Background
In humans, sebaceous glands associated with hair follicles are distributed throughout all the skin and found in greatest abundance on the face and scalp and are absent from the palms and soles [
1]. Sebaceous glands can also form independently from the hair follicle and form specialized glands such as Meibomian glands of the eyelid, ectopic sebaceous gland of the glans penis [
2] and Fordyce’s spots of the oral epithelium. Sebaceous glands are microscopic glands which secrete an oily substance (sebum) in the hair follicles to lubricate the skin and hair of animals [
3]. Their function within the epidermis is to prevent the skin from dehydration and protect the body against infections and physical, chemical and thermal assaults of the environment. The main components of human sebum are triglycerides and fatty acids (57.5%), wax esters (26%), and squalene (12%) [
4]. The production of sebum is regulated throughout life, and decreases dramatically with age [
5]. This is associated with increased dryness and fragility of the skin. Moreover, several human diseases, such as acne vulgaris, atopic dermatitis, seborrheic dermatitis and primary cicatricial alopecia are thought to be associated with deregulation of the sebaceous glands [
4,
6,
7].
There is a crucial interdependency of sebaceous glands with hair follicles and epidermis as sebocyte dysfunction results in degeneration of hair follicle structures and a defective skin barrier [
7,
8]. This is illustrated in the
asebia mutant mouse, which lacks the SCD1 enzyme that desaturates fatty acids. This mutant displays rudimentary sebaceous glands and alteration in the profile of skin surface lipids leading to chronic inflammatory reactions, alopecia and dermal scarring [
8].
Successful growth of primary human cells often constitutes a breakthrough in a specific area of human biology with important clinical implications. Tissue stem cells such as those of the blood and the epidermis have already been successfully used in clinics for decades [
9,
10]. In particular, epidermal cells (keratinocytes) can be cultured in vitro and can be efficiently manipulated to form a three dimensional epidermis [
11,
12]. Despite these advancements, the successful methods for culturing human primary sebocytes without the use of mouse feeder layers are not established. Selective cultivation of human sebocytes has been attempted in the past using mitomycin-treated 3T3 feeder layers by covering the microdissected sebaceous gland explant with glass slides but primary sebocytes survived only two passages after which they underwent differentiation [
13]. Human sebaceous gland cell lines have been established in the past from adult human facial skin and periauricular area [
14‐
17], but their immortalization with Simian virus-40 large T antigen or HPV16/E6E7 genes, which bypass the p53 and retinoblastoma protein mediated restriction point, results in cellular transformation that has limited their use for analyzing their cell cycle and differentiation regulation. Here, we culture human primary sebocytes using a novel method, which can in the future, be incorporated into skin reconstructs and provide a basis for understanding the molecular pathways which regulate human sebaceous gland biology.
A potential candidate for human sebocyte regulation suggested by several lines of evidence is Transforming Growth Factor β (TGFβ) [
18,
19] but the lack of primary human cultures has impaired an in-depth investigation of the molecular mechanism whereby TGF β signaling controls sebaceous gland differentiation. The TGF β pathway is ubiquitous and involved in the control of growth and differentiation of multiple cell and tissue types. The two major receptors of the TGFβ signaling pathway, TGFβ Receptor I (TGFβ RI) and TGFβ Receptor II (TGFβ RII), are expressed in mouse sebaceous glands [
20,
21]. In human and mouse epithelial cell lines, TGFβ acts as a potent inhibitor of proliferation mediated at least in part via down-regulation of c-Myc expression [
22,
23]. Intriguingly, c-Myc overexpression in a mouse model induces an increase in sebaceous gland size due to activation of sebocyte differentiation at the expense of hair differentiation [
16,
24]. Moreover, disruption of epidermal Smad4, the common mediator of TGFβ signaling, leads to hyperplasia of inter-follicular epidermis, hair follicle, and sebaceous glands through c-Myc upregulation [
25].
To determine the effect of TGFβ signaling on sebocyte differentiation, we investigated the effect of TGFβ ligands on the primary human sebocytes we established using a novel culture system and skin samples from pediatric donors.
Discussion
Here we have developed a novel method of culturing human sebocytes without transformation and using a feeder layer-free culture system to examine the role of the TGFβ pathway in the control of differentiation. Primary sebaceous gland cells do not express Keratin 8 in contrast to previously immortalized sebocytes. Keratin 8 is not normally expressed in normal sebaceous gland in vivo [
29] and our results indicate that the transformation process in the immortalized line has likely altered the expression of several fundamental cell markers. Moreover, we showed different responsiveness to linoleic acid and TGFβ1 treatment between the primary sebocytes and the immortalized cells (data not shown) suggesting that the cellular properties of those cells substantially differ.
