Testosterone and estradiol concentrations in serum, velvet skin, and growing antler bone of male white-tailed deer

JOURNAL OF EXPERIMENTAL ZOOLOGY 303A:186–192 (2005) Testosterone and Estradiol Concentrations inSerum, Velvet Skin, and Growing Antler Bone ofMale White-Tailed Deer GEORGE A. BUBENIK1, KARL V. MILLER2, ANDREA L. LISTER1,DAVID A. OSBORN2, LUDEK BARTOS3, AND GLEN J. VAN DER KRAAK11Department of Zoology, University of Guelph, Guelph, Ontario, Canada,N1G 2W1 2Warnell School of Forest Resources, University of Georgia, Athens,Georgia 30602 3Ethology Group, Research Institute of Animal Production, CZ 104 01 Prague,Czech Republic The growth and mineralization of antlers correlate with the seasonal variation of serum androgens. Whereas seasonal levels of testosterone (T) in plasma are well established, steroidconcentrations have not yet been determined in the tissues of growing antlers. Therefore, RIA wasused to determine T and 17b estradiol (E2) in serum, and three areas (tip, middle, and base) of theantler bone and the antler skin, called velvet. Blood and antler tissues of white-tailed deer(Odocoileus virginianus) were collected from May to August. The difference between levels of T andE2 among the sites was calculated using the square root transformation followed by a mixed modelanalysis with individual deer and an interaction of individual and year (individualnyear) as a randomfactor. Concentrations of T in serum (799782 pg/ml) were higher than T values in the velvet(589758 pg/ml, Po0.01) and in the antler bone (538758 pg/ml, Po0.001). Estradiol concentrationsdiffered among antler tissues and serum (Po0.001) and between years (Po0.01). Estradiolconcentrations in serum (25725pg/ml) were consistently lower than those in antler bone (208711pg/ml, Po0.001) and velvet (150712 pg/ml, Po0.001). The E2:T ratio in serum was 1:10–60. Thesame ratio for the antler bone was only 1:2–3 and for the velvet 1:3.5. It is concluded that higher Tand lower E2 concentrations found in plasma, as compared to antler bone or antler velvet, mayindicate a partial metabolism of systemic androgens into estrogens xin the tissues of growing antlers.
J. Exp. Zool. 303A:186–192, 2005.
androgens are essential for vigorous antler growth(Bubenik, ’82; Schams et al., ’87; Bartos et al., Development of antlers correlates with seasonal 2000). In addition, T was localized immunohisto- concentrations of reproductive hormones. Plasma logically in the growing antlers of white-tailed concentrations of testosterone (T) are minimal deer (Bubenik et al., ’74) and a cytosolic T receptor during antler growth, then rapidly increase during was detected in the growing antler tissue of sika antler mineralization and reach peak levels deer (Cervus nippon) (Li, ’87). However, because shortly before the breeding season (Lincoln et al., antlers are growing during a period of minimal ’70; Bubenik et al., ’75, ’82; Suttie et al., ’84; Muir blood concentrations of T, and antler growth is et al., ’88). It is generally accepted that androgens also initially vigorous in castrated deer, some are involved in the maturation and mineralization authors concluded that T is not essential for of antlers (Tachezy, ’56; Lincoln et al., ’70; Morrisand Bubenik, ’82). Conversely, the role of andro- This study was partly supported by a grant from the Ministry of Agriculture of the Czech Republic (MZE0002701402) and by J. William gens in antler growth remains controversial.
Deer antlers are a secondary sexual character- nCorrespondence to: George A. Bubenik, Department of Zoology, University of Guelph, 50 Stone Rd. E., Guelph, Ontario, Canada, N1G istic and testosterone may have a role in their development (Bartos et al., 2000). Several studies Received 20 May 2004; Accepted 8 November 2004Published online in Wiley InterScience (www.interscience.wiley.
indicate that minimal, threshold concentrations of TESTOSTERONE & ESTRADIOL IN DEER SERUM AND ANTLERS antler growth (Wislocki, ’43; Kolle et al., ’93; Suttie et al., ’98). Conversely, several other studiessupport the role of androgens in the antler growth.
