HHS Public Access Author manuscript Author Manuscript
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01. Published in final edited form as: J Invest Dermatol. 2016 September ; 136(9): 1792–1800. doi:10.1016/j.jid.2016.05.113.
Sebaceous gland atrophy in psoriasis: An explanation for psoriatic alopecia? Laure Rittié1, Trilokraj Tejasvi1,3, Paul W. Harms1,2, Xianying Xing1, Rajan P. Nair1, Johann E. Gudjonsson1, William R Swindell1, and James T. Elder1,3,*
Author Manuscript
1Department
of Dermatology, University of Michigan, Ann Arbor, MI 48109
2Department
of Pathology, University of Michigan, Ann Arbor, MI 48109
3Ann
Arbor Veterans Affairs Hospital, Ann Arbor, MI
Abstract
Author Manuscript
In a transcriptome study of psoriatic (PP) vs. normal (NN) skin, we found a co-expressed gene module (N5) enriched 11.5-fold for lipid biosynthetic genes. We also observed fewer visible hairs in PP skin, compared to uninvolved (PN) or NN skin (p 4, 54.1-fold). The intersection of PP-downregulated and SH-upregulated gene lists generated a gene expression signature consisting solely of module N5 genes, whose expression in PP vs. NN skin was inversely correlated with the signature of IL17-stimuated keratinocytes. Despite loss of visible hairs, morphometry identified elongated follicles in PP vs. PN skin (average 1.7 vs. 1.2 μm, p=0.020). These results document SG atrophy in non-scalp psoriasis, identify a cytokine-regulated set of SG signature genes, and suggest that loss of visible hair in PP skin may result from abnormal SG function.
Keywords Sebaceous glands; eccrine sweat glands; hair follicles; alopecia; psoriasis
Author Manuscript
Introduction Psoriasis is a common inflammatory and hyperplastic skin disease affecting approximately 1–2% of European-origin individuals (Parisi et al., 2013). Immunological and genetic studies are converging upon an immunopathogenesis involving altered antigen presentation
*
Corresponding author. 7412 Medical Sciences Building 1, University of Michigan, 1301 E. Catherine, Ann Arbor, Michigan 48109-5675, USA, phone (734) 647-8070, [email protected]. 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 citable 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.
Rittié et al.
Page 2
Author Manuscript
and dysregulated signal transduction involving NF-κB activation, interferon signaling, and the IL-23-Th17 axis (Harden et al., 2015). Based on findings from our recent RNA-seq transcriptome study (Li et al., 2014), we became interested in exploring the relationship between psoriasis, hair growth, and sebaceous gland health, a subject that has received only limited attention in the literature.
Author Manuscript
Our transcriptome study used weighted gene co-expression network analysis (Langfelder and Horvath, 2008) to identify coordinately expressed gene modules in lesional psoriatic (PP) skin and normal (NN) skin. We identified one module of coordinately-expressed NN transcripts (module N5) that was highly enriched for genes involved in the metabolism of lipids and lipoproteins (p = 3 x 10−40). Another module (N17) was enriched (p = 1.5 x 10−11) for genes in the peroxisome proliferator activator receptor-γ (PPAR-γ) signaling pathway, which plays a central role in adipocyte differentiation (Alestas et al., 2006, Dozsa et al., 2014, Tontonoz and Spiegelman, 2008). Both modules manifested a high proportion of significantly down-regulated genes (28.8% and 52.5%, respectively) (Li, Tsoi, 2014). Our study also identified clusters of muscle- and hair follicle (HF)-related genes whose expression was similarly down-regulated in PP vs. NN skin (Li, Tsoi, 2014). Because the pilosebaceous unit (PSU) consists of the sebaceous gland (SG) and the arrector pili muscle in addition to the HF itself, we wondered whether our observations might be related to alterations in the PSU in psoriasis. Because it is difficult to obtain PSUs in the plane of section from the single vertical sections that are typically available in clinical practice, we (i) carried out 3-dimensional (3D) morphometry on transverse sections of six paired biopsies of PP vs. PN skin, (ii) utilized biopsy site photographs to assess the number of visible hairs in PP vs. PN vs. NN skin, and (iii) compared the expression behaviors of previously-defined coexpression modules in PP vs. NN skin (Li, Tsoi, 2014) to newly-generated gene expression profiles generated from lesions of SH and NN skin.
Author Manuscript
Results
Author Manuscript
Research subjects were individuals affected with psoriasis vulgaris for at least 8 years (age = 44.7±18.7 years [mean±SD], n=6, 50% male, 50% female). We performed 3D reconstructions of horizontal sections encompassing the complete follicular unit for three pairs of PP and PN biopsies. As visualized in Figure 1A (animated in the online version), SGs were markedly atrophic in PP vs. PN skin. Morphometric analysis demonstrated a 91% overall average reduction in sebaceous gland volume (Figure 1B), with a consistentlydecreased SG volume for each HF (Figure 1A) (P = 0.031; n = 38 sebaceous glands from 3 psoriasis patients). The estimated sebaceous gland volume in PP lesions is 3.5% of that in PN lesions (95% confidence interval: 0.5% – 25.2%). In addition to the marked decrease in SG volume, SG count was abnormal (either 0 or >1 SG per HF) more frequently in PP compared to PN skin (50% for PP vs. 17% for PN skin; P = 0.044; n = 67 hair follicles from 3 psoriasis patients). The estimated odds ratio is 5.12 (95% confidence interval: 1.04 – 25.2), indicating that the odds of an abnormal sebaceous gland count is 5.12 times greater for a PP follicle than for a PN follicle (Figure 1C and Table S1). Utilizing PP/PN pairs from three additional psoriasis patients, we confirmed these results by histology of individual sections using Oil Red O (ORO), a dye with high affinity for neutral
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 3
Author Manuscript
triglycerides and lipids (Ramirez-Zacarias et al., 1992), (Figure 2A). As expected given the lower neutral lipid content of stratum corneum compared to SGs (see Discussion), ORO staining was less pronounced for stratum corneum compared to SGs, and did not appear to differ between PP and PN skin (Figure 2B). As illustrated in Figure 3A for a representative patient, visible hairs are strikingly reduced in psoriatic lesions. To quantify this phenomenon, we utilized photographs of biopsy sites from our previous RNA-seq study (Li, Tsoi, 2014). As shown in Figure 3B, we found that significantly fewer patients had visible hairs in the psoriatic lesion itself, compared to immediately perilesional skin, nonlesional (PN) skin located at least 5 cm away from the nearest plaque, and normal (N) skin from nonpsoriatic individuals (p < 0.0001 vs. PP skin for all three comparisons, by two-tailed Fisher’s exact test).
Author Manuscript Author Manuscript
In contrast to the marked reduction of visible hairs, our 3D reconstructions revealed no visible damage to HFs in PP vs. PN skin (Figure 1A, Video S1). Indeed, hair follicle mean length was significantly greater in PP than in PN skin Hair follicle length is significantly elevated in PP lesions as compared to uninvolved PN skin (P = 0.020; n = 38 hair follicles from 3 psoriasis patients). The mean hair follicle height in PP lesions was 45% greater than in PN lesions (95% confidence interval: 8.86% – 81.1%, Figure 4A). There was no significant difference in HF counts in PP vs. PN skin, and eccrine duct counts were also similar (Figure 4B). As expected, the volume of the epidermal compartment was markedly (5.83±0.87-fold) and significantly (p=0.011, two-tailed paired t-test) greater in PP vs. PN skin (Figure 4C). The increased HF length noted above combined with the expansion of the epidermis resulted in similar contributions of the HF compartment in PP vs. PN (0.62±0.39fold in PP vs. PN, p=0.26). Due to this expansion of the epidermal compartment, both SGs and eccrine sweat glands (ESGs) made a smaller contribution to the total epithelial volume on a percentage basis in PP vs. PN skin (p=0.0251 and 0.0454, respectively). However, the appendage contribution to total epithelial volume in PP vs. PN skin was much more markedly reduced in the case of SGs (0.1% vs. 3.7%, respectively, overall average reduction of 98%) compared to ESGs (0.4% vs. 1.3% of the epithelial compartment in PP vs PN, overall average reduction of 67%) (Figure 4D).
Author Manuscript
Noting this marked and disproportionate reduction in SG volume, we reasoned that the most markedly down-regulated genes in the PP vs. NN skin transcriptome might be enriched for genes expressed in SG. To test this hypothesis, we queried our psoriasis RNA-seq transcriptome dataset (Li, Tsoi, 2014) to identify co-expression modules enriched for highly down-regulated genes. As noted earlier, module N5 as a whole was unique with respect to its strong enrichment for lipid biosynthetic process genes (p = 2.77 × 10−12, Figure 5, top panel). By assessing enrichments for each module at various FC thresholds, we found that enrichment for module N5 genes increased progressively among genes most strongly downregulated in PP vs. NN skin (Figure S1). Notably, the set of 445 genes with FC < 0.25 in PP vs. NN skin overlapped significantly with module N5 genes (P = 3.72 × 10−45 by Fisher’s exact test; Figure 5, middle panel). In order to generate a SG gene expression signature, we prepared RNA from formalin-fixed, paraffin-embedded sections of three biopsies of NN skin and four biopsies of SH, a common
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 4
Author Manuscript
skin condition characterized by marked enlargement of multilobular SGs sharing a single central lumen (Eisen and Michael, 2009). In contrast to the progressive enrichment for module N5 genes with increasing down-regulation in PP vs. NN skin, module N5 enrichment progressively increased with increasing overexpression in SH vs. NN skin (Figure S1). As shown in the bottom panel of Figure 5, genes overexpressed by > 4-fold in SH compared to normal skin overlapped significantly with module N5 genes (P = 1.9 × 10−31 by Fisher’s exact test). This same level of enrichment was not observed among any of the other modules we evaluated. To generate an SG gene expression signature, we enumerated 12 genes whose expression was upregulated > 4-fold in SH lesions vs. NN skin, and down-regulated > 4-fold PP vs. NN skin (Table S2). Notable among these were genes involved in the synthesis and/or metabolism of triglycerides, very-long-chain fatty acids, sphingolipids, and wax esters, all major components of sebum (Box 1).
Author Manuscript
Box 1 Functional properties of SG signature genes AADACL3 (1p36.21) encodes arylacetamide deacetylase-like 3, for which no specific functional information is available. Arylacetamide deacetylase (AADAC) participates in the mobilization of triglyceride stores for very low density lipoprotein (VLDL) assembly (Gibbons et al., 2000). Hepatitis C virus (HCV) assembles on lipid droplets in the liver, and expression of AADAC is markedly reduced in the liver early in the course of HCV infection (Nourbakhsh et al., 2013)
Author Manuscript
AWAT1 (Xq13.1) encodes an acyl-CoA wax alcohol acyltransferase, which predominantly esterifies long chain (wax) alcohols with acyl-CoA-derived fatty acids to produce wax esters (Holmes, 2010). The enzyme is expressed in many human tissues but predominates in skin. In situ hybridization demonstrates a differentiation-specific expression pattern within the human sebaceous gland for AWAT1 (Turkish et al., 2005). DGAT2L6 (Xq13.1) maps to the X chromosome and encodes a member of the diacylglycerol acyltransferase 2 family that is structurally related to AWAT1 and AWAT2 (Holmes, 2010). The encoded protein is most likely involved in the synthesis of di- or triacylglycerol. FADS2 (11q12.2) encodes a member of the fatty acid desaturase gene family. Desaturase enzymes regulate unsaturation of fatty acids by introducing double bonds between defined carbons of the fatty acyl chain. FADS2 and several other genes involved in sebocyte differentiation are induced by dihydrotesterone in human sebocytes expressing the androgen receptor (Barrault et al., 2015)
Author Manuscript
FAR2 (12p11.22) encodes a member of the short chain dehydrogenase / reductase superfamily, which is a peroxisomal protein with reductase activity that converts fatty acids into fatty alcohols in the first step of wax biosynthesis (Cheng and Russell, 2004). GLDC (9q22) encodes glycine decarboxylase, which is critical for tumor-initiating cells in non-small cell lung cancer (NSCLC) by inducing dramatic changes in glycolysis and glycine/serine metabolism, leading to changes in pyrimidine metabolism to regulate cancer cell proliferation (Zhang et al., 2012).
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 5
Author Manuscript
MOGAT1 (2q36.2) encodes acyl-CoA:monoacylglycerol acyltransferase, which catalyzes the synthesis of diacylglycerols, the precursor of physiologically important lipids such as triacylglycerol and phospholipids. MOGAT1 is structurally related to DGA2L6, AWAT1, and AWAT2 (Holmes, 2010). PDZK1 (1q21) encodes a PDZ domain-containing scaffolding protein that plays an important role in cholesterol metabolism by regulating the HDL receptor, scavenger receptor class B type 1 (Kocher and Krieger, 2009). PM20D1 (1q32.1) encodes a protein of unknown function, which contains a peptidase domain.
