MOLECULAR REPRODUCTION AND D E V E L O P M E N T 33:1-6 (1992)

Extraplacental Human Fetal Tissues Express mRNA Transcripts Encoding the Human Chorionic Gonadotropin+ Subunit Protein PAULA A. ROTHMAN, VICTOR A. CHAO, MARK R. TAYLOR, ROBERT W. KUHN, ROBERT B. JAFFE, AND ROBERT N. TAYLOR Reproductive Endocrinology Center, Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, S a n Francisco ABSTRACT The glycoprotein hormone human chorionic gonadotropin (hCG) is synthesized in large quantities by the developing placenta, reaching peak concentrations in maternal blood during the late first trimester and early midtrimester of pregnancy. In general it is believed that the a-subunit of this dimeric hormone is expressed in pituitary gonadotropes, thyrotropes, and trophoblasts, while the p-subunit is expressed exclusively by trophoblasts. Studies from our laboratory and other laboratories have shown that some midtrimester human fetal tissues, in addition to the placenta, can synthesize proteins that appear to be very similar to the p-subunit of hCG. To define precisely the nature of this putative hCG-@-subunitin extraplacental fetal tissues, we have examined the mRNA from a variety of human fetal and adult tissues using nucleic acid hybridization and reverse transcription-polymerase chain reaction (PCR) methods. Our results demonstrate that midtrimester fetal kidney and adrenal tissues contain hCG-p mRNA transcripts at concentrations comparable to that of placenta, while fetal lung, brain, muscle, and adult adrenal contain only trace to undetectable levels of hCG-p mRNA. By restriction endonuclease mapping of PCR fragments from fetal tissue cDNAs, we show that the hCG-p transcript expressed in midtrimester human fetal organs is a bone fide copy of hCG-@ gene No. 5 of the p-subunit gene family located on chromosome 19. 0 1992 Wiley-Liss, Inc.

Key Words: Fetal gene expression, hCG-p, Polymerase chain reaction INTRODUCTION The pattern of endocrine gene expression during human fetal development is complex, and its regulation is understood poorly. Placental hormones elaborated during pregnancy are believed to modulate maternal metabolism and may affect growth and differentiation of the developing fetus itself. Recent findings in reproductive organ systems suggest that classical protein and steroid hormones mediate some of their biological effects indirectly, via the production of locally acting substances (Dickson and Lippman, 1987; Lingham et al., 1988; Nelson et al., 1991). Similar findings have been reported in the developing human fetus (Voutilainen

0 1992 WILEY-LISS, INC.

and Miller, 1988; Mesiano et al., 1991). The expression of several human placental growth factors and their cognate receptors is maximal in the late first trimester and early second trimester, contemporaneous with the period of maximal linear fetal growth (Deal et al., 1982; Shen et al., 1986; Taylor and Williams, 1988). Since this developmental period also correlates with a period of elevated human chorionic gonadotropin (hCG) production, we have hypothesized t h a t hCG, in a manner analogous to that of other trophic protein hormones (Voutilainen and Miller, 1988; Mesiano et al., 1991), may be involved in the regulation of fetal growth. That such a mechanism may be fundamental to rapidly proliferating tissues is suggested by the atavistic expression of this fetal gene product in certain nontrophoblastic neoplasms (Braunstein et al., 1973; Baylin and Mendelsohn, 1980). Prior studies from our laboratories demonstrated that several fetal organs contain high concentrations of hCG protein (Huhtaniemi et al., 1978) and that organ cultures of midtrimester human fetal kidneys produce de novo a gonadotropic substance with biological and immunological properties indistinguishable from those of hCG (McGregor et al., 1983). The latter observation indicated that the detection of hCG in extraplacental tissues was not a n artifact of blood contamination. Other investigators have verified that immunodetectable hCG was present in midtrimester fetal thymus (Fukayama et al., 1990) and pituitary cells (Ode11et al., 1990) and even in adult lymphocytes (Harbour-McMenamin e t al., 1986). Criticisms of these studies raised concerns that the immunological activities detected might represent nonspecific cross reactivity with the hCG antibodies employed in the assays and that expression of hCG by fetal tissue explants may reflect a n in vitro artifact. To extend our evidence that extraplacental fetal tissues express genes that encode this glycoprotein hormone, we have assessed the presence of specific mRNA precursors required for bone fide hCG production in vivo. Received January 30,1992; accepted March 2,1992. Address reprint requests to Robert N. Taylor, MD, PhD, Department of Obstetrics, Gynecology and Reproductive Sciences, School of Medicine, M-1489 University of California, San Francisco, CA 94143-0132.

