If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Division of Dermatology, Department of Internal Medicine, Washington University School of Medicine, Campus Box 8123, 660 South Euclid Avenue, St Louis, Missouri 63110, USA
Serum response factor (SRF) is a transcription factor that regulates the expression of growth-related immediate-early, cytoskeletal, and muscle-specific genes to control growth, differentiation, and cytoskeletal integrity in different cell types. To investigate the role for SRF in epidermal development and homeostasis, we conditionally knocked out SRF in epidermal keratinocytes. We report that SRF deletion disrupted epidermal barrier function leading to early postnatal lethality. Mice lacking SRF in epidermis displayed morphogenetic defects, including an eye-open-at-birth phenotype and lack of whiskers. SRF-null skin exhibited abnormal morphology, hyperplasia, aberrant expression of differentiation markers and transcriptional regulators, anomalous actin organization, enhanced inflammation, and retarded hair follicle (HF) development. Transcriptional profiling experiments uncovered profound molecular changes in SRF-null E17.5 epidermis and revealed that many previously identified SRF target CArG box-containing genes were markedly upregulated in SRF-null epidermis, indicating that SRF may function to repress transcription of a subset of its target genes in epidermis. Remarkably, when transplanted onto nude mice, engrafted SRF-null skin lacked hair but displayed normal epidermal architecture with proper expression of differentiation markers, suggesting that although keratinocyte SRF is essential for HF development, a cross-talk between SRF-null keratinocytes and the surrounding microenvironment is likely responsible for the barrier-deficient mutant epidermal phenotype.
Abbreviations
HF
hair follicle
PBS
phosphate-buffered saline
SRF
serum response factor
Introduction
The epidermis, a multilayered self-renewing tissue, functions as a barrier against environmental assaults and dehydration. The barrier is first established during embryonic development through the activation of a tightly regulated epidermal differentiation program (
). Disturbances in barrier function are linked to various pathologies, from congenital ichthyoses to inflammatory skin diseases including psoriasis and atopic dermatitis (reviewed in
). Upon sensing barrier breach, epidermal cells turn on the expression of stress-responsive genes initiating proinflammatory cytokine production and stimulation of local and systemic immune responses, which creates a microenvironment supporting epidermal hyperproliferation and wound response. Conversely, prolonged or aberrantly initiated lymphocyte activation and/or keratinocyte proliferation disturb epidermal homeostasis, and cause or aggravate barrier defects. Thus, determining the contributions from epithelial and immunologic components is critical for understanding the molecular mechanisms underlying development and diseases of the skin.
Serum response factor (SRF) transcription factor regulates the growth and differentiation gene expression programs (
). SRF is activated through the mitogen-activated protein kinase or RhoA pathways, and elicits rapid transcriptional response through the regulation of transcription factors, signaling proteins, and cytoskeletal components (
). Recently, several lines of epidermal-specific SRF knockout mice have been independently reported to display a spectrum of epidermal phenotypes including keratinocyte hyperproliferation (
Here we further explore the role of SRF in regulating epidermal development and function using a skin-grafting approach in addition to a conditional knockout model. We report several important findings. First, the differentiation/proliferation defects observed in SRF mutant skin are intimately coupled with the activation of immune/inflammatory response, as SRF-null epidermis develops normally when grafted onto nude mice. Second, SRF is required for proper development of hair follicles (HFs), and this requirement is intrinsic to epidermis and does not depend on lymphocytes. Finally, we demonstrate several genes/pathways potentially involved in mediating the SRF-regulated downstream events, and reveal a repressor role for SRF in regulating the expression of a subset of its target genes in epidermis.