Author Manuscript
SOAT1 (1q25) encodes an acyltransferase protein located in the endoplasmic reticulum, which catalyzes the formation of fatty acid-cholesterol esters. SOAT1 has been implicated in the formation of atherosclerotic plaques by controlling the equilibrium between free cholesterol and cytoplasmic cholesteryl esters (Wu et al., 2010). SOAT1 is strongly expressed in SGs of WT mice but not in SGs of mice homozygous for the hair interior defect (hid) allele, which involves a 118 bp deletion in SOAT1 (Wu, Potter, 2010) THRSP (11q14.1) is the human homologue of the rat S14 gene. Expression of S14 is controlled by nutritional and hormonal factors and is restricted to liver and adipose tissue, particularly in lipomatous nodules. Suggestive of a role in lipid metabolism, it is also expressed in lipogenic breast cancers (Moncur et al., 1998) and upregulates lipogenesis in bovine mammary epithelial cells (Cui et al., 2015)
Author Manuscript
TMPRSS11E (4q13.2) encodes a transmembrane serine protease also known as DESC1. Expression of TMPRSS11E is decreased in head and neck squamous cell carcinoma. Its expression is epithelial-specific and limited to tissues derived from the head and neck, in skin, prostate and testes. (Lang and Schuller, 2001).
Author Manuscript
Asebia (Sundberg et al., 2000) and Flake (Georgel et al., 2005) mice lack the enzyme stearoyl-CoA desaturase-1 involved in fatty acid desaturation, due to mutations in the Scd1 gene. Asebia mice manifest marked SG atrophy (Sundberg, Boggess, 2000), and both strains manifest alopecia, skin inflammation, and eye infections. Germ-line (Miyazaki et al., 2001) or skin-targeted (Sampath et al., 2009) deletion of the Scd1 gene results in similar phenotypes. Gene expression profiling of skin-targeted Scd1−/− mice revealed abnormal expression of genes involved in lipid biosynthesis as well as in retinol and steroid metabolism (Flowers et al., 2011). Using this microarray dataset, we observed enrichment for module N5 genes among transcripts downregulated > 4-fold (P = 1.89 × 10−3), in skintargeted Scd1−/− mice relative to WT controls (Figure S2). We next used a previously-defined microarray transcriptome dataset involving 237 paired samples of PP and PN skin (Swindell et al., 2015) to explore the potential roles of various cell types and cytokines in the development of SG atrophy. Applying a criterion of | r | > 0.2, p < 0.0008 to define a meaningful association in the context of multiple testing (Bonferroni correction of 0.05/59 = 0.00085 for 59 tests performed), we found that expression of SG signature genes in PP vs. PN skin was negatively correlated with the gene expression signature of keratinocytes treated with IL-17C (r = −0.23), whereas they were positively J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 6
Author Manuscript
correlated with keratinocyte gene expression responses to IL-4 (r = 0.39) and IL-4 plus IL-13 (r = 0.23) (Table S2, Figure S3). Gene expression signatures from keratinocytes, fibroblasts, and several inflammatory cell types were not inversely correlated with the SG signature, although the macrophage signature was positively correlated (r = 0.32, Table S2).
Discussion
Author Manuscript
There are close anatomic and functional relationships between the HF, the arrector pili muscle, and the SG, which together form the PSU (Fujiwara et al., 2011, Song et al., 2006). Moreover, muscle and HF-related gene co-expression networks are highly correlated in mouse skin (Quigley et al., 2009). Our interest in the status of HFs and SGs in psoriasis was initiated by transcriptome studies in which we identified modules of coordinately-expressed genes enriched in annotation for genes involved in adipogenesis, HFs, and muscle in NN skin, many of which were down-regulated in PP skin (Li, Tsoi, 2014). Based on these observations, we carried out a detailed morphometric analysis of SGs in paired, anatomically-matched samples of PP vs. PN skin. As part of the same study, we assessed the status of ESGs in PP vs. PN skin, a topic that has not received a systematic assessment to date. Sebaceous gland atrophy in psoriasis lesions
Author Manuscript
The most striking morphologic finding from this study was the pronounced atrophy of SGs in PP vs. PN skin, which averaged a 91% overall reduction in volume that was observed in all six biopsy pairs studied (Figures 1 and 2). While earlier studies have shown SG atrophy in scalp psoriasis (Bardazzi et al., 1999, Headington et al., 1989, Silva et al., 2012, Werner et al., 2008, Wilson et al., 1994), here we document SG atrophy in psoriatic lesions at body sites other than the scalp. Similar findings have been reported very recently utilizing sequential vertical sections of PN and PP skin from scalp and back sites (Liakou, 2015), thus confirming our observations. We validated our findings using ORO staining, which specifically stains triglycerides and cholesteryl oleate but no other lipids (Ramirez-Zacarias, Castro-Munozledo, 1992). While triglycerides represent 57% of human sebum lipids by weight (Wilkinson and Karasek, 1966), they appear to be absent from stratum corneum lipids, as assessed by analysis of epidermoid cysts (Wertz et al., 1987). Consistent with this observation, we observed no obvious differences in ORO staining in the stratum corneum of PP vs. PN skin (Figure 2B), in contrast to the marked reduction of ORO staining in SG.
Author Manuscript
As shown in Figure 5, module N5 is uniquely enriched for genes involved in lipid biosynthetic processes. Given (i) the pronounced SG atrophy in psoriatic lesions that we observed morphometrically (Figure 1) and by ORO staining (Figure 2A), we hypothesized that—as opposed to genes involved in lipogenesis in adipose or stratum corneum—lipid biosynthetic genes expressed specifically in SG might be particularly strongly downregulated in PP vs. NN skin due to SG atrophy. Indeed, we found that as the threshold for down-regulation was increased in stringency, the specificity of enrichment for module N5 genes increased (Figure S1).
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 7
Author Manuscript
In support of our hypothesis, we found that module N5 genes were also highly enriched among genes whose expression was the most strongly increased in lesions of SH, relative to normal skin (Figure 5, Figure S1). By intersecting the gene sets decreased > 4-fold in PP vs. NN skin, and increased > 4-fold in SH vs. NN skin, we generated a 12-gene SG signature, with all of the genes mapping to module N5 (Table S2). Literature search revealed a role in lipid biosynthesis and/or localization to SG for essentially all of these genes (Box 1). We found that less stringent choices for the FC cutoff could be used to yield larger SG signature gene sets, with many additional genes providing literature evidence for high expression in SG (data not shown).
Author Manuscript
As a further test of this hypothesis, we queried a microarray-based transcriptome dataset (Flowers, Paton, 2011) derived from the skin of adult mice homozygous, skin targeted deletion of Scd1 (Sampath, Flowers, 2009), again revealing enrichment for module N5 (Figure S2). These mice manifest not only SG atrophy and alopecia, but also substantial inflammatory epidermal hyperplasia. Thus, it is not surprising that genes down-regulated in K5-in skin-targeted Scd1−/− mice are also enriched in other co-expression modules that are also down-regulated in psoriasis (Li, Tsoi, 2014), including modules N14 (hair follicle), N16 (muscular contraction) and N17 (PPAR signaling pathway).
Author Manuscript
Just as epidermal hyperplasia appears to be a response to psoriatic inflammation (Gudjonsson and Elder, 2012), it is not unreasonable to hypothesize that the same might be true for SG atrophy in psoriasis. Consistent with this hypothesis, treatment of isolated SGs with IL-1 or TNF in vitro led to profound reductions in lipid biosynthesis (Guy et al., 1999). However, infiltrating inflammatory cells need not be the primary source of such cytokines. While Headington and colleagues noted that while SG atrophy in scalp psoriasis was associated with increased perifollicular inflammation, this inflammation was not localized to the immediate vicinity of or within SG, but rather to the infundibular area (Headington, Gupta, 1989). We confirmed the paucity of inflammation around SGs and the presence of inflammation surrounding the isthmus and infundibulum in our samples (data not shown). Other potential sources of inflammatory cytokines and other mediators could include keratinocytes, fibroblasts, constitutive tissue dendritic cells, and endothelial cells.
Author Manuscript
Using previously-defined transcriptome datasets across 237 samples of PP and NN skin (Swindell, Sarkar, 2015), we found that the gene expression signature of keratinocytes treated with IL-17C was inversely correlated with the SG signature (Table S2, Figure S3). Notably, IL-17 is not only a key cytokine in psoriasis pathogenesis (Harden, Krueger, 2015), but has also has been found to inhibit the differentiation of adipocyte precursors (Ahmed and Gaffen, 2013). Conversely, keratinocyte responses to the Th2 cytokine IL-4 was positively correlated with the SG signature, whether alone or in combination with IL-13 (Table S2). No gene expression signatures from various inflammatory cell types were inversely correlated with the SG signature, though the macrophage signature was positively correlated (Table S2). Additional mechanistic studies will be required to define the roles of specific cytokines and cell types in the genesis of psoriatic SG atrophy.
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 8
Alopecia in psoriatic lesions
Author Manuscript
Using biopsy site photographs taken in the course of our RNA-seq transcriptome study (Li, Tsoi, 2014), we identified a pronounced loss of visible hairs in psoriatic plaques, relative to immediately perilesional skin, more distant PN skin, or NN skin (Figure 3). While the presence of alopecia in psoriatic lesions has been recognized for over 40 years (Shuster, 1972), there have been few systematic studies of HF structure in psoriatic skin from nonscalp regions. Previous studies of psoriatic alopecia of the scalp found no evidence for significant reductions in total or telogen hair counts in PP vs. PN or normal skin (Bardazzi, Fanti, 1999, Headington, Gupta, 1989, Silva, Brown, 2012, Werner, Brenner, 2008, Wilson, Dean, 1994). In agreement with these observations, we found no significant difference in HF counts per biopsy between PN and PP skin (Figure 4B).
Author Manuscript
A relatively sparse literature, limited to scalp psoriasis, has addressed the question of whether hair loss in psoriatic alopecia is due to scarring (Ramos-e-Silva and Pirmez, 2013, Silva, Brown, 2012). Four reports were case series in which the subjects were pre-selected based on a clinical appearance of scarring alopecia (Bardazzi, Fanti, 1999, Cockayne and Messenger, 2001, van de Kerkhof and Chang, 1992, Wright and Messenger, 1990), whereas larger and unselected series (Runne and Kroneisen-Wiersma, 1992, Shuster, 1972, Werner, Brenner, 2008) revealed a more mixed picture in which non-scarring processes predominated. We found no evidence of misshapen follicles to suggest a scarring process (Figure 1), and it is a common clinical observation that hair re-grows after psoriatic lesions have resolved (Runne and Kroneisen-Wiersma, 1992, Shuster, 1990, Silva, Brown, 2012). Thus, it would appear that while scarring can develop in chronic lesions of psoriasis, scarring is not the primary event leading to loss of visible hairs in psoriatic lesions.
Author Manuscript Author Manuscript
Treatment of isolated HFs with IL-1 or TNF inhibited hair growth (Philpott et al., 1996). Additionally, abnormal keratinization was observed after IL-1 treatment of microdissected follicular infundibula (Guy and Kealey, 1998), consistent with in vivo observations of irregular expansions in the bulge region and abnormalities of the infundibulum in PP skin (Shuster, 1972, Wilson, Dean, 1994). It is of interest to consider how these HF alterations could be related to SG dysfunction. In asebia mice, loss of SG function is associated with elongated follicles that ultimately rupture, eventuating in a scarring alopecia (Sundberg, Boggess, 2000). The inner root sheath fails to degrade in this mutant, remaining attached to the emerging hair shaft. Studies of microdissected follicles have demonstrated a dependence upon SGs for separation of the inner root sheath from the hair shaft, leading to the hypothesis that the adherent sheath restrains the shaft from growing out of the follicle, causing the follicle to extend downwards towards the subcutis (Sundberg, Boggess, 2000). Consistent with this hypothesis, we found that HFs were significantly longer in PP than in PN skin (Figure 4A). While it remains possible that inflammatory cells and/or mediators might directly inhibit hair loss in psoriasis, under this scenario, shorter and/or damaged follicles might have been expected. Regarding those cases of psoriatic alopecia that do develop scarring, we would agree with other authors (Bardazzi, Fanti, 1999, Runne and Kroneisen-Wiersma, 1992) that secondary infection may play a prominent role, as sebum is known to contain an number of antimicrobial peptides and lipid species (Dahlhoff et al., 2016, Georgel, Crozat, 2005).
