Maturitas 33 (1999) 239 – 247 www.elsevier.com/locate/maturitas
Effects of postmenopausal hypoestrogenism on skin collagen Pietro Affinito *, Stefano Palomba, Claudia Sorrentino, Costantino Di Carlo, Giuseppe Bifulco, Maria Paola Arienzo, Carmine Nappi Clinical Department of Gynecology, Obstetrics and Pathophysiology of Human Reproduction, Uni6ersity of Naples ‘Federico II’, Via Pansini 5, 80131 Naples, Italy Received 22 March 1999; accepted 18 August 1999
Abstract Objecti6e: The aim of our study was to evaluate the effect of aging and postmenopausal hypoestrogenism on skin collagen content. Methods: Thirty-two women (mean age 48.7899.86; year9S.D., range 28 – 68), 14 in premenopause and 18 in postmenopause, underwent skin biopsies performed during laparotomic operation. The amount of collagen type I, III and type III/type I ratio was evaluated by immunohistochemistry and computerised image analysis, and was related to age and years of postmenopause. Results: In the postmenopausal patients, a significant (PB0.01) decrease of percentage of skin collagen type I, type III and type III/type I ratio was observed in comparison to premenopausal women. The percentages of collagen type I, type III and type III/I ratio of all patients studied was significantly (P B0.01) correlated with chronological age (r =0.88, 0.89 and 0.61, respectively). Considering only postmenopausal subjects, the correlation with chronological age was significant (P B 0.01) for collagen type I and type III of postmenopausal women (r= 0.59, r= 0.64, respectively), but not for the type III/I ratio (r =0.37, P=0.131). The percentages of collagen type I, type III and type III/I ratio of postmenopausal women showed a significant (PB 0.01) inverse correlation with years of postmenopause (r =0.76, 0.73 and 0.73, respectively). Conclusions: Our data suggest that the decrease of skin collagen is an estrogen-related phenomenon. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Menopause; Collagen; Hypoestrogenism; Skin; Sex hormones
1. Introduction The skin undergoes regressive changes after the menopause. It has been suggested that these changes are mainly related to a loss of skin collagen content [1]. However, it is not known if this
* Corresponding author. Tel.: +39-81-7462903; fax: + 3981-7462905.
phenomenon is only age-related or mainly dependent on postmenopausal hypoestrogenism. The data available in literature [1–4] only consider total skin collagen content. However, skin contains two types of collagen fiber, collagen I and collagen III, whose relative amounts determine the mechanical properties of the tissue [5]. The aim of our study was to evaluate by immunohistochemistry if and how the postmenopausal hypoestrogenism influences the
0378-5122/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 8 - 5 1 2 2 ( 9 9 ) 0 0 0 7 7 - 8
P. Affinito et al. / Maturitas 33 (1999) 239–247
240
amount of skin collagen type I, type III and the type III/type I ratio.
2. Materials and methods
postmenopausal women. The relationship between collagen type I, type III and type III/type I ratio with age was also considered. Indeed, the percentages of collagen type I, type III and type III/type I ratio of only postmenopausal subjects were related to age and years of postmenopause.
2.1. Subjects We selected 32 women (46913.4, years9 S.D.; range 22–65), candidates for laparotomic operation for benign ovarian cysts. The criteria for exclusion from the study included endocrine, dermatologic, renal and connective disease and prolonged treatments (over 1 month) with active drugs on collagen metabolism. Patients affected by uterine leiomyomata, genital prolapse, and urinary incontinence were also excluded. The patients in postmenopausal age (56.39 4.1, years9S.D.; range 50 – 65) were amenorrhoic for more than 12 months before enrollment (4.3 9 2.3, years9S.D.; range 1 – 9) in the study and presented follicle-stimulating hormone (FSH) and 17b-estradiol (E2) serum concentrations within the normal menopausal reference range (FSH\40 IU/l and E2 B20 pg/ml). The patients in premenopause age (32.797.8, years 9S.D.; range 22 – 44) regularly cycled. The characteristics of the patients are summarized in Table 1. The study was approved by the Ethics Committee of the University Federico II of Naples and all subjects gave their informed consent. We evaluated the differences in collagen type I, type III and type III/type I ratio in pre- and Table 1 Characteristics of subjects
Patients (n) Age (years9S.D., range) BMI (kg/m2 9S.D., range) Time since menopause (years 9S.D., range)
Premenopause
Postmenopause
14 32.79 7.8 (22– 44) 27.6 9 5.3 (22– 33) –
18 56.39 4.1 (50–65) 29.39 4.7 (23–33) 4.39 2.3 (1–9)
2.2. Skin biopsies Skin biopsies were performed during the laparotomic operations with skin cut about 4 cm over pubic symphysis, on the central point. Skin biopsies were frozen in liquid nitrogen and embedded in OCT sectioning.
