Tuberous sclerosis complex (TSC) is a tumor syndrome caused by pathogenic variants in TSC1 or TSC2. Individuals with TSC develop skin tumors with second-hit somatic mutations in TSC1 or TSC2 in fibroblast-like tumor cells. These somatic mutations in facial angiofibromas, but not fibrous tumors in non-sun-exposed areas, are commonly UV signature mutations, prompting recommendations to limit UV exposure to reduce the severity of TSC skin tumors. In order to investigate whether UV light induces additional genetic alterations that promote TSC tumor pathogenesis, we performed whole genome sequencing of DNA from patient peripheral blood and fibroblast-like cells grown from TSC skin tumors. Unique dual-indexed PCR-free libraries were sequenced as pools on an Illumina NovaSeq 6000 generating 2x150 read pairs with a Picard mean coverage of ∼38X for blood samples and ∼63X for tumor samples. TSC1 and TSC2 germline mutations were identified in 39 out of 48 patients (5 TSC1, 34 TSC2 including 12 mosaic). TSC1 and TSC2 somatic mutations were found in 27 out of 45 tumor samples (2 TSC1, 25 TSC2), including copy neutral LOH, point mutations, large deletions, and intra- and interchromosomal gene fusions. The mutation rate per Mb was higher in sun-exposed (median 5.7/Mb, IQR 4.1-11.8) than non-sun-exposed locations (median 1.7/Mb, IQR 1.5-2.0). C>T mutations and a UV signature were more frequent throughout the genome in sun-exposed than other sites, but there were no recurrent mutations in known driver genes. In summary, TSC skin tumors have relatively low somatic mutation rates that are increased in sun-exposed sites; however, somatic alteration of TSC1 or TSC2, whether random or UV-induced, combined with a TSC1/TSC2 germline mutation, appears to be sufficient to drive tumorigenesis.

Heterotypic spheroids are used to study mesenchymal-epithelial interactions important for skin development and regeneration. In order to investigate the behavior and interaction of human hair follicle dermal papilla cells (DPC) and keratinocytes during in vitro three-dimensional (3D) co-culture, we labeled DPC and keratinocytes with two different fluorescent cell tracking dyes and prepared spheroids in DPC medium by hanging-drop technique. After incubation for 1 or 2 days, spheroids were live embedded in an agarose cylinder and image stacks were generated in multi-view mode from several angles using dual-sided LSFM. In presence of 100 μg/ml rat tail collagen I, cells aggregated into a single 300-500 μm spheroid, whereas multiple spheroids of various sizes formed in absence of collagen. Cells appeared randomly distributed after 1 day, but after 2 days the vast majority of DPC were located in the core with most keratinocytes distributed in the periphery, with or without collagen. LSFM can be used to study self-assembly of DPC and keratinocytes in spheroids in order to gain insights into factors regulating organogenesis during skin regeneration and hair follicle neogenesis. The sub-micrometer resolution of the LSFM microscope in x-y dimension in conjunction with multi-view imaging allows for detailed quantification of distribution of cell populations in 3D space.

The genetic disorder tuberous sclerosis complex (TSC) results in benign tumor formation due to biallelic inactivating mutations in TSC1 or TSC2 and increased mechanistic target of rapamycin complex 1 (mTORC1) signaling. TSC skin tumors often display follicular abnormalities including variable enlargement, increased numbers, and abnormal structure. In an earlier abstract we reported that mice with mesenchymal-specific deletion of Tsc2 (Tsc2cKO) displayed increased dermal thickness, increased hairshaft diameter, and earlier hair regrowth at 7 weeks of age in regions where Tsc2 was disrupted. To identify Tsc2-dependent transcriptional programs, Tsc2cKO or WT mice were crossed with mice carrying the RiboTag allele, and RNA sequencing was performed on ribosome-associated transcripts in primary neonatal dermal fibroblasts isolated from these mice (Tsc2KO or WT cells) treated with or without 20nM rapamycin for 24hr, an inhibitor of mTORC1 used in the treatment of TSC. Gene enrichment analysis indicated that genes down-regulated in Tsc2KO cells were enriched for those involved in the response to transforming growth factor beta (TGFβ), a known regulator of both catagen and mesenchymal cell differentiation. Treatment with rapamycin normalized these transcripts. Biological responses to TGFβ1 in Tsc2-deficient fibroblasts were subsequently analyzed. Tsc2KO cells showed reduced TGFβ1-stimulated collagen gel contraction compared to control cells (P=0.026). Using a luciferase reporter, TGFβ1-induced Smad activity was reduced by about half in Tsc2KO cells. In summary, our results suggest defects in hair follicles observed in mice with mesenchymal Tsc2 disruption may be the result of an impaired TGFβ1 response. Due to the critical role of TGFβ in cell fate determination, overcoming reduced TGFβ may also be a novel treatment approach in TSC.