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 9
Eccrine sweat glands in psoriatic lesions
Author Manuscript
Sweating has long been known to be markedly reduced in psoriatic lesions (Johnson et al., 1970, Shuster and Johnson, 1969). As assessed by skin capacitance mapping, this appears to be due to a marked reduction of functioning acrosyringial ostia, which is confined to psoriasis lesions (Xhauflaire-Uhoda et al., 2006). We found a significant, 67% reduction in eccrine duct volume in PP vs. PN skin, when measured as a percentage of total epithelial volume (Figure 4D). However, consistent with a prior report (Headington, Gupta, 1989), we found no difference in the numbers of ESG per biopsy (Figure 4B). Rather, this reduction appears to be due to the marked increase in epidermal volume in PP vs. PN skin (Figure 4C). By contrast, the contribution of SGs to total epithelial volume was reduced to a much greater extent (98%) than that of eccrine ducts (Figure 4D). Thus, in marked contrast to SGs, ESGs do not appear to be structurally compromised in PP vs. NN skin, even though they may be functionally compromised due to abnormalities of the acrosyringium.
Author Manuscript
Taken together, our data suggest that psoriatic inflammation results in profound atrophy of SG. Moreover, IL-17 may be involved in this process, whether derived from T-cells (IL-17A) and/or keratinocytes (IL-17C). In turn, loss of SG function in psoriasis could result in defective separation of the inner root sheath from the hair shaft, resulting in infundibular hyperkeratosis, loss of visible hairs protruding through the epidermis, and HF elongation into the dermis and subcutis, which could lead to scarring alopecia if such elongation were to result in follicular rupture. Future studies of PSU morphology and gene expression in psoriatic patients treated with highly effective biologic therapies targeting IL-17 and other key psoriasis cytokines should allow further exploration of this hypothesis.
Materials and Methods Author Manuscript
Procurement of human tissue samples All procedures were approved by the Institutional Review Board of the University of Michigan Medical School, and all patients provided written informed consent prior to entering the study. Two 6-mm punch biopsies were taken from anatomically-matched PP and PN skin sites of psoriatic individuals under local anesthesia. PN sites were located at least 10 cm from PP sites and all biopsies were from either thigh or buttock. Skin samples were embedded in Tissue-Tek OCT compound (Miles Scientific Laboratories, Ltd., Naperville, IL), and stored at −80°C until pr ocessing. Histology and imaging
Author Manuscript
Consecutive 20 μm thick-frozen skin sections were collected on charged slides (ThermoFisher Scientific, Waltham, MA) by cutting skin samples horizontally (parallel to the dermal-epidermal junction), and were kept at −80°C until processing. Upon thawing, skin sections were fixed in 2% paraformaldehyde, and rinsed in distilled water. Routine hematoxylin and eosin staining was followed by tissue dehydration and mounting with Permount medium (ThermoFisher Scientific). Oil Red O staining was performed for 15 min with freshly prepared Oil Red O staining solution [obtained by diluting a 0.5% stock solution of Oil Red O (Sigma Aldrich, St. Louis, MO) in isopropanol with distilled water (3:5, v/v), allowed to stand for 10 minutes, and filtered]. Slides were then thoroughly rinsed
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 10
Author Manuscript
in 60% isopropanol, mounted with 90% glycerin (ThermoFisher Scientific, Waltham, MA), and sealed with clear nail polish. Digital imaging of horizontal sections was performed with a Leica MXFL III Stereo Microscope (Leica Microsystems, Inc., Buffalo Grove, IL; accessed at the Microscopy and Image-Analysis Laboratory, Biomedical Research Core Facilities, University of Michigan Medical School) or with an Axioskop 2 (Carl Zeiss Inc., Thornwood, NY) equipped with a Spot digital camera (Spot Imaging Solutions, Sterling Heights, MI). A microscope micrometer (ThermoFisher Scientific) was used during imaging for image calibration. Computer-assisted 3D-reconstructions and morphometric analysis
Author Manuscript
A series of ~200–400 transversely-cut, consecutive, stained, and imaged sections was generated for each biopsy. Digital images were imported in Reconstruct 1.1 software (Fiala, 2005), aligned, and reconstructed as previously described (Rittie et al., 2013). Whole biopsy reconstructs were generated from sections including the outermost layers of the stratum corneum, down to the end of the tallest HF, representing a total of 2 to 4mm. For the partial reconstructs presented in Figure 2, 50 sections (i.e. 1,000 μm) encompassing the SGs were utilized. Animated 3D renderings and still 3D images were generated using Blender 2.68 (http://www.blender.org). Morphometric measurements and quantifications were performed on calibrated traces and shapes exported from Reconstruct 1.1. HF length was calculated from the number of 20μm consecutive sections encompassing the topmost remarkable circle of the HF (before the HF merges with the epidermis on horizontal sections) down to the bottom of the HF bulb. Data are presented as mean ± SD, or as box-and-whisker graphs (boxes extend from 25th to 75th percentiles, lines are medians, whiskers are min to max), generated with GraphPad Prism 6.
Author Manuscript
Assessment of visible hairs Photographs of biopsy sites from our previous RNA-seq study (Li, Tsoi, 2014) were available for 71 of the cases. Of these, 52 also had photos of uninvolved (PN) skin biopsies, which were harvest from skin at least 10 cm away from the nearest plaque. Photographs were also available for 52 of the normal (NN) controls. Photographs of biopsy sites were obtained at a distance of 2–3 cm using the macro mode on a Sony Cyber-Shot 16-megapixel digital camera. Hairs were defined to be present at a specified site if at least 2 hairs were visible in the photographic field. Transcriptome analysis of SH lesions compared to normal skin
Author Manuscript
Formalin-fixed, paraffin-embedded blocks of SH (four different individuals) and NN skin (three different individuals) were accessed from the archives of the Department of Pathology at the University of Michigan, and diagnosis was confirmed by review of archival hematoxylin and eosin stained sections by a board-certified dermatopathologist (PWH). Five 20 micron microtome sections from paraffin-embedded SH samples (n=5) and normal skin (n=3) were used to isolate RNA with the E.Z.N.A. FFPE RNA kit (Omega bio-tek) according to manufacturer’s instructions. RNA quantity and quality assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies). Gene expression profiling of SH (n = 4) and normal skin (n = 3) samples was performed using the Affymetrix Human Gene 2.1 ST array platform. Robust multichip average (RMA) was used to normalize arrays and generate J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 11
Author Manuscript
expression summaries at the transcript level (R package: oligo) (Irizarry et al., 2003). Additional details are presented in the Supplementary Methods. Identification of SG signature genes SG signature genes were identified by taking the intersection of genes overexpressed in SH vs. NN skin, and genes whose expression was decreased in PP vs. NN skin, as determined previously (Li, Tsoi, 2014). A criterion of 4-fold differential expression (overexpression in SH vs. NN skin, and reduced expression in PP vs. NN skin), was selected by inspection of the plots shown in Figure S1. The expression behavior of this SG signature gene set in PP vs. NN skin was subsequently correlated with those of signature sets derived for various skin-resident cell types and keratinocytes treated with various cytokines, as described previously (Swindell et al., 2013, Swindell, Sarkar, 2015, Swindell et al., 2014). Additional details are presented in the Supplementary Methods.
Author Manuscript
Statistical Methods Differences between PP and PN skin biopsies with respect to SG count, SG volume and HF length were evaluated using maximum likelihood linear mixed-effect models, with biopsy type (PP or PN) treated as a fixed effect and patient treated as a random effect (R packages: lme4 and lmerTest). SG counts were classified as normal (1 gland per follicle) or abnormal (0 or ≥ 2 glands per follicle) and modeled as a binomial response variable (lme4 function: glmer with family = “binomial”). Sebaceous gland volumes (log10-transformed) and hair follicle lengths were treated as Gaussian response variables (lmerTest function: lmer).
Supplementary Material Author Manuscript
Refer to Web version on PubMed Central for supplementary material.
Acknowledgments We thank the volunteers who provided skin biopsies for this study, and Elyssa A. Farr and Jenna E. Youssif for their technical help. This research was supported by NIH R01 grants AR042742, AR050511, AR054966, AR062382, and AR065183 to JTE, as well as K08 AR060802 to JEG and K01 AR059678 to LR. JTE and TT are supported by the Ann Arbor Veterans Affairs Hospital, and JEG is supported by the Doris Duke Charitable Foundation Grant #2013106 and Taubman Medical Institute (as the Kenneth and Frances Eisenberg Emerging Scholar). WRS received support from the American Skin Association Carson Family Research Scholar Award in Psoriasis. The Microscopy and Image-analysis Laboratory is a multiuser imaging facility administered by the Biomedical Research Core Facilities, Office of Research, University of Michigan Medical School. We also acknowledge generous support from the Dawn and Dudley Holmes Memorial Fund and the Babcock Endowment Fund to the Department of Dermatology at the University of Michigan.
Author Manuscript
Abbreviations ESG
eccrine sweat gland
HF
hair follicle
NN
normal skin
ORO
Oil Red O
PN
non-lesional psoriatic skin
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 12
Author Manuscript
PP
lesional psoriatic skin
PSU
pilosebaceous unit
SD
standard deviation
SG
sebaceous gland
SH
sebaceous hyperplasia
IL-17
interleukin-17
References
Author Manuscript Author Manuscript Author Manuscript
Ahmed M, Gaffen SL. IL-17 inhibits adipogenesis in part via C/EBPalpha, PPARgamma and Kruppellike factors. Cytokine. 2013; 61:898–905. [PubMed: 23332504] Alestas T, Ganceviciene R, Fimmel S, Muller-Decker K, Zouboulis CC. Enzymes involved in the biosynthesis of leukotriene B4 and prostaglandin E2 are active in sebaceous glands. J Mol Med (Berl). 2006; 84:75–87. [PubMed: 16388388] Bardazzi F, Fanti PA, Orlandi C, Chieregato C, Misciali C. Psoriatic scarring alopecia: observations in four patients. International journal of dermatology. 1999; 38:765–8. [PubMed: 10561048] Barrault C, Garnier J, Pedretti N, Cordier-Dirikoc S, Ratineau E, Deguercy A, et al. Androgens induce sebaceous differentiation in sebocyte cells expressing a stable functional androgen receptor. The Journal of steroid biochemistry and molecular biology. 2015; 152:34–44. [PubMed: 25864624] Cheng JB, Russell DW. Mammalian wax biosynthesis. I. Identification of two fatty acyl-Coenzyme A reductases with different substrate specificities and tissue distributions. J Biol Chem. 2004; 279:37789–97. [PubMed: 15220348] Cockayne SE, Messenger AG. Familial scarring alopecia associated with scalp psoriasis. The British journal of dermatology. 2001; 144:425–7. [PubMed: 11251592] Cui Y, Liu Z, Sun X, Hou X, Qu B, Zhao F, et al. Thyroid hormone responsive protein spot 14 enhances lipogenesis in bovine mammary epithelial cells. In Vitro Cell Dev Biol Anim. 2015; 51:586–94. [PubMed: 25608868] Dahlhoff M, Zouboulis CC, Schneider MR. Expression of dermcidin in sebocytes supports a role for sebum in the constitutive innate defense of human skin. Journal of dermatological science. 2016; 81:124–6. [PubMed: 26718508] Dozsa A, Dezso B, Toth BI, Bacsi A, Poliska S, Camera E, et al. PPARgamma-mediated and arachidonic acid-dependent signaling is involved in differentiation and lipid production of human sebocytes. J Invest Dermatol. 2014; 134:910–20. [PubMed: 24129064] Eisen DB, Michael DJ. Sebaceous lesions and their associated syndromes: part I. J Am Acad Dermatol. 2009; 61:549–60. quiz 61–2. [PubMed: 19751879] Fiala JC. Reconstruct: a free editor for serial section microscopy. Journal of microscopy. 2005; 218:52–61. [PubMed: 15817063] Flowers MT, Paton CM, O’Byrne SM, Schiesser K, Dawson JA, Blaner WS, et al. Metabolic changes in skin caused by Scd1 deficiency: a focus on retinol metabolism. PLoS One. 2011; 6:e19734. [PubMed: 21573029] Fujiwara H, Ferreira M, Donati G, Marciano DK, Linton JM, Sato Y, et al. The basement membrane of hair follicle stem cells is a muscle cell niche. Cell. 2011; 144:577–89. [PubMed: 21335239] Georgel P, Crozat K, Lauth X, Makrantonaki E, Seltmann H, Sovath S, et al. A toll-like receptor 2responsive lipid effector pathway protects mammals against skin infections with gram-positive bacteria. Infect Immun. 2005; 73:4512–21. [PubMed: 16040962] Gibbons GF, Islam K, Pease RJ. Mobilisation of triacylglycerol stores. Biochimica et biophysica acta. 2000; 1483:37–57. [PubMed: 10601694]