2

P.A. ROTHMAN ET AL.

MATERIALS AND METHODS Placentae and fetal kidneys, adrenals, lungs, brain (cerebral cortex), and striated muscle tissues were obtained from elective midtrimester pregnancy terminations at 16-20 gestational weeks (as determined from last menses) in accordance with a protocol approved by the UCSF Committee on Human Research. The organs were identified visually and placed immediately into separate containers of ice-cold saline. Due to the high degree of sensitivity inherent in the PCR amplification, great care was taken to rinse and dissect each specific organ from loosely adherent connective tissue to avoid contamination of the fetal organs by trophoblastic cells. The dissected organs were frozen in liquid nitrogen in individual containers and transported to the laboratory for RNA preparation. Rat adrenal tissue was kindly provided by Dr. Synthia Mellon and adult human adrenal tissue by Dr. Sam Mesiano (Reproductive Endocrinology Center, UCSF). RNA was prepared from the tissues by homogenization in guanidinium isothiocyanate and purified by centrifugation through a 5.7 M CsCl cushion (Chirgwin et al., 1979). Slot blots were performed using Nytran (Schleicher & Schuell, Keene, NH) filters after denaturing 3 bg total RNAislot in 4.8 M formaldehyde: 7.5 x SSC. Northern analyses were performed using 10 pg total RNA/lane of agarose-formaldehyde gels run in MOPS buffer and transferred to Nytran membranes (Ausubel et al., 1989). Random primer-labelled ["PIhCG-P cDNA was hybridized to the filters as described previously (Taylor e t al., 1991). Similar results were obtained in three Northern blots. Purified RNA samples (0.1 pg each) were reverse transcribed, and the resultant cDNAs were amplified for 35 cycles using a Perkin Elmer Cetus DNA thermal cycler (Norwalk, CT) according to the methods of Rappolee et al. (1989). A step program (95"C, 20 sec; 63"C, 4 sec; 75"C, 70 sec) was followed by a 10 min final extension reaction at 75°C. The PCR products were separated on 4% agarose (3% NuSieve GTG and 1%SeaKem ME; FMC Bioproducts, Rockland ME) gels and visualized by ethidium bromide staining. To verify the specificity of the hCG-P gene product, analytic restriction endonuclease cleavages were performed with Hue111 and TuqI (Boehringer-Mannheim, Indianapolis, IN). The results shown in the figures were representative of a t least four independent PCR amplifications. RESULTS Slot blot analysis of total RNA, using a 0.6 kb hCG-p cDNA probe under stringent hybridization and washing conditions (Fig. l A ) , initially was used to identify hCG-P mRNA in midtrimester placenta and fetal kidney. Quantitatively, both tissues appeared to contain similar steady-state concentrations of hCG-P-like mRNA transcripts, while a n equal amount of human fibroblast RNA showed no detectable hybridization with the hCG-P cDNA probe. To characterize the molecular size of the fetal kidney-derived transcripts, Northern analyses were performed using similar hy-

Fig. 1. A Slot blot analyses of hCG-B mRNA. RNA samples (3 kg/slot) from 18-week placenta, 18-week fetal (F.) kidney, and cultured human fibroblasts were denatured in formaldehyde and filtered onto Nytran. A buffer control without RNA also was included. The filter was hybridized with L"P1hCG-B cDNA and washed under stringent conditions before autoradiography. B: Northern analysis of fetal kidney hCG-p mRNA. Total RNA (10 pg) prepared from a pair of 19-week fetal kidneys was denatured in formaldehyde, subjected to electrophoresis in a 1%formaldehyde-agarose gel, and blotted to Nytran. The filter was baked and hybridized with IT'PP]hCG-p cDNA. ["'P]PM2 DNA cut withHindII1 was used to provide molecular weight standards shown at right.

