PDFlib PLOP: PDF Linearization, Optimization, Protection Page inserted by evaluation version www.pdflib.com – [email protected]

Steroid Chemistry and Hormone Controls during the Hair Follicle Cycle h4ARTY E. SAWAYA Department of Dermatology and Biochemistry University of Miami School of Medicine Miami, Florida 33101 The human hair follicle (HF) is an interesting structure to study in skin because of its variable response to target tissue active androgens, such as testosterone (T) and dihydrotestosterone (DHT).Androgens stimulate terminal hair growth in certain body regions, such as axillary, pubic, and beard hair; whereas in HF of scalp androgens have an effect of shortening the anagen growth phase of the hair cycle, causing the HF to regress and recede in genetically susceptible men and women. Our investigative efforts focus on the androgen-mediated biochemical mechanisms influencing hair growth primarily in androgenetic alopecia (AGA). AGA is a hereditary condition that occurs in both men and women and is most likely inherited as an autosomal dominant trait with variable expression.’ Questions often asked regarding men and women who suffer with AGA are: (1) Are the biochemical mechanisms or processes that take place in men with AGA the same as in women who suffer from AGA? (2) Why is it that men have different patterns of hair loss than their female counterparts? (3) Are the HF in on region of the scalp or body “programmed” differently, or do they have a different sensitivity level to androgens? (4) Do blood hormone levels always factor into the “cause” of hair loss in women with AGA? Many of the answers to these questions are not clearly known, and with ongoing research the answers may not be as straightforward as hoped. That is, there could be various steps in the androgen metabolic pathway that could be at fault or the cause of turning off hair growth in scalp and yet stimulate hair growth in other body sites. Our research efforts focus on studying the androgen enzyme-receptor systems in HF of both men and women, and in various areas on the scalp. For example, we have analyzed the enzyme systems that mediate androgen action in various areas of the scalp where the HFs are sensitive to the effects of androgens, such as the frontal (Fr) to vertex areas, versus those areas where HF are most resistant to androgens, such as the occipital region (Occ). Investigators in the past293 have shown that the pilosebaceous unit, which comprises the HF and sebaceous gland (SG), enzymatically convert weak androgens, such as dehydroepiandrosterone (DHEA) and androstenedione (AD), to more potent target tissue androgens, such as T and DHT, by various enzymes, as shown in FIGURE 1. The enzymes in the pathway (FIG.1) are the delta-5, 3B-hydroxysteroid dehydrogenase (3B-HSD), 17B-hydroxysteroid dehydrogenase (17B-HSD), and Sa-reductase (5a-R), which convert weak androgens such as DHEA and AD to T and DHT; whereas the aromatase (A enzyme) converts T to estrogens, namely estradiol. The 3a-hydroxysteroid dehydrogenase (3a-HSD) converts DHT to a less active androgenic metabolite, with less potency than DHT, called 3a-androstanediol (AL). AL also has affinity for the androgen receptor protein (ARP) and can stimulate androgen actions similar to those seen with T. The conversion of androgens by these enzymes is dependent on oxidized-reduced pyridine cofactors, NAD, NADH, and NADPH. 376



Many of these androgen-converting enzymes shown in the FIGURE 1 pathway have been identified and characterized in whole human skin from face to genital areas. Differences in biochemical characteristics have been described for 5a-R from dermal papilla cells in HFs taken from one body region to another, suggesting the possibility that two separate 5a-R enzymes exist.4 Other ~ t u d i e s 2 ~have ~ ~ 5shown that levels of steroid metabolic enzymes differ in one body site to another. Recently, HFs in Fr versus Occ scalp of men and women with AGA revealed significant differences in levels of some of these androgenmediated enzymes.5 This suggests that HFs are independently regulated and that differences exist in each separate HF, locally and systemically. We have also studied the cytochrome P-450aromatase enzyme (A enzyme) in HF of men and women.5 This enzyme was thought to be located primarily in adipose tissue? The A enzyme, as mentioned above, converts androgens, such as AD and T to estrone and estradiol, respectively. The A enzyme has been studied by who showed it to be widely distributed in reproductive tissues, such as ovary, placenta, uterus, breast and who even showed to be of great importance in the brain. An enzyme has also been recognized in mediating tumor-associated processes in some of these tissues.9 The A enzyme is regulated by various cellular factors, such as CAMP, growth factors, gonadotropins, and even the substrates that utilize it.lo We have performed biochemical studies to assess the enzymes in the androgen metabolic pathway-the 3a-HSD, 3B-HSD, 17a-HSD, 5a-R, and A enzyme-using intact microdissected HF and sebaceous glands from scalp biopsy or hair transplant surgery from 10 men and 10 women between 18 and 35 years of age. The methods are stated in the legends of FIGURES 2 and 3.








