Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Dec 14.
Published in final edited form as: J Invest Dermatol. 2008 Oct 16;129(4):984–993. doi: 10.1038/jid.2008.318

Analysis of the temporal requirement for Eda in hair and sweat gland development

Chang-Yi Cui 1, Makoto Kunisada 1, Diana Esibizione 1,2, Eric Douglass 1, David Schlessinger 1,*
PMCID: PMC5155335  NIHMSID: NIHMS833044  PMID: 18923450

Abstract

EDA signaling plays a pivotal role in skin appendage initiation. Its possible involvement in appendage subtype determination and post-induction stage appendage development, however, has not been studied systematically. To address these issues we manipulated Eda-A1 transgene expression in a tetracycline-regulated conditional mouse model, where the transgene is the only source of active ectodysplasin (Eda). We find that Eda-A1 restores sweat glands and all hair subtypes in Tabby, but each requires its action at an idiosyncratic time of development: by E17 for guard, by E19 for awl, and starting at E18 for zigzag/auchen hair. Guard and awl hairs were indistinguishable from their wild-type counterparts; but restored zigzag and auchen hairs, while recognizable, were somewhat smaller and lacked characteristic bends. Notably, secondary hair follicle formation of awl, auchen, and zigzag hairs required higher Eda-A1 expression level than did guard hair or sweat glands. Furthermore, Eda-A1 expression is required until the early dermal papilla stage for guard hair germs to make follicles, but is dispensable for their maturation. Similarly, sweat gland pegs require Eda-A1 at an early stage to form mature glands. Thus we infer that EDA signaling is needed for the determination and development of various skin appendages at spatiotemporally restricted intervals.

Keywords: Ectodysplasin, EDAR, hair follicle subtype, sweat glands, Shh

Introduction

The development of skin appendages can be divided into three phases: induction, progression and maturation. Skin appendage development is achieved by well-controlled reciprocal signaling between mesenchyme and epithelium (Hardy, 1992). General morphogenetic signaling pathways play a key role. Canonical Wnt and BMP pathways were shown to be the first mesenchymal initiation signals for appendage induction (reviewed in Millar, 2002; Botchkarev and Paus, 2003), and Shh was suggested to be the secondary epithelial signal for appendage germ progression and subsequent dermal papilla formation (St-Jacques et al., 1998; Chiang et al., 1999). Notch, Whn and Hox genes also participate in hair follicle maturation (Millar, 2002).

In addition to more general signaling pathways, a skin-restricted critical role in skin appendage initiation is played by ectodysplasin, the product of the EDA gene (Kere et al., 1996; Srivastava et al., 1997). Guard hair and sweat gland germs are not formed in Eda mutant Tabby mice, but were restored by an Eda-A1 transgene or recombinant ectodysplasin (Cui et al., 2003; Gaide and Schneider, 2003; Mustonen et al., 2004). The EDA pathway genes EDA, EDAR and EDARADD are all expressed in the epithelial portion of skin appendages, where they activate NF-kBs (reviewed in Cui and Schlessinger, 2006). The EDA pathway has therefore been suggested to function immediately downstream of the mesenchymal initiation signals, regulating appendage germ formation by modulating more general morphogenetic signals. Consistent with this notion, EDA was shown to operate downstream of inductive Wnt signaling (Durmowicz et al., 2002; Laurikkala et al., 2002). Further supporting evidence has come from findings that in contrast to mice lacking Wnt, in which hair follicles were not initiated, mutant mice lacking the ectodysplasin receptor “Edar” start to form guard hair follicles, but fail to make functional hair germs (Andl et al., 2002; Schmidt-Ullrich et al., 2006).

Further action of Eda later in skin appendage development was inferred from recent findings that EDA pathway genes are highly expressed in developing hair follicles, and Shh, the critical regulator for skin appendage progression, is the most prominent target of EDA signaling (Cui et al., 2006; Schmidt-Ullrich et al., 2006). Observations of mutant Tabby mice also implied that Eda signaling is involved in hair subtype determination (Vielkind and Hardy, 1996). However, remaining questions include: 1) is there a time window for the action of EDA for each type of skin appendage? and 2) is there a differential dose dependence on EDA for the determination of numbers or subtype of hair or sweat glands during development?

Skin appendages certainly develop along defined time courses and patterns. In general, guard hair develops starting around E14.5; awl hair somewhat later around E16.5; and zigzag hair around E18.5 in mouse back skin (Vielkind and Hardy, 1996). Follicle down-growth is then complete by postnatal day 8 (Blanpain and Fuchs, 2006). To assess possible time windows of EDA action in appendage subtype determination more precisely, we generated a tetracycline-regulated conditional Eda-A1 transgenic mouse model (Cui et al., 2003). Using this model, we show that Eda-A1 signaling at defined times is sufficient to specify the determination of hair subtypes and the initiation and progression, but not maturation, of hair follicles and sweat glands. We also find that the development of some skin appendages needs relatively higher levels of Eda-A1 activity.

Results

To gauge the involvement of EDA signaling at different developmental phases of hair follicles and sweat glands, we used a tet-regulated Eda-A1 transgenic mouse model to selectively turn the transgene on or off in Tabby background at different embryonic stages, and scored the mice at age 2 months for hair follicle subtypes and sweat gland formation (Figure S1a, and see Materials and Methods).

Doxycycline regulates Eda-A1 transgene expression and transgene-induced skin appendage phenotypes in a tet-regulated conditional mouse model

The conditional transgenic mice bear two transgenes in their genome: an MMTV-driven regulatory tet-off gene that constitutively produces a transactivator tTA; and a tTA responsive TRE-Eda-A1 transgene encoding the mouse Eda-A1 isoform (Cui et al. 2003). The tTA binds to TRE and activates Eda-A1 transgene transcription only when doxycycline is absent. Therefore, transgene expression can be selectively turned on or off at any time by doxycycline food in vivo. The MMTV-driven tet-off gene is highly expressed in skin epidermis and various secretory tissues, and therefore provides an excellent system for skin and skin appendage studies (Hennighausen et al., 1995).

We confirmed that doxycycline food completely blocks skin appendage phenotypes in Eda-A1 transgenic Tabby mice (see Materials and Methods). The tight regulation of Eda-A1 transgene expression and transgene induced skin appendage phenotypes by doxycycline food allowed us to carry out the experiments described below.

