- Email: [email protected]
Hair Cycle-Specific Immunolocalization of Retinoic Acid Synthesizing Enzymes Aldh1a2 and Aldh1a3 Indicate Complex Regulation
Hair Cycle-Specific Immunolocalization of Retinoic Acid Synthesizing Enzymes Aldh1a2 and Aldh1a3 Indicate Complex Regulation Helen B. Everts, Lloyd E. King Jr.,w John P. Sundberg,wz and David E. Ong
Department of Biochemistry and wDivision of Dermatology, Vanderbilt University Medical Center, Nashville, Tennessee, USA; zThe Jackson Laboratory,
Bar Harbor, Maine, USA
Retinoic acid has long been known to alter skin and hair growth but an exact mechanism is unclear. This study was performed to examine the sites of endogenous retinoic acid synthesis in the cycling hair follicle to better understand the role retinoic acid plays in this process. Retinal dehydrogenases (Aldh1a1, 2, and 3, formerly Raldh 1, 2, and 3) are the enzymes responsible for the last step in retinoic acid synthesis. Immunohistochemistry was performed on adult C57BL/6J mouse skin sections with antibodies against Aldh1a2 and Aldh1a3. Aldh1a2 expression was seen primarily in the outer root sheath and basal/spinous layer during all stages of the hair cycle, and in the bulge during anagen and early catagen, whereas Aldh1a3 expression was primarily in the dermal papilla, pre-cortex, and hair shaft during mid–late anagen. The expression patterns of these two similar retinoic acid synthesizing enzymes at specific follicular sites suggest that they mediate and are regulated by different epithelial proliferation and differentiation signaling pathways.
Key words: sebaceous gland/bulb/bulge/hair follicle/immunohistochemistry J Invest Dermatol 123:258 – 263, 2004 Vitamin A and its derivatives (retinoids) are critically important in the development and maintenance of multiple tissues, including skin and hair, as shown by the detrimental effects of vitamin A deficiency (Wolbach and Howe, 1925; Frazier and Hu, 1931). These observations have led to the widespread use of retinoids in various skin disorders (Stuttgen, 1986). All-trans retinoic acid (atRA) was proven in vivo (Werner and DeLuca, 2001) and in vitro (Kurlandsky et al, 1994) to be the most physiologically relevant retinoid in the skin. But, due to toxicity issues, 13-cis retinoic acid (13cRA, i.e., Accutane, Roche Laboratories, Inc., Nutley, New Jersey), which can slowly isomerize to atRA, is used more frequently in clinically abnormal states (Stuttgen, 1986). The regulation of endogenous atRA synthesis and its precise tissue/cellular localization is a subject of active investigation. AtRA was shown to be synthesized in cultured keratinocytes (Kurlandsky et al, 1994) and degraded in dermal fibroblasts (Randolph and Simon, 1998). This degradation of atRA by the dermal fibroblasts would limit the delivery of physiological concentrations of plasma atRA
to the epidermis and hair follicle making local atRA synthesis essential for normal function. The synthesis of atRA from circulating retinol occurs by the action of two enzyme families (Napoli, 1999). Retinol dehydrogenases (Rdh) convert retinol to retinal, whereas retinal dehydrogenases (Aldh1a) convert retinal to RA. Two human epidermal Rdh family members, hRODH-E and hRODHE2, have been identified (Jurukovski et al, 1999; Markova et al, 2003). Currently, there are three known Aldh1a (1–3) that convert all-trans retinal to atRA. Recent studies with Aldh1a1 null mice suggest that this enzyme is involved in the catabolism of excess retinol (Fan et al, 2003). The expressions of the cellular retinol binding protein and cellular retinoic acid binding protein type II have also been associated with the ability to make RA (i.e., Bucco et al, 1997; Napoli, 1999). Therefore, expression of these binding proteins and enzymes can be used as markers for sites of RA synthesis. Recently, Markova et al (2003) showed that hRODH-E2 transcripts are expressed in sebaceous and sweat glands, matrix cells, and the Huxley cells of the inner root sheath (IRS) of the infundibulum. But little data are available on the expression pattern of Aldh1a in the hair follicle and sebaceous gland. Most of the studies with Aldh1a2 and Aldh1a3 have focused on embryogenesis. The temporal and spatial expression patterns of retinal dehydrogenases have been used as an initial step in dissecting the sites of RA synthesis in the mouse embryo, the regulation of RA synthesis, and the precise role of RA during embryonic development. The expression of Aldh1a2 and Aldh1a3 is seen at distinct times and locations throughout embryogenesis
Abbreviations: Aldh1a, retinal dehydrogenase; atRA, all-trans retinoic acid; 13cRA, 13-cis retinoic acid; ILORS, inner layer of the outer root sheath; IRS, inner root sheath; Lef/Tcf, lymphoid enhancer factor/T-cell factor; ORS, outer root sheath; Rara,b,c, retinoic acid receptor a, b, g; RA, retinoic acid; Shh, sonic hedgehog; Wnt, wingless related MMTV The mouse genome nomenclature was used (www.informatics. jax.org). Genes and RNA message expression is italicized, while proteins are not. First letter capitalized followed by lowercase indicates a mouse or rat gene, while all capital letters indicates the human gene or protein.
Copyright r 2004 by The Society for Investigative Dermatology, Inc.
258
123 : 2 AUGUST 2004
FOLLICULAR LOCALIZATION OF Aldh1a2 AND Aldh1a3
259
(Blentic et al, 2003). In the developing hair follicle Aldh1a3 expression was reported at the base of the hair follicle at 5 d, but not 3 wk, post-partum (Niederreither et al, 2002a). The sites of RA synthesis suggested in these expression studies were consistent with developmental problems in the null mice. Aldh1a2/ mice die in utero due to defects in heart development (Niederreither et al, 1999), and Aldh1a3/ mice die within 10 h of birth due to defects in nasal development (Dupe et al, 2003). These defects can be only partially restored with maternal RA administration, suggesting that precise timing of RA synthesis at specific sites is essential. Neither Aldh1a2 nor Aldh1a3 expression patterns in the sebaceous gland or hair follicle have been well described in the adult mammal. Thus, this report describes the expression pattern of Aldh1a2 and Aldh1a3 in the adult mouse hair follicle, as an initial step in dissecting the role of endogenous RA in mammalian hair cycling.
Results and Discussion To examine the possible sites of RA synthesis, immunohistochemistry was performed on adult C57BL/6J mouse skin with antibodies against Aldh1a2 and Aldh1a3. Dot blot analysis of these antibodies against recombinant proteins shows that these antibodies are specific (Fig 1). The antiAldh1a2 antibody did not cross-react with Aldh1a3. There was a slight cross-reaction of the anti-Aldh1a3 antibody with Aldh1a2, but this antibody showed a 64-fold selectivity for Aldh1a3 over Aldh1a2. Very different anatomical sites of Aldh1a2 versus Aldh1a3 expression were observed in mouse hair follicles and their associated sebaceous gland at various stages in the hair cycle (Tables I and II and Figs 2
Figure 1 Dot blot analysis of antibodies against recombinant whole proteins. Recombinant rat proteins Aldh1a2 (2) and Aldh1a3 (3) (0–25 pmol) were blotted onto nitrocellulose membranes, cut into strips, and exposed to affinity-purified antibodies against Aldh1a2 (a) or Aldh1a3 (b). Detection was by enhanced chemiluminescence. The image shows the two films for the two antibodies overlaying the strips of dot blotted protein.
and 3). Each follicular site reflects a specific function in the regulation of proliferation and differentiation during the hair cycle. This differential localization of two proteins with the
Table I. Aldh1a2 expression in the stages of the hair cyclea Stage
Basal Spin.
Seb.
ORS
Bulge
Prol. Ker.
ILORS/CL
ES
Germ Caps.