FIGURE 1. Steroid metabolic pathway.



The results are shown in FIGURES 2 and 3 for analysis of the HSD enzymes in the pilosebaceous structures (PS unit) of HF and SG from scalp specimens taken from the receded, regressed frontal site and compared to the thick, anagen HFs taken from the occipital scalp. The specific activity for the enzymes is expressed as Vmax (pmol/min/mg protein). The analysis revealed that in the PS unit 3B-HSD activity differs from men to women. In men 3B-HSD is greater in the Occ versus the Fr region. In women the enzyme level is slightly elevated in Fr, and our findings show that the SG has greater levels of enzyme than HF. The interpretation of these findings is not clear, but shows that the HF and SG are capable of transforming the weak precursors, DHEA, to the more potent AD. As well, the 3B-HSD may be used to convert delta-5-androstenediol to T (FIG.1).Studies to compare the kinetic constants, Km values, for the two substrates, DHEA, and deltad-androstenediol have not been assessed to determine the primary pathway for these steroids in the human PS unit. As well, more detailed studies of the cofactors influencing the kinetic constants of these reactions should be further investigated. The 17B-HSD enzyme (FIG.2) is elevated in Fr HF of both men and women with AGA, showing men to have much greater levels than their female counterparts. It is also known that this enzyme reversibly converts AD to T, by use of reduced NAD; and that when the oxidized cofactor is used, the reaction favors the formation of AD. Once T is formed, it is converted to DHT by 5a-R (and NADPH, as cofactor). The analysis of 5a-R (FIG.3) in HF alone revealed that women with AGA have


Vmrx (pmol/min/mgP)



160 100 60



FO Occ


HF &



Mol.: Fr

SG from patients

Enzymo Aotlvlty n-10




FIGURE 2. Hydroxysteroid dehydrogenases in the pilosebaceous unit in androgenetic alopecia. Homogenates of isolated human SG and HFs from 10 men and women were incubated with tritium-labeled substrates DHEA, DHT, and AD to analyze for the 3BHSD, 3a-HSD, and 17B-HSD enzymes, respectively. Steroid transformation assays, thin-layer chromatography, kinetic studies, and statistical analysis of enzyme activity were done to determine the specific activity, Vmax (pmoVmin/mgP) value, as by previously published methods.’ The figures given are the mean 2 SE of all patients, assessed in duplicate, with Vmax shown on the y-axis. The abbreviations on the x-axis are the PS structures taken from female occipital (Fe:Occ), female frontal (Fe:Fr), male occipital (Male:Occ), male frontal (Male:Fr) biopsy or transplant scalp plugs analyzed.










F:. Fr MaleOcc HF Samplor from patkntr


Miaroromal 1 ractIon Aromataro onzymo


n 4 0 prtlentr

FIGURE 3. Aromatase and Sa-reductase levels in men and women with androgenetic alopecia. Microsomal fractions of human HFs were obtained from 10 men and women with AGA and were assessed for specific activity (Vmax, pmol/min/mgP) of aromatase enzyme and Sa-reductase. The steroid transformation assays, chromatographic methods, and kinetic and statistical methods are by previous method^.^*^*^^ Results are expressed as the mean 2 SE of all patients, with samples performed in duplicate and expressed as Vmax on the y-axis, from the female occipital (Fe:Occ), female frontal (Fe:Fr), male occipital (Male:Occ), male frontal (ma1e:Fr) biopsy or transplant plugs obtained.