Guard hair germs develop to guard hair follicles if Eda-A1 is still active after germ formation

Requirement of EDA signaling for guard hair follicle initiation has been clear from the observations that the first wave of hair follicle formation, which generates guard hair germs at embryonic day 14.5 (E14.5), is absent in Tabby mice and is restored by an Eda-A1 transgene (Cui et al., 2003; Mustonen et al., 2004).

When the transgene is turned off at E15, just after follicle germs had formed (E15-off mice, Figure S1a), the adult mice were slightly darker (see below) but otherwise identical to Tabby; they failed to restore guard hair, as shown by microscopic observation (Fig. 1a-c). These results suggested that EDA signaling is required both at induction and at post-induction stages of guard hair development.

Fig. 1.

Fig. 1

Open in a new tab

Restoration of guard hair in conditional Eda-A1 transgenic mice. a, Guard hair, which has two rows of medullary granules, was restored in TaTG and in E17-off (and E19-off, not shown) conditional transgenic mice. b, The contribution of restored guard hair to total hair number in TaTG, E17-off and E19-off mice was comparable to wild-type mice. No guard hair was restored in E15-off or E18-on mice. c, E15-off mice were similar to Tabby; the hair coat of E17-off mice was darker and bald patches behind ears had been filled in; and E19-off mice were very similar to TaTG mice in which the transgene was expressed throughout life. E18-on mice, in which the transgene was only activated at that time, were similar to Tabby, though some hairs were slightly darker and thinner.

To assess the time and duration of the window for the Eda-A1 requirement, we turned off the transgene at successively later times. If the transgene was turned off at E17 (E17-off), when the early dermal papillae form, or at E19 (E19-off), when the hair shafts are nascent, the mice were still able to fill in bald patches that are evident behind the ears of Tabby mice (Fig. 1a; cf. Figure S1a). Microscopic analysis of hair from adult animals revealed fully restored, structurally normal guard hair with two rows of medullary granules in both the E17-off and E19-off mice (Fig. 1b). Its representation in total hair was comparable to that of wild-type (Fig. 1c). Also, when Eda-A1 was expressed until E19, the guard hairs lay at a more vertical angle compared to wild-type (similar to mice in which the Eda-A1 transgene was expressed throughout life (“TaTG mice”) (Fig. 1a, and Cui et al., 2003). The length of restored guard hairs in conditional mice was comparable to wild-type or TaTG mice, though slightly shorter in E17-off mice (Figure S1b). Therefore we infer that EDA signaling is required until the formation of early dermal papilla, but not for the subsequent full development of guard hair.

A few minor differences were noted. In E15-off and E17-off mice, there were also about 10% of dark hairs that are structurally identical to abnormal awl-like Tabby hair, and which may contribute to the somewhat darker coat appearance in these mice (Fig. 1a and data not shown). A similar phenotype was also observed in Tabby mice injected with recombinant ectodysplasin protein (Gaide and Schneider, 2003).

In a control experiment, we turned on the transgene only at E18 (E18-on mice). In that case no guard hair follicles were restored, as expected (see below). Therefore, Eda-A1 cannot rescue guard hair after the critical period at E15-E17, the guard hair induction and progression stages.

Eda-A1 transgene is able to restore normal awl hair in Tabby mice

The second wave of hair follicle formation, resulting eventually in awl hair, starts to form follicles around E16.5. Mature awl hair is characterized by the presence of 3 to 4 rows of medullary granules (Fig. 2a). This wave occurs even in the absence of Eda, but the hair produced is abnormally straight and short, as seen in Tabby mice. Further, the hairs are characterized by the presence of 1 or 2 rows of medullary granules arranged in an alternating pattern, and the hairs are much thinner (Fig. 2a).

Fig. 2.

Fig. 2

Open in a new tab

Restoration of awl hair by Eda-A1. a, In contrast to Tabby hair, restored awl hairs in TaTG and E19-off mice, as in wild-type mice, have 3 rows of medullary granules. b, The contribution of restored awl hair in TaTG and E19-off mice was similar to wild-type. Awl hair was not restored in E15-off, E17-off or E18-on mice.

Previous studies have suggested that an Eda-A1 transgene or its cognate recombinant protein is sufficient to restore guard hair in Tabby mice, but that awl or zigzag hairs are not restored (Cui et al., 2003; Gaide and Schneider, 2003). By more extended microscopic analysis, however, we now find that most awl hair in TaTG mice is restored to show 3 rows of medullary granules (Fig. 2a); and as in wild-type mice, the restored awl hair constitutes about 10% of total hair in TaTG mice (Fig. 2b). Thus an Eda-A1 transgene is sufficient to restore awl hair.

The contribution of Eda-A1 was again time-sensitive. No awl hair was restored in E15-off and E17-off mice; however, in E19-off mice we found about 10% awl-like hair with 2-3 rows of medullary granules, structurally similar to the restored awl hair in TaTG mice (Fig. 2a, b, “E19-off”). These results suggested that as in the case of guard hair follicles, EDA signaling is required for both initiation and progression of awl hair follicles.

Restoration of zigzag hair in Eda-A1 transgenic Tabby mice

A third wave of follicle formation generates zigzag hair, the most frequent subtype in mice, which is completely missing in Tabby mice. Consistent with previous reports, no obviously zigzag hairs were found in TaTG mice, grossly or under the dissection microscope (Cui et al., 2003; Gaide and Schneider, 2003). However, once again, under higher magnification we found a large number of hairs with zigzag hair characteristics (Fig. 3a-c), including a single row of medullary granules and 2 or 3 putative bending (constriction) sites where medullary granules were absent (Fig. 3b). This class constituted up to 62.9 (± 5.8) % of the total hair (Fig. 3d, TaTG). Apparently, these are partially restored zigzag-like hairs in TaTG mice, though slightly thinner and shorter than the normal equivalent and lacking obvious bends at the “bending sites” (Fig. 3b).

Fig. 3.

Fig. 3

Open in a new tab

Restoration of zigzag and auchen hair in Eda-A1 transgenic mice. a, Abnormal awl-like Tabby hair is straight and has one or two layers of medullary granules in a staggered pattern. b, Restored zigzag hair in transgenic Tabby mice (TaTG) has one layer of medullary granules and two putative bending sites where the medullary granules are absent (higher magnification); however, no bends are seen. c, Wild-type zigzag hair is like TaTG hair, but shows the characteristic bends. d, Zigzag hair was restored in TaTG mice and E18-on mice, but not in conditional transgenic mice in which the transgene was turned off earlier or later. e, Auchen-like hair was restored only in E18-on mice.

