The Action of Testosterone and Oestradiol on the Sebaceous Glands and Epidermis of the Rat

That androgens increase the size of sebaceous glands is well known (de Graaf, 1943; Ebling, 1948; Montagna & Kenyon, 1949; Hamilton & Montagna, 1950; Haskin, Lasher, & Rothman, 1953). This effect might be due to increased cell proliferation, since it has been shown that testosterone stimulates mitosis in the epidermis of the mouse (Bullough & Van Oordt, 1950) or to increased cell size. But these are not the only possible explanations. Oestrogen has a marked effect in reducing epidermal thickness and the size of sebaceous glands without reducing the incidence of mitoses and it has been suggested that this is due to accelerated loss of cells (Hooker & Pfeiffer, 1943; Ebling, 1948, 1954, 1955). Thus testosterone might produce opposite effects by decreasing the rate of cell loss. Moreover, it must not be assumed that sex hormones have the same effect in the rat as in the mouse; e.g. the effect of oestrogen in stimulating epidermal mitosis in the mouse, reported by Bullough (1946, 1950 a, b), appears to be lacking in the adult rat (Carter, 1953; Ebling, 1954, 1955). For these reasons it seemed desirable to investigate the mechanism of the action of testosterone on the epidermis and sebaceous glands of the rat.

Forty female Wistar rats from a small randomly mated colony were used for the experiments, which were designed so that the rats were killed when about 25 weeks old. All rats were spayed 19—22 days before death. On the day when they were killed at 1500 hours they were injected intraperitoneally at 1000 hours with 0·1 mg. colchicine in water per 100 g. body-weight. The use of colchicine for the determination of the incidence of epidermal mitoses in the rat has already been discussed (Ebling, 1954).

At the time of spaying the rats were divided into four groups of ten which were treated as follows:

• I. Untreated controls.

• II. Implanted subcutaneously with about 8 mg. of testosterone.

• III. Implanted with 8 mg. testosterone and about 5 mg. of 1:9 oestradiol / cholesterol.

• IV. Implanted with 5 mg. of 1:9 oestradiol/cholesterol.

The implants were weighed after removal; the average absorption of testo sterone appeared to have been roughly of the order of 0·3 mg./day, and of oestradiol 2–4 μg. / day. These figures are in general agreement with amounts calculated from the known rates of absorption in man in relation to the surface area of the implants. The data of the manufacturers (Organon Laboratories) suggest that, on average, 0·2 mg. of testosterone and 3 μg. oestradiol would be absorbed each day.

The skins were removed and fixed as described in a previous paper (Ebling, 1954). Samples taken from near the mid-line in the anterior region of the back were embedded in ester wax (Steedman, 1947), sectioned sagittally at 7 μ, and stained in Ehrlich’s haematoxylin and alcoholic eosin.

Whole mounts were prepared from similar pieces of skin about 1 cm. square by staining them for about 24 hours in a mixture of absolute alcohol, sodium hydroxide, and Sudan IV. Subcutaneous fat and the panniculus carnosus were dissected off, and the skin was mounted in glycerine jelly (Badertscher, 1940).

If a sphere is cut into thin slices of equal thickness its volume is equal to the product of the total area of the slices and their thickness.

The number of alveoli per square millimetre was also estimated in whole mounts of skin from counts of ten fields (2·01 mm. in diameter) for each rat.

The thickness of the stratum germinativum plus stratum granulosum or epidermal thickness (measured from base of stratum germinativum to base of stratum corneum), the number of arrested metaphases per fifty sebaceous gland sections, and the number of metaphases per 1 cm. length of stratum germinativum, were all determined as previously described (Ebling, 1954).

In Table 1 the actual observations are summarized; the transformed figures shown in Tables 2 and 4 are calculated not from the means shown in Table 1 but from the individual results for each rat.

It is clear (Table 2) that alveolar volume was very significantly increased, on average by about 80 per cent., in thè rats which were implanted with testosterone, whereas oestradiol reduced the mean volume by about one-half. In no groups were there any significant changes in the alveolar number as calculated from sagittal sections. Since the glands tend to be sac-shaped rather than spherical, an exaggeration of alveolar volume with a corresponding underestimate of alveolar number might have been expected. An examination of the alveolar numbers as estimated from whole mounts (Table 3) shows that such an error is very slight and significant only in the testosterone-treated group in which the glands were very large. This suggests that the estimates of alveolar volume in Table 2 are reasonably accurate. In Table 3 alveolar number appears to have been slightly higher in rats treated with both testosterone and oestradiol than in the other groups. However, this may well be because the glands of the rats treated with testosterone stained much more satisfactorily than the others.

The effect of testosterone on alveolar volume is in part due to a significant increase (about 16 per cent, on average) in cell volume (Table 2). By dividing cell volume into alveolar volume an estimate of the number of cells per gland may be made. This shows that testosterone caused a very significant increase in cell number, amounting to about 60 per cent. (Table 4). The question of whether this is due to increased cell proliferation or decreased cell loss now arises. It appears from Table 1 that the incidence of mitoses per fifty gland sections per 5 hours was significantly increased. The total number of cells visible in fifty gland sections can be calculated from the sebaceous cell and alveolar counts and thus it is possible to estimate the proportion of cells undergoing mitosis in 5 hours and from this to make a comparative estimate of the number of cells produced per gland per day (Table 4). As these figures make the unwarranted assumption that the incidence of cell division does not vary diurnally, and as no absolute value can be assigned to mitotic counts made with the use of colchicine, they must be regarded of value only for comparative purposes. If the total number of cells per gland is divided by the number produced per day it is possible to make an estimate of the time of cell survival, shown in Table 4 as the ‘estimated cell life’. The results show that testosterone more than doubled the mean rate of cell proliferation in the glands, but elicit the interesting fact that at the same time the length of cell life was decreased, i.e. the rate of cell loss was increased.