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 13
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Gudjonsson, JE.; Elder, JT. Psoriasis. In: Goldsmith, L.; Katz, SI.; Gilchrest, BA.; Paller, AS.; Woolf, K.; Leffell, DJ., editors. Dermatology in General Medicine. 8. New York: McGraw-Hill; 2012. p. 197-231. Guy R, Downie M, Kealey T. The organ maintained human sebaceous gland. Experimental dermatology. 1999; 8:315–7. [PubMed: 10439245] Guy R, Kealey T. The effects of inflammatory cytokines on the isolated human sebaceous infundibulum. J Invest Dermatol. 1998; 110:410–5. [PubMed: 9540984] Harden JL, Krueger JG, Bowcock AM. The immunogenetics of Psoriasis: A comprehensive review. Journal of autoimmunity. 2015; 64:66–73. [PubMed: 26215033] Headington JT, Gupta AK, Goldfarb MT, Nickoloff BJ, Hamilton TA, Ellis CN, et al. A morphometric and histologic study of the scalp in psoriasis. Paradoxical sebaceous gland atrophy and decreased hair shaft diameters without alopecia. Arch Dermatol. 1989; 125:639–42. [PubMed: 2637673] Holmes RS. Comparative genomics and proteomics of vertebrate diacylglycerol acyltransferase (DGAT), acyl CoA wax alcohol acyltransferase (AWAT) and monoacylglycerol acyltransferase (MGAT). Comparative biochemistry and physiology Part D, Genomics & proteomics. 2010; 5:45– 54. Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, Speed TP. Summaries of Affymetrix GeneChip probe level data. Nucleic acids research. 2003; 31:e15. [PubMed: 12582260] Johnson C, Dawber R, Shuster S. Surface appearances of the eccrine sweat duct by scanning electron microscopy. The British journal of dermatology. 1970; 83:655–60. [PubMed: 5492476] Kocher O, Krieger M. Role of the adaptor protein PDZK1 in controlling the HDL receptor SR-BI. Current opinion in lipidology. 2009; 20:236–41. [PubMed: 19421056] Lang JC, Schuller DE. Differential expression of a novel serine protease homologue in squamous cell carcinoma of the head and neck. British journal of cancer. 2001; 84:237–43. [PubMed: 11161383] Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008; 9:559. [PubMed: 19114008] Li B, Tsoi LC, Swindell WR, Gudjonsson JE, Tejasvi T, Johnston A, et al. Transcriptome analysis of psoriasis in a large case-control Sample: RNA-Seq rovides insights into disease mechanisms. J Invest Dermatol. 2014; 134:1828–38. [PubMed: 24441097] Liakou, AI. Volume and Differentiation of Sebaceous Glands in Psorisis [MD]. Berlin: Charité Medical University; 2015. Miyazaki M, Man WC, Ntambi JM. Targeted disruption of stearoyl-CoA desaturase1 gene in mice causes atrophy of sebaceous and meibomian glands and depletion of wax esters in the eyelid. The Journal of nutrition. 2001; 131:2260–8. [PubMed: 11533264] Moncur JT, Park JP, Memoli VA, Mohandas TK, Kinlaw WB. The “Spot 14” gene resides on the telomeric end of the 11q13 amplicon and is expressed in lipogenic breast cancers: implications for control of tumor metabolism. Proc Natl Acad Sci U S A. 1998; 95:6989–94. [PubMed: 9618526] Nourbakhsh M, Douglas DN, Pu CH, Lewis JT, Kawahara T, Lisboa LF, et al. Arylacetamide deacetylase: a novel host factor with important roles in the lipolysis of cellular triacylglycerol stores, VLDL assembly and HCV production. Journal of hepatology. 2013; 59:336–43. [PubMed: 23542347] Parisi R, Symmons DP, Griffiths CE, Ashcroft DM, et al. Identification, Management of P. Global epidemiology of psoriasis: a systematic review of incidence and prevalence. J Invest Dermatol. 2013; 133:377–85. [PubMed: 23014338] Philpott MP, Sanders DA, Bowen J, Kealey T. Effects of interleukins, colony-stimulating factor and tumour necrosis factor on human hair follicle growth in vitro: a possible role for interleukin-1 and tumour necrosis factor-alpha in alopecia areata. The British journal of dermatology. 1996; 135:942–8. [PubMed: 8977716] Quigley DA, To MD, Perez-Losada J, Pelorosso FG, Mao JH, Nagase H, et al. Genetic architecture of mouse skin inflammation and tumour susceptibility. Nature. 2009; 458:505–8. [PubMed: 19136944] Ramirez-Zacarias JL, Castro-Munozledo F, Kuri-Harcuch W. Quantitation of adipose conversion and triglycerides by staining intracytoplasmic lipids with Oil red O. Histochemistry. 1992; 97:493–7. [PubMed: 1385366]
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 14
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Ramos-e-Silva M, Pirmez R. Disorders of hair growth and the pilosebaceous unit: facts and controversies. Clinics in dermatology. 2013; 31:759–63. [PubMed: 24160282] Rittie L, Sachs DL, Orringer JS, Voorhees JJ, Fisher GJ. Eccrine sweat glands are major contributors to reepithelialization of human wounds. Am J Pathol. 2013; 182:163–71. [PubMed: 23159944] Runne U, Kroneisen-Wiersma P. Psoriatic alopecia: acute and chronic hair loss in 47 patients with scalp psoriasis. Dermatology. 1992; 185:82–7. [PubMed: 1421635] Sampath H, Flowers MT, Liu X, Paton CM, Sullivan R, Chu K, et al. Skin-specific deletion of stearoyl-CoA desaturase-1 alters skin lipid composition and protects mice from high fat dietinduced obesity. J Biol Chem. 2009; 284:19961–73. [PubMed: 19429677] Shuster S. Psoriatic alopecia. The British journal of dermatology. 1972; 87:73–7. [PubMed: 5043203] Shuster S. Psoriatic alopecia. Arch Dermatol. 1990; 126:397. Shuster S, Johnson C. The abnormality of sweat duct function in psoriasis. The British journal of dermatology. 1969; 81:846–50. [PubMed: 5359903] Silva CY, Brown KL, Kurban AK, Mahalingam M. Psoriatic alopecia - fact or fiction? A clinicohistopathologic reappraisal. Indian journal of dermatology, venereology and leprology. 2012; 78:611–9. Song WC, Hwang WJ, Shin C, Koh KS. A new model for the morphology of the arrector pili muscle in the follicular unit based on three-dimensional reconstruction. Journal of anatomy. 2006; 208:643–8. [PubMed: 16637886] Sundberg JP, Boggess D, Sundberg BA, Eilertsen K, Parimoo S, Filippi M, et al. Asebia-2J (Scd1(ab2J)): a new allele and a model for scarring alopecia. Am J Pathol. 2000; 156:2067–75. [PubMed: 10854228] Swindell WR, Johnston A, Voorhees JJ, Elder JT, Gudjonsson JE. Dissecting the psoriasis transcriptome: inflammatory- and cytokine-driven gene expression in lesions from 163 patients. BMC genomics. 2013; 14:527. [PubMed: 23915137] Swindell WR, Sarkar MK, Stuart PE, Voorhees JJ, Elder JT, Johnston A, et al. Psoriasis drug development and GWAS interpretation through in silico analysis of transcription factor binding sites. Clin Transl Med. 2015; 4:13. [PubMed: 25883770] Swindell WR, Stuart PE, Sarkar MK, Voorhees JJ, Elder JT, Johnston A, et al. Cellular dissection of psoriasis for transcriptome analyses and the post-GWAS era. BMC medical genomics. 2014; 7:27. [PubMed: 24885462] Tontonoz P, Spiegelman BM. Fat and beyond: the diverse biology of PPARgamma. Annual review of biochemistry. 2008; 77:289–312. Turkish AR, Henneberry AL, Cromley D, Padamsee M, Oelkers P, Bazzi H, et al. Identification of two novel human acyl-CoA wax alcohol acyltransferases: members of the diacylglycerol acyltransferase 2 (DGAT2) gene superfamily. J Biol Chem. 2005; 280:14755–64. [PubMed: 15671038] van de Kerkhof PC, Chang A. Scarring alopecia and psoriasis. The British journal of dermatology. 1992; 126:524–5. [PubMed: 1610698] Werner B, Brenner FM, Boer A. Histopathologic study of scalp psoriasis: peculiar features including sebaceous gland atrophy. The American Journal of dermatopathology. 2008; 30:93–100. [PubMed: 18360109] Wertz PW, Swartzendruber DC, Madison KC, Downing DT. Composition and morphology of epidermal cyst lipids. J Invest Dermatol. 1987; 89:419–25. [PubMed: 3668284] Wilkinson DI, Karasek MA. Skin lipids of a normal and mutant (asebic) mouse strain. J Invest Dermatol. 1966; 47:449–55. [PubMed: 5924301] Wilson CL, Dean D, Lane EB, Dawber RP, Leigh IM. Keratinocyte differentiation in psoriatic scalp: morphology and expression of epithelial keratins. The British journal of dermatology. 1994; 131:191–200. [PubMed: 7522513] Wright AL, Messenger AG. Scarring alopecia in psoriasis. Acta dermato-venereologica. 1990; 70:156– 9. [PubMed: 1969203] Wu B, Potter CS, Silva KA, Liang Y, Reinholdt LG, Alley LM, et al. Mutations in sterol Oacyltransferase 1 (Soat1) result in hair interior defects in AKR/J mice. J Invest Dermatol. 2010; 130:2666–8. [PubMed: 20574437] J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 15
Author Manuscript
Xhauflaire-Uhoda E, Pierard-Franchimont C, Pierard GE. Skin capacitance mapping of psoriasis. Journal of the European Academy of Dermatology and Venereology : JEADV. 2006; 20:1261–5. [PubMed: 17062043] Zhang WC, Shyh-Chang N, Yang H, Rai A, Umashankar S, Ma S, et al. Glycine decarboxylase activity drives non-small cell lung cancer tumor-initiating cells and tumorigenesis. Cell. 2012; 148:259–72. [PubMed: 22225612]
Author Manuscript Author Manuscript Author Manuscript J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 16
Author Manuscript Author Manuscript Author Manuscript
Figure 1. Sebaceous glands are atrophic in lesional (PP) vs. non-lesional (PN) psoriatic skin
Author Manuscript
A) (Printed version) 3D reconstruction of PP and PN skin sample pair #1 (figure shown is representative of rotational views of three PP/PN pairs, which are shown in the electronic version of this manuscript). Note SG atrophy in the PP skin sample relative to its paired PN sample. SGs are colored light blue, HFs purple, hair canals dark blue, eccrine ducts magenta, and epidermis yellow. A) (Electronic version) Rotational views of 3D reconstructions of three PP/PN pairs showing sebaceous gland atrophy in each PP skin sample relative to its paired PN sample. SG are colored light blue, HFs purple, hair canals dark blue eccrine ducts magenta, and epidermis yellow. B) Quantification of SG volumes in the three pairs shown in (A). C) Distribution of hair follicles with normal (i.e. 1) vs. abnormal SG count (0 or >1 per hair follicle). Data complement the contingency Table S1; n = 6 pairs.
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 17
Author Manuscript Author Manuscript
Figure 2. Oil Red O staining is markedly reduced in SGs, but not stratum corneum, of PP vs. PN skin
Author Manuscript
A) SG staining in PP vs. PN skin. Top row: 3D reconstruction of ORO-stained PSUs (representative images of 1 pilosebaceous ‘group’ per subject are shown; 3 groups were found in each biopsy). Hair follicles are in blue, SGs are in cyan. Middle and bottom rows: aspect of ORO staining in sections corresponding to the colored lines on reconstructs (red and purple in middle and bottom, respectively). Scale bars indicate 100 μm B) Stratum corneum staining in PP vs. PN skin. Note the similar degree of stratum corneum staining between PP and PN skin (top panels, in contrast to markedly lower ORO staining in SG from the same individual, horizontal section). Very similar results were obtained in the other two individuals studied. Scale bar indicates 100 μm.
Author Manuscript J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 18
Author Manuscript Author Manuscript Author Manuscript Figure 3. Loss of visible hairs in PP vs. PN skin
Author Manuscript
A) Example of a clinical photograph of lesional and surrounding perilesional skin. Circles indicate two biopsy sites chosen in lesional skin. B) Quantitation of hairs. Visible hairs were assessed in photographs of PP (lesional) skin from 71 cases, of whom 52 also had photos of PN (non-lesional) skin biopsy sites, and in 52 NN (controls) skin biopsies. Asterisks indicate p < 0.0001 by Fisher’s exact test, with PP skin as the referent.
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 19
Author Manuscript Author Manuscript Author Manuscript
Figure 4. Expansion of the epidermis reduces the relative contribution of appendageal structures to the epithelial compartment of psoriatic skin
Quantitation of A) hair follicle length, B) appendage count for eccrine sweat glands (ESGs), hair follicles (HFs), C) epidermal volume per biopsy, and D) percentage of total epithelial volume occupied by ESGs, HFs, SGs and epidermis in 3D reconstructs of PP vs. PN skin (n = 3 pairs for all). A) box-and-whisker graphs (boxes extend from 25th to 75th percentiles, lines are medians, whiskers are min to max). B, D) Mean±SD. *: p
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01. Published in final edited form as: J Invest Dermatol. 2016 September ; 136(9): 1792–1800. doi:10.1016/j.jid.2016.05.113.
