DOI: 10.1111/exd.12700 www.wileyonlinelibrary.com/journal/EXD
Patterns of Expression
Mapping the expression of epithelial hair follicle stem cell-related transcription factors LHX2 and SOX9 in the human hair follicle Talveen S. Purba1, Iain S. Haslam1, Asim Shahmalak2, Ranjit K. Bhogal3 and Ralf Paus1,4 1 Centre for Dermatology Research, Institute of Inflammation and Repair, University of Manchester, Manchester, UK; 2Crown Clinic, Manchester, UK; 3Unilever R&D Colworth, Bedfordshire MK44 1LQ, UK; 4Department of Dermatology, University of M€ unster, M€ unster, Germany Correspondence: Ralf Paus, Centre for Dermatology Research, Institute of Inflammation & Repair, University of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT, UK, e-mail: [email protected]
Abstract: In the murine hair follicle (HF), the transcription factors LHX2 and SOX9 are implicated in epithelial hair follicle stem cell (eHFSC) self-renewal and the maintenance of eHFSC niche characteristics. However, the exact expression patterns of LHX2 and SOX9 in the human HF are unclear. Therefore, we have quantitatively mapped the localisation of known human eHFSC markers keratin 15 (K15) and keratin 19 (K19) in the outer root sheath (ORS) of male occipital scalp anagen HFs and related this to the localisation of LHX2 and SOX9 protein expression. As expected, K15+ and K19+ cells represented two distinct progenitor cell populations in the bulge and in the proximal bulb ORS (pbORS). Interestingly, cell fluorescence for K19 was significantly stronger within the pbORS versus the bulge, and vice versa for K15, describing a hitherto unrecognised differential expression pattern. LHX2 and SOX9 expressing cells were distributed throughout the ORS, including the bulge, but were not restricted to it. SOX9 expression was most prominent
in the ORS immediately below the human bulge, whereas LHX2+ cells were similarly distributed between the sub-bulge and pbORS, that is compartments not enriched with quiescent eHFSCs. During catagen development, the intensity of LHX2 and SOX9 protein expression increased in the proximal HF epithelium. Double immunostaining showed that the majority of SOX9+ cells in the human anagen HF epithelium did not co-express K15, K19 or LHX2. This expression profile suggests that LHX2 and SOX9 highlight distinct epithelial progenitor cell populations, in addition to K15+ or K19+ cells, that could play an important role in the maintenance of the human HF epithelium.
Introduction
distribution and functional relevance for HF biology (14–18), our understanding of human HF stem cell markers remains rudimentary. For example, eHFSC transcription factor SRY (sex determining region Y)-box 9 (SOX9), which is downstream of SHH signalling (19,20), has recently been shown to regulate activin/pSMAD2 signalling (16) and is known to be required in eHFSC maintenance and ORS differentiation in the murine hair follicle (16,19). SOX9 is also expressed within progenitor cells of multiple tissues including the murine intestinal and lung epithelium (21,22), human fetal pancreas (23) and the basal layer of human epidermis (24). Its multifaceted roles as a transcription factor, whose biological effects depend on its partnering with other transcription factors (25), involve the regulation of somatic progenitor cell proliferation and their respective differentiation programmes (21–23). Moreover, SOX9 has been implicated in the regulation of diverse processes such as chondrogenesis (26) and mammalian sex determination (27). Another transcription factor known to be important in murine epithelial HF stem/progenitor cell biology is Lim homeobox 2 (LHX2) (17,28,29), which regulates SOX9 (30). It has recently been shown that LHX2 maintains the organisation of the murine eHFSC niche and regulates known eHFSC genes as well as cytoskeletal and cell adhesion-related proteins, which could help mediate asymmetric self-renewal of eHFSCs (17). Beyond the HF system, LHX2 has also been implicated in other, namely developmental (31), contexts such as the development of the eye (32),
The human hair follicle (HF) is a veritable stem cell repository and uniquely suited to study various human progenitor cell populations, namely epithelial HF stem cells (eHFSCs), within their natural habitat (1–4). eHFSCs reside in the bulge region of the HF’s outer root sheath (ORS) and allow remodelling and tissue renewal during HF cycling through growth (anagen), regression (catagen) and relative rest (telogen). Murine studies have shown that eHFSCs are controlled by locally derived micro-environmental regulatory signalling (5–8). Cells expressing known markers of human epithelial stem and progenitor cells are found throughout the ORS of the HF isthmus and suprabulbar regions (4), which can be further subdivided into the bulge, sub-bulge and proximal bulb ORS (pbORS) regions. The epithelial stem/progenitor cell markers keratin 15 (K15) and keratin 19 (K19) localise to the bulge and pbORS of the human anagen HF (9–11). Other useful epithelial stem/progenitor markers including CD200, PHLDA1 and negative expression for the gap junction protein connexin 43 (CX43) demarcate the bulge, whereas CD34 and p75NTR mark the suprabulbar ORS area (4,9– 13). This differential expression of eHFSC markers may represent multiple functionally distinct cell populations, where bulge cells are thought to be relatively quiescent, as opposed to the relatively activated and fate-restricted progenitor cells in the sub-bulge and pbORS (4). However, while the currently available markers for murine eHFSCs are becoming increasingly well defined, both in their
462
Key words: eHFSCs – epithelial human hair follicle stem cells – K15 – K19 – LHX2 – SOX9
Accepted for publication 17 March 2015
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
LHX2 and SOX9 expression in human hair follicles
lineage specification and maintenance of haematopoietic progenitor cells (33) and neuronal patterning and specification (34,35). However, the distribution of LHX2+ or SOX9+ cells within the human HF remains relatively uncharted territory. The exact intrafollicular location and expression characteristics of epithelial progenitor cell markers in the human HF are an important prerequisite for understanding their functional role in human HF biology. Thus, we have mapped the localisation of cells expressing LHX2 and SOX9 in the human HF ORS relative to the established markers K15 and K19 by quantitative immunohistomorphometry on occipital scalp human HFs.
Methods Tissue preparation and immunofluorescence Occipital scalp tissue was obtained from male patients undergoing transplantation surgery after informed consent under ethical and institutional approval from the University of Manchester in accordance with Human Tissue Act guidelines. HFs were isolated and embedded in optimal cutting temperature compound (OCT) and frozen in liquid nitrogen. HF cryosections (at a thickness of 7 lM) were prepared for immunofluorescence protocols. Tissue sections were fixed for 10 min in ice-cold ( 20°C) acetone and treated overnight with primary antibody diluted (see optimised dilutions below) in phosphate-buffered saline (PBS) with either 10% normal goat serum (NGS) or normal donkey serum (NDS) at 4°C in a humid chamber. Slides were treated with secondary antibody for 45 min diluted in 1:200 in PBS + 10% NGS or 10% NDS and counterstained for 2 min in DAPI (1 lg/ml). Following each step, slides were washed in PBS (3 9 5 min per wash). Slides were mounted using fluoromount G (Southern Biotech, Birmingham, AL, USA) and imaged via fluorescence microscopy (Keyence BZ-8000, Osaka, Japan). Primary antibodies anti-K15 (1:200, LHK15; Abcam, Cambridge, UK) and anti-K19 (1:10, Ks19.1; Abcam) were visualised using goat anti-mouse Alexa Fluor 488/594 (Abcam) secondary antibodies. Anti-SOX9 (1:50, Santa Cruz H-90, Heidelberg, Germany) and anti-LHX2 (1:50; Santa Cruz C-20) were visualised with goat anti-rabbit Alexa Fluor 594 (Abcam) and donkey antigoat Alexa Fluor 488 (Abcam) secondary antibodies, respectively. Double immunofluorescence was achieved via simultaneous incubation of antibodies. Where species cross-reactivity was indicated, staining was conducted in sequential steps. For K15, K19 and SOX9 immunohistology, HF cryosections from 6 different patients (3 HFs per patient) were examined, and for LHX2 immunohistology, HF cryosections from 5 different patients (2–3 HFs per patient) were examined. Double immunostains were run on HF sections from at least 3 different patients. See (Figure S1a–c) and associated legend for details on the positive and negative controls.
where K15+/K19+ expression was markedly decreased and irregular, (iii) the pbORS region was defined as the ORS below this region and above the bulb, where K15+ and K19+ cell fluorescence increases again in intensity relative to the sub-bulge zone. Values for positive cell number and cell fluorescence were quantified and pooled per region. Region by region comparisons were statistically assessed using the Mann–Whitney U-test suitable where parametric assumptions are not met. Data handling was performed using GraphPad Prism V 6.0 (GraphPad Software, La Jolla, California, USA).