Mosaicism has been suggested to explain tuberous sclerosis complex (TSC) manifesting as unilateral angiofibromas or as mild disease and no mutation identified using standard genetic testing. To identify mosaicism and its phenotypic spectrum in TSC, we performed next generation sequencing (NGS) on DNA isolated from TSC skin tumors, including angiofibromas, ungual fibromas, shagreen patch, oral fibroma, and fibrous cephalic plaque. Mosaicism was identified in 17/30 patients, first by identification of mutations in TSC2 in DNA from 28 skin tumors from the 17 patients, of which 16 had two mutations consistent with Knudson's two-hit hypothesis. The mutant allele fractions in whole tumors or cultured tumor cells ranged from 1 to 54%. A shared mutation in TSC2 was identified in every case where more than one tumor was studied from a single patient. The median mutant allele fraction was 5% in blood samples (range 0-19, n=11) and 1.4% in other control tissues (range 0-13, n=13), well below the 50% allelic fraction expected for germline mutations. Examination of these mosaic adult women with TSC revealed that 4 had unilateral or asymmetric angiofibromas, 3 had fewer than 20 angiofibromas, and 10 had numerous bilateral facial angiofibromas. Overall these results show that: 1) analysis of DNA from whole skin tumors or cultured tumor cells enables identification of low-level mosaicism using NGS, and 2) the phenotype of mosaic TSC patients varies from those with unilateral angiofibromas and mild disease to those with a disseminated phenotype that may be indistinguishable from those with germline mutations.

Tuberous sclerosis complex (TSC) is an autosomal genetic disorder in which tumors form in multiple organs due to biallelic inactivating mutations in TSC1 or TSC2 and increased mTORC1 signaling. In order to discover biomarkers for TSC, we studied our mouse model of TSC with conditional deletion of Tsc2 in dermal mesenchyme (Tsc2cKO), testing the potential clinical relevance of findings using human TSC skin tumors. RNA-sequencing analysis of cultured Tsc2-null and wild-type mouse dermal fibroblasts identified galectin-3 as overexpressed by the mutant cells. Levels of both intracellular and secreted galectin-3 protein were approximately 3-fold greater in Tsc2-null fibroblasts than wild-type fibroblasts. Tsc2-null cells incubated in the presence of 20 nM sirolimus, an mTORC1-inhibitor, showed decreased galectin-3 expression but not to the level of wild-type fibroblasts, suggesting that overexpression of galectin-3 is only partially mTORC1-dependent. Serum levels of galectin-3 were 68% greater (p=0.0015) in Tsc2 cKO than wild-type mice and sirolimus treatment for 4 weeks decreased galectin-3 serum levels 25% (p=0.036) in the Tsc2cKO. Fibroblast-like cells grown from human TSC skin tumors, including angiofibromas, fibrous cephalic plaque, and periungual fibromas, showed greater intracellular and secreted galectin-3 levels than patient normal fibroblasts. Immunohistochemical analysis showed abundant galectin-3 positive dermal cells in TSC skin tumors but nearly absent dermal staining in paired samples of patient normal-appearing skin. Our results demonstrate that loss of TSC2 in dermal fibroblasts is associated with increased galectin-3 and suggest that this molecule should be investigated as a biomarker for TSC and possibly other diseases associated with increased mTORC1 activation.

