The results of aromatase activity assays were analyzed by split plot analysis of variance where tissue was the main plot and treatment with inhibitors formed the ...
BIOLOGY OF REPRODUCTION 54, 497-505 (1996)
Functional Aromatase Expression in Porcine Adrenal Gland and Testis' A.J. Conley, 2 ' 3' 4 CJ. Corbin, 3 '4 M.M. Hinshelwood, 5 Z. Liu,5 E.R. Simpson, 5 JJ. Ford,6 and N. Harada 7 North Dakota State University,4 Fargo, North Dakota 58105 University of Texas Southwestern Medical Center,5 Dallas, Texas 75235-9051 USDA-ARS Meat Animal Research Center,6 Clay Center, Nebraska 68933 Fujita Health University,7 Toyoake, Aichi 4 70-11, Japan ABSTRACT The expression of aromatase cytochrome P450 (P450a,om) in the adrenal glands, testes, and placentas of fetal and newborn pigs was investigated. Western immunoblot analysis detected a single 48-50-kDa protein band in these tissues as well as in other porcine tissues known to express P45 0a,om including Day 12 tubular conceptuses, theca interna, and granulosa. Slight differences in migration suggested that the P450a,om protein expressed in the testis was larger than that in the adrenal gland, which was, in turn, larger than that in placenta, theca, and granulosa. Consistent with P450a,om expression in these tissues, a cDNA encoding porcine P450a,om hybridized to a 2.3-kb transcript in Northern analyses of porcine blastocysts, placentas, and fetal and newborn adrenal glands and testes, as well as intheca and granulosa tissues from preovulatory follicles. No differences in transcript size were detectable among tissues. The identity of P450rO. transcripts was confirmed by sequence analysis of partial cDNA clones amplified from porcine fetal adrenal glands, testes, and placentas according to the RACE procedure. The sequences of the adrenal and testis clones were identical but differed from the placental sequence, which represented the first 85 amino acids of porcine P450a,om. Specifically, the adrenal and testis clones expressed transcripts that resembled the ovarian isoform of porcine P450o, rather than the porcine placental isoform, predicting atwo-amino acid deletion and 12 predicted amino acid substitutions. P450aom activity was examined to further define expression in these tissues. Activity in adrenal, testis, and placental homogenates was inhibited by 4-hydroxyandrostenedione (40H-A4) whereas inhibition by etomidate was demonstrated in the adrenal and testis homogenates but not inthe placental homogenates. The sensitivity of activity in the newborn porcine adrenal glands and testes to inhibition by etomidate was similar to that of ovarian P450,,, activity. The level of P450,,,,om activity was highest in the placenta and lowest in the adrenal gland, and no effect of fetal sex was noted in either tissue. Immunocytochemical studies localized the expression of P450a,,,om in the adrenal gland of newborns to cells at the corticomedullary junction and, with greater intensity, to cells around the developing medullary lobules. The same cells expressed cytochrome P450 17a-hydroxylase, the expression of which also extended throughout the zona fasciculata. The interstitial cells were the site of P450,,,om expression in the testis, but no expression could be detected in any cells within the spermatic tubules. These data demonstrate that fetal and newborn porcine adrenal glands and testes express an active P450,om that resembles the isoform expressed in the ovary. The localization of adrenal expression suggests a possible role in medullary maturation and function inthe fetal and newborn pig.
INTRODUCTION Aromatase cytochrome P450 (P 50,,arom) catalyses the syn-
the fetal and neonatal brain that are involved in sexual behavior [4, 5]. Differentiation of the human reproductive tract
thesis of estrogens from androgens. The gonads, principally the ovaries, are the primary sites of estrogen synthesis during periods of reproductive activity in most vertebrates [1]. Other extra-ovarian sites of significance are highly speciesdependent and may include adipose tissue, testis, and fetal liver [2]. Trophoblastic tissues of several species, including
ficiency is associated with pseudohermaphrodism in afflicted female infants due to the accumulation of androgens derived from fetal adrenal C19 steroids [6, 7]. Finally, aromatase activity in the gonadal primordia is thought to determine female differentiation in several reptilian species
4
is also dependent on P450arom, and congenital P450arom de-
[8, 9]. Clearly, P450arom is expressed in a broad range of tis-
the pig, horse, and rabbit, also express P450rom, and the
sues, and it is a critical component of normal development and function in mammals and other vertebrates. Less effort has been directed toward the tissue-specific
local production of estrogen is thought to be important in implantation [3]. Additionally, subsequent fetal development may rely on P4 50arom expression in these and other species. For instance, P4 50arom expression influences neural
aspects of P450rom expression in the pig than in other spe-
cies studied. The pig presents some unusual features with regard to the synthesis of steroids in the gonads and the synthesis of estrogen by conceptus tissues during gestation. Unlike the human, the rat, or even the ruminant ovary,
differentiation in rodents when expressed in the areas of Accepted September 29, 1995. Received August 2, 1995. 'Supported by NDSU Grant-in-Aid to AJ.C. and PHS grant HD13234 to E.R.S. Mention of a trade name does not constitute a guarantee or warranty of the product by the United States Department of Agriculture. 2Correspondence. FAX: (916) 752-4278. 3Current address: department of Population Health and Reproduction, School of Veterinary Medicine, University of California at Davis, Davis, CA 95616.
