Given the potential for manipulation of the course of skin aging by systemic signals, there has been an effort to study these signals in a controlled context, allowing for thorough analyses of hormone levels and preliminary tests of pharmaceutical interventions to reverse aging phenotypes. One method to test the role of the systemic milieu on skin aging involves experiments in which the circulation of old and young animals are surgically linked [8]. In these so-called parabiosis experiments, tissues from a linked heterochronic pair consisting of one old and one young mouse are compared to tissues of control isochronic pairs consisting of two young mice or two old mice. Thus, the effect of a young systemic milieu on the aged tissue phenotype and the effect of an aged systemic milieu on the young tissue phenotype can be evaluated.

In early experiments employing this technique, the regeneration of aged skeletal muscle and hepatic tissue in response to acute injury were enhanced by exposure to a young systemic environment [9]. In the brain, parabiosis studies demonstrated that neurogenesis in young mice is impaired by hormones present in the circulation of old mice [10]. Although the effects of parabiosis on skin aging have not been extensively explored, recent investigation of the impact of age on mouse hair follicle regeneration revealed enhancement of the in vitro colony-forming potential of hair follicle stem cells (HFSCs) from aged mice following heterochronic parabiosis [11].

In exploring the mechanisms by which an aging systemic milieu can suppress the regenerative potential of young tissue, an excess of Wnt signaling activity has been identified in aged serum [8]. That pathologic Wnt signaling plays a role in mammalian aging was supported by the observation that mutant klotho mice, which display phenotypes of precocious aging, also have abnormally high levels of circulating Wnt [12]. This exaggerated Wnt signaling can be attributed to the loss of a Wnt inhibitory factor encoded by the klotho allele. Skin atrophy and impaired wound-healing are among the aging phenotypes of klotho mutants [13], suggesting that Wnt is a circulating factor that can dictate the functional phenotypes of aging skin. Indeed, mice engineered to express Wnt at supraphysiologic levels display markers of senescence in skin cells akin to what is observed in klotho mutants [12]. Excess Wnt seemed to be particularly harmful to HFSCs, which were depleted from the follicle in mice lacking klotho. Human klotho exhibits 86 % amino-acid sequence similarity with the mouse protein. Although there are no studies examining the specific role of klotho in human skin aging, a recent population-based association study has suggested a correlation between functionally distinct KLOTHO alleles and both life expectancy and longevity [14]. Understanding whether these KLOTHO variants predict, or even determine, characteristics of human skin aging will be an important area for future investigation.

Tissue-specific stem cells may play an especially important role in the manifestation of skin aging phenotypes. This association arises from the observation that aging often reflects impairment of homeostatic tissue repair, which is generally the job of resident adult stem cells [15]. Furthermore, stem cells are uniquely sensitive to hormonal growth factors that are commonly dysregulated in aged animals [16]. In addition to their sensitivity to increased Wnt signaling, HFSCs show an age-dependent decline in hair follicle regeneration that is linked to their over-stimulation by bone morphogenetic proteins (BMPs) [11]. In studies by Fuchs and colleagues, increases in the expression of Bmp2, Bmp4, and Bmp6 in the skin of aged mice were found to cause abnormal activation of the transcription factor NFATc1. The dysregulated BMP-NFATc1 axis impairs the activation of aged HFSCs in response to follicle depilation. Treatment of old mice with systemic inhibitors of BMP or NFATc1 returned HFSCs’ youthful capacity for hair regrowth, indicating that alterations in BMP-NFATc1 signaling caused reversible impairment of aged HFSCs.

The influence of BMP signaling on the function of skin stem cells during aging may relate in part to its regulation by hormonal TGF-β signaling. In healthy, young skin, TGF-β normally inhibits the repressive effects of BMP signaling through the activation of Tmeff1 [17]. In aged skin, however, TGF-β levels are markedly reduced, and this reduction is associated with a decrease in the level of type I collagen production [18]. Reducing the level of TGF-β in young skin is also sufficient to reduce local collagen content. This effect of TGF-β may be driven by a synergistic interaction with connective tissue growth factor (CTGF), which augments the increased expression of collagen induced by TGF-β and also decreases during aging [19]. Although TGF-β levels have not been linked directly to the decline in skin stem cell function during aging, changes in the efficiency of skin wound healing seen in aged mammals have been attributed to low TGF-β. These studies revealed a decline in expression of all three TGF-β isoforms within regenerating wounds of aged mice [20]. In agreement with these findings, analyses of mutants lacking functional TGF-β1 [21], TGF-β3 [22], or TGF-β receptor type II [23], revealed impaired wound healing, including reduced collagen deposition and poor wound contraction. Whether pharmacologic stimulation of TGF-β signaling is sufficient to improve wound healing in the context of aging is yet to be studied. Nonetheless, the addition of TGF-β3 to wounds increases the rate of healing even in young animals [24]. Also of note, exogenous erythropoietin accelerates wound healing by inducing expression of TGF-β1 [25]. In humans, exposure to ultraviolet radiation, known for inducing aging-associated skin damage, impairs TGF-β signaling and thus causes a reduction in collagen expression [26, 27]. Decreased epidermal TGF-β signaling may also have a role independent of ultraviolet radiation in intrinsic skin aging [28]. Although stimulation of TGF-β signaling in human skin fibroblasts in vitro is sufficient to induce collagen expression [29], future clinical trials are required to evaluate the effects of exogenous TGF-β on human skin. An important consideration in the therapeutic modulation of TGF-β signaling for enhanced wound healing or the prevention of skin aging, however, is the concern that excess activation of this pathway may contribute to the formation of hypertrophic scarring or fibrosis [30–32].