To assess when EDA signaling is required for zigzag hair formation, we turned off the transgene at E15 and E17, well before zigzag hair germ formation. No zigzag-like hair at all was restored (Fig. 3d, E15-off and E17-off). Also, when the transgene was turned off after zigzag hair germ formation had began, at E19, neither normal zigzag nor zigzag-like hairs were formed (Fig. 3d, E19-off).

To see if a sharper “pulse” of Eda-A1 at the induction phase might fully restore the morphology of zigzag hair, we turned on the transgene at E18 (E18-on mice), just before the start of zigzag hair formation. In these conditions, the genetic program for proper guard hair and awl hair formation were suppressed, so that the E18-on mice lack guard and awl hairs (Fig. 1c, 2b). Instead, a large number of zigzag-like hairs, structurally identical to those restored in TaTG mice, were formed. The zigzag-like hairs reached up to 53.6% (± 10.5) of total hair, ranging from 40% to nearly 70% (Fig. 3d, E18-on). Correspondingly, the numbers of abnormal Tabby hairs were sharply reduced. It was previously inferred that zigzag follicles are initiated on time, but are unable to make normal zigzag hair in Tabby mice (Hammerschmidt and Schlake, 2007). Our results are consistent with a requirement for Eda-A1 for normal zigzag hair shaft formation but not for zigzag follicle initiation. In addition, in the E18-on mice, the numbers of auchen-like hairs were also restored to the wild-type level of about 12.1% (± 1.7) of the total (Fig. 3e, E18-on).

Thus, zigzag- and auchen-like hairs, though still abnormal, are restored independent of guard and awl hair formation by time-dependent action of Eda-A1. To our knowledge this is previously unreported condition found in which a transgene can restore auchen hair, in keeping with nuanced requirements for Eda action (see Discussion).

Secondary hair follicle development requires higher Eda activity

We have generated two transgenic mouse models that encode the Eda-A1 isoform: conventional Eda-A1 transgenic Tabby mice, in which the transgene was expressed at about the endogenous level of Eda (Srivastava et al., 2001); and a tet-regulated conditional Eda-A1 transgenic Tabby mice that has transgene expression up to 100-fold higher than normal endogenous levels (Cui et al., 2003, 2005). To look for any dosage effect on EDA signaling for hair follicle subtype determination, we reanalyzed hair subtypes in the conventional transgenic mice compared to the conditional TaTG mice. Guard hair was fully restored in both strains. However, in contrast to the conditional TaTG mice, in the conventional transgenic mice awl and auchen hairs were not at all restored and zigzag-like hair was restored to only about 15% of the total population (Table 1). The “zigzag hair” that did form was structurally identical to the imperfect zigzag hair formed in conditional TaTG mice. These data suggested that guard hair and some zigzag hair (and sweat glands, see Srivastava et al., 2001) can be restored even at lower levels of expression of Eda-A1, but full restoration of secondary hair follicles (including awl, auchen and zigzag) requires a higher level of EDA signaling.

Table 1.

Only guard hair was fully restored in conventional Eda-A1 transgenic Tabby mice

Hair subtypes (% in total hair number)
Strains G Awl Au Z Awl-like* Z-like**
Conv-Eda-A1 TG 1.9±0.3 0 0 0 84.1±6.1 14.0±5.9
Ta 0 0 0 0 100 0
Open in a new tab
*

: Represents abnormal awl-like Tabby hair.

**

: Identical to z-like hair restored in conditional TaTG mice (Fig. 3).

Eda-A1 action disrupted bent hair structures in wild-type mice

In a previous study, we noted that bent zigzag and auchen hair were sharply reduced in transgenic mice with constant high expression of Eda-A1 in wild-type mice (WTTG mice; Cui et al., 2003). Upon further analysis we found that the structure and numbers of straight guard and straight awl hairs in WTTG mice were essentially unaffected (Table 2). [Some awl hairs were longer than normal, which may be related to hair cycling time rather than an effect on morphogenesis (Mustonen et al., 2003; Fessing et al., 2006).] By contrast, the structure of bent hair types was severely disrupted in the WTTG mice and we could only classify about 10%; of the individual hairs more than 80% could not be classified into subtypes, consistent with previous reports from other groups (Mustonen et al., 2003; Zhang et al., 2003).

Table 2.

Bent auchen and zigzag hairs were disrupted in WTTG mice

Hair subtypes (% in total hair number)
Strains G Awl Au Z Awl-like Au-like Z-like-B* Z-like-S**
WTTG 2.8±0.6 8.8±0.4 0 0 9.9±2.1 5.2±2.1 36.8±3.8 36.5±6.9
WT 2.2±0.2 11.1±1.2 12.9±1.9 73.8±1.0 0 0 0 0
Open in a new tab
*

: Z-like hair, slightly bent.

**

: Z-like hair, grossly straight.

Of the total restored hairs, 10% resembled abnormal Tabby hair; about 10% were auchen-like hair; 35% were slightly bent zigzag hair [as seen in Shh transgenic mice (Hammerschmidt and Schlake, 2007)]; and 35% were straight zigzag-like hair similar to the zigzag hair restored in TaTG or E18-on mice (Table 2). It may be that the bent hair formation in late embryonic life is more sensitive to dose- and time-dependent action of EDA signaling than is the earlier formation of straight hair (see Discussion).

