Proliferative actions of natriuretic peptides on neuroblastoma cells. Involvement of guanylyl cyclase and non-guanylyl cyclase pathways.

To identify neural tumor cell lines that could be used as models to study growth-related natriuretic peptide actions, we determined the effects of these peptides on the proliferation of human and rodent neuroblastoma cell lines. Subnanomolar concentrations of atrial natriuretic peptide (ANP) and type C natriuretic peptide (CNP) stimulated proliferation in all four cell lines. These actions were associated with cGMP elevation and were blocked by a protein kinase G inhibitor. These data imply the involvement of guanylyl cyclase (GC)-coupled natriuretic receptors. However, higher concentrations of ANP and CNP, and low concentrations of des-[Gln(18),Ser(19),Gly(20),Leu(21),Gly(22)]-ANP(4-23)-NH(2) (desANP(4-23)) (analog for NPR-C receptor) exerted antiproliferative actions in three of the cell lines. These effects were insensitive to a protein kinase G inhibitor and to HS-142-1, suggesting that growth-inhibitory actions involved a non-GC receptor. They did not appear to involve cAMP, protein kinase A, protein kinase C, or calcium mobilization but were abolished when constitutive mitogen-activated protein kinase activity was inhibited. Radioligand binding experiments revealed the presence of a uniform class of binding sites in NG108 cells and multiple binding sites in Neuro2a cells. Northern and reverse transcriptase-polymerase chain reaction analyses revealed differential gene expression for NPR-A/B/C in NG108 and Neuro2a cells. The results indicate that natriuretic peptides stimulate neuroblastoma cell proliferation through type NPR-A/B (GC) receptors. Higher concentrations of ANP and CNP exerted a mitogen-activated protein kinase-dependent antiproliferative action mediated by a non-GC receptor that interacts with desANP(4-23) with relatively high affinity.

Natriuretic peptides constitute a family of structurally related hormones that includes atrial natriuretic peptide (ANP), 1 brain natriuretic peptide, and the type C natriuretic peptide (CNP) (1,2). These peptides interact with overlapping specificity on three known receptors (3,4). Receptor subtypes A and B (NPR-A and NPR-B, respectively) contain a single transmembrane domain and possess intrinsic GC activity (5,6). The type C receptor (NPR-C) is similar to types A and B but lacks the intracellular GC domain (7). Despite widespread expression of NPR-C, initial studies were unable to identify its signaling pathways. Because NPR-C was found to internalize after binding natriuretic peptides, it was proposed that the receptor is involved in removing ANP peptides from the circulation (8). It has thus been commonly referred to as the "clearance" receptor. However, more recent studies that have utilized relatively specific ligands for NPR-C suggest that it may be positively coupled to phospholipase C (9) and adenylyl cyclase (10) or negatively coupled to adenylyl cyclase (reviewed in Ref. 4) and MAP kinase pathways (11).
Natriuretic peptides have been shown to inhibit the proliferation of several cell types, including vascular smooth muscle cells (12), kidney mesangial cells (13)(14)(15), chondrocytes (16), osteoblast-like cells (17), and hepatoblastoma cells (18). In addition, a recent report indicated that the NPR-C-mediated inhibition of astrocyte proliferation occurs in association with decreased MAP kinase activation (11). In contrast to these inhibitory actions, ANP stimulated the proliferation of embryonic cardiomyocytes (19), and both CNP and brain natriuretic peptide were found to stimulate longitudinal bone growth in vitro (20). Moreover, NPR-C knockout mice and transgenic mice overexpressing brain natriuretic peptide exhibit pronounced skeletal overgrowth (21,22). Data suggest that elevation of cGMP may be involved in some of type A and B receptormediated proliferative responses (13,15,17,19), possibly through regulation of the MAP kinase-selective phosphatase MPK-1 (23) and/or phosphorylation of platelet-derived growth factor receptor (24).