Changes in the systemic milieu during aging will by definition have the potential to impact tissues throughout the body. Therefore, it stands to reason that a broader survey of age-dependent pathology will identify shared mechanisms that point to hormonal dysregulation. Employing this logic, Chang and colleagues performed an elegant analysis of genome-wide mRNA expression data from 365 microarrays encompassing nine different human and mouse tissues, including skin, to determine which cis-regulatory motifs were over-represented among those genes for which expression changes significantly with age [33]. Remarkably, the motif most strongly associated with age-dependent gene expression changes was that of NF-κB, a transcription factor responsible for conveying the message of numerous hormonal signals. Among the known regulators of NF-κB activity are tumor necrosis factor alpha (TNF-α), interleukin 1 beta (IL-1β), epidermal growth factor (EGF), fibroblast growth factor (FGF), platelet-derived growth factor (PDGF), progesterone, and glucocorticoids [34]. The implication of this finding is that increased NF-κB-dependent transcription is responsible, at least in part, for the aging phenotype of skin, among other tissues. To test this model, NF-κB was inhibited in aged skin by expressing a dominant-negative NF-κB allele [33, 35]. Of the 414 genes that showed age-dependent changes in expression in mouse skin, 225 returned to the level observed in young mice upon inhibition of NF-κB. Abrogating NF-κB in aged skin also increased skin thickness and epidermal cell proliferation, while decreasing the amount of cell senescence normally observed during aging. Although elevated NF-κB signaling in aged skin has yet to be linked to signaling by a particular hormone, previous studies have shown dysregulation of TNF-α [36, 37], PDGF [20], FGF [38, 39], and cortisol [40] levels during aging.

Among human hormones thought to be involved in skin aging, the sex hormones have received the bulk of attention in clinical studies. The prevalence of studies examining the effects of estrogens and androgens may be attributed to a well-understood phenomenon of sex hormone reduction in both genders with age, and to the popularity of sex hormone replacement therapy. In women, menopause marks the cessation of estrogen production by the ovaries, and a consequent drop in circulating hormone levels. Estradiol, the predominant estrogen during reproductive years in terms of both estrogenic activity and absolute serum levels, decreases by 80 % during menopause [41]. Menopause thus mimics a natural loss-of-function experiment for studying the role of estrogens in skin biology. Phenotypes of aging skin linked to declining estrogen levels during menopause include atrophy, xeroderma, wrinkling, fragility, and poor wound-healing [42].

Clinically, many of these phenotypes are abrogated with estrogen replacement. Following menopause, the dermal collagen content declines by 1–2 % per year [43], and this decline correlates better with age relative to menopause than with absolute age [44]. In two randomized, placebo-controlled studies, estrogen replacement for 6–12 months increased the skin collagen content of postmenopausal women [45, 46]. Estrogen replacement also led to a 33 % increase in skin thickness over a period of 12 months of therapy [47]. Skin wrinkling and laxity are two phenomena of aging thought to reflect a loss of tissue elasticity. Bolognia et al. identified degenerative changes in the elastic fibers of menopausal women [48], and subsequent studies revealed enhanced size and concentration of elastic fibers with estrogen replacement [49, 50]. Others have used computerized suction devices to measure the extensibility of skin in postmenopausal women receiving estrogen therapy or placebo, and have shown consistently that estrogens prevent the increase in skin extensibility, and decrease in elasticity, normally seen during menopause [51, 52]. Unfortunately, in vitro studies have not easily translated into clearly effective treatments for aging-related skin pathologies. To date, small clinical trials of topical estrogen cream on skin aging parameters such as rhytides have not consistently demonstrated significant effects, and hence this treatment modality is not the standard of care in general dermatologic practice [53, 54].