Only primitive sweat gland-like structures were formed in Tabby mice when Eda-A1 was turned off at the early progression stage

In Tabby mice sweat glands are not formed; but like guard hair they are restored by an Eda-A1 transgene (Cui et al., 2003). We further assessed the possible involvement of EDA during post-induction stages. During normal development, sweat gland germs are discernible at E17.5 in mouse footpads (data not shown). At E19, sweat gland germs are formed in both wild-type and conditional Eda-A1 transgenic Tabby embryos (TaTG, Fig. 4a upper panels). The germs are like stage 3 hair follicles (Muller-Rover et al., 2001), though sweat glands lack dermal condensations/dermal papillae. They further develop to mature sweat glands if EDA signaling is still active (Fig. 4a, middle panels). As expected, when the Eda-A1 transgene was turned off at E15 (data not shown) or E17, before sweat gland germ formation, sweat glands failed to develop (Fig. 4a, lower left panel). If the transgene was turned off at E19 (E19-off), at the early progression stage, primitive sweat gland-like structures were formed but failed to mature further (Fig. 4a, lower middle and right panels). In these sections, a cluster of keratinocyte-like cells in a narrow zone of dermis formed vertically positioned duct-like structures that were associated with connective tissue, including blood vessels deep in the dermis. Consistent with an arrest in development, iodine-starch sweat tests for E19-off mice were negative (data not shown). Thus, as in guard hair formation, Eda-A1 is required for both initiation and post-initiation stage sweat gland development.

Fig. 4.

Fig. 4

Open in a new tab

Mature sweat gland-like structures required Eda-A1 expression in transgenic Tabby mice throughout the early progression stage. a, In Hematoxylin & Eosin stained sections, sweat gland germs were formed in wild-type and TaTG embryos at E19 (upper panels). Sweat glands were clearly not formed in Tabby mice (Ta), but were restored by an Eda-A1 transgene in TaTG mice, in which the transgene was expressed throughout life (middle panels). Sweat glands were not restored in E17-off mice, but primitive sweat gland-like structures were seen in E19-off mice (lower panels). Higher magnification (lower right) showed that these structures consist of keratinocyte-like cells. Scale bars, 100μm. b, Immunofluorescent staining showed that keratin 14 protein (green, with DAPI staining of DNA in blue) was localized in sweat gland germs in wild-type and TaTG embryos at E19, but was absent in Ta (upper panels, arrows indicate sweat gland germs). Keratin 14 was positive in peripheral cells of sweat duct and myoepithelial cells of the secretory portion of wild-type (middle left panel), but was absent in E19-off sweat gland-like structures (center and right middle panels). Keratin 8 (red) was expressed in luminal cells (lower left panel) of sweat glands in adult stage wild-type mice. Cells in sweat gland-like structures in E19-off mice were devoid of keratin 8. Arrows in middle and lower panels indicate sweat gland ducts in wild-type (WT) and sweat gland-like structures in E19-off mice. Scale bars, 50μm.

To characterize the sweat gland-like structures further, we carried out immunostaining with characteristic sweat gland markers. We confirmed that keratin 14 expression was seen in both wild-type and Eda-A1 transgenic Tabby embryos at E19 from basal cells of epidermis to sweat gland germs (Fig. 4b, upper panels, arrows indicate sweat gland germs). In adult mice, K14 was seen in basal cells of epidermis, peripheral cells of sweat ducts, and the outermost myoepithelial cells of the secretory portion of sweat glands in wild-type mice (Fig. 4b, middle panels; see also Langbein et al., 2005). In contrast, no keratin 14 expression was found in the sweat gland-like structures in E19-off mice (Fig. 4b, middle panels). Another keratin family member, keratin 8, is generally restricted in expression to more differentiated luminar cells of sweat ducts and the secretory portion of glands in adult stage wild-type skin (Fig. 4b, lower left panel). It was not yet expressed in wild-type sweat gland germs at E19, and in E19-off mice, no keratin 8 expression developed (Fig. 4b, lower middle and right panels, and data not shown).

Vimentin, a marker for mesenchyme-derived cells, was expressed in dermal fibroblasts, endothelial cells of blood vessels, and the outermost layers of secretory portions of sweat glands in wild-type mice (Figure S2). In the sweat gland-like structures of E19-off mice, however, most cells showed only a background level, with a few positive cells in the periphery (Figure S2). The sweat gland-like structures in the absence of EDA action may thus result from dysplastic differentiation of sweat gland pegs at post-induction stages.

Meibomian and preputial glands were not restored in Eda-A1 transgenic Tabby mice in any condition tested

Among the exocrine glands missing in Tabby mice, sweat glands were fully restored in both conventional and conditional Eda-A1 transgenic mice, but meibomian glands and preputial glands were not (Srivastava et al., 2001; Cui et al., 2003, 2005). Restoration of either gland type was again not seen in any of various conditions of conditional expression in the transgenic mice (Figure S3a, b). Interestingly, Gaide and Schneider (2003) found that injected recombinant ectodysplasin restored meibomian glands, perhaps reflecting differential levels of Eda activity from the two delivery routes.

Discussion

Critical EDA signaling (Kere et al., 1996; Srivastava et al., 1997; Headon et al., 1999, 2001) supplements more general morphogenetic pathways specifically in skin. There it regulates skin appendage development through a canonical NF-kB cascade (Ezer et al., 1999; Elommaa et al., 2001; Cui and Schlessinger, 2006). The requirement for EDA in skin appendage initiation has been well established; here its further involvement at later stages is delineated in a mouse model. In general, an auxiliary or alternative role in the maintenance of skin appendages has not been excluded; but our results are more simply consistent with a continuing requirement for EDA action in later stages of skin appendage formation.

EDA signaling is involved in skin appendage development with defined time windows

Primary guard hair germs are formed at E14.5; engulfment of dermal condensation/dermal papillae by growing hair pegs is discernible at E16.5; and generation of inner root sheath can be recognized at E18.5 in mice. Here we demonstrate that EDA signaling is required for progression as well as initiation of guard hair germs. However, once juvenile dermal papillae are formed, EDA signaling is no longer needed. Similarly, awl hairs were also shown to require EDA signaling for initiation and initial progression. In contrast, auchen and zigzag hairs most likely require EDA during hair shaft formation.

A requirement for EDA signaling at initiation and post-initiation stages is also shown in sweat gland development. Even when we turned off the Eda-A1 transgene slightly later in the early progression stage (E19), only primitive sweat gland-like structures were formed. Thus, the requirement of EDA signaling in sweat gland development is formally very similar to that for guard hair, though at distinct times and in footpads rather than back skin. Both are fully restored in transgenic mice expressing Eda-A1 at low levels (Srivastava et al., 2001).