Castration and inhibition of androgen receptors by Serum and antler tissue samples were collected cyproterone acetate result in antlers covered either from captive bucks immobilized with a 1:1 permanently in velvet and in the loss of the combination of xylazine (AnaSed, Lloyd Labora- species-specific antler shape. When castrates are tories, Shenandoah, Iowa) and ketamine (Ketaset, treated with T (which causes mineralization and Fort Dodge Animal Health, Iowa) at 1–2 mg, or casting of old antlers), the shape of new antlers from free ranging bucks euthanized by gunshot.
grown in a subsequent year will return to their The captive group was part of the research herd species-specific shape (Lincoln, ’75; Bubenik, ’82, maintained at the Daniel B. Warnell School of Forest Resources, University of Georgia, Athens, Several studies indicate that growing antlers are Georgia. Free ranging deer were collected from metabolically active tissue producing growth-pro- Chatham, Clarke, Murry, and Newton Counties in moting chemicals such as alkaline phosphatase Georgia; Grenada County in Mississippi; and (Bubenik et al., ’87), epidermal growth factor (Ko Charleston and Jasper Counties in South Caroli- et al., ’86), and vitamin D (Sempere et al., ’89).
na. Blood was taken from the jugular vein of Conversely, velvet antler tissue may utilize var- captive deer and from the heart ventricle or chest ious growth-promoting hormones, such as triio- cavity of free ranging deer. The blood was allowed dothyronine (Bubenik et al., ’87). One of the to clot and the separated serum was transferred hormones produced in the growing antlers may be into cryogenic vials, and stored at À201C until it was assayed for steroids. Strips of antler velvet The role of E2 in the development of antlers has (i.e. epidermis and dermis) measuring 2.0 Â 1.0 cm been investigated in several studies. Estradiol is were collected from near the base, mid-line, and the most potent steroid causing maturation and tip of the main beam of one antler of each deer.
mineralization of antlers in intact and castrated Antler bone samples (cross section of approxi- deer (Blauel, ’35; Tachezy, ’56; Goss, ’68; Morris mately 1cm thick) were also collected from near and Bubenik, ’82) but its mode of action has not the base, mid-line, and tip of each free ranging been determined. Seasonal concentrations of E2 in deer. Antler samples were stored in physiological white-tailed deer plasma peak during antler saline and frozen at À201C until assayed for growth (May) and again during the rut (Novem- ber). This second peak of E2 coincides with thepeak concentrations of T in plasma (Bubenik et al., ’79). Blockade of E2 receptors by CI–628 andMER–25, administered during the antler growth, Segments of velvet (0.3 g of tissue) were cut to severely impaired formation of compact ivory bone tiny pieces by a razor blade and combined with and delayed shedding of antler velvet (Bubenik 1 ml of double-distilled water. The tissue was homogenized using Beckman Polytron, trans- the velvet of growing antlers (Barrel et al., ’87), ferred into 20 ml glass tube, sonicated for one providing additional evidence of the role of E minute, and the steroids extracted for RIA. Antler the antler growth. Finally, an early study (Bube- bone (0.3g) was finely ground using a nutmeg nik GA, Raeside J. – personal communication) grinder. After adding 1 ml of double-distilled water, the treatment was identical to the techni- outflowing antler vein were higher than the que described above. Serum (0.4 ml) was mixed with 0.6 ml of double-distilled water prior to therefore hypothesized that similarly to other target organs for sexual steroids, antler bone is capable of producing E2, by using T as a substrate (Bubenik, ’90). To better understand the role ofsteroid hormones in the tissues of growing antlers, Plasma concentrations of T (N¼13) and E2 and to test the above hypothesis, concentrations of (N¼11) were measured by RIA after extraction T and E2 were determined by RIA in the serum, with diethyl ether according to the methods velvet, and the antler bone of male white-tailed described by Van Der Kraak and Chang (’90) and McMaster et al. (’92). Antler bone and velvet homogenates (1ml) and serum samples (400ml) account. Therefore, least-squares-means (LSM) diluted in 1 ml of ultra-pure water were mixed were used instead. LSM are, in effect, within- with 5 ml of diethyl ether and vortexed three times group means appropriately adjusted for the other for 30 seconds. After snap-freezing the aqueous effects in the model. LSM (further referred to as components in cold acetone, the ether phase was ‘adjusted means’) were computed for each class decanted into glass vials for evaporation. Samples and differences between classes were tested by t- were reconstituted in 1 ml of phosgel (5.75 g test. The Tukey-Kramer adjustment was used for Na2HPO4, 1.28 g NaH2PO4ÀH2O, 1.0 g gelatin, 0.1 thimersol, 1 L ultrapure water) and stored at Two tests were performed for both T and E2 À201C until analysis by RIA using tritiated T or levels. First, a mean value was used for bone and E2 (Amersham Biosciences, Baie d’Urfe, PQ, velvet samples; these were compared with serum Canada) and specific antibodies (Mediocorp Inc., levels. Second, samples from all sites (i.e., from the PQ, Canada). The interassay variabilities for the T base, middle, and tip of the bone and from the and E2 were less than 15%. Assay sensitivity was velvet skin) were used and values were compared less than 3 pg and intraassay and interassay among themselves and with the serum levels.
