Abstract
N-Acetylated α-linked acidic dipeptidase (NAALADase) is a neuropeptidase that may modulate glutamatergic neurotransmission. Independent of its characterization in the nervous system, one form of NAALADase was shown to be expressed at high levels in human prostatic adenocarcinomas, and it was designated the prostate-specific membrane antigen (PSMA). The NAALADase/PSMA gene is known to produce multiple mRNA splice forms, and based on previous immunohistochemical evidence, it had been assumed that the human brain and prostate expressed different isoforms of the enzyme. Because PSMA is being actively pursued as a target for autoimmune and cytotoxic targeting strategies to treat prostate cancer, the rigorous comparison of the two forms of the enzyme remained an important but untested question. To assess similarities and/or differences between human brain NAALADase and PSMA, we compared the two molecules using criteria of activity, immunoreactivity and sequences of the corresponding mRNAs. NAALADase from human cerebellar isolates displayed a kinetic profile and pharmacological sensitivities similar to PSMA. Also, Northern hybridization to PSMA cDNA detected indistinguishable sets of 2.8-, 4.0- and 6.0-kb RNA species in human brain and the LNCaP prostatic tumor cell line. In addition, the monoclonal antibody 7E11-C5 directed against the prostatic form of the enzyme immunoprecipitated 82% of human cerebellar NAALADase activity. Moreover, reverse transcription-polymerase chain reaction cloning of cerebellar cDNAs indicated that the human brain and prostate express a common mRNA splice form. Therefore, we conclude that the form of NAALADase also known as PSMA is expressed in brain and comprises a significant fraction of brain NAALADase activity.
NAALADase and its neuropeptide substrate NAAG have been implicated in the regulation of excitatory signaling in the nervous system, and alterations in levels of NAALADase and NAAG have been observed in disorders that are thought to involve abnormalities in glutamatergic neurotransmission (Tsai et al., 1991, 1995; Meyerhoffet al., 1992). NAALADase activity and protein have also been detected in non-neural tissues, including the kidney and sexual organs (Slusher et al., 1990). NAALADase (EC 3.4.17.21) possesses glutamate carboxypeptidase activity in vitro against NAAG, α-aspartylglutamate, α-glutamylglutamate, γ-glutamylglutamate, and methotrexate poly-γ-glutamates (Robinson et al., 1987;Serval et al., 1990; Slusher et al., 1990; Pintoet al., 1996). In vivo hydrolysis of NAAG by NAALADase in brain has also been demonstrated (Stauch et al., 1989; Serval et al., 1992).
The molecular characterization of the enzyme or enzymes responsible for NAALADase activity has been of considerable interest. The cDNA for one form of human NAALADase, also known as PSMA (Israeli et al., 1993), has been isolated from human prostatic tumor cells (Carteret al., 1996). The precise molecular relationship between PSMA and the human brain enzyme, however, has remained unclear. The PSMA antigen was identified originally as the ligand of the monoclonal antibody 7E11-C5, whose histological profile of immunoreactivity demonstrated a high degree of selectivity for prostatic tissue (Horoszewicz et al., 1987). Thus, PSMA has been proposed as a prostate cancer marker antigen and, accordingly, is being examined for possible use in the diagnosis of prostatic malignancies and detection and treatment of metastatic prostatic disease. Indium 111 and yttrium 90 radioimmunoconjugate derivatives of the monoclonal antibody 7E11-C5 (CYT-356) are currently in clinical trials for radiologic imaging and irradication of prostate-derived tumor cells, respectively (Kahn et al., 1994; Deb et al., 1996). In addition, clinical trials of PSMA-targeted dendritic cell immunotherapies are under way (Tjoa et al., 1997). The diagnostic and prognostic values of PSMA-based immunoassay and RT-PCR assay also are under evaluation (Heston, 1995; Murphy et al., 1996).