Estrogen also seems to have beneficial effects on skin wound healing. In one study, men and women treated with transdermal estrogen at the site of a punch biopsy had faster wound repair and increased wound collagen levels [55]. Estrogen replacement also prevented the onset of chronic venous stasis or pressure ulcers [56]. The beneficial effects of estrogen on wound healing may relate to stimulation of TGF-β secretion at the site of the wound [57], and to the suppression of pro-inflammatory macrophage migration inhibitory factor [58]. Despite its overall favorable impact on aging phenotypes in skin, the regular use of oral estrogen replacement therapy remains limited by its association with increased risk of venous thromboembolism, stroke, heart disease, and breast cancer [59, 60].

Selective estrogen receptor modulators (SERMs), which both activate and block estrogen receptors, have also been studied for their effects on human skin. In human skin fibroblasts, collagen biosynthesis has been observed to increase in response to oral raloxifene, a SERM [61]. In a small study in postmenopausal women, raloxifene has also been reported to improve skin elasticity [62]. Whether these improvements are clinically significant and whether the risk–benefit ratio of SERMs is favorable for antiaging indications remains to be determined.

Estrogen depletion at menopause is thought to explain the accelerated rate of skin aging, especially in terms of skin atrophy, observed in females as compared to males [63]. However, men too undergo a so-called “andropause” associated with a decline in testosterone levels with age [64]. Many of the age-related changes associated with postmenopausal estrogen reduction, such as decreases in skin thickness and elasticity, have also been linked to low testosterone [65]. Likewise, testosterone replacement can lead to rejuvenation of the skin of aged men, inducing increased epidermal thickness and dermal mucopolysaccharide content [66]. Perhaps the most recognizable effect of androgens is their regulation of hair growth. In contrast to the involvement of increased androgen signaling in the androgenic alopecia seen in younger men, aged men undergo a separate process of hair loss associated with dwindling testosterone levels [67]. Women also exhibit beneficial hair regrowth in response to androgen replacement therapies [68]. Interestingly, the diverse effect of androgens on hair follicle function may reflect anatomical heterogeneity in follicle hormone receptor expression [69, 70]. Thus, the age-dependent decline in androgen levels is predicted to manifest as a variable hair loss dictated in part by positional differences in androgen sensitivity.

Glucocorticoids represent another class of steroid hormone for which there have been ample clinical studies owing to the prevalence of glucocorticoid-based treatments of dermatologic and non-dermatologic conditions. The association of glucocorticoid signaling with aging stems from the striking similarity between the phenotypes observed in aged individuals and in individuals with glucocorticoid excess. It is now known that both aging and glucocorticoid excess cause dermal and epidermal thinning [71, 72], flattening of the rete ridges at the dermal–epidermal junction [73, 74], poor wound healing [75, 76], and reduced collagen content [77–80]. These observations have led to the hypothesis that age-related skin pathology can be attributed to an increase in glucocorticoid signaling.

As mentioned, some have attributed an excess of glucocorticoid signaling in adults to elevated circulating cortisol [40]. However, in these studies of healthy human volunteers, significant variation in cortisol levels were noted, and a subset of individuals had a decrease in cortisol levels with age. Given this discrepancy, some have argued that excess glucocorticoid signaling is instead related to local enzymatic regulation of cortisol availability [81]. In support of this model, investigators found that the expression of 11-β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) in the skin of aged humans and mice is higher than that observed in young subjects. 11β-HSD1 is a cellular oxoreductase that converts cortisone to cortisol. Additional studies have shown that 11β-HSD1 expression is also increased in photo-exposed relative to photo-protected skin, implying that this mechanism may account also for skin photoaging [82]. Remarkably, genetic loss of 11β-HSD1 leads to reduced skin atrophy and increased skin collagen content in aged mice [81]. Likewise, application of a small-molecule inhibitor of 11β-HSD1 to the skin of aged mice enhances the rate of wound healing. Thus, excess glucocorticoid signaling appears to be a clinically tractable feature of skin aging.

Recently, techniques for global gene expression analysis have been used to study the molecular changes undergone by skin during aging that are likely to underlie the phenotypes we associate with aged skin. Although these studies fundamentally reflect changes within the skin cells themselves, it follows that these intrinsic characteristics reflect alterations in the systemic environment, and therefore may be used to shed light on the mechanisms by which systemic signals affect skin physiology during aging. One such study examined global gene expression using RNA sequencing in five young (age <30 years) and five old (age >50 years) subjects before and after treatment of skin with broadband light [83]. With this study design, investigators were able to identify changes in gene expression patterns associated with aging and those associated with rejuvenation, or a return to youthful levels, induced by treatment with broadband light. Among the genes for which expression was rejuvenated by broadband light, there were a number associated with the intracellular transduction of hormonal signaling.