Telogen
þþþ
þþþ
þ /
NA
NA
NA
þþþ
Anagen I
þþ
þ /
þþ
þþ
þþ
NA
NA
NA
Anagen II
þþ
þ
þ
þþ
þþ
NA
NA
NA
Anagen IIIa
þþ
þ
þþ
þþ
þþ
NA
NA
NA
Anagen IIIb
þþ
þ /
þ þ in
þþ
þþ
þþ
NA
NA
Anagen IIIc
þþ
þ /
þ þ is
þþ
NA
þþ
NA
NA
Anagen IV
þþ
þ /
þ þ is,in
þþ
NA
þþ
NA
NA
Anagen V
þþ
þ /
þ þ þ þ bin
þþþ
NA
þþ
NA
NA
Anagen VI/catagen I
þ
þ /
0 þ bin
þþ
NA
þ /
NA
NA
Catagen II
þþ
0 þ þ bin
þþ
NA
þ
NA
NA
Catagen III
þþ
0 þ þ bin
þþ
NA
þþ
NA
NA
Catagen IV
þ
þ /
0 þ þ bin
þþ
NA
þþ
NA
NA
Catagen V
þþ
þ /
þþþ
þ /
NA
þþ
þ
þ
Catagen VI
þþ
þ /
þþþ
þ /
NA
þþ
þ
þ
Catagen VII
þþ
þþþ
þ /
NA
NA
þþ
þþ
Catagen VIII
þþ
þ /
þþþ
þ /
NA
NA
NA
þþ
a Negative (), faint ( þ /), weak ( þ ), strong ( þ þ ), and very strong ( þ þ þ ) staining is indicated. Basal/Spin, basal/spinous layer; Seb., sebocyte; ORS, outer root sheath; Prol. Ker, proliferating keratinocyte strand between dermal papilla and club hair in early anagen; ILORS/CL, inner layer of the outer root sheath/companion layer; ES, epithelial strand; Germ Caps., germ capsule; in, infundibulum; is, isthmus; b, bulb; NA, not applicable.
260 EVERTS ET AL
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY Table II. Aldh1a3 expression in the stages of the hair cyclea
Stage
Seb gland
Seb duct
ILORS/CL
Hair Fiber
Pre cortex
DP
Germ Caps.
Telogen
þ /
NA
Anagen I
þþ þ
NA
NA
NA
þ /
NA
Anagen II
þ
þþ þ
NA
NA
NA
NA
Anagen IIIa
þ
þ
NA
NA
þ
þ
NA
þþ
þþ
NA
Anagen IIIb
þþ
þþ
NA
Anagen IIIc
þþ
þþ þ
þþ
NA
Anagen IV
þþ
þ þ þ 0 b–in
þþ
þ
NA
Anagen V
þ þ þ 0 bin
þ
þþþ
NA
Anagen VI/catagen I
þ
þ þ þ 0 bin
þþþ
þ /
NA
Catagen II
þ þ þ 0 bin
þþ
NA
Catagen III
þ
þ / 0 bin
NA
Catagen IV
þ
þ / 0 bin
NA
NA
Catagen V
þ
þþ
NA
þ
Catagen VI
þ /
þ
þþ
NA
þ
IIIa–b
Catagen VII
þ /
NA
NA
þ
Catagen VIII
þ /
NA
NA
a Negative (), faint ( þ /), weak ( þ ), strong ( þ þ ), and very strong ( þ þ þ ) staining is indicated. Seb., sebocyte; ILORS/CL, inner layer of the outer root sheath/companion layer; Hair fiber (aka shaft); Precortex, cells just above the matrix; DP, dermal papilla; Germ Caps., germ capsule; in, infundibulum; b, bulb; NA, not applicable.
Figure 2 Aldh1a2 expression in adult mouse hair follicles. Immunohistochemical localization of Aldh1a2 in adult mouse hair follicles during telogen (a), anagen IIIa (b), anagen V (c), catagen V (d), and catagen VII (e). Bar ¼ 1 mm. Insets are higher magnifications of their respective follicles (arrows). Bar ¼ 0.1 mm all insets.
apparent same enzymatic function allows RA signaling to be differentially regulated at each unique site. Dermal papilla: induction and termination of anagen, catagen, and telogen There are three main stages of the hair cycle: anagen (growth); catagen (regression); and
telogen (rest). Aldh1a2 was expressed strongly during all stages of the hair cycle (Fig 2), whereas Aldh1a3 was expressed primarily during mid–late anagen (Fig 3). During telogen, Aldh1a2 was expressed strongly in the outer root sheath (ORS), bulge and germ capsule (Table I, Fig 2a, and inset i). In early anagen (I–IIIb) Aldh1a2 was expressed in the
123 : 2 AUGUST 2004
FOLLICULAR LOCALIZATION OF Aldh1a2 AND Aldh1a3
261
Figure 3 Aldh1a3 expression in adult mouse hair follicles. Immunohistochemical localization of Aldh1a3 in adult mouse hair follicles during anagen IIIa (a), IIIa–b (b), IIIc (c), V (d), and early VI (e). Bar ¼ 1 mm. Insets are higher magnifications of their respective follicles (small arrows). Bar ¼ 0.1 mm all insets. Large arrows point to the sebaceous gland duct.
bulge region and the proliferating keratinocytes between the club hair and the bulb region (Table I, Fig 2b, inset iii). Starting in mid-anagen, Aldh1a3 but not Aldh1a2 was expressed in the bulb region (Table II, Figs 2b and c, 3a–e, and Fig 3 insets ii, iv, v, vii–x and data not shown). This Aldh1a3 expression appeared to differ during the stages of anagen. In anagen IIIa–b, IIIc, and V Aldh1a3 localized to the dermal papilla, but then in anagen IIIa–b, IIIb, IIIc, and VI it was strongly expressed and anagen IV and V weakly expressed in the pre-cortical region (Table II, Fig 3a–e and inset ii, iv, v, vii–x and data not shown). The hair fiber (aka shaft) expressed Aldh1a3 beginning in anagen IIIb and increased in intensity through catagen II, then reduced its intensity in catagen III and IV as hair growth ceases (Table II, Fig 3c–e, insets v, viii, x, and data not shown). During late anagen a gradient of Aldh1a3 expression is seen in the fiber with the strongest expression seen in the pre-cortex region and progressively weaker expression is seen as the fiber reaches the skin surface (Fig 3e, insets x and xi). As was seen for telogen and early anagen, the bulge region continues to express Aldh1a2 throughout anagen (Table I, Fig 2c, inset v). In addition, the ORS expression of Aldh1a2 is strongest in the isthmus near the bulge throughout anagen, with a peak in expression seen during anagen V and a drop at anagen VI/catagen I (Table I, Fig 2c, inset v and data not shown). These expression patterns suggest that Aldh1a2 may be involved in the pathways responsible for overall maintenance of the hair follicle, whereas Aldh1a3 may be a positive regulator of anagen growth, through epithelial–mesenchymal interactions. The hair cycle is regulated through the interactions of growth factors between the bulge area stem cells, the dermal papilla, the epithelial matrix and pre-cortical cells (reviewed in Stenn and Paus, 2001). All of these sites expressed either Aldh1a2 or Aldh1a3 at some stage of the hair cycle, suggesting a role of RA in these regulatory
pathways. Two of these growth factor signaling systems, wingless related MMTV (Wnt) and sonic hedgehog (Shh), are regulated by RA in other systems (Easwaran et al, 1999; Power et al, 1999; Niederreither et al, 2002b), providing further support for a possible role of RA in hair cycle regulation. The localization of Aldh1a2 was consistent with that of the Wnt transcription factor T cell factor 3 (Tcf3) in the ORS and bulge, and the Shh receptor Ptc1 in the proliferating keratinocytes between the DP and club hair during early anagen and ORS in late anagen (Callahan and Oro, 2001; Alonso and Fuchs, 2003; Oro and Higgins, 2003). The expression of Aldh1a3 was similar to that of the Wnt transcription factor lymphoid enhancer factor 1 (Lef1) in the dermal papilla and pre-cortical cells, and Ptc 1 in the dermal papilla. Tcf3 and Lef1 are in the same family of transcription factors and it has been suggested that Tcf3 acts to inhibit transcription in the bulge during Wnt stimulation, whereas Lef1 activates transcription in the pre-cortex when stimulated by Wnt. This differential expression of Aldh1a and Lef/ TcF family members at different sites within the hair follicle would allow for site-specific regulation. Future studies could examine whether RA has different effects on Lef1 versus TcF3. In addition, do either Lef1 or TcF3 regulate Aldh1a2 or Aldh1a3? A role for RA in stem cell regulation was first noted by Wolbach and Howe in 1925. The expression of Aldh1a2 in the bulge region throughout anagen (Table I, Fig 2b and c, insets iii and v) provides further support for such a role of RA in stem cell maintenance. During all stages of catagen, Aldh1a2 expression occurs in the ORS, especially near the bulge (Table I, Fig 2d and e, insets vii and viii). But the outer cells of the bulge do not express Aldh1a2 in catagen V–VIII (Table I, Fig 2d and e, insets vii and viii). Aldh1a2 also localizes to the germ capsule cells as they form in catagen V–telogen (Table I, Fig 2a, d, and e, inset vi). Weak expression of Aldh1a3 was also seen in the germ capsule during late catagen (Table II, data
