Abstract In vitro evidence suggests that natriuretic peptide receptors (NPR)-B and NPR-C inhibit vascular smooth muscle (VSM) proliferation. NPR-B is guanylate cyclase-coupled and selectively activated by C-type natriuretic peptide (CNP)-(1-22). NPR-C is not guanylate cyclase-coupled and, unlike NPR-B, avidly binds atrial natriuretic peptide (ANP)-(1-28) as well as CNP-(1-22). Here, we investigate these receptors during the VSM proliferation and neointimal formation found 5, 7, and 20 days after compressing the central ear artery of the rabbit. Receptors were mapped autoradiographically using [125I-Tyr0]CNP-(1-22), which binds NPR-B and NPR-C, and 125I-ANP-(1-28), which binds NPR-C and NPR-A, another guanylate cyclase-coupled receptor. Normal tunica media had NPR-B–like binding sites, and the level of these did not change significantly after compression. Consistent with this, CNP-(1-22) stimulated cGMP production equally with membranes from normal or damaged arteries and was more effective than ANP-(1-28). Neointima, which became evident 5 to 7 days after arterial damage, expressed NPR-C–like sites and no detectable NPR-B–like binding. NPR-C–like sites also appeared on the media for the first time between 5 and 7 days after compression. Immunohistochemistry for proliferating cell nuclear antigen revealed widespread mitosis in VSM at 5 days after compression, but mitosis was virtually restricted to the neointima at and beyond 7 days after compression. Thus, whereas levels of NPR-B did not change significantly after arterial injury and NPR-A was not detected, NPR-C–like receptors were unregulated as mitosis declined in the media and as a prominent neointima formed.
Growth and remodeling of VSM is a key feature of diseases such as atherosclerosis and is crucial to restenosis after angioplasty.1 Several factors, including the natriuretic peptides, probably regulate these important processes.1 2 The heart secretes ANPs and BNPs into the plasma.3 Additionally, VSM itself may synthesize ANPs, and the vascular endothelium secretes two CNPs, the 22-amino acid CNP-(1-22) and the 53-amino acid CNP-(1-53), which contains the sequence of CNP-(1-22) at its carboxylic acid terminal.4 5 6 ANP-(1-28) is the major circulating form of ANP,3 and both ANP-(1-28) and CNP-(1-22) inhibit the hyperplastic responses of cultured adult VSM to serum and growth factors.2 7 8 9 Moreover, ANP-(1-28) inhibits the hypertrophic effects of angiotensin II and transforming growth factor-β on postconfluent VSM cultured from adult rat aorta.10 The effects of BNPs and CNP-(1-53) on the growth of VSM are not known. In explaining the trophic actions of the natriuretic peptides, attention has focused on guanylate cyclase-coupled NPRs2 6 7 8 because cGMP powerfully inhibits VSM hyperplasia and hypertrophy.11 However, recent work suggests that the antiproliferative effect of natriuretic peptides on cultured VSM may not correlate with their power to stimulate cGMP production.2 9
Two groups of high-affinity receptors for the natriuretic peptides are known. The first consists of two proteins, designated NPR-A and NPR-B. These proteins are monomers of 120 and 130 kD, which have been cloned and shown to contain constitutive guanylate cyclase domains.12 NPR-A and NPR-B are conservative in their ligand selectivity. NPR-A has high affinity for ANP(1-28) but virtually none for CNP-(1-22) or for synthetic analogues of ANP-(1-28) such as des[Gln18,Ser19 ,Gly20,Leu21,Gly22]ANP-(4-23)-amide (C-ANP).12 13 14 15 16 17 18 19 20 NPR-B has high affinity for CNP-(1-22) but little for ANP-(1-28)12 13 14 15 16 17 18 ; its affinity for C-ANP has not been established. The other group of high-affinity receptors for natriuretic peptides is characterized by a disulfide-linked homodimeric protein of 60 to 70 kD units. This protein has been cloned from the cow, lacks any recognized constitutive enzymatic domain, and is designated NPR-C.12 Cloned NPR-C binds a wide range of ligands, including ANP-(1-28), CNP-(1-22), and C-ANP, all with high affinities.12 15 16 It has yet to be demonstrated that NPR-C can modulate any intracellular event when it is expressed by transfection into heterologous cells.12 Nevertheless, C-ANP competitively inhibits the cellular binding and internalization of ANP-(1-28) for lysosomal destruction in a number of tissues.12 20 As a result, cloned NPR-C has been identified as the clearance receptor of ANP-(1-28).12 Recently, however, two NPR-C–like proteins have been identified in the rat.20 These have similar apparent molecular masses, and both bind ANP-(1-28) as well as C-ANP. One of these proteins has very low affinities for CNPs and is probably the natriuretic peptide clearance receptor. The other protein has much higher affinities for CNPs and may be coupled to the inhibition of cellular levels of cAMP. Consequently, ligands such as C-ANP, previously thought to be specific for NPR-C, bind to a mixture of NPR-C–like receptor proteins.20
The mitotically quiescent VSM of adult rat aorta contains mRNA for both NPR-A and NPR-B.21 VSM subcultured from adult rat aortic media expresses mRNA for NPR-B but little or none for NPR-A.21 Consistent with this, CNP-(1-22) stimulates cGMP production and inhibits proliferation of subcultured medial VSM more powerfully than does ANP-(1-28).7 8 Exogenous CNP-(1-22) also inhibits neointimal formation in vivo.22 Such facts, together with the endothelial synthesis of CNP-(1-22),5 6 have led to the suggestion that NPR-B is selectively expressed by dividing VSM in vivo and that CNP-(1-22) acts through this receptor as a paracrine regulator of VSM hyperplasia.6 However, CNP-(1-22) has recently been found to inhibit the growth of subcultured, adult, medial VSM through NPR-C–like receptor sites rather than through the guanylate cyclase–coupled NPR-B.9 Moreover, in vitro observations of receptor expression may not translate to the in vivo situation. The population of VSM cells that proliferate from adult media in culture may be different from that involved in the proliferative response to arterial injury in vivo.23 24 25 In any case, humoral factors and interactions between different types of cell, which are absent in vitro, may modify VSM receptor expression or activity in vivo. Nothing is known of the NPRs of proliferating VSM in vivo. Consequently, we have investigated the local NPRs after a compressive injury to the central artery of the rabbit ear. We studied binding with [125I-Tyr0]CNP-(1-22) and 125I-ANP-(1-28), radioligands that respectively share the receptor selectivities of CNP-(1-22) and ANP-(1-28).15 16 17 18 19 20 26 We also examined the effects of natriuretic peptides on membrane-bound guanylate and adenylate cyclases. The results show that an NPR-C–like receptor with high affinities for both CNPs and ANPs is heavily expressed by the neointima and that this type of arterial injury does not detectably alter other classes of NPR.