Real-time quantitative-PCR Isolated occipital scalp HFs were dissected and stored in RNALater (Life Technologies, Carlsbad, California, USA). Tissue was homogenised, RNA was extracted as directed by manufacturer’s protocol (RNeasy mini-kit; Qiagen, Crawley, UK) and converted to cDNA [Tetro cDNA Synthesis Kit (Bioline, London, UK)]. Real-time quantitative-PCR was performed to detect LHX2 and SOX9 transcripts (see Figure S1d for further details). PCRs utilised Taqman reagents and probes (LHX2 – Hs00180351_m1, SOX9 – Hs01001343_g1; Life Technologies).
Results Keratin 15 and keratin 19 demarcate progenitor/stem cells in the human bulge and the pbORS, but show distinct expression patterns First, we investigated the localisation of the widely accepted human HF epithelial progenitor/stem cell markers, K15 and K19 (2–4,9–11,36), by quantitative immunohistomorphometry. The number of K15+ or K19+ cells in defined human HF compartments had not been quantitatively determined before. These data served as a framework to accurately report relative LHX2 and SOX9 protein expression. Significantly more K15+ cells were seen within the bulge compared with the sub-bulge and pbORS (P < 0.001), with the subbulge containing the lowest number of K15+ cells (Fig. 1a,b) (Figure S2a). K15 cell fluorescence correlated with regional cell number, with the most intense staining found in the bulge, lower intensity in the sub-bulge ORS and markedly increased intensity within the pbORS (P ≤ 0.001) (Fig. 1c). In line with previous reports (10,11) K19+ cells were found at the highest density within the bulge and the pbORS regions, with lower numbers in the sub-bulge zone (P < 0.001) (Fig. 1d,e) (Figure S2b). The number of K19+ positive cells in the bulge versus the pbORS was not significantly different (Fig. 1e). The intensity of K19 cell fluorescence was highest in the pbORS compared with the bulge and sub-bulge regions (P < 0.001) (Fig. 1f). This contrasts with K15, where cell fluorescence was most prominent within the bulge, thus highlighting a previously unrecognised distinct expression pattern of K15 and K19 in the human HF.
Analysis
SOX9 and LHX2 are predominantly expressed in cells outside the human bulge
Using ImageJ (NIH, Bethesda, Maryland, USA) software, individual cell fluorescence intensity and positive cell number for markers K15, K19, SOX9 and LHX2 were quantified and compared across the bulge, sub-bulge and proximal bulb ORS regions. Regions to be analysed were differentiated using the following defined criteria: (i) The bulge region was morphologically defined as the K15+/K19+ within isthmus region below the sebaceous gland duct and above the arrector pili muscle attachment site [where CD200 is positive and CX43 is negative (4)]. (ii) The sub-bulge region was defined as the point
Next, we examined the epithelial progenitor cell-associated transcription factors, SOX9 and LHX2. Expression of transcripts was first detected by RT-qPCR analysis of whole human HF mRNA extracts (n = 3 patients, n ≥ 5 HFs per patient). As shown in Figure S1d, LHX2 and SOX9 transcripts were detected, confirming that these progenitor cell marker genes are transcribed in the human HF. SOX9 and LHX2 protein expression and localisation were subsequently characterised. LHX2+ cells were found throughout the human ORS, in the basal and initial suprabasal cell layers,
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
463
Purba et al.
(a)
(b)
(a)
(c)
(b)
(c)
(d) (d)
(e)
(f) (e)
Figure 1. K15 and K19 show distinct expression patterns within the human hair follicle (HF) epithelium (a–c) K15 cell fluorescence and cell number is significantly greater within the bulge versus the sub-bulge and pbORS (d) K19 localises to the bulge and pbORS, with diffuse K19+ cells in-between. (e) The number of K19+ cells in the bulge and pbORS is comparable and significantly decreased within the subbulge (f) K19 cell fluorescence was most prominent within the pbORS compartment. Results for K15 and K19 expression in human HF were consistent with prior accounts [reviewed in Ref. (4)]. Analysis: n = 6 patient, (3 HFs per patient). Error bars are SEM. Comparisons were statistically assessed using the Mann–Whitney U-test. ***P ≤ 0.001. SG - sebaceous gland. APM - Arrector Pili Muscle. CTS - Connective Tissue Sheath. ORS - Outer Root Sheath. Cp Companion Layer. IRS - Inner Root Sheath. HS - Hair Shaft. DP - Dermal Papilla. Please see Figure S2 for representative immunofluorescence images.
yet no LHX2+ cells were seen within the innermost ORS layers (Fig. 2a) (Figure S3a). Variable LHX2 immunoreactivity within the bulge was observed between different HFs, including HFs from the same patient (Figure S3b) Bulge staining was demonstrated via serial sections, showing colocalisation with K15+, CD200+ or CX43-negative cells in the bulge area (Figure S3c). However, significantly more LHX2+ cells were found outside the bulge, within the sub-bulge and pbORS (P ≤ 0.001) (Fig. 2b). The intensity of LHX2 cell fluorescence was also significantly higher in cells of the pbORS than the bulge (P ≤ 0.001) (Fig. 2c), suggesting that pbORS cells upregulate LHX2 protein expression. SOX9+ cells were also found throughout the human ORS, with the highest number of SOX9+ cells found in the sub-bulge region (P ≤ 0.001). Here, cell fluorescence was also significantly more pronounced compared with the bulge (P < 0.001) and pbORS regions (P ≤ 0.01) (Fig. 2d–f) (Figure S4a).
464
(f)
Figure 2. LHX2 and SOX9 expression in the human anagen hair follicle (HF) ORS (a–c) LHX2 is expressed throughout the HF ORS, associating with the basal and adjacent suprabasal layer. The number of LHX2+ cells is significantly increased outside of the bulge. LHX2 cell fluorescence was significantly decreased within the bulge compared with the pbORS, but not sub-bulge. There was no significant difference for LHX2-positive cell number and cell fluorescence between the subbulge and pbORS. (d–f) SOX9 is also expressed throughout cells of the HF ORS. SOX9 cell fluorescence and positive cell number were significantly increased in the sub-bulge region. Error bars are SEM. Comparisons were statistically assessed using the Mann-Whitney U-test. ***P ≤ 0.001. **P ≤ 0.01. LHX2 analysis: n = 5 patients (2–3 HFs per patient), SOX9 analysis: n = 6 patients, 3 HFs per patient. SG sebaceous gland. APM - Arrector Pili Muscle. CTS - Connective Tissue Sheath. ORS - Outer Root Sheath. Cp - Companion Layer. IRS - Inner Root Sheath. HS - Hair Shaft. DP - Dermal Papilla. Please see Figures S3 and S4 for representative immunofluorescence images.
SOX9+ cells showing immunoreactivity within the cytoplasm were primarily located in the outermost layers of the ORS (including the basal layer ORS). This was most apparent in the SOX9+ cell population in the sub-bulge region (Fig. 3a). In contrast, cells showing mainly nuclear SOX9 immunoreactivity were found within the innermost, differentiated ORS layers (Fig. 3a) (Figure S4b); however, nuclear SOX9 immunoreactivity could also be observed in the basal layer (Figure S4b).
SOX9 marks a distinct cell population from K15+ or K19+ cells To determine whether SOX9 labelled distinct cell populations from those demarcated by K15 and/or K19, we performed double-
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
LHX2 and SOX9 expression in human hair follicles
(a)
immunofluorescence microscopy of SOX9 with K15 and K19. SOX9+ cells were found at an overlapping junctional zone with K15+ bulge cells, where K15 expression began to diminish, SOX9 protein expression was upregulated (Fig. 3a). At this junction, very few cells were SOX9 and K15 double positive (Figure S5a), while immediately below this zone, only SOX9 single-positive cells were found. SOX9+ cells also colocalised with the K15+ population in the pbORS, where a small number of SOX9+ cells co-expressed K15 (Figure S5b). SOX9+ cells also colocalised with the irregular, non-contiguous K19+ cell population seen in the sub-bulge region of the ORS (Fig. 3b). Very few sub-bulge K19+ cells also expressed SOX9, and these few double-positive cells were exclusively seen in the basal layer ORS. Similarly, only a small number of SOX9+ cells coexpressed K19 in the pbORS (Figure S5c). These expression data suggest that SOX9 single-positive cells in the human HF represent a distinct epithelial progenitor cell population from K15+ or K19+ cells while SOX9/K15 or SOX9/K19 double-positive cells may represent functionally distinct intra-follicular epithelial progenitor cells.