Results
Characterization of mice lacking Srf in epidermis
SRF is expressed in neonatal mouse interfollicular epidermis and HFs (Supplementary Figure S1a online). Conditional mutants with epidermal-specific Srf ablation K14-Cre+/−;Srffl/fl (Srf-cKO) (Supplementary Figure S1b online), generated by crossing mice homozygous for a floxed Srf allele (Srffl/fl) (
), were born at expected Mendelian ratio. The efficient epidermal Srf deletion was confirmed by immunoblot (Figure 1e). Srf-cKO mutants lacked visible skin lesions including scaling, peeling, or erosions, but exhibited wrinkled skin, eye-open-at-birth phenotype (Figure 1a; Supplementary Figure S1c and d online), malformation of hind limbs, including splaying and hemorrhaging of the digits (Figure 1b, lower panel; Supplementary Figure S1e online), and lack of whiskers (Figure 1b, upper panel). Mutants died or were eliminated by their mothers within 6hours of natural birth. Upon E18.5 caesarean delivery, unfed Srf-cKOs survived for at least 31hours at room temperature, and died of apparent dehydration (Figure 1d).
Figure 1Phenotypic and biochemical changes in Srf-cKO epidermis (epi). (a) Postnatal day 0 (P0) wild-type (wt) and mutant (mt) mice. (b) EOB phenotype, lack of whiskers, and abnormal hind limb digits in mutants. E18.5, embryonic day 18.5. (c) Acidic X-gal dye penetration assay. (d) Dehydration assay. Steady weight loss exhibited by mutants but not by wt control littermates, suggesting defective “inside-to-outside” barrier in mutants. (e) Immunoblot confirms serum response factor (SRF) ablation in mutant epidermis. Tubulin immunoblot verifies equal protein loading. (f) H&E and (g–m) immunofluorescent staining of E17.5 wt and mt littermate skin sections, using the indicated antibodies, color coded according to secondary antibodies. Secondary antibody alone was used as a negative control for each primary antibody (not shown). Nuclei were counterstained with Hoechst dye. βcat, β-catenin; der, dermis; Intβ4, integrin β4; Ker, keratin; Lor, loricrin. Bar=50μm.
). Srf-cKO mutants showed normal ability to exclude penetration of X-gal through their skin, except around digits of hind limbs (Figure 1c), indicating that the “outside-to-inside” barrier was not generally affected by epidermal Srf loss. Hourly assessment of body weight of unfed E18.5 mice revealed a steady weight loss in mutants, but not in wild-type littermate controls, throughout the measurements period (within about 30hours after birth) (Figure 1d), indicating a possibility of increased transepidermal water loss because of impaired “inside-to-outside” epidermal barrier function in mutants. Indeed, Srf-cKO skin acquired a markedly desiccated appearance during this period (not shown). These data suggest that mutant mice die perinatally from dehydration because of a failure to acquire a functional epidermal permeability barrier. Similar perinatal lethalities have been reported for other barrier-deficient mouse models (
Epidermal Srf loss resulted in an abnormal epidermal architecture with marked hyperplasia and a lack of a discernible granular layer (Figure 1f). Srf-cKOs displayed microblistering phenotype (Supplementary Figure S1f online). E17.5 mutant epidermis exhibited abnormal hyperproliferation in the basal layer as evidenced by an increased number of BrdU-positive basal keratinocytes (a mean of 58±5% of BrdU-positive cells in mutant, versus 19±4% in control, per field at × 40 magnification; n=3 embryos, four to six fields per embryo for each genotype; P=1.37 × 10−9; Figure 1g), increased number of keratin 14-expressing layers (Figure 1g), and upregulation of keratins 6 and 17, the indicators of perturbed epidermal homeostasis (Figure 1l; Supplementary Figure S2c online). In addition, extensive apoptosis was detected in suprabasal layers of SRF-null epidermis (Figure 1h). Loss of Srf led to perturbed expression patterns of terminal epidermal differentiation markers, including loricrin, keratin 10, and involucrin (Figures 1i and 2a,b).
Figure 2Abnormal expression of keratin 10, involucrin, and p63 in the epidermis of Srf-cKO mice. Immunohistochemical analysis of wild-type (wt) control and SRF-null mutant (mt) postnatal day 1 (P1) mouse skin sections for (a) keratin 10 and (b) involucrin expression. Positive staining is shown by brown staining. Secondary antibody alone was used as a control (No 1° Ab). Original magnification, × 40. (c) Immunofluorescent staining using antibodies specific for the indicated proteins in wt control and SRF-null mt P1 mouse skin sections. Secondary antibody alone was used as a negative control for each primary antibody (not shown). Nuclei were counterstained with Hoechst dye (DNA). Bar=50μm.