Materials and Methods
Unlabeled peptides were obtained from Peninsula Laboratories Ltd, as was [125I-Tyr0]CNP-(1-22) (specific activity, 1000 to 1400 Ci/mmol). 125I-ANP-(1-28) (specific activity, 2000 Ci/mmol) was obtained from Amersham Ltd, as were 125I standards and Hyperfilm 3H for autoradiography. Kits for radioimmunoassay of cAMP and cGMP were also from Amersham. NTB3 photographic emulsion was from Kodak Ltd. Hypnorm was obtained from Janssen Pharmaceutical Ltd. Rabma pellet rabbit food was from SDS Ltd. D-19 developer was from Kodak Ltd, and Amifix fixative was from Champion Photochemistry Ltd. Antibodies were from Dako Ltd. All other reagents were from Sigma Chemical Co.
Preparation of Arterial Lesions
All animal experiments were in accord with the guidelines for animal care of the University of Cambridge and had Home Office permission. Arterial lesions were prepared according to the technique of Banai et al,27 with slight modifications. Male New Zealand White rabbits (3 to 3.5 kg) were anesthetized with 0.5 mL/kg IM Hypnorm (fentanyl citrate [0.315 mg/mL] and fluanisone [10 mg/mL]). Segments of the central artery of one ear were compressed between polyethylene disks (10×5×2 mm) applied on either side. The disks and intervening artery were compressed for 45 minutes by Kelly clamps (Biomedical Research Instruments, Inc). Three sites, equally spaced in the middle third of the ear, were compressed. The other ear in each rabbit provided undamaged artery from positions corresponding to the compressed segments in the first ear. On recovering from the anesthetic, rabbits were allowed free access to water and food (Rabma pellets). Rabbits were killed 5, 7, or 20 days later by an overdose of pentobarbital (50 mg/kg IV), and the arterial segments were used as described below.
Quantitative In Vitro Autoradiography
We have previously described this technique in detail.19 26 Briefly, 15-μm frozen sections of ear tissue containing the lesioned and control arterial segments were prepared from 6 rabbits each on the 5th and 7th days after compression and from 7 rabbits on the 20th day after compression. These sections were used for binding studies with [125I-Tyr0]CNP-(1-22). Comparable sections were obtained from a further 19 rabbits (6 on the 5th day, 6 on the 7th day, and 7 on the 20th day after compression) for parallel studies with 125I-ANP-(1-28). Sections were preincubated at 20°C for 10 minutes in 50 mmol/L Tris-HCl (pH 7.4) containing (mmol/L) NaCl 100, MgCl2 5, and phenanthroline 1. Sections were then immersed in 150 mmol/L sodium chloride plus 0.5% acetic acid (pH 5.0) at 4°C for 10 minutes. This step dissociates natriuretic peptides from their receptors.19 20 26 The sections were subsequently washed in three changes of preincubation buffer, each for 2 minutes at 4°C, before being incubated with either 250 pmol/L [125I-Tyr0]CNP-(1-22) or 100 pmol/L 125I-ANP-(1-28) in fresh preincubation buffer plus 0.5% (wt/vol) BSA (fraction V) at 20°C for 30 minutes. Competitive inhibition of radioligand binding by unlabeled peptides was assessed in consecutive sections by including various concentrations (1 pmol/L to 1 μmol/L) of unlabeled CNP-(1-22), ANP-(1-28), or C-ANP. Sections from each rabbit were also incubated with radioligand plus 1 μmol/L each CNP-(1-22) and ANP-(1-28), radioligand plus 1 μmol/L each CNP-(1-22) and C-ANP, or radioligand plus 1 μmol/L each ANP-(1-28) and C-ANP. Binding of [125I-Tyr0]CNP-(1-22) in the presence of 1 μmol/L CNP-(1-22) or binding of 125I-ANP-(1-28) in the presence of 1 μmol/L ANP-(1-28) was considered to be nonspecific. The amino acid sequences of the peptides used were those of the rat peptides. Preliminary studies established that the specific binding of both radioligands had achieved equilibrium by 30 minutes under these conditions of incubation. Specificity of binding was assessed in the presence of angiotensin II, vasopressin, or neuropeptide Y (all 1 μmol/L). Radioligand and ligand degradation during incubation was measured in the incubation fluids at the end of the incubations by using HPLC with detection by gamma counts or absorption at 210 nm, as previously described.20 26 At the end of the incubations, sections were washed in preincubation buffer for 5 minutes at 4°C and rinsed with distilled water at 4°C before being air-dried for exposure to Hyperfilm 3H with 125I standards.19 26 Autoradiograms were developed as described previously.19 26 Sections were then fixed in formaldehyde and stained with hematoxylin and eosin. Binding was measured microdensitometrically, comparing autoradiograms of sections and 125I standards.19 26 Regional binding was assessed by measuring the dimensions of the lumen, media, and any neointima as separated from the media by the internal elastic lamina for each stained section. These dimensions were then used to separate medial and any neointimal regions on the corresponding autoradiograms. Specific gravity was measured gravimetrically, and protein content was analyzed from six further control arterial segments. These measurements were used to convert values of bound ligand from femtomoles per square millimeter to femtomoles per milligram protein.19 26
Binding of [125I-Tyr0]CNP-(1-22) was also assessed in sections from the control and damaged ear arteries of another 6 rabbits 20 days after compression. These sections were incubated with increasing concentrations (0 to 1 nmol/L) of [125I-Tyr0]CNP-(1-22), with or without 1 μmol/L CNP-(1-22).