(b)
(c)
Figure 3. SOX9 double immunofluorescence with K15, K19 and LHX2 (a) SOX9 expression is more pronounced below the K15 bulge compartment, with few cells showing co-expression at the bulge/sub-bulge border. Note SOX9 cytoplasmic- and nuclear-staining patterns in the outermost and suprabasal ORS layers, respectively. (b) SOX9 can colocalise with K19 in the human hair follicle (HF), but few cells coexpress these proteins. (c) SOX9 colocalises with LHX2 in the human hair follicle, but only a small number of cells co-express these proteins. Please see Figure S5e for quantitative analysis. Double-immunofluorescence staining was carried out in HFs from three independent patients. ORS – outer root sheath. CTS – connective tissue sheath. HS – hair shaft. Scale bars are 40 lm. Please also see Figure S5 for further double-immunofluorescence images.
SOX9 NS
20
(b)
*
Cell fluourescence
15 10 5 0
***
40 20
Proximal bulb
S
S
R
R en
pb O
pb O ag
na
at C
ge n
bbu A
Catagen
(d)
CTS
lg e
ul ge
Su en ag
ge n na A
Catagen
Proximal bulb
CTS
LHX2
Figure 4. SOX9 and LHX2 are upregulated in the pbORS of the Catagen hair follicle (HF). (a, b) SOX9 and LHX2 cell fluorescence is unchanged in the sub-bulge but is significantly increased in the pbORS of cultured catagen HFs compared with cultured anagen HFs (*P ≤ 0.05. ***P ≤ 0.001, respectively). Error bars are SEM. Comparisons were statistically assessed using the Mann-Whitney U-test. Analysis conducted on isolated HFs cultured [as described previously (48)] for ~5 days to allow approximately 50% of HFs to enter catagen. HFs were staged and subsequently processed as described in Methods. HFs pooled from three patients (two female, one male). 14 anagen HFs versus 21 catagen HFs and 10 anagen HFs versus 11 catagen HFs analysed for SOX9 and LHX2, respectively. (c, d) SOX9 and LHX2 shows immunoreactivity in the proximal bulb regressing catagen epithelium in a late catagen HF (serial sections). Scale bars are 40 lm. Please see Figure S6 for comparative anagen/catagen immunofluorescence images.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
Double-immunofluorescence revealed that most cells in the subbulge and pbORS were either SOX9 or LHX2 single positive (Fig. 3c) (Figure S5d), suggesting that these transcription factors demarcate distinct progenitor cell populations in the human HF. However, approximately 17% of cells in these regions were SOX9/LHX2 double positive (Figure S5e), demarcating another phenotypically distinct epithelial progenitor cell population outside of the bulge. Whether these phenotypic differences in SOX9, LHX2, K15 and K19 expression also translate into functional differences between these distinct subpopulations remains to be dissected.
During catagen development, LHX2 and SOX9 protein expression increases in the proximal hair bulb epithelium
60
Su bb
S
pb O R S
R pb O
C at ag en
A na ge n
bu lg e
lg e
Su b-
bbu C
at
ag en
Su A na ge n
NS
0
(c)
SOX9
LHX2 80
C at
Cell fluourescence
(a)
LHX2 labels a subpopulation distinct from SOX9+ cells
SOX9 and LHX2 cell fluorescence were significantly heightened in the pbORS (but not the sub-bulge region) of HFs that had spontaneously switched from anagen to catagen in HF organ culture, (Fig. 4a,b) (Figure S6a–d). Correspondingly, prominent LHX2 and SOX9 protein expression was evident in cultured late catagen HFs in the pbORS (Fig. 4c,d). This pattern was reminiscent of sustained K15 and K19/SOX9 staining patterns identified in the epithelial strand of a freshly isolated (as opposed to cultured) late catagen HF. SOX9 (Figure S7).
Discussion K15 and K19 are well-characterised human bulge eHFSC and pbORS markers (2–4,9–11,36). We found that the expression levels of K15 and K19 in human anagen HFs are quite distinct, with K15 expression strongest in the bulge, whereas maximal K19 protein expression was observed in the pbORS. This confirms previous reports that K15 and K19 label different cell populations (4,9), yet further provides a clear dissection of K15 and K19 expression differences throughout the human HF epithelium. This K15 and K19 protein expression profile serves as an important basis for defining the specific expression and localisation of murine eHFSC-related transcription factors SOX9 and LHX2 within the human HF ORS. Despite previous reports of SOX9 expression in the human ORS (37–39), detailed characterisation of SOX9+ cell localisation
465
Purba et al.
has been lacking. In addition, reports of LHX2 expression within the human HF are conflicting, with some authors claiming underrepresentation of LHX2 within the human bulge (2,10), while others report bulge localisation (30). Our current data demonstrate that LHX2+ cells are distributed throughout the human ORS, including the bulge region, but are primarily located in the subbulge and pbORS. SOX9+ cells were also distributed across the ORS and were located mainly in the sub-bulge region. The expression of LHX2 and SOX9 in the human anagen HF shares similarities with the distribution reported in the murine anagen HF, with the majority of studies reporting LHX2+ and SOX9+ cells distributed across the murine ORS, including the bulge (16,17,19,30,40). One conflicting report does, however, suggest that LHX2 is largely absent in the mouse anagen bulge (29). Although we report here the most representative expression patterns, the observed interfollicular variability in SOX9 and LHX2 protein expression could result from differences in intrinsic or extrinsic micro-environmental eHFSC signalling cues, cell cycle activity or otherwise due to differences in the duration in which a HF has been in anagen VI at the time of analysis. In fact, our observation that both SOX9 and LHX2 cell fluorescence are significantly increased in the pbORS during catagen development, also raises the possibility that the observed interfollicular differences in SOX9 and LHX2 expression may have reflected that some anagen HFS were closer to catagen entry than others. K15/SOX9, K19/SOX9 and LHX2/SOX9 double immunostaining in the human anagen HF demonstrated that most SOX9+ cells in human HF epithelium do not co-express K15, K19 or LHX2. The finding that LHX2 and SOX9 demarcate distinct cell populations has been documented before in the murine hair placode, where LHX2 and SOX9 show distinctive basal and suprabasal expression patterns respectively (30,40). Moreover, the LHX2+ and SOX9+ non-bulge ORS compartments of the human HF are not enriched with quiescent eHFSCs [i.e. they are not known to be label retaining, unlike the bulge (36), and show greater proliferative activity than the bulge (4,11)]. Therefore, the expression of stem cell-related markers in these human HF regions suggests the presence of non-quiescent stem/progenitor cell populations. SOX9 is associated with the regulation of proliferation and differentiation in various tissues (21–23,38) and is expressed in hyperproliferative disorders such as psoriasis and skin carcinomas (24). In human cultured ORS cells, SOX9 expression confers greater colony-forming efficiency and proliferative capacity and has been claimed to promote K15 expression (37). On this basis, and alongside limited co-expression with other eHFSC markers, it is reasonable to speculate that the SOX9+ (and SOX9/K15+, SOX9/K19+ and SOX9/LHX2+) cells that can be visualised in distinct compartments of the human ORS represent different epithelial cell progenitor subpopulations that are in distinct stages of keratinocyte proliferation and/or differentiation. However, patients with hypertrichosis terminalis with gingival hyperplasia have been identified to carry mutations upstream of
References 1 Cotsarelis G. J Invest Dermatol 2006: 126: 1459–1468. 2 Ohyama M, Terunuma A, Tock C L et al. J Clin Invest 2006: 116: 249–260.