greater levels of the enzyme in Fr versus Occ and yet nearly half of the enzyme level of men with AGA. Men with AGA have nearly two-fold greater 5a-R in Fr HF than in Occ HF. Results from analysis of A enzyme are also shown in FIGURE 3. The Fr HF in both men and women had less A enzyme than the thick Occ HFs. Overall, HF from women had nearly three to five times the level of A enzyme in Fr and Occ HFs in comparison to men. In women, where the frontal hairline was spared, the Fr HF had higher levels than receded, regressed neighboring Fr balding sites. Men with AGA have minimal levels of A enzyme in Fr HF. Some conclusions that can be drawn from the analysis include: (a) females with AGA have greater than twofold Sa-R in Fr HF versus Occ HF, and yet, nearly half of the amount of enzyme as men with AGA. @) Men with AGA have nearly twofold greater 5a-R in Fr HF, with minimal A enzyme. (c) Females may have sparing of the frontal hairline due to increased A enzyme, limiting the formation of 5a-reduced substrates that may bind to the ARP to initiate androgenic cellular events. (d) The elevated 5a-R and low A enzyme level seen in men with AGA may be due to formation of Sa-reduced products decreasing the amount of available substrate for the A enzyme to utilize; as it is known that substrate levels regulate levels of the A enzyme.10 Thus, similar hormone enzyme mechanisms may be working in men and women with AGA; however, the levels or amounts of enzymes differ between men and women, perhaps giving some explanation of the varied patterns of hair loss noted in men versus women displaying a diffuse frontal-to-vertex thinning.



Immunohistochemical studies recently done by SawayalOa revealed that A enzyme to be located in the lower portion of the external root sheath of HF, with accentuated staining in anagen HFs. Finding the A enzyme in the external root sheath may suggest that circulating hormones are metabolized or processed prior to entry into the hair bulb or dermal papilla cells, where ARPs are most concentrated and are likely to be the target area stimulating gene expression. We hypothesize that A enzyme may be minimizing or protecting the HF from synthesizing DHT by conversion of circulating systemic levels of AD and T to estrogens, thus sparing the frontal hairline in women with AGA due to high levels of A enzyme, minimizing the presence of androgens for binding to ARP, which may alter or affect hair growth. Finding A enzyme in the outer external root sheath suggests that several other steroid enzymes-that is, 5a-R or the HSD enzymes-may be located there as well and could have a role in hormone processing/metabolism prior to entry into the dermal papilla cells, therefore regulating hair cycles. The 3a-HSD enzyme (FIG. 2) has not been studied in as much detail. Previous work describes the primary metabolite formed from this enzyme-namely, 3aandrostanediol from T and DHT metabolism (FIG.1). The results from our studies show that there were slightly elevated levels for the 3a-HSD in the Fr PS of both men and women. AL is also a potent androgen, which has affinity for binding to the ARP and inducing androgenic processes. The oxidized-reduced pyridine cofactors are also of importance, since these cofactors are required for the steroid enzymic conversions to take place, and they determine the direction of synthesis for the 17B-HSD, a reversible reaction. The cofactors are synthesized in other metabolic pathways, such as the TCA cycle, the pentose phosphate pathway, and others. The cofactors are also utilized in numerous biochemical reactions systemically, and it has been suggested that levels of oxidizedreduced cofactors may be influenced by androgens, perhaps affecting the direction for some of the local enzyme reactions-that is, the formation of AD or T by the 17B-HSD. Once DHT is formed, it binds to a specific intracellular receptor protein, namely the ARP, which forms an “activated” hormone-receptor complex (HRC). The ARP is dependent on other “activation” factors found in the cell to form the HRC. It is thought that steroid receptors are phosphoproteins, phosphorylated by specific kinases, and it is thought that the receptor must be in sulfhydryl-reduced state for optimum hormone binding to occur.ll The phosphorylating and sulfhydryl-reducing enzymes necessary for receptor “activation” to form the HRC are of great importance in understanding the processes taking place in AGA. The sulfhydryl-reducing factor, believed to be a thioredoxin enzyme system (TR) influences the intramolecular disulfide bonding of the ARP, which affects hormone binding to its binding site.’* The TR is found in high levels in the active, anagen growing HFs in scalp and is found to be diminished in regressed, receded HF. It is not certain if finding this factor in low amounts is a reflection of the “balding” process or represents HF that are in a metabolically reduced state, such as normal resting telogen HF. The TR also utilizes the reduced cofactor NADPH. The TR has been shown by other investigators to be important in ~ k i n . ’ The ~ J ~TR functions as an electron transfer protein common to all living cells, with its main functions including: reducing free radicals at the surface of the skin, an antioxidant in reduction of methionine residues, reducing disulfide links in proteins, and an electron donor in ribonucleotide reductases for RNA synthesis. Previous investigators have shown the TR system to be important in “activation” of some of the steroid receptors. Our work has shown it to be important for activation