Further sections from each rabbit were treated as above but incubated with each radioligand in the presence or absence of a single concentration (1 μmol/L) of CNP-(1-22), ANP-(1-28), or C-ANP. Once these sections were washed, they were fixed in Bouin’s solution for 30 seconds at 4°C, washed in distilled water for 1 minute at 4°C, and dried. They were then dipped in Kodak NTB3 photographic emulsion and kept in darkness at 4°C. After exposure, the emulsion was developed by immersing each slide in Kodak D-19 developer for 4 minutes and then fixative for 10 minutes at 20°C. Slides were washed in a stream of tap water for 10 minutes and counterstained with hematoxylin and eosin.
Immunohistochemistry of PCNA
Eighteen more male New Zealand White rabbits were killed, 6 at 5 days, 6 at 7 days, and 6 at 20 days after arterial compression. The crushed and control arterial segments were fixed in formaldehyde and embedded in paraffin. Sections (10 μm) were cut every 100 μm along the length of the crushed segments to determine the midpoint of each lesion. The midpoint sections were dewaxed and rehydrated in graded ethanols. Immunohistochemistry was performed as previously described.28 Endogenous peroxidases were quenched with 1% (vol/vol) H2O2 in absolute methanol for 30 minutes. Sections were then rinsed in PBS (pH 7.2) and incubated with mouse anti-PCNA monoclonal antibody diluted 1:200 for 20 hours at 20°C. Sections were then washed in PBS (pH 7.2) and incubated with peroxidase-coupled rabbit anti-mouse IgG diluted 1:150 for 1 hour at 20°C. Sections were again washed in PBS (pH 7.2), and bound secondary antibody was localized by incubating each section with 0.06% diaminobenzidine (wt/vol) in PBS (pH 7.2) containing 0.03% H2O2 (vol/vol) for 8 minutes at 20°C. Sections were counterstained with iron hematoxylin, dehydrated in graded ethanols, cleared in xylene, and mounted in DPX. Areas on sections were measured by microplanimetry, and PCNA-positive nuclei were counted per square millimeter.
Arteries from groups of three rabbits were excised 20 days after injury, pooled, and homogenized in HBSS (mmol/L: NaCl 137, KCl 5.4, MgSO4 0.4, Na2HPO4 0.34, CaCl2 1.26, NaHCO3 4.17, KH2PO4 0.44, and MgCl2 0.49) plus 10 mmol/L HEPES, 1 mmol/L phenanthroline, 1 mmol/L phenylmethylsulfonyl fluoride, 0.1 mg/mL aprotinin, 10 μmol/L leupeptin, and 0.1 mg/mL bacitracin (pH 7.2) at 4°C. Homogenates were centrifuged at 1200g for 10 minutes at 4°C. The supernatants were recentrifuged at 40 000g for 60 minutes at 4°C. The resulting pellets were washed three times with 1 mL of PBS containing 1 mmol/L phenanthroline, 1 mmol/L phenylmethylsulfonyl fluoride, 0.1 mg/mL aprotinin, 10 μmol/L leupeptin, and 0.1 mg/mL bacitracin (pH 7.2) (solution B) and then resuspended in this solution at 1 mg protein/mL for storage at −80°C. Aliquots (120 μg) of membrane protein were incubated in 0.5 mL of solution B plus 0.2% gelatin and 100 pmol/L 125I-ANP-(1-28), with or without 1 μmol/L unlabeled ANP-(1-28), CNP-(1-22), or C-ANP for 1 hour at 4°C. Membranes were also incubated with radioligands plus 1 μmol/L angiotensin II, vasopressin, or gastrin. Bound ligand was cross-linked to its receptor by 1 mmol/L bis(sulfosuccinimidyl) suberate for 40 minutes at 4°C. Cross-linking was stopped, and membranes were dissolved in SDS with or without 50 mmol/L dithiothreitol as described previously.20 Solubilized samples and standard proteins (myosin, 205 kD; β-galactosidase, 116 kD; phosphorylase B, 97.4 kD; BSA, 66 kD; ovalbumin, 45 kD; and carbonic anhydrase, 29 kD) were resolved by SDS-PAGE using a discontinuous buffer system on 7.5% acrylamide separating gels. Gels were stained by Coomassie blue, and autoradiograms were prepared on GRI-AX film as previously described.20
Six more male New Zealand White rabbits (3 to 3.5 kg) were killed 5 days after arterial compression, and another 6 were killed 20 days after arterial compression. Crushed and control arterial segments were isolated by dissection. Arterial tissue from different rabbits was treated separately, so that a complete set of cGMP assays was performed for each rabbit. Crushed and control arterial segments from each rabbit were respectively pooled, finely diced, and homogenized at 4°C in PBS (pH 7.4) by three 30-second bursts of 27 000 rpm (Polytron), with each burst separated by 30 seconds. Each homogenate was centrifuged at 1200g for 10 minutes at 4°C, and the supernatant was recentrifuged at 40 000g for 60 minutes at 4°C. The membrane pellet was washed three times with 50 mmol/L Tris-HCl (pH 7.4) and resuspended in this solution at 1 mg membrane protein per milliliter. Aliquots of 10 μg protein of the suspension were incubated for 20 minutes at 37°C in a final volume of 250 μL of 50 mmol/L Tris-HCl (pH 7.6) containing 1 mmol/L isobutylmethylxanthine, 4 mmol/L MgCl2, 0.5 mmol/L ATP (Mg2+ salt), 1 mmol/L GTP (Mg2+ salt), 15 mmol/L creatine phosphate, and 80 μg/mL creatine phosphokinase, plus a range of concentrations (0 to 10 μmol/L) of either CNP-(1-22) or ANP-(1-28). Incubations with CNP-(1-22) were paired to those with ANP-(1-28) for each rabbit. Single incubations were performed for each concentration of each peptide for each rabbit. Incubations were stopped by adding 0.75 mL of 50 mmol/L sodium acetate (pH 5.8) (ice cold) and boiling for 4 minutes. Samples were then centrifuged at 4000g for 10 minutes at 4°C. cGMP was measured in supernatants by radioimmunoassay.19 26 All samples from one rabbit were measured in one assay. Ligand degradation during incubations of 10 μmol/L CNP-(1-22) or 10 μmol/L ANP-(1-28) was measured by HPLC, detecting UV absorption of the natriuretic peptides at 210 nm, as previously described.20 Preliminary experiments showed that basal and natriuretic peptide-stimulated cGMP production by membranes from crushed or control arteries increased linearly with time up to at least 30 minutes and with the dose of membrane protein up to at least 20 μg.