Studies assessing the presence of PSMA-like species in the brain have yielded varying results. The absence of PSMA-like immunoreactivity in human brain sections has been reported in two studies (Horoszewiczet al., 1987; Silver et al., 1997). In contrast, the presence of PSMA-like RNAs in brain has been detected by Northern analysis (Carter et al., 1996), RNase protection assay (Israeli et al., 1994) and RT-PCR (Urı́a et al., 1997). The molecular characterization of the human brain NAALADase enzyme is important for two distinct reasons. First, establishing the molecular identity of human brain NAALADase is paramount to the exploration of its relationship to glutamatergic neurotransmission. Second, knowledge of the extent of similarity between the brain and prostate forms of the enzyme would be important in addressing the possible neurological ramifications of using PSMA-targeted imaging and cytotoxic agents.
Because the importance of its molecular description had become clear, we designed a strategy to characterize human brain NAALADase. Based on the results of previous RNA analyses, we hypothesized that PSMA and brain NAALADase would possess molecular similarities. Thus, we characterized the NAALADase expressed in human cerebellum using molecular probes directed against PSMA. This has resulted in the detection of a PSMA-like NAALADase protein and the isolation of a human brain cDNA, which confirm that PSMA is indeed expressed in the human nervous system.
Materials and Methods
Chemicals.
General chemical reagents were obtained from Fisher Scientific (Pittsburgh, PA) or Sigma Chemical (St. Louis, MO). Dr. Barbara Slusher (Guilford Pharmaceuticals, Baltimore, MD) kindly provided the NAALADase inhibitor PMPA (Jackson et al., 1996).
Cell lines and culture medium.
The LNCaP and PC3 tumor cell lines were obtained from the American Type Culture Collection (Rockville, MD). LNCaP cells were grown in RPMI 1640 supplemented with 2 mM glutamine, nonessential amino acids, penicillin/streptomycin (50 units/50 μg/ml) and 5% fetal bovine serum. PC3 cells were cultured in minimum essential medium supplemented with penicillin/streptomycin and 10% fetal bovine serum. All medium reagents were obtained from GIBCO BRL (Bethesda, MD).
Human brain tissue.
Control human brain tissue (case numbers 3262 and 3547) was obtained from the McLean Hospital Brain Tissue Resource Center (Belmont, MA). Brain membranes were prepared for immunoprecipitation studies and NAALADase assays as described previously (Tsai et al., 1995). Protein concentration of the solubilized samples was determined by BCA assay using the enhanced protocol with bovine serum albumin as standard (Pierce, Rockford, IL).
RNA.
Total RNA from the LNCaP cell line was prepared according to the method of Chirgwin et al. (1979), and human brain poly(A)+ RNA was obtained from Clonetech (Palo Alto, CA).
Northern blotting.
RNA was resolved by electrophoresis through a 1.2% agarose gel containing 3% formaldehyde, electrophoretically transferred to a nylon membrane and hybridized to a random-primed32P-radiolabeled cDNA probe (S.A. = 6.0 × 109 dpm/μg) prepared using a Prime-It kit (Stratagene, La Jolla, CA) at 42°C overnight. Final high-stringency washes were performed with 0.1× SSC plus 0.1% SDS at 65°C. Hybridization was detected by autoradiography using a Molecular Dynamics PhosphorImager (Sunnyvale, CA).
Enzyme assays.
Radioenzymatic assays, which measured the release of [3H]glutamate from N-acetylaspartyl-[3H]glutamate, were conducted in triplicate as described by Tsai et al. (1995) and included protein-free blanks as negative controls. N-Acetylaspartyl-[3H]glutamate was purchased from DuPont-New England Nuclear (Boston, MA). NAAG hydrolysis was allowed to proceed for a duration of 2 hr to maximize signal. This did not affect the linearity of the assay; no more than 20% of the substrate was consumed during this period. Assays comparing human brain and cloned enzyme samples were normalized for NAALADase content by enzyme activity [Vmax of NAAG hydrolysis = 175 ± 25 fmol/min as determined by preliminary experiments]; total protein in the assay samples was 2.5 μg (PSMA2), 15 μg (3242) or 25 μg (3547)). Results shown are the mean ± S.E.M. activities of triplicate assays. Curve fit was determined by a nonlinear least-squares minimization algorithm.
