TC #
BIOLOGY OF REPRODUCTION 59, 621–625 (1998)
Differential Expression of a Tumor Necrosis Factor Receptor-Related Transcript in Gestational Trophoblastic Diseases in Women1 Catherine I. Dumur, Nicola´s P. Koritschoner, Alfredo Flury, Graciela Panzetta-Dutari, Jose´ L. Bocco, and Luis C. Patrito2 Departamento de Bioquı´mica Clı´nica, Facultad de Ciencias Quı´micas, Universidad Nacional de Co´rdoba, Co´rdoba, Argentina ABSTRACT
liferative state of trophoblasts and to be accompanied by some risk of progression to choriocarcinoma. A careful postmolar gonadotropin follow-up is necessary to ensure sustained remission and also for early detection of CHMderived persistent GTT and/or choriocarcinoma [4]. It has been demonstrated that CHM has some biochemical and molecular alterations in comparison to normal early placenta (NEP) and normal term placenta (NTP). For example, we have shown a low aromatase activity in microsomes from the molar tissue [5] and a down-regulation of pregnancy-specific glycoproteins (PSG) genes in CHM [6] in comparison with what is seen in both normal early and term placentas. It has also been demonstrated that some protooncogenes as well as p53, epidermal growth factor, and its receptor genes showed different expression levels, which were higher in CHM and choriocarcinoma than in normal placenta [7–10]. The simultaneous expression of epidermal growth factor and its receptor in trophoblastic tissues suggests that it may act in an autocrine-paracrine manner, playing important roles in proliferation, differentiation, and development of the normal and neoplastic trophoblasts [10]. In the present study our aim was to identify other differentially expressed genes in CHM as compared to normal placenta in an effort to better understand the molecular genetic alterations associated with the pathogenesis of trophoblastic diseases. By differential screening of a molar cDNA library we have isolated and characterized a gene, complete hydatidiform mole-specific-1 (CHMS-1), whose corresponding transcript is enriched in the molar tissue. This transcript was undetectable in CHM-derived persistent GTT, as well as in the human choriocarcinoma-derived cell line JEG-3, as revealed by Northern blotting analysis. The features encountered in CHMS-1 suggested that it encodes a new member of the tumor necrosis factor (TNF) receptor superfamily bearing a putative death domain, like TNF-R1 [11], Fas [12], and the Drosophila reaper protein [13, 14]. In these proteins, the death domain plays a crucial role activating programmed cell death or apoptosis [15]. Most interestingly, TNF-R1 and CHMS-1 shared an identical expression profile in all the trophoblastic tissues and cell lines analyzed. Collectively, our findings are in line with the idea that CHMS-1 may participate in the TNF-dependent apoptotic signaling pathway probably involved in the natural course of spontaneous regression followed by the majority of hydatidiform moles.
Gestational trophoblastic diseases comprise a group of interrelated neoplasms, including complete hydatidiform mole (CHM), persistent gestational trophoblastic tumor (GTT), and choriocarcinoma. To better define the molecular features of these diseases, a CHM cDNA library was constructed and a novel cDNA sequence, named CHMS-1, was isolated by differential screening. The CHMS-1 sequence showed a 62% homology with the tumor necrosis factor receptor (TNF-R2) cDNA, and its amino acid deduced sequence shared a high level of homology with the ‘‘death domain’’ region found in various proteins, including two members of the TNF receptor superfamily, the TNF-R1 and Fas. We also determined the CHMS-1, TNF-R1, and TNF-R2 expression patterns among different CHM tissues and cell lines of trophoblastic (JEG-3) and nontrophoblastic (HeLa and COS-7) origin. Our results indicated that the CHMS-1 transcript is highly expressed in CHM in comparison with both normal early and term placenta and that it exhibits an expression profile identical to that of TNF-R1. Furthermore, the CHMS-1 transcript was undetectable in CHM-derived GTT and in the human choriocarcinoma-derived JEG-3 cells, suggesting that its expression is down-regulated in the malignant transformation of trophoblast. The presence of a potential ‘‘death domain’’ in CHMS-1, together with its high expression level in CHM, strongly suggests that the CHMS-1 gene encodes a protein that might be involved in tumor regression processes occurring at later stages of molar development.