2.3. Procedure 2.3.1. Antisera Mouse monoclonal anti-type I collagen (Biogenesis, Temecula, CA) and rabbit anti-type III collagen (Chemicon, Poole, Dorset, UK) antisera were used at dilutions of 1/100 and 1/1000, respectively. Biotinylated goat anti-mouse IgG (Becton & Dickinson, Milan, Italy) and goat anti-rabbit IgG (Calbiochem, Milan, Italy) were used as secondary antisera, at dilutions of 1/200 and 1/20 000, respectively. Avidin-biotin-peroxidase complex (Vector Laboratories, Burlingame, CA) was used in both cases to reveal the antigens. For titolation of collagen antibodies, leiomyomata tissue was used because of its known collagen type III/I ratio close to 1. 2.3.2. Immunoenzymatic method Eight-micron cryostat sections were mounted on poly-L-lysine coated slides and air dried for about 24 h. The sections were fixed with acetone (4°C) for 10 min, air dried and circled with waxpen (Dako, Milan, Italy). Endogenous peroxidase was inhibited with 0.3% hydrogen peroxide in phosphate-buffered saline (PBS) for 30 min. Positive control pancreas sections were processed in the same way. Sections were then rinsed in PBS at pH 7.3 (3× 5 min) and incubated for 10 min with normal goat serum (NGS) at the dilution of 1/20.
P. Affinito et al. / Maturitas 33 (1999) 239–247
The sections were incubated overnight at room temperature with the primary antisera. Negative control sections were incubated with normal rabbit serum (NRS), while unrelated antisera such as monoclonal mouse anti-insulin 1:1500 (Sigma, St Louis, MO) and polyclonal rabbit anti-glucagon 1:3000 (Chemicon, Milan, Italy) were used to stain pancreatic tissue, in order to obtain positive controls for the immunoenzymatic reaction. Nickel-sulfate enhancement with di-amino-benzidina chromogen (DAB), in the presence of 0.03% H2O2, was used to enzymatically reveal the antigens. Sections were counterstained with 2% methyl-green.
2.3.3. Computerised image analysis To quantify collagen type I and type III, advanced image analysis software (Image-Pro Plus, Media Cybernetics) was used. The images were acquired directly by optic microscope with enlargement 16 × using a special video camera. The histologic areas of tools were digitized and successively processed. The optic quality of these areas was optimized modifying the brightness and the contrast. The areas which were immunoreactive to the collagen were highlighted by the program on the basis of their levels of gray and measured. Reading of the levels of gray could be modified by varying the intensity of the range to highlight the real areas which were immunoreactive to the collagen. The surface occupied by collagen was expressed as a percentage of areas positive to the immunohistochemical reaction in comparison with the total examined area (% positive pixels/total pixels). This measurement was carried out on five sections of each skin biopsy. Six fields were analyzed in each section, three from the papillary dermas and three from reticular dermas. The mean value of collagen amount derived by the analysis of all areas in the five sections was reported. 2.4. Statistical analysis Statistical analysis of the differences in contents of collagen type I, type III and type III/type I ratio in patients in premenopause and in postmenopause was performed with Student’s t-test
241
for unpaired data. The correlation among percentage of collagen type I, type III and type III/type I ratio, and age and years of postmenopause was carried out by linear regression analysis. The statistical analysis was performed using StatSoft for Windows (Release 4.5, 1992). Statistical significance was defined as PB 0.01.
