Aging Skin

JEAN L. BOLOGNIA, M.D., New Haven,Connecticut

.;:

Aging of the skin is a composite of actinic

damage, chronologic aging, and hormonal influences. The majority of changes associated
with aging, such as wrinkles and solar lentigines (liver spots), are due to photoaging
and reflect cumulative sun exposure as well
as skin pigmentation.
Classically, chronologic
aging*includes those cutaneous changes that
occur in non-sun-exposed
areas, such as the
buttocks, and are observed in both men and
women. A clinical example would be soft tissue sagging due to elastic fiber degeneration.
In women, investigations
into the effect of
hormones on aging of the skin have concentrated on estrogens; in men, there have been
a limited number of studies on the influence
of testosterone. The latter have shown an
age-dependent decrease in tissue androgens
in pubic skin, but not scrotal or thigh skin.
To date, age has not been shown to have an
effect on androgen receptor binding, although
a decrease in foreskin 5a-reductase activity
with increasing age has been described. In
fibroblast cultures from foreskins, there have
been conflicting results as to whether 5areductase activity decreases in an age-dependent manner. Some of the skin changes that
have been categorized as secondary to chronologic aging, such as decreased sebaceous
gland activity and decreased hair growth, may
actually represent a decline in the concentration of tissue androgens with increasing age.
The influence of androgens on age-related
changes in keratinocyte
and fibroblast function remains speculative.

Fromthe Departmentof Dermatology,

Yale University School of Medicine, New

Haven, Connecticut.
Requests for reprints should be addressed
to Jean L. Bolognia, M.D., Department of Dermatology,
Yale MedIcal School, 500 LCI, 333 Cedar Street, New
Haven, Connecticut
06520.

lthough photoaging plays a major role in skin

aging, it is also important to examine the role
of chronologic aging as well as the influence of circulating hormones, including estrogens, progestins,
and androgens. The skin signs most commonly associated with aging such as wrinkling, furrowing, telangiectasias, and solar lentigines (liver spots) are
a reflection of cumulative skin damage (Table I) [l].
It has been estimated that the regular use of a sunscreen with a skin protective factor (SPF) of 15 during the first 18 years of life would decrease the lifetime incidence of nonmelanoma skin cancers by 78%
[2]. Obviously, the other manifestations of photoaging would also be reduced by such an intervention.
A number of cutaneous changes have been attributed to chronologic aging, including thinning of the
epidermis and dermis, [3] decreased sebum production, [4] and decreased growth of body hair (Table
II) [5]. Such changes are attributed
to chronologic
aging based on their presence in non-sun-exposed
skin as well as sun-exposed skin and their occurrence in a similar pattern in women and men. More
recently, however, it has been show that several of
these changes are influenced by circulating levels of
estrogen. For example, in a study of postmenopausal women (n = 148), a decrease was observed in the
collagen content of thigh skin of approximately
2%
per postmenopausal
year (for up to 15 years) in
those who had not received estrogen replacement
therapy [6]. Of note, there was no correlation between skin collagen content and chronologic age in
this group of women [7].
These initial studies pointing to a relationship
between dermal collagen content and menopause
led to further investigations involving women who
were receiving hormone replacement therapy (implants of 50 mg estradiol plus 100 mg testosterone
every 6 months for 2-10 years). Such therapy was
shown to prevent the decrease in thigh skin collagen content that was previously observed in untreated women [8]. Because the patients in this
study received both estrogen and testosterone as
replacement therapy, it is important to note that an
increase in dermal collagen was also described in
women who received estrogen (orally or transdermally) following hysterectomy and bilateral oophorectomy [9]. Moreover, there was a report of premature aging of dermal elastic fibers in women with
untreated premature menopause [lo].

January16, 1995 The AmericanJournal of Medicine Volume 98 kuppl 1A)

lA-99S

SYMPOSIUM ON ANDROGENSAND WOMENS HEALTH/ BOLOGNIA

re$valuate the assumption that the changes outlined in Table II are independent of the sex or hormonal status of the patient. The possibility exists
that some of the changes ascribed to chronologic
aging are actually reflections of decreasing levels of
circulating androgens.

