Induction of Collagen by Estradiol: Difference Between Sun-Protected and Photodamaged Human Skin In Vivo

Objective  To evaluate the effectiveness of topical estradiol in stimulating collagen I and III production in naturally aged and photoaged human skin of postmenopausal women and age-matched men.

Design  Vehicle-controlled treatment followed by biochemical and immunohistochemical analyses of skin biopsy specimens.

Setting  Academic referral center.

Participants  Seventy healthy volunteers (40 postmenopausal women with a mean age of 75 years, and 30 men with a mean age of 75 years) with photodamaged skin.

Interventions  Topical application of estradiol, 0.01%, 0.1%, 1%, or 2.5% or vehicle on aged or photoaged skin, with biopsy specimens taken after last treatment.

Main Outcome Measures  De novo synthesis of collagen by quantitative polymerase chain reaction, immunohistochemistry, and enzyme-linked immunosorbent assay.

Results  Topical estradiol increased procollagen I and III messenger RNA and collagen I protein levels in sun-protected aged hip skin in postmenopausal women and, to a lesser extent, in age-matched men. Surprisingly, no significant changes in production were observed in women or men after 2-week estradiol treatment of photoaged forearm or face skin, despite similar expression of estrogen receptors (ER-α, ER-β, and GPR30) in aged and photoaged skin. Estradiol treatment induced the estrogen-responsive gene GREB1, indicating that penetration of topical estradiol and genomic response to estrogen were similar in the 3 anatomic sites.

Conclusions  Two-week topical estradiol treatment stimulates collagen production in sun-protected hip skin, but not in photoaged forearm or face skin, in postmenopausal women and aged-matched men. These findings suggest that menopause-associated estrogen decline is involved in reduced collagen production in sun-protected skin. Interestingly, alterations induced by long-term sun exposure hinder the ability of topical 2-week estradiol to stimulate collagen production in aged skin.

Trial Registration  clinicaltrials.gov Identifier: NCT00113100

Skin aging is associated with reduced skin function, increased skin fragility, and compromised wound healing.¹ On areas of the body that are often not covered by clothing (ie, face, nape of the neck, hands, and forearms), long-term UV exposure from the sun damages the skin, causing it to look prematurely old. Photoaging is the main process responsible for the worsening appearance of aged skin.

Clinical signs of natural aging include fine wrinkles, skin laxity, and sagging, while photoaged skin appears dry, with coarse wrinkles and uneven pigmentation. Histologically, epidermis is thin in aged skin and rather irregular in photoaged skin.² However, both processes share major biochemical features such as reduced collagen content in the dermis, resulting from increased collagen degradation and decreased procollagen synthesis.³ Collagen I is the major structural protein of the dermis, and it provides strength and resiliency to the skin. Collagen I is primarily produced by dermal fibroblasts. Collagen production is regulated by a variety of mediators including growth factors, cytokines, hormones, and mechanical tension.⁴ In naturally aged skin, the combination of reduced fibroblast number, reduced fibroblast metabolic activity, and loss of mechanical tension results in a 70% decrease in new collagen (procollagen) production by fibroblasts.⁵ In photoaged skin, loss of mechanical tension appears to be primarily responsible for reduced procollagen production.^6-8 This loss of mechanical tension is due to accumulation of fragmented collagen, which is generated by repeated exposures to UV radiation (reviewed by Rittié et al⁹). Overall, decreased quantity and quality of collagen fibers in the dermis are associated with aged appearance of human skin. Hence, antiaging therapies include drugs (such as tretinoin) or procedures (such as carbon dioxide laser resurfacing) that stimulate collagen production in human skin^9,10 and thereby improve skin strength, resiliency, and function.