INTRODUCTION
The gestational trophoblastic disease complete hydatidiform mole (CHM) is characterized in the human conceptus by hyperplasia of the trophoblast cells, edematous swelling of the chorionic villi, and complete absence of fetal tissue [1]. Both dispermic and monospermic CHMs are diploid, and all the pairs of chromosomes are paternally derived [2, 3]. It is important to distinguish CHM from other types of fetal wastage because the risk of developing persistent gestational trophoblastic tumor (GTT) increases after CHM and patients require subsequent chemotherapy. CHM could also be considered to be characteristic of a disorderly proAccepted April 28 1998. Received February 24, 1998. 1 This work was supported by grants from the Consejo Nacional de Investigaciones Cientı´ficas y Tecnolo´gicas de Argentina (CONICET), the Consejo de Investigaciones Cientı´ficas y Tecnolo´gicas de la Provincia de Co´rdoba (CONICOR) and the Secretarı´a de Ciencia y Te´cnica de la Universidad Nacional de Co´rdoba (SECyT). The nucleotide sequence reported in this work has been submitted to the GenBank Data Bank under the accession number: AF040257. 2 Correspondence: Luis C. Patrito, Departamento de Bioquı´mica Clı´nica, Facultad de Ciencias Quı´micas, Universidad Nacional de Co´rdoba, Agencia Postal N8 4. C.C. 61 (5000) Co´rdoba Argentina. FAX: 54 51– 334174; e-mail:
[email protected]
MATERIALS AND METHODS
Tissue Collection and Cell Lines
The tissues employed in this study were obtained after therapeutic abortion (CHMs and NEP) from seven patients with amenorrhea of about a 120-day duration and after nor621
TC #
622
DUMUR ET AL.
mal delivery (NTP). Six individuals’ CHMs were identified on the basis of gross morphology and histopathology. One of the cases collected had persistently raised serum concentrations of hCG on subsequent follow-up, and singleagent chemotherapy was given. The tissues were repeatedly washed with ice-cold 137 mM NaCl, cut into small pieces, and stored at 2708C within 30 min after the material was available. HeLa, COS-7, and JEG-3 cells were grown in Dulbecco’s Modified Eagle’s medium supplemented with 5% fetal calf serum, streptomycin (0.1 mg/ml), and penicillin (100 U/ml).
blunt-end ligated to a synthetic EcoRI linker, and cut with EcoRI endonuclease. The cDNA was separated from unligated linkers by size exclusion chromatography on a sepharose 4B column and ligated to precut dephosphorylated lgt10 arms. Phage packaging was performed using the Packagene System (Promega, Madison, WI) as described by the manufacturer. The library was used to transfect Escherichia coli C600 and plated on large Luria broth plates. About 1 3 105 plaque-forming recombinants containing inserts 2 kilobases (kb) in average size were obtained from the CHM library. Preparation of cDNA Probes
Human CG Assay
The hCG was measured in urine using direct agglutination method (detection limit 0.3 IU/ml). The latex particles were coated with a monoclonal antibody anti-b subunit (Embriolatex, Santa Fe, Argentina). When the hCG was undetectable by this method, the hormone was measured in serum using an immunoradiometric assay (IRMA D.P.C.; Diagnostic Products Corporation, Los Angeles, CA). RNA Purification and Northern Blot Analysis