Natriuretic binding sites are reported to be present in the proliferative zones of the embryonic brain (25) and in neuroblastoma tumor cell lines (26). Neuroblastoma cells have also been reported to express natriuretic peptides (27) and to secrete certain endopeptidases (28) that are capable of cleaving natriuretic peptides. Despite these findings, relevant biological actions of natriuretic peptides in embryonic neuroblasts and neuroblastoma cells have not been elucidated. Here, we investigated the hypothesis that natriuretic peptides regulate the proliferation of neuroblastoma cells through specific ANP receptor subtypes. These studies utilized ANP, which binds with high affinity to NPR-A and NPR-C; CNP, which binds with highest affinity to NPR-B; and an NPR-C-specific angonist des-[Gln 18 ,Ser 19 ,Gly 20 ,Leu 21 ,Gly 22 ]-ANP 4 -23 -NH 2 (desANP 4 -23 ) (29). We also studied the sensitivity of the proliferative actions to various kinase inhibitors and characterized at molecular and pharmacological levels the type of ANP receptors that were expressed. Finally, we measured the peptide-induced changes in the levels of cyclic second messengers, cAMP and cGMP, by radioimmunoassay and intracellular calcium by fura-2 cytofluorometry. The data obtained indicate that natriuretic peptides stimulate proliferation through NPR-A and/or NPR-B (GC) receptors, whereas antiproliferative actions appear to be mediated by a non-GC natriuretic receptor that interacts with desANP 4 -23 with relatively high affinity.
Proliferation Studies-Cells were seeded in 24-well plates (80,000 cells/well) in 1 ml of medium and cultured for 24 h. Medium was then replaced with fresh serum-free medium. After 1.5 h at 37°C, cells were preincubated for an additional 1 h with vehicle or signal transduction inhibitors (GF109203X, H89, and Rp-8-pCTP-cGMPS from Calbiochem and PD98059 from New England Biolabs) or the NPR-A/NPR-B antagonist HS-142-1 (kindly provided by Kyowa Hakko Kogyo Co., Shizuoka, Japan). Inhibitor efficiencies and specificities were tested in preliminary studies across a range of concentrations (from 1 to 100 M) using stimulators of the specific kinases (PACAP for protein kinase A, platelet-derived growth factor for mitogen-activated protein kinase, and ANP for G-kinase). The lowest doses causing full but specific inhibitions were used in subsequent assays. After preincubation with the specified inhibitors, peptides (desANP 4 -23 , ANP, and CNP from Sigma) were added for 1 h prior to distribution of [ 3 H]thymidine (1 Ci/well). Four hours after radiotracer addition, cells were harvested, and [ 3 H]thymidine incorporation was determined as previously described (30).
Cyclic AMP and GMP Measurements-Cells (80,000/well) were cultured for 3 days in 24-well plates. Medium was replaced with serumfree medium containing 0.1 mM isobutylmethylxanthine and, in the case of cAMP measurement, with or without 10 M forskolin (Sigma). After incubation for 15 min at 37°C, peptides were added and cells were incubated further for 15 min at 37°C. Cells were lysed in 6% trichloroacetic acid solution, and radioimmunoassays were performed according to the manufacturer's protocol (PerkinElmer Life Sciences).
Binding Studies-125 I-ANP (2200 Ci/mmol) was obtained from PerkinElmer Life Sciences. Time course and saturation studies were performed in six-well plates (Falcon) seeded 24 h earlier with 400,000 cells/well. Displacement experiments were conducted on freshly suspended cells in microcentrifuge tubes (800,000 cells/ml). Incubations in all cases were performed in binding medium consisting of culture medium (Dulbecco's modified Eagle's medium) supplemented with bovine serum albumin, HEPES, and protease inhibitors (o-phenanthroline and bacitracin). Saturation studies were performed in binding medium containing increasing concentrations (from 1 pM to 0.15 nM) of radioligand in the presence or absence of 5 M native ANP (to determine nonspecific and total binding, respectively). In competition and kinetic studies, binding medium contained 36 pM 125 I-ANP and the specified concentrations of competitive analogs. Incubation time for competition and saturation experiments was fixed at 120 min based on the time course study. All incubations were performed at 4°C under gentle agitation. Rinses were performed, and bound radioactivity measurements were determined as previously described (30).
Purification of mRNA-Neuro2a and NG108 cells were cultured in seven flasks (75 cm 2 ) and harvested at 90% confluence with 0.05% trypsin, 0.02% EDTA. Trypsin was neutralized with the tissue culture medium (containing 8% serum), and cells were centrifuged and then rinsed with phosphate-buffered saline. Total RNA from control tissues (brain, kidney, and lung) was obtained from two male ND4 mice. Total RNA from cells and tissues was extracted and purified according to the method of Chomczynski and Sacchi (31). Subsequent poly(A)-selection of mRNA was accomplished using the Poly(A)Pure™ kit (Ambion) according to the manufacturer's instructions.