466
SOX9, which reportedly reduce its expression (38). This counterintuitive observation emphasises both the importance and complexity of SOX9 activity in the normal regulatory control of cell proliferation/differentiation in the human HF, which may entirely depend on context (e.g. hair matrix versus outer root sheath epithelial cells, differences in the local signalling milieu and activity level of partner transcription factors). Possibly, SOX9 silencing experiments in HF organ culture using previously published siRNA methodology (41,42) can clarify SOX9 function in human HF physiology. It cannot be excluded that these transcription factors regulate the fate of other progenitor cell types that are present within the ORS and other HF compartments. For example, melanoblasts/amelanotic melanocytes in the human ORS (43) may be regulated by SOX9, as this transcription factor has been implicated in the differentiation of epidermal melanoblasts in vitro (44). Furthermore, SOX9 is also expressed within hair matrix melanocytes (39), where it could be involved in the regulation of melanogenesis [analogous to its role in epidermal melanocytes (45)], which is coupled to the anagen phase (46). Epithelial hair follicle stem cell markers K15, K19, LHX2 and SOX9 show unique expression patterns in the human HF, both in and outside of the quiescent stem cell niche. Therefore, it is important to evaluate all ORS compartments, and not just the bulge, when studying the human HF during observational or experimental conditions that are hypothesised to affect HF stem/ progenitor cell biology. Moreover, it is important for researchers to recognise and precisely define which compartment of the human HF they are studying. For example, it is all too easy to mistake K15+ cells that are actually located in the pbORS with K15+ cells in the bulge region. Study of each ORS subcompartment is necessary if we are to accurately determine the role of these proteins in human HF progenitor cell maintenance and differentiation/lineage specification. Furthermore, this may help elucidate the respective contributions of progenitor cells to general human HF homoeostasis and tissue remodelling/renewal through HF cycling and in turn will assist us on our quest to explore the uses of distinct HF stem/progenitor subpopulations clinically, such as in the context of wound repair (47).
Acknowledgements This work was supported by a UNILEVER BBSRC iCASE PhD studentship (recipient: T.P.; award: R.P.) in collaboration with Unilever, Colworth, UK. Weiping Li, Stella Pearson, Derek Pye and Dr. Jonathan Hardman are gratefully acknowledged for technical support and all members of the Paus laboratory for constructive criticism.
Author contributions T.P performed the experiments, data analysis and prepared the manuscript. A.S contributed essential research material. I.H, R.B and R.P contributed research design, planning and manuscript preparation, R.P. conceived the study and edited the manuscript.
Conflict of interest The authors have declared no conflicting interests.
3 Tiede S, Koop N, Kloepper J E et al. Stem Cells 2009: 27: 2793–2803. 4 Purba T S, Haslam I S, Poblet E et al. BioEssays 2014: 36: 513–525.
5 Festa E, Fretz J, Berry R et al. Cell 2011: 146: 761–771. 6 Sennett R, Rendl M. Semin Cell Dev Biol 2012: 23: 917–927.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
LHX2 and SOX9 expression in human hair follicles
7 Brownell I, Guevara E, Bai C B et al. Stem Cell 2011: 8: 552–565. 8 Kandyba E, Leung Y, Chen Y-B et al. Proc Natl Acad Sci USA 2013: 110: 1351–1356. 9 Pontiggia L, Biedermann T, Meuli M et al. J Invest Dermatol 2009: 129: 480–490. 10 Kloepper J E, Tiede S, Brinckmann J et al. Exp Dermatol 2008: 17: 592–609. 11 Commo S, Gaillard O, Bernard B A. Differentiation 2000: 66: 157–164. 12 Inoue K, Aoi N, Sato T et al. Lab Invest 2009: 89: 844–856. 13 Sellheyer K, Krahl D. Br J Dermatol 2011: 164: 141–147. 14 Jaks V, Barker N, Kasper M et al. Nat Genet 2008: 40: 1291–1299. 15 Keyes B E, Segal J P, Heller E et al. Proc Natl Acad Sci USA 2013: 110: E4950–E4959. 16 Kadaja M, Keyes B E, Lin M et al. Genes Dev 2014: 28: 328–341. 17 Folgueras A R, Guo X, Pasolli H A et al. Cell Stem Cell 2013: 13: 314–327. 18 Jaks V, Kasper M, Toftg ard R. Exp Cell Res 2010: 316: 1422–1428. €tzkendorf S 19 Vidal V P I, Chaboissier M-C, Lu et al. Curr Biol 2005: 15: 1340–1351. 20 Vidal V P I, Ortonne N, Schedl A. J Cutan Pathol 2008: 35: 373–379. 21 Bastide P, Darido C, Pannequin J et al. J Cell Biol 2007: 178: 635–648. 22 Rockich B E, Hrycaj S M, Shih H P et al. Proc Natl Acad Sci 2013: 110: E4456– E4464. 23 McDonald E, Li J, Krishnamurthy M et al. Int J Biochem Cell Biol 2012: 44: 72–83. 24 Shi G, Sohn K-C, Li Z et al. PLoS One 2013: 8: e54355.
25 Kamachi Y, Kondoh H. Development 2013: 140: 4129–4144. 26 Bi W, Deng J M, Zhang Z et al. Nat Genet 1999: 22: 85–89. 27 Moniot B, Declosmenil F, Barrionuevo F et al. Development 2009: 136: 1813–1821. 28 Rhee H, Polak L, Fuchs E. Science 2006: 312: 1946–1949. €rnqvist G, Sandberg A, H€agglund A et al. 29 To PLoS Genet 2010: 6: e1000904. 30 Mardaryev A N, Meier N, Poterlowicz K et al. Development 2011: 138: 4843–4852. 31 Porter F D, Drago J, Xu Y et al. Development 1997: 124: 2935–2944. 32 Roy A, de Melo J, Chaturvedi D et al. J Neurosci 2013: 33: 6877–6884. 33 Kitajima K, Kawaguchi M, Iacovino M et al. Stem Cells 2013: 31: 2680–2689. 34 Peukert D, Weber S, Lumsden A et al. PLoS Biol 2011: 9: e1001218. 35 Berghard A, H€agglund A-C, Bohm S et al. FASEB J 2012: 26: 3464–3472. 36 Lyle S, Christofidou-Solomidou M, Liu Y, Elder D E, Al-beida S, Cotsarelis G. The C8/144B monoclonal antibody recognizes cytokeratin 15 and defines the location of human hair follicle stem cells. J Cell Sci 1998: 111: 3179–3188. 37 Shi G, Sohn K-C, Kim S-Y et al. Ann Dermatol 2011: 23: 138–143. 38 Fantauzzo K A, Kurban M, Levy B et al. PLoS Genet 2012: 8: e1003002. 39 Krahl D, Sellheyer K. Br J Dermatol 2010: 162: 1294–1301. 40 Nowak J A, Polak L, Pasolli H A et al. Cell Stem Cell 2008: 3: 33–43. 41 Samuelov L, Sprecher E, Tsuruta D et al. J Invest Dermatol 2012: 132: 2332–2341.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
42 Hardman J A, Tobin D J, Haslam I S et al. J Invest Dermatol 2014: 135: 1053–1064. DOI: 10.1038/jid.2014.442. 43 Nishimura E K. Pigment Cell Melanoma Res 2011: 24: 401–410. 44 Cook A L, Smith A G, Smit D J et al. Exp Cell Res 2005: 308: 222–235. 45 Passeron T, Valencia J C, Bertolotto C et al. Proc Natl Acad Sci USA 2007: 104: 13984–13989. 46 Slominski A, Paus R. J Invest Dermatol 1993: 101: 90S–97S. 47 Jimenez F, Poblet E, Izeta A. Exp Dermatol 2014: 24: 91–94. DOI: 10.1111/exd.12521. 48 Philpott M P, Sanders D, Westgate G E et al. J Dermatol Sci 1994: 7: S55–S72.
Supporting Information Additional supporting data may be found in the supplementary information of this article. Figure S1. Immunofluorescence microscopy controls & mRNA transcript detection. Figure S2. K15 and K19 immunofluorescence in the human HF ORS. Figure S3. LHX2 immunofluorescence in the human HF ORS. Figure S4. SOX9 immunofluorescence in the human HF ORS. Figure S5. SOX9 double-immunofluorescence with K15, K19 and LHX2. Figure S6. SOX9 and LHX2 immunofluorescence in cultured anagen versus catagen HFs. Figure S7. K19+SOX9 and K15 immunofluorescence in the late catagen HF.