262 EVERTS ET AL
not shown). The epithelial strand also showed Aldh1a2 expression during catagen V–VII (Table I, Fig 2d and e). In the bulbar region Aldh1a3 expression is reduced as catagen progresses (Table II, data not shown). The dermal papilla only faintly expresses Aldh1a3 in anagen VI/catagen I (Fig 3e, inset ix) and is completely negative through catagen and telogen. The pre-cortex ceases to express Aldh1a3 by catagen III when these cells regress. Catagen is marked by an arrest of proliferation in the bulbar region and massive apoptosis (Stenn and Paus, 2001). Thus, loss of Aldh1a3 produced RA in the dermal papilla and Aldh1a2 produced RA in the bulge may be involved in this arrest in proliferation. In addition, added RA has been found to initiate apoptosis in differentiating non-tumorogenic mouse keratinocytes (Islam et al, 2000). The strong expression of Aldh1a2 within the regressing catagen hair follicle suggests that RA synthesis is involved in this catagen-induced apoptosis. To confirm this, future studies with Tunel, Caspase 3 double labeling on frozen sections can be done. Sheath–fiber interactions The inner layer of the outer root sheath (ILORS, aka companion layer) expressed Aldh1a2 strongly during anagen IIIb–V and catagen IV–VI, but weakly during anagen VI–catagen III (Table I, Fig 2c, inset iv). Aldh1a3 localized to the ILORS during anagen IV, anagen VI, and catagen III–VI (Table II, data not shown). The ORS expressed Aldh1a2 during all stages of the hair cycle with the strongest expression in anagen V in the isthmus and catagen V–telogen the whole length of the follicle (Table I, Fig 2a, c, and e, insets i, v–viii). After anagen initiation occurs, the newly made IRS and fiber must traverse the ORS to move to the skin surface (Stenn and Paus, 2001). This is thought to involve a proteolytic processs and the expression of the tissue inhibitor of metalloproteinases (TIMP) in Henle’s layer of the IRS and plasminogen activator inhibitor-2 (PAI-2) in the ILORS. The expression of Aldh1a3 in the ILORS suggests a role for RA in this process. This hypothesis is supported by recent microarray data in our laboratory that found both TIMP and PAI-2 to be upregulated by RA in the rat uterus (Li and Ong, unpublished observation). The sebaceous glands expressed both Aldh1a2 and Aldh1a3 during anagen II–IIIa (Tables I and II, Fig 2b, inset iii, Fig 3a and b, insets i, iii). There was light color in the sebaceous glands at other times, but that may be nonspecific as a light background was sometimes noted (Fig 3c, inset vi). Aldh1a3 also localized to cells lining the sebaceous gland duct during anagen I–IIIa and catagen V– telogen (Table II, Fig 3a and b, insets i and iii large arrow, and data not shown). Thus, RA synthesis may occur within the sebocyte or in the isthmus where the sebaceous gland duct opens and affect sebaceous gland function. In vitro studies have shown that a factor produced by the sebaceous gland, duct, or associated isthmus is required for the hair fiber to be released from the IRS (Williams and Stenn, 1994; Williams et al, 1996). These studies have been supported in vivo by studies with the asebia mouse, which have small, hypoplastic sebaceous glands, defects in the release of the hair fiber, and progressive scarring alopecia (Sundberg et al, 2000). RA has been shown to be important for sebaceous gland function and hair sheath–fiber interac-
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
tions in a concentration-dependent manner (Zouboulis et al, 1993; Williams et al, 1996). Vitamin A deficiency in cultured human sebocytes results in decreased cell proliferation and lipogenesis that can be partially restored with low doses of 13cRA. At concentrations above 107 M 13cRA and atRA inhibit proliferation, lipogenesis, and sheath growth (Zouboulis et al, 1993; Williams et al, 1996). The localization of Aldh1a3 to sites just proximal to the sebaceous gland duct and Aldh1a2 and Aldh1a3 within the sebocyte suggests that RA made at these sites are important for sebocyte function and the normal growth and release of the hair fiber. Interfollicular epidermis and presumptive water barrier Lipids made within the sebocyte are also essential for water barrier formation (Madison, 2003). Thus, Aldh1a3 synthesized RA may play a role in water barrier formation by altering lipid metabolism in the sebocyte. This role of RA in sebocytes is supported by studies of dominant negative Rara transgenic mice, which have defects in water barrier formation and lipid metabolism (Attar et al, 1997). Also, RA has been found to activate PPAR b/d (Shaw et al, 2003), whereas the drug, clofibrate, has been found to increase barrier function (Hanley et al, 1999). Aldh1a2 is localized to the ORS of the infundibulum in all stages (Table I, Fig 2a–e, inset ii). But Aldh1a3 was not seen in the ORS of the infundibulum at any stage of the hair cycle (Fig 3a–e, data not shown). The expression of Aldh1a2 within the basal/spinous layers suggested that RA is also involved in differentiation in these cells directly and not just in sebocyte lipid production. The fact that RA synthesis at these two sites involved in water barrier formation occurs by two separate enzymes suggests that they are differentially regulated. This report demonstrates that Aldh1a2 and Aldh1a3 are expressed at different sites within the cycling hair follicle. Their expression patterns suggest that locally produced RA is important in all stages of the mammalian hair cycle. Aldh1a3 expression may be involved in epithelial–mesenchymal interactions, whereas Aldh1a2 expression may be involved in stem cell maintenance. Having different RA synthesizing enzymes involved at each site and stage allow differential regulation of each step in the hair cycle. Future experiments will examine more closely the regulation of these enzymes, as well as other components of RA synthesis, during the normal hair cycle. Aldh1a expression patterns in mice null for Wnts and Shh will help to establish regulatory pathways in normal and abnormal skin and hair growth. Materials and Methods Tissue acquisition Archived paraffin blocks of wax stripped dorsal skin (used from previous hair cycle studies) from C57BL/ 6J wild-type mice fixed with Tellyzneski–Fekete’s acid–alcohol– formalin fixative were sectioned at 6 mm. The Paus system of determining the stages of the hair cycle was used without additional staining (Muller-Rover et al, 2001). Antibodies Rabbit polyclonal antibodies against Aldh1a2 and 3 were developed. For the anti-Aldh1a2 antibody the peptide GGKGLGRKGFFIEP was chosen. This peptide was synthesized by PeptidoGenic (Livermore, California), conjugated to Imject Activated KLH (Pierce, Rockford, Illinois), and injected intradermally into rabbits with Hunter’s Titermax Gold (CytRx Corp., Norcross,
123 : 2 AUGUST 2004
FOLLICULAR LOCALIZATION OF Aldh1a2 AND Aldh1a3
Georgia). Antiserum was first purified on a Protein A Sepharose column (Pierce, Rockford, Illinois) to obtain an immunoglobulin (Ig) G fraction. Then the IgG fraction was affinity purified using a histidine-tagged whole protein rat Aldh1a2 (generous gift of Marcia Newcomer, PhD Dept. Biol. Sci., LSU) immobilized on a CNBractivated Sepharose 4B column (Amersham Biosciences, Piscataway, New Jersey). The Aldh1a3 antibody was made by a similar procedure except a recombinant whole rat protein was used for the immunization and it was affinity purified using the recombinant whole protein immobilized on a Sulfolink column (Pierce, Rockford, Illinois).