Estrogens exert a profound influence on skin, as highlighted by regressive cutaneous changes that occur after menopause in women. Diminution of circulating estrogens that accompanies menopause is associated with increased skin dryness¹¹ and slackness,¹² and decreased skin elasticity,¹³ dermal thickness of thigh skin,¹⁴ and skin collagen content as measured in thigh,¹⁵ abdomen,¹⁶ and pubis.¹⁷ Conversely, various studies have shown that estrogen therapy (ET) users have better skin hydration^18-21 and elasticity,^20-23 higher density of collagen fibers,²⁴ fewer fine wrinkles,^19,21 and thicker skin^20,25 than nonusers. However, other studies have failed to show any effect of ET on the prevalence or severity of cutaneous signs of menopause on the photoaged face.¹¹ Reasons for this discrepancy are unknown.

Effects of estrogens on skin collagen content are also not clear. While some studies demonstrated that ET significantly increased skin and dermal thickness in postmenopausal women^26-28 (as measured on thigh and abdomen), a study of 3875 women did not demonstrate any impact of ET on postmenopausal-associated skin atrophy¹⁹ (unspecified site). Similarly, some studies demonstrated greater skin collagen content after ET^{15,16,24,27-29} (as measured on thigh, lower abdomen, upper inner arm, and unspecified site), while others did not show any beneficial effects of ET on collagen synthesis^30,31 (upper arm and unspecified site).

A few studies have used topical estrogens on human skin. Topical estradiol (17β-estradiol) induces procollagen in abdominal skin of postmenopausal women (38% increase in total collagen³² and hydroxyproline³¹ content in estradiol-treated skin after 3 months) and hip skin of elderly men and women (58% increase in procollagen after the use of estradiol cream, 0.01%, twice a week for 2 weeks).³³ When applied to the photoaged face, estrogens increase skin thickness (7.8% increase vs placebo),³⁴ increase epidermal thickness,³⁵ and decrease fine wrinkles³⁴ but have no effect on epidermal hydration.³⁶ To our knowledge, the effects of topical estrogens on procollagen production in photoaged skin have not been reported. However, despite the lack of scientific data, estrogen-based treatments are largely believed to be beneficial for maintaining a youthful appearance of photoaged skin.^19,37-41

Estrogens mediate their effects by interacting with and activating specific estrogen receptors (ERs), which function as ligand-activated transcription factors.⁴² Activated ERs bind to specific regions in target gene promoters and modulate their expression.⁴² Recently, estrogens have been described to transmit some of their effects through membrane-associated G-coupled protein receptor GPR30,⁴³ the presence of which in skin has not been reported. Mechanisms by which estradiol regulates collagen production in human skin remain largely unknown. Procollagen I gene promoter lacks estrogen response elements,⁴⁴ and a direct effect of estrogens on procollagen I gene expression has yet to be described.

In the present study, we quantified procollagen after vehicle or estradiol treatment of healthy volunteers (postmenopausal women and age-matched men) on sun-protected or photoaged skin in vivo. We found that 1-week topical estradiol treatment increased procollagen I and III expression in sun-protected hip skin in a dose-dependent manner. Surprisingly, 2-week estradiol treatment had no effect on procollagen synthesis in photoaged skin (forearm or face). We found that ERs were expressed at similar levels in sun-protected and photoaged skin and that estradiol penetration and genomic activity were similar in all skin sites, as assessed by quantification of estrogen target gene expression. Taken together, our results demonstrate that estradiol-mediated collagen stimulation is indirect and that alterations induced by long-term sun exposure hinder the ability of topical estradiol to stimulate collagen production in aged human skin in vivo.

All procedures involving human subjects were approved by the University of Michigan Institutional Review Board, and all subjects provided written informed consent. Healthy volunteers, not taking ET, were enrolled in the study with no racial or ethnic distinction. Participants' age distribution, degree of photodamage, and years after menopause are given in Table 1.