Through our analysis, we have identified that certain markers of sebocytes are differentially expressed depending upon the location on the body (scalp, chest, face), and localization within the sebaceous gland. These results highlight the need for studies covering a range of patient ages to fully comprehend the regulation of the sebaceous glands. However, our work shows that the effect of TGFβ1 activation on sebocyte differentiation is similar in sebocytes derived from three areas (scalp, breast and face) suggesting the specificity of that effect is independent of location. Previous reports have largely focused on cells and glands derived from older adults and post-menopausal women [
14‐
16]. While we have not identified differences in sex, the age of the individual from which the sebaceous gland is derived seems to be of significance. It is known that the sebaceous glands undergo dramatic changes over the course of one’s lifespan, with high sebum production occurring in infancy, a reduction during early childhood, followed by a steady increase through puberty into early adulthood. Using pediatric donors we ensured that the skin is not exposed to the hormonal changes that adult or old donor skin goes through. In the future it may be interesting to use our novel method to isolate sebocytes from old donors to examine the effect of age on TGFβ responsiveness in sebocytes.
We have begun to unravel one mechanism of differentiation of human sebaceous glands that culminates in sebum production. Our data suggest that TGFβ signaling maintains sebocytes in an undifferentiated state by decreasing the expression of
FADS2 and
PPARγ thereby decreasing lipid accumulation through the TGFβ RII-Smad2 dependent pathway. The successful growth of these primary human sebocytes has important clinical application such as the possibility of designing new strategies of culturing engineered skin to enable and maintain the presence of sebaceous glands in skin grafts for burn victims [
10,
39]. In addition to cell autonomous regulators and signals inducing proliferation and maturation among sebaceous cells, the complex microenvironment surrounding the sebaceous gland might have a profound effect on homeostasis of the tissue. Molecular crosstalk between the dermis and the epithelial cells is crucial for the initiation and maintenance of the hair follicles [
40]. It seems most likely that similar mechanisms of communication between sebocytes and the surrounding dermal tissue exist. For instance, in the mouse, TGFβ1 is known to be released by the inner root sheath of the hair follicle, thereby providing a means for a bidirectional interaction between the sebaceous gland and the hair follicle epithelium [
20]. Similarly, in the dermis, human fibroblasts secrete TGFβ [
41,
42] which may then act on keratinocytes and sebocytes. Another component in the microenvironment that could also be part of this crosstalk are the arrector pili muscle cells recently shown to be controlled by bulge stem cells in mouse [
43]. Being located in close proximity to the sebaceous gland, arrector pili muscles could help release sebum onto the skin surface [
44].
Impairment of the skin barrier due to the deregulation of sebum production when associated with bacteria colonization and inflammation, can be the cause of serious skin conditions in people. For instance, hyperseborrhea combined with the presence of
Propionibacterium acnes and inflammation can lead to acne vulgaris [
4] and
Staphylococcus aureus can aggravate atopic dermatitis [
6]. Sebocytes can produce antimicrobial peptides such as defensin-1 and −2 upon exposure to
Propionibacterium acnes or lipopolysaccharides [
45,
46] to prevent from bacteria colonization [
47] and from an upregulation of sebum production [
48]. Studies have revealed that TGFβ induces the expression of human β-defensin-2 in endothelial cells [
49] and influences inflammatory response [
50]. Therefore it will be interesting to further investigate the impact of TGFβ on immune responses in sebaceous gland and its implication in antimicrobial peptides secretion by sebocytes. With the novel isolation strategy we described here, different interactions with the microenvironment can now be investigated.
Methods
Cell Culture
The sebaceous gland populations were generated from human scalp (SSG3), face, chest and breast from both male and female donors. The skin samples were collected as a surgical waste with information provided regarding the age and sex of the donors with Institutional Review Board (IRB) approval at Cincinnati Children’s Hospital Medical Center. Cincinnati Children’s Hospital is a Pediatric Hospital that allowed us to collect samples from donors ranging 9 months old to 12 years old. The IRB determined that the research does not meet the regulatory criteria for research involving human subjects as there were no interaction with the donors and no identifiable private information. After treating the skin with dispase overnight at 4°C, intact sebaceous glands were isolated with microsurgical instruments under a dissecting microscope (Figure
1b). To mimic the microenvironment of the sebaceous gland, the explants were sandwiched between glass coverslips coated with human fibronectin (10 μg/ml, Millipore, Billerica, MA) (Figure
1d). The explants were cultivated in sebocyte medium as described [
15] (DMEM/Ham’s F-12 (3:1), Epidermal Growth Factor (EGF 3 ng/ml, Austral Biologicals, San Ramon, CA), cholera toxin (1.2×10-10M, Sigma, St. Louis, MO), adenine (24 μg/ml, Sigma, St. Louis, MO), insulin (10 ng/ml Sigma), hydrocortisone (45.2 ng/ml, Sigma), FBS (2.5% Hyclone, San Jose, CA), antibiotic/antimycotic (100×, Invitrogen, Grand Island, NY). After 1–2 weeks of growth in culture, cellular outgrowth became apparent from the periphery of the gland lobules. The explants were removed and the isolated cells cultured on the fibronectin-coated coverslips.
