were enriched for formation of the cornified envelope/keratinization and crosslinking of collagen fibrils (Fig. 2c).

3. Analysis of enriched transcription factor binding sites in DEG promoters

Next, we searched for transcription factors (TFs) potentially involved in DEG regulation in skin and HSEs. We specifically focused on TF binding sites (TFBS) in the composite modules, the genomic areas in promoters where multiple TFs could bind in cooperative manner to regulate DEG expression (Waleev et al., 2006) (see Supplement for details). The clusters of enriched TFBS in DEG promoters (−1000bp upstream and +100bp downstream of transcription start site) were identified by F-Match (Kel et al., 2006) and CMA (Composite Module Analyst) algorithms (Waleev et al., 2006) using the TRANSFAC® database (Matys et al., 2006), release 2019.2. Among top TFs potentially involved in the regulation of overexpressed DEGs in AA skin (Supplemental Table 7A) were RELA, major subunit of pro-inflammatory NF-kB factor; POU2F1/OCT1 that regulates immunoglobulins; KLF4 required for skin barrier development (https://www.genecards.org/) and androgen receptor.

Among top TFs potentially involved in the regulation of overexpressed DEGs in AA HSE were RELA, steroid hormone receptors including progesterone and glucocorticoid receptors, and CEBPB involved in immune/inflammatory responses, metabolism, keratinocyte survival (Supplemental Table 7B).

4. Similarity between DEGs in AA healthy skin/HSE and molecular signatures of AD and psoriasis

The pro-inflammatory signaling and alteration of skin barrier functions are typical for many inflammatory skin diseases (Brunner et al., 2017, de Guzman Strong et al., 2010, Schwingen et al., 2020, Tsoi et al., 2019). Thus, we compared AA DEGs with DEGs typical for AD and psoriasis skin that changed in the same direction (Tsoi et al., 2019). We revealed a significant overlap between genes up- or down-regulated in healthy AA skin/AA keratinocytes and molecular signatures of AD and psoriasis (Fig. 3a,b, Supplemental Table S4). Most of these shared DEGs were common for both inflammatory skin diseases (Supplemental Fig. S3a,b). Accordingly, GSEA for AA skin DEGs overlapping with molecular signatures of AD or psoriasis revealed common pro-inflammatory pathways mediated by cytokines, chemokines, interferons (Fig. 3c). Importantly, in addition, AA skin DEGs overlapping with AD molecular signature in skin were enriched by IL13 and IL4 signaling central for AD (Fig. 3c). Further analysis (see Supplemental Materials and Methods) revealed that DEGs down-regulated in AA skin, show a modest but consistent shift towards lower expression in lesional skin (p=4.3×10−10 in psoriasis and p=3.2×10−5 in AD, using the non-parametric Wilcoxon rank-sum test, Supplemental Fig. S4). Genes that were upregulated in AA HSEs had higher expression in AD lesional skin (p=4.45×10−5).

Figure 3. Comparison of AA-specific DEGs in healthy skin and HSEs with molecular signatures of lesional skin in AD and psoriasis patients.

Figure 3.

Open in a new tab

(a). Venn diagrams showing overlap between molecular signature of lesional skin in psoriasis patients (Tsoi et al., 2019) and either AA HSE DEGs or AA skin DEGs. (b). Venn diagrams showing overlap between molecular signature of lesional skin in AD patients (Tsoi et al., 2019) and either AA HSE DEGs or AA skin DEGs. We compared only DEGs whose expression has changed in the same direction in AA skin/AA keratinocytes and AD or psoriasis patient skin. (c). GSEA Reactome pathway enrichment analysis of AA skin DEGs shared with AD molecular signature; and AA skin DEGs shared with psoriasis molecular signature. Note: the enrichment of AA skin DEGs overlapping with AD DEGs by IL4 and IL13 signaling central for atopic dermatitis.

5. AA HSE cultures are more responsive to TNFα effects

Next, we assessed whether activated intrinsic pro-inflammatory signaling makes AA HSE more sensitive to pro-inflammatory cytokine TNFα.