Northern Blot Analyses-Poly(A)-selected mRNA from Neuro2a and NG108 neuroblastoma cells, as well as control mouse brain, kidney, and liver tissues were loaded (6 g/lane) on a 0.8% agarose 2% formaldehyde MOPS gel (30,32). After transfer to nylon membranes (MSI), blots were prehybridized for 2 h and then hybridized for 16 h in Ultrahyb TM hybridization solution (Ambion) at 44°C. Northern hybridizations were performed sequentially (NPR-C, NPR-B, and NPR-A, respectively) using 200,000 cpm/ml of each probe. Probes consisted of EcoRI-excised cDNA inserts (see below) labeled by random incorporation of [ 32 P]dCTP (Random Primers kit; Life Technologies, Inc.). Membranes were washed twice in 2ϫ SSC, 0.1% SDS solution at 44°C for 5 min, and then washed twice in 0.1ϫ SSC, 0.1% SDS for 15 min at 44°C. Blots were exposed for 3 days and signals detected with a PhosphorImager (Molecular Dynamics, Inc.).
Natriuretic receptor probes were obtained by RT-PCR using mouse kidney total RNA as template. Primers (Life Technologies, Inc.) were designed using the on-line Primer3 software (33) based on mouse or rat natriuretic peptide receptor sequences published in the NCBI data base (GenBank TM accession numbers L31932, X14177, and D78175). The sense primers for natriuretic receptor subtypes A, B, and C were 5Ј-ATTTGTGGGAGCTTGTACCG-3Ј, 5Ј-GTGTACCCTGCTGCCTCTG-T-3, and 5Ј-CTTCCAGGTGGCCTACGAA-3Ј, and antisense primers were 5Ј-GGCAATTTCCTGAAGGATGA-3Ј, 5Ј-CCGCAGATATACACA-ATGCG-3Ј, and 5Ј-GGCACACATGATCACCACTC-3Ј, respectively. These were designed to generate PCR fragments of 389, 379, and 492 base pairs, corresponding to nucleotides 1790 -2179, 376 -755, and 279 -771 for natriuretic receptor subtypes A, B, and C respectively. Amplifications were carried out for 35 cycles of denaturation (94°C, 50 s), annealing (54°C, 45 s), and extension (72°C, 45 s). PCR was finished by an incubation for 5 min at 72°C. RT-PCR reactions yielded products of the expected size. To validate the nature of the PCR products, one-fifth of the PCR products were run on 2% agarose gels, followed by overnight transfer to nylon membrane (Magna, MSI). Membranes were baked, UV-cross-linked, and then preincubated for 6 h at 37°C in the hybridization solution described (30). Hybridizations were performed at 37°C for 6 h, using the following 32 P-end-labeled internal oligonucleotides: 5Ј-GCGTGGTAGATGGACGTTTT-3Ј, 5Ј-GT-ATCTGGATGCTCGCACAG-3Ј, and 5Ј-CTGGACGACATAGTGCGCT-A-3Ј. These corresponded to the nucleotides 1996 -2015, 579 -598, and 715-734 of the sequences of natriuretic receptor subtypes A, B, and C, respectively. Membranes were washed three times at 42°C for 20 min in 1ϫ SSC buffer containing 0.1% SDS. Signals were detected using a PhosphorImager (Molecular Dynamics) (exposure time from 2 h to overnight). In parallel, PCR products were subcloned into PCR2.1 vector using the TA cloning kit (Invitrogen, San Diego, CA) and sequenced to confirm their identity.
RT-PCR Analysis of Natriuretic Peptide Receptor Gene Expression in Neuro2a Cells-RT-PCR was performed on RNA from Neuro2a cells and NG108 cells and positive control such as mouse brain and mouse embryonic neural tube after DNase treatment (DNA-free TM kit; Ambion). The same RT-PCR conditions were used as those used to obtain the hybridization probes from mouse kidney (see above). Southern analysis of one-fifth of the PCR products utilized 32 P-labeled oligonucleotides internal to the PCR primers (i.e. the same as those used above to validate the cDNA probes used on Northern blots). Another fraction of the PCR products was subcloned and sequenced.