RNA samples were purified from JEG-3, HeLa, and COS-7 cells and tissues essentially as described by Chomczynski and Sacchi [16] with slight modifications. For Northern blotting, RNA samples were fractionated by electrophoresis through 1.2% agarose/formaldehyde gels and transferred overnight to nylon membranes in a solution of 10-strength SSC (single-strength SSC is 0.15 M sodium chloride and 0.015 M sodium citrate). The nylon filters were subjected to UV cross-linking for 4 min on a Bio-Rad (Richmond, CA) 300-nm UV transilluminator. The filters were then prehybridized for at least 4 h at 428C in prehybridization solution (50% formamide, 5-strength SSC [0.75 M NaCl, 0.075 M sodium citrate, pH 7.0], 5-strength Denhardt’s solution [0.5% Ficoll, 0.5% polyvinylpyrrolidone, 0.5% BSA], 0.1% SDS, and 100 mg/ml heat-denatured salmon sperm DNA). Filters were hybridized for 16 h at 428C in hybridization solution (same as prehybridization solution plus 5 ng/ml 32P-labeled cDNA probe). Complementary DNA probes were labeled with [32P]dATP to a specific activity of 5 3 108 to 3 3 109 cpm/mg DNA using the random hexamer procedure. After hybridization, filters were washed twice at room temperature (15 min each) in double-strength SSC/0.2% SDS; this was followed by two washes at 428C for 30 min each and two washes at 658C for 30 min with the same solution. The filters were air dried and exposed for autoradiography. For reprobing, bound probes were removed by boiling in deionized water for 5 min. Complete removal of the cDNA probes was controlled by autoradiography. In some experiments, the ubiquitously expressed gene coding for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a probe to normalize RNA loading. Construction of CHM cDNA Library
Poly(A)1 RNA was isolated from CHM using oligo(dT)cellulose (Pharmacia, Piscataway, NJ) chromatography columns and standard techniques. Poly(A)1 RNA (5 mg) was used as a template for cDNA synthesis using murine Moloney leukemia virus (MMLV) reverse transcriptase and Escherichia coli DNA polymerase I method. Doubledstranded cDNA was methylated using EcoRI methylase,
32P-Labeled cDNA was synthesized from poly(A)1 RNA purified from either CHM or NTP. In a total reaction volume of 50 ml, 100 mCi [32P]dATP, 7 mg poly(A)1 RNA, 50 mg/ml oligo(dT), reaction buffer (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 10 mM dithiothreitol, 3 mM MgCl2), 0.5 mM dGTP, 0.5 mM dTTP, 0.5 mM dCTP, 25 U RNAsin (Promega), and 200 U MMLV reverse transcriptase (Promega) were added. After first-strand synthesis for 120 min at 378C, 12.5 ml of 1 M NaOH was added, followed by an incubation for 20 min at 658C and neutralization with 12.5 ml of 1 N HCl and 10 ml of 1 M Tris-HCl, pH 7.6. The labeled cDNA was separated from unincorporated nucleotides by size exclusion chromatography on a 5-ml G-50 Sephadex (Pharmacia) column, and the radioactivity in 0.3 ml fractions was measured by scintillation counting. The fractions containing cDNA were pooled, and equal counts from the two reactions were used for hybridizations. Typically, the filters were hybridized with 1–3 3 106 cpm/ filter for at least 16 h at 428C.