Sebaceous gland atrophy in psoriasis: An explanation for psoriatic alopecia? Laure Rittié1, Trilokraj Tejasvi1,3, Paul W. Harms1,2, Xianying Xing1, Rajan P. Nair1, Johann E. Gudjonsson1, William R Swindell1, and James T. Elder1,3,*
Author Manuscript
1Department
of Dermatology, University of Michigan, Ann Arbor, MI 48109
2Department
of Pathology, University of Michigan, Ann Arbor, MI 48109
3Ann
Arbor Veterans Affairs Hospital, Ann Arbor, MI
Abstract
Author Manuscript
In a transcriptome study of psoriatic (PP) vs. normal (NN) skin, we found a co-expressed gene module (N5) enriched 11.5-fold for lipid biosynthetic genes. We also observed fewer visible hairs in PP skin, compared to uninvolved (PN) or NN skin (p 4, 54.1-fold). The intersection of PP-downregulated and SH-upregulated gene lists generated a gene expression signature consisting solely of module N5 genes, whose expression in PP vs. NN skin was inversely correlated with the signature of IL17-stimuated keratinocytes. Despite loss of visible hairs, morphometry identified elongated follicles in PP vs. PN skin (average 1.7 vs. 1.2 μm, p=0.020). These results document SG atrophy in non-scalp psoriasis, identify a cytokine-regulated set of SG signature genes, and suggest that loss of visible hair in PP skin may result from abnormal SG function.
Keywords Sebaceous glands; eccrine sweat glands; hair follicles; alopecia; psoriasis
Author Manuscript
Introduction Psoriasis is a common inflammatory and hyperplastic skin disease affecting approximately 1–2% of European-origin individuals (Parisi et al., 2013). Immunological and genetic studies are converging upon an immunopathogenesis involving altered antigen presentation
*
Corresponding author. 7412 Medical Sciences Building 1, University of Michigan, 1301 E. Catherine, Ann Arbor, Michigan 48109-5675, USA, phone (734) 647-8070, [email protected]. 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 citable 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.
Rittié et al.
Page 2
Author Manuscript
and dysregulated signal transduction involving NF-κB activation, interferon signaling, and the IL-23-Th17 axis (Harden et al., 2015). Based on findings from our recent RNA-seq transcriptome study (Li et al., 2014), we became interested in exploring the relationship between psoriasis, hair growth, and sebaceous gland health, a subject that has received only limited attention in the literature.
Author Manuscript
Our transcriptome study used weighted gene co-expression network analysis (Langfelder and Horvath, 2008) to identify coordinately expressed gene modules in lesional psoriatic (PP) skin and normal (NN) skin. We identified one module of coordinately-expressed NN transcripts (module N5) that was highly enriched for genes involved in the metabolism of lipids and lipoproteins (p = 3 x 10−40). Another module (N17) was enriched (p = 1.5 x 10−11) for genes in the peroxisome proliferator activator receptor-γ (PPAR-γ) signaling pathway, which plays a central role in adipocyte differentiation (Alestas et al., 2006, Dozsa et al., 2014, Tontonoz and Spiegelman, 2008). Both modules manifested a high proportion of significantly down-regulated genes (28.8% and 52.5%, respectively) (Li, Tsoi, 2014). Our study also identified clusters of muscle- and hair follicle (HF)-related genes whose expression was similarly down-regulated in PP vs. NN skin (Li, Tsoi, 2014). Because the pilosebaceous unit (PSU) consists of the sebaceous gland (SG) and the arrector pili muscle in addition to the HF itself, we wondered whether our observations might be related to alterations in the PSU in psoriasis. Because it is difficult to obtain PSUs in the plane of section from the single vertical sections that are typically available in clinical practice, we (i) carried out 3-dimensional (3D) morphometry on transverse sections of six paired biopsies of PP vs. PN skin, (ii) utilized biopsy site photographs to assess the number of visible hairs in PP vs. PN vs. NN skin, and (iii) compared the expression behaviors of previously-defined coexpression modules in PP vs. NN skin (Li, Tsoi, 2014) to newly-generated gene expression profiles generated from lesions of SH and NN skin.
Author Manuscript
Results
Author Manuscript
Research subjects were individuals affected with psoriasis vulgaris for at least 8 years (age = 44.7±18.7 years [mean±SD], n=6, 50% male, 50% female). We performed 3D reconstructions of horizontal sections encompassing the complete follicular unit for three pairs of PP and PN biopsies. As visualized in Figure 1A (animated in the online version), SGs were markedly atrophic in PP vs. PN skin. Morphometric analysis demonstrated a 91% overall average reduction in sebaceous gland volume (Figure 1B), with a consistentlydecreased SG volume for each HF (Figure 1A) (P = 0.031; n = 38 sebaceous glands from 3 psoriasis patients). The estimated sebaceous gland volume in PP lesions is 3.5% of that in PN lesions (95% confidence interval: 0.5% – 25.2%). In addition to the marked decrease in SG volume, SG count was abnormal (either 0 or >1 SG per HF) more frequently in PP compared to PN skin (50% for PP vs. 17% for PN skin; P = 0.044; n = 67 hair follicles from 3 psoriasis patients). The estimated odds ratio is 5.12 (95% confidence interval: 1.04 – 25.2), indicating that the odds of an abnormal sebaceous gland count is 5.12 times greater for a PP follicle than for a PN follicle (Figure 1C and Table S1). Utilizing PP/PN pairs from three additional psoriasis patients, we confirmed these results by histology of individual sections using Oil Red O (ORO), a dye with high affinity for neutral
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 3
Author Manuscript
triglycerides and lipids (Ramirez-Zacarias et al., 1992), (Figure 2A). As expected given the lower neutral lipid content of stratum corneum compared to SGs (see Discussion), ORO staining was less pronounced for stratum corneum compared to SGs, and did not appear to differ between PP and PN skin (Figure 2B). As illustrated in Figure 3A for a representative patient, visible hairs are strikingly reduced in psoriatic lesions. To quantify this phenomenon, we utilized photographs of biopsy sites from our previous RNA-seq study (Li, Tsoi, 2014). As shown in Figure 3B, we found that significantly fewer patients had visible hairs in the psoriatic lesion itself, compared to immediately perilesional skin, nonlesional (PN) skin located at least 5 cm away from the nearest plaque, and normal (N) skin from nonpsoriatic individuals (p < 0.0001 vs. PP skin for all three comparisons, by two-tailed Fisher’s exact test).
Author Manuscript Author Manuscript
In contrast to the marked reduction of visible hairs, our 3D reconstructions revealed no visible damage to HFs in PP vs. PN skin (Figure 1A, Video S1). Indeed, hair follicle mean length was significantly greater in PP than in PN skin Hair follicle length is significantly elevated in PP lesions as compared to uninvolved PN skin (P = 0.020; n = 38 hair follicles from 3 psoriasis patients). The mean hair follicle height in PP lesions was 45% greater than in PN lesions (95% confidence interval: 8.86% – 81.1%, Figure 4A). There was no significant difference in HF counts in PP vs. PN skin, and eccrine duct counts were also similar (Figure 4B). As expected, the volume of the epidermal compartment was markedly (5.83±0.87-fold) and significantly (p=0.011, two-tailed paired t-test) greater in PP vs. PN skin (Figure 4C). The increased HF length noted above combined with the expansion of the epidermis resulted in similar contributions of the HF compartment in PP vs. PN (0.62±0.39fold in PP vs. PN, p=0.26). Due to this expansion of the epidermal compartment, both SGs and eccrine sweat glands (ESGs) made a smaller contribution to the total epithelial volume on a percentage basis in PP vs. PN skin (p=0.0251 and 0.0454, respectively). However, the appendage contribution to total epithelial volume in PP vs. PN skin was much more markedly reduced in the case of SGs (0.1% vs. 3.7%, respectively, overall average reduction of 98%) compared to ESGs (0.4% vs. 1.3% of the epithelial compartment in PP vs PN, overall average reduction of 67%) (Figure 4D).
Author Manuscript
Noting this marked and disproportionate reduction in SG volume, we reasoned that the most markedly down-regulated genes in the PP vs. NN skin transcriptome might be enriched for genes expressed in SG. To test this hypothesis, we queried our psoriasis RNA-seq transcriptome dataset (Li, Tsoi, 2014) to identify co-expression modules enriched for highly down-regulated genes. As noted earlier, module N5 as a whole was unique with respect to its strong enrichment for lipid biosynthetic process genes (p = 2.77 × 10−12, Figure 5, top panel). By assessing enrichments for each module at various FC thresholds, we found that enrichment for module N5 genes increased progressively among genes most strongly downregulated in PP vs. NN skin (Figure S1). Notably, the set of 445 genes with FC < 0.25 in PP vs. NN skin overlapped significantly with module N5 genes (P = 3.72 × 10−45 by Fisher’s exact test; Figure 5, middle panel). In order to generate a SG gene expression signature, we prepared RNA from formalin-fixed, paraffin-embedded sections of three biopsies of NN skin and four biopsies of SH, a common
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 4
Author Manuscript
skin condition characterized by marked enlargement of multilobular SGs sharing a single central lumen (Eisen and Michael, 2009). In contrast to the progressive enrichment for module N5 genes with increasing down-regulation in PP vs. NN skin, module N5 enrichment progressively increased with increasing overexpression in SH vs. NN skin (Figure S1). As shown in the bottom panel of Figure 5, genes overexpressed by > 4-fold in SH compared to normal skin overlapped significantly with module N5 genes (P = 1.9 × 10−31 by Fisher’s exact test). This same level of enrichment was not observed among any of the other modules we evaluated. To generate an SG gene expression signature, we enumerated 12 genes whose expression was upregulated > 4-fold in SH lesions vs. NN skin, and down-regulated > 4-fold PP vs. NN skin (Table S2). Notable among these were genes involved in the synthesis and/or metabolism of triglycerides, very-long-chain fatty acids, sphingolipids, and wax esters, all major components of sebum (Box 1).
Author Manuscript
Box 1 Functional properties of SG signature genes AADACL3 (1p36.21) encodes arylacetamide deacetylase-like 3, for which no specific functional information is available. Arylacetamide deacetylase (AADAC) participates in the mobilization of triglyceride stores for very low density lipoprotein (VLDL) assembly (Gibbons et al., 2000). Hepatitis C virus (HCV) assembles on lipid droplets in the liver, and expression of AADAC is markedly reduced in the liver early in the course of HCV infection (Nourbakhsh et al., 2013)
Author Manuscript
AWAT1 (Xq13.1) encodes an acyl-CoA wax alcohol acyltransferase, which predominantly esterifies long chain (wax) alcohols with acyl-CoA-derived fatty acids to produce wax esters (Holmes, 2010). The enzyme is expressed in many human tissues but predominates in skin. In situ hybridization demonstrates a differentiation-specific expression pattern within the human sebaceous gland for AWAT1 (Turkish et al., 2005). DGAT2L6 (Xq13.1) maps to the X chromosome and encodes a member of the diacylglycerol acyltransferase 2 family that is structurally related to AWAT1 and AWAT2 (Holmes, 2010). The encoded protein is most likely involved in the synthesis of di- or triacylglycerol. FADS2 (11q12.2) encodes a member of the fatty acid desaturase gene family. Desaturase enzymes regulate unsaturation of fatty acids by introducing double bonds between defined carbons of the fatty acyl chain. FADS2 and several other genes involved in sebocyte differentiation are induced by dihydrotesterone in human sebocytes expressing the androgen receptor (Barrault et al., 2015)
Author Manuscript
FAR2 (12p11.22) encodes a member of the short chain dehydrogenase / reductase superfamily, which is a peroxisomal protein with reductase activity that converts fatty acids into fatty alcohols in the first step of wax biosynthesis (Cheng and Russell, 2004). GLDC (9q22) encodes glycine decarboxylase, which is critical for tumor-initiating cells in non-small cell lung cancer (NSCLC) by inducing dramatic changes in glycolysis and glycine/serine metabolism, leading to changes in pyrimidine metabolism to regulate cancer cell proliferation (Zhang et al., 2012).
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 5
Author Manuscript
MOGAT1 (2q36.2) encodes acyl-CoA:monoacylglycerol acyltransferase, which catalyzes the synthesis of diacylglycerols, the precursor of physiologically important lipids such as triacylglycerol and phospholipids. MOGAT1 is structurally related to DGA2L6, AWAT1, and AWAT2 (Holmes, 2010). PDZK1 (1q21) encodes a PDZ domain-containing scaffolding protein that plays an important role in cholesterol metabolism by regulating the HDL receptor, scavenger receptor class B type 1 (Kocher and Krieger, 2009). PM20D1 (1q32.1) encodes a protein of unknown function, which contains a peptidase domain.
