0013-7227/91/1281-0349$02.00/0 Endocrinology Copyright© 1991 by The Endocrine Society

Vol. 128, No. 1 Printed in U.S.A.

Androgen Receptor Messenger Ribonucleic Acid (mRNA) in the Rat Liver: Changes in mRNA Levels during Maturation, Aging, and Calorie Restriction* C. S. SONG, T. R. RAO, W. F. DEMYAN, M. A. MANCINI, B. CHATTERJEE, AND A. K. ROY Departments of Cellular and Structural Biology (C.S.S., W.F.D., M.A.M., B.C., A.K.R.) and Obstetrics and Gynecology (T.R.R., W.F.D., M.A.M., A.K.R.), University of Texas Health Science Center, San Antonio, Texas 78284

the dietary calorie intake. Analysis of poly (A)-containing RNA from two liver cell populations, hepatocytes and nonhepatocytes, revealed that only the hepatocytes that express a2u-globulin gene contain AR mRNA. From these results and our earlier observation of in vitro induction of a2u-globulin in isolated rat liver, we conclude 1) that androgen can act directly on hepatocytes to promote a2u-globulin synthesis; 2) that changes in the hepatic androgen sensitivity during maturation and aging are reflections of the age-dependent expression of the receptor gene; and 3) that retardation of the age-dependent loss of androgen sensitivity by calorie restriction is due to a concomitant delay in the decline of the hepatic AR mRNA level. {Endocrinology 128: 349-356, 1991)

ABSTRACT. By means of RNAase protection assay with an antisense cRNA probe, we have shown that the liver of the young adult male rat contains androgen receptor (AR) mRNA to a level of 4% compared to the prostate. Steady state levels of AR mRNA in the liver show both sex and age specificity. Compared to that of the male, the female liver contains a markedly reduced amount of AR mRNA. AR mRNA is almost undetectable in livers of prepubertal male (750 days old) rats. Both prepubertal and senescent animals are relatively insensitive to the androgenic induction of a2u-globulin, a hepatic secretory protein. The agedependent decline in hepatic androgen sensitivity and AR mRNA level can be delayed considerably by a 40% reduction in

T

affinity labeling with [3H]R-1881, followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, showed this cytoplasmic androgen-binding (CAB) protein to be only 31 kDa, which makes it about one third the size of any steroid hormone receptor (10). Thus, it is likely that CAB is not structurally related to AR. The presence of CAB may have been the complicating factor in the unequivocal demonstration of AR activity in the rat liver (11). The subject of androgen action in the rat liver has been further clouded by the hypothesis that androgens act indirectly on the liver by regulating the pattern of GH secretion by the pituitary gland (12, 13). This hypothesis has gained considerable support from the observation that less pulsatile (feminine)-type GH secretion results in the decreased hepatic synthesis of cytochrome P-450h, carbonic anhydrase III and a2u-globulin (14-16). The transgenic male mouse containing metallothionein-GH fusion gene has a high circulating level of GH with low episodic pulses (i.e. similar to normal females). Based on the observation of decreased hepatic synthesis of the male specific urinary protein in the GH transgenic mouse, it has been suggested that in the liver GH may function as a sex hormone (17).

HE MAMMALIAN liver shows a considerable degree of sexual dimorphism (1). In the rat, androgenic induction of a number of hepatic enzymes and other proteins, such as carbonic anhydrase III, cytochrome P450h, and a2u-globulin, and repression of senescence marker protein-2 have been well documented (2-7). Despite these examples of androgenic influence on hepatic gene expression, the existence of the androgen receptor (AR) in the rat liver has not been clearly established. By means of sucrose density gradient analysis, we provided the initial evidence for the presence of an androgenbinding protein in the rat liver and considered it to be the AR (8, 9). Although this binding protein has the same specificity for steroidal hormones as that of the AR and binds 5a-dihydrotestosterone with a Ka of 10~8 M, it fails to translocate into the nucleus. In addition, photoReceived June 27,1990. Address all correspondence and requests for reprints to: A. K. Roy, Ph.D., Division of Molecular Genetics, Department of Obstetrics and Gynecology, University of Texas Health Science Center, San Antonio, Texas 78284-7836. * This work was supported by NIH Grants DK-14744 and AG-03527 and a grant from the Morrison Trust. Some of the results contained in this manuscript were presented at the 72nd Meeting of The Endocrine Society.