Screening of CHM Library by Differential Hybridization
The CHM library was screened by differential hybridization. About 200 plaque-forming units per plate were grown on Luria broth medium by incubation for 16 h at 378C. Plaques were transferred in duplicate to nitrocellulose filters. The filters were washed sequentially with 0.5 M NaOH for 1 min, 0.5 M Tris-HCl (pH 7.5) for 5 min, and double-strength SSC for 5 min and were air dried. Finally, the filters were baked for 2 h at 808C and then washed with 6-strength SSC. The filters were incubated in 5 ml/filter of prehybridization solution for 4 h at 428C. Duplicate filters were hybridized with 32P-labeled cDNA made from purified poly(A)1 RNA from either CHM or NTP. The filters were washed as described above and subjected to autoradiography. RESULTS
Isolation and Characterization of CHM cDNA Clones
In order to obtain mRNA species predominantly expressed in CHM but not in NTP, we constructed a cDNA library from CHM. Replica filters of this library were screened with 32 P-labeled cDNA probes made from poly(A)1 RNA that had been isolated from either CHM or NTP. After screening of 2000 recombinants, those plaques that preferentially hybridized with the CHM probe were selected and purified by several screening steps. We focused on a CHM-specific clone exhibiting the strongest hybridization signal, named CHMS-1, with an insert of 1 kb. The tissue distribution of this clone was examined by Northern blot analysis using total RNA from CHM and NTP. Since CHM is usually evacuated at the end of the first
TC #
A NEW TNF-R-RELATED GENE IN TROPHOBLASTIC DISEASES
623
FIG. 1. Tissue specificity of the CHMS-1 clone. Total RNA was extracted from CHM, NTP, and NEP, run on a formaldehyde gel, and transferred to a nylon membrane filter. The filter was hybridized with the CHMS-1 clone as a probe (A) and was stained with ethidium bromide to examine the quality and the equivalence of RNA loads (B). Lengths of the bands were estimated (kb) on the basis of the mobilities of the rRNAs.
trimester of gestation, we also compared the expression pattern of CHMS-1 in NEP of the same gestational age. Using CHMS-1 clone as a probe in Northern blotting, a strong signal was detected as a single band of 4.3 kb in CHM RNA and a weak signal in NTP and NEP RNA (Fig. 1). Essentially the same results were obtained when different regions of the 936-base pair CHMS-1 clone were employed as probes in similar Northern blot assays, indicating that the CHMS-1 cDNA is actually expressed as a single transcript in CHM and related tissues. A similar pattern was observed using poly(A)1 RNA instead of total RNA. The CHMS-1 insert was PCR amplified using specific primers located in lgt10 arms and subcloned into a plasmid vector (pGEM-T). The 936 nucleotides were determined, and the sequence did not match other known genes and hence was considered as a new sequence not previously identified. However, the sequence analysis revealed a significant 62% homology with the TNF-R2 cDNA [17] over the 936 nucleotides, while the homology at the nucleotide level in the 39 noncoding region was greater than 65%. We found a single open reading frame in the 59 end of CHMS1 clone. The 57 amino acid deduced sequence showed no relevant homology with the C-terminal domain of TNF-R2, but surprisingly it had 37% similarity and 21% identity with
FIG. 3. Human CG regression curves. Postevacuation curves of three CHM cases; A and B represent patients with spontaneous hCG normalization, while C represents a patient with persistent GTT before start of chemotherapy.
the C-terminal domain of TNF-R1, where a functional region named death domain has been described [15]. Moreover, some critical residues that can abolish the apoptotic function in TNF-R1 when mutated [18] are conserved in the CHMS-1 predicted amino acid sequence (Fig. 2A). The identity level rose to 42% when the amino acid sequences of other death domain-bearing proteins involved in the signaling pathway of apoptosis, such as Fas, FADD (Fas-associated death domain) [19], TRADD (TNF-R1-associated death domain) [20], the death receptors DR3 [21] and DR5 [22], and reaper, were aligned together with those of TNFR1 and CHMS-1 using the ClustalX (EMBL, Heidelberg, Germany) multialignment program (Fig. 2B). Expression Profile of CHMS-1 and TNF-R Genes in Various CHMs and Human Term Placenta
To gain insights into the potential biological role of CHMS-1, its expression pattern was determined by Northern blot in CHM tissues that showed different hCG regression curves in the weeks after tumor evacuation. As shown in Figure 3, two representative cases of CHM, named A
FIG. 2. Comparison of the predicted amino acid sequence of CHMS-1 and death domains. CHMS-1 amino acid residues that are identical or similar are boxed in black or gray, respectively. A) CHMS-1 amino acid sequence was aligned with TNFR1 death domain. The following comparisons between residues are indicated: identical (*), weak similarity (.), and strong similarity (:). Amino acids in TNF-R1 that are crucial to its apoptotic function are also indicated (†). B) Multialignment of the CHMS-1 sequence and other death domain-bearing proteins. CHMS-1 amino acid residues identical or similar to one or several proteins are indicated.