SRF-null E17.5 skin showed increased density of infiltrating CD45+ cells in the dermis, indicating increased inflammation compared with wild-type littermate control skin (Figure 1j). Furthermore, Srf loss in keratinocytes caused cortical actin cytoskeleton disruption, abnormal increase in suprabasal expression of hemidesmosomal integrin β4 (Figure 1k), and decrease in suprabasal β-catenin expression at cell–cell contacts (Figure 1m).
The expression patterns of transcription factors implicated in control of epidermal gene expression are perturbed in mutant epidermis
Srf-cKOs displayed perturbed expression patterns of transcription factors implicated in the regulation of epidermal transcription and terminal differentiation, including ΔNp63α, AP1, and C/EBP. We observed a partial loss of ΔNp63α from the nuclei of basal keratinocytes in mutant epidermis (Figure 2c). Srf ablation led to altered expression and localization of Fra-1, JunB, and c-Jun AP1 family members, and C/EBPα transcription factor (Figure 3). As reported (
Differential expression of the fos and jun family members c-fos, fosB, Fra-1, Fra-2, c-jun, junB and junD during human epidermal keratinocyte differentiation.
), in wild-type epidermis, Fra-1 expression was cytoplasmic and involved all epidermal layers, except basal. In contrast, in Srf-cKO, Fra-1 was localized mainly to basal nuclei (Figure 3a). In wild-type epidermis, JunB was expressed in the suprabasal nuclei, whereas in mutant staining was more intense in the basal nuclei (Figure 3b). c-Jun expression was observed in the nuclei of basal and suprabasal keratinocytes in wild-type, but exhibited intermittent appearance in mutant epidermis (Figure 3c). C/EBPα was highly expressed in suprabasal, and weakly in some basal keratinocytes in wild-type epidermis, as reported by
Expression of CCAAT/enhancer binding proteins (C/EBP) is associated with squamous differentiation in epidermis and isolated primary keratinocytes and is altered in skin neoplasms.
, and the expression was markedly reduced in SRF-null epidermis (Figure 3d).
Figure 3Aberrant patterns of expression of the AP1 family members Fra-1, JunB and c-Jun, and C/EBPα transcription factors in epidermis (epi) lacking serum response factor. Immunofluorescence (upper panel) and immunohistochemical analyses of wild-type (wt) and Srf-cKO mutant (mt) E18.5 mouse skin sections for the expression of the indicated transcription factors. Nuclei were counterstained with Hoechst dye (upper panel). Positive staining for JunB, c-Jun, and C/EBPα is shown by brown staining. Secondary antibody alone was used as a negative control for each primary antibody (not shown). der, dermis. Bar=50μm.
Our analyses showed that the downgrowth of HFs in mutants was apparently less extensive than that of littermate controls (Figure 1f–m). We next performed alkaline phosphatase staining on E18.5 control and Srf-cKO skin sections to determine the position of dermal papilla cells, which mark the end of HF. Indeed, the alkaline phosphatase–stained dermal papilla cells in mutants are much closer to epidermis than their wild-type counterparts (Supplementary Figure S3a online), especially considering that Srf-cKO epidermis is thicker than wild type. In addition, the overall number of HFs in E18.5 mutant skin is 25% lower relative to that in controls (Supplementary Figure S3b online) and HFs in mutant skin are at the earlier developmental stages compared with those in their wild-type counterparts (Supplementary Figure S3c online).
Engrafted SRF-null skin lacks hair but exhibits normal epidermal architecture and proper expression of differentiation markers
Srf-cKO embryonic and neonatal skin exhibited typical features of human inflammatory skin pathology including barrier breach, hyperkeratosis, and enhanced lymphocytic infiltration. However, these phenotypes are often interconnected and each feature might be a part of the positive-feedback loop. Thus, we sought to graft full-thickness skins from E17.5 mutants and wild-type control littermates onto the backs of nude mice, to distinguish the contribution from epithelial compartment and infiltrated lymphocytes, and to circumvent early neonatal lethality of mutants. We observed robust hair growth in wild-type grafts throughout the course of the experiment (14 weeks). In contrast, SRF-null mutant grafts showed very limited and abnormal hair growth at all stages examined during the same period (Figure 4a).