Another 6 male New Zealand White rabbits (3 to 3.5 kg) were killed 20 days after arterial compression, as described above. Membranes were individually prepared from the crushed and control arterial segments of each rabbit as described above, and aliquots of 7.5 μg of membrane protein were incubated for 12 minutes at 37°C in a final volume of 200 μL of 50 mmol/L Tris-HCl (pH 7.5) containing 1 mmol/L isobutylmethylxanthine, 6.4 mmol/L MgCl2, 0.02 mmol/L guanosine triphosphate (Mg2+ salt), 1 mmol/L ATP (Mg2+ salt), 0.1 mg/mL myokinase, 0.2 mg/mL creatine phosphokinase, and 20 mmol/L creatine phosphate with or without 50 μmol/L forskolin and with or without 1 μmol/L CNP-(1-22), 1 μmol/L ANP-(1-28), or 1 μmol/L C-ANP. One incubation was made for each combination of added reagents for each rabbit, and all incubations for one rabbit were performed as a single experiment. Incubations were stopped by adding ice-cold trichloroacetic acid (final concentration, 6% [wt/vol]) and centrifuging at 4000g for 15 minutes at 4°C. Supernatants were extracted with water-saturated ether, and supernatant cAMP was then measured by radioimmunoassay.20 All measurements from one rabbit were performed in a single assay. Peptide degradation during incubations was measured by HPLC and UV detection as described above. Preliminary experiments showed that basal and forskolin-stimulated cAMP production by membranes from both crushed and control arteries increased linearly with time up to at least 30 minutes and with the dose of membrane protein up to at least 20 μg.
Protein contents were measured by a bicinchoninic acid assay (Sigma).
Apparent association constants (Ka) and maximum binding capacities (Bmax) were estimated by using the ligand program.29 Models with up to three classes of binding sites were compared by the extra sum-of-squares principle.29 A model with more than one class was accepted over a simpler model if the F-test criterion showed that the more complex model improved the goodness of fit to the data more than expected by chance at the P>.05 level.29 Estimates of Ka are more nearly log-normally than normally distributed about their true mean.30 Consequently, each value of Ka (liters per mole) is given as −log10 (pKa) of the geometric mean and its SEM.20 30 For reference, values of Ka are also given in Tables 1 through 3⇓⇓⇓ as their reciprocal, the dissociation constant (Kd). Other results are arithmetic mean±SEM. Dose-response curves were modeled as logistics by the allfit program, and their parameters were compared by the extra sum-of-squares principle using allfit.31 Remaining results were compared by paired or unpaired t tests, as appropriate.32
Development of Neointima
The neointima developed as originally described27 for this preparation. Control arteries had no cells between the endothelium and internal elastic lamina. The neointima was still absent at many points around the arterial lumen 5 days after arterial damage but was three to nine cells thick by 7 days and had greatly narrowed the lumen by 20 days (Figs 1⇓ and 2⇓). PCNA immunoreactivity was used as a marker of the S phase of mitosis.28 Control arteries contained no PCNA-positive cells. Smooth muscle nuclei (475±104/mm2) of sectioned media were PCNA positive 5 days after injury, but only 24±16/mm2 were positive at 7 days. PCNA-positive nuclei were virtually absent in media at 20 days (Fig 1⇓). PCNA-positive nuclei (2330±300/mm2) occurred in 7-day neointima, and 34±10/mm2 were present in neointima at 20 days (Fig 1⇓).
Autoradiographic Binding Studies
The binding of 250 pmol/L [125I-Tyr0]CNP-(1-22) was inhibited differently by 1 μmol/L CNP-(1-22), 1 μmol/L ANP-(1-28), or 1 μmol/L C-ANP (Fig 3⇓). Further information suggested that this resulted from the substantially lower affinities of ANP-(1-28) and C-ANP compared with CNP-(1-22) for a single class of binding sites. Thus, there were no significant differences between the inhibitions of the binding of 250 pmol/L [125I-Tyr0]CNP-(1-22) by 1 μmol/L CNP-(1-22), 1 μmol/L CNP-(1-22) plus 1 μmol/L ANP-(1-28), or 1 μmol/L CNP-(1-22) plus 1 μmol/L C-ANP (not shown). This implies that neither ANP-(1-28) nor C-ANP inhibited nonspecific radioligand binding.19 20 26 Moreover, the inhibitions of specific radioligand binding by increasing concentrations of CNP-(1-22), ANP-(1-28), or C-ANP were all consistent with a single class of sites (P>.24 for improvements of fit by two-class versus one-class models for each ligand) (Fig 4⇓). This class bound CNP-(1-22), ANP-(1-28), and C-ANP with Bmax values that did not differ significantly from each other. It was avid for CNP-(1-22) and less so for ANP-(1-28) (P<.01), and it bound C-ANP even less strongly than ANP-(1-28) (P<.01) (Fig 4⇓, Table 1⇑). Its low affinity for ANP-(1-28) was confirmed in that no high-affinity binding of ANP-(1-28) contributed detectably to the low level of specific binding of 100 pmol/L 125I-ANP-(1-28) on control media (Fig 5⇓, Table 2⇑). These ligand selectivities are consistent with the presence of NPR-B in the normal tunica media of the ear artery of the rabbit.
Arterial Media 5 Days After Compression
As with control arteries, the binding of 250 pmol/L [125I-Tyr0]CNP-(1-22) was inhibited differently by 1 μmol/L CNP-(1-22), 1 μmol/L ANP-(1-28), or 1 μmol/L C-ANP (Fig 3⇑), but there were no significant differences between the inhibitions of the binding of 250 pmol/L [125I-Tyr0]CNP-(1-22) by 1 μmol/L CNP-(1-22), 1 μmol/L CNP-(1-22) plus 1 μmol/L ANP-(1-28), or 1 μmol/L CNP-(1-22) plus 1 μmol/L C-ANP (not shown). The effects of increasing concentrations of CNP-(1-22), ANP-(1-28), and C-ANP on specific radioligand binding were all consistent with a single class of sites (P>.30 for improvements of fit by two-class versus one-class models for each ligand) (Fig 4⇑). This class bound CNP-(1-22), ANP-(1-28), and C-ANP with Bmax values that did not differ significantly from each other. The binding constants for the unlabeled ligands did not differ significantly from their constants in control arteries (Table 1⇑). Thus, ANP-(1-28) bound less strongly than CNP-(1-22) (P<.01), and C-ANP bound even less well than ANP-(1-28) (P<.01). Finally, no high-affinity binding of ANP-(1-28) was detected 5 days after compression with 100 pmol/L 125I-ANP-(1-28) as the radioligand (Fig 5⇑, Table 2⇑). Consequently, the medial binding of natriuretic peptides 5 days after injury does not differ significantly from that of normal arteries and resembles NPR-B.