dermal–epidermal composites dermal papilla hair follicle neonatal foreskin keratinocytes TO THE EDITOR Tissue-engineered skin substitutes are used in the clinic to treat chronic wounds and burns and in the laboratory to advance our understanding of wound healing, skin biology, and skin disease. One type of skin substitute, dermal–epidermal composites (DECs), also known as skin equivalents or bilayered living skin constructs, comprises dermal fibroblasts embedded in a matrix such as collagen and overlaid with keratinocytes (Veves et al., 2001Veves A. Falanga V. Armstrong D.G. et al.Graftskin, a human skin equivalent, is effective in the management of noninfected neuropathic diabetic foot ulcers: a prospective randomized multicenter clinical trial.Diabetes Care. 2001; 24: 290-295Crossref PubMed Scopus (587) Google Scholar). DECs promote wound healing (Falanga and Sabolinski, 1999Falanga V. Sabolinski M. A bilayered living skin construct (APLIGRAF) accelerates complete closure of hard-to-heal venous ulcers.Wound Repair Regen. 1999; 7: 201-207Crossref PubMed Scopus (316) Google Scholar) and have been used to model skin development and diseases (Carretero et al., 2011Carretero M. Guerrero-Aspizua S. Del Rio M. Applicability of bioengineered human skin: from preclinical skin humanized mouse models to clinical regenerative therapies.Bioeng Bugs. 2011; 2: 203-207Crossref PubMed Scopus (11) Google Scholar; Kamsteeg et al., 2011Kamsteeg M. Bergers M. Boer R. et al.Type 2 helper T-cell cytokines induce morphologic and molecular characteristics of atopic dermatitis in human skin equivalent.Am J Pathol. 2011; 178: 2091-2099Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar), but their use has been limited by the inability of the skin constructs to regenerate hair follicles (HFs). During embryogenesis, mesenchymal cells signal the overlying epithelium to induce HF formation, and, in adults, a specialized group of mesenchymal cells, the dermal papilla (DP) cells, have been shown to retain the capacity to induce HF regeneration (Hardy, 1992Hardy M.H. The secret life of the hair follicle.Trends Genet. 1992; 8: 55-61Abstract Full Text PDF PubMed Scopus (806) Google Scholar; Reddy et al., 2001Reddy S. Andl T. Bagasra A. et al.Characterization of Wnt gene expression in developing and postnatal hair follicles and identification of Wnt5a as a target of Sonic hedgehog in hair follicle morphogenesis.Mech Dev. 2001; 107: 69-82Crossref PubMed Scopus (373) Google Scholar; Gharzi et al., 2003Gharzi A. Reynolds A.J. Jahoda C.A. Plasticity of hair follicle dermal cells in wound healing and induction.Exp Dermatol. 2003; 12: 126-136Crossref PubMed Scopus (113) Google Scholar). DP cells from rodents induce HFs in a variety of assays (reviewed in Ohyama et al., 2010Ohyama M. Zheng Y. Paus R. et al.The mesenchymal component of hair follicle neogenesis: background, methods and molecular characterization.Exp Dermatol. 2010; 19: 89-99Crossref PubMed Scopus (123) Google Scholar), but it has been difficult to grow human DP cells that maintain inductive capacity in culture (Higgins et al., 2010Higgins C.A. Richardson G.D. Ferdinando D. et al.Modelling the hair follicle dermal papilla using spheroid cell cultures.Exp Dermatol. 2010; 19: 546-548Crossref PubMed Scopus (89) Google Scholar). Recent technological advances have enabled the use of human cells to form chimeric HFs, for example, by combining human keratinocytes and rodent mesenchymal cells in chamber assays (Ehama et al., 2007Ehama R. Ishimatsu-Tsuji Y. Iriyama S. et al.Hair follicle regeneration using grafted rodent and human cells.J Invest Dermatol. 2007; 127: 2106-2115Abstract Full Text Full Text PDF PubMed Scopus (91) Google Scholar), human scalp dermal papilla cells and mouse epidermal keratinocytes in flap grafts (Qiao et al., 2009Qiao J. Zawadzka A. Philips E. et al.Hair follicle neogenesis induced by cultured human scalp dermal papilla cells.Regen Med. 2009; 4: 667-676Crossref PubMed Scopus (52) Google Scholar), or by injecting human DP cells, grown as spheroids, together with mouse epidermal cells in reconstitution or “patch” assays (Kang et al., 2012Kang B.M. Kwack M.H. Kim M.K. et al.Sphere formation increases the ability of cultured human dermal papilla cells to induce hair follicles from mouse epidermal cells in a reconstitution assay.J Invest Dermatol. 2012; 132: 237-239Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar). However, to date, complete and entirely human HFs formed from normal cultured cells have not been reported. Recently, the potential for human HF development in grafted DECs was demonstrated using composites containing human neonatal foreskin keratinocytes (NFKs) and fibroblast-like cells derived from tuberous sclerosis skin hamartomas (Li et al., 2011Li S. Thangapazham R.L. Wang J.A. et al.Human TSC2-null fibroblast-like cells induce hair follicle neogenesis and hamartoma morphogenesis.Nat Commun. 