349

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 25 July 2016. at 08:32 For personal use only. No other uses without permission. . All rights reserved.

HEPATIC AR mRNA

350

Earlier studies from our laboratory have shown that the rat liver attains androgen sensitivity only after puberty (~40 days), when it can respond to the androgenic induction of «2u-globulin (18). The androgen sensitivity of the liver declines during advanced age, and beyond 750 days very little hepatic androgen responsiveness can be observed, as evidenced by a2u-globulin gene expression (5). We have also shown that in the in vitro liver perfusion system, androgen infusion into the liver of the young adult rat rapidly increases the synthesis of «2u-globulin (19). This observation indicates a direct effect of androgen on the liver, rather than through the hypothalamohypophysial system. Two groups of investigators have cloned the cDNA corresponding to the mammalian AR, and the cloned probes are available for exploring the presence of AR mRNAs in various target tissues (20, 21). We have used the rat cDNA clone to examine the presence of the AR mRNA in the liver. Our results show that the AR gene is expressed in the rat hepatocyte, and the steady state level of AR mRNA correlates with the three phases of hepatic androgen sensitivity, i.e. androgen insensitivity during prepuberty, androgen responsiveness during young adulthood, and the gradual loss of androgen sensitivity during aging. Materials and Methods Animals and treatment Fischer 344 rats were used for this study. Rats were individually housed in a barrier facility with a 12-h light, 12-h dark cycle. For aging studies, animals were killed at appropriate ages, the livers were quickly removed and frozen in liquid nitrogen, and the frozen tissues were stored at -80 C until further processing. Rats of 25-35 days of age were considered prepubertal, those 80-120 days of age as young adult, and those 750-800 days of age as senescent animals. Various tissues from senescent animals were routinely examined histopathologically for tumors and other lesions. Only livers from healthy animals were used for this study. The University of Texas Health Science Center at San Antonio has an AAALAC-certified animal care facility, and all animal care and treatment protocols were approved by the Institutional Animal Care Committee. For nutritional studies animals were housed individually. Calorie restriction was initiated at 6 weeks of age, when animals were divided into two groups, one of which had free access to food; the other was fed 60% of the food consumed by the ad libitum fed animals. The diet used for this study contained the following calorie composition: 57% carbohydrate, 22% fat, and 21% casein (22). Animals were killed at the appropriate ages, and livers were quick frozen in liquid nitrogen for further processing. Preparation of the radiolabeled antisense cRNA probe to rat AR mRNA The rat AR cDNA clone (20) was obtained from Dr. S. Liao of the University of Chicago. The 5' hypervariable end of the

Endo • 1991 Voll28«Nol

rat AR cDNA was subcloned in the plasmid pGEM-3Z (Promega, Madison, WI), and the cDNA sequence was subsequently transcribed in vitro to prepare the cRNA probe for RNAase protection studies. The 444-nucleotide long EcoRl-Pvull fragment of the original 2.8-kilobase rat AR cDNA was obtained by appropriate restriction endonuclease digestion, agarose gel electrophoresis, electroelution, and ethanol precipitation. This DNA fragment was then subcloned in pGEM-3Z at the Smal site. The orientation of this fragment is such that in vitro transcription in the presence of SP6 polymerase generates antisense receptor RNA transcripts. The plasmid was grown in DH5a, and the purified plasmid DNA was linearized by EcoRI digestion. In vitro transcription in the presence of SP6 polymerase, [a-32P]CTP (800 Ci/mmol), and the other three unlabeled ribonucleotides was carried out under conditions previously described (23). After DNAase digestion, the RNA transcripts were purified from the reaction mixture by phenol extraction and repeated ethanol precipitation. Isolation of liver cells, extraction of RNA, and Northern blot analysis Hepatocyte and nonhepatocyte cell populations were isolated from young adult male rats by collagenase perfusion, followed by Percoll gradient centrifugation (24). Hepatocytes were recovered as a pellet by centrifugation (50 X g; 4 min at 4 C). The supernatant fraction was recentrifuged for another 15 min to remove any remaining hepatocytes, which were then discarded. The supernatant fraction from the second centrifugation step was further centrifuged at 300 x g for 10 min. Aliquots of hepatocyte and nonhepatocyte cell suspensions were plated in Dulbecco's Modified Eagle's Medium (Sigma, St. Louis, MO) containing 10% fetal bovine serum. Hepatocytes were identified by their characteristic morphology, and the purity of separated cell fractions was quantified by direct cell counts from photographs of the culture plates. Based on the counts of a representative field, the hepatocyte population was 98% pure (291 hepatocytes within 297 total cells), and nonhepatocytes were 99% free of hepatocytes (5 hepatocytes within 500 total cells). The nonhepatocytic cells constituted primarily endothelial and macrophage-like cells. Immediately after their isolation, aliquots of hepatocyte and nonhepatocyte cell pellets were extracted for total RNA using Chomczynski and Sacchi's single step guanidinium thiocyanate-phenol method (25). Total RNA from the tissue samples (liver, seminal vesicle, prostate, and female intestine) was extracted by the phenol-sodium dodecyl sulfate procedure (26). Total poly(A)+ RNAs were selected through 2 cycles of oligo(dT)-cellulose chromatography (27). Poly (A)-enriched RNA samples (5 ng) from total liver, isolated hepatocytes, and nonhepatocytes were used for Northern blot analyses. RNA samples were fractionated on a 1.5% agarose gel containing 2.2 M formaldehyde (28). RNA blotting and fixing on Nytran (Schleicher and Schuell, Keene, NH) filters were performed under standard conditions. Prehybridization, hybridization with 32P-labeled cDNA probes, and posthybridization washing of filters were carried out as described previously (7). Cloned cDNA inserts for a2u-globulin (29) and cytochrome oxidase were radiolabeled in the presence of [a-32P] dCTP by random priming (30).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 25 July 2016. at 08:32 For personal use only. No other uses without permission. . All rights reserved.