Acute TNFα effects on transcriptome in AA and WNH HSE

TNFα treatment for 24 h resulted in a broad genomic response, largely consistent with previous results (Banno et al., 2004, Chiricozzi et al., 2011). However, the dynamics and amplitude of pro-inflammatory signaling were rather different in two types of HSE. At 24 h TNFα induced 260 DEGs (115 up- and 145 down-regulated) in AA and 122 DEGs (77 up- and 45 down-regulated) in WNH HSEs (P≤0.01, FC ≥2) indicating an overall two-fold stronger AA HSE response (Fig. 4a).

Figure 4. Increased response of AA 3D human skin equivalents (HSE) to short-term TNFα treatment.

Figure 4.

Open in a new tab

AA and WNH HSEs maintained at air-liquid interface for 11 days for epidermal maturation, were treated with vehicle (0.1% BSA/phosphate-buffered saline, Control) or 10 ng/ml TNFα for 24 h. (a). Venn diagram showing common DEGs (P<0.01, FC >2) upregulated or down-regulated in AA and WNH HSEs after TNFα. (b). Results of Reactome GSEA pathway enrichment analysis of DEGs in AA HSEs and WNH HSEs. (c). qRT-PCR validation of selected target genes differentially regulated by TNFα in AA (red) compared to WNH (blue) HSEs. Y axis is expression fold change in TNFα-treated HSEs versus endogenous control. Statistical analysis was performed using unpaired two-tailed t test. (d). Analysis of Filaggrin expression in HSEs by immunofluorescence. Scale bar is 10 μM. Note: strong down-regulation of FLG in WNH HSEs.

Among DEGs upregulated in AA HSEs were numerous chemokines (CXCL3, CXCL5, CXCL11, CCL5, CCL20), interleukins (IL8/CXCL8, IL32, IL36G), IL receptors (IL1R2, IL13RA2, IL2RG). The number of inflammation-related DEGs was much lower in WNH TNFα-treated cultures (Fig. 4, Supplemental Table S5). The down-regulated DEGs were presented by functionally diverse genes with ~ 20 TFs, including the clock genes (PER2, PER3, CRY2) important for epidermal barrier and immune responses (Nakao, 2020, Vaughn et al., 2018). Interestingly, we also observed that TNFα significantly inhibited FLG mRNA expression only in WNH HSE cultures which was further confirmed by immunofluorescence analysis (Fig. 4c,d).

GSEA (Reactome) analysis revealed that seven of 9 gene sets in AA HSE related to inflammation. Specifically, AA HSE cultures developed multiple inflammatory pathways including innate immune system, cytokine and chemokine signaling, neutrophil degranulation (Fig. 4b). One of the pathways was IL13/IL4 signaling linked to the increased expression of IL13RA2 (Fig. 4b, Suppl. Table S5A). In contrast, in WNH HSEs only three of 9 gene sets were inflammation-related, with a much lower number of DEGs in each of them. Instead, DEGs in TNFα-treated WNH cultures were linked to ECM, collagen fibers organization/degradation, MMP activation gene categories (Fig. 4b).

Effects of Chronic TNFα treatment on transcriptome in AA and WNH HSEs

Continuous HSE exposure to TNFα for 168 h resulted in a dramatic increase in the DEG numbers (P≤0.01, FC ≥2): to 1172 in AA and 1060 in WNH HSEs, with ~ 50% of DEGs shared between two types of cultures (Fig. 5a, Supplemental Table S6).

Figure 5. Differential response of AA and WNH human skin equivalents (HSE) to chronic TNFα treatment.

Figure 5.

Open in a new tab

AA and WNH HSEs maintained at air-liquid interface for 11 days for epidermal maturation, were treated with vehicle (0.1% BSA/phosphate-buffered saline, Control) or 10 ng/ml TNFα for 168 h.

(a). Venn diagram showing common DEGs (FC >2, P<0.01) upregulated or down-regulated in AA and WNH HSEs after TNFα-treatment. (b). Results of GSEA Reactome pathway enrichment analysis of DEGs in AA and WNH HSEs. (c). qRT-PCR validation of target genes regulated by TNFα in individual AA (red) and WNH (blue) HSEs. RPL27 was used as a cDNA normalization control. Y axis is fold change of expression in TNFα-treated versus endogenous control HSEs. Statistical analysis was performed using unpaired two-tailed t test.