Volunteers were treated with estradiol and vehicle (95% ethanol–propylene glycol, 7:3, vol/vol), under occlusion, on sun-protected hip or photodamaged forearm skin sites. Treatment (0.1%-2.5%, ie, 5.4-135 μg of estradiol per square centimeter) was administered 3 times, every other day. Twenty-four hours after the last treatment, 4-mm full-thickness punch biopsy specimens were obtained from each site. For face treatment, estradiol was freshly dissolved in propylene glycol and incorporated into moisturizing cream (Neutrogena; Neutrogena Corp, Los Angeles, California) to a final concentration of 0.2% (wt/vol). Estradiol treatment (570 μL of cream on the whole surface of the face; approximately 5.7 μg of estradiol per square centimeter) was applied twice a day for 14 days (participants washed their face with soap and water before applying the cream in the morning and at bedtime). Biopsy specimens (2 mm) were taken in the crow’s-foot area of the face before and 24 hours after the last treatment. For photoaging studies, all participants were asked to avoid sun exposure at least 48 hours before and during treatment of photoaged areas. Face studies (without occlusion) were performed in Michigan between November and February to limit sun exposure during treatment.

Immediately after biopsy, skin samples were embedded in low-temperature embedding medium (Tissue-Tek OCT compound; Miles, Naperville, Illinois), frozen in liquid nitrogen, and stored at −80°C until processing.

Total RNA was extracted from whole punch skin biopsy specimens by means of a commercial kit (RNeasy; Qiagen, Chatsworth, California) and quantified with dye (RiboGreen; Invitrogen, Carlsbad, California). Real-time reverse transcription–polymerase chain reaction (RT-PCR) was performed on 100 ng of total RNA as described previously,⁴⁵ with the use of custom primers and probe for collagen I and 36B4,⁴⁵ and for collagen III (COL3A1 sense primer, 5′-TCT-TGG-TCA-GTC-CTA-TGC-GGA-TA-3′; antisense primer, 5′-TCC-TAG-GCA-AGA-GAC-GCT-AC-3′; probe, 5′-CCA-GAA-CCA-TGC-CAA-ATA-TGT-GTC-TGT-GAC-T-3′). All other primer-probe sets were validated gene expression assays (TaqMan; Applied Biosystems, Foster City, California). Efficiency of real-time RT-PCR reactions was determined by using complementary DNA standards for ER-α (ESR1, UniGene Hs.208124) and ER-β (ESR2, UniGene Hs.660607) (generous gifts from Peter J. Kushner, PhD, University of California, San Francisco). Results were normalized to the level of housekeeping gene 36B4 (UniGene Hs.546285) (internal control). Results are presented as normalized fold change in estradiol-treated vs vehicle or untreated skin sample, or fold vs 36B4 (= 2^{−[CTtarget−CT36B4]}, where CT indicates cycle threshold and is the end point of the real-time PCR reaction).

Serial frozen sections of 7, 200, and 7 μm were prepared from OCT-embedded skin biopsy specimens. Dermal areas of 7-μm sections were measured with Image ProPlus software (Media Cybernetics, Bethesda, Maryland) and used to calculate the volume of the 200-μm sample. Soluble proteins were extracted from the 200-μm sample in ice-cold extraction buffer (50mM Tris hydrochloride, pH 7.4; 0.15M sodium chloride; 1% Triton X-100; protease inhibitors [Complete Mini; Roche Diagnostics, Indianapolis, Indiana]). After 5-minute centrifugation at 10 000g at 4°C, supernatants were assayed for procollagen I by means of a commercial enzyme-linked immunosorbent assay kit (Panvera, Madison, Wisconsin). This enzyme-linked immunosorbent assay, which detects the C-terminal domain of procollagen I, quantifies only newly made collagen I. Procollagen I concentrations were normalized to the volume of tissue used for each sample.

Frozen skin sections (7 μm) were fixed in 2% paraformaldehyde and immunohistochemistry was performed by means of anti–procollagen I antibody (Millipore, Billerica, Massachusetts) and an immunohistochemistry detection system (Link-Label; BioGenex Laboratories Inc, San Ramon, California). Slides were counterstained with hematoxylin (BioCare, Concord, California) and mounted with medium (Supermount; BioGenex).

Laser capture microdissection was performed as previously described⁴⁶ to separate interfollicular epidermis (without hair follicle infundibulum), dermis, and appendages (whole hair follicles, sebaceous glands, and sweat glands in a defined length of section).

Blood estradiol measurements were obtained by radioimmunoassay by the Clinical Chemistry Laboratory at the University of Michigan Hospital.