P4 50arom is expressed in the theca interna of porcine pre-
ovulatory follicles at levels that are comparable to, and correlated with, those expressed in the stratum granulosum [10]. In addition, 19-norandrogens, which are found in porcine follicular fluids [11], are less commonly found in other 497
498
CONLEY ET AL.
species, and their enzymatic origin and fate have yet to be explained. Porcine blastocysts also synthesize estrogen for a short period during trophoblastic elongation, which occurs on Day 11 or 12 of development [12]. We have demonstrated that the estrogen content of porcine blastocysts is correlated with the transient expression of P450aro, [13], which occurs in the embryonic trophectoderm and hypoblast cell layers [14] before differentiation and outgrowth of the mesoderm. The porcine blastocyst, together with the rabbit and equine blastocysts, which also synthesize estrogen, may represent the earliest stage of mammalian embryonic development in which P450,,,,, expression has been shown to occur. Maternal concentrations of estrogen rise throughout the second half of gestation in the pig presumably from placental synthesis [15]. Fetal, neonatal, and mature porcine testes are yet another significant source of estrogen in this species [16]. Because it synthesizes estrogen during development and in a broad range of tissues as an adult, the pig is a unique animal model in which to investigate the regulation of P450,,,,,,, expression. The study reported here was conducted to examine the expression of P4 50arom in the adrenal glands, testes, and placentas of fetal and newborn piglets. To the authors' knowledge, this is the first report to document the physiological expression of an active aromatase enzyme system in the adrenal gland of any species. MATERIALS AND METHODS Tissues All animals used in the study were from the university swine herd and either were experiencing normal estrus or were time-bred. All tissues were collected on ice after slaughter of animals at a federally inspected facility. Theca interna and granulosa tissues were dissected from preovulatory porcine follicles [101, and blastocyst, fetal, and placental tissues were collected on various days of pregnancy or at the time of delivery [13, 17]. Western Analysis Porcine tissues were homogenized in PBS containing 1% cholate and 0.1% SDS. Protein concentrations were estimated by use of BCA Protein Assay Reagent (Pierce Chemical Co., Rockford, IL) and BSA as a standard. Homogenates (50 jlg protein/lane) were electrophoresed on an 8% polyacrylamide-SDS (Amresco, Solon, OH) gel in buffer containing 50 mM Tris, 383 mM glycine, 10% SDS, and 0.4 mM EDTA. Separated proteins were electroblotted onto a nitrocellulose membrane and immunoblotted with antisera (1:100 dilution) raised against recombinant human P450,.,,, protein [18]. Immunoblotting was carried out at room temperature with use of a chemiluminescent detection system according to the manufacturer's instructions (ECL, Amersham, Arlington Heights, IL).