Arterial Media at 7 and 20 Days After Compression
Only NPR-B–like sites were detected on normal media and on media 5 days after compression, but NPR-C–like sites were added at 7 and 20 days after injury. Thus, [125I-Tyr0]CNP-(1-22) labeled two classes of medial sites at both 7 and 20 days after injury. Both classes at each time had similarly high affinities for CNP-(1-22) but differed in their binding of ANP-(1-28) (P<.05 versus a single class at both times) and C-ANP (P<.05 versus a single class at both times) (Fig 4⇑, Table 1⇑). One class at each time matched the NPR-C–like sites of neointima in having high affinities for ANP-(1-28) and C-ANP. The affinities of this medial class exceeded those of the NPR-B–like sites of, for example, 20-day control media for these ligands (P<.01 for either ligand at both 7 and 20 days). The other class of sites on lesioned media at both 7 and 20 days had affinities for ANP-(1-28) and C-ANP that were significantly lower than those of the NPR-C–like sites of 20-day neointima (P<.05 and P<.01 for either ligand at 7 and 20 days, respectively). This class resembled the NPR-B–like sites of control media in its affinities for ANP-(1-28) and C-ANP. The Bmax of the medial classes of NPR-B–like binding at 5, 7, and 20 days after compression tended to be greater than for control media (Table 1⇑). However, this Bmax for injured media at any one time did not differ significantly from that of contemporary control media. Finally, the specific binding of increasing concentrations of [125I-Tyr0]CNP-(1-22) to media 20 days after injury identified only one class of sites (P>.23 for improvement of fit by two-class versus one-class models) (Fig 6⇓). This class had an affinity for the radioligand (pKa=9.01±0.63) that did not differ significantly from the affinity of media for CNP-(1-22) (Table 1⇑).
The effects of increasing concentrations of ANP-(1-28), CNP-(1-22), or C-ANP on the specific binding of 125I-ANP-(1-28) were consistent with one class of sites at both 7 days (P>.14 for improvements of fit by two-class versus one-class models for each ligand) and 20 days (P>.45 for improvements of fit by two-class versus one-class models for each ligand) (Fig 5⇑). These sites at both 7 and 20 days after injury bound CNP-(1-22), ANP-(1-28), and C-ANP with high affinities (Fig 5⇑, Table 2⇑). At neither time did the Bmax values for ANP-(1-28), CNP-(1-22), and C-ANP differ significantly from each other (Table 2⇑), nor were there significant differences in the inhibitions of the binding of 100 pmol/L 125I-ANP-(1-28) produced by 1 μmol/L ANP-(1-28) alone, 1 μmol/L ANP-(1-28) plus 1 μmol/L CNP-(1-22), or 1 μmol/L ANP-(1-28) plus 1 μmol/L C-ANP (not shown). These results with 125I-ANP-(1-28) confirm that NPR-C–like sites occur on the media at 7 and 20 days after injury.
Specific binding of [125I-Tyr0]CNP-(1-22) occurred throughout the neointima but tended to be most concentrated near the arterial lumen by 20 days (Figs 2⇑ and 3⇑). Neointimal binding of 250 pmol/L [125I-Tyr0]CNP-(1-22) at 20 days was not inhibited differently by 1 μmol/L CNP-(1-22), 1 μmol/L CNP-(1-22) plus 1 μmol/L ANP-(1-28), or 1 μmol/L CNP-(1-22) plus 1 μmol/L C-ANP (not shown). Moreover, the inhibitions of specific radioligand binding caused by increasing concentrations of CNP-(1-22), ANP-(1-28), or C-ANP were consistent with a single class of sites on the neointima as a whole (P>.40 for improvements of fit by two-class versus one-class models for each ligand) (Fig 7⇓). This class of sites bound CNP-(1-22), ANP-(1-28), and C-ANP with high affinities and with Bmax values that did not differ significantly from each other (Fig 7⇓, Table 3⇑). This was confirmed by the specific binding of 125I-ANP-(1-28), which was also almost uniformly distributed across the 20-day neointima. Thus, there were no significant differences between the inhibitions of the binding of 100 pmol/L125I-ANP-(1-28) caused by 1 μmol/L ANP-(1-28), 1 μmol/L ANP-(1-28) plus 1 μmol/L CNP-(1-22), or 1 μmol/L ANP-(1-28) plus 1 μmol/L C-ANP (not shown). Moreover, the inhibitions of the specific binding of 125I-ANP-(1-28) produced by increasing concentrations of ANP-(1-28), CNP-(1-22), or C-ANP were consistent with one class of sites (P>.32 for improvements of fit by two-class versus one-class models for each ligand), which bound these unlabeled ligands with high affinities and Bmax values that did not differ significantly from each other (Fig 7⇓, Table 3⇑). Finally, there was no significant difference between the Bmax for ANP-(1-28) measured by its displacement of 125I-ANP-(1-28) and that for CNP-(1-22) measured with [125I-Tyr0]CNP-(1-22) (Table 3⇑).
The specific binding of increasing concentrations of [125I-Tyr0]CNP-(1-22) revealed a single class of neointimal sites (P>.45 for improvement of fit by two-class versus one-class models), with an affinity for the radioligand of pKa=9.63±0.07 (Fig 6⇑). This did not differ significantly from the neointimal affinity for CNP-(1-22) (Table 3⇑). The affinity of 20-day neointima for CNP-(1-22) (pKa=9.52±0.24) was similar to that of the NPR-B–like sites of undamaged media (eg, pKa=9.18±0.74 for undamaged arteries of the 20-day control group), but the neointima had higher affinities for ANP-(1-28) (pKa=9.72±0.21 versus 7.51±0.30 for 20-day controls, P<.01) and C-ANP (pKa=9.68±0.14 versus 6.71±0.50 for 20-day controls, P<.01). Neointimal binding 20 days after injury was therefore consistent with NPR-C rather than NPR-B or NPR-A.