epithelial cells correlates with their synthesis of retinoic acid. Biochemistry 36:4009–4014, 1997 Callahan CA, Oro AE: Monstrous attempts at adnexogenesis: Regulating hair follicle progenitors through Sonic hedgehog signaling. Curr Opin Genet Dev 11:541–546, 2001 Dupe V, Matt N, Garnier JM, Chambon P, Mark M, Ghyselinck NB: A newborn lethal defect due to inactivation of retinaldehyde dehydrogenase type 3 is prevented by maternal retinoic acid treatment. Proc Natl Acad Sci USA 100:14036–14041, 2003 Easwaran V, Pishvaian M, Shah S, Byers S: Cross-regulation of beta-catenin-Lef/ Tcf and retinoid signaling pathways. Curr Biol 9:1415–1418, 1999 Fan X, Molotkov A, Manabe S, et al: Targeted disruption of Aldh1a1 (Raldh1) provides evidence for a complex mechanism of retinoic acid synthesis in the developing retina. Mol Cell Biol 23:4637–4648, 2003 Frazier CN, Hu C-K: Cutaneous lesions associated with a deficiency in vitamin A in man. Arch Intern Med 48:507–514, 1931 Hanley K, Komuves LG, Bass NM, et al: Fetal epidermal differentiation and barrier development in vivo is accelerated by nuclear hormone receptor activators. J Invest Dermatol 113:788–795, 1999 Islam TC, Skarin T, Sumitran S, Toftgard R: Retinoids induce apoptosis in cultured keratinocytes. Br J Dermatol 143:709–719, 2000 Jurukovski V, Markova NG, Karaman-Jurukovska N, Randolph RK, Su J, Napoli JL, Simon M: Cloning and characterization of retinol dehydrogenase transcripts expressed in human epidermal keratinocytes. Mol Genet Metab 67:62–73, 1999 Kurlandsky SB, Xiao JH, Duell EA, Voorhees JJ, Fisher GJ: Biological activity of all-trans retinol requires metabolic conversion to all-trans retinoic acid and is mediated through activation of nuclear retinoid receptors in human keratinocytes. J Biol Chem 269:32821–32827, 1994 Madison KC: Barrier function of the skin: ‘‘La raison d’etre’’ of the epidermis. J Invest Dermatol 121:231–241, 2003 Markova NG, Pinkas-Sarafova A, Karaman-Jurukovska N, Jurukovski V, Simon M: Expression pattern and biochemical characteristics of a major epidermal retinol dehydrogenase. Mol Genet Metabol 78:119–135, 2003 Muller-Rover S, Handjiski B, van der Veen C, et al: A comprehensive guide for the accurate classification of murine hair follicles in distinct hair cycle stages. J Invest Dermatol 117:3–15, 2001 Napoli JL: Interactions of retinoid binding proteins and enzymes in retinoid metabolism. Biochim Biophys Acta 1440:139–162, 1999 Niederreither K, Fraulob V, Garnier JM, Chambon P, Dolle P: Differential expression of retinoic acid-synthesizing (Raldh) enzymes during fetal development and organ differentiation in the mouse. Mech Dev 110:165– 171, 2002a Niederreither K, Subbarayan V, Dolle P, Chambon P: Embryonic retinoic acid synthesis is essential for early mouse post-implantation development. Nat Genet 21:444–448, 1999 Niederreither K, Vermot J, Schuhbaur B, Chambon P, Dolle P: Embryonic retinoic acid synthesis is required for forelimb growth and anteroposterior patterning in the mouse. Development 129:3563–3574, 2002b Oro AE, Higgins K: Hair cycle regulation of hedgehog signal reception. Dev Biol 255:238–248, 2003 Power SC, Lancman J, Smith SM: Retinoic acid is essential for Shh/Hoxd signaling during rat limb outgrowth but not for limb initiation. Dev Dyn 216:469–480, 1999 Randolph RK, Simon M: Dermal fibroblasts actively metabolize retinoic acid but not retinol. J Invest Dermatol 111:478–484, 1998 Shaw N, Elholm M, Noy N: Retinoic acid is a high affinity selective ligand for the peroxisome proliferator-activated receptor b/d. J Biol Chem 278:41589– 41592, 2003 Stenn KS, Paus R: Controls of hair follicle cycling. Physiol Rev 81:449–494, 2001 Stuttgen G: Historical perspectives of tretinoin. J Am Acad Dermatol 15:735–740, 1986 Sundberg JP, Boggess D, Sundberg BA, Eilertsen K, Parimoo S, Filippi M, Stenn K: Asebia-2J (Scd1ab2J): A new allele and a model for scarring alopecia. Am J Pathol 156:2067–2075, 2000 Werner EA, DeLuca HF: Metabolism of a physiological amount of all-trans-retinol in the vitamin A-deficient rat. Arch Biochem Biophys 393:262–270, 2001 Williams D, Siock P, Stenn K: 13-cis-retinoic acid affects sheath-shaft interactions of equine hair follicles in vitro. J Invest Dermatol 106:356–361, 1996 Williams D, Stenn KS: Transection level dictates the pattern of hair follicle sheath growth in vitro. Dev Biol 165:469–479, 1994 Wolbach SB, Howe PR: Tissue changes following deprivation of fat-soluble A vitamin. J Exp Med 42:753–777, 1925 Zouboulis CC, Korge BP, Mischke D, Orfanos CE: Altered proliferation, synthetic activity, and differentiation of cultured human sebocytes in the absence of vitamin-A and their modulation by synthetic retinoids. J Invest Dermatol 101:628–633, 1993
Dot blot analysis Recombinant whole protein rat Aldh1a2 was a generous gift of Marcia Newcomer, PhD Dept. Biol. Sci., LSU. Histidine-tagged whole protein rat Aldh1a3 was expressed in Escherichia coli and purified using Ni Agarose according to the manufacturer’s instructions (Qiagen, Valencia, California). These proteins (0–25 pmol) were blotted onto a nitrocellulose membrane (BioRad, Hercules, California) using the Bio-Dot microfiltration apparatus (BioRad) according to the manufactures instructions. Strips of dots blots were cut and blocked with 5% dry milk in Tris buffered saline containing 0.1% Tween. Blots were then incubated in primary antibody for 1 h, washed with Tris buffered saline with 0.1% Tween, incubated with an antirabbit antibody conjugated with horseradish peroxidase (Jackson ImmunoResearch Laboratories, Inc., West Grove, Pennsylvania), washed again and detected with enhanced chemiluminescence (Amersham Biosciences). Immunohistochemistry Sections were pre-treated with 3% hydrogen peroxide, blocked with 3% BSA plus 1.28% normal goat serum, incubated with the primary antibody overnight at 41C, then incubated with a biotin conjugated secondary antibody, an antibiotin IgG conjugated with horseradish peroxidase, and then stained with AEC þ (Dako, Carpinteria, California; Bucco et al, 1997). Hydrogen peroxide and BSA were obtained from Sigma (St. Louis, Missouri). Normal goat serum, secondary, and tertiary antibodies were obtained from Jackson ImmunoResearch Laboratories, Inc. (West Grove, Pennsylvania). All mouse work was approved by The Jackson Laboratory Institutional Animal Care and Use Committee.