TC #
624
DUMUR ET AL.
FIG. 4. CHMS-1 expression pattern in different CHM cases. Northern blot analysis of total RNA extracted from CHMs from cases A, B, and C and NTP, hybridized with CHMS-1, TNF-R1, TNF-R2, and GAPDH probes, as indicated by arrows.
and B, regressed to normal hCG values according to normal postevacuation regression curves [23], while case C showed a persistent rise or plateau in the hCG regression curve, indicating that the patient had developed persistent GTT and required chemotherapy. We observed a higher expression level of CHMS-1 transcript in cases A and B than in case C, where it was completely undetectable (Fig. 4). To establish a correlation between the expression pattern of CHMS-1 with that of the TNF receptor genes, the Northern blots were hybridized with radiolabeled specific probes for CHMS-1, TNF-R1, and TNF-R2. Both CHMS-1 and TNF-R1 transcripts were strongly expressed in CHMs A and B, while they were undetectable in case C; a weak expression was observed in NTP. In contrast, TNF-R2 expression in all the CHMs analyzed was almost identical and was stronger than in the NTP RNA sample (Fig. 4). These findings indicate that CHMS-1 and TNF-R1 displayed the same expression profile in the trophoblastic tissues analyzed. Expression Profile of CHMS-1 and TNF-R Genes in JEG3, a Human-Derived Choriocarcinoma Cell Line
The results described above indicate that the expression profiles of both CHMS-1 and TNF-R1 transcripts are associated with the most benign form of trophoblastic transformation. To verify these findings, we also investigated the gene expression pattern of CHMS-1 and TNF receptors in the most malignant trophoblast neoplasm, the choriocarcinoma. In this study we used the human choriocarcinomaderived cell line, JEG-3, as a source of total RNA for the Northern blot assays and total RNA extracted from HeLa and COS-7 as nontrophoblastic cell lines. The results of these assays are shown in Figure 5, where the CHMS-1 and TNF-R1 transcripts were clearly detected in human nontrophoblastic HeLa cells. TNF-R1 expression was also de-
FIG. 5. CHMS-1 expression pattern in trophoblastic and nontrophoblastic cell lines. Northern blot analysis of total RNA extracted from HeLa, COS-7, and JEG-3 cells, hybridized with CHMS-1 and TNF-R1 (A), TNFR2 (B), and GAPDH (C) probes.
tected in the nonhuman COS-7 cell line, albeit at lower amounts than in HeLa cells, whereas no CHMS-1 expression was observed in this cell line (Fig. 5A). Almost no CHMS-1 transcript was detected in the human-derived choriocarcinoma cell line JEG-3 (Fig. 5A), whereas TNF-R2 seems to be constitutively expressed in the three cell lines analyzed (Fig. 5B). These results, together with those shown in Figure 4, confirm that CHMS-1 and TNF-R1 mRNAs are preferentially expressed in the benign form of transformed trophoblast cells and are diminished in the more malignant forms such as persistent GTT and choriocarcinoma. DISCUSSION
In this study we describe a novel member of the TNF receptor superfamily that is highly expressed in CHM in comparison with NTP. By applying differential screening to a CHM-derived cDNA library, we obtained a cDNA clone encoding a gene predominantly expressed in this GTT. The tissue specificity of this CHM-specific clone, named CHMS-1, was confirmed by Northern blotting of CHM, NTP, and NEP samples. Analysis of the 1-kb insert of CHMS-1 did not reveal identity with sequences stored in the GenBank database. Nevertheless, a high homology level was found with members of the TNF receptor superfamily. Even more importantly, the amino acid deduced sequence showed 42% identity when multialigned with the death domain of several proteins, including the members of this superfamily, TNFR1 and Fas. Those proteins bearing the so-called death domain are responsible for the induction of apoptosis in the cells or tissues that overexpress them [15]. A new TNF receptor protein bearing such a death domain is not unex-