467
Patterns of Expression
Mapping the expression of epithelial hair follicle stem cell-related transcription factors LHX2 and SOX9 in the human hair follicle Talveen S. Purba1, Iain S. Haslam1, Asim Shahmalak2, Ranjit K. Bhogal3 and Ralf Paus1,4 1 Centre for Dermatology Research, Institute of Inflammation and Repair, University of Manchester, Manchester, UK; 2Crown Clinic, Manchester, UK; 3Unilever R&D Colworth, Bedfordshire MK44 1LQ, UK; 4Department of Dermatology, University of M€ unster, M€ unster, Germany Correspondence: Ralf Paus, Centre for Dermatology Research, Institute of Inflammation & Repair, University of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT, UK, e-mail: [email protected]
Abstract: In the murine hair follicle (HF), the transcription factors LHX2 and SOX9 are implicated in epithelial hair follicle stem cell (eHFSC) self-renewal and the maintenance of eHFSC niche characteristics. However, the exact expression patterns of LHX2 and SOX9 in the human HF are unclear. Therefore, we have quantitatively mapped the localisation of known human eHFSC markers keratin 15 (K15) and keratin 19 (K19) in the outer root sheath (ORS) of male occipital scalp anagen HFs and related this to the localisation of LHX2 and SOX9 protein expression. As expected, K15+ and K19+ cells represented two distinct progenitor cell populations in the bulge and in the proximal bulb ORS (pbORS). Interestingly, cell fluorescence for K19 was significantly stronger within the pbORS versus the bulge, and vice versa for K15, describing a hitherto unrecognised differential expression pattern. LHX2 and SOX9 expressing cells were distributed throughout the ORS, including the bulge, but were not restricted to it. SOX9 expression was most prominent
in the ORS immediately below the human bulge, whereas LHX2+ cells were similarly distributed between the sub-bulge and pbORS, that is compartments not enriched with quiescent eHFSCs. During catagen development, the intensity of LHX2 and SOX9 protein expression increased in the proximal HF epithelium. Double immunostaining showed that the majority of SOX9+ cells in the human anagen HF epithelium did not co-express K15, K19 or LHX2. This expression profile suggests that LHX2 and SOX9 highlight distinct epithelial progenitor cell populations, in addition to K15+ or K19+ cells, that could play an important role in the maintenance of the human HF epithelium.
Introduction
distribution and functional relevance for HF biology (14–18), our understanding of human HF stem cell markers remains rudimentary. For example, eHFSC transcription factor SRY (sex determining region Y)-box 9 (SOX9), which is downstream of SHH signalling (19,20), has recently been shown to regulate activin/pSMAD2 signalling (16) and is known to be required in eHFSC maintenance and ORS differentiation in the murine hair follicle (16,19). SOX9 is also expressed within progenitor cells of multiple tissues including the murine intestinal and lung epithelium (21,22), human fetal pancreas (23) and the basal layer of human epidermis (24). Its multifaceted roles as a transcription factor, whose biological effects depend on its partnering with other transcription factors (25), involve the regulation of somatic progenitor cell proliferation and their respective differentiation programmes (21–23). Moreover, SOX9 has been implicated in the regulation of diverse processes such as chondrogenesis (26) and mammalian sex determination (27). Another transcription factor known to be important in murine epithelial HF stem/progenitor cell biology is Lim homeobox 2 (LHX2) (17,28,29), which regulates SOX9 (30). It has recently been shown that LHX2 maintains the organisation of the murine eHFSC niche and regulates known eHFSC genes as well as cytoskeletal and cell adhesion-related proteins, which could help mediate asymmetric self-renewal of eHFSCs (17). Beyond the HF system, LHX2 has also been implicated in other, namely developmental (31), contexts such as the development of the eye (32),
The human hair follicle (HF) is a veritable stem cell repository and uniquely suited to study various human progenitor cell populations, namely epithelial HF stem cells (eHFSCs), within their natural habitat (1–4). eHFSCs reside in the bulge region of the HF’s outer root sheath (ORS) and allow remodelling and tissue renewal during HF cycling through growth (anagen), regression (catagen) and relative rest (telogen). Murine studies have shown that eHFSCs are controlled by locally derived micro-environmental regulatory signalling (5–8). Cells expressing known markers of human epithelial stem and progenitor cells are found throughout the ORS of the HF isthmus and suprabulbar regions (4), which can be further subdivided into the bulge, sub-bulge and proximal bulb ORS (pbORS) regions. The epithelial stem/progenitor cell markers keratin 15 (K15) and keratin 19 (K19) localise to the bulge and pbORS of the human anagen HF (9–11). Other useful epithelial stem/progenitor markers including CD200, PHLDA1 and negative expression for the gap junction protein connexin 43 (CX43) demarcate the bulge, whereas CD34 and p75NTR mark the suprabulbar ORS area (4,9– 13). This differential expression of eHFSC markers may represent multiple functionally distinct cell populations, where bulge cells are thought to be relatively quiescent, as opposed to the relatively activated and fate-restricted progenitor cells in the sub-bulge and pbORS (4). However, while the currently available markers for murine eHFSCs are becoming increasingly well defined, both in their
462
Key words: eHFSCs – epithelial human hair follicle stem cells – K15 – K19 – LHX2 – SOX9
Accepted for publication 17 March 2015
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
LHX2 and SOX9 expression in human hair follicles
lineage specification and maintenance of haematopoietic progenitor cells (33) and neuronal patterning and specification (34,35). However, the distribution of LHX2+ or SOX9+ cells within the human HF remains relatively uncharted territory. The exact intrafollicular location and expression characteristics of epithelial progenitor cell markers in the human HF are an important prerequisite for understanding their functional role in human HF biology. Thus, we have mapped the localisation of cells expressing LHX2 and SOX9 in the human HF ORS relative to the established markers K15 and K19 by quantitative immunohistomorphometry on occipital scalp human HFs.
Methods Tissue preparation and immunofluorescence Occipital scalp tissue was obtained from male patients undergoing transplantation surgery after informed consent under ethical and institutional approval from the University of Manchester in accordance with Human Tissue Act guidelines. HFs were isolated and embedded in optimal cutting temperature compound (OCT) and frozen in liquid nitrogen. HF cryosections (at a thickness of 7 lM) were prepared for immunofluorescence protocols. Tissue sections were fixed for 10 min in ice-cold ( 20°C) acetone and treated overnight with primary antibody diluted (see optimised dilutions below) in phosphate-buffered saline (PBS) with either 10% normal goat serum (NGS) or normal donkey serum (NDS) at 4°C in a humid chamber. Slides were treated with secondary antibody for 45 min diluted in 1:200 in PBS + 10% NGS or 10% NDS and counterstained for 2 min in DAPI (1 lg/ml). Following each step, slides were washed in PBS (3 9 5 min per wash). Slides were mounted using fluoromount G (Southern Biotech, Birmingham, AL, USA) and imaged via fluorescence microscopy (Keyence BZ-8000, Osaka, Japan). Primary antibodies anti-K15 (1:200, LHK15; Abcam, Cambridge, UK) and anti-K19 (1:10, Ks19.1; Abcam) were visualised using goat anti-mouse Alexa Fluor 488/594 (Abcam) secondary antibodies. Anti-SOX9 (1:50, Santa Cruz H-90, Heidelberg, Germany) and anti-LHX2 (1:50; Santa Cruz C-20) were visualised with goat anti-rabbit Alexa Fluor 594 (Abcam) and donkey antigoat Alexa Fluor 488 (Abcam) secondary antibodies, respectively. Double immunofluorescence was achieved via simultaneous incubation of antibodies. Where species cross-reactivity was indicated, staining was conducted in sequential steps. For K15, K19 and SOX9 immunohistology, HF cryosections from 6 different patients (3 HFs per patient) were examined, and for LHX2 immunohistology, HF cryosections from 5 different patients (2–3 HFs per patient) were examined. Double immunostains were run on HF sections from at least 3 different patients. See (Figure S1a–c) and associated legend for details on the positive and negative controls.
where K15+/K19+ expression was markedly decreased and irregular, (iii) the pbORS region was defined as the ORS below this region and above the bulb, where K15+ and K19+ cell fluorescence increases again in intensity relative to the sub-bulge zone. Values for positive cell number and cell fluorescence were quantified and pooled per region. Region by region comparisons were statistically assessed using the Mann–Whitney U-test suitable where parametric assumptions are not met. Data handling was performed using GraphPad Prism V 6.0 (GraphPad Software, La Jolla, California, USA).