Data are expressed as mean and standard error of the mean. Comparisons among groups were made with the paired t test. All P values are 2-tailed and were considered significant when less than .05.

To assess the effects of topical estradiol on procollagen I and III expression, postmenopausal women were treated on hip skin with vehicle or various doses of estradiol for 1 week, and procollagen messenger RNA (mRNA) and protein levels were measured in skin biopsy specimens. As shown in Figure 1A, estradiol increased procollagen I and III mRNA levels in a dose-dependent manner: 0.1%, 1%, and 2.5% increased procollagen I mRNA levels by 2.87-, 3.54-, and 3.67-fold vs vehicle, respectively (n = 6-18; P < .05 for all concentrations). Similarly, 0.1%, 1%, and 2.5% estradiol treatment increased procollagen III mRNA levels by 2.10-, 3.19-, and 3.25-fold vs vehicle, respectively (P < .05 for all concentrations). A concentration of 0.01% estradiol did not significantly increase procollagen I or III mRNA levels (n = 6; P = .14 and .45, respectively). Estradiol treatment was also accompanied by a dose-dependent increase in procollagen I protein (Figure 1B): 0.1%, 1%, and 2.5% estradiol stimulated procollagen I production by 1.53-, 2.33-, and 2.75-fold vs vehicle, respectively (n = 6; P < .05 for 1% and 2.5% estradiol).

Collagen-producing cells were localized by immunohistochemistry. As shown in Figure 2, estradiol, 1%, treatment induced a marked increase in procollagen I protein expression throughout the reticular and papillary dermis in sun-protected skin of postmenopausal women in vivo.

Histologic analyses of human skin sections did not show any effect of topical estradiol treatment on epidermal or dermal thickness, as measured by image analysis on hematoxylin-eosin–stained skin sections (Figure 2 and data not shown).

Topical estradiol also increased procollagen expression in sun-protected hip skin of men, although to a lesser extent. As shown in Figure 3A, procollagen I mRNA levels were increased by topical estradiol by 1.69-, 1.58-, and 2.02-fold vs vehicle after 0.1%, 1%, and 2.5% estradiol treatment, respectively (n = 12; P < .05 for 0.1% and 1% estradiol). Similarly, procollagen III mRNA levels were increased by topical estradiol by 1.87-, 1.64-, and 2.11-fold vs vehicle after 0.1%, 1%, and 2.5% estradiol treatment, respectively (n = 12; P < .05 for all concentrations). Topical estradiol also induced procollagen I protein (Figure 3B) by 1.25-, 1.3-, and 2.3-fold vs vehicle after 0.1%, 1%, and 2.5% estradiol treatment, respectively (n = 9; P < .05 for 1% and 2.5%).

To determine the effect of estradiol on procollagen expression in photoaged skin, we next treated postmenopausal women on photodamaged forearms with vehicle or estradiol for 1 week and measured procollagen I and III mRNA levels in skin biopsy specimens by real-time RT-PCR. As shown in Figure 4A, 1% estradiol treatment did not significantly alter procollagen I or III mRNA levels in photodamaged forearm skin of postmenopausal women in vivo (n = 18). In contrast, estradiol treatment of sun-protected hip skin increased procollagen I mRNA levels by 3.54 (0.56)–fold (mean [SEM]) vs baseline (n = 18; P < .05, Figure 1A). Similarly, 2.5% estradiol did not alter procollagen mRNA levels in photodamaged skin of postmenopausal women (1.61 [0.44]– and 1.37 [0.16]–fold vs vehicle for procollagen I and III, respectively; n = 6). Because type I collagen regulation includes multiple posttranslational steps,⁴⁷ procollagen I protein levels were measured in estradiol- and vehicle-treated forearm skin samples of postmenopausal women. As shown in Figure 4B, estradiol treatment had no significant effect on procollagen I protein levels (n = 12).

Results obtained in photoaged forearm skin of men were similar to those obtained in postmenopausal women. Topical treatment with 1% or 2.5% estradiol did not alter procollagen I or III mRNA expression (Figure 4C; n = 12) or procollagen I protein (Figure 4D; n = 6) compared with vehicle.