bridization conditions (Fig. 1B). [32P]hCG-PcDNA hybridization to fetal kidney RNA identified a 1.1 kb mRNA transcript that comigrated with the placental transcript and is characteristic of hCG-P mRNA in trophoblast cells (Milsted et al., 1987; Ringler et al., 1989; Taylor et al., 1991). This evidence suggested, but did not prove, that fetal kidney expressed a n authentic hCG-P mRNA. The identity of placental, fetal kidney, and other fetal organ hCG-P mRNA transcripts was confirmed using the reverse transcription-PCR (RT-PCR) method. Several pairs of oligonucleotide primers derived empirically from the hCG-P gene sequence (Fiddes and Goodman, 1980) did not reproducibly generate discrete amplification products. A successful primer pair was deduced using the Oligo 3.4 program (W. Rychlik, 1989). These primers, shown in Figure 2 alongside the cognate genomic DNA sequence of hCG-6, span the second intron (234 nucleotides) of this gene. The primers were designed to yield a n amplification product from hCG-P cDNA of 294 nucleotides in length or a genomic DNA product of 528 bases. Advantage was taken of a dinucleotide rearrangement a t bases 1977 and 1978, in which a TG doublet in hCG-P cDNA is replaced by CC in the closely related luteinizing hormone (LH)-P cDNA sequence, to yield a 3' primer that would preferentially hybridize with the hCG-P cDNA

EXTRAPLACENTAL hCG-p GENE EXPRESSION

3

hCG-5 LH

(1451) CCACCCTGGC TGTGGAGAAG GAGGGCTGCC CCGTGTGCAT CACCGTCAAC ACCACCATCT GTGCCGGCTA (1451) CCACCCTGGC TGTGGAGAAG GAGGGCTGCC CCGTGTGCAT CACCGTCAAC ACCACCATCT GTGCCGGCTA 5' PCR primer

hCG-5 LH

(1521) CTGCCCCACC ATG.. ( I N T R O N , 234 4).. (1768) ACC CGCGTGCTGC AGGGGGTCCT GCC-CTG (1521) CTGCCCCACC ATG.. ( I N T R O N . 234 4}..(1768) Atg CGCGTGCTGC AGGGGGTCCT GCCGcCCCTG HaeIII site

hCG-5 LH

(1801) CCTCAGGTGG TGTGCAACTA CCGCGATGTG CGCT-GT CCATCCGGCT CCCTGGCTGC CCGCGCGGCG (1801) CCTCAGGTGG TGTGCAcCTA CCGtGATGTG C G C T W G T CCATCCGGCT CCCTGGCTGC CCGCGtGGCG T a q I site

hCG-5 LH

(1871) TGAACCCCGT GGTCTCCTAC GCCGTGGCTC TCAGCTGTCA ATGTGCACTC TGCCGCCGCA GCACCACTGA (1871) TGgACCCCGT GGTCTCCTtC CCtGTGGCTC TCAGCTGTCg cTGTGgACcC TGCCGCCGCA GCACCtCTGA

hCG-5 LH

(1941) CTGCGGGGGT CCCAAGGACC ACCCCTTGAC CTGTGATGAC CCCGGCTTCC (1941) CTGtGGGGGT CCCAAaGACC ACCCCTTGAC CTGTGAccAC CCCCGCTTCC 3 ' PCR primer

Fig. 2. Partial DNA sequences of hCG-5 and LH @-subunitgenes showing oligonucleotide primer designs. Oligonucleotide primers (23mers) were deduced using the Oligo 3.4 program as described in the text. The primers span the second intron of the hCG-p gene No. 5 (hCG-5), and the numerical nucleotide positions (indicated in parentheses to the left of the corresponding base) refer to the GenBank designation. To convert these values to the nucleotide positions defined by Fiddes and Talmadge (19841, subtract 446 bases. Going from 5' to 3', the underscored sequences in the figure indicate: 11 the 5' PCR

primer sequence, homologous in the hCG-P and LH-p coding regions; 2) the 234 base (b) second intron separating exons 2 and 3 of both P-subunit genes; 3 ) a n HaeIII site (nucleotides 1794-1797) present in the hCG-p cDNA but absent in LH-p cDNA; 4) a TuqI site (nucleotides 1835-1838) present in both hCG-@and LH-@cDNA; and 5)the 3' PCR primer, which is homologous to the hCG-@sequence but differs from LH-P cDNA a t nucleotides 1977 and 1978. Nonhomologous nucleotides in the LH-0 cDNA sequence are indicated with lowercase letters.