RESULTS
Action of ANP Analogs on Proliferation-Thymidine incorporation was determined during the four final hours of a 5-h period of peptide treatment. These studies revealed that natriuretic peptides modulated proliferation in all neuroblastoma cell lines tested. Natriuretic peptides stimulated NG108 growth dose-dependently (Fig. 1A). Proliferation was significantly stimulated by ANP at 0.1 nM and reached a nearly maximum 33% increase in proliferation at 10 nM. CNP also stimulated the growth of these cells, although less potently. In contrast, the NPR-C-selective analog desANP 4 -23 (29) had no effect on NG108 cells. In all other neuroblastoma cell lines, a biphasic action of ANP and CNP was observed ( Fig. 1, B-D). ANP and CNP slightly but significantly stimulated proliferation in all of these cell lines at subnanomolar concentrations but inhibited cell proliferation at higher doses. Treatment with desANP 4 -23 did not stimulate proliferation at any concentration but inhibited growth beginning at 0.1 nM. Stimulatory and inhibitory effects of natriuretic peptides were unaffected by o-phenanthroline and bacitracin (data not shown), which have been shown to potently inhibit endopeptidase activities that cleave natriuretic peptides in other neuroblastoma cells (28). This indicates that the actions of natriuretic analogs were not likely to be due to degradation products.
To obtain information regarding the signaling pathways used by natriuretic peptides to control cell proliferation, peptide actions were studied in the presence and absence of various protein kinase inhibitors. NG108 and Neuro2a cells were selected for these studies. To determine the role of cGMP signaling, cells were pretreated with Rp-8-pCTP-cGMPS (20 M), a protein kinase G (PKG) inhibitor (34). Stimulatory effects of natriuretic peptides in both cell lines were completely blocked by Rp-8-pCTP-cGMPS (Fig. 2, A and B). In Neuro2a cells, blockade of this stimulation revealed more potent and pronounced growth inhibitory effects of ANP and CNP (Fig. 2B versus Fig. 1D).
In contrast to the PKG inhibitor, the MEK1/2 kinase inhibitor PD98059 (35) did not block the induction of proliferation by ANP or CNP, although it significantly decreased the basal rate of proliferation of NG108 cells at the concentration of 20 M (Fig. 3A). The same was true in Neuro2a cells (Fig. 3B). Growth-inhibitory actions of higher doses of ANP, CNP, and desANP 4 -23 in Neuro2a cells were no longer observed in the presence of PD98059. Thus, the antiproliferative actions of natriuretic peptides appeared to require basal activity of a MEK1/2 pathway. Inhibitors of protein kinase A (H89, 20 M) (36) and PKC (GF109203X, 10 M) (37) did not significantly alter the inhibitory actions of any of the natriuretic analogs tested (data not shown).
One possible explanation for the inhibitory actions of high concentrations of ANP and CNP in SK-N-SH subclones and Neuro2a cells is that NPR-A and/or NPR-B switch from GC activation to MAP kinase inhibition when presented with higher concentrations of ligands and thereby inhibit proliferation. To determine if both effects were mediated by NPR-A and/or NPR-B, DNA synthesis was determined in Neuro2a and NG108 cells incubated with natriuretic peptides in the presence or absence of the NPR-A/NPR-B-selective antagonist HS-142-1 (38) (Fig. 4). The stimulatory actions of the natriuretic peptides ANP and CNP on NG108 cells were antagonized by HS-142-1, confirming that the growth-stimulatory actions of natriuretic peptides were mediated by GC-coupled receptors (Fig. 4, A and B). In contrast, the antiproliferative actions of ANP, CNP, and desANP 4 -23 in Neuro2a cells were unaffected by this drug (Fig. 4C), indicating that the growth-inhibitory effects were probably not mediated by NPR-A or NPR-B. HS-142-1 (20 g/ml), by itself, had a negligible effect on DNA synthesis in untreated Neuro2a and NG108 cells.
Coupling to Cyclic Nucleotide Messengers and Intracellular Calcium-To better delineate the dual actions of natriuretic peptides on Neuro2a cell proliferation, we measured natriuretic peptide-induced changes in intracellular levels of various second messengers. Changes in cGMP levels were measured in the presence of 0.1 mM isobutylmethylxanthine to inhibit phosphodiesterase activity. ANP and CNP potently triggered an approximate 6-fold increase of cGMP levels in Neuro2a cells, whereas desANP 4 -23 did not produce any measurable change (Fig. 5). In agreement with the pharmacological results, these data suggested that ANP and CNP stimulated growth by a cGMP/PKG pathway, presumably via NPR-A and/or NPR-B. Alternatively, NPR-C (the only known receptor with which desANP 4 -23 interacts) has been shown in some systems to couple to adenylyl cyclase modulation and calcium influx (9, 10, 29). Thus, natriuretic peptide-induced changes in cAMP levels were measured in Neuro2a cells in the presence of isobutylmethylxanthine with or without forskolin. None of the ANP analogs affected cAMP levels under these conditions (data not shown). Changes in intracellular calcium levels were measured in Neuro2a cells grown on glass coverslips, in response to 0.1, 10, or 20 nM concentrations of natriuretic peptides ANP, CNP, and desANP 4 -23 . None of the natriuretic peptides induced changes in intracellular calcium levels (data not shown). In contrast, other neuropeptides (1 nM PACAP-38, 10 nM PACAP-27, or 20 nM VIP), used as internal positive controls, stimulated increases in intracellular calcium, presumably due to action of these peptides on the PACAP-preferring PAC 1 receptor in these cells (30).