of the glucocorticoid receptor (GCR) in skin, with important findings of showing elevated or suppressed levels of TR in certain skin diseases; for instance, elevated levels were found in psoriasis and keloids, but TR is suppressed in alopecia areata.13.15.16 Another factor isolated from HF called “inhibitor protein” (IP), an 18-kDa protein, was found in human anagen HF.17The IP was shown to be important in regulating hormone binding to the ARP ligand binding site. Various low-molecularweight proteins have been found in other tissues to affect steroid binding to their respective receptors. IP may bind near the hormone binding site to alter the conformational shape of the ARP, discouraging the steroid to bind to its ligand binding site, hence giving an overall effect of limiting hormone binding. It is recognized that there are many complex steps required for specific hormone binding to the ARP to form the HRC. The HRCs are primarily found to be concentrated in the nucleus of the cell, where they are important for binding to nuclear chromatin hormone response element (HRE) sites, which signal neighboring regulatory or structural gene sites to induce transcription, affecting cell expression-in this case, hair growth. Our most recent investigations have focused on the complexities of the specific binding interaction of the HRCs to the HRE. A nuclear matrix-associated acceptor protein (NAP), a 1ZkDa protein, was recently discovered and found to mediate high-affinity binding of HRCs to DNA.lSNAP was isolated and extracted from a pool of other nonhistone nuclear proteins and found to be specific and unique in its specificity of mediating saturable binding of HRCs to DNA. Present studies show that removing NAP from chromatin using detergents revealed loss of high-affinity binding. The NAP was found in HF of men and women. Future work to characterize the DNA fragments associated with NAP will be important in finding the specific DNA sequences necessary for HRC binding and the specific androgen-regulated genes signaling synthesis of cellular proteins altering HF growth in AGA. Cellular changes occur during the “aging” process and can also alter hair growth during the maturation stages of life. Age-related molecular protein^'^ have been found to affect gene expression by “masking” specific positive and negative HRE sites important for hair growth. These age-related nucleoacidic, nonhistone proteins are thought to mask and occlude certain gene sites that can alter cell growth, as can occur in the aging process.*O Thus, there may be age-related cellular processes that may be independent of the androgenic pathways that affect hair growth. In conclusion, future research efforts focused on finding better agents to stimulate hair growth for the treatment of AGA may be better formulated based on research efforts understanding the biochemical processes involved with hair growth. As is evident from this presentation, the cellular processes involved in hair growth regulation are complex. However, with present technology and with further endeavors in the field, the potential to develop more suitable treatment options is favorable for the millions of men and women who suffer from hair diseases such as AGA.

SUMMARY Human hair follicles contain several steroid enzymes capable of transforming weak androgens, such as dehydroepiandrosterone, into more potent target tissue androgens, such as testosterone and dihydrotestosterone. Kinetic constants have been evaluated for the 3-alpha, 3-beta, and 17-beta hydroxysteroid dehydrogenase enzymes, Sa-reductase, and the aromatase enzyme in isolated human HF from scalp of



men and women with androgenetic alopecia. The apparent K,,, values did not differ for each enzyme whether present in bald, receded HF or thick, anagen H F of men or women. However, levels of specific activity varied greatly in the frontal versus occipital HF analyzed. The androgen receptor content and activation factors also differ between men and women. The steroid mechanisms influencing AGA in men and women may be similar, but differences in the specific activity/amounts of enzymes, receptors, and activation factors differ between men and women. These findings may explain the varied clinical presentations of men and women with AGA, and may shape treatment options for the future.