3. Results Fig. 1 shows that the percentages of skin collagen type I, type III and type III/type I ratio in postmenopausal patients were significantly (P B 0.01) reduced in comparison to percentages obtained in fertile age patients (5.799 0.75 vs 2.699 0.73, 3.07 9 0.54 vs 0.91 90.39, 0.53 9 0.08 vs 0.369 0.15, respectively). The percentages of collagen type I, type III and type III/I ratio of all patients showed a significant (PB 0.01) inverse correlation with chronological age (r= 0.88, 0.89, and 0.61, respectively) (Fig. 2). Considering only postmenopausal subjects, the correlation with chronological age was significant (PB 0.01) for collagen type I and type III of postmenopausal women (r= 0.59 and 0.64, respectively), but not for the type III/I ratio (r= 0.37, P= 0.131) (Fig. 3). The percentages of collagen type I, type III and type III/I ratio of postmenopausal women showed a significant (PB 0.01) inverse correlation with years of postmenopause (r= 0.76, 0.73 and 0.73, respectively) (Fig. 4).
4. Discussion The reduction of collagen is traditionally considered the main factor in the pathogenesis of skin atrophy. Moreover, the change in the amount of specific skin collagen component may be responsible for alterations of mechanical properties of skin. Indeed, the specific skin collagen types have different functions, collagen type I being responsible of the strength of the tissue and collagen type III contributing more to the elastic properties of the skin. However, estrogen-related reduction of
242
P. Affinito et al. / Maturitas 33 (1999) 239–247
Fig. 1. Percentage of skin collagen type I, type III and type III/I ratio in pre- and postmenopausal subjects. P B0.01 vs. premenopause.
P. Affinito et al. / Maturitas 33 (1999) 239–247
Fig. 2. Linear correlation between skin collagen type I, type III and type III/I ratio, and chronological age.
243
244
P. Affinito et al. / Maturitas 33 (1999) 239–247
Fig. 3. Linear correlation between skin collagen type I, type III and type III/I ratio of postmenopausal subjects and chronological age.
P. Affinito et al. / Maturitas 33 (1999) 239–247
245
Fig. 4. Linear correlation between skin collagen type I, type III and type III/I ratio of postmenopausal subjects and years of postmenopause.
246
P. Affinito et al. / Maturitas 33 (1999) 239–247
skin collagen has an importance not limited to skin. Indeed, this reduction may determine regressive change also in bone [1,4] and urinary tract [6 – 8], contributing to the pathogenesis of osteoporosis and stress incontinence. In this study, the population investigated was homogeneous with regard to the gynecological pathology requiring laparotomic operations. The choice to exclude patients with irregular cycles or with uterine leiomyomata, vaginal prolapse, and urinary incontinence was determined by the need to avoid possible interference on collagen metabolism. Indeed, the presence of uterine leiomyomata may be correlated with an increased amount of collagen while the presence of vaginal prolapse and urinary incontinence, and of oligohypoamenorrhoic cycles may be associated with a decreased concentration of collagen type I and type III, respectively [6]. The postmenopausal reduction of skin collagen content was previously reported using different methods for the evaluation of the collagen [1 – 5]. To our knowledge, our study is the first report on this issue using morphologic analysis of skin collagen. This method allows a direct and specific identification of different types of collagen. The results of this study show that postmenopausal women present a significantly lower amount of skin collagen type I, type III and type III/type I ratio in comparison with those of fertile age. Whether the postmenopausal loss of skin collagen is due mainly to postmenopause or chronological age is a controversial issue. A prospective study on skin collagen has shown that collagen was lost as a result of duration of estrogen deficiency after menopause and that it was possible to restore collagen to premenopausal levels within 6 months of initiating hormone replacement therapy (HRT) [2]. On the other hand, Castelo Branco et al. [3] have reported a close correlation of skin collagen loss with chronological age rather than time since menopause. However, in this study the category of the postmenopausal women had spent little time in postmenopause and therefore there was only a short period to observe the effects of estrogen deficiency.