Immunoprecipitation experiments.
After initial titration experiments to determine optimal precipitation, 30 μg of purified 7E11-C5 monoclonal antibody (a gift of Drs. Gerald Murphy and Alton Boynton, Northwest Hospital, Seattle, WA) was incubated with 100 mg of Protein A/trisacryl (Pierce) at 4°C overnight in 300 μl of 25 mM Tris·Cl, 0.9% NaCl and 0.25% Triton X-100, pH 7.0 (TTS). The Protein A-resin/immunoglobulin complexes were then precipitated by centrifugation (400 × g for 2 min). Subsequently, the supernatants were removed, and the pellets were washed three times with 1 ml of TTS. The resultant pellets were then incubated overnight at 4°C with 12.5 μg of Triton X-100-solubilized human brain membranes that had been previously centrifuged for 5 min at 400 ×g to remove insoluble material. Pellets were separated from the supernatant fraction of the sample (100 μl) by centrifugation at 400 × g for 2 min, and the supernatants were saved for enzymatic assay. The pellets were then washed twice with 100 μl of TTS, and these washes were saved and assayed. Final pellet fractions were resuspended in 100 μl of TTS and assayed in parallel. A series of control samples, which contained the brain membranes and Protein A/trisacryl, but no antibody, distinguished specific immunoprecipitation from nonspecific binding. NAAG hydrolysis by the first supernatant and two wash fractions were added together and considered the total supernatant activity. Negative and positive control samples included 100 μl of the immunoprecipitation buffer (TTS).
RT-PCR cloning of brain cDNAs.
RT reactions were conducted at 47° to 50°C for 2 hr using Superscript II reverse transcriptase (BRL) according to the manufacturer’s recommended conditions with the addition of 40 units/25 μl of recombinant RNasin (Promega, Madison, WI) and 3.33 mM dimethylsulfoxide in the RNA denaturation step (equivalent to 2 mM final concentration in the transcription reaction) and a PSMA-derived oligonucleotide primer of the sequence: TTTATATATAATATTCAAC. The reverse transcriptase then was heat-inactivated by incubation at 70°C for 15 min, and RNA-cDNA hybrids were treated with RNase H (Promega) for 10 min at 55°C. Single-stranded cDNA was isolated with PCR Purification Cartridges (Advanced Genetic Technologies, Gaithersburg, MD) before use in the subsequent PCR amplification. Primers used for the amplification of the 5′ segment of the cDNA were of the sequences TGCAGGGCTGATAAGCGAG and CCTCTGAGAGTACCTATCACA. The amplification of the 3′ segment of the cDNA was performed with primers of the sequences TCATCCAATTGGATACTATG and TCTTTCTGAGTGACATAC. Thermal cycling parameters consisted of: denaturation at 94°C for 5 min (initial) or 1 min (subsequent); annealing at 60°C (5′ cDNA) or 50°C (3′ cDNA) for 1 min; and extension at 72°C for 3 min. Each round of amplification was comprised of 35 cycles. DNA sequence analysis was performed by the dideoxy method using the Pfu (exo-) Cyclist system (Stratagene) according to the manufacturer’s instructions.
Results
Kinetic and pharmacological comparisons of NAAG hydrolysis by cloned and native human brain NAALADases.