Western blotting
Proteins were separated by electrophoresis on 8-10% acrylamide gels, transferred to nitrocellulose membranes and subjected to immunoblotting. Membranes were blocked for one hour with 5% non-fat milk or 5% BSA in PBS containing 0.1% Tween-20. Primary antibodies were used at concentrations described below and HRP-coupled secondary antibodies were used at 1/2,000 in 5% non-fat milk. Immunoblots were developed using standard ECL (Amersham, Pittsburg, PA) and Luminata TM crescendo and classico (Millipore, Billerica, MA). Two-color immunoblot detection was performed using LI-COR Odyssey CL× (Biosciences, Lincoln, NE). Membranes were blocked in Odyssey blocking buffer (Li-Cor) and secondary antibodies conjugated to IRDye 680LT and 800CW were used (1/10,000; Li-Cor). Protein levels were quantified using the Odyssey Infrared Imaging System (Li-Cor).
Retroviral Infection
To ablate TGFβ RII in SSG3 cells, we used shRNA vectors from the CCHMC Heart Institute lenti-shRNA library core (shRNA TGFβ RII #197031 and 194992 and a shRNA control). The human library was purchased from Sigma-Aldrich (MISSION shRNA). Lentivirus was produced by the Viral Vector Core at the Translational Core Laboratories, Cincinnati Children’s Hospital Research Foundation. Cells were grown to 80% confluency in 6-well plates before being infected with the lentivirus for 48 h. Infected cells were selected with 1 μg/ml puromycin (Sigma, St. Louis, MO) for 48 h. Following selection, TGFβ RII knock down cells were grown in regular media for 48 h before being activated with 5 ng/ml TGFβ1 for 24 h.
Histology and Immunofluorescence
Human tissues were frozen unfixed in OCT compound (Tissue-Tek, Sakura, Torrance, CA) for cryosectioning. Immunostainings were performed as previously described [
51].
Antibodies
Primary antibodies against the following proteins were used at the dilution indicated: PPARγ (Santa-Cruz Biotechnology Inc., Santa Cruz, CA, H-100 1/250 for immunofluorescence, 1/500 for immunoblot), Blimp1 (Cell Signaling Technology, Beverly, MA, 1/500 for immunofluorescence, 1/1,000 for immunoblot), Fibronectin (Santa-Cruz Biotechnology Inc., Santa Cruz, CA, EP5 1/150), Muc1 (Millipore, Billerica, MA 1/500), cMyc (Cell Signaling Technology, Beverly, MA, 1/800 for immunofluorescence, 1/1,000 for immunoblot), TGFβ RII (Santa-Cruz Biotechnology Inc., Santa Cruz, CA, sc-220 1/1,000), p-Smad2 (Cell Signaling Technology, Beverly, MA, 1/100 for immunofluorescence, 1/1,000 for immunoblot), Smad2/3 (BD Biosciences, San Jose, CA, 1/500), α6 integrin (CD49f, BD Biosciences, San Jose, CA, 1/100), Keratin 8 (this antibody, developed by Dr. Brulet and Dr. Kemler, was obtained from the NICHD Developmental Studies Hybridoma Bank maintained by the University of Iowa, 1/1,000), β-actin (Sigma, St. Louis, MO, 1/2,000), Keratin 7 (Cell Signaling Technology, Beverly, MA, 1/1,000), 4′,6-diamidino-2-phenylindole (DAPI) was utilized as a marker of cell nuclei (Sigma Chemical Co., St. Louis, MO, 1/5,000). Secondary antibodies Alexa Fluor 488 or 555 (Molecular Probes, Grand Island, NY) were used at a dilution of 1/1,000. Fluorescence images were acquired with a fluorescent microscope AxioImager M1 (Zeiss, Oberkochen, Germany) and pictures were taken with an axioCam MRm camera (Zeiss, Oberkochen, Germany).