The number of cytokines, chemokines, interleukins, TNFα Receptor associated/related genes and growth factors in AA HSEs strongly increased with time from 12 (at 24 h) to more than 40 (at 168 h), and many initially activated pro-inflammatory genes remained overexpressed. The pro-inflammatory TNFα molecular signature similarly increased with time in WNH HSEs. In addition, upregulation of S100 calcium-binding proteins including S100A4 and S100A7A/Koebnerisin was observed only in AA HSEs (Fig. 5c).

It is known that TNFα positively regulates its own expression via feed-forward loop mediated by NF-κB and other mechanisms (Furue et al., 2019). We found a higher level of endogenous TNFα protein following modest induction of TNFα mRNA, a trend observed mostly in AA HSEs (Fig. 6a,b).

Figure 6. Induction of endogenous TNFα protein in 3D human skin equivalents (HSE) by TNFα treatment.

Figure 6.

Open in a new tab

(a). TNFα expression analysis by qRT-PCR in individual AA (red) and WNH (blue) HSEs treated with 10ng/ml of TNFα or vehicle control (C) for 24h or 168h. RPL27 was used as a cDNA normalization control. (b). Western blot analysis of TNFα expression in individual AA and WNH HSEs treated with TNFα (10 ng/ml) or vehicle control (C) for 24 h or 168 h. Tubulin was used as loading control.

The prolonged TNFα treatment also resulted in drastic down-regulation of dozens of structural genes, keratins, LCE genes, and integrins in both HSE types. The LCEs gene regulation by TNFα is a new observation from this deep RNA-seq approach. These changes coincided with further down-regulation of FLG, FLG2, and LOR critical for skin barrier, which was more apparent in WNH cultures (Fig. 5c, Supplemental Table S6) and consistent with the less pronounced FLG down-regulation in lesional skin in AA patients with AD (Sanyal et al., 2019). In contrast, the down-regulation of multiple collagen genes (COL1A1, COL5A2, COL9A3, COL18A1; COL21A1, COl27A1) was more pronounced in AA HSEs (Fig. 5c, Supplemental Table S6). We also observed down-regulation of metabolic genes encoding cytochrome P450, fatty acid-binding, phospholipase, transglutaminase, and serine protease inhibitor (SERPIN) families in HSE of both types.

In addition, GSEA analysis revealed large gene sets related to development and lipid/fatty acid metabolism (Fig. 5b). The unique for AA DEGs Reactome gene set was posttranslational protein modifications including O-linked glycosylation (Fig. 5b). Even though the major inflammation-related gene sets were the same in AA and WNH HSE (innate immune system, cytokine signaling in immune system, neutrophil degranulation), the number of inflammation-related genes were still lower in WNH HSE (Fig. 5b).

DISCUSSION

To our knowledge the differential gene expression in healthy skin of color was not previously reported. The analysis of RNA-seq data revealed 570 DEGs in AA as compared to WNH skin; only four of these DEGs were melanogenesis/skin pigment related. Interestingly, the previous attempt to analyze baseline gene expression in epidermis of different self-identified racial groups (Asian, African and European/Caucasian) also revealed significant differences in gene expression between African and Caucasian or African and Asian skin, even though the number of DEGs was rather limited possibly because the epidermis was isolated using skin suction blisters (Yin et al., 2014).

DEGs in AA skin included a large number of pro-inflammatory genes, immunoglobulins and their receptors, EDC genes, keratins. Some of these DEGs could be linked to known clinical findings or morphological features of skin/appendages in populations of African ancestry (Quaresma et al., 2015, Westgate et al., 2017). For example, the increase in Ig gene expression in AA skin correlated with previously reported increase in IgG amount in serum and cerebrospinal fluid of healthy AA volunteers and AA patients with multiple sclerosis, periodontitis, and HIV along with overall higher IgG synthesis rate in AA individuals (Albandar et al., 2002, Lucey et al., 1992, Rinker et al., 2007). In addition, observed overexpression of IgE receptors, FCER1G and FCER2 in AA skin may contribute to the increased IgE signaling when combined with the increased IgE serum concentration typical for AA patients with AD (Brunner et al., 2017). Further, the increased expression of late keratinocyte differentiation markers LCEs and SPRRs could be one of the mechanisms underlying increased stratum corneum thickness, stronger corneocytes cohesion, a better barrier function in AA skin (Girardeau-Hubert et al., 2019, Muizzuddin et al., 2010). On the other hand, down-regulation of multiple high molecular weight keratins, keratin-associated proteins and trichohyalin (TCHH) could partially explain the known fragility, decreased tensile strength, frequent hair shaft splits of curly AA hair (Westgate et al., 2017).