variabilities were 6.7 and less than 10%, respec- Associations between testosterone and estradiol tively. Recovery of exogenous testosterone added concentrations were estimated by fitting a random to the bone and velvet extracts were within 5% among the two tissues and the three antler version 9.0) as described by Tao et al. (2002) with fixed effects testosterone concentration and ‘Year’,and with the SUBJECT ‘Site’. With this randomcoefficient model predicted E calculated and plotted against testosterone con- Associations between testosterone or estradiol centrations with predicted regression lines for concentrations, age, and date of collecting the samples (month and year) were tested, usingmultivariate General Linear Mixed Model (GLLM)with testosterone or estradiol concentration as the dependent variable and the different seasonal or individual variables described above as indepen- was not significant (o0.05) in all models, so Age dent variables. To account for the repeated was not included in the final model. Adjusted measures on the same individuals across different mean values of T concentrations (7S.E.) in antler sites of the growing antler and serum, all analyses tissue sites differed from serum (Figure 1, DF¼2, were performed using mixed model analysis withindividual deer and an interaction of individualand year (individualnyear) as a random factor, significance of each fixed effect in the mixed GLLM was assessed by the F-test, on sequential dropping of the least significant effect, starting with a full model. Independent variables wereclasses (‘Site’ – serum, bone, velvet, or serum, bone base, bone middle, bone tip, velvet base, velvet middle, velvet tip – see below; ‘Month’– May, June, July, and August; and ‘Year’ – 2001 and 2002) and a continuous variable ‘Age’ ranging from 2 to 6 years. ‘Age,’ a fixed effect, surprisingly did not reach level of significance for bothtestosterone and estradiol concentrations in any of the models applied; these were thus droppedfrom the final model. In unbalanced designs with more than one effect, the arithmetic mean for a Comparison of testosterone (T) and 17b estradiol group may not accurately reflect response for that (E2) concentrations in serum, bone, and velvet (adjusted group, since it does not take other effects into means 7S.E.) of male white-tailed deer.
TESTOSTERONE & ESTRADIOL IN DEER SERUM AND ANTLERS 123, F¼7.08, Po0.001); month of collection affected T concentrations (DF¼3, 22.1, F¼5.61, Po0.01). Values of T in serum were higher than in bone (Po0.001) or velvet (Po0.01). However, T concentrations did not differ between bone and Estradiol concentrations differed among antler tissues and serum (Figure 1, DF¼2, 112, F¼24.55, F¼13.34, Po0.01), but not between months.
Estradiol concentrations in serum were consis-tently lower than those in antler bone (Po0.001) and velvet (Po0.001). Concentrations in bone were higher than in velvet (Po0.001).
Testosterone concentrations differed among the Monthly testosterone concentrations (adjusted antler tissue sites and serum (Figure 2 top, DF¼6, means 7S.E.) in serum of male white-tailed deer.
123, F¼3.50, Po0.01) and in different months (Figure 3, DF¼3, 22.1, F¼5.61, Po0.01). Concen- trations of T in serum were significantly higher (Po0.05) than in any antler tissue site, bone or velvet. Concentrations of T in different antler sites Estradiol concentrations differed among the antler tissue sites and serum (Figure 2 bottom, DF¼6, 111, F¼11.24, Po0.001) and between years (DF¼1, 20.8, F¼13.34, Po0.001). Estradiol con- centrations in serum were lower (Po0.01) than concentrations in antler tissue collection sites.
Among antler tissue collection sites, E2 concentra-tions in the bone tip were greatest, approaching statistical significance, when compared with bone base (Po0.058), velvet base (Po0.001), velvet middle (Po0.001), and velvet tip (Po0.01, but after Tukey-Kramer adjustment it was nonsigni- ficant). Also, the middle section of the antler bone had higher E2 concentrations than velvet base(Po0.05) and velvet middle (Po0.073).