HEPATIC AR mRNA RNAase protection of the AR mRNA RNAase protection assays were performed as described by Zinn et al (31). Ribosomal RNA, transfer RNA (tRNA), and DNA contents of total RNA preparations from various tissues are known to vary considerably. To obtain a more precise comparative index we have used poly(A)+ RNA for the RNAase protection assay. The poly(A)+ RNAs derived from appropriate tissues and cells were mixed with 5 X 105 cpm freshly synthesized antisense cRNA transcripts in a 30-^1 hybridization mixture containing 80% deionized formamide, 40 mM piperazineiV,AT'-bis(2-ethane sulfonic acid) (pH 6.4), 0.4 M NaCl, and 1 mM EDTA. The hybridization mixture containing the mRNAs and the cRNA probe was heated at 85 C for 5 min and immediately brought to 45 C. Annealing was carried out at 45 C for 8 h. Subsequently, the RNA-RNA hybrid was digested with RNAase-A (40 Mg/ml) and RNAase-T! (2 /xg/ml) at 30 C for 60 min. The protected RNA hybrids were recovered from the reaction mixture after proteinase-K digestion, phenol-chloroform extraction, and ethanol precipitation. The recovered RNA-RNA hybrids were then dissolved in a denaturing solution [80% (vol/vol) deionized formamide, 1 mM EDTA (pH 8.0), 0.5% bromophenol blue, and 0.1% xylene cyanol], heated at 85 C for 15 min, quickly cooled on ice, and electrophoresed on a 5% sequencing gel. After electrophoresis, the gel was dried and autoradiographed on a Kodak AR-5 x-ray film (Eastman Kodak, Rochester, NY). Each experiment was repeated with RNA preparations from at least three different animals. In addition to the specifically protected band of 444 bases, with certain batches of the labeled antisense probe a major band with approximately the same electrophoretic mobility as that of the undigested transcript was seen in both experimental and control lanes. The intensities of the upper bands appear to be same regardless of the type of unlabeled RNA sample used for the protection experiment. In vitro transcription of the linearized vector with SP6 polymerase may also produce a trace amount of the labeled sense transcript, which can hybridize with the antisense probe. Several minor bands of lower size ranges were also observed both in tRNA control and mRNA sample lanes. When assaying for rare mRNAs, other investigators have observed an intense nonspecific upper band (same size as that of the undigested transcript) and several less intense bands of lower size ranges (32).

351

amplification program involved denaturation of the reaction mixture at 95 C for 3 min, primer annealing at 60 C for 2 min, and extension at 72 C for 3 min for the first cycle. In all subsequent cycles denaturation was performed at 94 C for 1 min. Ten-microliter aliquots of each PCR mixture (after 35 cycles of amplification) were electrophoresed on a 1.5% agarose gel, stained with ethidium bromide, and photographed under UV illumination. After photographic recording, the gel was subjected to Southern transfer, and the filter was hybridized with AR cDNA probe labeled with 32P by random priming (30).