Figure 4Engrafted serum response factor–null skin lacks hair but exhibits normal epidermal architecture and proper expression of differentiation markers. Wild-type (wt) and mutant (mt) E17.5 skins were engrafted onto the backs of nude mice. (a) General appearance of the grafts at the indicated days post grafting. (b) Histological and immunohistochemical analyses, using the indicated antibodies, of the 97-day (97-d) grafts. Positive staining is shown by brown staining. (c) Immunofluorescence analysis of TSLP expression. Secondary antibody alone was used as a negative control in both b and c (No 1° Ab). (d) TUNEL assay reveals no TUNEL-positive cells in the 97-d wt or mt grafts. In contrast, E17.5 mt sample used as a positive control shows numerous TUNEL-positive nuclei in epidermis. Nuclei were counterstained with Hoechst dye. d, days; Inv, involucrin; Ker, keratin; Lor, loricrin. (b–d) Bar=50μm.
As nude mice also exhibit defective HF development, we evaluated whether grafted epidermis is solely derived from SRF-null cells. In nude mice, a well-characterized mutation in FoxN1 gene results in a disruption of FoxN1 nuclear localization signal preventing a truncated protein from entering the nucleus (
). Conversely, Srf-null mutants carry wild-type FoxN1 alleles, which produce protein localized to nucleus. We detected nuclear FoxN1 expression in all epidermal cells from grafted mutant skin, but not in host nude mouse skin (Supplementary Figure S4 online), indicating that grafted epidermis is solely derived from SRF-null cells.
Srf-cKO grafts displayed largely normal epidermal architecture similar to that of grafted skin from control mice, apart from the lack of HFs (Figure 4b). Intriguingly, the expression patterns of differentiation markers involucrin and loricrin, both dysregulated in Srf-cKO skin at late embryonic stages, were indistinguishable between grafted skin of different genotypes (Figure 4b). Moreover, keratin 6, a perturbed epidermal homeostasis indicator, highly induced in E17.5 mutant epidermis (Figure 1l), was undetectable in Srf-cKO graft (not shown). Similarly, expression of TSLP, a sensitive measure of the barrier integrity (
), was upregulated in embryonic but absent from grafted mutant skin (Figure 4c). In addition, increased epidermal thickness and apoptosis, the prominent features of Srf-cKO embryonic epidermis, were not observed in grafted mutant skin (Figure 4d). Collectively, no difference in epidermal histology or marker expression was detected between mutant and control grafted skin. These data are in contrast with previous findings showing severe psoriasiform changes in SRF-deficient young postnatal mice (
), and suggest that inflammatory activity has a key role in mediating the progression of the phenotype in SRF-null skin. Conversely, the lack of hair in grafted skin indicates that HF deficiency is a manifestation of intrinsic epidermal anomalies in the absence of SRF, and not the result of inflammatory or immunological processes.
Global gene expression changes in SRF-null epidermis
To explore the molecular mechanisms underlying the effects of SRF loss on epidermal homeostasis and barrier acquisition, we conducted microarray studies using E17.5 SRF-null and wild-type control epidermal RNA samples. Our analysis identified 1124 and 342 probe sets upregulated and downregulated, respectively, by more than 2-fold in mutant samples (P<0.1; Supplementary Table S1 online). Figure 5a and b shows selected genes differentially expressed in mutant epidermis, grouped into functional categories. Changes in expression of selected genes were validated by reverse transcriptase–PCR, in situ hybridization, western blot, or immunofluorescence (Supplementary Figure S2 online; Figure 1i, k, data not shown).
Figure 5Changes in the expression of selected genes in serum response factor (SRF)–null epidermis. Selected genes (a) upregulated or (b) downregulated in mutant E17.5 epidermis, grouped into functional categories essential for skin barrier function. Numbers in parenthesis correspond to fold (a) increase or (b) decrease in Srf-cKO mutant versus wild-type (wt) control epidermis. P-value is less than 0.1. (c) Heat map shows both downregulation and upregulation of SRF target CArG-containing genes. A summary of SRF target genes, containing known and novel CArG sequences, generated using bioinformatics screen was published by
. About 60% of the putative SRF target genes have been functionally validated in that study. *Known CArG-containing SRF target genes, as reported in the aforementioned study.