To begin our characterization of the human brain enzyme, samples of human cerebellar membranes from two control cases were assayed for NAAG-hydrolytic activity in parallel with lysates of PSMA-transfected PC3 cells. A substrate-velocity relationship was established for each of the three samples, as shown in figure 1 and table1. Km values for the brain samples were similar to cloned PSMA (87 and 100 nMvs. 146 nM) and the NAALADase activity measured previously in the LNCaP prostate cancer cell line (65 nM; Carter et al., 1996). To assess further the similarities of the brain and cloned prostate forms of NAALADase, their relative sensitivities to the NAALADase inhibitors quisqualic acid (Robinson et al., 1987), β-NAAG (Serval et al., 1990) and 2-(phosphonomethyl)pentanedioic acid (Jackson et al., 1996) were examined. As shown in figure 2, brain and transfected cell samples displayed comparable concentration-dependent profiles of sensitivity to each of the three compounds with >95% inhibition of NAAG hydrolysis observed at the highest inhibitor concentrations.
Northern analyses of human brain RNAs.
To compare the RNA species expressed in the human brain with those observed in human prostatic cells (LNCaP cell line), poly(A)+ RNA from human brain was hybridized to the NAALADase/PSMA cDNA. As shown in figure 3, similar sets of three RNA bands were observed in all brain and prostatic cell samples. The predominant of these, a 2.8-kb species, is reported to contain the NAALADase/PSMA mRNA represented by the cloned PSMA cDNA (Israeli et al., 1993). Also, the relative amounts of NAALADase-like RNA detected among LNCaP, whole brain, and two region-specific brain isolates corresponded to their respective levels of NAALADase activity: LNCaP > hippocampus > whole brain > cerebellum (Carter et al., 1996; Passani et al., 1997).
Immunoprecipitation of NAALADase activity from brain homogenates using anti-PSMA antibody.
Having detected an mRNA in human brain that was approximately the same size as and similar in sequence to PSMA, we explored the detection of the presence of NAALADase protein using the anti-PSMA antibody 7E11-C5. To determine what fraction of total brain NAALADase activity might be composed by PSMA-like immunoreactive species, we assessed the extent of immunoprecipitation of activity by 7E11-C5/Protein A/trisacryl complexes. This involved solubilization of cerebellar membranes with Triton X-100 and subsequent examination of the distribution of NAALADase activity in the supernatant vs. pellet fractions after immunoprecipitation. As shown in figure 4, 82% of NAALADase was removed from the supernatant fraction of the cerebellar membrane isolates via binding to the 7E11-C5 antibody. Control supernatant fractions from which the antibody was omitted retained full activity. Furthermore, NAALADase activity was recovered in the pellets of 7E11-C5-containing samples but not in the antibody-free controls.
Cloning of human brain NAALADase cDNAs.
To characterize the brain NAALADase species in greater molecular detail, we analyzed brain mRNA [poly(A)+-RNA] using RT-PCR. Two primer sets, designed to amplify the 5′ and 3′ coding regions of the PSMA cDNA, respectively, were tested for their ability to produce specific RT-PCR amplicons. Parallel RT-PCR reactions in which only the brain RNA was omitted served as negative controls. As illustrated in figure5, two specific products of the expected sizes were amplified by this procedure (one from each primer set). Because the NAALADase/PSMA mRNA is derived from multiple exons (see Discussion), our RT-PCR amplification of these cDNA sequences from brain was not due to chromosomal DNA contamination (the amplification of which would yield much larger products). Subsequent subcloning and sequencing of these cDNAs revealed that they are identical in sequence to the previously cloned PSMA species. These results indicate that human brain NAALADase is composed, at least in part, of a polypeptide identical to PSMA.
Discussion
PSMA is expressed in human brain.