Supported by NIH Grants DK32642 (to D.E.O), DK26657 (to the Clinical Nutrition Research Unit Protein and Immunology core), P30AR41943 (SDRCC, to LEK), AR43801 (to JPS), RR173 (to JPS), CA34196 (to JPS), the Wauford Foundation (to HBE) and the North American Hair Research Society (to HBE). DOI: 10.1111/j.0022-202X.2004.23223.x Manuscript received December 2, 2003; revised April 1, 2004; accepted for publication April 5, 2004 Address correspondence to: David E. Ong, Department of Biochemistry, Vanderbilt University Medical Center, Nashville, TN 37232. Email: [email protected]
References Alonso L, Fuchs E: Stem cells in the skin: Waste not, Wnt not. Genes Dev 17:1189–1200, 2003 Attar PS, Wertz PW, McArthur M, Imakado S, Bickenbach JR, Roop DR: Inhibition of retinoid signaling in transgenic mice alters lipid processing and disrupts epidermal barrier function. Mol Endocrinol 11:792–800, 1997 Blentic A, Gale E, Maden M: Retinoic acid signalling centres in the avian embryo identified by sites of expression of synthesising and catabolising enzymes. Dev Dyn 227:114–127, 2003 Bucco RA, Zheng WL, Davis JT, SierraRivera E, Osteen KG, Chaudhary AK, Ong DE: Cellular retinoic acid-binding protein (II) presence in rat uterine
263
Department of Biochemistry and wDivision of Dermatology, Vanderbilt University Medical Center, Nashville, Tennessee, USA; zThe Jackson Laboratory,
Bar Harbor, Maine, USA
Retinoic acid has long been known to alter skin and hair growth but an exact mechanism is unclear. This study was performed to examine the sites of endogenous retinoic acid synthesis in the cycling hair follicle to better understand the role retinoic acid plays in this process. Retinal dehydrogenases (Aldh1a1, 2, and 3, formerly Raldh 1, 2, and 3) are the enzymes responsible for the last step in retinoic acid synthesis. Immunohistochemistry was performed on adult C57BL/6J mouse skin sections with antibodies against Aldh1a2 and Aldh1a3. Aldh1a2 expression was seen primarily in the outer root sheath and basal/spinous layer during all stages of the hair cycle, and in the bulge during anagen and early catagen, whereas Aldh1a3 expression was primarily in the dermal papilla, pre-cortex, and hair shaft during mid–late anagen. The expression patterns of these two similar retinoic acid synthesizing enzymes at specific follicular sites suggest that they mediate and are regulated by different epithelial proliferation and differentiation signaling pathways.
Key words: sebaceous gland/bulb/bulge/hair follicle/immunohistochemistry J Invest Dermatol 123:258 – 263, 2004 Vitamin A and its derivatives (retinoids) are critically important in the development and maintenance of multiple tissues, including skin and hair, as shown by the detrimental effects of vitamin A deficiency (Wolbach and Howe, 1925; Frazier and Hu, 1931). These observations have led to the widespread use of retinoids in various skin disorders (Stuttgen, 1986). All-trans retinoic acid (atRA) was proven in vivo (Werner and DeLuca, 2001) and in vitro (Kurlandsky et al, 1994) to be the most physiologically relevant retinoid in the skin. But, due to toxicity issues, 13-cis retinoic acid (13cRA, i.e., Accutane, Roche Laboratories, Inc., Nutley, New Jersey), which can slowly isomerize to atRA, is used more frequently in clinically abnormal states (Stuttgen, 1986). The regulation of endogenous atRA synthesis and its precise tissue/cellular localization is a subject of active investigation. AtRA was shown to be synthesized in cultured keratinocytes (Kurlandsky et al, 1994) and degraded in dermal fibroblasts (Randolph and Simon, 1998). This degradation of atRA by the dermal fibroblasts would limit the delivery of physiological concentrations of plasma atRA
to the epidermis and hair follicle making local atRA synthesis essential for normal function. The synthesis of atRA from circulating retinol occurs by the action of two enzyme families (Napoli, 1999). Retinol dehydrogenases (Rdh) convert retinol to retinal, whereas retinal dehydrogenases (Aldh1a) convert retinal to RA. Two human epidermal Rdh family members, hRODH-E and hRODHE2, have been identified (Jurukovski et al, 1999; Markova et al, 2003). Currently, there are three known Aldh1a (1–3) that convert all-trans retinal to atRA. Recent studies with Aldh1a1 null mice suggest that this enzyme is involved in the catabolism of excess retinol (Fan et al, 2003). The expressions of the cellular retinol binding protein and cellular retinoic acid binding protein type II have also been associated with the ability to make RA (i.e., Bucco et al, 1997; Napoli, 1999). Therefore, expression of these binding proteins and enzymes can be used as markers for sites of RA synthesis. Recently, Markova et al (2003) showed that hRODH-E2 transcripts are expressed in sebaceous and sweat glands, matrix cells, and the Huxley cells of the inner root sheath (IRS) of the infundibulum. But little data are available on the expression pattern of Aldh1a in the hair follicle and sebaceous gland. Most of the studies with Aldh1a2 and Aldh1a3 have focused on embryogenesis. The temporal and spatial expression patterns of retinal dehydrogenases have been used as an initial step in dissecting the sites of RA synthesis in the mouse embryo, the regulation of RA synthesis, and the precise role of RA during embryonic development. The expression of Aldh1a2 and Aldh1a3 is seen at distinct times and locations throughout embryogenesis
Abbreviations: Aldh1a, retinal dehydrogenase; atRA, all-trans retinoic acid; 13cRA, 13-cis retinoic acid; ILORS, inner layer of the outer root sheath; IRS, inner root sheath; Lef/Tcf, lymphoid enhancer factor/T-cell factor; ORS, outer root sheath; Rara,b,c, retinoic acid receptor a, b, g; RA, retinoic acid; Shh, sonic hedgehog; Wnt, wingless related MMTV The mouse genome nomenclature was used (www.informatics. jax.org). Genes and RNA message expression is italicized, while proteins are not. First letter capitalized followed by lowercase indicates a mouse or rat gene, while all capital letters indicates the human gene or protein.
Copyright r 2004 by The Society for Investigative Dermatology, Inc.
258
123 : 2 AUGUST 2004
FOLLICULAR LOCALIZATION OF Aldh1a2 AND Aldh1a3
259
(Blentic et al, 2003). In the developing hair follicle Aldh1a3 expression was reported at the base of the hair follicle at 5 d, but not 3 wk, post-partum (Niederreither et al, 2002a). The sites of RA synthesis suggested in these expression studies were consistent with developmental problems in the null mice. Aldh1a2/ mice die in utero due to defects in heart development (Niederreither et al, 1999), and Aldh1a3/ mice die within 10 h of birth due to defects in nasal development (Dupe et al, 2003). These defects can be only partially restored with maternal RA administration, suggesting that precise timing of RA synthesis at specific sites is essential. Neither Aldh1a2 nor Aldh1a3 expression patterns in the sebaceous gland or hair follicle have been well described in the adult mammal. Thus, this report describes the expression pattern of Aldh1a2 and Aldh1a3 in the adult mouse hair follicle, as an initial step in dissecting the role of endogenous RA in mammalian hair cycling.
Results and Discussion To examine the possible sites of RA synthesis, immunohistochemistry was performed on adult C57BL/6J mouse skin with antibodies against Aldh1a2 and Aldh1a3. Dot blot analysis of these antibodies against recombinant proteins shows that these antibodies are specific (Fig 1). The antiAldh1a2 antibody did not cross-react with Aldh1a3. There was a slight cross-reaction of the anti-Aldh1a3 antibody with Aldh1a2, but this antibody showed a 64-fold selectivity for Aldh1a3 over Aldh1a2. Very different anatomical sites of Aldh1a2 versus Aldh1a3 expression were observed in mouse hair follicles and their associated sebaceous gland at various stages in the hair cycle (Tables I and II and Figs 2
Figure 1 Dot blot analysis of antibodies against recombinant whole proteins. Recombinant rat proteins Aldh1a2 (2) and Aldh1a3 (3) (0–25 pmol) were blotted onto nitrocellulose membranes, cut into strips, and exposed to affinity-purified antibodies against Aldh1a2 (a) or Aldh1a3 (b). Detection was by enhanced chemiluminescence. The image shows the two films for the two antibodies overlaying the strips of dot blotted protein.
and 3). Each follicular site reflects a specific function in the regulation of proliferation and differentiation during the hair cycle. This differential localization of two proteins with the
Table I. Aldh1a2 expression in the stages of the hair cyclea Stage
Basal Spin.
Seb.
ORS
Bulge
Prol. Ker.
ILORS/CL
ES
Germ Caps.