Notably, microarray data revealed that many purported SRF target genes are markedly upregulated in SRF-null epidermis (Figure 5c), suggesting a potential repressor function for SRF for a subset of its epidermal targets (
). As expected, the CArG-containing genes, that is, target genes for the MADS transcription factor family, are the most prominent ones in the downregulated category (Supplementary Table S2 online). In contrast, the reticuloendotheliosis viral oncogene homologs (REL) family targets are the most prominent among the upregulated genes (Supplementary Table S2 online), reflecting the inflammatory and immune responses elicited in SRF-null epidermis and the role of NFκB in this process (
Analysis of the ontological categories of regulated genes pinpointed a Chromosome 3 cytoband 45.2cM locus as highly enriched (Figure 5a, Supplementary Table S3 online); this region contains 11 small proline-rich proteins, the cornified envelope components. This upregulation may be responsible for the efficient “outside-to-inside” barrier of SRF-null epidermis. The “keratinization” category also contains keratin genes 6 and 17, markers of hyperproliferative epidermis (
Transient synthesis of K6 and K16 keratins in regenerating rabbit corneal epithelium: keratin markers for an alternative pathway of keratinocyte differentiation.
), but not any of the paradigmatic differentiation markers, such as keratins 1/10, filaggrin, loricrin, or involucrin. The genes associated with wounding and inflammatory responses are prominently upregulated in SRF-null epidermis (Figure 5a; Supplementary Tables S3 and S4 online). tumor necrosis factor-α, one of the primary upregulators of inflammation and inflammation-associated genes in human epidermis (
), is among the induced genes in this category (Supplementary Table S4 online). The most significant categories among the downregulated genes include amino acid metabolism, lipoproteins, and lipid metabolism (Supplementary Table S3 online). This downregulation of lipid metabolism and lipoproteins may be responsible for the defective “inside-to-outside” barrier and water loss of SRF-null mutants.
Importantly, we detected the downregulation of the genes linked to HF development and differentiation, including Wnt/β-catenin pathway members, Shh pathway member Ptch1, hair-associated keratins Krt32 and Krt34, and stem cell-associated Krt15, in E17.5 mutant epidermis (Figure 6a). These findings were confirmed by in situ hybridization (Lef1, Ptch1, and Shh, Figure 6b, d) and immunohistochemistry (Lef1, Figure 6c). Notably, Lef1 expression in mutant E15.5 epidermis was similar to that in littermate control epidermis (Supplementary Figure S5 online). Conversely, the expression of negative regulators of Wnt/β-catenin signaling, such as Wif1, Sfrp1, and Gsk3b, was upregulated in mutant E17.5 epidermis (Figure 6a). Notably, the promoters of Lef1, Fzd2, and Ptch1 genes contain conserved CArG boxes, and the promoters of Krt34, Krt32, Gata3, and Edar genes contain conserved CArG box–like sequences (a consensus sequence with one base-pair deviation) (
), indicating a potential for SRF to directly bind and transactivate the transcription of these genes. Importantly, we observed a decrease in the Lef1 promoter-luciferase reporter activity in cultured mouse keratinocytes upon endogenous SRF depletion (Supplementary Figure S6a and c online), indicating a downregulation of Wnt activity. In contrast, activation of the SRE-luciferase reporter (a construct that contains a multimerized CArG box and adjacent E twenty-six binding site) was increased following SRF depletion (Supplementary Figure S6b online), supporting the notion that SRF can function as a cell autonomous transcriptional repressor of certain target genes in epidermal keratinocytes (Figure 5c).