sequence (Fig. 2). As a n internal control of RNA yield, integrity, and efficiency of reverse transcription, parallel RNA samples were assayed for amplification of the glyceraldehyde-3-phosphatedehydrogenase (GAPDH) mRNA transcript (241 base pair product). The primers for GAPDH span no intron, so genomic and cDNA amplification products are not distinguished by this method. RNA isolated from midtrimester human placenta, kidney, adrenal, lung, brain, and muscle was purified through CsCl gradients and reverse transcribed as described in Materials and Methods. PCR amplification of the cDNAs prepared from these fetal tissues yielded the expected 294 base fragment, indicating the presence of hCG-P mRNA in placenta, kidney, and adrenal preparations. The densities of the 294 base band in these tissues were similar, suggesting comparable steadystate concentrations of this transcript. By contrast, however, fetal lungs reproducibly contained less hCG-P mRNA than the other fetal tissues (Fig. 3A), and fetal brain and muscle contained no detectable hCG+ mRNA transcripts (Fig. 3B, lanes 7 and 11).Positive controls for the fetal brain and muscle cDNAs are provided by the presence of GAPDH products (Fig. 3B, lanes 5 and 9). Negative control lanes, in which no cDNA was added to the PCR reaction, were consistently blank (Fig. 3B,even-numbered lanes). In some experiments, genomic DNA contamination of the RNA pellet was purposefully introduced during centrifugation by diluting the concentration of CsCl in the cushion to 4.0 M. Under the latter conditions, PCR amplification of reverse-transcribed fetal adrenal nu-

cleic acid using the same primer pair yielded a 528 nucleotide hCG-p product as predicted for the genomic sequence (Fig. 4, lane 1). The GAPDH product, which does not span a n intron, ran as the expected 241 nucleotide fragment (Fig. 4 , lane 2). Human fetal adrenal RNA, appropriately purified from genomic DNA through a 5.7 M CsCl cushion, yielded a PCR product of 294 bases (Fig. 4, lane 3). Adrenal RNA from the rat, a species that has no chorionic gonadotropin genes (Tepper and Roberts, 19841, was subjected to identical reverse transcription and PCR protocols to provide a negative control for hCG-P mRNA detection (Fig. 4, lane 5). However, the GAPDH primers yielded a n amplification product of this phylogenetically conserved control transcript (Fig. 4, lane 6). To determine if the expression of hCG-P mRNA by the human adrenal gland was confined to fetal life, adult human adrenal tissue was analyzed using the same technique. With adrenals from two different adult men, PCR amplification yielded trace to undetectable amounts of hCG-P DNA fragments (Fig. 5, lanes 4 and 6). To confirm the identity of the amplified human placental, fetal kidney, and fetal adrenal sequences, the bands were excised from the gel, eluted, and digested with restriction endonuclease HueIII. The full-length 294 base hCG+ cDNA PCR product (Fig. 6, lane 1)was cleaved by HueIII to yield subfragments of 189 and 105 nucleotides (Fig. 6, lanes 2-4) and by TaqI to yield a n apparent doublet of 149 and 145 base pairs (individual fragments were not resolved in the 4% gel; data not shown). In addition to corroborating the sequence spec-

4

P.A. ROTHMAN ET AL. vl

$

n r t n

Placenta

F. brain

F. muscle

2 - n 2 ,94 bp -241 bp

*. 1 2 3 4 5 6

294, 241

-

6.

Fig. 3. Reverse transcription-polymerase chain reaction (RT-PCR) analyses of fetal mRNAs. A Placenta and fetal (F.) kidney and lung RNA from a 19-week pregnancy were assessed by RT-PCR and ethidium bromide staining in 4% agarose gels. Amplification of GAPDH cDNA (241 base pairs) served as a n internal control for each tissue (lanes 1, 3, 5, respectively). The 294 nucleotide product of hCG-P cDNA was observed in all three fetal tissues (lanes 2,4,6); however, significantly lower concentrations of this transcript were consistently noted in fetal lung (lane 6). B: Placenta and fetal (F.)brain and muscle

ificity of the amplification reaction products, the former cleavage pattern distinguished the hCG-P cDNA sequence from the closely related LH-P cDNA sequence (see Fig. 2), which lacks a n HaeIII site.

1 2 3 4 5 6 7 8 9 101112 RNA from an 18-weekpregnancy were assessed by RT-PCR and ethidium bromide staining in 4% agarose gels. Amplification of GAPDH cDNA (241 base pairs) served as an internal positive control for each tissue (lanes 1, 5,9, respectively), while exclusion of cDNA from the reaction mixes provided negative controls (lanes 2,4,6,8,10,11). The 294 nucleotide product of hCG+ cDNA was observed only in placenta (lane 3) and was absent in fetal brain and muscle (lanes 7 and 11, respectively 1.