We confirmed dimorphic gene expression and the existence of intrinsic pro-inflammatory circuits, activation of epithelial-mesenchymal transition (EMT) and alteration of skin barrier genes expression in AA skin using HSE made from neonatal AA and WNH keratinocytes. The significant differences in global gene and protein expression, enriched by processes related to terminal differentiation and lipid metabolism, were also observed in 3D skin models made out of African and White keratinocytes isolated from adult female skin (all volunteers were from France, (Girardeau-Hubert et al., 2019).

This baseline increase in pro-inflammatory signaling made AA keratinocytes very sensitive to inflammatory stimulus, TNFα, involved in the development of different inflammatory skin diseases (Banno et al., 2004, Chiricozzi et al., 2011). The finding that genes differentially expressed in healthy AA skin and AA keratinocytes have overlap with both AD and psoriasis skin molecular signatures (Tsoi et al., 2019), also pointed towards an intrinsic inflammatory bias and perturbation of skin barrier in healthy AA population. Moreover, the unique enrichment of AA skin DEGs overlapping with AD by IL13/IL4 signaling (Fig. 3c), suggests that endogenous pro-inflammatory signaling in healthy AA skin is moderately skewed towards Th2 immune pathways typical for AD.

We acknowledge that our analysis has some limitations due to the relatively small size of volunteer cohort. In addition, we still have to extend these studies towards the analysis of protein expression. Nevertheless, we were able to reveal several hundred genes differentially expressed in AA skin that were enriched by pro-inflammatory signaling. This major observation was further confirmed by experiments with 3D HSE cultures. Importantly, when our manuscript was under review, another group of authors using meta-analysis of RNA-seq data from 25 tissue-types uncovered the widespread enrichment of pathways linked to inflammation in AA healthy tissues (Singh et al., 2021)

Overall, the existence of intrinsic pro-inflammatory circuits in healthy AA skin/keratinocytes, could account for the increased sensitivity of AA population to the development of certain inflammatory skin diseases/conditions such as AD, hidradenitis suppurativa, acne. However, the specific molecular mechanisms that trigger shift of this “ambivalent” pre-disease signaling towards specific inflammatory skin diseases remain to be investigated. When found, these molecular drivers could serve as novel targets for the development of targeted skin disease prevention and treatment approaches for AA population.

Materials and Methods

Human skin biopsies

Four mm full thickness skin biopsies were provided by Northwestern Skin Biology and Diseases Resource-Based Center (SBDRC) as described (Lili et al., 2019). Samples were from the upper inner arm (relatively sun-protected skin) of 17 healthy volunteers without history of skin diseases. Subjects included 9 volunteers with darkly pigmented skin (Fitzpatrick score V-VI, self-identified as AA), 8 volunteers with lightly pigmented skin (Fitzpatrick score I-II, self-identified as WNH) of both sexes (specifically 5 females and 3 males in WNH group; 5 females and 4 males in AA group) ages 26–58 years (see Supplemental Table S1 for details). In all cases self-report and Fitzpatrick score were concordant. Informed consent was obtained in writing, following the principles of the WMA Declaration of Helsinki and the NIH Belmont Report. All studies were approved by the Northwestern University IRB.

Even though it is now well understood that there is a considerable diversity within the census defined Black and White populations in US, and that self-identification of volunteers as “Black/African American” or “White/Caucasian” could be inaccurate, we standardized our terminology to be more inclusive and consistent with the previous literature reports that discussed similar issues.

3D Human Skin Equivalents (HSE)

HSE provided by the SBDRC STEM Core were prepared with NHEKs isolated from AA and WNH neonatal foreskins (grouping was based on mothers self-reported race/ethnicity) and cultured as described earlier (Simpson et al., 2010).