The analysis of the mode of action of oestradiol is not quite so clear cut. Cell volume was not significantly affected (Table 2). The number of cells per gland was approximately halved (Table 4), although the number of mitoses per fifty gland sections was not affected (Table 1). The possibility that cell proliferation was slightly reduced cannot, however, be ruled out (Table 4), and it seems likely that the cell life was lowered, although the difference in the means is not statistically significant.

In the rats treated with both testosterone and oestradiol the mean volume of the sebaceous glands was only about 20 per cent, greater than in the controls, and the difference is not statistically significant. Moreover, the number of cells per gland is not significantly above normal (Table 4). Nevertheless, it appears that each hormone retained its individual effects. The cell volume was equal to that of rats treated with testosterone alone and significantly higher than that of the controls. The rate of cell proliferation was double that of untreated rats, and the estimated cell life significantly less.

The results are shown in Table 5. Testosterone did not affect epidermal thickness, nor was the incidence of mitoses significantly changed. It is not impossible, however, that testosterone produced a slight increase in the incidence of mitoses in both groups II and III. Oestradiol caused a significant decrease in epidermal thickness which was clearly not due to decreased cell proliferation; therefore it seems that cell loss must have been accelerated. A decrease in epidermal thickness probably occurred in the rats treated with both hormones, although the difference in the means is not statistically significant.

From the present study it appears that the enlargement of the sebaceous glands, induced in the rat by testosterone, is caused partly by an increase in cell volume, which accounts for about one-quarter of the effect, but mainly by an increase in cell proliferation. At the same time the rate of breakdown of the cells is accelerated, that is to say their length of life is reduced. These effects on cell size and cell proliferation seem to retain their identity even when oestradiol, which by itself reduces the size of the glands, is given at the same time as testosterone, even though such glands may not be enlarged.

The mechanism by which oestradiol reduces gland size is not quite so clearly demonstrated. It does not affect cell size in adult spayed rats. Although in these experiments the reduction in mean cell life was not statistically significant, there seems to be little doubt that oestradiol does accelerate cell breakdown in both sebaceous glands and epidermis. It has, for example, been shown that in immature rats these structures can be greatly reduced in size at the same time as the incidence of mitoses is increased (Ebling, 1954) and other evidence in support of such an action has already been presented (Ebling, 1955). But it does seem possible that a decrease in cell proliferation may account for a minor part of the reduction in gland size in the present study.

Lapière (1953) has studied the effect of testosterone and oestradiol applied in oil to the skin of mice. The present observations on the rat are in keeping with his conclusion that testosterone increased the diameter of the sebaceous glands by increasing both cell size and mitotic index. He claims, however, that the mode of action of oestradiol is exactly opposite to that of testosterone, i.e. that it reduces both cell volume and the mitotic index. My evidence does not support either of these conclusions. Cell volume was not significantly altered, and it seems unlikely that a lowered incidence of mitoses could explain more than a small part of the effect of oestradiol on the sebaceous glands of the rat. Moreover, even in the mouse such a conclusion may not be a correct inference from Lapière’s data. The ‘mitotic index’, which is estimated without the use of colchicine, is the number of cells per 1,000 that are undergoing mitosis at the instant of fixation of the tissue. This index depends, therefore, not only on the number of new cells being produced but also on the duration of mitosis. Thus a lowered mitotic index could be due to a speeding up of mitosis, so that there were fewer cells in division at any instant, and need not indicate a lower production of cells. Bullough ( 1950b) has presented evidence that oestrogens do increase the speed of mitosis in this way. Finally, even if it were true that oestradiol did not alter the duration of mitosis, it would then follow that a reduced mitotic index must indicate a reduced rate of cell loss as well as reduced cell proliferation. It has been shown that such an effect is unlikely.

The results may throw some light on the suggestion that the human complaint acne vulgaris, which involves enlarged sebaceous glands, is due to an increase in the androgen / oestrogen ratio as put forward, for example, by Lawrence & Werthessen (1940) and Aron-Brunetière (1953). It would appear that although the size of the sebaceous glands depends on the androgen/oestrogen ratio, the amount of sebum actually secreted, which is dependent only on cell proliferation and cell volume, is affected mainly by the amount of androgen. A rough comparative estimate of glandular output in the experiments can be made from the data in Tables 2 and 4. Testosterone increased the mean output from about 29,000 cu. p to over 80,000 cu. p per gland per day. Oestradiol caused a slight reduction to about 21,000 cu. p. In rats treated with both hormones the mean output was 62,000 cu. p. These figures emphasize that although superficially oestradiol and testosterone appear to affect gland size in opposite ways, their individual effects are in fact maintained when both hormones are administered at the same time.

I should like to thank Miss Elizabeth Johnson for helping with some of the measurements and Mr. John Skinner for technical help. My grateful thanks are due to Dr. W. J. Tindall of Organon Laboratories Ltd. for the implants of testosterone and oestradiol/cholesterol.