Lack of effect of estradiol on collagen levels in photodamaged forearm skin was confirmed by immunohistochemistry in postmenopausal women (Figure 5) and age-matched men (not shown).

To determine whether the observed lack of effect of estrogen on collagen synthesis in photoaged forearms is related to anatomic site or photoaging, we evaluated the effect of topical estradiol on procollagen synthesis in the photoaged face. To this end, 5 postmenopausal women and 5 age-matched men with clinical photoaging were treated on the face with 0.2% estradiol cream twice daily for 2 weeks. The treatment regimen matched the quantity of estradiol per surface area that was used on forearms and hips. Levels of procollagen were compared in skin samples obtained before and after estradiol application. No change in procollagen mRNA (Figure 6A) or protein (Figure 6B) expression was observed after 2-week treatment of photoaged faces of men or postmenopausal women. In parallel, blood estradiol concentrations were measured before and after treatment. Topical estradiol treatment significantly increased estradiol blood levels in all treated subjects (Figure 6C). The increase in estradiol blood levels was greater in postmenopausal women than in men (12- vs 2-fold, respectively; P = .006 and P = .001, respectively). These results demonstrate that the 0.2% estradiol penetrated the skin and was chemically stable throughout the treatment.

To determine whether lack of response to estradiol treatment in photoaged skin was due to a lower level of ERs, we quantified ER-α, ER-β, and GPR30 (GPER; UniGene Hs.20961) mRNA levels in vehicle- and estradiol-treated skin samples from hip, forearm, and face of postmenopausal women and age-matched men. The ER-α, ER-β, and GPR30 mRNA levels were readily detected in human skin of all tested subjects (Figure 7). We did not find any differences in the levels of expression of ER-α (Figure 7A), ER-β (Figure 7B), or GPR30 (Figure 7C) between sun-protected hip and photoaged forearm or face skin samples, in women or men, nor were these levels affected by estradiol treatment. Interestingly, men had significantly less ER-α in hip skin (−49%) and less ER-β in forearm skin (−49%) than did site-matched women (n = 6; P < .05). Direct comparison of Figure 7A, B, and C shows that ER-α mRNA levels were approximately 10 times more abundant than those of ER-β or GPR30 in both men and women.

Despite readily detectable ER mRNA levels, we were not able to reliably detect the presence of ER protein in skin samples by immunohistochemistry, using various different antibodies (data not shown). We therefore used laser capture microdissection to localize ER mRNA expression in normal human skin. Expression of each of the 3 genes was extremely low in the epidermis (Figure 7D). The ER-α and GPR30 were primarily expressed in dermal cells, while expression was near the limit of detection in appendages. In contrast, ER-β expression was localized in appendages and, to a lesser extent, in dermal cells. These data indicate that all ERs are expressed in the dermal compartment of human skin in vivo.

Sex steroid concentrations in human tissues are tightly regulated, and excess estrogens are biologically inactivated by specific enzymes.⁴⁸ To determine whether lack of response to estradiol treatment in photoaged skin was due to increased estrogen catabolism, we measured the levels of estrogen inactivating enzymes in human skin treated with vehicle or estradiol. Estrogen-inactivating enzymes can be subdivided into 3 groups: (1) 17-β-hydroxysteroid dehydrogenases (HSD17B) 2, 4, and 8 catalyze oxidation of the biologically active estradiol into weaker estrone⁴⁹; (2) sulfotransferase (SULT) 1A1, SULT2A1, and SULT1E1 catalyze sulfation of estradiol, preventing it from binding to, and thereby activating, ERs^50,51; and (3) cytochrome P450 1B1 (CYP1B1) is the major CYP450 enzyme inactivating estradiol by 4-hydroxylation.⁵² We first assessed gene expression of each of these enzymes in hip, forearm, and face skin in vivo. The HSD17B4 was the most highly expressed (Table 2). The HSD17B2, HSD17B8, CYP1B1, and SULT1E1 were expressed at similar levels, approximately 5 times less than HSD17B4. The SULT1A1 mRNA levels were 3000 times lower than those of HSD17B4, whereas SULT2A1 mRNA was not detected in human skin in vivo (not shown).