Results The rat AR (RAR) DNA construct and the assay procedure for RNAase protection analysis are schematically presented in Fig. 1. The RAR construct shown in Fig. IB, when linearized with EcoRl and transcribed with SP6 polymerase, is expected to produce a 509-base cRNA transcript (444 bases of RAR plus 65 bases of the vector sequence). Considering a 5-10% slower mobility of the RNA compared to the DNA size markers (35), the size of the transcribed band matches that of the expected .—EcoRl

RAR 444 bp

B

EcoRl

Smal

LINEARIZED VECTOR

SP6 PROMOTER

cRNA PROBE

ARmRNA

Detection of AR mRNA by polymerase chain reaction (PCR) Two oligonucleotide primers, spanning bases 771-795 (5'GTG GAA GCA CTG GAA CAT CTG AGT C3') of the sense, strand and 1625-1649 (5'CTG GGG TGG GAA GTA ATA GTC GAT G3') of the antisense strand of the rat AR cDNA (33), were used for the amplification reaction. Two micrograms of poly(A)+ mRNA from each sample (male adult liver and prostate) were reverse transcribed with Moloney murine leukemia virus reverse transcriptase (BRL, Gaithersburg, MD) as previously described (34). In the control reaction, reverse transcriptase was heat denatured (95 C for 4 min) before the start of incubation. After first strand cDNA synthesis, nucleic acids were extracted with phenol-chloroform (1:1) and ethanol precipitated. The pellet was dissolved in 100 n\ H2O, and a 10-fi\ aliquot was used for PCR amplification. The

444 bp RNAase PROTECTED FRAGMENT

FIG. 1. Schematic drawings of the AR cDNA construct and RNAase protection analysis. A, Map of the full-length rat AR cDNA clone in the expression vector pGEM-3Z. B, Map of the truncated construct containing a 444-basepair fragment from the 5' end of the AR cDNA spanning between EcoRl and Pvull sites. This construct, when linearized with EcoRl and transcribed with SP6 polymerase, is expected to produce a 509-base long antisense transcript (444 RAR plus 65 vector sequence). C, The extent of the SP6 transcription and the size of the anticipated RNAase protected AR fragment are shown.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 25 July 2016. at 08:32 For personal use only. No other uses without permission. . All rights reserved.

HEPATIC AR mRNA

352

product (Fig. 2). In contrast to the intestine and tRNA control lanes (lanes 5 and 6), both prostate and adult male liver RNAs (lanes 3 and 4) show specifically protected bands of the size expected for that of the cloned RAR fragment (i.e. 444 bases). Densitometric analysis of the specifically protected band indicates that, compared to the prostate (100%), young adult male liver contains 4% AR mRNA. The presence of AR mRNA in the liver of young adult male rats was substantiated independently by qualitative PCR. Reverse transcripts of the poly(A)+ RNAs from different tissues were used as templates for PCR. The primers chosen for PCR are expected to yield a 878-base pair RAR fragment. With both prostate and young adult male liver, a single amplified product of the expected size (Fig. 3A, lanes 2 and 4) was evident. Hybridization of the amplified product with labeled AR cDNA probe confirmed the authenticity of the PCR product (Fig. 3B). Thus, both the RNAase protection assay and PCR independently establish the presence of AR mRNA in the liver of the young adult male rat. Earlier studies in our laboratory have shown that prepubertal and senescent male rats are insensitive to androgenic induction of a2u-globulin. It is, therefore, of great interest to examine the possibility of an age-dependent program in expression of the AR gene in the rat liver. The autoradiogram in Fig. 4 shows results of protection analysis of liver RNAs derived from animals of three different age groups (prepubertal, 30 days old;

910 —

235 — FIG. 2. RNAase protection analysis of AR mRNA in RNA preparations from prostate and young adult male liver. Lanes are: 1, size marker (Ddel digestion products of pUCl9 DNA); 2, undigested cRNA probe (1 103 cpm); 3, prostate RNA (4 fig); 4, young adult male liver RNA (20 fig); 5, intestine RNA (20 ng); and 6, tRNA (20 iig). The specifically protected AR bands in the prostate and liver samples are shown with arrowheads.