REFERENCES 1. SMITH, M. A., R. S. WELLS.1964. Male-type alopecia, alopecia areata, and normal hair in women. Arch. Dermatol. 8 9 95-98. 2. SCHWEIKART, H. U. & J. D. WILSON.1974. Regulation of human hair growth by steroid hormones. 11. Androstenedionemetabolism in isolated hairs. J. Clin. Endocrinol. Metab. 3 9 1012-1019. 3. SAWAYA, M. E., I. S. HONIG,I. D. GARLAND& S. L. HSIA.1988. -3B-Hydroxysteroid dehydrogenase activity in sebaceous glands of scalp in male pattern baldness. J. Invest. Dermatol. 91: 101-105. S., S. KURATA, T. SONODA & S. TAKAYASU. 1991. Characterization of 5a-reductase 4. ITAMI, in cultured human dermal papilla cells from beard and occipital scalp hair. J. Invest. Dermatol. In press. 5 . SAWAYA, M. E., V. H. PRICE,K. A. HARRIS, R. S. KIRSNER & S. L. HSIA.1990. Human hair follicle aromatase activity in females with androgenetic alopecia. J. Invest. Dermatol. 9 4 575. 6. EVANS,C. T., C. J. CORBIN, C. T. SAUNDERS, J. C. MERRILL, E. R. SIMPSON& C. R. MENDELSON. 1987. Regulation of estrogen biosynthesis in human adipose stromal cells. J. Biol. Chem. 262: 6914-6920. J., A. FOIDART & N. HARADA. 1990. Immunocytochemical localization of 7. BALTHAZART, aromatase in the brain. Brain Res. 514: 327-333. 8. RYAN,K. J. 1982. Biochemistry of aromatase: Significance to female reproductive physiology. Cancer Res. Suppl. 42: 3342-3344. S. J., M. E. BRANDT, D. F. COVEY & D. PUETT.1987. Inhibition of aromatase 9. ZIMNISKI, activity and endocrine responsive tumor growth by 10-propargylestr-4-ene3, 17 dione, and its proprionate derivative. Steroids 50: 135-146. 10. BERKOVITZ, G. C., M. FUJIMOTO, T. R. BROWN, A. M. BRODIE& C. J. MIGEON. 1984. Aromatase activity in cultured human genital skin fibroblasts. J. Clin. Endocrinol. Metab. 59: 665-671. 10a. SAWAYA, M. E. & N. S. PENNEYS. 1992. Immunohistochemicaldistribution of aromatase and 3B-hydroxysteriod dehydogenase in human hair follicle and sebaceous gland. J. Cutaneous Pathol. In press. 11. PELEG,S., W. T. SCHRADER& B. W. O'MALLEY.1988. Sulfhydryl group content of chicken progesterone receptor: Effect of oxidation on DNA binding activity. Biochemistry 27: 358-367. 12. SAWAYA, M. E., L. A. LEwis & S. L. HSIA.1989. Presence of a converting factor for androgen receptor proteins in isolated human hair follicles and sebaceous glands. FASEB J. 2: 4765. 1987. Anthralin inhibits elevated levels of 13. SCHALLREUTER, K. U. & M. D. PITTELKOW. thioredoxin reductase in psoriasis. Arch. Dermatol. 123: 1494-1498. 14. SCHALLREUTER, K. U. & J. M. WOOD.1986. The role of thioredoxin reductase in the reduction of free radicals at the surface of the epidermis. Biochem. Biophys. Res. Commun. 136 630-637. R. J. COHEN, K. A. HARRIS & L. A. SCHACHNER. 1990. 15. SAWAYA, M. E., M. K. HORDINSKY, Elevated unoccupied glucocorticoid receptors in scalp of patients with alopecia areata. J. Invest. Dermatol. 94: 574.



16. SAWAYA, M.E.,R. S. KIRSNER, A. J. NEMETH,D. S. WEISS& S.L. HSIA.1990.Elevated type I1 glucocorticoid receptor binding in keloids and hypertrophic scars. J. Invest. Dermatol. 94: 575. 17. SAWAYA, M.E.,A. I. MENDEZ & S. L. HSIA.1988.Presence of an inhibitor to androgen binding to receptor protein in human sebaceous gland and hair follicle. J. Invest. Dermatol. 90: 605. M.E.,C. A. KRAFFERT & S.L. HSIA.1991.A nuclear matrix associated acceptor 18. SAWAYA, protein involved in the chromatin binding of the androgen receptor regulating human hair follicle growth in androgenetic alopecia. J. Invest. Dermatol. Submitted for publication. 19. SAWAYA, M. E.,C. A. KRAFFERT, L. A. LEWIS, M. IRIONDO & S.L. HSIA.1990.Age related molecular proteins affecting hair growth in men with androgenetic alopecia. J. Invest. Dermatol. 94 575. 20. CHUKNYISKA, R. S. & G. S. ROTH. 1985. Decreased estrogenic stimulation of RNA polymerase I1 in aged rat uteri is apparently due to reduced nuclear binding of receptorestradiol complexes. J. Biol. Chem. 15: 8661-8663.