TABLEI
Skin ChangesDueto Photoaging
* Winkles and furrows

* Solarlentigines(liverspots)
e Telangiectasias
* Mottledpigmentation
* Senilepurpura
* Actinickeratoses
@Basalcellandsquamouscellcarclnamas
* A subsetof melanomas

INFLUENCEOF ANDROGENS
Whole Skin

When androgen receptor binding was measured

in human preputial skin cytosol, there was no effect
of age (range, newborn to 65 years) on androgenbinding capacity [121. Unfortunately,
the number of
skin biopsies performed in women (n = 4) was too
small to examine the effect of age in this group. In
skin slices of prepuce, a significant age-associated
decrease in 5cY-reductase activity was observed
over a span of 80 years [13]. Also, the enzyme activity was significantly higher in genital skin as compared with nongenital skin in both men and women.
In more recent studies, Deslypere and Vermeulen
[ 141 demonstrated
an age-associated decrease in
the concentration of several androgens (testosterone, 5a-androstane-lip-ol-3-one
[DHT], and 5cyandrostaneSa,l7P-diol)
in the pubic skin of 24 male
subjects, ages 20-82 years. However, no decrease
was observed in either scrotal or thigh skin.

TABLEII
Skin ChangesAssociatedwith ChronologicAging
* Dryness
* Increasedfragilfty
* Decreasedepidermalthickness(atrophyI,t
* Decreaseindermalthickness(a$ophy),t
* Fragmentation
of elasticfibers(saggingit
* Decreasedsebumproductton
8 Decreasednumberandfunction of apocrine glands8
* Decreased growth of body hairs
nRuenced by circulatinglevels of estrogens.
lay be Influenced by crrculating levels of androgens.
robably Influenced by circulating levels of estrogens.
robably Influenced by levels of circulahng androgens.

TABLEIll
Localizationof SorReductaseActivity and Androgen
Receptorsin the Skin*,t
5&eductese
Activity

Hair

Androgen
Receptors

Epidermal keratlnocfles

Dermalfibroblasts
Hairfollicles
External root sheath
Matrix epithelrum

+%
+%

+
+I1

+I

t
t

t
t

Dermal paptlla cells

Sebaceousglands
Apocmeglands

iaReductase activity has also been found In eccrine glands l261.

lndrogen receptors have also been localized to coil and ductal cells of eccnne glands and endothe_a1 cells of dermal
,. ,. bloqd,~e!sse!s!! 7-191.
rrresent in lsolarea nalr ~ollrcles 1161.
Two of three studies [17-191.
IlOne of three studies 117-191.

In contrast to their effect on the dermis, the influence of circulating estrogens on the structure
and function of the epidermis is rather confusing
given the opposite effects-atrophy
(thinning) as
well as thickening-that
have been reported in
postmenopausal women receiving estrogen replacement therapy [ll]. In summary, it is necessary to
lA-100s

January

16,1995

The American Journal of Medicme

There is a general decrease in the number of hair

follicles on the body as one ages, and atrophy as
well as fibrosis of hair follicles has been observed
[5]. In addition, an increase in the number of hair
follicles in the telogen, or resting, phase of the hair
cell cycle has been described [15]. Growth of axillary hair begins to decline after the fourth decade,
especially in women [15]. Androgen receptors have
been isolated from hair follicles and they have been
shown to exist in two major forms, an active monomer form (62 kilodaltons) and an inactive tetramer
form [16].
In immunohistochemical
studies using a polyclonal antibody against the human androgen receptor, Blguer et al [17] demonstrated specific binding
in the cells of the outer root sheath. Liang et al [ 181,
using a monoclonal antibody against the human
androgen receptor, noted specific staining in three
components of the hair follicle: (a) the external root
sheath; (b) the hair bulb epithelium, or germinative
matrix cells; and (c) dermal papilla cells (Table III).
A third study utilizing a mouse monoclonal antibody against a synthetic androgen receptor peptide
confirmed the presence of specific staining for the
androgen receptor in the dermal papilla cells but

Volume 98 (suppl IA)

not in the cells of the external

or outer root sheath

cw.
The possible importance of the papilla cells is
underscored by transplantation
studies in animals
in which the origin of the papilla cells (not the origin
of the hair epithelium) determined the type of hair
produced [ZO]. It has been possible to isolate and
culture dermal papilla cells, and experiments have
shown that the papilla cells isolated from beard hair
have higher 5Lu-reductase activity than those isolated from occipital scalp hair or reticular dermal
fibroblasts [21]. This suggests that dermal papilla
cells may play a role in androgen-dependent
hair
growth.