Pharmacological Characterization of Natriuretic Receptors-The radiotracer 125 I-ANP bound specifically to both NG108 and Neuro2a cells, in a time-dependent and saturable fashion. Nonspecific binding was less than 25% of the total binding at the equilibrium. Kinetic (Figs. 6, A and B) and saturation (Figs. 6, C and D) experiments were performed in both Neuro2a and NG108 cells to assess binding parameters. In absence of native peptide, total binding in both cell lines increased nonlinearly with time and did not reach a maximum by 200 min. However, after subtraction of nonspecific binding, binding equilibrium was reached after a 50 -100-min incubation at 4°C in NG108 and Neuro2a, respectively. Subsequent experiments were done after a 120-min incubation at 4°C. Under these conditions, specific 125 I-ANP binding increased with concentration of the radiotracer and appeared saturable in both cell lines (Fig. 6, C  and D). Linearization of the time course plots (data not shown) revealed one class of binding sites in NG108 cells, characterized by an apparent association constant (K obs ) of 0.034 min Ϫ1 , whereas two classes of binding sites were revealed in Neuro2a cells. These two classes are characterized by different K obs (0.018 and 0.06 min Ϫ1 ). Scatchard plots (data not shown) allowed a direct access to the binding parameters. However, to determine more precisely the values of B max and K D , a nonlinear regression was run using Graphprism TM software. In NG108 cells, a single class of binding sites was observed with a K D of 42 Ϯ 1.3 pM and a B max of about 650 sites/cell. Neuro2a had high (18.6 pM) and low (K D ϭ 0.64 nM) affinity binding sites. The density of these binding sites is given by B max values of about 210 and 310 sites/cell for high and low affinity sites, respectively.
Pharmacological profiles of ANP binding sites were determined on NG108 and Neuro2a cells, using displacements of 125 I-ANP binding by increasing concentrations of ANP, CNP, and ANP 4 -23 . For NG108 cells, displacement curves (Fig. 7A) and derived IC 50 values (Table I) indicated that natriuretic peptides displaced 125 I-ANP with the following potency: ANP Ͼ CNP Ͼ desANP 4 -23 . Hill values were near unity for all analogs, suggesting that ligand/receptor interactions occurred on an apparent single 125 I-ANP binding site. The observed analog rank potency suggests that the main receptor on NG108 cells is NPR-A (3).
In Neuro2a cells, displacement curves revealed a more complex situation. Both CNP and desANP 4 -23 were as potent as unlabeled ANP in displacing 125 I-ANP (Fig. 7B). Using a onesite competition equation to graph the displacement curves, Hill values obtained were below 0.7 (Table I). This suggested that natriuretic peptides interact on Neuro2a cells through more than one binding site. This is more apparent in Fig. 8A, which shows the full set of data points for displacement by CNP and desANP 4 -23 (fewer data points were available for ANP displacement; these are all shown in Fig. 7B). Using a partial F-test to compare one-site versus two-site models (39), it was found that CNP and desANP 4 -23 displacement curves were statistically better described using a two-site competitive binding equation (Fig. 8A). Using this model, displacement of radiolabeled ANP by CNP revealed high affinity (IC 50 ϭ 74 Ϯ 2.06 pM) and low affinity (IC 50 ϭ 8.1 Ϯ 0.45 nM) sites. The high affinity sites represented about 70 Ϯ 0.6% of the total binding. The opposite situation was observed using desANP 4 -23 as a displacing agent. Approximately 35% of the total binding sites were high affinity (IC 50 ϭ 9 Ϯ 0.26 pM), whereas the remaining 65% were low affinity (IC 50 ϭ 1.5 Ϯ 0.18 nM). Taken together, these data suggest that two distinguishable binding sites are present in Neuro2a cells.