Immunohistochemical analyses with the 7E11-C5 antibody indicated that the expression of PSMA was limited to prostatic and a few other non-neural tissues (Horoszewiczet al., 1987, Silver et al., 1997). These studies implied that human brain NAALADase would be composed of a different polypeptide, which would not be 7E11-C5 reactive. In contrast, molecular studies have provided considerable evidence of similarities between the two species (Israeli et al., 1994; Troyeret al., 1995; Carter et al., 1996, Urı́aet al., 1997). Previous interpretations of these disparate results led to the hypothesis that a different splice isoform of the enzyme might be expressed in brain. In the present study, the immunoprecipitation and the RT-PCR data clearly indicate that the PSMA polypeptide is expressed in human brain and that the majority of human cerebellar NAALADase is 7E11-C5 immunoreactive. The discovery that PSMA is expressed in brain should not curtail efforts to use PSMA as a prostatic cancer marker and targeting functionality. However, this identity does suggest that the design of therapeutic agents that use PSMA to target cytotoxins or to induce autoimmune responses should take into account any possible adverse effects to the nervous system. An autoimmune therapy for prostate cancer using autologous PSMA peptide-stimulated dendritic cells has thus far shown to be efficacious and without adverse neurological consequences (Tjoa et al., 1997).
Molecular characteristics of human brain NAALADase.
By virtue of its newly discovered identity to PSMA, structural information on the human brain NAALADase molecule can be obtained from previous analyses of the prostatic form of the enzyme. The primary amino acid sequence of NAALADase predicts an 84-kDa polypeptide, and Western blots and glycosylation consensus motifs predict a native glycoprotein of ∼100 kDa in mass (Israeli et al., 1993; Troyer et al., 1995). NAALADase demonstrates sequence homology to the M28 metallopeptidase family (Rawlings and Barrett, 1997) and thus is inferred to be a cocatalytic zinc peptidase. Homology to the M28 family has also allowed predictions of the overall structure and functional groups of NAALADase within its active site (Rawlings and Barrett, 1997).
The identity of brain and prostatic NAALADases also suggests that their expression arises from a common gene. Two NAALADase-related human gene loci have been identified on chromosome 11 (11p11.1–13/11cen-p12 and 11q14; Leek et al., 1995; Rinker-Schaeffer et al., 1995). Partial sequence information from these loci indicates that they are composed of one NAALADase/PSMA gene, containing at least 19 exons (GenBank accession nos. AF007411-AF007413, AF007544, AF011896,AF016826-AF016830), and one pseudogene (GenBank accession nos. U93598,U93599) or related peptidase gene (Shneider et al., 1997). The large number of exons in the NAALADase/PSMA gene indicates a potential for multiple splice isoforms. Northern analyses in the present study show the expression of three populations of RNA (as differentiated by electrophoretic migration), the 2.8-kb PSMA mRNA and two higher-molecular-weight bands of hybridization. Whether the 4.0 and 6.0 species on the blot represent incompletely processed hnRNAs or differential splice forms or give rise to different protein species remains to be determined. It is also possible that each band of hybridization seen in the present study is composed of more than one unique RNA (of similar molecular weights). Limited information is available at present on the diversity of products that arise from the NAALADase/PSMA gene. 7E11-C5-immunoreactive proteins larger than PSMA have been observed in Western blot experiments, but they are reported to be dimers (Troyer et al., 1995). A smaller PSMA-like protein species and two RNA splice heterogeneities have also been described (Su et al., 1995; Murphy et al., 1996;Bzdega et al., 1997). The cytosolic splice isoform PSM′ appears to be the predominant gene product in normal prostate (Suet al., 1995; Murphy et al., 1996), but it is probably not expressed in brain because brain NAALADase activity appears to be entirely membrane bound (Robinson et al., 1987; Slusher et al., 1990). The splice isoform identified by Bzdega et al. (1997) is expressed in brain, but the activity encoded by this splice form has not been characterized to date.
The role of NAALADase in the nervous system.