Author Manuscript
SOAT1 (1q25) encodes an acyltransferase protein located in the endoplasmic reticulum, which catalyzes the formation of fatty acid-cholesterol esters. SOAT1 has been implicated in the formation of atherosclerotic plaques by controlling the equilibrium between free cholesterol and cytoplasmic cholesteryl esters (Wu et al., 2010). SOAT1 is strongly expressed in SGs of WT mice but not in SGs of mice homozygous for the hair interior defect (hid) allele, which involves a 118 bp deletion in SOAT1 (Wu, Potter, 2010) THRSP (11q14.1) is the human homologue of the rat S14 gene. Expression of S14 is controlled by nutritional and hormonal factors and is restricted to liver and adipose tissue, particularly in lipomatous nodules. Suggestive of a role in lipid metabolism, it is also expressed in lipogenic breast cancers (Moncur et al., 1998) and upregulates lipogenesis in bovine mammary epithelial cells (Cui et al., 2015)
Author Manuscript
TMPRSS11E (4q13.2) encodes a transmembrane serine protease also known as DESC1. Expression of TMPRSS11E is decreased in head and neck squamous cell carcinoma. Its expression is epithelial-specific and limited to tissues derived from the head and neck, in skin, prostate and testes. (Lang and Schuller, 2001).
Author Manuscript
Asebia (Sundberg et al., 2000) and Flake (Georgel et al., 2005) mice lack the enzyme stearoyl-CoA desaturase-1 involved in fatty acid desaturation, due to mutations in the Scd1 gene. Asebia mice manifest marked SG atrophy (Sundberg, Boggess, 2000), and both strains manifest alopecia, skin inflammation, and eye infections. Germ-line (Miyazaki et al., 2001) or skin-targeted (Sampath et al., 2009) deletion of the Scd1 gene results in similar phenotypes. Gene expression profiling of skin-targeted Scd1−/− mice revealed abnormal expression of genes involved in lipid biosynthesis as well as in retinol and steroid metabolism (Flowers et al., 2011). Using this microarray dataset, we observed enrichment for module N5 genes among transcripts downregulated > 4-fold (P = 1.89 × 10−3), in skintargeted Scd1−/− mice relative to WT controls (Figure S2). We next used a previously-defined microarray transcriptome dataset involving 237 paired samples of PP and PN skin (Swindell et al., 2015) to explore the potential roles of various cell types and cytokines in the development of SG atrophy. Applying a criterion of | r | > 0.2, p < 0.0008 to define a meaningful association in the context of multiple testing (Bonferroni correction of 0.05/59 = 0.00085 for 59 tests performed), we found that expression of SG signature genes in PP vs. PN skin was negatively correlated with the gene expression signature of keratinocytes treated with IL-17C (r = −0.23), whereas they were positively J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 6
Author Manuscript
correlated with keratinocyte gene expression responses to IL-4 (r = 0.39) and IL-4 plus IL-13 (r = 0.23) (Table S2, Figure S3). Gene expression signatures from keratinocytes, fibroblasts, and several inflammatory cell types were not inversely correlated with the SG signature, although the macrophage signature was positively correlated (r = 0.32, Table S2).
Discussion
Author Manuscript
There are close anatomic and functional relationships between the HF, the arrector pili muscle, and the SG, which together form the PSU (Fujiwara et al., 2011, Song et al., 2006). Moreover, muscle and HF-related gene co-expression networks are highly correlated in mouse skin (Quigley et al., 2009). Our interest in the status of HFs and SGs in psoriasis was initiated by transcriptome studies in which we identified modules of coordinately-expressed genes enriched in annotation for genes involved in adipogenesis, HFs, and muscle in NN skin, many of which were down-regulated in PP skin (Li, Tsoi, 2014). Based on these observations, we carried out a detailed morphometric analysis of SGs in paired, anatomically-matched samples of PP vs. PN skin. As part of the same study, we assessed the status of ESGs in PP vs. PN skin, a topic that has not received a systematic assessment to date. Sebaceous gland atrophy in psoriasis lesions
Author Manuscript
The most striking morphologic finding from this study was the pronounced atrophy of SGs in PP vs. PN skin, which averaged a 91% overall reduction in volume that was observed in all six biopsy pairs studied (Figures 1 and 2). While earlier studies have shown SG atrophy in scalp psoriasis (Bardazzi et al., 1999, Headington et al., 1989, Silva et al., 2012, Werner et al., 2008, Wilson et al., 1994), here we document SG atrophy in psoriatic lesions at body sites other than the scalp. Similar findings have been reported very recently utilizing sequential vertical sections of PN and PP skin from scalp and back sites (Liakou, 2015), thus confirming our observations. We validated our findings using ORO staining, which specifically stains triglycerides and cholesteryl oleate but no other lipids (Ramirez-Zacarias, Castro-Munozledo, 1992). While triglycerides represent 57% of human sebum lipids by weight (Wilkinson and Karasek, 1966), they appear to be absent from stratum corneum lipids, as assessed by analysis of epidermoid cysts (Wertz et al., 1987). Consistent with this observation, we observed no obvious differences in ORO staining in the stratum corneum of PP vs. PN skin (Figure 2B), in contrast to the marked reduction of ORO staining in SG.
Author Manuscript
As shown in Figure 5, module N5 is uniquely enriched for genes involved in lipid biosynthetic processes. Given (i) the pronounced SG atrophy in psoriatic lesions that we observed morphometrically (Figure 1) and by ORO staining (Figure 2A), we hypothesized that—as opposed to genes involved in lipogenesis in adipose or stratum corneum—lipid biosynthetic genes expressed specifically in SG might be particularly strongly downregulated in PP vs. NN skin due to SG atrophy. Indeed, we found that as the threshold for down-regulation was increased in stringency, the specificity of enrichment for module N5 genes increased (Figure S1).
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 7
Author Manuscript
In support of our hypothesis, we found that module N5 genes were also highly enriched among genes whose expression was the most strongly increased in lesions of SH, relative to normal skin (Figure 5, Figure S1). By intersecting the gene sets decreased > 4-fold in PP vs. NN skin, and increased > 4-fold in SH vs. NN skin, we generated a 12-gene SG signature, with all of the genes mapping to module N5 (Table S2). Literature search revealed a role in lipid biosynthesis and/or localization to SG for essentially all of these genes (Box 1). We found that less stringent choices for the FC cutoff could be used to yield larger SG signature gene sets, with many additional genes providing literature evidence for high expression in SG (data not shown).
Author Manuscript
As a further test of this hypothesis, we queried a microarray-based transcriptome dataset (Flowers, Paton, 2011) derived from the skin of adult mice homozygous, skin targeted deletion of Scd1 (Sampath, Flowers, 2009), again revealing enrichment for module N5 (Figure S2). These mice manifest not only SG atrophy and alopecia, but also substantial inflammatory epidermal hyperplasia. Thus, it is not surprising that genes down-regulated in K5-in skin-targeted Scd1−/− mice are also enriched in other co-expression modules that are also down-regulated in psoriasis (Li, Tsoi, 2014), including modules N14 (hair follicle), N16 (muscular contraction) and N17 (PPAR signaling pathway).
Author Manuscript
Just as epidermal hyperplasia appears to be a response to psoriatic inflammation (Gudjonsson and Elder, 2012), it is not unreasonable to hypothesize that the same might be true for SG atrophy in psoriasis. Consistent with this hypothesis, treatment of isolated SGs with IL-1 or TNF in vitro led to profound reductions in lipid biosynthesis (Guy et al., 1999). However, infiltrating inflammatory cells need not be the primary source of such cytokines. While Headington and colleagues noted that while SG atrophy in scalp psoriasis was associated with increased perifollicular inflammation, this inflammation was not localized to the immediate vicinity of or within SG, but rather to the infundibular area (Headington, Gupta, 1989). We confirmed the paucity of inflammation around SGs and the presence of inflammation surrounding the isthmus and infundibulum in our samples (data not shown). Other potential sources of inflammatory cytokines and other mediators could include keratinocytes, fibroblasts, constitutive tissue dendritic cells, and endothelial cells.
Author Manuscript
Using previously-defined transcriptome datasets across 237 samples of PP and NN skin (Swindell, Sarkar, 2015), we found that the gene expression signature of keratinocytes treated with IL-17C was inversely correlated with the SG signature (Table S2, Figure S3). Notably, IL-17 is not only a key cytokine in psoriasis pathogenesis (Harden, Krueger, 2015), but has also has been found to inhibit the differentiation of adipocyte precursors (Ahmed and Gaffen, 2013). Conversely, keratinocyte responses to the Th2 cytokine IL-4 was positively correlated with the SG signature, whether alone or in combination with IL-13 (Table S2). No gene expression signatures from various inflammatory cell types were inversely correlated with the SG signature, though the macrophage signature was positively correlated (Table S2). Additional mechanistic studies will be required to define the roles of specific cytokines and cell types in the genesis of psoriatic SG atrophy.
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 8
Alopecia in psoriatic lesions
Author Manuscript
Using biopsy site photographs taken in the course of our RNA-seq transcriptome study (Li, Tsoi, 2014), we identified a pronounced loss of visible hairs in psoriatic plaques, relative to immediately perilesional skin, more distant PN skin, or NN skin (Figure 3). While the presence of alopecia in psoriatic lesions has been recognized for over 40 years (Shuster, 1972), there have been few systematic studies of HF structure in psoriatic skin from nonscalp regions. Previous studies of psoriatic alopecia of the scalp found no evidence for significant reductions in total or telogen hair counts in PP vs. PN or normal skin (Bardazzi, Fanti, 1999, Headington, Gupta, 1989, Silva, Brown, 2012, Werner, Brenner, 2008, Wilson, Dean, 1994). In agreement with these observations, we found no significant difference in HF counts per biopsy between PN and PP skin (Figure 4B).
Author Manuscript
A relatively sparse literature, limited to scalp psoriasis, has addressed the question of whether hair loss in psoriatic alopecia is due to scarring (Ramos-e-Silva and Pirmez, 2013, Silva, Brown, 2012). Four reports were case series in which the subjects were pre-selected based on a clinical appearance of scarring alopecia (Bardazzi, Fanti, 1999, Cockayne and Messenger, 2001, van de Kerkhof and Chang, 1992, Wright and Messenger, 1990), whereas larger and unselected series (Runne and Kroneisen-Wiersma, 1992, Shuster, 1972, Werner, Brenner, 2008) revealed a more mixed picture in which non-scarring processes predominated. We found no evidence of misshapen follicles to suggest a scarring process (Figure 1), and it is a common clinical observation that hair re-grows after psoriatic lesions have resolved (Runne and Kroneisen-Wiersma, 1992, Shuster, 1990, Silva, Brown, 2012). Thus, it would appear that while scarring can develop in chronic lesions of psoriasis, scarring is not the primary event leading to loss of visible hairs in psoriatic lesions.
Author Manuscript Author Manuscript
Treatment of isolated HFs with IL-1 or TNF inhibited hair growth (Philpott et al., 1996). Additionally, abnormal keratinization was observed after IL-1 treatment of microdissected follicular infundibula (Guy and Kealey, 1998), consistent with in vivo observations of irregular expansions in the bulge region and abnormalities of the infundibulum in PP skin (Shuster, 1972, Wilson, Dean, 1994). It is of interest to consider how these HF alterations could be related to SG dysfunction. In asebia mice, loss of SG function is associated with elongated follicles that ultimately rupture, eventuating in a scarring alopecia (Sundberg, Boggess, 2000). The inner root sheath fails to degrade in this mutant, remaining attached to the emerging hair shaft. Studies of microdissected follicles have demonstrated a dependence upon SGs for separation of the inner root sheath from the hair shaft, leading to the hypothesis that the adherent sheath restrains the shaft from growing out of the follicle, causing the follicle to extend downwards towards the subcutis (Sundberg, Boggess, 2000). Consistent with this hypothesis, we found that HFs were significantly longer in PP than in PN skin (Figure 4A). While it remains possible that inflammatory cells and/or mediators might directly inhibit hair loss in psoriasis, under this scenario, shorter and/or damaged follicles might have been expected. Regarding those cases of psoriatic alopecia that do develop scarring, we would agree with other authors (Bardazzi, Fanti, 1999, Runne and Kroneisen-Wiersma, 1992) that secondary infection may play a prominent role, as sebum is known to contain an number of antimicrobial peptides and lipid species (Dahlhoff et al., 2016, Georgel, Crozat, 2005).
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 9
Eccrine sweat glands in psoriatic lesions
Author Manuscript
Sweating has long been known to be markedly reduced in psoriatic lesions (Johnson et al., 1970, Shuster and Johnson, 1969). As assessed by skin capacitance mapping, this appears to be due to a marked reduction of functioning acrosyringial ostia, which is confined to psoriasis lesions (Xhauflaire-Uhoda et al., 2006). We found a significant, 67% reduction in eccrine duct volume in PP vs. PN skin, when measured as a percentage of total epithelial volume (Figure 4D). However, consistent with a prior report (Headington, Gupta, 1989), we found no difference in the numbers of ESG per biopsy (Figure 4B). Rather, this reduction appears to be due to the marked increase in epidermal volume in PP vs. PN skin (Figure 4C). By contrast, the contribution of SGs to total epithelial volume was reduced to a much greater extent (98%) than that of eccrine ducts (Figure 4D). Thus, in marked contrast to SGs, ESGs do not appear to be structurally compromised in PP vs. NN skin, even though they may be functionally compromised due to abnormalities of the acrosyringium.