Figure 6Serum response factor(SRF)–null E17.5 epidermis exhibits reduced expression of Lef1, Ptch1, and Shh. (a) Changes in the expression of selected genes linked to hair follicle morphogenesis in SRF-null E17.5 mouse epidermis. Numbers in parenthesis correspond to fold change in Srf-cKO mutant (mt) versus wild-type (wt) control epidermis; negative values represent fold decrease. Fold change>1.5, P-value cutoff: 0.1. (b, d, and e) 35S in situ hybridization on wt and mt E17.5 mouse skin sections using indicated probes. Nuclei were counterstained with Hoechst dye (blue). (c) Immunohistochemical analysis of wt and mt E17.5 mouse skin sections for Lef1 expression. Positive staining is shown by brown staining. Secondary antibody alone was used as a negative control for each primary antibody (not shown). (b–e) Bar=200μm.
Here we analyzed the role of SRF in epidermal development and function by characterizing epidermal-specific Srf-cKO mutant. In part, our findings are consistent with other studies (
), which demonstrated a marked morphological change in SRF-deficient epidermis. Nonetheless, by showing that SRF-null skin develops normally when grafted onto immune-compromised nude mice, we provide evidence that most of these epidermal changes are dependent on inflammatory activity. In addition, we demonstrate an obligatory role for SRF in the development of HFs. Moreover, global gene expression profiling revealed an unexpected repressor role for SRF in regulating a subset of its downstream epidermal targets. The notion that SRF can function as a cell-autonomous transcriptional repressor is supported by our finding that the activity of SRE promoter is upregulated following SRF depletion in cultured keratinocytes. Alternatively, an antagonistic regulation of distinct subsets of SRF target genes could stem from a recently described negative-feedback circuit between the actin-MAL and the mitogen-activated protein kinase–ternary complex factor pathways, both of which regulate the SRF-dependent transcription (
). Koegel et al. reported a sustained inflammation in young mice with patchy Srf depletion, and proposed that increased keratinocyte proliferation is regulated by lymphocyte activation, whereas adhesion defects reflect keratinocyte SRF cell-autonomous requirement in regulating actin cytoskeleton. Luxenburg et al. reasoned that spindle defects and abnormal mitotic division caused by cortical actin network deficiency are the underlying mechanisms for the proliferation and differentiation defects observed in mutant embryonic epidermis, as these defects occur before the infiltration of immune cells. We studied Srf-cKO skin in both embryonic/neonatal stages and in a skin graft model, and demonstrated that most embryonic anomalies, other than the HF defect, do not persist or progress when grafted onto nude mice. In grafted mutant skin, both proliferation and differentiation are comparable to wild-type controls 14 weeks post grafting. This contrast between grafted skin and skin from neonatal or young mutant mice (
) suggests a role for inflammation in the progression of the psoriasis-like phenotype. Indeed, the largest functional group of genes upregulated in SRF mutant epidermis are those involved in wound response, including many proinflammatory cytokines and chemokines produced by either keratinocytes or infiltrated lymphocytes. It is plausible that SRF mutant skin initially develops mild abnormalities including cell shape changes reported by
. In turn, keratinocytes sense these changes and initiate wound response, which sustains and aggravates abnormal epidermal phenotype. The induction of proinflammatory cytokines, notably tumor necrosis factor-α, and the activation of REL family transcription factors (Supplementary Tables S3, S4 online) support this idea. Whether the barrier disturbance is the cause or the result of immune activation remains to be determined. Future experiments crossing Srf-cKO to Rag1-deficient mutant lacking mature B and T lymphocytes will provide insight into this question. Epithelial–mesenchymal interaction might also contribute to abnormal Srf-cKO epidermal phenotype offering additional levels of complexity.