Endo • 1991 Voll28-Nol

1 2

3 4 5

1 2

3

4

752596-

FlG. 3. Identification of the AR mRNA with PCR. Total poly(A)+containing RNA samples extracted from prostate and young adult male rat liver were reverse transcribed to cDNA and amplified with Taq polymerase, as described in Materials and Methods. The lefthand frame (A) displays the ethidium bromide-stained gel. The lanes are: 1, size markers (a 3.2-kilobase a2u-globulin genomic clone in pSV2 CAT plasmid digested with EcoRl, PstI, and Sail); 2, amplification products of the reverse transcribed prostate RNA; 3, PCR products of the prostate RNA in the presence of heat-denatured reverse transcriptase; 4, amplification products of reverse transcribed young adult male liver RNA; and 5, same as lane 4, but with heat-denatured reverse transcriptase. The righthand frame (B) shows the autoradiographic picture of the Southern blot of the gel shown in frame A and hybridized with AR cDNA probe. RNA samples used as initial templates in various lanes are: 1, prostate; 2, prostate with heat-denatured reverse transcriptase; 3, young adult male liver; 4 young adult male liver with heat-denatured reverse transcriptase.

young adult, 110 days old; senescent, 790 days old) compared to positive controls (prostate and seminal vesicle) and a negative control (tRNA). In contrast to the young adult male, the liver RNA from the prepubertal male contains virtually undetectable levels of AR mRNAs, and the level of this mRNA is also greatly reduced in the liver of the senescent male. These results are in concordance with those derived from in vivo studies which show that the inductive response is limited to the liver of young adult animals. Estrogens are known to inhibit the androgenic induction of «2u-globulin, and androgen treatment of nonovariectomized female rats fails to induce this protein (36) (Roy, A. K., unpublished results). It is, therefore, likely that in the normal female estrogens inhibit hepatic expression of the AR gene. This prompted us to compare AR mRNA levels in livers of young adult male and female rats. As seen in Fig. 5, the female liver contains markedly reduced (almost undetectable) levels of AR mRNA. Within the liver lobule, the expression of the a2uglobulin gene is markedly influenced by the location of

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 25 July 2016. at 08:32 For personal use only. No other uses without permission. . All rights reserved.

HEPATIC AR mRNA

1 2

3

4

5 6

353

the hepatocyte. The androgenic induction of a2u-globulin initially begins in hepatocytes that are attached to the wall of the central vein (perivenous hepatocytes) (37). This finding raises the issue of whether androgens initially act on the endothelial or other nonhepatocytic cells of the liver. Such an effect, in turn, can produce an indirect paracrine mediator responsible for the induction of a2u-globulin. We have addressed this question by separating hepatocytes from nonhepatocytic cells of the liver and assaying the RNA samples from these two cell populations for a2u-globulin and AR mRNAs. Results of these experiments are presented in Fig. 6. By means of collagenase perfusion and Percoll density gradient it is possible to obtain a more than 96% pure hepatocyte cell

FIG. 4. RNAase protection assay of the AR mRNA in the livers of animals of different ages. Lanes 1 and 2, Positive controls (prostate and seminal vesicles, respectively); lanes 3, 4, and 5, poly(A)+ hepatic RNA samples from prepubertal (30 days), young adult (110 days), and senescent (790 days) male rats, respectively; lane 6, negative control (20 ng tRNA). Prostate and seminal vesicle reaction mixtures contained 10 ng poly(A)+ RNA, whereas all of the liver samples contained 20 ng poly(A)+ RNA. Specifically protected AR mRNA bands are marked with arrowheads.

1 2

3 4

FlG. 5. Gender-specific expression of the AR mRNA in the rat liver. The photograph shows RNAase protection analysis of poly(A)+ RNA samples from livers of young adult rats. Lanes are: 1, prostate (4 ng RNA); 2, young adult male (20 ng RNA); 3, young adult female (20 fig RNA); and 4, tRNA (20 ^g) control. The specifically protected AR band is indicated with an arrowhead.