Telogen
þþþ
þþþ
þ /
NA
NA
NA
þþþ
Anagen I
þþ
þ /
þþ
þþ
þþ
NA
NA
NA
Anagen II
þþ
þ
þ
þþ
þþ
NA
NA
NA
Anagen IIIa
þþ
þ
þþ
þþ
þþ
NA
NA
NA
Anagen IIIb
þþ
þ /
þ þ in
þþ
þþ
þþ
NA
NA
Anagen IIIc
þþ
þ /
þ þ is
þþ
NA
þþ
NA
NA
Anagen IV
þþ
þ /
þ þ is,in
þþ
NA
þþ
NA
NA
Anagen V
þþ
þ /
þ þ þ þ bin
þþþ
NA
þþ
NA
NA
Anagen VI/catagen I
þ
þ /
0 þ bin
þþ
NA
þ /
NA
NA
Catagen II
þþ
0 þ þ bin
þþ
NA
þ
NA
NA
Catagen III
þþ
0 þ þ bin
þþ
NA
þþ
NA
NA
Catagen IV
þ
þ /
0 þ þ bin
þþ
NA
þþ
NA
NA
Catagen V
þþ
þ /
þþþ
þ /
NA
þþ
þ
þ
Catagen VI
þþ
þ /
þþþ
þ /
NA
þþ
þ
þ
Catagen VII
þþ
þþþ
þ /
NA
NA
þþ
þþ
Catagen VIII
þþ
þ /
þþþ
þ /
NA
NA
NA
þþ
a Negative (), faint ( þ /), weak ( þ ), strong ( þ þ ), and very strong ( þ þ þ ) staining is indicated. Basal/Spin, basal/spinous layer; Seb., sebocyte; ORS, outer root sheath; Prol. Ker, proliferating keratinocyte strand between dermal papilla and club hair in early anagen; ILORS/CL, inner layer of the outer root sheath/companion layer; ES, epithelial strand; Germ Caps., germ capsule; in, infundibulum; is, isthmus; b, bulb; NA, not applicable.
260 EVERTS ET AL
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY Table II. Aldh1a3 expression in the stages of the hair cyclea
Stage
Seb gland
Seb duct
ILORS/CL
Hair Fiber
Pre cortex
DP
Germ Caps.
Telogen
þ /
NA
Anagen I
þþ þ
NA
NA
NA
þ /
NA
Anagen II
þ
þþ þ
NA
NA
NA
NA
Anagen IIIa
þ
þ
NA
NA
þ
þ
NA
þþ
þþ
NA
Anagen IIIb
þþ
þþ
NA
Anagen IIIc
þþ
þþ þ
þþ
NA
Anagen IV
þþ
þ þ þ 0 b–in
þþ
þ
NA
Anagen V
þ þ þ 0 bin
þ
þþþ
NA
Anagen VI/catagen I
þ
þ þ þ 0 bin
þþþ
þ /
NA
Catagen II
þ þ þ 0 bin
þþ
NA
Catagen III
þ
þ / 0 bin
NA
Catagen IV
þ
þ / 0 bin
NA
NA
Catagen V
þ
þþ
NA
þ
Catagen VI
þ /
þ
þþ
NA
þ
IIIa–b
Catagen VII
þ /
NA
NA
þ
Catagen VIII
þ /
NA
NA
a Negative (), faint ( þ /), weak ( þ ), strong ( þ þ ), and very strong ( þ þ þ ) staining is indicated. Seb., sebocyte; ILORS/CL, inner layer of the outer root sheath/companion layer; Hair fiber (aka shaft); Precortex, cells just above the matrix; DP, dermal papilla; Germ Caps., germ capsule; in, infundibulum; b, bulb; NA, not applicable.
Figure 2 Aldh1a2 expression in adult mouse hair follicles. Immunohistochemical localization of Aldh1a2 in adult mouse hair follicles during telogen (a), anagen IIIa (b), anagen V (c), catagen V (d), and catagen VII (e). Bar ¼ 1 mm. Insets are higher magnifications of their respective follicles (arrows). Bar ¼ 0.1 mm all insets.
apparent same enzymatic function allows RA signaling to be differentially regulated at each unique site. Dermal papilla: induction and termination of anagen, catagen, and telogen There are three main stages of the hair cycle: anagen (growth); catagen (regression); and
telogen (rest). Aldh1a2 was expressed strongly during all stages of the hair cycle (Fig 2), whereas Aldh1a3 was expressed primarily during mid–late anagen (Fig 3). During telogen, Aldh1a2 was expressed strongly in the outer root sheath (ORS), bulge and germ capsule (Table I, Fig 2a, and inset i). In early anagen (I–IIIb) Aldh1a2 was expressed in the
123 : 2 AUGUST 2004
FOLLICULAR LOCALIZATION OF Aldh1a2 AND Aldh1a3
261
Figure 3 Aldh1a3 expression in adult mouse hair follicles. Immunohistochemical localization of Aldh1a3 in adult mouse hair follicles during anagen IIIa (a), IIIa–b (b), IIIc (c), V (d), and early VI (e). Bar ¼ 1 mm. Insets are higher magnifications of their respective follicles (small arrows). Bar ¼ 0.1 mm all insets. Large arrows point to the sebaceous gland duct.
bulge region and the proliferating keratinocytes between the club hair and the bulb region (Table I, Fig 2b, inset iii). Starting in mid-anagen, Aldh1a3 but not Aldh1a2 was expressed in the bulb region (Table II, Figs 2b and c, 3a–e, and Fig 3 insets ii, iv, v, vii–x and data not shown). This Aldh1a3 expression appeared to differ during the stages of anagen. In anagen IIIa–b, IIIc, and V Aldh1a3 localized to the dermal papilla, but then in anagen IIIa–b, IIIb, IIIc, and VI it was strongly expressed and anagen IV and V weakly expressed in the pre-cortical region (Table II, Fig 3a–e and inset ii, iv, v, vii–x and data not shown). The hair fiber (aka shaft) expressed Aldh1a3 beginning in anagen IIIb and increased in intensity through catagen II, then reduced its intensity in catagen III and IV as hair growth ceases (Table II, Fig 3c–e, insets v, viii, x, and data not shown). During late anagen a gradient of Aldh1a3 expression is seen in the fiber with the strongest expression seen in the pre-cortex region and progressively weaker expression is seen as the fiber reaches the skin surface (Fig 3e, insets x and xi). As was seen for telogen and early anagen, the bulge region continues to express Aldh1a2 throughout anagen (Table I, Fig 2c, inset v). In addition, the ORS expression of Aldh1a2 is strongest in the isthmus near the bulge throughout anagen, with a peak in expression seen during anagen V and a drop at anagen VI/catagen I (Table I, Fig 2c, inset v and data not shown). These expression patterns suggest that Aldh1a2 may be involved in the pathways responsible for overall maintenance of the hair follicle, whereas Aldh1a3 may be a positive regulator of anagen growth, through epithelial–mesenchymal interactions. The hair cycle is regulated through the interactions of growth factors between the bulge area stem cells, the dermal papilla, the epithelial matrix and pre-cortical cells (reviewed in Stenn and Paus, 2001). All of these sites expressed either Aldh1a2 or Aldh1a3 at some stage of the hair cycle, suggesting a role of RA in these regulatory
pathways. Two of these growth factor signaling systems, wingless related MMTV (Wnt) and sonic hedgehog (Shh), are regulated by RA in other systems (Easwaran et al, 1999; Power et al, 1999; Niederreither et al, 2002b), providing further support for a possible role of RA in hair cycle regulation. The localization of Aldh1a2 was consistent with that of the Wnt transcription factor T cell factor 3 (Tcf3) in the ORS and bulge, and the Shh receptor Ptc1 in the proliferating keratinocytes between the DP and club hair during early anagen and ORS in late anagen (Callahan and Oro, 2001; Alonso and Fuchs, 2003; Oro and Higgins, 2003). The expression of Aldh1a3 was similar to that of the Wnt transcription factor lymphoid enhancer factor 1 (Lef1) in the dermal papilla and pre-cortical cells, and Ptc 1 in the dermal papilla. Tcf3 and Lef1 are in the same family of transcription factors and it has been suggested that Tcf3 acts to inhibit transcription in the bulge during Wnt stimulation, whereas Lef1 activates transcription in the pre-cortex when stimulated by Wnt. This differential expression of Aldh1a and Lef/ TcF family members at different sites within the hair follicle would allow for site-specific regulation. Future studies could examine whether RA has different effects on Lef1 versus TcF3. In addition, do either Lef1 or TcF3 regulate Aldh1a2 or Aldh1a3? A role for RA in stem cell regulation was first noted by Wolbach and Howe in 1925. The expression of Aldh1a2 in the bulge region throughout anagen (Table I, Fig 2b and c, insets iii and v) provides further support for such a role of RA in stem cell maintenance. During all stages of catagen, Aldh1a2 expression occurs in the ORS, especially near the bulge (Table I, Fig 2d and e, insets vii and viii). But the outer cells of the bulge do not express Aldh1a2 in catagen V–VIII (Table I, Fig 2d and e, insets vii and viii). Aldh1a2 also localizes to the germ capsule cells as they form in catagen V–telogen (Table I, Fig 2a, d, and e, inset vi). Weak expression of Aldh1a3 was also seen in the germ capsule during late catagen (Table II, data