Finally, the associations between T and E2 concentrations were estimated by fitting a random coefficient model. A likelihood ratio test (w2 o17.94, Po0.001) indicated that the randomcoefficient model fitted was better than the null model. Therefore, a plot was constructed showingpredicted regression lines for each site of the sample collection (Figure 4). The association was different in all three sites. While there was almost no association between T and E2 concentrations in velvet, associations in serum and bone differed Comparison of T (top) and E2 (bottom) concentra- remarkably. In serum there was a negative tions in serum, antler base, antler middle, and antler tip, velvet base, velvet middle, and velvet tip (adjusted means 7S.E.). Note that proportionally, there is more E centration. In bone the relationship between both and velvet than in blood. The opposite is true for T.
However, the authors did not specify the period during which the antlers were harvested. Theyobserved that the antler bone exhibited a stronger stimulatory effectin their testes bioassay than did the serum extract.
No androgenic effect was detected in the epididy- mal bioassay when the alcoholic extract of growing antler bone was tested. Finally, no estrogen-like effect was detected in the aqueous extract made from growing antler bone, using the Allen-Doisy reaction. Apparently, the bioassay technique em- ployed by Putchkov and coworkers was notsufficiently sensitive to detect androgens in the Predicted E2 concentrations plotted against T concentrations. Plot shows predicted regression lines for each alcoholic extract. Obviously, the aqueous medium is not suitable for extraction of androgens. There-fore, the stimulatory effect observed in thetesticular bioassay could have been due to the presence of growth factors, such as the epidermalgrowth factor (Ko et al., ’86) or the fibroblast Seasonal serum concentrations of T and E2 growth factor (Sunwoo et al., ’97), which were determined in this study were within the range detected recently in the growing antler bone.
previously reported for male white-tailed deer Since mineralization of antler bone occurs (Bubenik et al.,’79, ’82). Results appear to confirm during periods of rising serum levels of T, and the hypothesis that growing antler tissues are administration of T initiates calcificiation of utilizing serum androgens to produce estrogens.
growing antlers, it was assumed that this andro- Testosterone levels in serum of these deer were gen is directly involved in this process. However, almost 50% higher than corresponding levels in administration of T derivatives to castrated white- the velvet and in the antler bone. Conversely, tailed deer has a differential effect on their antler serum levels of E2 were 6–10 Â lower than corre- growth. Dihydrotestosterone (DHT) caused desic- sponding concentrations in antler bone. Interest- cation of the velvet, whereas 5b-androstanediol ingly, whereas the E2 :T ratio in the serum was and 5a-androsterone induced a variable degree of 1:10–60, the same ratio for the velvet was 1:3.5 osteon formation and antler mineralization (Mor- and for the antler bone 1:2–3. Increasing T ris and Bubenik, ’82). Another study indicated concentrations in the bone were mirrored by that T derivative, androstenedione, which is increasing E2 concentrations, suggesting that T produced in the testes and in the adrenal gland, is converted into E2. Thus it is likely that antler may be responsible for the stimulation of antler velvet and antler bone are not just passively growth in castrates (Bubenik et al., ’87). Conver- utilizing androgens but are active in steroid sion of T to various steroid derivatives is common and the metabolites, such as androstenedione, Circulating levels of androgens and estrogens estradiol, estrol (DHT), or androstanediol are then have been determined by RIA in several cervid utilized in various target tissues, such as ovaries, species, but steroid levels have not been deter- mammary gland, external and internal genitalia, mined in growing antler bone tissues, using skin, prostate gland, and brain (Martin, ’78; modern detection techniques. In the only other Kennedy et al., ’97). T then acts as a prohormone, study attempting to determine the levels of which is metabolized to its active derivatives androgens in the growing antlers, Putchkov and coworkers (’38), used bioassay and reported find- Estradiol may be important in the formation ings contradictory to this study. The authors and mineralization of antler bone. Plasma con- measured the stimulatory effect of the aqueous centrations of E2 in reindeer bulls during the rut and alcoholic extracts, made from the growing exceed by far the levels seen in reindeer cows antler bone and serum of maral deer (Cervus (Bubenik et al., ’97). The origin of that estrogen elaphus maral), on the size of seminiferous has not been elucidated yet. Several studies tubules and the development of germinal epithe- revealed that E2 is up to 50 times more effective lium of rat and mice testes (Putchkov et al., ’38).
in mineralization of growing antlers than T TESTOSTERONE & ESTRADIOL IN DEER SERUM AND ANTLERS (Blauel, ’35; Goss, ’68; Bubenik, ’90). Further- Bubenik GA, Brown GM, Bubenik AB, Grota LJ. 1974.
more, serial injection of 19–OH–testosterone, Immuno-histological localization of testosterone in growing which is converted to estradiol, will also cause antler of white-tailed deer. Calcif Tiss Res 14:121–130.