Sebaceous and Apocrine Glands

The initial studies on sebum production reported

that in women sebaceous secretion decreased gradually after the menopause, with little change after
the seventh decade, while in men sebum levels remained essentially unchanged until the age of 80
[22]. However, these earlier studies did not control
for the sebum reservoir in the pilosebaceous apparatus. Although sebum production declines in old
age, the size of the follicular reservoir may not because the sebaceous glands undergo hypertrophy in
older individuals [23,24]. Therefore, accurate sebum
secretion measurement
requires a long equilibration period during which the reservoir is depleted
[231. When this is done, a steady decline in sebum
production throughout
adult life is observed for
both men and women [4].
Androgens cause an increase in the secretion of
sebum; thus, one explanation for the decline in
sebum production with age is the decline in circulating levels of androgens, including dehydroepiandrosterone sulfate (DHEAS) [23,25]. Androgen receptors have been purified from isolated human
sebaceous glands. They exist in two major forms (as
in the hair follicle), an active monomer form (62 kilodaltons) and an inactive tetramer form [16]. Androgen receptors have also been detected in human
sebocytes (both basal and glandular cells) by immunohistochemical methods in the skin of adult women
and men [17-191. Using antipeptide
antibodies
(against amino acids 227-240), researchers have
detected 5a-reductase in sebaceous glands [26]; it is
the opinion of several investigators that the sebaceous gland is the major site of 5a-reductase activity in the skin [27].
Apoerine glands tend to decrease in size with increasing age, and a limited number of studies have
shown a decrease in the function of the apocrine
glands in the elderly [28]. Utilizing
a monoclonal
antibody against a synthetic androgen receptor

peptide, Choudhry et al [19] found positive immunocytochemical

staining in the luminal epithelial
cells of apocrine glands. In addition to localizing the
androgen receptor to the apocrine gland, earlier
studies from the 1970s involving
dissection of
freeze-dried skin nicely demonstrated the presence
of 5cY-reductase activity in these glands 1291. The
age-related decrease in the function of the apocrine
glands may be related to circulating
androgens
(Table II).

Fibroblasts

In the mid 1970s specific 5a-dihydrotestosterone

binding to cultured human skin fibroblasts was initially described [30,31]. Cells obtained from genital
skin had mean binding values three times those of
cells from nongenital skin; no differences between
males and females were observed 1301. In addition,
the values did not correlate with donor age [301.
Cultured skin fibroblasts from both genital (foreskin, labium majus) and nongenital areas demonstrated 5a-reductase activity [32], with fibroblasts
from genital areas having considerably more activity than those from nongenital skin [32,33]. However, because of significant variability,
the 5crreductase activity in the labial strains did overlap
with that of the nongenital strains 1321.
In a study utilizing foreskin fibroblast cultures
(range, newborn to 81 years), age had no effect on
the specific binding of DHT, but there was a trend
toward decreased 5a-reductase activity [34]. Unfortunately, only three lines were established from
genital skin in women, and the women were the
same age (25 years). These three lines did exhibit
decreased 5a-reductase activity when compared
with foreskin strains. Not all investigators
have
reported a decrease in 5a-reductase activity in fibroblasts from male genital skin as a reflection of
increasing age [33,35]. However, the age ranges in
these two studies were newborn to 25 years and
newborn to 42 years, respectively.
More recently, 5a-reductase cDNAs have been
cloned from skin tibroblast cDNA libraries. Nucleotide sequence analysis has shown that type I is the
predominant
form of the enzyme present in fibroblasts [26]. Using a polyclonal antibody against the
androgen receptor, Blauer et al 1171 demonstrated
specific staining in the nuclei of dermal fibroblasts
in both genital (male) and nongenital (male and female) skin. Localization of the androgen receptor to
dermal fibroblasts was also observed in studies utilizing monoclonal antibodies against the androgen
receptor [l&19]. These results all point to a relationship between dermal fibroblasts and androgens,
including the finding that 24 women receiving an-

January 16, 1995

The American Journal of Medicine

Volume 98 (suppl 1A)