HSE were treated with 10 ng/ml TNFα (R&D Systems, Minneapolis, MN) or 0.1% BSA (Sigma Aldrich, St Louis, MO)/PBS for 24 h or 168 h. Separated from collagen scaffold epidermis was used for RNA and protein extraction.

RNA extraction

Total RNA from whole human skin and HSE epidermis was isolated as described elsewhere (Lili et al., 2019).

RNA-seq data analysis for differentially expressed genes (DEGs) in skin and HSE

The human skin RNA-seq was performed on the Illumina NextSeq500 platform (single-end reads, 75 bp) (Lili et al., 2019). The results were previously published with a focus on the effects of topical glucocorticoids. Here we analyzed the gene expression in non-treated AA and WNH skin using RNA-seq data for control samples. To identify DEGs, a linear model (adjusted for sex, age and batch) was implemented (limma R package (Ritchie et al., 2015)). DEGs were considered at p-value ≤ 0.05 and FC >1.5.

The HSE RNA-seq was performed by LC Sciences (Houston, TX) on the Illumina Novaseq™ 6000 platform (paired-end reads, 150 bp). Per-base quality was assessed with FastQC software (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Alignment of fastq files to the human genome (ftp://ftp.ensembl.org/pub/release-96/fasta/homo_sapiens/dna/) was done with HISAT2 (Kim et al., 2015). The assembly of mapped reads of each sample and estimation of the expression levels was performed using StringTie (Pertea et al., 2015). DEGs were considered at p-value ≤ 0.05 and FC >1.5 for control HSE, and at p-value ≤ 0.001 and FC >2 for TNFα experiments.

GEO numbers for the RNA seq datasets: for human skin: GSE120783; for HSE/human skin equivalents: GSE163711.

Validation of RNA-seq data

Validation of RNA-seq data was done by qRT-PCR analysis as described (Lili et al., 2019). The comparisons between control AA and WNH samples were assessed by t-test. TNFα effect was compared to endogenous controls using HSE made from the same primary NHEK cultures.

Functional pathway enrichment for DEGs

Pathway enrichment analysis was done using the online GSEA tool (Broad Institute) utilizing the Reactome and the Hallmark database gene sets (http://software.broadinstitute.org/gsea/msigdb/index.jsp).

Analysis of TF binding sites in DEG enhancers.

TF binding sites (TFBS) in promoters of DEGs (−1000 to +100 bp of start codon) were analyzed using known DNA-binding motifs (Kel et al., 2003, Matys et al., 2006) as described earlier (Orekhov et al., 2020). See Supplement for details.

Western blot analysis

Whole-cell protein extracts were isolated from HSE epidermis and used for Western blot analysis with TNFα primary antibodies (R&D, Minneapolis, MS, USA) as previously described (Baida et al., 2015).

Filaggrin expression analysis by immunofluorescence

Frozen sections (4 μm) of 3D HSE were fixed in ice cold 100% methanol and stained with anti-filaggrin antibody (dilution 1:500, Poly19058, Biolegend Inc., San Diego, CA) using standard procedures. Nuclei were counterstained with DAPI.

Supplementary Material

1

ACKNOWLEDGEMENTS

We acknowledge Northwestern University Genomics Facility/Sequencing Core, and Northwestern University SBDRC (5 P3 0AR075049) Skin Tissue Engineering and Morphology (STEM) Core for technical support. We also acknowledge RNA-seq analysis in HSE by LC Sciences (Houston, TX).

Abbreviations:

AA

African American

AD

atopic dermatitis

DEG

differentially expressed gene

EMT

epithelial-mesenchymal transition

FLG

filaggrin

FC

fold change

GSEA

gene set enrichment analysis

HSE

3D human skin equivalent

Ig

immunoglobulin

LCE

late cornified envelope

LOR

loricrin

NHEK

normal human epidermal keratinocytes

qRT-PCR

quantitative real-time reverse-transcription polymerase chain reaction

RNA-seq

RNA sequencing

TF

transcription factor

TNFα

tumor necrosis factor alpha

WNH

White Non-Hispanic

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

CONFLICT OF INTEREST

Dr. Kel is an employee of the geneXplain GmbH Company. Other authors state no conflict of interest.