Northern Analysis Total RNA was isolated according to the method of Chirgwin et al. [19]. Briefly, tissue was homogenized in 4 M guanidinium thiocyanate, layered over a 5.7 M CsCl2 cushion, and centrifuged for 3 h in a Beckman TL-100 tabletop ultracentrifuge (Beckman Instruments, Palo Alto, CA), equipped with a TL100.3 fixed-angle rotor, at 80 000 rpm. Poly(A) + RNA was prepared by using oligo(dT) cellulose spin columns (5Prime-3Prime Inc., Boulder, CO). The resulting RNA (10 pig per lane) and RNA ladder (Gibco BRL, Gaithersburg, MD) were electrophoresed on a 1.1% formaldehyde agarose gel in 3-(N-morpholino)propanesulfonic acid buffer and electroblotted onto nylon membrane (Magna; MSI, Westboro, MA) in 0.5 M Tris. Loading was verified by visualization of ribosomal RNA bands after ethidium bromide staining. Hybridization was carried out overnight at 42°C according to MSI specifications with either a full-length porcine P450,,,, cDNA [20] or a 1.5-kb cDNA corresponding to the 3' end of bovine P450ar,,,, [21] labeled with [32 P]dATP (Multiprime; Amersham, Arlington Heights, IL). Hybridizing species were detected by autoradiography at -70°C with intensifying screens. RACE Procedure Partial cDNA sequence encoding porcine P450 ,,1,mwas obtained by the RACE procedure [22]. A single-stranded cDNA was synthesized with M-MLV reverse transcriptase (Gibco BRL), with 1.0 pg RNA used as a template and with priming from a degenerate oligonucleotide based on human, bovine, and rat P4 50aro,,, corresponding to base pairs 443-465 of bovine P450,,,(, [21]. A poly(A) tail was then added to the 3' end of the cDNA with terminal deoxynucleotidyl transferase (Gibco BRL), and polymerase chain reaction was used subsequently to amplify a region between the poly(A) tail and a nested degenerate oligonucleotide corresponding to base pairs 396-416 of bovine P450,,,, [21]. The amplified products were subcloned into pCRII vector (Invitrogen Corp., San Diego, CA) and sequenced according to the Sequenase double-stranded chain termination method (.S. Biochemicals, Cleveland, OH). Aromatase Activity Analysis Tissues were homogenized on ice in 50 mM P0 4, 1 mM EDTA buffer, and protein concentration was estimated by using Bradford protein reagent (Pierce Chemical Co.) with BSA as a standard. Activity was assayed by measuring the incorporation of label into tritiated water from [1J33 H]androstenedione (New England Nuclear, Wilmington, DE) as described [20, 23]. Tissue homogenates (100 jg protein) were incubated in duplicate with 150 nM [3 H]androstenedione in the presence of an NADPH-generating system containing 1 mM NADPH, 2 mM NADP, 17 mM glucose-6-phosphate, and 1 unit of glucose-6-phosphate
PORCINE ADRENAL AROMATASE
499
dehydrogenase. The incubations were carried out in a 37°C water bath in the presence or absence of 1 giM 4-hydroxyandrostenedione (40H-A4; a P450,,,a inhibitor) and 1 IM etomidate (a general inhibitor of mitochondrial steroidogenic cytochromes P450). Proteins were precipitated with trichloroacetic acid (10% final concentration), and the steroid was extracted with chloroform. Charcoal/dextran was used to purify 3H-water, and an aliquot was quantified by scintillation counting. Controls were included to verify that no tritiated water was released in the absence of protein or a generating system. Verification that tritiated water release was associated with estrogen production in this assay system was obtained by thin layer chromatography [20]. Immunohistochemistry Immunohistochemistry was performed as described previously [24], using a specific rabbit polyclonal antibody raised against purified recombinant human P450,,,. Pread-
sorbed antisera was not available to verify specificity of staining; however, this antisera recognizes a single protein expressed in Cosl cells transfected with cDNAs encoding
FIG. 1. Western immunoblot of porcine tissues (50 g) using antibody raised against recombinant human P450,,rom. The immunoblot shown was left underexposed to emphasize differences in electrophoretic migration of immunodetectable P450,aro in these tissues. Data were replicated in immunoblots performed on tissues collected from several different pregnancies and are consistent with previous immunoblot analyses with respect to expression in Day 12 conceptuses [131 and theca interna and granulosa components of preovulatory ovaries [101. B, tubular blastocyst of Day 12 conceptus; F,filamentous blastocyst of Day 12 conceptus; P, placenta; A, fetal adrenal; T,fetal testis; G, granulosa; I, theca interna.