Author Manuscript
Taken together, our data suggest that psoriatic inflammation results in profound atrophy of SG. Moreover, IL-17 may be involved in this process, whether derived from T-cells (IL-17A) and/or keratinocytes (IL-17C). In turn, loss of SG function in psoriasis could result in defective separation of the inner root sheath from the hair shaft, resulting in infundibular hyperkeratosis, loss of visible hairs protruding through the epidermis, and HF elongation into the dermis and subcutis, which could lead to scarring alopecia if such elongation were to result in follicular rupture. Future studies of PSU morphology and gene expression in psoriatic patients treated with highly effective biologic therapies targeting IL-17 and other key psoriasis cytokines should allow further exploration of this hypothesis.
Materials and Methods Author Manuscript
Procurement of human tissue samples All procedures were approved by the Institutional Review Board of the University of Michigan Medical School, and all patients provided written informed consent prior to entering the study. Two 6-mm punch biopsies were taken from anatomically-matched PP and PN skin sites of psoriatic individuals under local anesthesia. PN sites were located at least 10 cm from PP sites and all biopsies were from either thigh or buttock. Skin samples were embedded in Tissue-Tek OCT compound (Miles Scientific Laboratories, Ltd., Naperville, IL), and stored at −80°C until pr ocessing. Histology and imaging
Author Manuscript
Consecutive 20 μm thick-frozen skin sections were collected on charged slides (ThermoFisher Scientific, Waltham, MA) by cutting skin samples horizontally (parallel to the dermal-epidermal junction), and were kept at −80°C until processing. Upon thawing, skin sections were fixed in 2% paraformaldehyde, and rinsed in distilled water. Routine hematoxylin and eosin staining was followed by tissue dehydration and mounting with Permount medium (ThermoFisher Scientific). Oil Red O staining was performed for 15 min with freshly prepared Oil Red O staining solution [obtained by diluting a 0.5% stock solution of Oil Red O (Sigma Aldrich, St. Louis, MO) in isopropanol with distilled water (3:5, v/v), allowed to stand for 10 minutes, and filtered]. Slides were then thoroughly rinsed
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 10
Author Manuscript
in 60% isopropanol, mounted with 90% glycerin (ThermoFisher Scientific, Waltham, MA), and sealed with clear nail polish. Digital imaging of horizontal sections was performed with a Leica MXFL III Stereo Microscope (Leica Microsystems, Inc., Buffalo Grove, IL; accessed at the Microscopy and Image-Analysis Laboratory, Biomedical Research Core Facilities, University of Michigan Medical School) or with an Axioskop 2 (Carl Zeiss Inc., Thornwood, NY) equipped with a Spot digital camera (Spot Imaging Solutions, Sterling Heights, MI). A microscope micrometer (ThermoFisher Scientific) was used during imaging for image calibration. Computer-assisted 3D-reconstructions and morphometric analysis
Author Manuscript
A series of ~200–400 transversely-cut, consecutive, stained, and imaged sections was generated for each biopsy. Digital images were imported in Reconstruct 1.1 software (Fiala, 2005), aligned, and reconstructed as previously described (Rittie et al., 2013). Whole biopsy reconstructs were generated from sections including the outermost layers of the stratum corneum, down to the end of the tallest HF, representing a total of 2 to 4mm. For the partial reconstructs presented in Figure 2, 50 sections (i.e. 1,000 μm) encompassing the SGs were utilized. Animated 3D renderings and still 3D images were generated using Blender 2.68 (http://www.blender.org). Morphometric measurements and quantifications were performed on calibrated traces and shapes exported from Reconstruct 1.1. HF length was calculated from the number of 20μm consecutive sections encompassing the topmost remarkable circle of the HF (before the HF merges with the epidermis on horizontal sections) down to the bottom of the HF bulb. Data are presented as mean ± SD, or as box-and-whisker graphs (boxes extend from 25th to 75th percentiles, lines are medians, whiskers are min to max), generated with GraphPad Prism 6.
Author Manuscript
Assessment of visible hairs Photographs of biopsy sites from our previous RNA-seq study (Li, Tsoi, 2014) were available for 71 of the cases. Of these, 52 also had photos of uninvolved (PN) skin biopsies, which were harvest from skin at least 10 cm away from the nearest plaque. Photographs were also available for 52 of the normal (NN) controls. Photographs of biopsy sites were obtained at a distance of 2–3 cm using the macro mode on a Sony Cyber-Shot 16-megapixel digital camera. Hairs were defined to be present at a specified site if at least 2 hairs were visible in the photographic field. Transcriptome analysis of SH lesions compared to normal skin
Author Manuscript
Formalin-fixed, paraffin-embedded blocks of SH (four different individuals) and NN skin (three different individuals) were accessed from the archives of the Department of Pathology at the University of Michigan, and diagnosis was confirmed by review of archival hematoxylin and eosin stained sections by a board-certified dermatopathologist (PWH). Five 20 micron microtome sections from paraffin-embedded SH samples (n=5) and normal skin (n=3) were used to isolate RNA with the E.Z.N.A. FFPE RNA kit (Omega bio-tek) according to manufacturer’s instructions. RNA quantity and quality assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies). Gene expression profiling of SH (n = 4) and normal skin (n = 3) samples was performed using the Affymetrix Human Gene 2.1 ST array platform. Robust multichip average (RMA) was used to normalize arrays and generate J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 11
Author Manuscript
expression summaries at the transcript level (R package: oligo) (Irizarry et al., 2003). Additional details are presented in the Supplementary Methods. Identification of SG signature genes SG signature genes were identified by taking the intersection of genes overexpressed in SH vs. NN skin, and genes whose expression was decreased in PP vs. NN skin, as determined previously (Li, Tsoi, 2014). A criterion of 4-fold differential expression (overexpression in SH vs. NN skin, and reduced expression in PP vs. NN skin), was selected by inspection of the plots shown in Figure S1. The expression behavior of this SG signature gene set in PP vs. NN skin was subsequently correlated with those of signature sets derived for various skin-resident cell types and keratinocytes treated with various cytokines, as described previously (Swindell et al., 2013, Swindell, Sarkar, 2015, Swindell et al., 2014). Additional details are presented in the Supplementary Methods.
Author Manuscript
Statistical Methods Differences between PP and PN skin biopsies with respect to SG count, SG volume and HF length were evaluated using maximum likelihood linear mixed-effect models, with biopsy type (PP or PN) treated as a fixed effect and patient treated as a random effect (R packages: lme4 and lmerTest). SG counts were classified as normal (1 gland per follicle) or abnormal (0 or ≥ 2 glands per follicle) and modeled as a binomial response variable (lme4 function: glmer with family = “binomial”). Sebaceous gland volumes (log10-transformed) and hair follicle lengths were treated as Gaussian response variables (lmerTest function: lmer).
Supplementary Material Author Manuscript
Refer to Web version on PubMed Central for supplementary material.
Acknowledgments We thank the volunteers who provided skin biopsies for this study, and Elyssa A. Farr and Jenna E. Youssif for their technical help. This research was supported by NIH R01 grants AR042742, AR050511, AR054966, AR062382, and AR065183 to JTE, as well as K08 AR060802 to JEG and K01 AR059678 to LR. JTE and TT are supported by the Ann Arbor Veterans Affairs Hospital, and JEG is supported by the Doris Duke Charitable Foundation Grant #2013106 and Taubman Medical Institute (as the Kenneth and Frances Eisenberg Emerging Scholar). WRS received support from the American Skin Association Carson Family Research Scholar Award in Psoriasis. The Microscopy and Image-analysis Laboratory is a multiuser imaging facility administered by the Biomedical Research Core Facilities, Office of Research, University of Michigan Medical School. We also acknowledge generous support from the Dawn and Dudley Holmes Memorial Fund and the Babcock Endowment Fund to the Department of Dermatology at the University of Michigan.
Author Manuscript
Abbreviations ESG
eccrine sweat gland
HF
hair follicle
NN
normal skin
ORO
Oil Red O
PN
non-lesional psoriatic skin
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 12
Author Manuscript
PP
lesional psoriatic skin
PSU
pilosebaceous unit
SD
standard deviation
SG
sebaceous gland
SH
sebaceous hyperplasia
IL-17
interleukin-17
References
Author Manuscript Author Manuscript Author Manuscript
Ahmed M, Gaffen SL. IL-17 inhibits adipogenesis in part via C/EBPalpha, PPARgamma and Kruppellike factors. Cytokine. 2013; 61:898–905. [PubMed: 23332504] Alestas T, Ganceviciene R, Fimmel S, Muller-Decker K, Zouboulis CC. Enzymes involved in the biosynthesis of leukotriene B4 and prostaglandin E2 are active in sebaceous glands. J Mol Med (Berl). 2006; 84:75–87. [PubMed: 16388388] Bardazzi F, Fanti PA, Orlandi C, Chieregato C, Misciali C. Psoriatic scarring alopecia: observations in four patients. International journal of dermatology. 1999; 38:765–8. [PubMed: 10561048] Barrault C, Garnier J, Pedretti N, Cordier-Dirikoc S, Ratineau E, Deguercy A, et al. Androgens induce sebaceous differentiation in sebocyte cells expressing a stable functional androgen receptor. The Journal of steroid biochemistry and molecular biology. 2015; 152:34–44. [PubMed: 25864624] Cheng JB, Russell DW. Mammalian wax biosynthesis. I. Identification of two fatty acyl-Coenzyme A reductases with different substrate specificities and tissue distributions. J Biol Chem. 2004; 279:37789–97. [PubMed: 15220348] Cockayne SE, Messenger AG. Familial scarring alopecia associated with scalp psoriasis. The British journal of dermatology. 2001; 144:425–7. [PubMed: 11251592] Cui Y, Liu Z, Sun X, Hou X, Qu B, Zhao F, et al. Thyroid hormone responsive protein spot 14 enhances lipogenesis in bovine mammary epithelial cells. In Vitro Cell Dev Biol Anim. 2015; 51:586–94. [PubMed: 25608868] Dahlhoff M, Zouboulis CC, Schneider MR. Expression of dermcidin in sebocytes supports a role for sebum in the constitutive innate defense of human skin. Journal of dermatological science. 2016; 81:124–6. [PubMed: 26718508] Dozsa A, Dezso B, Toth BI, Bacsi A, Poliska S, Camera E, et al. PPARgamma-mediated and arachidonic acid-dependent signaling is involved in differentiation and lipid production of human sebocytes. J Invest Dermatol. 2014; 134:910–20. [PubMed: 24129064] Eisen DB, Michael DJ. Sebaceous lesions and their associated syndromes: part I. J Am Acad Dermatol. 2009; 61:549–60. quiz 61–2. [PubMed: 19751879] Fiala JC. Reconstruct: a free editor for serial section microscopy. Journal of microscopy. 2005; 218:52–61. [PubMed: 15817063] Flowers MT, Paton CM, O’Byrne SM, Schiesser K, Dawson JA, Blaner WS, et al. Metabolic changes in skin caused by Scd1 deficiency: a focus on retinol metabolism. PLoS One. 2011; 6:e19734. [PubMed: 21573029] Fujiwara H, Ferreira M, Donati G, Marciano DK, Linton JM, Sato Y, et al. The basement membrane of hair follicle stem cells is a muscle cell niche. Cell. 2011; 144:577–89. [PubMed: 21335239] Georgel P, Crozat K, Lauth X, Makrantonaki E, Seltmann H, Sovath S, et al. A toll-like receptor 2responsive lipid effector pathway protects mammals against skin infections with gram-positive bacteria. Infect Immun. 2005; 73:4512–21. [PubMed: 16040962] Gibbons GF, Islam K, Pease RJ. Mobilisation of triacylglycerol stores. Biochimica et biophysica acta. 2000; 1483:37–57. [PubMed: 10601694]
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 13