Real-time quantitative-PCR Isolated occipital scalp HFs were dissected and stored in RNALater (Life Technologies, Carlsbad, California, USA). Tissue was homogenised, RNA was extracted as directed by manufacturer’s protocol (RNeasy mini-kit; Qiagen, Crawley, UK) and converted to cDNA [Tetro cDNA Synthesis Kit (Bioline, London, UK)]. Real-time quantitative-PCR was performed to detect LHX2 and SOX9 transcripts (see Figure S1d for further details). PCRs utilised Taqman reagents and probes (LHX2 – Hs00180351_m1, SOX9 – Hs01001343_g1; Life Technologies).
Results Keratin 15 and keratin 19 demarcate progenitor/stem cells in the human bulge and the pbORS, but show distinct expression patterns First, we investigated the localisation of the widely accepted human HF epithelial progenitor/stem cell markers, K15 and K19 (2–4,9–11,36), by quantitative immunohistomorphometry. The number of K15+ or K19+ cells in defined human HF compartments had not been quantitatively determined before. These data served as a framework to accurately report relative LHX2 and SOX9 protein expression. Significantly more K15+ cells were seen within the bulge compared with the sub-bulge and pbORS (P < 0.001), with the subbulge containing the lowest number of K15+ cells (Fig. 1a,b) (Figure S2a). K15 cell fluorescence correlated with regional cell number, with the most intense staining found in the bulge, lower intensity in the sub-bulge ORS and markedly increased intensity within the pbORS (P ≤ 0.001) (Fig. 1c). In line with previous reports (10,11) K19+ cells were found at the highest density within the bulge and the pbORS regions, with lower numbers in the sub-bulge zone (P < 0.001) (Fig. 1d,e) (Figure S2b). The number of K19+ positive cells in the bulge versus the pbORS was not significantly different (Fig. 1e). The intensity of K19 cell fluorescence was highest in the pbORS compared with the bulge and sub-bulge regions (P < 0.001) (Fig. 1f). This contrasts with K15, where cell fluorescence was most prominent within the bulge, thus highlighting a previously unrecognised distinct expression pattern of K15 and K19 in the human HF.
Analysis
SOX9 and LHX2 are predominantly expressed in cells outside the human bulge
Using ImageJ (NIH, Bethesda, Maryland, USA) software, individual cell fluorescence intensity and positive cell number for markers K15, K19, SOX9 and LHX2 were quantified and compared across the bulge, sub-bulge and proximal bulb ORS regions. Regions to be analysed were differentiated using the following defined criteria: (i) The bulge region was morphologically defined as the K15+/K19+ within isthmus region below the sebaceous gland duct and above the arrector pili muscle attachment site [where CD200 is positive and CX43 is negative (4)]. (ii) The sub-bulge region was defined as the point
Next, we examined the epithelial progenitor cell-associated transcription factors, SOX9 and LHX2. Expression of transcripts was first detected by RT-qPCR analysis of whole human HF mRNA extracts (n = 3 patients, n ≥ 5 HFs per patient). As shown in Figure S1d, LHX2 and SOX9 transcripts were detected, confirming that these progenitor cell marker genes are transcribed in the human HF. SOX9 and LHX2 protein expression and localisation were subsequently characterised. LHX2+ cells were found throughout the human ORS, in the basal and initial suprabasal cell layers,
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
463
Purba et al.
(a)
(b)
(a)
(c)
(b)
(c)
(d) (d)
(e)
(f) (e)
Figure 1. K15 and K19 show distinct expression patterns within the human hair follicle (HF) epithelium (a–c) K15 cell fluorescence and cell number is significantly greater within the bulge versus the sub-bulge and pbORS (d) K19 localises to the bulge and pbORS, with diffuse K19+ cells in-between. (e) The number of K19+ cells in the bulge and pbORS is comparable and significantly decreased within the subbulge (f) K19 cell fluorescence was most prominent within the pbORS compartment. Results for K15 and K19 expression in human HF were consistent with prior accounts [reviewed in Ref. (4)]. Analysis: n = 6 patient, (3 HFs per patient). Error bars are SEM. Comparisons were statistically assessed using the Mann–Whitney U-test. ***P ≤ 0.001. SG - sebaceous gland. APM - Arrector Pili Muscle. CTS - Connective Tissue Sheath. ORS - Outer Root Sheath. Cp Companion Layer. IRS - Inner Root Sheath. HS - Hair Shaft. DP - Dermal Papilla. Please see Figure S2 for representative immunofluorescence images.
yet no LHX2+ cells were seen within the innermost ORS layers (Fig. 2a) (Figure S3a). Variable LHX2 immunoreactivity within the bulge was observed between different HFs, including HFs from the same patient (Figure S3b) Bulge staining was demonstrated via serial sections, showing colocalisation with K15+, CD200+ or CX43-negative cells in the bulge area (Figure S3c). However, significantly more LHX2+ cells were found outside the bulge, within the sub-bulge and pbORS (P ≤ 0.001) (Fig. 2b). The intensity of LHX2 cell fluorescence was also significantly higher in cells of the pbORS than the bulge (P ≤ 0.001) (Fig. 2c), suggesting that pbORS cells upregulate LHX2 protein expression. SOX9+ cells were also found throughout the human ORS, with the highest number of SOX9+ cells found in the sub-bulge region (P ≤ 0.001). Here, cell fluorescence was also significantly more pronounced compared with the bulge (P < 0.001) and pbORS regions (P ≤ 0.01) (Fig. 2d–f) (Figure S4a).
464
(f)
Figure 2. LHX2 and SOX9 expression in the human anagen hair follicle (HF) ORS (a–c) LHX2 is expressed throughout the HF ORS, associating with the basal and adjacent suprabasal layer. The number of LHX2+ cells is significantly increased outside of the bulge. LHX2 cell fluorescence was significantly decreased within the bulge compared with the pbORS, but not sub-bulge. There was no significant difference for LHX2-positive cell number and cell fluorescence between the subbulge and pbORS. (d–f) SOX9 is also expressed throughout cells of the HF ORS. SOX9 cell fluorescence and positive cell number were significantly increased in the sub-bulge region. Error bars are SEM. Comparisons were statistically assessed using the Mann-Whitney U-test. ***P ≤ 0.001. **P ≤ 0.01. LHX2 analysis: n = 5 patients (2–3 HFs per patient), SOX9 analysis: n = 6 patients, 3 HFs per patient. SG sebaceous gland. APM - Arrector Pili Muscle. CTS - Connective Tissue Sheath. ORS - Outer Root Sheath. Cp - Companion Layer. IRS - Inner Root Sheath. HS - Hair Shaft. DP - Dermal Papilla. Please see Figures S3 and S4 for representative immunofluorescence images.
SOX9+ cells showing immunoreactivity within the cytoplasm were primarily located in the outermost layers of the ORS (including the basal layer ORS). This was most apparent in the SOX9+ cell population in the sub-bulge region (Fig. 3a). In contrast, cells showing mainly nuclear SOX9 immunoreactivity were found within the innermost, differentiated ORS layers (Fig. 3a) (Figure S4b); however, nuclear SOX9 immunoreactivity could also be observed in the basal layer (Figure S4b).
SOX9 marks a distinct cell population from K15+ or K19+ cells To determine whether SOX9 labelled distinct cell populations from those demarcated by K15 and/or K19, we performed double-
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
LHX2 and SOX9 expression in human hair follicles
(a)
immunofluorescence microscopy of SOX9 with K15 and K19. SOX9+ cells were found at an overlapping junctional zone with K15+ bulge cells, where K15 expression began to diminish, SOX9 protein expression was upregulated (Fig. 3a). At this junction, very few cells were SOX9 and K15 double positive (Figure S5a), while immediately below this zone, only SOX9 single-positive cells were found. SOX9+ cells also colocalised with the K15+ population in the pbORS, where a small number of SOX9+ cells co-expressed K15 (Figure S5b). SOX9+ cells also colocalised with the irregular, non-contiguous K19+ cell population seen in the sub-bulge region of the ORS (Fig. 3b). Very few sub-bulge K19+ cells also expressed SOX9, and these few double-positive cells were exclusively seen in the basal layer ORS. Similarly, only a small number of SOX9+ cells coexpressed K19 in the pbORS (Figure S5c). These expression data suggest that SOX9 single-positive cells in the human HF represent a distinct epithelial progenitor cell population from K15+ or K19+ cells while SOX9/K15 or SOX9/K19 double-positive cells may represent functionally distinct intra-follicular epithelial progenitor cells.