To further analyze the nature of the binding sites on Neuro2a cells, we performed displacement of 125 I-ANP with the NPR-A/NPR-B antagonist HS-142-1. This compound was able to maximally inhibit about 65% of 125 I-ANP-specific binding, with an IC 50 of about 7 g/ml (Fig. 8B). This suggested that about 65% of the binding was due to interaction of 125 I-ANP with NPR-A and/or NPR-B, while the remaining 35% was due to interaction with another site. To determine if the residual HS-142-1-insensitive 125 ANP binding sites could be displaced by desANP 4 -23 , cells were preincubated with 25 g/ml HS-142-1, a concentration that maximally inhibited 125 I-ANP displacement by native ANP. Under these conditions, desANP 4 -23 was able to displace 125 I-ANP binding with high affinity (IC 50 of about 22 Ϯ 3.5 pM (Fig. 8C)).
Natriuretic Peptide Receptor Gene Expression in Neuroblastoma Cells-Northern analysis on poly(A)-selected RNA was used to detect the three known natriuretic receptor mRNAs in NG108 and Neuro2a cells. Positive controls for hybridization were brain, kidney, and lung samples. NG108 cells displayed an appropriately sized band for NPR-A, two weak bands for NPR-B (surrounding the expected size observed in control tissues), and no signals for NPR-C (Fig. 9). Neuro2a cells also exhibited weak bands for NPR-A and NPR-B and no signal for NPR-C. To confirm the Northern data and to detect very low levels of gene expression for these receptors, we performed RT-PCR on poly(A)-selected RNA using the same receptorspecific primer sets that generated cDNA probes used in the Northern analysis. In NG108 cells, bands of the expected sizes were obtained in NPR-A, NPR-B, and NPR-C lanes (Fig. 10). Southern analysis with internal probes confirmed specificities of hybridization signals for NPR-A and NPR-C (Fig. 10). Because NPR-B could not be detected in NG108 cells in this experiment, RT-PCR was repeated on these cells. Cloning and sequencing of PCR products finally revealed the presence of a low level of NPR-B receptor gene expression in NG108 (data not shown). Thus, GC-coupled receptors (primarily NPR-A, but also perhaps a low level of NPR-B) are likely to mediate the natriuretic peptide-induced stimulation of NG108 proliferation. Although NPR-C mRNA was expressed at low levels in NG108 cells, the pharmacological data showed no evidence that a functional NPR-C was present in these cells.
RT-PCR on Neuro2a cell mRNA resulted in bands for NPR-A and NPR-B but no band corresponding to NPR-C (Fig. 10). NPR-A and NPR-B bands were confirmed by Southern analysis on PCR products. However, because the NPR-C receptor analog desANP 4 -23 inhibited proliferation in Neuro2a cells, we assayed further for expression of NPR-C using multiple primer pairs, which spanned most of the cDNA (Fig. 11). These primer sets consistently demonstrated the presence of NPR-C receptors in mouse neural tube (used as control) but failed to show the presence of an NPR-C mRNA in Neuro2a cells.

DISCUSSION
The data described here indicate that natriuretic peptides regulate the proliferation of neuroblastoma cell lines in a cellspecific manner. In one cell line (NG108), ANP and CNP induced purely a dose-dependent increase in thymidine incorporation by a mechanism that involved cGMP-dependent PKG. Modest but significant growth-stimulatory effects were also observed in SK-N-SH subclones and Neuro2a cells with low concentrations of natriuretic peptides. At higher peptide concentrations, however, the growth-stimulatory effects in these cells were apparently overridden by a growth-inhibitory mechanism. This inhibitory effect was sensitive to the MEK1/2 inhibitor PD98059, which also significantly decreased proliferation in the absence of natriuretic peptides.
The interpretation of experiments in NG108 cells seems relatively straightforward. The dose-dependent stimulation of proliferation was completely blocked by the PKG inhibitor Rp-8-pCTP-cGMP. This suggests the involvement of NPR-A or NPR-B receptors. These contain a GC domain in the intracellular portion of the receptor that is activated by agonist bind-

FIG. 6. Binding of 125 I-ANP in NG108 and Neuro2a cells: kinetic (A and B) and saturation (B and C) analyses.