FIG. 6. Specific localization of the AR mRNA in the hepatocytes of the young adult male rat. Frame A shows the light microscopic picture of the isolated nonhepatocyte (1) and hepatocyte (2) cell populations. Frame B shows the RNAase protection analysis for the AR mRNA within the total RNA preparations (10 /xg polyA+ RNAs) derived from nonhepatocyte (1) and hepatocyte (2) cell populations. The specifically protected AR mRNA band in lane 2 is marked with an arrowhead. Frame C shows the Northern blot analysis of the same RNA preparation as that in B hybridized with the a2u-globulin probe (upper panel) and cytochrome oxidase probe (bwer panel). Lane 1, Nonhepatocyte RNA; lane 2, hepatocyte RNA. a2u-Globulin mRNA can be detected only in the hepatocyte RNAs that also contain the AR mRNA.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 25 July 2016. at 08:32 For personal use only. No other uses without permission. . All rights reserved.

HEPATIC AR mRNA

354

population (Fig. 6A). Analysis of AR mRNA through the RNAase protection assay shows that this mRNA is selectively localized in the hepatocytes (Fig. 6B). Compared to AR mRNAs within total hepatic mRNAs of young adult male rat liver (presented in Figs. 2, 4, and 5), the RNA samples from isolated hepatocytes show a slight enrichment for AR mRNA. Such a minor enrichment is consistent with the fact that hepatocytes constitute 65% of the total liver cells (38). Northern blot analysis of the a2u-globulin mRNA within the RNAs from these two cell populations (Fig. 6C) provides another indication of the purity of the hepatocyte and nonhepatocyte cell populations. We have shown that a 40% reduction in the daily calorie intake, initiated after puberty, can markedly delay the age-dependent loss of hepatic androgen sensitivity (22). Whether such a prolongation of androgen sensitivity is due to continued expression of the AR gene is addressed in Fig. 7. Samples of hepatic RNA from animals at different ages with and without calorie restriction were assayed for the AR mRNA. In animals that were fed ad libitum, AR mRNA was reduced to an almost undetectable level beyond 18 months of age. However, in the calorie-restricted rats a high level of AR mRNA was found even in 27-month-old animals. These results, therefore, clearly show that the prolongation of hepatic androgen sensitivity in senescent rats by calorie restriction is due to the continued expression of AR mRNA. Discussion Although sexual dimorphism of the rat liver is uncontroversial, the mechanism of androgen action on this organ has remained unclear. This has partly been due to

1 2

3 4 5 6

FIG. 7. Retardation of the age-dependent decline in AR mRNA by calorie restriction. The same amounts of polyA+ hepatic RNA (20 ng) were subjected to RNAase protection analysis. Even-numbered lanes are from animals fed a calorie-restricted diet. Lanes are: 1 and 2, 12 months old; 3 and 4, 18 months old; 5 and 6, 20 months old; and 7 and 8, 27 months old. Lanes 7 and 8 are from a different run and, therefore, are not exactly of the same exposure conditions. The specifically protected AR mRNA band is identified with the arrowhead.

Endo • 1991 Vol 128 • No 1

the fact that many of the androgenic effects on the liver are synergistically influenced by GH (5), and the hormone-binding assay of the liver extract has failed to provide an unequivocal demonstration of AR activity. In fact, the search for the hypothetical pituitary feminizing factor has led to GH being the possible mediator of sexual dimorphism of hepatic steroid metabolism (12). Since then, it has been shown in a number of instances, that the feminine pattern (i.e. higher baseline level and lower pulse height) of GH secretion can blunt the androgen responsiveness of the liver (14-17). The exact mechanism of this phenomenon is presently unclear. Regardless of the modifying influence of GH on hepatic gene expression, the induction of «2u-globulin in the in vitro liver perfusion system clearly indicates that androgen exerts a direct effect on the liver (19). In the context of our present understanding of the mechanism of steroid hormone action, it seems likely that such an in vitro effect of the hormone should be mediated through the AR (39). Results presented in this article show that the gene for AR is indeed expressed in the rat liver. However, the steady state level of AR mRNA in the young adult liver is rather low, i.e. only about 4% of the level expressed in the prostate, one of the most highly sensitive androgen targets. Not only is AR mRNA expressed in the liver, but its relative level during maturation and aging also correlates with differential hepatic sensitivity to androgens during the lifespan of the rat. For example, the liver of the prepubertal rat is insensitive to androgenic induction of a2u-globulin (18), and the AR mRNA in the prepubertal liver is beyond the detectibility limit of the RNAase protection assay. Similarly, development of androgen insensitivity during senescence correlates with a marked reduction of AR mRNA in the liver of the old rat. Sexspecific differential expression of AR mRNA in the liver also indicates a role for sex steroids in AR regulation. Regulation of the AR activity in the target tissue by the sex steroids may be mediated through complex molecular interactions. In fact, both up- and down-regulation of the AR activity and AR mRNA have been reported (4042). At this point we can only surmise that estrogen down-regulates AR mRNA in the liver. The interesting observation concerning cellular interactions in the regulation of a2u-globulin has suggested an important role of the nonparenchymal cells in the initial inductive response (37). Such an observation has led to the suggestion that a certain type of nonparenchymal cells, which constitute about 35% of the total hepatic cell population, may be the primary target for androgen action. This, in turn, produces a paracrine mediator that induces a2u-globulin. The above hypothesis had its origin in the mesenchymal-epithelial interaction involved in the expression of prostatic genes during development