Consistent with observations in Tabby mice carrying a full-length Eda-A1 transgene (Cui et al., 2003), an injected soluble portion of ectodysplasin ligand linked to an IgG1 Fc domain (Fc:EDA1) decisively restored guard hair, teeth and sweat glands in Tabby mice (Gaide and Schneider, 2003). Interestingly, Fc:EDA1 injected into pregnant dams before or at the sweat gland germ formation stage of embryos was sufficient to restore sweat glands (Gaide and Schneider, 2003). This is somewhat inconsistent with our findings of a requirement for Eda-A1 transgene action at post-induction stages as well. One possible explanation is that the portion of ectodysplasin with an Fc stalk is more stable than native ectodysplasin, and especially with high injection dosages may extend the lifespan of Fc:EDA1 through progression/maturation stages. But the striking ability of Fc:EDA1 to restore sweat glands even after birth, further suggests that sweat gland progenitor cells are susceptible to EDA action for a longer period during development than other skin appendages. This could increase the feasibility of clinical intervention (Gaide and Schneider, 2003; Casal et al., 2007).

Due to the low yield of positive progeny, we have not been able to analyze the fate of guard hair germs formed in E15-off mice, awl hair germs in E17-off mice, and zigzag hair germs in E19-off mice. However, in adult stage E15-off mice, we found no trace of guard hair follicles or structures reminiscent of them. It seems likely that in the absence of EDA signaling, juvenile hair germs apoptose, as observed in Edar “Downless” mutant mice (Schmidt-Ullrich et al., 2006).

EDA signaling is sufficient to restore all hair subtypes in Tabby, but additional subsidiary signals may be required for restoration of bends in hair

In Eda mutant Tabby mice, guard, awl, auchen, and zigzag hair subtypes are all missing, and instead abnormal awl-like straight hair dominates. Previous studies suggested that Eda-A1 transgene restored guard hair and sweat glands but not zigzag hair (Cui et al., 2003; Gaide and Schneider, 2003). However, further analyses here indicate that Eda-A1 is itself sufficient to determine all hair subtypes in mice in a recognizable if not perfect form.

Nevertheless, several hair types are slightly abnormal. In particular, unlike straight hair bent hair was imperfectly restored by Eda-A1. Bent hair formation may be a more delicate process, a notion supported in the case of WTTG mice, in which added levels of Eda even disrupted bent hair formation in a wild-type mouse background (Table 2). Consistent with a need for greater nuance, Shh and IGF signaling regulate axial polarity of hair follicles and zigzag hair bending in a complex way, with periodic asymmetric expression of these pathway genes in the hair follicle matrix (Schlake, 2005; Hammerschmidt and Schlake, 2007). Expression of genes from both those pathways is impaired in Tabby hair follicles, and was partially restored by Eda-A1 transgene (Hammerschmidt and Schlake, 2007). Therefore, finer regulation of Eda-A1 transgene expression dosage, time and pattern may be required to effect asymmetric spatiotemporal expression of Shh and IGF pathway genes required for the bending process.

Alternatively, additional signals may be required for bent hair formation, as well as for the meibomian and preputial gland development that fail decisively in Eda-A1 transgenic mice (Schmidt-Ullrich et al., 2006; Cui and Schlessinger, 2006; Hashimoto et al., 2008). For example, XEDAR and Troy signaling may play a subsidiary but necessary role for full restoration of bent hair formation. XEDAR mediates EDA signaling by binding to the Eda-A2 isoform (Yan et al., 2000). Unlike EDAR or Troy, which are highly conserved in vertebrates from fish to human, XEDAR is much more variable. XEDAR is highly conserved in mammals but varies considerably in lower animals, and an intracellular Death Domain has been formed only in chicken, hinting at a more specific possible function of XEDAR acquired during evolution (Drew et al., 2007; Knecht et al., 2007; Pantalacci et al., 2008). Consistent with such a possibility, an XEDAR polymorphism has been associated with male pattern baldness in Sardinians (Prodi et al., 2008).

In relevant studies, XEDAR and Troy were both suggested to be involved in chicken feather follicle formation in organ cultures, and are highly expressed in murine hair follicles during bent hair follicle formation (Yan et al., 2000, Schmidt-Ullrich et al., 2006; Drew et al., 2007). However, discordant with the expectation of their significant involvement, XEDAR knockout and Troy mutant mice show no discernible hair follicle phenotypes (Newton et al., 2004; Shao et al., 2005; Hashimoto et al., 2008). Based on their very similar protein structures, XEDAR and Troy might compensate for one another in some functions, but that speculation has not been tested.

In the current model, hair follicle induction seems to be fixed by early dermal signals, with Wnt action most likely determining the overall number of hairs and their future potential positions (Millar, 2002; Sick et al., 2006). Subsequent EDA signaling is involved in hair follicle placode formation/down-growth and hair subtype determination, at least in part by suppressing BMP signaling and activating Shh and LTβ pathways (Cui et al., 2006; Pummila et al., 2007). Thus, Wnt EDA regulated BMP, Shh and LTβ are probably major participants in normal skin appendage induction and progression, and appendage subtype determination is mainly regulated by Eda-A1 spatiotemporally, with fixed time windows for each subtype, as summarized in Fig. 5. This placement for Eda action is consistent with recent observations that an EDAR gene polymorphism affects hair shaft diameter, distinguishing the Asian population from European and African, but with no effect on number of hairs (Fujimoto et al., 2008).

Fig. 5.

Fig. 5

Open in a new tab

Schematic representation of the temporal windows of EDA action for each skin appendage subtype during development. EDA signaling is required for guard hair formation (G) until E17, and for awl hair formation (Aw) until E19. EDA signaling regulates zigzag (Z) and auchen (Au) hair formation from E18, and sweat gland (SWG) formation from E17.

As for its critical action, some EDA function in skin appendage processes is likely mediated by NF-kB activation of Shh, which is tightly regulated by EDA signaling in hair follicles and sweat glands from the early pre-germ stages (Laurikkala et al., 2002; Cui et al., 2006). In Shh knockout mice, guard hair germs can be normally initiated at the expected time, but down-growth to form hair follicles fails (St-Jacques et al., 1998; Chaing et al., 1999). This formulation is consistent with the findings in mice and in human patients with Ectodermal Dysplasia, Anhidrotic, the inherited condition associated with EDA inactivation. But if Shh is a common intermediate signal, it remains to be determined how it regulates hair subtype formation in cooperation with additional effectors during development. The programmed temporal and spatial effects of EDA depend on susceptibility to its action at the germ site during each wave of appendage generation. Once maturation is begun, additional signals, including Notch, Whn and Hox genes, likely take over from EDA by an unknown mechanism.