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Gudjonsson, JE.; Elder, JT. Psoriasis. In: Goldsmith, L.; Katz, SI.; Gilchrest, BA.; Paller, AS.; Woolf, K.; Leffell, DJ., editors. Dermatology in General Medicine. 8. New York: McGraw-Hill; 2012. p. 197-231. Guy R, Downie M, Kealey T. The organ maintained human sebaceous gland. Experimental dermatology. 1999; 8:315–7. [PubMed: 10439245] Guy R, Kealey T. The effects of inflammatory cytokines on the isolated human sebaceous infundibulum. J Invest Dermatol. 1998; 110:410–5. [PubMed: 9540984] Harden JL, Krueger JG, Bowcock AM. The immunogenetics of Psoriasis: A comprehensive review. Journal of autoimmunity. 2015; 64:66–73. [PubMed: 26215033] Headington JT, Gupta AK, Goldfarb MT, Nickoloff BJ, Hamilton TA, Ellis CN, et al. A morphometric and histologic study of the scalp in psoriasis. Paradoxical sebaceous gland atrophy and decreased hair shaft diameters without alopecia. Arch Dermatol. 1989; 125:639–42. [PubMed: 2637673] Holmes RS. Comparative genomics and proteomics of vertebrate diacylglycerol acyltransferase (DGAT), acyl CoA wax alcohol acyltransferase (AWAT) and monoacylglycerol acyltransferase (MGAT). Comparative biochemistry and physiology Part D, Genomics & proteomics. 2010; 5:45– 54. Irizarry RA, Bolstad BM, Collin F, Cope LM, Hobbs B, Speed TP. Summaries of Affymetrix GeneChip probe level data. Nucleic acids research. 2003; 31:e15. [PubMed: 12582260] Johnson C, Dawber R, Shuster S. Surface appearances of the eccrine sweat duct by scanning electron microscopy. The British journal of dermatology. 1970; 83:655–60. [PubMed: 5492476] Kocher O, Krieger M. Role of the adaptor protein PDZK1 in controlling the HDL receptor SR-BI. Current opinion in lipidology. 2009; 20:236–41. [PubMed: 19421056] Lang JC, Schuller DE. Differential expression of a novel serine protease homologue in squamous cell carcinoma of the head and neck. British journal of cancer. 2001; 84:237–43. [PubMed: 11161383] Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008; 9:559. [PubMed: 19114008] Li B, Tsoi LC, Swindell WR, Gudjonsson JE, Tejasvi T, Johnston A, et al. Transcriptome analysis of psoriasis in a large case-control Sample: RNA-Seq rovides insights into disease mechanisms. J Invest Dermatol. 2014; 134:1828–38. [PubMed: 24441097] Liakou, AI. Volume and Differentiation of Sebaceous Glands in Psorisis [MD]. Berlin: Charité Medical University; 2015. Miyazaki M, Man WC, Ntambi JM. Targeted disruption of stearoyl-CoA desaturase1 gene in mice causes atrophy of sebaceous and meibomian glands and depletion of wax esters in the eyelid. The Journal of nutrition. 2001; 131:2260–8. [PubMed: 11533264] Moncur JT, Park JP, Memoli VA, Mohandas TK, Kinlaw WB. The “Spot 14” gene resides on the telomeric end of the 11q13 amplicon and is expressed in lipogenic breast cancers: implications for control of tumor metabolism. Proc Natl Acad Sci U S A. 1998; 95:6989–94. [PubMed: 9618526] Nourbakhsh M, Douglas DN, Pu CH, Lewis JT, Kawahara T, Lisboa LF, et al. Arylacetamide deacetylase: a novel host factor with important roles in the lipolysis of cellular triacylglycerol stores, VLDL assembly and HCV production. Journal of hepatology. 2013; 59:336–43. [PubMed: 23542347] Parisi R, Symmons DP, Griffiths CE, Ashcroft DM, et al. Identification, Management of P. Global epidemiology of psoriasis: a systematic review of incidence and prevalence. J Invest Dermatol. 2013; 133:377–85. [PubMed: 23014338] Philpott MP, Sanders DA, Bowen J, Kealey T. Effects of interleukins, colony-stimulating factor and tumour necrosis factor on human hair follicle growth in vitro: a possible role for interleukin-1 and tumour necrosis factor-alpha in alopecia areata. The British journal of dermatology. 1996; 135:942–8. [PubMed: 8977716] Quigley DA, To MD, Perez-Losada J, Pelorosso FG, Mao JH, Nagase H, et al. Genetic architecture of mouse skin inflammation and tumour susceptibility. Nature. 2009; 458:505–8. [PubMed: 19136944] Ramirez-Zacarias JL, Castro-Munozledo F, Kuri-Harcuch W. Quantitation of adipose conversion and triglycerides by staining intracytoplasmic lipids with Oil red O. Histochemistry. 1992; 97:493–7. [PubMed: 1385366]
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 14
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Ramos-e-Silva M, Pirmez R. Disorders of hair growth and the pilosebaceous unit: facts and controversies. Clinics in dermatology. 2013; 31:759–63. [PubMed: 24160282] Rittie L, Sachs DL, Orringer JS, Voorhees JJ, Fisher GJ. Eccrine sweat glands are major contributors to reepithelialization of human wounds. Am J Pathol. 2013; 182:163–71. [PubMed: 23159944] Runne U, Kroneisen-Wiersma P. Psoriatic alopecia: acute and chronic hair loss in 47 patients with scalp psoriasis. Dermatology. 1992; 185:82–7. [PubMed: 1421635] Sampath H, Flowers MT, Liu X, Paton CM, Sullivan R, Chu K, et al. Skin-specific deletion of stearoyl-CoA desaturase-1 alters skin lipid composition and protects mice from high fat dietinduced obesity. J Biol Chem. 2009; 284:19961–73. [PubMed: 19429677] Shuster S. Psoriatic alopecia. The British journal of dermatology. 1972; 87:73–7. [PubMed: 5043203] Shuster S. Psoriatic alopecia. Arch Dermatol. 1990; 126:397. Shuster S, Johnson C. The abnormality of sweat duct function in psoriasis. The British journal of dermatology. 1969; 81:846–50. [PubMed: 5359903] Silva CY, Brown KL, Kurban AK, Mahalingam M. Psoriatic alopecia - fact or fiction? A clinicohistopathologic reappraisal. Indian journal of dermatology, venereology and leprology. 2012; 78:611–9. Song WC, Hwang WJ, Shin C, Koh KS. A new model for the morphology of the arrector pili muscle in the follicular unit based on three-dimensional reconstruction. Journal of anatomy. 2006; 208:643–8. [PubMed: 16637886] Sundberg JP, Boggess D, Sundberg BA, Eilertsen K, Parimoo S, Filippi M, et al. Asebia-2J (Scd1(ab2J)): a new allele and a model for scarring alopecia. Am J Pathol. 2000; 156:2067–75. [PubMed: 10854228] Swindell WR, Johnston A, Voorhees JJ, Elder JT, Gudjonsson JE. Dissecting the psoriasis transcriptome: inflammatory- and cytokine-driven gene expression in lesions from 163 patients. BMC genomics. 2013; 14:527. [PubMed: 23915137] Swindell WR, Sarkar MK, Stuart PE, Voorhees JJ, Elder JT, Johnston A, et al. Psoriasis drug development and GWAS interpretation through in silico analysis of transcription factor binding sites. Clin Transl Med. 2015; 4:13. [PubMed: 25883770] Swindell WR, Stuart PE, Sarkar MK, Voorhees JJ, Elder JT, Johnston A, et al. Cellular dissection of psoriasis for transcriptome analyses and the post-GWAS era. BMC medical genomics. 2014; 7:27. [PubMed: 24885462] Tontonoz P, Spiegelman BM. Fat and beyond: the diverse biology of PPARgamma. Annual review of biochemistry. 2008; 77:289–312. Turkish AR, Henneberry AL, Cromley D, Padamsee M, Oelkers P, Bazzi H, et al. Identification of two novel human acyl-CoA wax alcohol acyltransferases: members of the diacylglycerol acyltransferase 2 (DGAT2) gene superfamily. J Biol Chem. 2005; 280:14755–64. [PubMed: 15671038] van de Kerkhof PC, Chang A. Scarring alopecia and psoriasis. The British journal of dermatology. 1992; 126:524–5. [PubMed: 1610698] Werner B, Brenner FM, Boer A. Histopathologic study of scalp psoriasis: peculiar features including sebaceous gland atrophy. The American Journal of dermatopathology. 2008; 30:93–100. [PubMed: 18360109] Wertz PW, Swartzendruber DC, Madison KC, Downing DT. Composition and morphology of epidermal cyst lipids. J Invest Dermatol. 1987; 89:419–25. [PubMed: 3668284] Wilkinson DI, Karasek MA. Skin lipids of a normal and mutant (asebic) mouse strain. J Invest Dermatol. 1966; 47:449–55. [PubMed: 5924301] Wilson CL, Dean D, Lane EB, Dawber RP, Leigh IM. Keratinocyte differentiation in psoriatic scalp: morphology and expression of epithelial keratins. The British journal of dermatology. 1994; 131:191–200. [PubMed: 7522513] Wright AL, Messenger AG. Scarring alopecia in psoriasis. Acta dermato-venereologica. 1990; 70:156– 9. [PubMed: 1969203] Wu B, Potter CS, Silva KA, Liang Y, Reinholdt LG, Alley LM, et al. Mutations in sterol Oacyltransferase 1 (Soat1) result in hair interior defects in AKR/J mice. J Invest Dermatol. 2010; 130:2666–8. [PubMed: 20574437] J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 15
Author Manuscript
Xhauflaire-Uhoda E, Pierard-Franchimont C, Pierard GE. Skin capacitance mapping of psoriasis. Journal of the European Academy of Dermatology and Venereology : JEADV. 2006; 20:1261–5. [PubMed: 17062043] Zhang WC, Shyh-Chang N, Yang H, Rai A, Umashankar S, Ma S, et al. Glycine decarboxylase activity drives non-small cell lung cancer tumor-initiating cells and tumorigenesis. Cell. 2012; 148:259–72. [PubMed: 22225612]
Author Manuscript Author Manuscript Author Manuscript J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 16
Author Manuscript Author Manuscript Author Manuscript
Figure 1. Sebaceous glands are atrophic in lesional (PP) vs. non-lesional (PN) psoriatic skin
Author Manuscript
A) (Printed version) 3D reconstruction of PP and PN skin sample pair #1 (figure shown is representative of rotational views of three PP/PN pairs, which are shown in the electronic version of this manuscript). Note SG atrophy in the PP skin sample relative to its paired PN sample. SGs are colored light blue, HFs purple, hair canals dark blue, eccrine ducts magenta, and epidermis yellow. A) (Electronic version) Rotational views of 3D reconstructions of three PP/PN pairs showing sebaceous gland atrophy in each PP skin sample relative to its paired PN sample. SG are colored light blue, HFs purple, hair canals dark blue eccrine ducts magenta, and epidermis yellow. B) Quantification of SG volumes in the three pairs shown in (A). C) Distribution of hair follicles with normal (i.e. 1) vs. abnormal SG count (0 or >1 per hair follicle). Data complement the contingency Table S1; n = 6 pairs.
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 17
Author Manuscript Author Manuscript
Figure 2. Oil Red O staining is markedly reduced in SGs, but not stratum corneum, of PP vs. PN skin
Author Manuscript
A) SG staining in PP vs. PN skin. Top row: 3D reconstruction of ORO-stained PSUs (representative images of 1 pilosebaceous ‘group’ per subject are shown; 3 groups were found in each biopsy). Hair follicles are in blue, SGs are in cyan. Middle and bottom rows: aspect of ORO staining in sections corresponding to the colored lines on reconstructs (red and purple in middle and bottom, respectively). Scale bars indicate 100 μm B) Stratum corneum staining in PP vs. PN skin. Note the similar degree of stratum corneum staining between PP and PN skin (top panels, in contrast to markedly lower ORO staining in SG from the same individual, horizontal section). Very similar results were obtained in the other two individuals studied. Scale bar indicates 100 μm.
Author Manuscript J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 18
Author Manuscript Author Manuscript Author Manuscript Figure 3. Loss of visible hairs in PP vs. PN skin
Author Manuscript
A) Example of a clinical photograph of lesional and surrounding perilesional skin. Circles indicate two biopsy sites chosen in lesional skin. B) Quantitation of hairs. Visible hairs were assessed in photographs of PP (lesional) skin from 71 cases, of whom 52 also had photos of PN (non-lesional) skin biopsy sites, and in 52 NN (controls) skin biopsies. Asterisks indicate p < 0.0001 by Fisher’s exact test, with PP skin as the referent.
J Invest Dermatol. Author manuscript; available in PMC 2017 September 01.
Rittié et al.
Page 19
Author Manuscript Author Manuscript Author Manuscript
Figure 4. Expansion of the epidermis reduces the relative contribution of appendageal structures to the epithelial compartment of psoriatic skin
Quantitation of A) hair follicle length, B) appendage count for eccrine sweat glands (ESGs), hair follicles (HFs), C) epidermal volume per biopsy, and D) percentage of total epithelial volume occupied by ESGs, HFs, SGs and epidermis in 3D reconstructs of PP vs. PN skin (n = 3 pairs for all). A) box-and-whisker graphs (boxes extend from 25th to 75th percentiles, lines are medians, whiskers are min to max). B, D) Mean±SD. *: p