As shown in Table 2, mRNA levels of estradiol metabolizing enzymes were similar in hip and forearm skin and were not altered by topical estradiol treatment in postmenopausal women. Expression of HSD17B2 tended to be higher in face skin than in hip skin, although this difference (approximately 10-fold) did not reach statistical significance (P = .09 and .06 for vehicle- and estradiol-treated sites, respectively; n = 5-6). In addition, CYP1B1 mRNA levels were 2.7-fold higher in face skin than in hip skin (P = .04). The level of HSD17B8 was approximately 4 times lower in face skin than in hip skin of postmenopausal women (P < .001).

All of the estradiol metabolizing enzymes that were expressed in women were also expressed in men, and their levels were not altered by estradiol treatment of hip, forearm, or face skin. Interestingly, levels of CYP1B1 mRNA were significantly higher in hip and forearm skin of men compared with women (3.7- and 3.8-fold, respectively; P = .03 for hip).

To assess topical estradiol biological activity, we measured induction of GREB1 (gene regulated by estrogen in breast cancer 1; Unigene Hs.467733), an estrogen-responsive gene that is directly regulated by ERs.⁵³ As shown in Figure 8, topical estradiol treatment of postmenopausal women induced GREB1 mRNA levels with a 3.6-, 4.7-, and 9.9-fold increase vs control in hip, forearm, and face skin, respectively (P < .05). In men, GREB1 mRNA levels were significantly increased after estradiol treatment in hip skin (2.25-fold; n = 6; P = .005) but not in forearm (n = 12) or face (n = 5) skin (Figure 8).

In the present study, we evaluated the effects of estradiol treatment on collagen production in sun-protected aged and photoaged skin of postmenopausal women and age-matched men. We demonstrated that 1-week topical treatment with estradiol stimulates procollagen I and III expression in sun-protected aged skin of postmenopausal women in vivo. Surprisingly, we found that estradiol has no effect on procollagen expression in photoaged skin (forearm or face) of postmenopausal women, despite similar levels of expression of estrogen receptors in sun-protected and sun-exposed skin sites, and similar levels of expression of estradiol-inactivating enzymes in hip and forearm skin. In addition, we provide evidence that topical estradiol penetration and activity are similar in sun-protected and sun-exposed skin sites, as assessed by quantification of induction of GREB1 gene expression, which is directly regulated by ERs.⁵³ Taken together, these findings support the concept that topical estradiol regulates procollagen I and III production in sun-protected skin in women. Our results also indicate that alterations induced by long-term sun exposure hinder the ability of topical estradiol to stimulate collagen production in aged human skin in vivo.

The study of estradiol responses in male skin yielded interesting differences compared with women. Topical estradiol increased procollagen production in naturally aged hip skin of men to a lesser extent than in postmenopausal women. Induction of GREB1 gene expression was also weaker in men than in women. In contrast, topical estradiol did not alter procollagen or GREB1 expressions in photoaged forearm or face skin of men. Differences in estradiol responsiveness between men and women were not due to differences in ER expression. Interestingly, CYP1B1, which inactivates 17β-estradiol by 4-hydroxylation, was significantly more highly expressed in photoaged forearm skin of men than of women, suggesting that reduced responsiveness of men to estradiol may be related to catabolism.

Our data are in agreement with previously published studies that demonstrated that topical estrogen treatment stimulates procollagen production in abdominal skin of postmenopausal women³² and hip skin of elderly men and women.³³ We did not observe any correlation between subject age and magnitude of estradiol-mediated induction of collagen mRNA or protein (n = 18).