5

C

a,

DISCUSSION These experiments confirm that the hCG synthesized by extraplacental human fetal tissues results from the expression of the authentic hCG-P gene in vivo. This conclusion is consistent with our previous immunohistochemical study (Goldsmith et al., 1983) and in vitro experiments demonstrating t h a t the hCG-like protein synthesized by fetal kidney explants had immunologic, chromatographic, and biological properties similar to those of placental hCG (McGregor et al., 1983). Moreover, the fetal tissue pattern of hCG-P mRNA expression in vivo paralleled the de novo synthesis of hCG subunit proteins by organ cultures in vitro, with placenta and fetal kidney expressing considerably higher concentrations than fetal lung and muscle (McGregor et al., 1983). Although these studies do not address the expression of a-glycoprotein subunit mRNA, the de novo synthesis of biologically active, intact hCG by fetal kidney minces (McGregor et al., 1983) indicates that this precursor must be transcribed and translated. Eight homologous genes or pseudogenes located on human chromosome 19 were thought initially to make up the hCG-P gene family (Naylor et al., 1983; Fiddes and Talmadge, 1984); however, the data of Graham et al. (1987) suggest that one of these genes (hCG-P gene No. 6) is a cloning artifact. The predicted amino acid sequence encoded by hCG-p gene No. 5 agrees com-

- 528 bp

-294 bp 1 2 4 1 bp 1 2 3 4 5 6 Fig. 4. Reverse transcription-polymerase chain reaction (RT-PCR) analyses of fetal adrenal mRNA with controls for genomic contamination. Human fetal (f.)adrenal RNA contaminated with genomic DNA (lanes 1 and 2; see Results), purified human fetal adrenal RNA (lanes 3 and 41, and rat adrenal RNA (lanes 5 and 6) were assessed by RT-PCR as in Figure 3. The GAPDH control amplification was positive in all cases (lanes 2, 4, 6). Amplification of human fetal adrenal nucleic acid in the presence of contaminating genomic sequences yielded a 528 base pair product, as predicted from the additional size of the second intron of the hCG-P gene (lane 1). Purified human fetal adrenal RNA (lane 3) gave the expected 294 base pair hCG+ cDNA product. Rat adrenal RNA (lane 5) had no transcripts homologous to the hCG-p sequence, although the highly conserved GAPDH sequence was efficiently amplified (lane 61, verifying the integrity of the RNA in this preparation.

EXTRAPLACENTAL hCG-P GENE EXPRESSION

-c m

?

U

cu m c

\

a,

LL

‘nn-

,294

bp

2

-241

bp

1 2 3 4 5 6 Fig. 5. RT-PCR analyses of human fetal and adult adrenal tissues. RNA from a 17-week fetal adrenal and two different adult adrenal specimens was analyzed using GAPDH primers (lanes 1, 3, and 5, respectively) and hCG-P primers (lanes 2, 4, and 6, respectively), demonstrating significantly lower concentrations of hCG-p mRNA in the adult tissues.

T i r uu -294 bp - 189 bp

- 105 bp 1 2 3 4 Fig. 6. Restriction endonuclease mapping of RT-PCR amplified human fetal hCG-p mRNA transcripts. The sequence of the RT-PCR generated hCG-p cDNA products was confirmed by mapping with H a d 1 endonuclease (see Fig. 21. The 294 base pair RT-PCR product from placental RNA (lane 1) yielded subfragments of 189 and 105 nucleotides following digestion with Hue111 (lane2). Identical restriction enzyme-generated fragments were noted when fetal (F.) kidney and adrenal RT-PCR products were cut with HaeIII, corroborating their identification as bone fide hCG-p sequences (lanes3,4).

pletely with the established amino acid sequence of the purified protein (Morgan et al., 1975). By transient transfection analyses in COS cells, only hCG-p genes Nos. 3 and 5 appear to have the necessary structural components for transcriptional expression (Talmadge et al., 1984). Gene No. 4 of this family encodes the LH-P subunit, which is normally regulated in a tissue-specific manner, with its expression confined to the pituitary gland (Fiddes and Talmadge, 1984). For the analyses performed in the current study, we used the DNA