The neointima did not encircle the lumen at 5 days after injury and was still thin at 7 days. However, it was possible to measure the binding of [125I-Tyr0]CNP-(1-22) in the 2 rabbits with the thickest neointima at 7 days. In both, CNP-(1-22), ANP-(1-28), and C-ANP bound with uniformly high affinities as though to a single class of NPR-C–like sites similar to that of the neointima at 20 days (not shown).
Ligand Degradation During Binding Studies
Measurements showed that 98.1±0.3% of the initial amount of [125I-Tyr0]CNP-(1-22) and 97.8±0.4% of the initial amount of 125I-ANP-(1-28), added to sections from arteries injured 20 days before, survived the binding incubations. Similarly, 96.8±0.3% of CNP-(1-22) and 97.3±0.3% of ANP-(1-28) survived incubation after being added initially as 1 μmol/L solutions.
SDS-PAGE of Arterial Membrane Proteins
One protein was specifically labeled by 125I-ANP-(1-28) in membranes prepared from whole arteries 20 days after injury. The reduced protein had an Mr of ≈64 000, and its labeling was prevented by 1 μmol/L unlabeled ANP-(1-28), CNP-(1-22), and C-ANP (Fig 8⇓) but not 1 μmol/L angiotensin II, vasopressin, or gastrin (not shown). An additional band, with an apparent Mr of ≈130 000, was specifically labeled by 125I-ANP-(1-28) in the absence of dithiothreitol, and the ≈64-kD band became less heavily labeled (Fig 8⇓). No proteins were detectably labeled by 125I-ANP-(1-28) under the same conditions in membranes from control arteries. The results affirm that injured arteries preferentially express a protein with ligand selectivities and molecular characteristics of the disulfide-bridged homodimeric NPR-C.
Cyclic Nucleotide Production
Generation of cGMP
CNP-(1-22) produced dose-dependent increments in the rate of cGMP production by membranes prepared either from control arteries or from arteries 20 days after injury (Fig 9⇓). Basal cGMP production, as estimated by allfit, was higher in the lesioned arteries (1.4±0.2 versus 1.1±0.2 pmol/mg protein per minute, P<.05). Taking this into account, the maximum increments in cGMP production caused by CNP-(1-22) did not differ significantly between control and lesioned arteries (Fig 9⇓). Moreover, the concentration of CNP-(1-22) that achieved 50% of the maximum increment in particulate cGMP production (EC50) was not significantly different between control (1.1×10−7 mol/L) and damaged (3.2×10−8 mol/L) arteries. ANP-(1-28) also caused dose-related increases in cGMP production (Fig 9⇓). Basal cGMP production estimated by allfit in the experiments with ANP-(1-28) was 1.0±0.2 and 1.6±0.3 pmol/mg protein per minute for control and damaged arteries, respectively. These were not significantly different from the corresponding basal rates in the experiments with CNP-(1-22), but the maximum rates of cGMP production for control and damaged arterial membranes with ANP-(1-28) (2.0±0.4 and 2.4±0.8 pmol/mg protein per minute, respectively) were significantly lower than those with CNP-(1-22) (3.1±0.6 and 3.4±0.6 pmol/mg protein per minute, respectively) [both P<.01 for ANP-(1-28) versus CNP-(1-22)]. The results suggest that normal arteries and arteries 20 days after injury express similar levels of the CNP-stimulated cyclase-coupled NPR-B. This is consistent with the autoradiographic binding data, which show that injury does not significantly alter the level of arterial NPR-B–like binding sites.
Generation of cAMP
Basal rates of cAMP production and the increments achieved by 50 μmol/L forskolin were similar for membranes from control arteries and arteries 20 days after injury. Neither basal nor forskolin-stimulated rates were significantly affected by CNP-(1-22), ANP-(1-28), or C-ANP, whatever the source of membranes (Fig 10⇓).
Agonist Degradation During Enzymatic Studies
When 10 μmol/L CNP-(1-22) and 10 μmol/L ANP-(1-28) were incubated with control membranes to assay cGMP production, 90.7±3.1% and 93.9±7.9%, respectively, of the peptide survived incubation. These figures were 92.5±5.3% for CNP-(1-22) and 91.3±5.1% for ANP-(1-28) for membranes from arteries injured 20 days earlier. Similarly, 91.1±2.0% of 1 μmol/L CNP-(1-22) and 90.1±4.6% of 1 μmol/L ANP-(1-28) survived incubations to assay cAMP production with control membranes. These figures were 90.2±4.1% for CNP-(1-22) and 93.9±5.1% for ANP-(1-28) for membranes from arteries injured 20 days earlier. Consequently, differential degradation does not explain the relative agonist activities of CNP-(1-22) and ANP-(1-28).