DATA AVAILABILITY

All raw and pre-processed RNA-seq data for human skin used in this study are deposited in the GEO database with GEO accession number: GSE120783, and for HSE RNA-seq with GEO accession number: GSE163711. All other data generated or analyzed during this study are included in this article and in the Supplemental Files.

DECLARATIONS

Ethics approval and consent to participate

Informed consent of all human volunteers in this study was obtained in writing, following the principles established in the WMA Declaration of Helsinki and the NIH Belmont Report (http://www.hhs.gov/ohrp/humansubjects/guidance/belmont.html). Informed consent for neonatal foreskin tissue was not required as this tissue is de-identified and considered discarded material by Northwestern IRB policy.

All studies were approved by the Northwestern University Institutional Review Board (IRB protocol # STU00009443).

REFERENCES

  1. Agbai ON, Buster K, Sanchez M, Hernandez C, Kundu RV, Chiu M, et al. Skin cancer and photoprotection in people of color: a review and recommendations for physicians and the public. J Am Acad Dermatol 2014;70(4):748–62. [DOI] [PubMed] [Google Scholar]
  2. Albandar JM, DeNardin AM, Adesanya MR, Winn DM, Diehl SR. Associations of serum concentrations of IgG, IgA, IgM and interleukin-1beta with early-onset periodontitis classification and race. J Clin Periodontol 2002;29(5):421–6. [DOI] [PubMed] [Google Scholar]
  3. Alexis AF, Blackcloud P. Psoriasis in skin of color: epidemiology, genetics, clinical presentation, and treatment nuances. J Clin Aesthet Dermatol 2014;7(11):16–24. [PMC free article] [PubMed] [Google Scholar]
  4. Alexis AF, Rendon M, Silverberg JI, Pariser DM, Lockshin B, Griffiths CE, et al. Efficacy of Dupilumab in Different Racial Subgroups of Adults With Moderate-to-Severe Atopic Dermatitis in Three Randomized, Placebo-Controlled Phase 3 Trials. J Drugs Dermatol 2019;18(8):804–13. [PubMed] [Google Scholar]
  5. Baida G, Bhalla P, Kirsanov K, Lesovaya E, Yakubovskaya M, Yuen K, et al. REDD1 functions at the crossroads between the therapeutic and adverse effects of topical glucocorticoids. EMBO Mol Med 2015;7(1):42–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Banno T, Gazel A, Blumenberg M. Effects of tumor necrosis factor-alpha (TNF alpha) in epidermal keratinocytes revealed using global transcriptional profiling. J Biol Chem 2004;279(31):32633–42. [DOI] [PubMed] [Google Scholar]
  7. Berdyshev E, Goleva E, Bronova I, Dyjack N, Rios C, Jung J, et al. Lipid abnormalities in atopic skin are driven by type 2 cytokines. JCI Insight 2018;3(4). [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brunner PM, Guttman-Yassky E. Racial differences in atopic dermatitis. Ann Allergy Asthma Immunol 2019;122(5):449–55. [DOI] [PubMed] [Google Scholar]
  9. Brunner PM, Guttman-Yassky E, Leung DY. The immunology of atopic dermatitis and its reversibility with broad-spectrum and targeted therapies. J Allergy Clin Immunol 2017;139(4S):S65–S76. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cheret J, Bertolini M, Ponce L, Lehmann J, Tsai T, Alam M, et al. Olfactory receptor OR2AT4 regulates human hair growth. Nat Commun 2018;9(1):3624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chiricozzi A, Guttman-Yassky E, Suarez-Farinas M, Nograles KE, Tian S, Cardinale I, et al. Integrative responses to IL-17 and TNF-alpha in human keratinocytes account for key inflammatory pathogenic circuits in psoriasis. J Invest Dermatol 2011;131(3):677–87. [DOI] [PubMed] [Google Scholar]