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 25 July 2016. at 08:32 For personal use only. No other uses without permission. . All rights reserved.

HEPATIC AR mRNA (43). However, as shown in Fig. 6, the hepatocytes of the liver, and not the nonhepatocytes, are the site of expression of both AR and the a2u-globulin gene. Programmed loss of androgen sensitivity during aging may be linked to the overall regulatory dysfunction characteristic of the aging process. Calorie restriction is the only known experimental manipulation that can prolong longevity by delaying the onset of age-dependent dysfunctions (44). Therefore, it is highly intriguing that expression of the mRNA for a gene regulatory protein, i.e. the AR, can be modulated by calorie restriction. Further studies on factors involved in the age-dependent regulation of the AR gene are expected to provide important insights into the regulatory processes involving hormone action during aging. The above findings along with earlier observations of the in vitro androgenic induction of a2u-globulin in the perfused liver indicate a receptor-mediated direct effect of androgen on the regulation of a2u-globulin gene expression. Because a2u-globulin gene expression requires the synergistic influence of peptide hormones such as insulin and GH (5), it is conceivable that in addition to androgen, certain paracrine factors from the nonparenchymal cells may also be needed for the expression of this gene. In this context it is notable that a recent report shows binding of the prostatic AR with a 5' flanking fragment spanning —642 to —584 basepairs of the a2u-globulin gene (45). It is reasonable to assume that this type of direct effect of androgen is also operative on the regulation of other androgen-sensitive hepatic genes, such as carbonic anhydrase III and cytochrome P-450h. Thus, it is likely that androgens and GH act independently on the same subset of hepatic genes through two different signalling pathways.

20.

Acknowledgments

21.

The authors thank Drs. S. Liao and F. Sierra for kindly providing the rat AR and cytochrome oxidase cDNA clones, respectively.

22.

References

23.

1. Roy AK, Chatterjee B 1983 Sexual dimorphism in the liver. Annu Rev Physiol 45:37-50 2. Garg LC 1974 The effect of sex-hormones on rat liver carbonic anhydrase. J Pharmacol Exp Therap 197:297-302 3. Carter ND, Hewett-Emmett D, Jeffery S, Tashian RE 1981 Testosterone-induced sulfonamide resistant carbonic anhydrase isozyme of rat liver is indistinguishable from skeletal muscle carbonic anhydrase III. FEBS Lett 128:114-118 4. Waxman DJ, Dannan GA, Guengerich FP 1985 Regulation of rat hepatic cytochrome P-450: age-dependent expression, hormonal imprinting, and Xenobiotic inducibility of sex-specific isoenzymes. Biochemistry 24:4409-4417 5. Roy AK, Chatterjee B, Demyan WF, Milin BS, Motwani NM, Nath T Schiop MJ 1983 Hormone and age-dependent regulation of a2uglobulin gene expression. Recent Prog Horm Res 39:425-461 6. Dolan KP, Reinke R, Lynch KR, McLaughlin M, Unterman R, Nakhasi HL, Laperche Y, Feigelson P 1983 The application of

7. 8. 9.

10.

11.

12. 13. 14.

15. 16.

17. 18. 19.

24.

25. 26.

27. 28.