Bubenik GA, Brown GM, Bubenik AB, Wilson DA. 1975. The mineralization of antlers in castrates. When role of sex hormones in the growth of antler bone tissues. I.
conversion of 19–OH–T to E2 is blocked by Endocrine and metabolic effects of antiandrogen therapy.
administration of an aromatase inhibitor (1,4,6 androstatriene–3–17 dione), formation of an ivory Bubenik GA, Bubenik AB, Zamecnik J. 1979. The develop- bone mantle is prevented (Morris and Bubenik, ment of circadian rhythm of estradiol in plasma of white-tailed deer (Odocoileus virginianus). Comp Biochem Physiol ’82). Although the average concentrations of estrogens in plasma of male mammals are gen- Bubenik GA, Morris JM, Schams D. 1982. Photoperiodicity erally 10–50 Â lower than levels of androgens and circannual levels of LH, FSH, and testosterone in (Leymarie et al., ’74; Dohler and Wutke, ’75), normal and castrated male, white-tailed deer. Can J injected estrogens are 100–1000 Â more potent Bubenik GA, Sempere A, Hamr J. 1987. Developing antler, the than androgens in inducing male sexual behavior model for endocrine regulation of bone growth. 1. Concen- in rats (Gay and Dever, ’71) or deer (Fletcher and tration gradient of T3, T4, and alkaline phosphatase in Short, ’74). In addition, a reduced aggressive the antler-, jugular-, and saphenous veins. Calcif Tiss Int behavior was observed in estrogen receptor alpha knockout mice (Ogawa et al., 2000). Several Bubenik GA, Schams D, White RJ, Rowell J, Blake J, Bartos L.
1997. Seasonal levels of reproductive hormones and studies indicated the possible role of E2 in the their relationship to the antler cycle of male and female growth and mineralization of antlers. Blockage of reindeer (Rangifer tarandus). Comp Biochem Physiol 116B: estradiol receptors by two different antiestrogens resulted in the disruption of the formation of Dohler KD, Wutke W. 1975. Changes with age in levels of Haversian systems, a delay in velvet shedding and serum gonadotropins, prolactin, and gonadal steroids inprepubertal male and female rats. Endocrinology 97:898– a greatly reduced thickness of the compact (ivory) bone mantle (Bubenik and Bubenik, ’78; Bubenik, Fletcher TJ, Short RV. 1974. Restoration of libido in castrated ’90). Whether the availability of T substrate or the red deer stag (Cervus elaphus) with estradiol–17b. Nature aromatase activity in tissues is the rate-limiting Gay VL, Dever NW. 1971. Effects of testosterone propionate and estradiol benzoate – alone or in combination – on serum In conclusion, the current literature and the LH and FSH in orchidectomized rats. Endocrinology discovery of a different ratio of T to E2 in the serum, antler bone, and antler velvet supports the Goss RJ. 1968. Inhibition of growth and shedding of antlers by Kennedy RG, Davies T, Al-Azzawi F. 1997. Sexual interest in postmenopausal women is related to 5a-reductase activity.
histogenesis and mineralization of deer antlers.
Ko KM, Yip TT, Tsao SW, Kong YC, Fenessy P, Belew MC, Porath J. 1986. Epidermal growth factor from deer (Cervuselaphus) submaxillary gland and antler velvet. Gen Comp Bartos L, Schams D, Kierdorf U, Fischer K, Bubenik GA, Siler J, Losos L, Tomanek M, Lastovkova L. 2000. Cyproterone Kolle R, Kierdorf U, Fischer K. 1993. Effects of an antiandro- acetate reduced antler growth in surgically castrated fallow gen treatment on morphological characters and physiologi- cal functions of male fallow deer (Dama dama L.). J Exp Barrel GK, Lengoc CM, Muir FD. 1987. Receptors for estradiol in growing antlers of red deer stag. In: Bobek B, Perznowski Leymarie P, Roger M, Castanier M, Scholler R. 1974.