Our results suggest that the media of the central ear artery of the rabbit normally expresses NPR-B and also develops NPR-C–like receptors between 5 and 7 days after compressive injury. NPR-C–like receptors also occur on the neointima. These findings are important because both NPR-B6 7 8 and NPR-C–like receptors2 9 are implicated in controlling VSM growth and because recent evidence suggests that neointimal VSM produces CNP-(1-22).33 CNP-(1-22) may therefore have autocrine and paracrine effects through the NPR-C–like receptors of the damaged artery, even though the normal site of synthesis for this peptide, the endothelium, may be lost over neointima.27 34
NPR-A is the only guanylate cyclase-coupled NPR identified so far in the rabbit.35 36 37 38 However, we found no NPR-A–like binding sites on the rabbit ear artery, despite the washing of tissues with acid to expose receptors occluded by endogenous ligands.19 20 26 Instead, normal media showed a single class of natriuretic peptide binding sites with high affinity for CNP-(1-22) and low affinity for ANP-(1-28), consistent with the ligand selectivity of cloned NPR-B.14 15 16 17 18 CNP-(1-22), which is the prime agonist of NPR-B and has virtually no affinity for NPR-A,12 13 14 15 16 17 18 19 was also the main agonist of guanylate cyclase in membranes prepared from normal ear arteries. Consequently, it is likely that NPR-B mediated the cGMP response of arterial membranes to low levels of CNP-(1-22). ANP-(1-28) did increase cGMP production by arterial membranes, possibly through a pool of NPR-A too small to be detected autoradiographically. However, ANP-(1-28) is a partial agonist of cloned NPR-B.16 We did not examine whether ANP-(1-28) antagonized cGMP production stimulated by CNP-(1-22), but considering the partial agonism of ANP-(1-28) at NPR-B, it would be compatible with the presence of only NPR-B that ANP-(1-28) should stimulate guanylate cyclase activity in control membranes, albeit to a significantly lower maximum than achieved with CNP-(1-22). Thus, both binding and enzymatic data suggest that the central ear artery of the rabbit normally expresses NPR-B rather than other NPRs. This contrasts with normal rabbit aorta, in which both NPR-A and NPR-C–like receptors are readily detected.35 38 39 The presence of NPR-B rather than NPR-A on the ear artery would explain why this artery, unlike the rabbit aorta, hardly relaxes in response to ANPs.40 Moreover, the rabbit ear artery is the first structure in vivo that has been shown to have both binding sites and a particulate guanylate cyclase that match the properties of cloned NPR-B.19 26 This is significant because the mapping of expressed NPR-B by its mRNA is plagued by discrepancies between the levels of mRNA and functionally expressed protein for this receptor.19 26 41
Rat aortic media normally expresses abundant NPR-A and little NPR-B,21 but NPR-A becomes undetectable and NPR-B substantially increases when rat aortic VSM adopts a synthetic phenotype and proliferates to confluence in culture.21 This has led to the suggestion that NPR-B is upregulated whenever VSM proliferates, for example, in damaged arteries.6 7 21 However, we detected no significant change in the level of NPR-B–like binding sites or in cGMP production from injured arteries compared with control arteries. Moreover, NPR-B–like sites were found only in media, whereas PCNA immunoreactivity confirmed that VSM proliferation after injury increased and then declined over a wide range in both the media and neointima.1 27 42 Consequently, measurable changes of NPR-B are not necessarily associated with changes in VSM proliferation in vivo. Species- and vessel-specific (ear artery versus aorta) differences might underlie this discrepancy from rat aortic VSM multiplying in vitro. However, we have recently shown that it is NPR-C–like receptors that become expressed in the media and neointima of the rat carotid artery after balloon angioplasty, whereas no NPR-B is detected.33 Upregulation of NPR-C–like receptors, rather than of NPR-B, may therefore be a general characteristic of the neointimal response to injury. The neointimal and medial VSM of damaged arteries in vivo is not only exposed to humoral agents but may also be influenced by surviving endothelium and cells such as platelets, macrophages, and lymphocytes,1 which are absent in culture. NPRs on VSM are homologously regulated,43 and they are also subject to differential heterologous regulation by, for example, angiotensin II.44 In addition to the circulating natriuretic peptides, VSM and macrophages can produce ANPs locally,4 45 whereas endothelium secretes CNPs5 6 and probably also ANP.46 Angiotensin II arises in the vascular wall as well as the plasma.47 Last, cytokines, in the form of transforming growth factor-β1, have been shown to control the expression of NPRs.41 Therefore, it is unlikely that the NPRs of VSM proliferating in vitro will directly reflect those of VSM in damaged arteries. In addition, the proliferative response to arterial injury in vivo may involve a specific subpopulation of VSM that differs from that usually selected by in vitro culture.23 24 25 This could also underlie the divergent regulation of NPR-B in the VSM of damaged arteries compared with VSM in cultures from normal rat aorta.
NPR-C–like receptor sites were not found in undamaged media, which is mitotically quiescent, or in media on the 5th day after injury, when previous work has shown that the proliferation of medial VSM is at its peak.27 The proportion of VSM cells that divide in response to arterial injury may be small,42 so that NPR-C–like receptor sites might be too few to be detected if restricted to these cells. NPR-C–like sites were discovered in the media at 7 and 20 days, but it is not known whether this was because the progeny of dividing cells accumulate locally42 or because medial cells individually develop more NPR-C–like sites at these times. However, the fact that we found NPR-C–like sites throughout the neointima at 7 and 20 days, times of widely different mitotic activity, suggests that these receptor sites occur on neointimal VSM whether or not it is dividing. Evidence has recently accrued indicating that neointimal VSM in the rat derives from a particular subpopulation of phenotypically distinct VSM cells, which form only a small fraction of the adult media.23 24 25 It may be that a similar phenomenon explains the accumulation of cells expressing NPR-C–like receptors in the neointima of the rabbit, although such cells are not apparent in the normal media.
The NPR-C–like receptor sites of rabbit neointima have high affinity for CNP-(1-22) but, unlike those with high affinity for this peptide on rat glomeruli,20 are not related to the control of cAMP levels. Nevertheless, both ANP-(1-28) and CNP-(1-22) can inhibit the proliferation of rat VSM in vitro through an NPR-C–like receptor that does not modulate cAMP.2 9 This implies that the NPR-C–like sites that occur after arterial damage could limit the proliferative response of VSM to the injury. In keeping with this, we have recently demonstrated that intravenous infusions of exogenous ANP-(1-28) or CNP-(1-22) inhibit neointimal development in the present type of arterial damage.22 Consequently, endogenous natriuretic peptides may also be significant in controlling this arterial response to injury, particularly at times when other inhibitors of VSM proliferation, such as prostacyclin and nitric oxide, are diminished by endothelial damage. Such a role for NPR-C–like receptors in the control of neointimal growth would have important implications for the future design of receptor-selective agents that could modulate VSM growth therapeutically.
Selected Abbreviations and Acronyms
|ANP||=||atrial natriuretic peptide|
|BNP||=||brain natriuretic peptide|
|CNP||=||C-type natriuretic peptide|
|HPLC||=||high-performance liquid chromatography|
|NPR||=||natriuretic peptide receptor|
|PCNA||=||proliferating cell nuclear antigen|
|VSM||=||vascular smooth muscle|
This study was supported by the British Heart Foundation and Newton Trust. The allfit program was donated by Dr P.J. Munson.
Previously published in part in abstract form [J Physiol (Lond). 1994;475:57P].
- Received March 24, 1995.
- Accepted July 24, 1995.
- © 1995 American Heart Association, Inc.