TC #
A NEW TNF-R-RELATED GENE IN TROPHOBLASTIC DISEASES
pected, since this domain is considered a single module able to fuse to different proteins through an exon-shuffling mechanism [24]. The specific expression of CHMS-1 gene in CHMs with normal hCG postevacuation curves, but not in CHM-derived persistent GTT or in the JEG-3 cell line, suggests that CHMS-1 may play a relevant biological role in molar trophoblastic tissue regression. Furthermore, there is a direct correlation between the expression of CHMS-1 gene and that of the TNF-R1 in different CHM cases. The TNF-R1 seems to play an essential role in triggering a complex network of signal transduction pathways that lead to apoptosis in some cells via activation of interleukin 1b-converting enzyme (ICE)-like proteases [25] or survival and proliferation of other cells via NF-kB activation [20]. These two distinct signaling pathways share a common factor named TRADD that mediates the interaction with FADD, leading to ICE-like protease activation. Additionally, TRADD interacts with TRAF2 (TNF receptor-associated factor 2) [26], a factor that is responsible for NF-kB activation [27]. The structural features of CHMS-1, and the expression pattern observed in the different trophoblastic tissues analyzed, strongly suggest that CHMS-1 may mediate both responses in a time-dependent manner. These functions could be exerted by CHMS-1 per se or in association with the death domain of TNF-R1, which is coexpressed in CHM that will not progress to any malignant process such as persistent GTT or choriocarcinoma. Thus, the CHMS-1 death domain may be responsible for cell proliferation at the first stages of molar development, due to NF-kB activation, leading to the tumorigenic aspect of CHM. In addition, CHMS-1 may also be involved in triggering the sudden regression observed in CHM toward the end of the first trimester of gestation, as well as in the remaining molar tissue in the weeks following tumor evacuation. Such a phenomenon is likely to involve a complex network of proteins whose actions may induce the apoptotic program in CHM cells. Accordingly, the high levels of normal p53 protein found in CHM [7] may also explain the normal molar tissue regression by an apoptotic mechanism. Finally, the identification of proteins that interact with a potential CHMS-1 death domain will improve understanding of the physiological role of such a molecule in CHM and perhaps in other malignancies. ACKNOWLEDGMENTS We are grateful to Dr. Bruno Chatton for discussions and critical reading of the manuscript and to Dr. Eduardo Artz for generously providing the TNF receptor cDNA probes. C.I.D. and N.P.K. thank CONICOR, SECyT, and CONICET for the fellowships granted.
REFERENCES 1. Szulman AE, Surti U. The syndromes of hydatidiform mole. II. Morphologic evolution of the complete and partial mole. Am J Obstet Gynecol 1978; 20:20–27. 2. Kajii T, Ohama K. Androgenetic origin of hydatidiform mole. Nature 1977; 268:633–634. 3. Wake N, Takagi N, Sasaki M. Androgenesis as a cause of hydatidiform mole. J Natl Cancer Inst 1978; 60:51–57. 4. Berkowitz RS, Goldstein DP. Gestational trophoblastic disease. Cancer Suppl 1981; 76:2079–2085.