Materials and Methods

Generation of conditional and conventional Eda-A1 transgenic mice

The generation of the tet-off Eda-A1 transgenic mouse strain was previously reported (Cui et al., 2003). Briefly, mouse Eda-A1 cDNA was cloned into the pTRE vector (Clontech) and was microinjected into pronuclei of C57BL/6J mouse embryos. Embryos were then implanted into pseudo-pregnant female mice. Potential founders were mated to C57BL/6J mice to establish transgenic lines. Male Eda-A1 transgenic mice were then mated to Tabby females to obtain Eda-A1 positive Tabby females. Tabby female progeny carrying an Eda-A1 transgene were then mated with tet-off positive C57BL/6J-TgN (MMTV-tTA) mice (Jackson Laboratory) (Hennighausen et al., 1995) to activate Eda-A1 transgene transcription. Because the EDA gene is X-linked, we analyzed male mice that carry two transgenes, Eda-A1 and tet-off, in a Tabby or wild-type background. The Tabby genotype was scored by PCR amplification and enzyme digestion as described previously (Cui et al., 2006).

Complete suppression of Eda-A1 transgene transcription by doxycycline in vitro and in vivo was confirmed previously (Cui et al., 2003). To further confirm the block of skin appendage development when the transgene was inactive, pregnant female mice were given food containing 200mg/kg doxycycline (BioServe) continuously after fertilization. In the absence of doxycycline, transgene expression fully restored guard hair and sweat glands in Tabby, but when doxycycline food was given, Eda-A1 transgenic Tabby mice were grossly indistinguishable from Tabby, with no restoration of guard hair or sweat glands detected (Figure S4).

Generation of conventional Eda-A1 transgenic mice was described in our previous report (Srivastava et al., 2001). A construct with the Eda-A1 transgene driven by a CMV promoter was microinjected into pronuclei of C57BL/6J mouse embryos, and positive progeny were crossed with Tabby females. The transgene expression level in Tabby background was similar to the endogenous Eda level in wild-type mice.

Treatment of animals was in compliance with Institutional Guidelines, and all animal study protocols were approved by an Institutional Review Board.

Timed mating and scheduled expression of Eda-A1 transgene

Timed matings were set up with Eda-A1 positive Tabby females and tet-off positive male mice to yield transgenic mice with a controllable promoter in a Tabby background.

Doxycycline is rapidly absorbed orally and reaches peak serum concentrations after 2 hours in human (Klein and Cunha, 1995). In cell culture systems, transgene expression was completely blocked within 9 hours of application (Clontech). Therefore, the tet-off system provides rapid and complete shutdown of the transgene. We fed Dox chow (200mg/kg) to pregnant mice at E14.5, E16.5 and E18.5 to shut down Eda-A1 transgene expression around E15, E17 and E19 (designated as E15-off, E17-off and E19-off mice; Figure S1) the germ stages for guard hair, awl hair, and zigzag hair/sweat glands, respectively. At age 2 months, the conditional transgenic mice with time-delimited expression were then compared to TaTG mice, in which the Eda-A1 transgene was expressed throughout life.

The half life of doxycycline is long (18-22 hours in human serum; Klein and Cunha, 1995). Therefore, changing a transgene from off to on with the tet-off system takes time. To assess the possible function of Eda-A1 in zigzag hair formation, we removed Dox food at E15.5 to reduce serum doxycycline concentration in the mice to less than 1/10 of original dosage by E18 and induce Eda-A1 expression by that time (designated as E18-on mice).

The numbers and genotypes of analyzed mice are listed in Table S1.

Hair phenotypes, and histological and immunohistochemical analyses

Hair was shaved from the back skin of each transgenic mouse, and hair shaft morphology and hair type composition were analyzed under a dissection microscope or at 100x magnification. At least 300 hairs were analyzed for each mouse. Back skin or footpads were taken from 2 month-old conditional transgenic mice, and were processed for paraffin or OCT (Optimal cutting temperature) compound embedding for regular histology and immunohistochemistry. Primary antibodies for K14 (Covance), K8 (Progen) and Vimentin (Sigma) were incubated with frozen footpad sections and interacted with AlexaFluor secondary antibodies (Invitrogen) for visualization by DeltaVision reconstruction microscopy (Applied Precision). Sweat tests were carried out on mice from all time points as previously reported (Cui et al., 2003). Meibomian and preputial glands were also scored on histological sections.

Supplementary Material

Supplemental

Acknowledgments

Authors thank R. Nagaraja for critical reading of the manuscript, and A. Butler for animal management. This work was supported by the IRP of the NIH, National Institute on Aging.

Abbreviations

TaTG mice

Tabby mice expressing Eda-A1 transgene for lifetime

EDA

anhidrotic ectodermal dysplasia

EDAR

ectodysplasin receptor

Troy

TNFRSF19

G

guard hair

Aw

awl hair

Au

auchen hair

Z

zigzag hair

Footnotes

Conflict of Interest:

No conflicts declared.