2011; 2: 235Crossref PubMed Scopus (35) Google Scholar). Therefore, we used the conditions developed in these experiments to test for HF formation in DECs using normal human DP cells. Human DP cells isolated from temporal scalp dermis (Promocell, Heidelberg, Germany) from six donors were propagated in vitro according to the manufacturer’s recommendations. Alkaline phosphatase activity, a DP marker that correlates with hair-inducing capacity (Ohyama et al., 2010Ohyama M. Zheng Y. Paus R. et al.The mesenchymal component of hair follicle neogenesis: background, methods and molecular characterization.Exp Dermatol. 2010; 19: 89-99Crossref PubMed Scopus (123) Google Scholar), was measured in vitro using the BCIP/NBT substrate (Sigma-Aldrich, St Louis, MO) on passage 5 DP cells. Alkaline phosphatase activity was variable between samples, with cells from three of the donors showing alkaline phosphatase activity in more than 50% of the cells (Table 1). DECs were constructed by combining DP cells with rat tail collagen type 1, adding NFKs on top and bringing the constructs to the air–liquid interface for 2 days before grafting onto female nude mice (Supplementary Figure S1 online and Supplementary Methods online). Eight weeks after grafting, HFs were observed in mice grafted with the three human DP cells with higher alkaline phosphatase activity (Table 1; Figure 1a). HFs had a bulb, dermal sheath, hair matrix, and cortex (Figure 1b). Epithelial compartments of the HFs were intact with concentric layers of inner and outer root sheaths, sebaceous glands, and hair shaft (Figure 1c–e). Fluorescence in situ hybridization showed the hybridization of a human-specific Alu probe (green) to the nuclei of both epithelial and dermal cells within the graft, including dermal sheath and dermal papilla, confirming their human origin (Figure 1f and g). An antibody reactive with human but not mouse COX IV stained follicular epithelium and dermal papilla/dermal sheath of grafts (Figure 1h). Fluorescence in situ hybridization showed the hybridization of a human-specific, pan-centromeric probe (green) to the nuclei of both epithelial and dermal components (Figure 1i), whereas a human-specific Y-chromosome probe (red) hybridized to nuclei in the epidermis and the follicular epithelium (Figure 1j), consistent with the origin of dermal and epidermal cells from female and male donors, respectively. HFs also stained for markers of specific compartments of a fully developed human HF. Cells in the region of the DP and lower DS displayed alkaline phosphatase activity (Figure 1k), normal reactivity with specific antibodies to human nestin (Figure 1l) and versican (Figure 1m). As expected, anagen HFs had more concentrated immunoreactivity to Ki-67 in the region of the hair matrix relative to the overlying epidermis (Figure 1n). The companion layer as identified by keratin 75 staining was present between the inner and the outer root sheaths (Figure 1o). The basal layer of the outer root sheath was immunoreactive for keratin 15, a marker of HF stem cells located in the bulge region (Figure 1p).Table 1Human dermal papilla cells with higher alkaline phosphatase activity form hair follicles when combined with neonatal foreskin keratinocytes in grafted dermal–epidermal constructsDermal papilla cellsHDP47HDP44HDP41HDP43HDP60HDP52Percentage of cells positive for alkaline phosphatase activity (mean±SD, n=6)75±367±352±435±129±417±4Number of grafted constructs with hair follicles/total number of grafts9/116/64/50/60/40/5 Open table in a new tab Download .pdf (.71 MB) Help with pdf files Supplementary Methods In summary, we report that cultured specialized human cells such as DP cells can induce complete pilosebaceous units in vivo in the grafted DEC model. Human HF formation may have been enabled by particular features of our experimental methods, such as the use of DP cells from the temporal scalp, use of an occlusive dressing for a long period after grafting, and a long duration for maturation of the grafts. It is not yet known whether these or other factors are critical to enable human HF formation in DECs, but success using the conditions described appears to require a starting population of DP cells in which the majority show alkaline phosphatase activity. This model could be used to evaluate the trichogenicity of various types of dermal cells in combination with different keratinocyte populations, evaluate hair loss therapies and may be adaptable to examine the regeneration of other skin appendages and the formation of skin adnexal neoplasms. Next-generation skin substitutes that promote HF neogenesis are expected to promote healing, normal skin function, and appearance and can be used to study of human HF neogenesis and regeneration with cultured adult cells. This work was supported by a grant to Thomas Darling, M.D., Ph.D., from the Defense Medical Research and Development Program and the University of Pennsylvania Skin Disease Research Center through grant 5-P30-AR-057217-03. Supplementary material is linked to the online version of the paper at http://www.nature.com/jid