This confusing factor may also explain the results of Haapasaari et al. [9]. In this study, 1 year of treatment with systemic estrogen alone or combined with progestin did not change the amount of collagen or the rate of collagen synthesis in postmenopausal women. However, the median menopausal age of the patients was 12 months. In comparison with other studies on this issue evaluating the total skin collagen content, our study is the first to investigate the effects of age and postmenopause on type of skin collagen and their relative ratio. Our data show that the amount of collagen type I, type III, and type III/type I ratio have a closed correlation with chronological age if we consider all patients (Fig. 2). Moreover, if we studied the relation between skin collagen and age only in postmenopausal subjects, we observe that the correlation results are significant (P B 0.01) for the collagen type I and type III, but not for the type III/I ratio. Indeed, the percentages of collagen type I, type III and type III/I ratio of postmenopausal women showed a significant (PB 0.01) inverse correlation with years of postmenopause. This correlation was clearer for years of postmenopause than for chronological age (r= 0.76 vs 0.59 for collagen type I, r=0.73 vs 0.64 for collagen type III, r= 0.73 vs 0.37 for collagen type III/type I ratio). These data show that the decrease in skin collagen in women of postmenopausal age involves both types of collagen and that it may not be an exclusively age-related phenomenon, but probably controlled by estrogen levels. Finally, we can hypothesize also that the collagen type III/type I ratio is under estrogen control and that the rapid postmenopausal change in the percentage of type of collagen involves specifically a reduction in the collagen type III. Confirmation of this hypothesis comes from data showing the rapid increase in the collagen type III/type I ratio in patients undergoing HRT [5].
References [1] Brincat M, Kabalan S, Studd JWW, Moniz CF, de Trafford J, Montgomery J. A study of the decrease of skin
P. Affinito et al. / Maturitas 33 (1999) 239–247
[2]
[3]
[4]
[5]
collagen content, skin thickness and bone mass in the postmenopausal women. Obstet Gynecol 1987;70:840–5. Brincat M, Versi E, Moniz CF, Magos A, de Trafford J, Studd JWW. Skin collagen changes in postmenopausal women receiving different regimens of estrogen therapy. Obstet Gynecol 1987;70:123–7. Castelo-Branco C, Duran M, Gonzales-Merlo J. Skin collagen and bone changes related to age and hormone replacement therapy. Maturitas 1992;15:113–9. Castelo-Branco C, Pons F, Gratacos E, Fortuny A, Vanrell JA, Gonzales-Merlo J. Relationship between collagen and bone change during aging. Maturitas 1994;18:199–204. Savvas M, Bishop J, Laurent G, Watson N, Studd J. Type III collagen content in skin of postmenopausal women receiving oestradiol and testosterone implants. Br J Obstet Gynaecol 1993;100:154–8.
247
[6] Bergman A, Elia G, Cheung D, Perelman N, Nimni ME. Biochemical composition of collagen in continent and stress urinary incontinent women. Gynecol Obstet Invest 1994;37:48 – 51. [7] Falconer C, Ekman G, Malmstrom A, Ulmster U. Decrease collagen synthesis in stress incontinent women. Obstet Gynecol 1994;84:583 – 6. [8] Versi E, Cardozo LD, Brincat M, et al. Correlation of urethral physiology and skin collagen in post-menopausal women. Br J Obstet Gynaecol 1988;95:147 – 52. [9] Haapasaari K-M, Raudaskoski T, Kallioinen M, SuvantoLuukkonen E, Kauppila A, Laara E, Risteli J, Oikarinen A. Systemic therapy with estrogen or estrogen with progestin has no effect on skin collagen in postmenopausal women. Maturitas 1997;27:153 – 62.
. .