(b)
(c)
Figure 3. SOX9 double immunofluorescence with K15, K19 and LHX2 (a) SOX9 expression is more pronounced below the K15 bulge compartment, with few cells showing co-expression at the bulge/sub-bulge border. Note SOX9 cytoplasmic- and nuclear-staining patterns in the outermost and suprabasal ORS layers, respectively. (b) SOX9 can colocalise with K19 in the human hair follicle (HF), but few cells coexpress these proteins. (c) SOX9 colocalises with LHX2 in the human hair follicle, but only a small number of cells co-express these proteins. Please see Figure S5e for quantitative analysis. Double-immunofluorescence staining was carried out in HFs from three independent patients. ORS – outer root sheath. CTS – connective tissue sheath. HS – hair shaft. Scale bars are 40 lm. Please also see Figure S5 for further double-immunofluorescence images.
SOX9 NS
20
(b)
*
Cell fluourescence
15 10 5 0
***
40 20
Proximal bulb
S
S
R
R en
pb O
pb O ag
na
at C
ge n
bbu A
Catagen
(d)
CTS
lg e
ul ge
Su en ag
ge n na A
Catagen
Proximal bulb
CTS
LHX2
Figure 4. SOX9 and LHX2 are upregulated in the pbORS of the Catagen hair follicle (HF). (a, b) SOX9 and LHX2 cell fluorescence is unchanged in the sub-bulge but is significantly increased in the pbORS of cultured catagen HFs compared with cultured anagen HFs (*P ≤ 0.05. ***P ≤ 0.001, respectively). Error bars are SEM. Comparisons were statistically assessed using the Mann-Whitney U-test. Analysis conducted on isolated HFs cultured [as described previously (48)] for ~5 days to allow approximately 50% of HFs to enter catagen. HFs were staged and subsequently processed as described in Methods. HFs pooled from three patients (two female, one male). 14 anagen HFs versus 21 catagen HFs and 10 anagen HFs versus 11 catagen HFs analysed for SOX9 and LHX2, respectively. (c, d) SOX9 and LHX2 shows immunoreactivity in the proximal bulb regressing catagen epithelium in a late catagen HF (serial sections). Scale bars are 40 lm. Please see Figure S6 for comparative anagen/catagen immunofluorescence images.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
Double-immunofluorescence revealed that most cells in the subbulge and pbORS were either SOX9 or LHX2 single positive (Fig. 3c) (Figure S5d), suggesting that these transcription factors demarcate distinct progenitor cell populations in the human HF. However, approximately 17% of cells in these regions were SOX9/LHX2 double positive (Figure S5e), demarcating another phenotypically distinct epithelial progenitor cell population outside of the bulge. Whether these phenotypic differences in SOX9, LHX2, K15 and K19 expression also translate into functional differences between these distinct subpopulations remains to be dissected.
During catagen development, LHX2 and SOX9 protein expression increases in the proximal hair bulb epithelium
60
Su bb
S
pb O R S
R pb O
C at ag en
A na ge n
bu lg e
lg e
Su b-
bbu C
at
ag en
Su A na ge n
NS
0
(c)
SOX9
LHX2 80
C at
Cell fluourescence
(a)
LHX2 labels a subpopulation distinct from SOX9+ cells
SOX9 and LHX2 cell fluorescence were significantly heightened in the pbORS (but not the sub-bulge region) of HFs that had spontaneously switched from anagen to catagen in HF organ culture, (Fig. 4a,b) (Figure S6a–d). Correspondingly, prominent LHX2 and SOX9 protein expression was evident in cultured late catagen HFs in the pbORS (Fig. 4c,d). This pattern was reminiscent of sustained K15 and K19/SOX9 staining patterns identified in the epithelial strand of a freshly isolated (as opposed to cultured) late catagen HF. SOX9 (Figure S7).
Discussion K15 and K19 are well-characterised human bulge eHFSC and pbORS markers (2–4,9–11,36). We found that the expression levels of K15 and K19 in human anagen HFs are quite distinct, with K15 expression strongest in the bulge, whereas maximal K19 protein expression was observed in the pbORS. This confirms previous reports that K15 and K19 label different cell populations (4,9), yet further provides a clear dissection of K15 and K19 expression differences throughout the human HF epithelium. This K15 and K19 protein expression profile serves as an important basis for defining the specific expression and localisation of murine eHFSC-related transcription factors SOX9 and LHX2 within the human HF ORS. Despite previous reports of SOX9 expression in the human ORS (37–39), detailed characterisation of SOX9+ cell localisation
465
Purba et al.
has been lacking. In addition, reports of LHX2 expression within the human HF are conflicting, with some authors claiming underrepresentation of LHX2 within the human bulge (2,10), while others report bulge localisation (30). Our current data demonstrate that LHX2+ cells are distributed throughout the human ORS, including the bulge region, but are primarily located in the subbulge and pbORS. SOX9+ cells were also distributed across the ORS and were located mainly in the sub-bulge region. The expression of LHX2 and SOX9 in the human anagen HF shares similarities with the distribution reported in the murine anagen HF, with the majority of studies reporting LHX2+ and SOX9+ cells distributed across the murine ORS, including the bulge (16,17,19,30,40). One conflicting report does, however, suggest that LHX2 is largely absent in the mouse anagen bulge (29). Although we report here the most representative expression patterns, the observed interfollicular variability in SOX9 and LHX2 protein expression could result from differences in intrinsic or extrinsic micro-environmental eHFSC signalling cues, cell cycle activity or otherwise due to differences in the duration in which a HF has been in anagen VI at the time of analysis. In fact, our observation that both SOX9 and LHX2 cell fluorescence are significantly increased in the pbORS during catagen development, also raises the possibility that the observed interfollicular differences in SOX9 and LHX2 expression may have reflected that some anagen HFS were closer to catagen entry than others. K15/SOX9, K19/SOX9 and LHX2/SOX9 double immunostaining in the human anagen HF demonstrated that most SOX9+ cells in human HF epithelium do not co-express K15, K19 or LHX2. The finding that LHX2 and SOX9 demarcate distinct cell populations has been documented before in the murine hair placode, where LHX2 and SOX9 show distinctive basal and suprabasal expression patterns respectively (30,40). Moreover, the LHX2+ and SOX9+ non-bulge ORS compartments of the human HF are not enriched with quiescent eHFSCs [i.e. they are not known to be label retaining, unlike the bulge (36), and show greater proliferative activity than the bulge (4,11)]. Therefore, the expression of stem cell-related markers in these human HF regions suggests the presence of non-quiescent stem/progenitor cell populations. SOX9 is associated with the regulation of proliferation and differentiation in various tissues (21–23,38) and is expressed in hyperproliferative disorders such as psoriasis and skin carcinomas (24). In human cultured ORS cells, SOX9 expression confers greater colony-forming efficiency and proliferative capacity and has been claimed to promote K15 expression (37). On this basis, and alongside limited co-expression with other eHFSC markers, it is reasonable to speculate that the SOX9+ (and SOX9/K15+, SOX9/K19+ and SOX9/LHX2+) cells that can be visualised in distinct compartments of the human ORS represent different epithelial cell progenitor subpopulations that are in distinct stages of keratinocyte proliferation and/or differentiation. However, patients with hypertrichosis terminalis with gingival hyperplasia have been identified to carry mutations upstream of
References 1 Cotsarelis G. J Invest Dermatol 2006: 126: 1459–1468. 2 Ohyama M, Terunuma A, Tock C L et al. J Clin Invest 2006: 116: 249–260.