262 EVERTS ET AL
not shown). The epithelial strand also showed Aldh1a2 expression during catagen V–VII (Table I, Fig 2d and e). In the bulbar region Aldh1a3 expression is reduced as catagen progresses (Table II, data not shown). The dermal papilla only faintly expresses Aldh1a3 in anagen VI/catagen I (Fig 3e, inset ix) and is completely negative through catagen and telogen. The pre-cortex ceases to express Aldh1a3 by catagen III when these cells regress. Catagen is marked by an arrest of proliferation in the bulbar region and massive apoptosis (Stenn and Paus, 2001). Thus, loss of Aldh1a3 produced RA in the dermal papilla and Aldh1a2 produced RA in the bulge may be involved in this arrest in proliferation. In addition, added RA has been found to initiate apoptosis in differentiating non-tumorogenic mouse keratinocytes (Islam et al, 2000). The strong expression of Aldh1a2 within the regressing catagen hair follicle suggests that RA synthesis is involved in this catagen-induced apoptosis. To confirm this, future studies with Tunel, Caspase 3 double labeling on frozen sections can be done. Sheath–fiber interactions The inner layer of the outer root sheath (ILORS, aka companion layer) expressed Aldh1a2 strongly during anagen IIIb–V and catagen IV–VI, but weakly during anagen VI–catagen III (Table I, Fig 2c, inset iv). Aldh1a3 localized to the ILORS during anagen IV, anagen VI, and catagen III–VI (Table II, data not shown). The ORS expressed Aldh1a2 during all stages of the hair cycle with the strongest expression in anagen V in the isthmus and catagen V–telogen the whole length of the follicle (Table I, Fig 2a, c, and e, insets i, v–viii). After anagen initiation occurs, the newly made IRS and fiber must traverse the ORS to move to the skin surface (Stenn and Paus, 2001). This is thought to involve a proteolytic processs and the expression of the tissue inhibitor of metalloproteinases (TIMP) in Henle’s layer of the IRS and plasminogen activator inhibitor-2 (PAI-2) in the ILORS. The expression of Aldh1a3 in the ILORS suggests a role for RA in this process. This hypothesis is supported by recent microarray data in our laboratory that found both TIMP and PAI-2 to be upregulated by RA in the rat uterus (Li and Ong, unpublished observation). The sebaceous glands expressed both Aldh1a2 and Aldh1a3 during anagen II–IIIa (Tables I and II, Fig 2b, inset iii, Fig 3a and b, insets i, iii). There was light color in the sebaceous glands at other times, but that may be nonspecific as a light background was sometimes noted (Fig 3c, inset vi). Aldh1a3 also localized to cells lining the sebaceous gland duct during anagen I–IIIa and catagen V– telogen (Table II, Fig 3a and b, insets i and iii large arrow, and data not shown). Thus, RA synthesis may occur within the sebocyte or in the isthmus where the sebaceous gland duct opens and affect sebaceous gland function. In vitro studies have shown that a factor produced by the sebaceous gland, duct, or associated isthmus is required for the hair fiber to be released from the IRS (Williams and Stenn, 1994; Williams et al, 1996). These studies have been supported in vivo by studies with the asebia mouse, which have small, hypoplastic sebaceous glands, defects in the release of the hair fiber, and progressive scarring alopecia (Sundberg et al, 2000). RA has been shown to be important for sebaceous gland function and hair sheath–fiber interac-
THE JOURNAL OF INVESTIGATIVE DERMATOLOGY
tions in a concentration-dependent manner (Zouboulis et al, 1993; Williams et al, 1996). Vitamin A deficiency in cultured human sebocytes results in decreased cell proliferation and lipogenesis that can be partially restored with low doses of 13cRA. At concentrations above 107 M 13cRA and atRA inhibit proliferation, lipogenesis, and sheath growth (Zouboulis et al, 1993; Williams et al, 1996). The localization of Aldh1a3 to sites just proximal to the sebaceous gland duct and Aldh1a2 and Aldh1a3 within the sebocyte suggests that RA made at these sites are important for sebocyte function and the normal growth and release of the hair fiber. Interfollicular epidermis and presumptive water barrier Lipids made within the sebocyte are also essential for water barrier formation (Madison, 2003). Thus, Aldh1a3 synthesized RA may play a role in water barrier formation by altering lipid metabolism in the sebocyte. This role of RA in sebocytes is supported by studies of dominant negative Rara transgenic mice, which have defects in water barrier formation and lipid metabolism (Attar et al, 1997). Also, RA has been found to activate PPAR b/d (Shaw et al, 2003), whereas the drug, clofibrate, has been found to increase barrier function (Hanley et al, 1999). Aldh1a2 is localized to the ORS of the infundibulum in all stages (Table I, Fig 2a–e, inset ii). But Aldh1a3 was not seen in the ORS of the infundibulum at any stage of the hair cycle (Fig 3a–e, data not shown). The expression of Aldh1a2 within the basal/spinous layers suggested that RA is also involved in differentiation in these cells directly and not just in sebocyte lipid production. The fact that RA synthesis at these two sites involved in water barrier formation occurs by two separate enzymes suggests that they are differentially regulated. This report demonstrates that Aldh1a2 and Aldh1a3 are expressed at different sites within the cycling hair follicle. Their expression patterns suggest that locally produced RA is important in all stages of the mammalian hair cycle. Aldh1a3 expression may be involved in epithelial–mesenchymal interactions, whereas Aldh1a2 expression may be involved in stem cell maintenance. Having different RA synthesizing enzymes involved at each site and stage allow differential regulation of each step in the hair cycle. Future experiments will examine more closely the regulation of these enzymes, as well as other components of RA synthesis, during the normal hair cycle. Aldh1a expression patterns in mice null for Wnts and Shh will help to establish regulatory pathways in normal and abnormal skin and hair growth. Materials and Methods Tissue acquisition Archived paraffin blocks of wax stripped dorsal skin (used from previous hair cycle studies) from C57BL/ 6J wild-type mice fixed with Tellyzneski–Fekete’s acid–alcohol– formalin fixative were sectioned at 6 mm. The Paus system of determining the stages of the hair cycle was used without additional staining (Muller-Rover et al, 2001). Antibodies Rabbit polyclonal antibodies against Aldh1a2 and 3 were developed. For the anti-Aldh1a2 antibody the peptide GGKGLGRKGFFIEP was chosen. This peptide was synthesized by PeptidoGenic (Livermore, California), conjugated to Imject Activated KLH (Pierce, Rockford, Illinois), and injected intradermally into rabbits with Hunter’s Titermax Gold (CytRx Corp., Norcross,
123 : 2 AUGUST 2004
FOLLICULAR LOCALIZATION OF Aldh1a2 AND Aldh1a3
Georgia). Antiserum was first purified on a Protein A Sepharose column (Pierce, Rockford, Illinois) to obtain an immunoglobulin (Ig) G fraction. Then the IgG fraction was affinity purified using a histidine-tagged whole protein rat Aldh1a2 (generous gift of Marcia Newcomer, PhD Dept. Biol. Sci., LSU) immobilized on a CNBractivated Sepharose 4B column (Amersham Biosciences, Piscataway, New Jersey). The Aldh1a3 antibody was made by a similar procedure except a recombinant whole rat protein was used for the immunization and it was affinity purified using the recombinant whole protein immobilized on a Sulfolink column (Pierce, Rockford, Illinois).