One of these genes, IGF1R, is the tyrosine kinase receptor for insulin-like growth factor 1 and 2. IGF1/2 signaling is known to be a critical regulator of aging that is conserved among humans and model organisms. Notably, disabling mutations in the C. elegans homolog of IGF1R, daf-2, were the first to be associated with lifespan extension [84]. Regulation of IGF signaling has since been shown to influence lifespan in yeast, flies, and mice [85]. Though controlled manipulations of IGF signaling have not yet been studied in humans, functional mutations of IGF1R that generate decreased IGF signaling have been identified in exceptionally long-lived individuals [86]. Furthermore, circulating levels of IGF-1 are lower in the offspring of human centenarians compared to age-matched offspring controls [87]. The correlation with decreased IGF signaling and tissue youthfulness is similar in human skin, where broadband light reverses the elevated expression of IGFR1 seen in aged subjects [83]. Interestingly, in one study of otherwise healthy young women (aged 22–41), increased IGF-1 levels in the skin were correlated with poor cosmetic characteristics such as increased facial pore area, sebum production, and epithelial disorganization [88]. The connection between IGF signaling and skin aging has been reinforced in multiple subsequent genome-wide association studies. In a case-control genome-wide association study of aged individuals who lack global features of skin aging, skin youthfulness was associated with a single-nucleotide polymorphism near EDEM1 [89], the expression of which correlates with expression of IGF1 in a mouse model of longevity [90]. In a similar genome-wide association study of skin aging in middle-aged women, a polymorphism that influences expression of FBXO40, a regulator of IGF signaling, was noted [91]. Taken together, these studies hint strongly at a role for IGF signaling in the progression of human skin aging, although the precise mechanisms by which alterations in IGF signaling lead to aging phenotypes remain to be elucidated.

Studies of gene expression in human skin have also added support for the aforementioned role of NF-κB in the aging process. In the group of genes exhibiting photo-reversible changes in expression with age, targets of NF-κB are significantly overrepresented [83]. It should be noted, however, that interactions between NF-κB and IGF signaling have been identified, and that the observation of age-related alterations in these pathways may therefore reflect their interdependence.

Advances in the study and clinical application of hormones and signaling pathways in skin aging will be dictated in large part by the development of new technologies. Among these technologies, improvements in the speed and accuracy of nucleic acid sequencing, and the related reduction in sequencing costs, will enable a broadening of sequencing studies. Although initial studies involving relatively small numbers of sequenced subjects have shed much light on the role of specific signaling pathways in skin aging, expansion of these data sets will allow us to determine how variation in hormonal signals influence aging at the level of individuals. Such studies would contribute to our currently limited understanding of aging heterogeneity—how the process of aging manifests uniquely in each subject. Clinically, more individualized data would allow intervention to be tailored to the needs of a specific patient. For instance, one patient may benefit more from estrogen replacement, another from BMP-NFAT suppression. Eventually, routine testing of patients may involve genome and RNA sequencing from a small skin biopsy.

Initial efforts to sequence the human genome identified numerous so-called “orphan” receptors, which were predicted on the basis of sequence to function as receptors, but which did not have a known ligand. Though the intervening years have seen the “adoption” of hundreds of such receptors, many orphans remain. Recent advances in techniques for determination of receptor structure, particularly in the case of transmembrane receptors, offer a new avenue for the prospective identification of ligands. Because previous screening approaches lacked the sensitivity to detect subtle or unexpected downstream responses [92], new analytical technologies including low-input RNA sequencing or mass spectrometry offer new opportunities for ligand screening. The likelihood that select orphan receptors have roles in skin biology that are relevant to age-related pathology makes this an interesting direction for future study.

Given that stem cell populations mediate many of the effects of age-associated hormonal changes in skin, future advances will benefit from a recent exponential increase in our understanding of skin stem cell biology. New insight into the identity of stem cells in the epidermis and hair follicle have allowed for novel methods to prospectively purify stem cell populations from mouse and human tissue. These isolation techniques enable unprecedented studies of the molecular mechanisms of stem cell-mediated maintenance of skin homeostasis. Future application of these techniques for comparison of stem cell characteristics in young and aged subjects will indicate specifically how skin stem cells are affected by systemic signals in aged individuals. As our knowledge of skin stem cells has grown, so has our ability to manipulate these cells ex vivo for clinical applications. If these cells can be grown in the laboratory, studies involving manipulation of hormonal signaling can be achieved much more efficiently and in a controlled setting. These studies could also be done in a patient-specific manner. Importantly, the successful ex vivo growth of skin stem cells will likely rely on a reconstitution of the hormonal milieu that these cells encounter in vivo. Techniques for efficient growth of stem cells will enable high-throughput screens of functional readouts of hormone signals, including combinatorial screens that look for interactions between hormonal signaling pathways. The most exciting promise of our expanding toolbox for manipulation of skin stem cells is of course the use of these cells for clinical repair of damaged or degenerated tissue. Insight into hormonal regulation of skin stem cell behavior may advance techniques for promoting stem cell engraftment, survival, and self-renewal during transplantation.