K, Regelin W, editors. Global Trends in Wildlife Manage- Circadian variation of testosterone in normal man. A study ment. Krakow-Warszawa: Swiat Press. p. 279–281.
by frequent sampling. J Steroid Biochem 5:167–172.
Blauel G. 1935. Beobachtungen uber die entstehung der Li C. 1987. Cytosol testosterone receptor in the antler tissue of perucke beim rehbock. Endocrinologie 15:321–329.
sika deer: An assay based on isoloelectric focusing in Bubenik GA. 1982. The endocrine regulation of the antler polyacrylamide gel. Chinese J Zool 22:23–25.
cycle. In: Brown RD, editor. Antler Development in Lincoln GA. 1975. An effect of the epididymis on the growth of Cervidae. Kingsville, Texas: Caesar Kleberg Wildlife Re- antlers of castrated red deer. J Reprod Fert 42:159–161.
Lincoln GA, Youngson RW, Short RW. 1970. The social and Bubenik GA. 1990. Neuroendocrine regulation of the antler sexual behavior of red deer stag. J Reprod Fertil Suppl cycle. In: Bubenik GA, Bubenik AB, editors. Horns, Pron- horns, and Antlers. Ch.19. New York: Springer Verlag.
Bubenik GA, Bubenik AB. 1978. The role of sex hormones in McMaster ME, Munkittrick KR, Van Der Kraak GJ. 1992.
the growth of antler bone tissue; Influence of antiestrogen Protocol for measurement circulating levels of gonadal sex therapy. Saugetierkundl Mitt 40:284–291.
steroids in fish. Can Technical Report of Fisheries and D during physiological bone growth: In vivo and in vitro Morris JM, Bubenik GA. 1982. The effects of androgens on the development of antler bone. In: Brown RD, editor. Antler Sunwoo JS, Nakano T, Sim JS. 1997. Effects of water soluble Development in Cervidae. Kingsville, TX: Caesar Kleberg extract from antler of wapiti (Cervus elaphus) on the growth of fibroblasts. Can J Anim Sci. 77:343–345.
Muir PD, Sykes AR, Barrell GK. 1988. Changes in blood Suttie JM, Lincoln GA, Kay RNB. 1984. Endocrine control content and histology during growth of antlers in red deer of antler growth in red deer stags. J Reprod Fertil 71: (Cervus elaphus) and their relationship to plasma testoster- Suttie JM, Li C, Sadighi M, Gunn J, Fleming S. 1998.
Ogawa S, Chester AE, Hewitt SC, Walker VR, Gustafsson JA, Physiological control of antler growth. In: Milne JA, editor.
Smithies O, Korach KS, Pfaff DW. 2000. Abolition of male Development in Deer Biology., Aberdeen, UK: Macalau sexual behavior in mice lacking estrogen receptor alpha Land Use Research Institute. p 189–196.
and beta (alpha beta ERKO). Proc Nat Acad Sci USA Tao J, Littel R, Patetta M, Truxillo C, Wolfinger R. 2002.
Mixed model analyses using the sas system course notes.
Putchkov NV, Aspissov PN, Gordenko AI. 1938. A study of the physiological effect of the horns (‘‘Pants’’) of the maral deer.
Tachezy R. 1956. Uber den einfluss der sexualhormone auf J Physiology of the USSR 24:1139–1144.
das geweihwachstum der cerviden. Saugetierkundl. Mitt Schams D, Barth D, Karg H. 1987. Is antler growth in adult roe deer (Capreolus capreolus) influenced by treatment with Van Der Kraak G, Chang J. 1990. Arachidonic acid stimulates estradiol 17–beta, an antiandrogen, or a prolactin inhibitor? steroidogenesis in goldfish preovulatory ovarian follicles.
In: Assenmacher I, Boissin J, editors. Endocrine regulation as adaptive mechanisms to the environment. Paris: Editions Wislocki GB. 1943. Studies on growth of deer antlers. In: Essays in Biology in Honor of Herbert M. Evans. Berkeley, Sempere AJ, Grimberg R, Silve C, Tau C, Garabedian M. 1989.
CA: University of California Press. p 631–653.
Evidence for extrarenal production of 1,25–dihydroxyvitam

Source: http://bartos.agrobiology.eu/publications/pdf/Bubenik_Miller_Lister_et_al_2005_JEZ.pdf