355

recombinant techniques to the study of the control of a2u -globulin gene expression. Biochem Act 10:1-19 Chatterjee B, Ozbilen O, Majumdar D, Roy AK 1987 Molecular cloning and characterization of the cDNA for androgen repressible rat liver protein SMP-2. J Biol Chem 262:822-825 Milin B, Roy AK 1973 Androgen receptor in rat liver: cytosol receptor deficiency in the pseudohermaphrodite male rat. Nature [New Biol] 242:248-250 Roy AK, Milin BS, McMinn DM 1974 Androgen receptor in rat liver: hormonal and developmental regulation of the cytoplasmic receptor and its correlation with the androgen-dependent synthesis of a2u-globulin. Biochim Biophys Acta 354:213-232 Sarkar FH, Sarkar PK, Watson S, Poulik MD Roy AK 1987 Cytoplasmic androgen binding protein of the rat liver: molecular characterization after photoaffinity labeling and functional correlation with a2u-globulin synthesis during maturation and aging. Biochemistry 26:3965-3970 Eagon PK, Seguiti SM, Rogerson BJ, McGuire TF, Porter LE, Seeley DH 1989 Androgen receptor in rat liver: characterization and separation from a male-specific estrogen-binding protein. Arch Biochem Biophys 268:161-175 Gustafasson JA, Mode A, Norstedt G, Skett P 1983 Sex steroid induced changes in hepatic enzymes. Annu Rev Physiol 45:51-60 Jansson J-O, Eden S, Isaksson O 1985 Sexual dimorphism in the control of growth hormone. Endocr Rev 6:128-150 Guzelian PS, Li D, Schuetz EG, Thomas P, Levin W, Mode A, Gustafsson J-A 1988 Sex change in cytochrome P-450 phenotype by growth hormone treatment of adult rat hepatocytes maintained in a culture system on matrigel. Proc Natl Acad Sci USA 85:97839787 Jeffery S, Wilson CA, Mode A, Gustafsson J-A, Carter ND 1986 Effects of hypophysectomy and growth hormone infusion on rat hepatic carbonic anhydrases. J Endocrinol 110:123-126 Husman B, Norstedt G, Mode A, Gustafsson J-A 1985 The mode of growth hormone administration is of major importance for the excretion of the major male rat urinary proteins. Mol Cell Endocrinol 40:205-210 Norstedt G, Palmiter R 1984 Secretory rhythm of growth hormone regulates sexual differentiation of mouse liver. Cell 36:805-812 Roy AK 1973 Androgenic induction of a2u-globulin in rat: androgen insensitivity in prepubertal animals. Endocrinology 92:957-960 Murty CVR, Rao KVS, Roy AK 1987 Rapid androgen stimulation of a^-globulin synthesis in perfused rat liver. Endocrinology 121:1814-1818 Chang C, Kokontis J, Liao S 1988 Molecular cloning of human and rat complementary DNA encoding androgen receptors. Science 240:324-326 Lubahn DB, Joseph DR, Sullivan PM, French FS, Wilson EM 1988 Cloning of human androgen receptor complementary DNA and localization to the X chromosome. Science 240:327-330 Chatterjee B, Fernandes G, Yu BP, Song CS, Kim JM, Demyan WF, Roy, AK 1989 Calorie restriction delays age-dependent loss in androgen responsiveness of the rat liver. FASEB J 3:169-173 Melton DA, Kreig PA, Rebagliati MR, Maniatis T, Zinn K, Green MR 1984 Efficient in vitro transcription of biologically active RNA and RNA hybridization probes from plasmids containing a bacteriophage SP6 promoter. Nucleic Acids Res 12:7035-7056 Braun L, Mead JE, Ranzica M, Mikumo R, Bell GI, Fausto N 1988 Transforming growth factor /3 mRNA increases during liver regeneration: a possible paracrine mechanism of growth regulation. Proc Natl Acad Sci USA 85:1539-1543 Chomczynski P, Sacchi N 1987 Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162:156-159 Rosenfeld GC, Comstock JP, Means AR, O'Malley BW 1972 Estrogen induced synthesis of ovalbumin messenger RNA and its translation in a cell free system. Biochem Biophys Res Commun 46:1695-1703 Aviv H, Leder P 1972 Purification of biologically active globin messenger RNA by chromatography on oligothymidylic acid-cellulose. Proc Natl Acad Sci USA 69:1408-1412 Sambrook J, Fretsch EF, Maniatis T 1989 Molecular Cloning-A

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 25 July 2016. at 08:32 For personal use only. No other uses without permission. . All rights reserved.

356

29. 30. 31. 32. 33. 34. 35.

36.

HEPATIC AR mRNA Laboratory Manual, ed 2. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, pp 7.43-52 Roy AK, Chatterjee B, Demyan WF, Nath TS, Motwani NM 1982 Pretranslational regulations of a2