lA-101s

SYMPOSIUM ON ANDROGENSAND WOMENS HEALTH/ BOLOGNIA

drogens for treatment of osteoporosis also experienced an increase in total skin collagen content [36].
Keratinocytes
Epidermal
turnover time, as assessed by tritiated thymidine uptake following intradermal
injection, has been shown to decrease by approximately
50% from age 20 to 70 years, and replacement time
for the stratum corneum is nearly doubled in the
elderly [37]. A relationship between androgens and
epidermal function has been known since the 1950s.
In castrated male mice, a decrease in epidermal
mitotic activity was observed when compared with
control animals, and subcutaneous injections of testosterone propionate restored the mitotic activity
to normal [381. Additional studies in castrated male
rats demonstrated an increase in the rate of epiderma1 cell proliferation
following the administration
of testosterone but no associated increase in the
thickness of the epidermis [39].
More recent studies have pointed to the presence
of androgen receptors as well as 5a-reductase activity in the keratinocytes of the epidermis. Immunohistochemical
studies using both polyclonal and
monodonal antibodies directed against the androgen receptor have demonstrated specific staining in
the nuclei of the cells of the epidermis in both genital (male) and nongenital (male and female) skin
[17-191. In one study, the proportion of skin specimens with positive staining was higher in men
(62%) than in women (23%) [19]. Cloning of 5~
reductase cDNAs from adult keratinocyte
cDNA
libraries has shown a predominance of the type I
form of the enzyme 1261. In addition to 5a-reductase
activity, other enzymatic activities found in epiderma1 keratinocytes include 3P-hydroxysteroid
oxidoreductase and 17phydroxysteroid
oxidoreductase,
suggesting that the epidermis represents a site of
synthesis of biologically potent androgens in human
skin [40].
The fact that no investigations into the effects of
testosterone alone on skin aging in women have
been reported points to the obvious need for studies
in this area.

ACKNOWLEDGMENT
I wish to thank Steven R. Kohn, M.D., for his
helpful suggestions.

REFERENCES
1. Leyden JJ. Clinical features of ageing skin. Br J Dermatol 1990; 122 (suppl 35):
l-3.
2. Stern RS, Weinstein MC, Baker SG. Risk reduction for nonmelanoma skin cancer
with childhood sunscreen use. Arch Dermatol 1986; 122: 537-45.
3. Branchet MC, Boisnic S, Frances C, Robert AM. Skin thickness changes in normal
aging skin. Gerontology 1990; 36: 28-35.

lA-102s

4. Jacobsen E, Billings JK, Frantz RA, Kinney CK, Stewart ME, Downing DT. Agerelated changes in sebaceous wax ester secretion rates in men and women. J Invest
Dermatol 1985; 85: 483-5.
5. Gilchrest BA. Age-associated changes in the skin. J Am Geriatr Sot 1982; 30:
139-43.
6. Brincat M, Kabalan S, Studd JWW, Moniz CF, de Trafford J, Montgomery J. A study
of the decrease of skin collagen content, skin thickness, and bone mass tn the
postmenopausal woman. Obstet Gynecol 1987; 70: 840-5.
7. Brincat M, Moniz CJ, Studde JWW, et al. Long-term effects of the menopause and
sex hormones on skin thickness. Br J Obstet Gynaecol 1985; 92: 256-9.
8. Brincat M, Moniz CF, Kabalan S, et al. Decline in skin collagen content and metacarpal index after the menopause and its prevention with sex hormone replacement.
Br J Obstet Gynaecol 1987; 94: 126-9.
9. Castelo-Branco C, Duran M, Gonzalez-Merlo J. Skin collagen changes related to
age and hormone replacement therapy. Maturitas 1992; 15: 113-9.
10. Bolognia JL, Braverman IM, Rousseau ME, Sarrel PM. Skin changes in men@
pause. Maturitas 1989; 11: 295-304.
11. Punnonen R. Effect of castration and peroral estrogen therapy on the skin. Acta
Obstet Gynaecol 1972; fsuppl21): 5-44.
12. Mowszowicz I, Riahi M, Wright F, Bouchard P, Kultenn F, Mauvais-Jarvis P. Androgen receptor in human skin cytosol. J Clin Endocrinol Metab 1981; 52: 338-44.
13. Wilson JD, Walker JD. The conversion of testosterone to 5a-androstane-17pol3-one (dihydrotestosterone) by skin slices of man. J Clin Invest 1969; 48: 371-9.
14. Deslypere JP, Vermeulen A. Aging and tissue androgens. J Clin Endocrinol Metab
1981; 53: 430-4.
15. Montagna W, Carlisle K. Structural changes in ageing skin. Br J Dermatol 1990;
122 (suppl 35): 61-70.
16. Sawaya ME. Purificatton of androgen receptors in human sebocytes and hair.
J Invest Dermatol 1992; 98: 92-6.
17. Blduer M, Vaalasti A, Paul1 S-L, Ylikomi T, Joensuu T, Tuohimaa P. Location of
androgen receptor in human skin. J Invest Dermatol 1991; 97: 264-8.
18. Liang T, Hoyer S, Yu R, et al. lmmunocytochemical localization of androgen
receptors in human skrn using monoclonal antibodies against the androgen receptor.
J Invest Dermatol 1993; 100: 663-6.
19. Choudhry R, Hodgins MB, Van der Kwast TH, Brinkmann AO, Boersma WJA.
Localization of androgen receptors in human skin by immunohistochemistry: implications for the hormonal regulation of hair growth, sebaceous glands and sweat
glands. J Endocrinol 1992; 133: 467-75.
20. Oliver RF, Jahoda CAB. Dermalepidermal interactions. Clin Dermatol 1988; 6:
74-82.
21. ltami S, Kurata S, Sonoda T, Takayasu S. Characterization of 5a-reductase in
cultured human dermal papilla cells from beard and occipital scalp hair. J Invest
Dermatol 1991; 96: 57-60.
22. Pochi PE, Strauss JS, Downing DT. Agerelated changes in sebaceous gland
activity. J Invest Dermatol 1979; 73: 108-11.
23. Downing DT, Stewart ME, Strauss JS. Changes in sebum production and the
sebaceous gland. Dermatol Clin 1986; 4: 419-22.
24. Plewig G, Kligman AM. Proliferative activity of the sebaceous gland of the aged.
J Invest Dermatol 1978; 70: 314-7.
25. Rozenberg S, Bosson 0, Peretz A, Caufriez A, Robyn C. Serum levels of gonadotrophins and steroid hormones in the postmenopause and later life. Maturitas 1988:
10: 215-24.
26. Luu-The V, Sugimoto Y, Puy L, et al. Characterization, expression, and immunohistochemical localization of Stieductase in human skin. 1 Invest Dermatol 1994;
102: 221-6.
27. Dijkstra AC, Goos CM, Cunliffe WJ, Sultan C, Vermorken Al. Is increased 5areductase activity a primary phenomenon in androgendependent skin disorders?
J invest Dermatol 1987; 89: 87-92.
28. Fenske NA, Lober CW. Structural and functional changes of normal aging skin.
1 Am Acad Dermatol 1986; 15: 571-85.
29. Ebling FJG. Functions of the skin. In: Champion RH, Burton JL, Ebling FJG, eds.
Textbook of Dermatology. London: Blackwell Scientific Publications, 1991: 125-55.
30. Kaufman M, Straisfeld C, Pinsky L. Expression of androgen-responsive properties in human skin fibroblast strains of genital and nongenital ongin. Somatic Cell
Genet 1977; 3: 17-25.
31. Keenan BS, Meyer WJ Ill, Hadjian AJ, Jones HW, Migeon CJ. Syndrome of andre
gen insensitivity in man: absence of 5&ihydrotestosterone
binding protein in skin
fibroblasts. J Clin Endocrinol Metab 1974; 38: 1143-6.
32. Pinsky L, Kaufman M, Straisfeld C, Zilahi 6, Hall CSG. Stieductase activity of