  12. de Guzman Strong C, Conlan S, Deming CB, Cheng J, Sears KE, Segre JA. A milieu of regulatory elements in the epidermal differentiation complex syntenic block: implications for atopic dermatitis and psoriasis. Hum Mol Genet 2010;19(8):1453–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Ewald DA, Malajian D, Krueger JG, Workman CT, Wang T, Tian S, et al. Meta-analysis derived atopic dermatitis (MADAD) transcriptome defines a robust AD signature highlighting the involvement of atherosclerosis and lipid metabolism pathways. BMC Med Genomics 2015;8:60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Feldman CH, Hiraki LT, Liu J, Fischer MA, Solomon DH, Alarcon GS, et al. Epidemiology and sociodemographics of systemic lupus erythematosus and lupus nephritis among US adults with Medicaid coverage, 2000–2004. Arthritis Rheum 2013;65(3):753–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Furue K, Ito T, Tsuji G, Kadono T, Furue M. Psoriasis and the TNF/IL23/IL17 axis. G Ital Dermatol Venereol 2019;154(4):418–24. [DOI] [PubMed] [Google Scholar]
  16. Girardeau-Hubert S, Deneuville C, Pageon H, Abed K, Tacheau C, Cavusoglu N, et al. Reconstructed Skin Models Revealed Unexpected Differences in Epidermal African and Caucasian Skin. Sci Rep 2019;9(1):7456. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Halder RM, Nootheti PK. Ethnic skin disorders overview. J Am Acad Dermatol 2003;48(6 Suppl):S143–8. [DOI] [PubMed] [Google Scholar]
  18. Harland DP, Plowman JE. Development of Hair Fibres. Adv Exp Med Biol 2018;1054:109–54. [DOI] [PubMed] [Google Scholar]
  19. Hombach S, Kretz M. The non-coding skin: exploring the roles of long non-coding RNAs in epidermal homeostasis and disease. Bioessays 2013;35(12):1093–100. [DOI] [PubMed] [Google Scholar]
  20. Hu G, Niu F, Humburg BA, Liao K, Bendi S, Callen S, et al. Molecular mechanisms of long noncoding RNAs and their role in disease pathogenesis. Oncotarget 2018;9(26):18648–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kaufman BP, Guttman-Yassky E, Alexis AF. Atopic dermatitis in diverse racial and ethnic groups-Variations in epidemiology, genetics, clinical presentation and treatment. Exp Dermatol 2018;27(4):340–57. [DOI] [PubMed] [Google Scholar]
  22. Kel A, Voss N, Jauregui R, Kel-Margoulis O, Wingender E. Beyond microarrays: find key transcription factors controlling signal transduction pathways. BMC Bioinformatics 2006;7 Suppl 2:S13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kel AE, Gossling E, Reuter I, Cheremushkin E, Kel-Margoulis OV, Wingender E. MATCH: A tool for searching transcription factor binding sites in DNA sequences. Nucleic Acids Res 2003;31(13):3576–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kim D, Langmead B, Salzberg SL. HISAT: a fast spliced aligner with low memory requirements. Nat Methods 2015;12(4):357–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lili LN, Klopot A, Readhead B, Baida G, Dudley JT, Budunova I. Transcriptomic Network Interactions in Human Skin Treated with Topical Glucocorticoid Clobetasol Propionate. J Invest Dermatol 2019;139(11):2281–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lucey DR, Hendrix CW, Andrzejewski C, Melcher GP, Butzin CA, Henry R, et al. Comparison by race of total serum IgG, IgA, and IgM with CD4+ T-cell counts in North American persons infected with the human immunodeficiency virus type 1. J Acquir Immune Defic Syndr (1988) 1992;5(4):325–32. [PubMed] [Google Scholar]
  27. Matys V, Kel-Margoulis OV, Fricke E, Liebich I, Land S, Barre-Dirrie A, et al. TRANSFAC and its module TRANSCompel: transcriptional gene regulation in eukaryotes. Nucleic Acids Res 2006;34(Database issue):D108–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Mei-Yen Yong A, Tay YK. Atopic Dermatitis: Racial and Ethnic Differences. Dermatol Clin 2017;35(3):395–402. [DOI] [PubMed] [Google Scholar]
  29. Muizzuddin N, Hellemans L, Van Overloop L, Corstjens H, Declercq L, Maes D. Structural and functional differences in barrier properties of African American, Caucasian and East Asian skin. J Dermatol Sci 2010;59(2):123–8. [DOI] [PubMed] [Google Scholar]