S. ITAMI(Medical College of Oita, Japan): I think estrogen may have some action on hair growth. Can estrogen act directly on hair growth? Otherwise, do hair follicles have estrogen receptors? Have you already checked this? M.E. SAWAYA: I haven’t looked for the estrogen receptor in hair follicle tissue. I’ve analyzed it for other tissue types, such as basal cell carcinoma and squamous cell carcinomas and melanomas, but not for hair follicle growth. The role of estrogens in hair follicle growth is not really well understood, even in the literature, and I’ve gone back to animal studies. Exactly what estrogens are doing even in combination with other types of hormones is unclear, so it leaves us a lot to study, especially in regards to aromatase. A lot of the work on the immunohistochemistry of aromatase was done with Neal Penneys in my department. He believes that aromatase has a lot to do with controlling the hair cycle from anagen to telogen and processing androgens to estrogens. We are looking at aromatase with regard to chemotherapeutic agents. ITAMI:Is the estrogen effect indirect? SAWAYA: Does it compete with androgen action? We don’t know what the estrogens are actually doing and whether they could have some other regulatory role. How clean are your hair follicle preparations? Have you UNiDENTiFiEo SPEAKER: done any histology to see if there are bits of dermis still stuck there? SAWAYA: Yes, we’ve done some histology on it, and they’re pretty clean. We have also analyzed the surrounding tissues for enzyme and receptor contact to show that they are more specifically located in the whole hair follicle and sebaceous gland rather than in tissues without pilosebaceous structures. SPEAKER: Have you measured the levels of these enzymes-aroUNIDENTIFIED matase and so forth-in nonbalding scalp? in different sites of nonbalding scalp? SAWAYA: In completely normal individuals, no. But, these are done from frontal versus occipital in women and men; so whether you want to call these normal in the back in the occipital, I’m sure they’re still hormonally changed in some way. We haven’t looked at “totally normal” individuals. These are fronts and backs of the same person. T. TAKEUCHI (Tohoku University,Sendai,Japan): When you measure the enzyme activity of 17-beta HSD, do you use hair follicle alone or together with sebaceous glands?



SAWAYA: The HSD enzymes are heavily located in sebaceous gland. Although we combined them, we do have data individually. The 5-alpha reductase reported here was just given for the hair follicle itself, since those are potent androgens and do have an effect directly on the hair follicle. UNIDENTIFIED SPEAKER: Since you work with whole pilosebaceous units, have you looked also at just isolated hair follicle. Since we know that the response of the sebaceous glands and the hair follicles are opposite, it is really very important to separate the components. SAWAYA: Right. No, we haven’t separated those, especially the HSDs; 3-beta HSD is intensely located in sebaceous glands and to a more minor extent in hair follicle; but yes, we do have that data individually done, too. I just did it for the sake of simplicity here. DR. HODGINS: I’m rather puzzled by your aromatase staining in the outer root sheath. When we originally reported about dermal papilla cells for aromatase activity we also noted that in spite of trying very hard, we could not detect aromatase activity in plucked hairs. M. B. HODGINS (Glasgow University, Glasgow, U.K.):Plucked anagen hairs have lots of outer root sheath. That’s the main cellular component. We have always checked our plucked hairs for outer root sheath. What I want to ask you is have you managed to detect aromatase activity in plucked hairs? And, two, what is the specificity of your aromatase antibody? Are you sure it isn’t cross-reacting with another member of cytochrome P450 found in enzymes? SAWAYA: To my knowledge there is a monoclonal and a polyclonal antibody for aromatase, and Dr. Penneys feels that this is specific for aromatase. He has analyzed hair follicles from different body regions. We have not analyzed plucked follicles. HODGINS: I would think it is absolutely essential to demonstrate that this enzyme activity is present in plucked hair follicles before we can accept that there is a significant amount of this enzyme activity in the outer root sheath.

Steroid chemistry and hormone controls during the hair follicle cycle.

Human hair follicles contain several steroid enzymes capable of transforming weak androgens, such as dehydroepiandrosterone, into more potent target t...
543KB Sizes 0 Downloads 0 Views