5

sequence of hCG-p gene No. 5 to generate our PCR primers. As revealed by restriction enzyme mapping of amplified placental, kidney, and adrenal cDNAs, the gene expressed in the fetal tissues is hCG-P and not LH-P. These studies demonstrate that the hCG+ gene is expressed in fetal kidney and adrenal during normal development, but that fetal lung, brain, and muscle (Fig. 3) and differentiated adult tissues (e.g., fibroblasts [Fig. lA] and adrenal gland [Fig. 51) contain undetectable to trace concentrations of hCG-p mRNA, despite the highly sensitive assays used in these analyses. De-repression and atavistic expression of this hormone are known to occur in certain tumors (Baylin and Mendelsohn, 1980) and may be induced in vitro by agents that demethylate DNA sequences (Wong e t al., 1984). Thus the expression of hCG in certain fetal tissues may be simply a fortuitous result of DNA hypomethylation. Alternatively, the local expression of hCG may be regulated in proliferative and/or differentiative functions of specific fetal organs. Possible endocrine/ paracrine roles of hCG in the placenta and fetal adrenal are supported by in vitro experiments (Menon and Jaffe, 1973; Seron-Ferre et al., 1978; North et al., 1991). In addition, Nomura et al. (1988) suggests that gonadotropins may play a role in rodent renal development. Thus the fetal organs that we have shown can synthesize hCG also may be targets of hCG action. Future studies will attempt to determine whether specific peptide mitogens or secondary messengers activated by hCG might mediate the growth and development of these human fetal organs in vivo.

ACKNOWLEDGMENTS We thank Drs. D. Rappolee and 2. Werb (Department of Anatomy, UCSF) for their expert advice on the establishment of the reverse transcription-polymerase chain reaction assays and Drs. J. Fiddes and K. Talmadge (California Biotechnology, Mountain View, CA) for providing the hCG-p cDNA clone. This work was supported by NIH grants HD 18726, HD 22873, and HD 11979 and a Basil O’Connor Scholar Award (5-810) to R.N.T. from the March of Dimes Birth Defects Foundation. REFERENCES Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struhl K (1989):Preparation and analysis of RNA. In: “Current Protocols in Molecular Biology.” New York: John Wiley & Sons, pp 491-494. Baylin SB, Mendelsohn G (1980): Ectopic (inappropriate) hormone production by tumors: Mechanisms involved and the biological and clinical implications. Endocrine Rev 1:45-76. Braunstein GD, Vaitukaitis JL, Carbone PP, Ross GT (1973):Ectopic production of human chorionic gonadotropin by neoplasms. Ann Intern Med 73:39-45. Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter W J (1979): Isolation of biologically-active ribonucleic acid from sources enriched in rihonuclease. Biochemistry 18:5294-5299. Deal CL, Guyda HJ, Lai WH, Posner BI (1982): Ontogeny of growth factor receptors in the human placenta. Pediatr Res 16:820-826.

6

P.A. ROTHMAN ET AL.

Dickson RB, Lippman ME (1987):Estrogenic regulation ofgrowth and polypeptide growth factor secretion in human breast carcinoma. Endocrine Rev 8:2943. Fiddes JC, Goodman HM (1980): The cDNA for the p-subunit of human chorionic gonadotropin suggests evolution of a gene by readthrough into the 3’-untranslatedregion. Nature 286:684487. Fiddes JC, Talmadge K (1984):Structure, expression, and evolution of the genes for the human glycoprotein hormones. Recent P r o p Hormone Res 40:43-78. Fukayama M, Hayashi Y, Shiozawa Y, Maeda Y, Koike M (1990): Human chorionic gonadotropin in the thymus: An immunocytochemical study on discordant expression of subunits. Am J Pathol 136:123-129. Goldsmith PC, McGregor WG, Raymoure WJ, Kuhn RW, Jaffe RB (1983): Cellular localization of chorionic gonadotropin in human fetal kidney and liver. J Clin Endocrinol Metab 57:654-661. Graham MY, Otani T, Boime I, Olson MV, Carle GF, Chaplin DD (1987): Cosmid mapping of the human chorionic gonadotropin p subunit genes by field-inversion gel electrophoresis. Nucleic Acids Res 15:44374448. Harbour-McMenamin D, Smith EM, Blalock J E (1986):Production of immunoreactive chorionic gonadotropin during mixed lymphocyte reactions: A possible selective mechanism for genetic diversity. Proc Natl Acad Sci USA 83:683&6838. Huhtaniemi IT, Korenbrot CC, Jaffe RB (1978): Content of chorionic gonadotropin in human fetal tissues. J Clin Endocrinol Metab 46:994-997. Lingham RB, Stance1 GM, Loose-Mitchell DS (1988):Estrogen regulation of epidermal growth factor receptor messenger ribonucleic acid. Mol Endocrinol2:230-235. McGregor WG, Kuhn RW, Jaffe RB (1983): Biologically active chorionic gonadotropin: Synthesis by the human fetus. Science 220:306308. Menon KMJ, Jaffe RB (1973): Chorionic gonadotropin-sensitive adenylyl cyclase in human term placenta. J Clin Endocrinol Metab 36:110&1109. Mesiano S, Mellon SH, Gospodarowicz D, DiBlasio AM, Jaffe RB (1991): Basic fibroblast growth factor expression is regulated by corticotropin in the human fetal adrenal: A model for adrenal growth regulation. Proc Natl Acad Sci USA 885428-5432. Milsted A, Cox RP, Nilson J H (1987):Cyclic AMP regulates transcription of the genes encoding human chorionic gonadotropin with different kinetics. DNA 6:213-219. Morgan FJ, Birkin S, Canfield RE (1975):The amino acid sequence of human chorionic gonadotropin. J Biol Chem 250:5247-5258. Naylor SL, Chin WW, Goodman HM, Lalley PA, Grzeschik K-H, Sakaguchi AY (1983): Chromosome assignment of genes encoding the a