625
5. Genti-Raimondi S, Alvarez CI, Patrito LC, Flury A. Low aromatase activity in microsomes from complete hydatidiform mole. J Clin Endocrinol Metab 1993; 76:108–111. 6. Bocco JL, Panzetta G, Flury A, Patrito LC. Expression of pregnancy specific b1-glycoprotein gene in human placenta and hydatidiform mole. Biochem Int 1989; 18:999–1008. 7. Park JS, Namkoong SE, Lee HY, Kim SJ, Hong KJ, Kim IS, Kim KU, Shim BS. Expression and amplification of cellular oncogenes in human developing placenta and neoplastic trophoblastic tissue. Asia Oceania J Obstet Gynecol 1992; 18:57–64. 8. Cheung AN, Srivastava G, Chung LP, Ngan HY, Man TK, Liu YT, Chen WZ, Collins RJ, Wong LC, Ma HK. Expression of the p53 gene in trophoblastic cells in hydatidiform moles and normal placentas. J Reprod Med 1994; 39:223–227. 9. Yasuda M, Kawai K, Serizawa A, Tang X, Osamura RY. Immunohistochemical analysis of expression of p53 protein in normal placentas and trophoblastic diseases. Appl Immunohistochem 1995; 3:132–136. 10. Ladines-Llave CA, Maruo T, Manalo AM, Mochizuki M. Decreased expression of epidermal growth factor and its receptor in the malignant transformation of trophoblasts. Cancer 1993; 71:4118–4123. 11. Loetscher H, Pan Y-C, Lahm H-W, Gentz R, Brockhaus M, Tabuchi H, Lesslauer W. Molecular cloning and expression of the human 55 kd tumor necrosis factor receptor. Cell 1990; 61:351–359. 12. Nagata S, Golstein P. The Fas death factor. Science 1995; 267:1449– 1456. 13. White K, Grether ME, Abrams JM, Young L, Farrel K, Steller H. Genetic control of programmed cell death in Drosophila. Science 1994; 264:677–683. 14. Golstein P, Marguet D, Depraetere V. Homology between reaper and the cell death domains of Fas and TNFR1. Cell 1995; 81:185–186. 15. Itoh N, Nagata S. A novel protein domain required for apoptosis. J Biol Chem 1993; 268:10932–10937. 16. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 1984; 162:156–159. 17. Smith CA, Davis T, Anderson D, Solam L, Beckmann MP, Jerzy R, Dower SK, Cosman D, Goodwin RG. A receptor for tumor necrosis factor defines an unusual family of cellular and viral proteins. Science 1990; 248:1019–1023. 18. Tartaglia LA, Ayres TM, Wong GHW, Goeddel DV. A novel domain within the 55 kd TNF receptor signals cell death. Cell 1993; 74:845– 853. 19. Boldin MP, Varfolomeev EE, Pancer Z, Mett IL, Camonish JH, Wallach D. A novel protein that interacts with the death domain of Fas/ APO1 contains a sequence motif related to the death domain. J Biol Chem 1995; 270:7795–7798. 20. Hsu H, Xiong J, Goeddel DV. The TNF receptor 1-associated protein TRADD signals cell death and NF-kB activation. Cell 1995; 81:495– 504. 21. Chinnaiyan AM, O’Rourke K, Lyons RH, Garg M, Duan R, Xing L, Gentz R, Ni J, Dixit VM. Signal transduction by DR3, a death domain-containing receptor related to TNF-R1 and CD95. Science 1996; 274:990–992. 22. Pan G, Ni J, Wei Y-F, Yu G-L, Gentz R, Dixit VM. An antagonist decoy receptor and a death domain-containing receptor for TRAIL. Science 1997; 277:815–818. 23. Yedema KA, Verheijen RH, Kenemans P, Schijf CP, Borm GF, Segers MF, Thomas CM. Identification of patients with persistent trophoblastic disease by means of a normal human chorionic gonadotropin regression curve. Am J Obstet Gynecol 1993; 168:787–792. 24. Golstein P, Marguet D, Depraetere V. Fas bridging cell death and cytotoxicity: the reaper connection. Immunol Rev 1995; 146:45–56. 25. Boldin MP, Goncharov TM, Goltsev YV, Wallach D. Involvement of MACH, a novel MORT1/FADD-interacting protease, in Fas/APO1and TNF receptor-induced cell death. Cell 1996; 85:803–815. 26. Hsu H, Shu H-B, Pan M-G, Goeddel DV. TRADD-TRAF2 and TRADD-FADD interactions define two distinct TNF receptor 1 signal transduction pathways. Cell 1996; 84:299–308. 27. Rothe M, Sarma V, Dixit VM, Goeddel DV. TRAF2-mediated activation of NF-kB by TNF receptor 2 and CD40. Science 1995; 269: 1424–1427.