both placental and ovarian isoforms of porcine P450arom but
not expressed in mock-transfected cells [20]. Staining was also conducted for cytochrome P450 17a-hydroxylase (P450c,7 ), kindly provided by Dr. Anita Payne, University of Michigan, Ann Arbor, to provide additional data relevant to the specificity of P450arom localization. In brief, after de-paraffinization, sections were preincubated in blocking buffer (Biomeda, Foster City, CA) for 5 min. Sections then were incubated for 1.5 h at 22°C with the primary antibody, which was diluted 1:100 (P450,,..) or 1:1000 (P450c 17) in PBS (50
mM phosphate, 150 mM NaCl, pH 7.6). Control sections were incubated with normal rabbit serum (1:250 in PBS; Vector Labs., Burlingame, CA) in place of primary antibody. After rinses with two changes of Tris buffer (25 mM Tris HC1, 150 mM NaCI, 10 mM Triton X, pH 7.6; 5 min/rinse), the primary antibody was detected by using a biotinylated secondary antibody (goat anti-rabbit IgG) and avidin-biotinperoxidase complex (both from Biomeda) with 3,3'-diaminbenzidine (Vector Labs.) as the substrate. Multiple sections of each tissue were examined from each animal. StatisticalAnalysis The results of aromatase activity assays were analyzed by split plot analysis of variance where tissue was the main plot and treatment with inhibitors formed the sub-plot. RESULTS 4 50
Cytochrome P
arom expression was examined by West-
ern immunoblot analysis in several tissues, and immunodetectable bands were observed in adrenal glands, testes, and placentas from fetal and newborn piglets. An intense band was also observed with tubular blastocysts but not
FIG. 2. Northern analysis of porcine tissues probed for P450rom. a) Total RNA (10 pg) from placenta (PI), fetal adrenal, and porcine fetal testis (T)probed with bovine cDNA. Numbers correspond to approximate days of gestation. b) Poly(A) + (5 g) from testis (T), granulosa (G), placenta (P), adult adrenal (A), theca interna (I),and blastocyst (B,Day 12 tubular) probed with porcine cDNA.
500
CONLEY ET AL. i
10
20
30
40
50
Ad/T MVLEMLNPM- -NISSMVSEAVLFGSIAILLLIGLLLWVWNYEDTSSIPGP Plac
.........
YYK.T....
60
V.P.A...V...T.F..LL .... N.....
70
S.
80
Ad/T GYFLGIGPLISHFRFLWMGIGSACNYYNIYGEFM Plac
...................................
FIG. 3. Predicted amino acid sequence comparison of partial cDNAs encoding porcine P450,rom obtained from porcine fetal adrenals and testes (Ad/T) and placenta (Plac) by RACE procedure. Amino acids are numbered sequentially from initiation methionine on the basis of placental sequence, and differences between adrenal/ testis and placenta clones are indicated relative to adrenal/testis (upper) sequence. Putative transmembrane spanning region 1311, which extends from residue A-20 to Y-4 of adrenal P450a,,,, sequence, is shown; potential N-linked glycosylation sites are underlined.
with the slightly more advanced filamentous blastocysts collected from mated gilts on Day 12 of pregnancy. Similarly, intense bands were detected in theca interna and granulosa from preovulatory follicles (Fig. 1). No immunodetectable protein was observed in kidney or liver (data not shown). Although the apparent size of the detected protein was 4850 kDa in all cases, migration on the gel differed slightly but perceptibly among the different tissues. The protein detected in the testis appeared to be slightly larger than that in the adrenal gland, which was in turn slightly larger than that expressed in the placenta, theca and granulosa tissues. The band detected in the tubular blastocysts appeared to be smaller than that observed in all other tissues. The finding of the expression of P4 50rom in newborn adrenal gland by Western immunoblot analysis was unexpected; therefore, confirmation of these results by Northern blot was sought. A cDNA encoding P450,,O,n hybridized to transcripts demonstrating a single band of approximately 2.3 kb in the adrenal gland at early and late stages of fetal development. Fetal testis and placenta hybridized to a band of similar size (Fig. 2a). Additional porcine conceptus and adult tissues were examined for evidence of P4 50rOm expression. As expected, P4 50ar,,,, expression was detected in Day 12 tubular blastocysts. Several adult tissues including theca interna, granulosa, and testis also demonstrated expression of a 2.3-kb transcript, but no hybridization was observed in RNA isolated from adult adrenal gland (Fig. 2b). Unlike the results of Western analyses, no differences were detectable in the relative sizes of hybridizing bands among the tissues expressing P4 501,..... TABLE 1. Basal and inhibited aromatase activities (pmol/mg/h) in tissues from newborn pigs. a Tissue Placenta Adrenal Testis Liver Kidney
Basal
40H-A4
Etomidate
581 + 93 41 + 17 83 + 24