In time course experiments, radiotracer (36 pM) was incubated for increasing times at 4°C. In saturation studies, cells were incubated for 120 min at 4°C, in binding buffer containing increasing concentrations of radioiodinated ANP. Cells were washed subsequently three times with 1 ml of cold phosphate-buffered saline containing 1% bovine serum albumin. Nonspecific binding was measured in the presence of 5 M of native ANP. Specific binding was calculated by subtraction of nonspecific from total binding. Data (mean Ϯ S.E.) are representative of two independent experiments each performed in duplicate. Using Graphpad TM software, data were graphed and fitted using nonlinear equations.  Table I. ing. The fact that ANP was significantly more potent than CNP in inducing proliferation and displacing 125 I-ANP binding in NG108 cells suggests that the primary receptor subtype expressed was NPR-A. This was corroborated by Northern analysis, which clearly showed the presence of NPR-A gene transcripts in NG108 cells. Natriuretic peptide stimulation of proliferation via cGMP-coupled receptors has also been observed in other cell types, for example chick embryonic cardiomyocytes (19). The GC pathway also appears to mediate the stimulatory effects of natriuretic peptides on bone growth (20).
As shown by RT-PCR followed by Southern analysis and/or DNA sequencing, low levels of NPR-B and NPR-C mRNAs were also present in NG108 cells. Such a low level of expression may explain why the encoded receptors could not be distinguished from NPR-A binding sites in pharmacological studies. As such, they seem unlikely to have much significance in these cells. Moreover, there was no pharmacological evidence that functional NPR-C receptors were expressed in NG-108 cells.
The actions of natriuretic peptides on the growth of SK-N-SH subclones and Neuro2a cells appear more complex. Low concentrations of ANP and CNP induced an increase in DNA synthesis that was sensitive to PKG blockade, again indicative of action on NPR-A and/or NPR-B (GC) natriuretic receptors. On the other hand, higher concentrations of ANP and CNP inhibited proliferation in a manner that was insensitive to PKG blockade. This inhibitory action was fully abolished when cells were also incubated with the MEK1/2 inhibitor PD98059. This may imply that natriuretic peptides act downstream of a growth-stimulatory MEK1/2 pathway that is constitutively active in these cells under the study conditions. In support of this possibility, PD98059 inhibited basal proliferation by 57% in control Neuro2a cell cultures (Fig. 3). Alternatively, it is possible that growth-inhibitory concentrations of natriuretic peptides act via MEK1/2 induction, although this seems unlikely.
NPR-A and NPR-B mRNAs were found to be present in Neuro2a cells, but at low levels. This expression was confirmed by hybridization and sequencing of RT-PCR products. In addi-TABLEI Displacement of ANP radioligand by unlabeled analogs in Neuro2a and NG108 cells Displacement parameters were extracted from curves given in Fig. 7  tion, radioligand displacement experiments confirmed the existence of high affinity 125 I-ANP binding sites consistent with NPR-A and/or NPR-B receptors. Thus, like in NG-108 cells, a GC-coupled receptor is likely to mediate the stimulatory action of natriuretic peptides on the proliferation of these cells. On the other hand, the identity of receptor that mediates the PKGindependent growth-inhibitory effects of natriuretic peptides is still unclear. One hypothesis is that NPR-A and/or NPR-B receptors mediate both growth-stimulatory and growth-inhibitory actions. For example, in the presence of higher concentrations of natriuretic peptides, these receptors might switch from proliferative to antiproliferative signaling pathways. To investigate whether or not NPR-A and/or NPR-B GC receptors might also mediate the growth-inhibitory actions, we tested whether or not these actions could be blocked by the NPR-A/NPR-Bselective antagonist HS-142-1. Although this reagent, as expected, blocked the growth-stimulatory action of natriuretic peptides, it did not prevent the growth-inhibitory effects (Fig.  4), suggesting that neither NPR-A nor NPR-B was involved in the antiproliferative actions.