epithelial cells correlates with their synthesis of retinoic acid. Biochemistry 36:4009–4014, 1997 Callahan CA, Oro AE: Monstrous attempts at adnexogenesis: Regulating hair follicle progenitors through Sonic hedgehog signaling. Curr Opin Genet Dev 11:541–546, 2001 Dupe V, Matt N, Garnier JM, Chambon P, Mark M, Ghyselinck NB: A newborn lethal defect due to inactivation of retinaldehyde dehydrogenase type 3 is prevented by maternal retinoic acid treatment. Proc Natl Acad Sci USA 100:14036–14041, 2003 Easwaran V, Pishvaian M, Shah S, Byers S: Cross-regulation of beta-catenin-Lef/ Tcf and retinoid signaling pathways. Curr Biol 9:1415–1418, 1999 Fan X, Molotkov A, Manabe S, et al: Targeted disruption of Aldh1a1 (Raldh1) provides evidence for a complex mechanism of retinoic acid synthesis in the developing retina. Mol Cell Biol 23:4637–4648, 2003 Frazier CN, Hu C-K: Cutaneous lesions associated with a deficiency in vitamin A in man. Arch Intern Med 48:507–514, 1931 Hanley K, Komuves LG, Bass NM, et al: Fetal epidermal differentiation and barrier development in vivo is accelerated by nuclear hormone receptor activators. J Invest Dermatol 113:788–795, 1999 Islam TC, Skarin T, Sumitran S, Toftgard R: Retinoids induce apoptosis in cultured keratinocytes. Br J Dermatol 143:709–719, 2000 Jurukovski V, Markova NG, Karaman-Jurukovska N, Randolph RK, Su J, Napoli JL, Simon M: Cloning and characterization of retinol dehydrogenase transcripts expressed in human epidermal keratinocytes. Mol Genet Metab 67:62–73, 1999 Kurlandsky SB, Xiao JH, Duell EA, Voorhees JJ, Fisher GJ: Biological activity of all-trans retinol requires metabolic conversion to all-trans retinoic acid and is mediated through activation of nuclear retinoid receptors in human keratinocytes. J Biol Chem 269:32821–32827, 1994 Madison KC: Barrier function of the skin: ‘‘La raison d’etre’’ of the epidermis. J Invest Dermatol 121:231–241, 2003 Markova NG, Pinkas-Sarafova A, Karaman-Jurukovska N, Jurukovski V, Simon M: Expression pattern and biochemical characteristics of a major epidermal retinol dehydrogenase. Mol Genet Metabol 78:119–135, 2003 Muller-Rover S, Handjiski B, van der Veen C, et al: A comprehensive guide for the accurate classification of murine hair follicles in distinct hair cycle stages. J Invest Dermatol 117:3–15, 2001 Napoli JL: Interactions of retinoid binding proteins and enzymes in retinoid metabolism. Biochim Biophys Acta 1440:139–162, 1999 Niederreither K, Fraulob V, Garnier JM, Chambon P, Dolle P: Differential expression of retinoic acid-synthesizing (Raldh) enzymes during fetal development and organ differentiation in the mouse. Mech Dev 110:165– 171, 2002a Niederreither K, Subbarayan V, Dolle P, Chambon P: Embryonic retinoic acid synthesis is essential for early mouse post-implantation development. Nat Genet 21:444–448, 1999 Niederreither K, Vermot J, Schuhbaur B, Chambon P, Dolle P: Embryonic retinoic acid synthesis is required for forelimb growth and anteroposterior patterning in the mouse. Development 129:3563–3574, 2002b Oro AE, Higgins K: Hair cycle regulation of hedgehog signal reception. Dev Biol 255:238–248, 2003 Power SC, Lancman J, Smith SM: Retinoic acid is essential for Shh/Hoxd signaling during rat limb outgrowth but not for limb initiation. Dev Dyn 216:469–480, 1999 Randolph RK, Simon M: Dermal fibroblasts actively metabolize retinoic acid but not retinol. J Invest Dermatol 111:478–484, 1998 Shaw N, Elholm M, Noy N: Retinoic acid is a high affinity selective ligand for the peroxisome proliferator-activated receptor b/d. J Biol Chem 278:41589– 41592, 2003 Stenn KS, Paus R: Controls of hair follicle cycling. Physiol Rev 81:449–494, 2001 Stuttgen G: Historical perspectives of tretinoin. J Am Acad Dermatol 15:735–740, 1986 Sundberg JP, Boggess D, Sundberg BA, Eilertsen K, Parimoo S, Filippi M, Stenn K: Asebia-2J (Scd1ab2J): A new allele and a model for scarring alopecia. Am J Pathol 156:2067–2075, 2000 Werner EA, DeLuca HF: Metabolism of a physiological amount of all-trans-retinol in the vitamin A-deficient rat. Arch Biochem Biophys 393:262–270, 2001 Williams D, Siock P, Stenn K: 13-cis-retinoic acid affects sheath-shaft interactions of equine hair follicles in vitro. J Invest Dermatol 106:356–361, 1996 Williams D, Stenn KS: Transection level dictates the pattern of hair follicle sheath growth in vitro. Dev Biol 165:469–479, 1994 Wolbach SB, Howe PR: Tissue changes following deprivation of fat-soluble A vitamin. J Exp Med 42:753–777, 1925 Zouboulis CC, Korge BP, Mischke D, Orfanos CE: Altered proliferation, synthetic activity, and differentiation of cultured human sebocytes in the absence of vitamin-A and their modulation by synthetic retinoids. J Invest Dermatol 101:628–633, 1993
Dot blot analysis Recombinant whole protein rat Aldh1a2 was a generous gift of Marcia Newcomer, PhD Dept. Biol. Sci., LSU. Histidine-tagged whole protein rat Aldh1a3 was expressed in Escherichia coli and purified using Ni Agarose according to the manufacturer’s instructions (Qiagen, Valencia, California). These proteins (0–25 pmol) were blotted onto a nitrocellulose membrane (BioRad, Hercules, California) using the Bio-Dot microfiltration apparatus (BioRad) according to the manufactures instructions. Strips of dots blots were cut and blocked with 5% dry milk in Tris buffered saline containing 0.1% Tween. Blots were then incubated in primary antibody for 1 h, washed with Tris buffered saline with 0.1% Tween, incubated with an antirabbit antibody conjugated with horseradish peroxidase (Jackson ImmunoResearch Laboratories, Inc., West Grove, Pennsylvania), washed again and detected with enhanced chemiluminescence (Amersham Biosciences). Immunohistochemistry Sections were pre-treated with 3% hydrogen peroxide, blocked with 3% BSA plus 1.28% normal goat serum, incubated with the primary antibody overnight at 41C, then incubated with a biotin conjugated secondary antibody, an antibiotin IgG conjugated with horseradish peroxidase, and then stained with AEC þ (Dako, Carpinteria, California; Bucco et al, 1997). Hydrogen peroxide and BSA were obtained from Sigma (St. Louis, Missouri). Normal goat serum, secondary, and tertiary antibodies were obtained from Jackson ImmunoResearch Laboratories, Inc. (West Grove, Pennsylvania). All mouse work was approved by The Jackson Laboratory Institutional Animal Care and Use Committee.
Supported by NIH Grants DK32642 (to D.E.O), DK26657 (to the Clinical Nutrition Research Unit Protein and Immunology core), P30AR41943 (SDRCC, to LEK), AR43801 (to JPS), RR173 (to JPS), CA34196 (to JPS), the Wauford Foundation (to HBE) and the North American Hair Research Society (to HBE). DOI: 10.1111/j.0022-202X.2004.23223.x Manuscript received December 2, 2003; revised April 1, 2004; accepted for publication April 5, 2004 Address correspondence to: David E. Ong, Department of Biochemistry, Vanderbilt University Medical Center, Nashville, TN 37232. Email: [email protected]
References Alonso L, Fuchs E: Stem cells in the skin: Waste not, Wnt not. Genes Dev 17:1189–1200, 2003 Attar PS, Wertz PW, McArthur M, Imakado S, Bickenbach JR, Roop DR: Inhibition of retinoid signaling in transgenic mice alters lipid processing and disrupts epidermal barrier function. Mol Endocrinol 11:792–800, 1997 Blentic A, Gale E, Maden M: Retinoic acid signalling centres in the avian embryo identified by sites of expression of synthesising and catabolising enzymes. Dev Dyn 227:114–127, 2003 Bucco RA, Zheng WL, Davis JT, SierraRivera E, Osteen KG, Chaudhary AK, Ong DE: Cellular retinoic acid-binding protein (II) presence in rat uterine
263