  30. Nakao A. Circadian Regulation of the Biology of Allergic Disease: Clock Disruption Can Promote Allergy. Front Immunol 2020;11:1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Orekhov AN, Nikiforov NG, Sukhorukov VN, Kubekina MV, Sobenin IA, Wu WK, et al. Role of Phagocytosis in the Pro-Inflammatory Response in LDL-Induced Foam Cell Formation; a Transcriptome Analysis. Int J Mol Sci 2020;21(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Pertea M, Pertea GM, Antonescu CM, Chang TC, Mendell JT, Salzberg SL. StringTie enables improved reconstruction of a transcriptome from RNA-seq reads. Nat Biotechnol 2015;33(3):290–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Price KN, Hsiao JL, Shi VY. Race and Ethnicity Gaps in Global Hidradenitis Suppurativa Clinical Trials. Dermatology 2019:1–6. [DOI] [PubMed] [Google Scholar]
  34. Quaresma MV, Martinez Velasco MA, Tosti A. Hair Breakage in Patients of African Descent: Role of Dermoscopy. Skin Appendage Disord 2015;1(2):99–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rinker JR 2nd, Trinkaus K, Naismith RT, Cross AH. Higher IgG index found in African Americans versus Caucasians with multiple sclerosis. Neurology 2007;69(1):68–72. [DOI] [PubMed] [Google Scholar]
  36. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015;43(7):e47. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sanyal RD, Pavel AB, Glickman J, Chan TC, Zheng X, Zhang N, et al. Atopic dermatitis in African American patients is TH2/TH22-skewed with TH1/TH17 attenuation. Ann Allergy Asthma Immunol 2019;122(1):99–110 e6. [DOI] [PubMed] [Google Scholar]
  38. Schwingen J, Kaplan M, Kurschus FC. Review-Current Concepts in Inflammatory Skin Diseases Evolved by Transcriptome Analysis: In-Depth Analysis of Atopic Dermatitis and Psoriasis. Int J Mol Sci 2020;21(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Simpson CL, Kojima S, Getsios S. RNA interference in keratinocytes and an organotypic model of human epidermis. Methods Mol Biol 2010;585:127–46. [DOI] [PubMed] [Google Scholar]
  40. Singh U, Hernandez KM, Aronow BJ, Wurtele ES. African Americans and European Americans exhibit distinct gene expression patterns across tissues and tumors associated with immunologic functions and environmental exposures. Sci Rep 2021;11(1):9905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Tsoi LC, Rodriguez E, Degenhardt F, Baurecht H, Wehkamp U, Volks N, et al. Atopic Dermatitis Is an IL-13-Dominant Disease with Greater Molecular Heterogeneity Compared to Psoriasis. J Invest Dermatol 2019;139(7):1480–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Vaughn AR, Clark AK, Sivamani RK, Shi VY. Circadian rhythm in atopic dermatitis-Pathophysiology and implications for chronotherapy. Pediatr Dermatol 2018;35(1):152–7. [DOI] [PubMed] [Google Scholar]
  43. Waleev T, Shtokalo D, Konovalova T, Voss N, Cheremushkin E, Stegmaier P, et al. Composite Module Analyst: identification of transcription factor binding site combinations using genetic algorithm. Nucleic Acids Res 2006;34(Web Server issue):W541–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Westgate GE, Ginger RS, Green MR. The biology and genetics of curly hair. Exp Dermatol 2017;26(6):483–90. [DOI] [PubMed] [Google Scholar]
  45. Yan D, Afifi L, Jeon C, Cordoro KM, Liao W. A cross-sectional study of the distribution of psoriasis subtypes in different ethno-racial groups. Dermatol Online J 2018;24(7). [PubMed] [Google Scholar]
  46. Yin L, Coelho SG, Ebsen D, Smuda C, Mahns A, Miller SA, et al. Epidermal gene expression and ethnic pigmentation variations among individuals of Asian, European and African ancestry. Exp Dermatol 2014;23(10):731–5. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1766483-supplement-1.pdf (2.6MB, pdf)

RESOURCES