Real-time PCR
Total RNA was isolated using a Qiagen Rneasy Mini Kit and used to produce cDNA (Maxima first strand cDNA synthesis kit, Fermentas, San Jose, CA). Reverse transcription (RT) reactions were diluted to 10 ng/μl and 1μl of each RT was used for real-time PCR. Real-time PCR was performed using the CFX96 real-time PCR System, CFX Manager Software and the SsoFast EvaGreen Supermix reagents (Biorad, Hercules, CA). All reactions were run in triplicate and analyzed using the ΔΔCT method with relative expression normalized to GAPDH. Primers used:
GAPDH-F: ACATCGCTCAGACACCATG, GAPDH-R: TGTAGTTGAGGTCAATGAAGGG
PPARγ-F: GAGCCCAAGTTTGAGTTTGC, PPARγ-R: GCAGGTTGTCTTGAATGTCTTC,
FADS2-F: TGTCTACAGAAAACCCAAGTGG, FADS2-R: TGTGGAAGATGTTAGGCTTGG,
TGFβ RII-F: CTGTGGATGACCTGGCTAAC, TGFβ RII-R: CATTTCCCAGAGCACCAGAG
Lipogenesis assays
For Nile red staining, cells or OCT sections were fixed 10 minutes at room temperature in 4% formaldehyde. After 3 washes in 1XPBS, Nile red staining was performed with 0.1 μg/ml of Nile red (Sigma, St. Louis, MO) in 0.15 M NaCl for 15 minutes at room temperature. For Oil red O staining, cells were fixed 15 minutes in 10% formalin, wash with water for 10 minutes and 60% isopropanol before being stained with Oil red O (0.7% in 60% isopropanol) for 45 minutes. Cells were rinsed with 60% isopropanol and the nuclei stained with haematoxylin. To trigger differentiation of sebocytes in vitro, 0.1 mM linoleic acid (Sigma, St Louis, MO) was added directly to sebocyte media. To prepare cells for extraction of lipids, 2-3×107 of cells were pelleted, washed with 1XPBS and lipids were preserved in the dark at −80°C under argon until analysis. The qualitative and quantitative composition of lipids in scalp-derived human sebocytes was determined using an Agilent 5973N Gas chromatograph/Mass spectrometer with a SPE cartridge (solid phase extraction) and was performed by Synelvia S.A.S (Labege, France).
Nile Red analysis by FACS
Cells were cultured in 6-well plates at 80% confluence and infected with the lentivirus expressing the shRNAs as previously described. After puromycin selection for 48 h, cells were washed in 1X PBS and treated with working medium with or without Linoleic acid (0.1 mM) for 24 h. The cells were trypsinized, washed once with 1X PBS and neutral lipids were labeled with the fluorescent dye Nile red (1 μg/ml in PBS). 10,000 cells per sample were analyzed using a FACS Canto I equipped with a blue laser (488 nm excitation).
Electron microscopy
Cells were grown at 80% confluency in sebocyte media and rinsed once with 0.175 M sodium cacodylate buffer. Cells were fixed in 3% glutaraldehyde/0.175 M cacodylate buffer (Electron Microscopy Sciences, Hatfield, PA) for 1 hour at 4°C. Dishes were washed twice with 0.175 M sodium cacodylate buffer. Cells were post fixed in 1% osmium tetroxide/cacodylate buffer for 1 hour at 4°C before being washed three times with 0.175 M sodium cacodylate buffer. After the final wash with 1.5 ml, cells were scraped and centrifuged for 5 min at 10,000 RPM. The cell pellet was then resuspended in 1 ml 1% agarose (Type IX ultra-low gelling tempt, Sigma) overnight at 4°C. The samples were then processed through a graded series of alcohols, infiltrated and embedded in LX-112 resin. After polymerization at 60°C for three days, ultrathin sections (100 nm) were cut using a Reichert-Jung Ultracut E microtome and counterstained in 2% aqueous uranyl acetate and Reynolds lead citrate. Images were taken with a transmission electron microscope (Hitachi H-6750) equipped with a digital camera (AMT 2k×2K tem CCD).
Statistics
Data are expressed as means +/− SD. Comparison between two cell types was performed using unpaired two-tailed student’s t test. Paired two-tailed student’s t test was used when we compared the effect of a treatment on the same cell type. p<0.05 was considered significant.
Competing interests
GG has no conflict of interest. JD, CB and LB are employed by L’Oreal and have conflict of interests.