Regardless of its use as a prostate cancer marker, NAALADase is an interesting target for pharmacotherapeutic intervention in the nervous system. NAALADase is believed to regulate excitatory neurotransmission by controlling the bioavailability of its abundant neuropeptide substrate NAAG and the peptide’s metabolite, glutamate. NAAG can negatively modulate excitatory signaling via activation of inhibitory mGluR3 receptors (Wroblewska et al., 1997) and possibly through inhibition of NMDA receptor channels, at which it is a weak agonist (Sekiguchi et al., 1989; Trombley and Westbrook, 1990;Puttfarcken et al., 1993; Burlina et al., 1994). Metabolism of the peptide by NAALADase yields glutamate itself, an agonist at all glutamatergic receptor subtypes, which conveys predominantly excitatory effects (reviewed in Hollmann and Heinemann, 1994). It is believed that NAALADase activity tips the balance of “NAAGergic” effects toward glutamate generation and thus, presumably, excitation. Consistent with this hypothesis, recent experiments in animal model systems show that NAALADase inhibitors are neuroprotective against excitatory insult (Orlando et al., 1997; Lu et al., 1997), and alterations in NAALADase activity that are consistent with inferred changes in glutamatergic neurotransmission have been observed in patients with amyotrophic lateral sclerosis (Tsai et al., 1991) and schizophrenia (Tsai et al., 1995) and in animal models of epilepsy (Meyerhoff et al., 1992).
To elucidate further the specific role of NAALADase in glutamatergic signaling, it is desirable to be able to study and genetically manipulate the expression of the peptidase. The rat brain enzyme has been extensively characterized by its purification, its molecular cloning and the detection of its expression in astrocytes and nonmyelinating Schwann cells (Slusher et al., 1990; Bergeret al., 1995; Bzdega et al., 1997; Luthi-Carteret al., 1998). Studies examining the effects of up- and down-regulation of NAALADase in rodent systems are under way. By facilitating future explorations of its expression in the human nervous system, the molecular characterization of human brain NAALADase should broaden our knowledge of the possible role of the enzyme in states of human disease.
Conclusions.
By demonstrating the identity between human brain NAALADase and PSMA, we hope to promote the refinement of PSMA-directed therapeutic strategies so as to minimize adverse consequences to the nervous system. The identity of the brain and prostatic tumor NAALADases also raises intriguing questions about the possibly similar functions of the enzyme in neurotransmission and carcinogenesis. Moreover, molecular information about the human brain enzyme may facilitate the development of NAALADase inhibitors as neuroprotective agents.
Acknowledgments
The authors are grateful for the advice and assistance of our colleagues Drs. Rachael Neve, Christine Konradi, Lucius Passani, Guochuan Tsai and Michael Leski. We are also appreciative of gifts of the NAALADase inhibitor PMPA from Dr. Barbara Slusher and of the 7E11-C5 antibody from Drs. Gerald Murphy and Alton Boynton. In addition, we thank Drs. Boynton, W.D.W. Heston and George Wright for helpful discussions about the expression and detection of PSMA.
Footnotes
-
Send reprint requests to: Dr. Joseph T. Coyle, McLean Hospital, 115 Mill Street, Belmont, MA 02178. E-mail:jcoyle{at}warren.med.harvard.edu
-
↵1 This study was supported by National Institutes of Health Grant MH-572901 and a National Alliance for Research on Schizophrenia and Depression Senior Investigator Award to J.T.C. The McLean Brain Tissue Resourse Center is supported in part by United States Public Health Service Grant MH/NS-31862.
-
↵2 Graduate Program in Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, 725 N. Wolfe St., Baltimore, MD 21205.
- Abbreviations:
- NAAG
- N-acetyl-α-aspartylglutamate
- β-NAAG
- N-acetyl-β-aspartylglutamate
- NAALADase
- N-acetylated α-linked acidic dipeptidase
- PMPA
- 2-(phosphonomethyl)pentanedioic acid
- PSMA
- prostate-specific membrane antigen
- RT
- reverse transcription
- PCR
- polymerase chain reaction
- Received January 7, 1998.
- Accepted April 20, 1998.
- The American Society for Pharmacology and Experimental Therapeutics