To our knowledge, the present study is the first to describe a lack of effect of topical estradiol treatment on procollagen production in photoaged skin. It is possible that a treatment time longer than 2 weeks would have demonstrated effects similar to those in sun-protected skin. Additional studies will be necessary to test this possibility. However, GREB1 levels rose significantly and similarly in all treated sites in postmenopausal women, indicating that estradiol penetrated the skin and activated ERs similarly in sun-protected and sun-exposed skin. These results demonstrate that photoaging is not associated with a loss of estrogen genomic response, at least in postmenopausal women. Conversely, our results suggest that the effect of estradiol on procollagen is indirect, and that photoaged skin lacks an essential component to the estrogen-mediated collagen response. This hypothesis is in agreement with findings of Meyer et al,⁵⁴ who showed that 17α-estradiol, which does not activate estrogen receptors, stimulated skin procollagen production in rodent and human skin. However, the mediators of estrogen action on skin collagen remain to be identified.

More recently, it was reported that the combination of ET and topical estradiol increased procollagen immunostaining in facial skin in women.⁵⁵ The study did not investigate the effects of either ET or topical estradiol alone. It is possible that ET triggers systemic responses that enable topical estradiol to promote collagen production in photoaged skin, or vice versa. This possibility raises interesting questions regarding mechanisms of estrogen actions in human skin.

The presence of estrogen binding sites in skin has been demonstrated by means of radioactive estradiol.^56,57 These studies found that thigh skin has approximately 3-fold lower ER levels than does face or abdominal skin (1.5, 4.3, and 4.9 fmol of receptor per milligram of protein, respectively^56,57). In the present study, we did not find any significant difference in the level of ESR1, ESR2, or GPR30 gene expression in face, forearm, or hip skin in postmenopausal women. In addition, we did not find any significant difference in ER expression between vehicle- and estradiol-treated samples, consistent with previous results.⁵⁷

Detection of ERs in human skin has yielded discordant results. Immunostaining-based studies reported that ER-α was not detected or was weakly detected, whereas ER-β was strongly detected in epidermis, eccrine glands, sebaceous glands, and hair follicles. On the basis of these reports, ER-β is often described as the predominant form of ER in adult human skin,^58-60 although the relative abundance of the 2 ER isoforms has not been formally quantified. We quantified ER-α, ER-β, and GPR30 mRNA in human skin and demonstrated that ER-α mRNA levels were approximately 10 times greater than those of ER-β or GPR30. We were unable to detect either protein by immunohistochemistry (using a variety of different antibodies) because of absent or nonspecific staining. Using laser capture microdissection, we found that most of the ER-α and GPR30 mRNA was expressed in dermal cells, whereas ER-β was expressed in both appendages and dermal cells. More studies are needed to precisely identify ER-expressing cells within dermis and appendages.

In summary, our study shows that estradiol stimulates collagen I and III production in sun-protected skin but not in photoaged skin of postmenopausal women and aged-matched men within 2 weeks. Because photoaging is superimposed on natural aging in sun-exposed areas of the skin, our results suggest that alterations induced by long-term sun exposure hinder the ability of topical estradiol to stimulate collagen production in aged human skin in vivo.

Correspondence: Gary J. Fisher, PhD, Department of Dermatology, University of Michigan Medical School, Medical Science I, Room 6447, 1301 E Catherine, Ann Arbor, MI 48109-0609 (gjfisher@umich.edu).

Accepted for Publication: April 8, 2008.

Author Contributions: Drs Rittié and Fisher had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Rittié, Kang, Voorhees, and Fisher. Acquisition of data: Rittié, Kang, and Fisher. Analysis and interpretation of data: Rittié, Kang, Voorhees, and Fisher. Drafting of the manuscript: Rittié. Critical revision of the manuscript for important intellectual content: Rittié, Kang, Voorhees, and Fisher. Statistical analysis: Rittié. Obtaining funding: Voorhees and Fisher. Study supervision: Voorhees and Fisher.

Financial Disclosure: Dr Voorhees was a consultant for Pfizer Inc and received consulting payments.

Funding/Support: This study was supported in part by a grant from Pfizer Inc (Dr Fisher).

Additional Contributions: Suzan Rehbine, LPN, helped with volunteer recruitment and tissue collection; Laura VanGoor, BFA, assisted with graphic material; and Stephanie Cooke, BS, Kenne Currie, BS, and Craig Hammerberg, PhD, provided technical help.