January16, 1995 The American Journal of Medicine Volume 98 (suppl 1AI

SYMPOSIUM ON ANDROGENSAND WOMENS HEALTH / BOLOGNIA

genital and nongenital skin fibroblasts from patients with 5reductase deficiency,
androgen insensitivity, or unknown forms of male pseudohermaphrodiiism. Am J Med
Genet 1978; 1: 407-16.
33. Wilson JD. Dihydrotestosterone formation in cultured human fibroblasts. J Biol
Chem 1975; 250: 3498-504.
34. Lamberigts G, Dierickx P, De Moor P, Verhoeven G. Comparison of the metabolism and receptor binding of testosterone and 17phydroxy-5nandrostaneSane
in
normal skin fibroblast cultures: influence of origin and passage number. J Clin Endocrinol Metab 1979; 48: 924-30.
35. Kaufman M, Pinsky L, Straisfeld C, Shanfield B, Zilahi B. Qualitative differences in
testosterone metabolism as an indication of cellular heterogeneity in fibroblast monolayers derived from human preputial skin. Exp Cell Res 1975; 96: 31-6.

36. Black MM, Shuster S, Bottoms E. Osteoporosis, skin collagen, and androgen. Br
Med J 1970; 4: 773-4.
37. Cerimele D, Celleno L, Serri F. Physiological changes in ageing skin, Br J Derma
tol 1990; 35 (suppI): 13-20.
38. Bullough WS, Van Oordt GJ. The mitogenic actions of testosterone propionate
and of oestrone on the eprdermis of the adult male mouse. Acta Endocrinol 1950; 4:
291-305.
39. Ebling FJ. The action of testosterone on the sebaceous glands and epidermis in
castrated and hypophysectomized male rats. J Endocrinol 1957; 15: 297-306.
40. Milewich L, Kaimal V, Shaw CB, Sontheimer RD. Epidermal keratinocytes: a
source of 5adihydrotestosterone
reduction in human skin. J Clin Endocrinol Metab
1986; 62: 739-46.

January 16, 1995

The American Journal of Medicine

Volume 98 fsuppl 1A)

lA-103s