References

  1. Andl T, Reddy ST, Gaddapara T, Millar SE. WNT signals are required for the initiation of hair follicle development. Dev Cell. 2002;2:643–653. doi: 10.1016/s1534-5807(02)00167-3. [DOI] [PubMed] [Google Scholar]
  2. Blanpain C, Fuchs E. Epidermal stem cells of the skin. Annu Rev Cell Dev Biol. 2006;22:339–373. doi: 10.1146/annurev.cellbio.22.010305.104357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Botchkarev VA, Paus R. Molecular biology of hair morphogenesis: development and cycling. J Exp Zoolog B Mol Dev Evol. 2003;298:164–180. doi: 10.1002/jez.b.33. [DOI] [PubMed] [Google Scholar]
  4. Casal ML, Lewis JR, Mauldin EA, Tardivel A, Ingold K, Favre M, et al. Significant correction of disease after postnatal administration of recombinant ectodysplasin A in canine X-linked ectodermal dysplasia. Am J Hum Genet. 2007;81:1050–1056. doi: 10.1086/521988. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chiang C, Swan RZ, Grachtchouk M, Bolinger M, Litingtung Y, Robertson EK, et al. Essential role for Sonic hedgehog during hair follicle morphogenesis. Dev Biol. 1999;205:1–9. doi: 10.1006/dbio.1998.9103. [DOI] [PubMed] [Google Scholar]
  6. Cui CY, Durmowicz M, Ottolenghi C, Hashimoto T, Griggs B, Srivastava AK, et al. Inducible mEDA-A1 transgene mediates sebaceous gland hyperplasia and differential formation of two types of mouse hair follicles. Hum Mol Genet. 2003;12:2931–2940. doi: 10.1093/hmg/ddg325. [DOI] [PubMed] [Google Scholar]
  7. Cui CY, Smith JA, Schlessinger D, Chan CC. X-linked anhidrotic ectodermal dysplasia disruption yields a mouse model for ocular surface disease and resultant blindness. Am J Pathol. 2005;167:89–95. doi: 10.1016/S0002-9440(10)62956-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cui CY, Schlessinger D. EDA signaling and skin appendage development. Cell Cycle. 2006;5:2477–2483. doi: 10.4161/cc.5.21.3403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cui CY, Hashimoto T, Grivennikov SI, Piao Y, Nedospasov SA, Schlessinger D. Ectodysplasin regulates the lymphotoxin-beta pathway for hair differentiation. Proc Natl Acad Sci U S A. 2006;103:9142–9147. doi: 10.1073/pnas.0509678103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Drew CF, Lin CM, Jiang TX, Blunt G, Mou C, Chuong CM, et al. The Edar subfamily in feather placode formation. Dev Biol. 2007;305:232–245. doi: 10.1016/j.ydbio.2007.02.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Durmowicz MC, Cui CY, Schlessinger D. The EDA gene is a target of, but does not regulate Wnt signaling. Gene. 2002;285:203–211. doi: 10.1016/s0378-1119(02)00407-9. [DOI] [PubMed] [Google Scholar]
  12. Elomaa O, Pulkkinen K, Hannelius U, Mikkola M, Saarialho-Kere U, Kere J. Ectodysplasin is released by proteolytic shedding and binds to the EDAR protein. Hum Mol Genet. 2001;10:953–962. doi: 10.1093/hmg/10.9.953. [DOI] [PubMed] [Google Scholar]
  13. Ezer S, Bayes M, Elomaa O, Schlessinger D, Kere J. Ectodysplasin is a collagenous trimeric type II membrane protein with a tumor necrosis factor-like domain and co-localizes with cytoskeletal structures at lateral and apical surfaces of cells. Hum Mol Genet. 1999;8:2079–2086. doi: 10.1093/hmg/8.11.2079. [DOI] [PubMed] [Google Scholar]
  14. Fessing MY, Sharova TY, Sharov AA, Atoyan R, Botchkarev VA. Involvement of the Edar signaling in the control of hair follicle involution (catagen) Am J Pathol. 2006;169:2075–2084. doi: 10.2353/ajpath.2006.060227. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fujimoto A, Kimura R, Ohashi J, Omi K, Yuliwulandari R, Batubara L, et al. A scan for genetic determinants of human hair morphology: EDAR is associated with Asian hair thickness. Hum Mol Genet. 2007 doi: 10.1093/hmg/ddm355. In press. [DOI] [PubMed] [Google Scholar]
  16. Gaide O, Schneider P. Permanent correction of an inherited ectodermal dysplasia with recombinant EDA. Nat Med. 2003;9:614–618. doi: 10.1038/nm861. [DOI] [PubMed] [Google Scholar]
  17. Hammerschmidt B, Schlake T. Localization of Shh expression by Wnt and Eda affects axial polarity and shape of hairs. Dev Biol. 2007;305:246–261. doi: 10.1016/j.ydbio.2007.02.010. [DOI] [PubMed] [Google Scholar]
  18. Hardy MH. The secret life of the hair follicle. Trends Genet. 1992;8:55–61. doi: 10.1016/0168-9525(92)90350-d. [DOI] [PubMed] [Google Scholar]
  19. Hashimoto T, Schlessinger D, Cui CY. Troy binding to lymphotoxin-α activates NF-kB mediated transcription. Cell Cycle. 2008;7:106–111. doi: 10.4161/cc.7.1.5135. [DOI] [PubMed] [Google Scholar]
  20. Headon DJ, Overbeek PA. Involvement of a novel Tnf receptor homologue in hair follicle induction. Nat Genet. 1999;22:370–374. doi: 10.1038/11943. [DOI] [PubMed] [Google Scholar]
  21. Headon DJ, Emmal SA, Ferguson BM, Tucker AS, Justice MJ, Sharpe PT, et al. Gene defect in ectodermal dysplasia implicates a death domain adapter in development. Nature. 2001;414:913–916. doi: 10.1038/414913a. [DOI] [PubMed] [Google Scholar]
  22. Hennighausen L, Wall RJ, Tillmann U, Li M, Furth PA. Conditional gene expression in secretory tissues and skin of transgenic mice using the MMTV-LTR and the tetracycline responsive system. J Cell Biochem. 1995;59:463–472. doi: 10.1002/jcb.240590407. [DOI] [PubMed] [Google Scholar]
  23. Kere J, Srivastava AK, Montonen O, Zonana J, Thomas N, Ferguson B, et al. X-linked anhidrotic (hypohidrotic) ectodermal dysplasia is caused by mutation in a novel transmembrane protein. Nat Genet. 1996;13:409–416. doi: 10.1038/ng0895-409. [DOI] [PubMed] [Google Scholar]
  24. Klein NC, Cunha BA. Tetracyclines. Med Clin North Am. 1995;79:789–801. doi: 10.1016/s0025-7125(16)30039-6. [DOI] [PubMed] [Google Scholar]