Radiotracer experiments in Neuro2a cells suggested that another natriuretic peptide receptor (in addition to NPR-A and NPR-B) might be present in these cells. Most notably, these studies demonstrated the existence of both HS-142-1-sensitive and HS-142-1-insensitive 125 I-ANP binding sites. This sug-gested that the antiproliferative actions of natriuretic peptides might be mediated by an HS-142-1-insensitive receptor. The NPR-C-specific analog desANP 4 -23 effectively competed these HS-142-1-insensitive 125 I-ANP binding sites, suggesting the presence of NPR-C or an NPR-C-like receptor in Neuro2a cells. That such a receptor might couple to growth is suggested by the fact that NPR-C-specific ligands such as desANP 4 -23 have been reported to regulate cell growth or other cellular processes by way of increased intracellular calcium, adenylyl cyclase inhibition, or decreased MAP kinase activity (reviewed in Ref. 4). Interestingly, we found that relatively low concentrations of desANP 4 -23 selectively mimicked the MEK1/2-sensitive antiproliferative actions of high concentrations of ANP and CNP in SK-N-SH subclones and Neuro2a cells. However, RT-PCR analysis revealed gene expression for NPR-A and NPR-B only, leaving unclear the molecular nature of the receptor that mediates the growth-inhibitory actions of the natriuretic peptides. Among the possibilities, 1) NPR-C is expressed in Neuro2a cells, but at levels below detection by Northern analysis and RT-PCR, and 2) a non-GC receptor other than NPR-C that binds both natural natriuretic peptides and desANP 4 -23 mediates the growth-inhibitory actions. Considering the sensitivity of RT-PCR technique and the multiple primer sets used, the first hypothesis seems highly improbable. Interestingly, NPR-C as well as an NPR-C-like receptor (termed "type D") was recently cloned from eels. Curiously, expression of the type D receptor in COS cells confirmed its high affinity for FIG. 10. Polymerase chain reaction amplification of natriuretic peptide receptor types A, B, and C in Neuro2a and NG108 neuroblastoma cells and in mouse control samples (adult brain and E10 embryo neural tube). Total RNA (800 ng) and mRNA from neuroblastoma cells (25 ng) were used as template for reverse transcription. Primers and PCR method are described under "Materials and Methods." Ten microliters of the amplified products (50 l total) were run on 2% agarose gel. After Southern transfer, membranes were hybridized with 32 P-end-labeled internal oligonucleotides as described under "Materials and Methods." FIG. 11. Polymerase chain reaction amplification of NPRC receptor in Neuro2a mouse neuroblastoma cells and in neural tubes isolated from embryonic day 10 mice. Various sets of primers (shown on the left of each gel panel) were designed to cover most of the receptor mRNA sequence (shown at the top). Amplified products were run on 2% agarose gels. After Southern transfer, membranes were hybridized with 32 P-end-labeled internal oligonucleotides as described under "Materials and Methods." Autoradiographs of the corresponding PCRs revealed specific hybridizations to the predicted size PCR reaction products from neural tubes samples, but never from Neuro2a RNA. For each sample, two different concentrations of MgCl 2 were used (A and B correspond to 2 and 3 mM, respectively). RT-PCRs performed in the absence of reverse transcriptase (NC) are shown as indicators of genomic DNA contamination. Expected sizes of the PCR products are given for each set of primers, whereas band sizes were estimated using a commercially available DNA ladder (M). desANP 4 -23 but revealed an unexpected sensitivity to HS-142-1 (40). Other groups have proposed the existence of atypical "NPR-C-like" receptors in rats (25,41).
Our interest in natriuretic peptide actions on neuroblastoma cells stems from previous studies on the proliferative actions of VIP-and PACAP-related peptides on both normal neuroblasts (32,42) and neuroblastoma cell lines (30,43). Because the actions of these peptides in Neuro2a cells could not be fully explained by the presence of known VIP and PACAP receptors (30), we postulated that the effects could be mediated by putative receptors that bind both VIP-and ANP-related peptides (44,45). In fact, we found that a small portion of 125 I-ANP binding sites in Neuro2a cells could be displaced by high concentrations of VIP-related neuropeptides (data not shown). Thus, VIP-and ANP-related peptides may in some way interact at the cell surface to control cellular functions (46). Studies that further address the interaction of VIP-and PACAP-related peptides with natriuretic peptides in neuroblastoma cells are in progress.
An interesting question raised by these studies is whether or not ANP-related ligands and their receptors also play a role in regulating the proliferation of neuroblasts in normal embryos or have other important actions in the developing nervous system. At least three different groups have shown by receptor autoradiography that natriuretic peptide binding sites are present in the ventricular (proliferative) zone of the embryonic rat brain (25,47,48). Furthermore, gene expression for CNP has been reported in the mouse brain at embryonic day 10.5 (49). Taken with the data obtained here in neuroblastoma cell lines, we propose that a natriuretic peptide ligand/receptor system functions in embryonic nervous system development.
Finally, the data reported here suggest that natriuretic peptides might be involved in neuroblastoma tumor growth. If the opposing actions of natriuretic peptides on neuroblastoma cell proliferation are mediated by different receptors, then the balance of these receptors might influence the overall growth rate of these cells in response to endogenous natriuretic peptides. A therapeutic approach might be to attempt to change the balance of these receptors or to utilize ligands that specifically antagonize NPR-A and/or NPR-B receptors or that stimulate non-GC growth receptors involved in growth inhibition.