SRF in mediating HF development
Defective HF development in Srf-cKO mutants reflects an autonomous requirement for SRF in epidermis as grafted mutant skin is largely hairless. In addition, Srf-cKOs displayed impaired HF downgrowth and a reduction in the overall number of follicles. E15.5 mutant skin exhibited normal HF structure and marker gene expression, indicating normal primary HF induction. However, at a later time point, mutant follicular cells showed decreased signaling activity and their development was arrested. The cluster of HF defects in Srf-cKO likely reflects a dynamic role for SRF in different cell types at different developmental stages. The molecular mechanisms underlying these defects remain to be further elucidated, as the current model does not allow us to dissect the function of SRF in different hair compartments at different developmental stages. Nonetheless, we have shown that the activity of the Wnt pathway is downregulated in Srf-cKOs. Canonical Wnt signaling is critical for HF development (
). Our data show that canonical Wnt ligands, mediators, and downstream targets are downregulated in Srf-cKO, and several Wnt inhibitors are upregulated, suggesting a specific inhibition of the pathway. Moreover, Wnt transcriptional activity is downregulated upon SRF depletion in cultured keratinocytes. In addition, a reduced expression of other key regulatory genes including Shh/Ptch1 and Edar/EdarADD are also noted in the mutant. It is plausible that the alteration of these important pathways contributes to the Srf-cKO HF phenotype. Hedgehog signaling pathway is of particular interest, as Shh is required for HF downgrowth (
). The reduction in Wnt activity and downregulation of Lef1 correlates well with the partial loss and developmental arrest of the HFs in Srf-cKO. Whether it also affects stem cell compartment and hair cycling remains to be probed.
The HF is a complex structure that contains multiple cell types and undergoes cyclic regeneration. Therefore, the role for SRF in HF development is likely to be dynamic and context dependent. Both Shh and Lef1 are mainly expressed in the matrix cells, a transit amplifying cell population giving rise to the hair shaft and the inner root sheath, which prompted us to test the function of SRF specifically in these cell populations using a previously described Doxycycline inducible matrix cell–specific Msx2-rtTA;tetO-Cre system (
). However, deletion of SRF at P0 caused only mild hair abnormalities in the first and second hair cycle (Supplementary Figure S7 online), indicating that a defect in matrix cells alone is not sufficient to cause a hairless phenotype, and suggesting potential key functions for SRF in other follicular cell types especially the outer root sheath, including bulge stem cells. Future analyses using cell type–specific conditional knockout approaches need to be performed to further dissect the function of SRF in different hair compartments.
Materials and Methods
Skin-targeted SRF knockout mice
Animal studies were approved by WUSM Animal Studies Committee. Srffl/fl (
). Briefly, unfixed newborn mice were rinsed in phosphate-buffered saline (PBS) and stained at 37oC for 12hours in X-gal solution (pH 4.5). The samples were fixed post staining in 4% paraformaldehyde in PBS at room temperature for 1hour, washed in PBS, and photographed with a digital camera. Dehydration assay was performed as previously described (
). Briefly, we monitored body weights of unfed E18.5 wild-type control and Srf-cKO mutant littermates hourly during the evaluation period, as shown in Figure 1d. The results are presented as a percentage of the initial weight.
Histology, immunofluorescence, and immunohistochemistry
Histology, immunofluorescence, and immunohistochemistry were performed as detailed in the Supplementary Materials online.
In situ hybridization
In situ hybridization was performed as described (
Female athymic nude mice recipients (Charles River) were anesthetized, 1cm2 piece of skin per graft was removed. E17.5 Srf-cKOs and littermate control donors were killed, similar-sized pieces of back skin removed and grafted onto recipients.
ACKNOWLEDGMENTS
We thank Raphael Kopan, Jeffrey H. Miner, and Shadmehr Demehri for helpful discussions and critical reading of the manuscript; Meei-Hua Lin for helpful discussions and technical advice; and Erin L. Gribben for mouse husbandry. This work was financially supported by the Division of Dermatology, Department of Internal Medicine, Washington University School of Medicine (T Efimova).
Expression of CCAAT/enhancer binding proteins (C/EBP) is associated with squamous differentiation in epidermis and isolated primary keratinocytes and is altered in skin neoplasms.
Transient synthesis of K6 and K16 keratins in regenerating rabbit corneal epithelium: keratin markers for an alternative pathway of keratinocyte differentiation.
Differential expression of the fos and jun family members c-fos, fosB, Fra-1, Fra-2, c-jun, junB and junD during human epidermal keratinocyte differentiation.
36156-X&id=gr1.jpg){kind=link}
36156-X&id=gr2.jpg){kind=link}
36156-X&id=gr3.jpg){kind=link}
36156-X&id=gr4.jpg){kind=link}
36156-X&id=gr5.jpg){kind=link}
36156-X&id=gr6.jpg){kind=link}