466
SOX9, which reportedly reduce its expression (38). This counterintuitive observation emphasises both the importance and complexity of SOX9 activity in the normal regulatory control of cell proliferation/differentiation in the human HF, which may entirely depend on context (e.g. hair matrix versus outer root sheath epithelial cells, differences in the local signalling milieu and activity level of partner transcription factors). Possibly, SOX9 silencing experiments in HF organ culture using previously published siRNA methodology (41,42) can clarify SOX9 function in human HF physiology. It cannot be excluded that these transcription factors regulate the fate of other progenitor cell types that are present within the ORS and other HF compartments. For example, melanoblasts/amelanotic melanocytes in the human ORS (43) may be regulated by SOX9, as this transcription factor has been implicated in the differentiation of epidermal melanoblasts in vitro (44). Furthermore, SOX9 is also expressed within hair matrix melanocytes (39), where it could be involved in the regulation of melanogenesis [analogous to its role in epidermal melanocytes (45)], which is coupled to the anagen phase (46). Epithelial hair follicle stem cell markers K15, K19, LHX2 and SOX9 show unique expression patterns in the human HF, both in and outside of the quiescent stem cell niche. Therefore, it is important to evaluate all ORS compartments, and not just the bulge, when studying the human HF during observational or experimental conditions that are hypothesised to affect HF stem/ progenitor cell biology. Moreover, it is important for researchers to recognise and precisely define which compartment of the human HF they are studying. For example, it is all too easy to mistake K15+ cells that are actually located in the pbORS with K15+ cells in the bulge region. Study of each ORS subcompartment is necessary if we are to accurately determine the role of these proteins in human HF progenitor cell maintenance and differentiation/lineage specification. Furthermore, this may help elucidate the respective contributions of progenitor cells to general human HF homoeostasis and tissue remodelling/renewal through HF cycling and in turn will assist us on our quest to explore the uses of distinct HF stem/progenitor subpopulations clinically, such as in the context of wound repair (47).
Acknowledgements This work was supported by a UNILEVER BBSRC iCASE PhD studentship (recipient: T.P.; award: R.P.) in collaboration with Unilever, Colworth, UK. Weiping Li, Stella Pearson, Derek Pye and Dr. Jonathan Hardman are gratefully acknowledged for technical support and all members of the Paus laboratory for constructive criticism.
Author contributions T.P performed the experiments, data analysis and prepared the manuscript. A.S contributed essential research material. I.H, R.B and R.P contributed research design, planning and manuscript preparation, R.P. conceived the study and edited the manuscript.
Conflict of interest The authors have declared no conflicting interests.
3 Tiede S, Koop N, Kloepper J E et al. Stem Cells 2009: 27: 2793–2803. 4 Purba T S, Haslam I S, Poblet E et al. BioEssays 2014: 36: 513–525.
5 Festa E, Fretz J, Berry R et al. Cell 2011: 146: 761–771. 6 Sennett R, Rendl M. Semin Cell Dev Biol 2012: 23: 917–927.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
LHX2 and SOX9 expression in human hair follicles
7 Brownell I, Guevara E, Bai C B et al. Stem Cell 2011: 8: 552–565. 8 Kandyba E, Leung Y, Chen Y-B et al. Proc Natl Acad Sci USA 2013: 110: 1351–1356. 9 Pontiggia L, Biedermann T, Meuli M et al. J Invest Dermatol 2009: 129: 480–490. 10 Kloepper J E, Tiede S, Brinckmann J et al. Exp Dermatol 2008: 17: 592–609. 11 Commo S, Gaillard O, Bernard B A. Differentiation 2000: 66: 157–164. 12 Inoue K, Aoi N, Sato T et al. Lab Invest 2009: 89: 844–856. 13 Sellheyer K, Krahl D. Br J Dermatol 2011: 164: 141–147. 14 Jaks V, Barker N, Kasper M et al. Nat Genet 2008: 40: 1291–1299. 15 Keyes B E, Segal J P, Heller E et al. Proc Natl Acad Sci USA 2013: 110: E4950–E4959. 16 Kadaja M, Keyes B E, Lin M et al. Genes Dev 2014: 28: 328–341. 17 Folgueras A R, Guo X, Pasolli H A et al. Cell Stem Cell 2013: 13: 314–327. 18 Jaks V, Kasper M, Toftg ard R. Exp Cell Res 2010: 316: 1422–1428. €tzkendorf S 19 Vidal V P I, Chaboissier M-C, Lu et al. Curr Biol 2005: 15: 1340–1351. 20 Vidal V P I, Ortonne N, Schedl A. J Cutan Pathol 2008: 35: 373–379. 21 Bastide P, Darido C, Pannequin J et al. J Cell Biol 2007: 178: 635–648. 22 Rockich B E, Hrycaj S M, Shih H P et al. Proc Natl Acad Sci 2013: 110: E4456– E4464. 23 McDonald E, Li J, Krishnamurthy M et al. Int J Biochem Cell Biol 2012: 44: 72–83. 24 Shi G, Sohn K-C, Li Z et al. PLoS One 2013: 8: e54355.
25 Kamachi Y, Kondoh H. Development 2013: 140: 4129–4144. 26 Bi W, Deng J M, Zhang Z et al. Nat Genet 1999: 22: 85–89. 27 Moniot B, Declosmenil F, Barrionuevo F et al. Development 2009: 136: 1813–1821. 28 Rhee H, Polak L, Fuchs E. Science 2006: 312: 1946–1949. €rnqvist G, Sandberg A, H€agglund A et al. 29 To PLoS Genet 2010: 6: e1000904. 30 Mardaryev A N, Meier N, Poterlowicz K et al. Development 2011: 138: 4843–4852. 31 Porter F D, Drago J, Xu Y et al. Development 1997: 124: 2935–2944. 32 Roy A, de Melo J, Chaturvedi D et al. J Neurosci 2013: 33: 6877–6884. 33 Kitajima K, Kawaguchi M, Iacovino M et al. Stem Cells 2013: 31: 2680–2689. 34 Peukert D, Weber S, Lumsden A et al. PLoS Biol 2011: 9: e1001218. 35 Berghard A, H€agglund A-C, Bohm S et al. FASEB J 2012: 26: 3464–3472. 36 Lyle S, Christofidou-Solomidou M, Liu Y, Elder D E, Al-beida S, Cotsarelis G. The C8/144B monoclonal antibody recognizes cytokeratin 15 and defines the location of human hair follicle stem cells. J Cell Sci 1998: 111: 3179–3188. 37 Shi G, Sohn K-C, Kim S-Y et al. Ann Dermatol 2011: 23: 138–143. 38 Fantauzzo K A, Kurban M, Levy B et al. PLoS Genet 2012: 8: e1003002. 39 Krahl D, Sellheyer K. Br J Dermatol 2010: 162: 1294–1301. 40 Nowak J A, Polak L, Pasolli H A et al. Cell Stem Cell 2008: 3: 33–43. 41 Samuelov L, Sprecher E, Tsuruta D et al. J Invest Dermatol 2012: 132: 2332–2341.
ª 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd Experimental Dermatology, 2015, 24, 462–467
42 Hardman J A, Tobin D J, Haslam I S et al. J Invest Dermatol 2014: 135: 1053–1064. DOI: 10.1038/jid.2014.442. 43 Nishimura E K. Pigment Cell Melanoma Res 2011: 24: 401–410. 44 Cook A L, Smith A G, Smit D J et al. Exp Cell Res 2005: 308: 222–235. 45 Passeron T, Valencia J C, Bertolotto C et al. Proc Natl Acad Sci USA 2007: 104: 13984–13989. 46 Slominski A, Paus R. J Invest Dermatol 1993: 101: 90S–97S. 47 Jimenez F, Poblet E, Izeta A. Exp Dermatol 2014: 24: 91–94. DOI: 10.1111/exd.12521. 48 Philpott M P, Sanders D, Westgate G E et al. J Dermatol Sci 1994: 7: S55–S72.
Supporting Information Additional supporting data may be found in the supplementary information of this article. Figure S1. Immunofluorescence microscopy controls & mRNA transcript detection. Figure S2. K15 and K19 immunofluorescence in the human HF ORS. Figure S3. LHX2 immunofluorescence in the human HF ORS. Figure S4. SOX9 immunofluorescence in the human HF ORS. Figure S5. SOX9 double-immunofluorescence with K15, K19 and LHX2. Figure S6. SOX9 and LHX2 immunofluorescence in cultured anagen versus catagen HFs. Figure S7. K19+SOX9 and K15 immunofluorescence in the late catagen HF.
467