and p subunits of glycoprotein hormones in man and mouse. Somat Cell Genet 9:757-770. Nelson KG, Takahashi T, Bossert N, Walmer DK, MacLachlan J A (1991): Epidermal growth factor replaces estrogen in the stimulation of female genital-tract growth and differentiation. Proc Natl Acad Sci USA 88:21-25. Nomura K, Tsunasawa S, Ohmura K, Sakiyama F, Shizume K (1988): Renotropic activity in ovine luteinizing hormone isoform(s). Endocrinology 123:700-712. North RA, Whitehead R, Larkins RG (1991): Stimulation by human chorionic gonadotropin of prostaglandin synthesis by early human placental tissue. J Clin Endocrinol Metab 73:60-70. Odell WD, Griffin J , Bashey HM, Snyder PJ (1990):Secretion ofchorionic gonadotropin by cultured human pituitary cells. J Clin Endocrinol Metab 71:1318-1321. Rappolee DA, Wang A, Mark D, Werb Z (1989): Novel method for studying mRNA phenotypes in single or small numbers of cells. J Cell Biochem 39:l-11. Ringler GE, Kao L-C, Miller WL, Strauss J F 111 (1989): Effects of 8-bromo-CAMP on expression of endocrine functions by cultured human trophoblast cells. Mol Cell Endocrinol61:13-21. Seron-Ferre M, Lawrence CC, Jaffe RB (1978): Role of hCG in the regulation of the fetal zone of the human fetal adrenal gland. J Clin Endocrinol Metab 46:834-837. Shen S-J, Wang C-Y, Nelson KK, Jansen M, Ilan J (1986):Expression of insulin-like growth factor I1 in human placentas from normal and diabetic pregnancies. Proc Natl Acad Sci USA 83:9179-9182. Talmadge K, Boorstein WR, Vamvakopoulos NC, Gething M-J, Fiddes J C (1984): Only three of the seven human chorionic gonadotropin beta genes can be expressed in the placenta. Nucleic Acids Res 12:8415-8436. Taylor RN, Newman ED, Chen S (1991):Forskolin and methotrexate induce a n intermediate trophoblast phenotype in cultured human choriocarcinoma cells. Am J Obstet Gynecol 164:20&210. Taylor RN, Williams LT (1988):Developmental expression of plateletderived growth factor and its receptor in the human placenta. Mol Endocrinol2:627-632. Tepper MA, Roberts JL (1984): Evidence for only one P-luteinizing hormone and no 6-chorionic gonadotropin gene in the rat. Endocrinology 115:385-391. Voutilainen R, Miller WL (1988):Developmental and hormonal regulation of mRNAs for insulin-like growth factor I1 and steroidogenic enzymes in human fetal adrenals and gonads. DNA 7:9-15. Wong DTW, Hartigan JA, Biswas DK (1984): Mechanism of induction of human chorionic gonadotropin in lung tumor cells in culture. J Biol Chem 259:10738-10744.