  25. Knecht AK, Hosemann KE, Kingsley DM. Constraints on utilization of the EDA-signaling pathway in threespine stickleback evolution. Evol Dev. 2007;9:141–154. doi: 10.1111/j.1525-142X.2007.00145.x. [DOI] [PubMed] [Google Scholar]
  26. Langbein L, Rogers MA, Praetzel S, Cribier B, Peltre B, Gassler N, et al. Characterization of a novel human type II epithelial keratin K1b, specifically expressed in eccrine sweat glands. J Invest Dermatol. 2005;125:428–444. doi: 10.1111/j.0022-202X.2005.23860.x. [DOI] [PubMed] [Google Scholar]
  27. Laurikkala J, Pispa J, Jung HS, Nieminen P, Mikkola M, Wang X, et al. Regulation of hair follicle development by the TNF signal ectodysplasin and its receptor Edar. Development. 2002;129:2541–2553. doi: 10.1242/dev.129.10.2541. [DOI] [PubMed] [Google Scholar]
  28. Millar SE. Molecular mechanisms regulating hair follicle development. J Invest Dermatol. 2002;118:216–225. doi: 10.1046/j.0022-202x.2001.01670.x. [DOI] [PubMed] [Google Scholar]
  29. Muller-Rover S, Handjiski B, van der Veen C, Eichmuller S, Foitzik K, McKay IA, et al. A comprehensive guide for the accurate classification of murine hair follicles in distinct hair cycle stages. J Invest Dermatol. 2001;117:3–15. doi: 10.1046/j.0022-202x.2001.01377.x. [DOI] [PubMed] [Google Scholar]
  30. Mustonen T, Pispa J, Mikkola ML, Pummila M, Kangas AT, Pakkasjarvi L, et al. Stimulation of ectodermal organ development by Ectodysplasin-A1. Dev Biol. 2003;259:123– 136. doi: 10.1016/s0012-1606(03)00157-x. [DOI] [PubMed] [Google Scholar]
  31. Mustonen T, Ilmonen M, Pummila M, Kangas AT, Laurikkala J, Jaatinen R, et al. Ectodysplasin A1 promotes placodal cell fate during early morphogenesis of ectodermal appendages. Development. 2004;131:4907–4919. doi: 10.1242/dev.01377. [DOI] [PubMed] [Google Scholar]
  32. Newton K, French DM, Yan M, Frantz GD, Dixit VM. Myodegeneration in EDA-A2 transgenic mice is prevented by XEDAR deficiency. Mol Cell Biol. 2004;24:1608–1613. doi: 10.1128/MCB.24.4.1608-1613.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pantalacci S, Chaumot A, Benoit G, Sadier A, Delsuc F, Douzery EJ, et al. Conserved features and evolutionary shifts of the EDA signaling pathway involved in vertebrate skin appendage development. Mol Biol Evol. 2008;25:912–928. doi: 10.1093/molbev/msn038. [DOI] [PubMed] [Google Scholar]
  34. Prodi DA, Pirastu N, Maninchedda G, Sassu A, Picciau A, Palmas MA, et al. EDA2R Is Associated with Androgenetic Alopecia. J Invest Dermatol. 2008 doi: 10.1038/jid.2008.60. Epub ahead of print. [DOI] [PubMed] [Google Scholar]
  35. Pummila M, Fliniaux I, Jaatinen R, James MJ, Laurikkala J, Schneider P, et al. Ectodysplasin has a dual role in ectodermal organogenesis: inhibition of Bmp activity and induction of Shh expression. Development. 2007;134:117–125. doi: 10.1242/dev.02708. [DOI] [PubMed] [Google Scholar]
  36. Schlake T. Segmental Igfbp5 expression is specifically associated with the bent structure of zigzag hairs. Mech Dev. 2005;122:988–997. doi: 10.1016/j.mod.2005.04.012. [DOI] [PubMed] [Google Scholar]
  37. Schmidt-Ullrich R, Tobin DJ, Lenhard D, Schneider P, Paus R, Scheidereit C. NF-kappaB transmits Eda A1/EdaR signalling to activate Shh and cyclin D1 expression, and controls post-initiation hair placode down growth. Development. 2006;133:1045–1057. doi: 10.1242/dev.02278. [DOI] [PubMed] [Google Scholar]
  38. Shao Z, Browning JL, Lee X, Scott ML, Shulga-Morskaya S, Allaire N, et al. TAJ/TROY, an orphan TNF receptor family member, binds Nogo-66 receptor 1 and regulates axonal regeneration. Neuron. 2005;45:353–359. doi: 10.1016/j.neuron.2004.12.050. [DOI] [PubMed] [Google Scholar]
  39. Sick S, Reinker S, Timmer J, Schlake T. WNT and DKK determine hair follicle spacing through a reaction-diffusion mechanism. Science. 2006;314:1447–1450. doi: 10.1126/science.1130088. [DOI] [PubMed] [Google Scholar]
  40. Srivastava AK, Pispa J, Hartung AJ, Du Y, Ezer S, Jenks T, et al. The Tabby phenotype is caused by mutation in a mouse homologue of the EDA gene that reveals novel mouse and human exons and encodes a protein (ectodysplasin-A) with collagenous domains. Proc Natl Acad Sci U S A. 1997;94:13069–13074. doi: 10.1073/pnas.94.24.13069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Srivastava AK, Durmowicz MC, Hartung AJ, Hudson J, Ouzts LV, Donovan DM, et al. Ectodysplasin-A1 is sufficient to rescue both hair growth and sweat glands in Tabby mice. Hum Mol Genet. 2001;10:2973–2981. doi: 10.1093/hmg/10.26.2973. [DOI] [PubMed] [Google Scholar]
  42. St-Jacques B, Dassule HR, Karavanova I, Botchkarev VA, Li J, Danielian PS, et al. Sonic hedgehog signaling is essential for hair development. Curr Biol. 1998;8:1058–1068. doi: 10.1016/s0960-9822(98)70443-9. [DOI] [PubMed] [Google Scholar]
  43. Vielkind U, Hardy MH. Changing patterns of cell adhesion molecules during mouse pelage hair follicle development. 2. Follicle morphogenesis in the hair mutants, Tabby and downy. Acta Anat (Basel) 1996;157:183–194. doi: 10.1159/000147880. [DOI] [PubMed] [Google Scholar]
  44. Yan M, Wang LC, Hymowitz SG, Schilbach S, Lee J, Goddard A, et al. Two-amino acid molecular switch in an epithelial morphogen that regulates binding to two distinct receptors. Science. 2000;290:523–527. doi: 10.1126/science.290.5491.523. [DOI] [PubMed] [Google Scholar]
  45. Zhang M, Brancaccio A, Weiner L, Missero C, Brissette JL. Ectodysplasin regulates pattern formation in the mammalian hair coat. Genesis. 2003;37:30–37. doi: 10.1002/gene.10230. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental
NIHMS833044-supplement-Supplemental.pdf (2.4MB, pdf)

RESOURCES