Bobick A, Campbell J. Vascular derived growth factors: cell biology, pathophysiology and pharmacology. Physiol Rev.. 1993;45:1-42.
Appel RG. Growth regulatory properties of atrial natriuretic factor. Am J Physiol.. 1992;262:F911-F918.
Westenfelder C, Baranowski EH. Cultured aortic smooth muscle cells (SMC) secrete atrial natriuretic factor (ANF). Kidney Int.. 1990;37:343. Abstract.
Stingo AJ, Clawell AL, Heublein DM, Wei C-M, Pittelkow MR, Burnett JC. Presence of C-type natriuretic peptide in cultured human endothelial cells and plasma. Am J Physiol.. 1992;263:H1318-H1321.
Suga S, Nakao K, Itoh H, Komatsu Y, Ogawa Y, Hama N, Imura H. Endothelial production of C-type natriuretic peptide and its marked augmentation by transforming growth factor-β J Clin Invest.. 1992;90:1145-1149.
Porter JG, Catalano R, McEnroe G, Lewicki JA, Protter AA. C-type natriuretic peptide inhibits growth factor dependent DNA synthesis in smooth muscle cells. Am J Physiol.. 1992;263:C1001-C1006.
Cahill PA, Hassid A. ANF-C-receptor-mediated inhibition of aortic smooth muscle cell proliferation and thymidine kinase activity. Am J Physiol.. 1994;266:R194-R203.
Itoh H, Pratt RE, Dzau VJ. Atrial natriuretic polypeptide inhibits hypertrophy of vascular smooth muscle cells. J Clin Invest.. 1990;86:1690-1697.
Garg UC, Hassid A. Nitric oxide-generating vasodilators and 8-bromo-cyclic guanosine monophosphate inhibit mitogenesis and proliferation of cultured rat vascular smooth muscle cells. J Clin Invest.. 1989;83:1774-1777.
Koller KJ, Lowe DG, Bennett GL, Minamino N, Kangawa K, Matsuo H, Goeddel DV. Selective activation of the B natriuretic peptide receptor by C-type natriuretic peptide (CNP). Science.. 1991;252:120-123.
Bennett BD, Bennett GL, Vitangcol RV, Lewett JRS, Burneer J, Henjel W, Lowe DG. Extracellular domain-IgG fusion proteins for three human natriuretic peptide receptors. J Biol Chem.. 1991;266:23060-23067.
Suga S, Nakao K, Hosoda H, Mukoyama M, Ogawa Y, Shirakami G, Arai H, Saito Y, Kambayashi Y, Inouye K, et al. Receptor selectivity of natriuretic peptide family, atrial natriuretic peptide, brain natriuretic peptide and C-type natriuretic peptide. Endocrinology.. 1992;130:229-239.
Brown J, Zuo Z. Renal receptors for atrial and C-type natriuretic peptides in the rat. Am J Physiol.. 1992;263:F89-F96.
Brown J, Zuo Z. Receptor proteins and biological effects of C-type natriuretic peptides in the renal glomerulus of the rat. Am J Physiol.. 1994;266:R1383-R1394.
Suga S, Nakao K, Kishimoto I, Hosoda K, Mukoyama M, Arai H, Shirakami G, Ogawa Y, Komatsu Y, Nakagawa O, et al. Phenotype-related alteration in expression of natriuretic peptide receptors in aortic smooth muscle cells. Circ Res.. 1992;71:34-39.
Brown J, Chen Q. Natriuretic peptides inhibit neointimal development in the rabbit. J Physiol (Lond).. 1995;483:144P. Abstract.
Walker LN, Bowen-Pope DF, Ross R, Reidy MA. Production of platelet-derived growth factor-like molecules by cultured arterial smooth muscle cells accompanies proliferation after arterial injury. Proc Natl Acad Sci U S A.. 1986;3:7311-7315.
Majesky MW, Reidy MA, Bowen-Pope DF, Hart CE, Wilcox JN, Schwartz SM. PDGF ligand and receptor gene expression during repair of arterial injury. J Cell Biol.. 1990;111:2149-2158.
Majesky MW, Giachelli CM, Reidy MA, Schwartz SM. Rat carotid neointimal smooth muscle cells reexpress a developmentally regulated mRNA phenotype during repair of arterial injury. Circ Res.. 1992;71:759-768.
Brown J, Zuo Z. C-type natriuretic peptide and atrial natriuretic peptide receptors of rat brain. Am J Physiol.. 1993;264:R513-R523.
Banai S, Shou M, Correa R, Jaklitsch MT, Douek PC, Bonner RF, Epstein SE, Unger EF. Rabbit ear model of injury-induced arterial smooth muscle cell proliferation: kinetics, reproducibility, and implications. Circ Res.. 1991;69:748-756.
Hall PA, Levison DA, Woods AL, Yu CC, Kellock DB, Watkins JA, Barnes DM, Gillett CE, Camplejohn R, Dover R, et al. Proliferating cell nuclear antigen (PCNA) immunolocalisation in paraffin sections: an index of cell proliferation with evidence of deregulated expression in some neoplasma. J Pathol.. 1990;162:285-294.
DeLean A, Hancock AA, Lefkowitz RJ. Validation and statistical analysis of a computer modeling method for quantitative analysis of radioligand binding data from mixtures of pharmacological receptor subtypes. Mol Pharmacol.. 1982;21:5-16.
DeLean A, Munson PJ, Rodbard D. Simultaneous analysis of families of sigmoidal curves: application to bioassay, radioligand assay and physiological dose-response curves. Am J Physiol.. 1978;235:E97-E102.
Snedecor GW, Cochran WG. Statistical Methods. Ames, Iowa: Iowa State University Press; 1980.
Brown J, Chen Q. An autocrine system for C-type natriuretic peptide (CNP) occurs in rat arterial neointima. J Physiol (Lond). Abstract. In press.
Vandlen RL, Arcur KE, Napier MA. Identification of a receptor for atrial natriuretic factor in rabbit aorta membranes by affinity cross-linking. J Biol Chem.. 1985;260:10889-10892.
Gunning M, Ballermann BJ, Silva P, Brenner BM, Zeidel ML. Characterization of atrial natriuretic peptide receptors in rabbit inner medullary collecting duct cells. Am J Physiol.. 1988;255:F324-F330.
Gunning M, Silva P, Brenner BM, Zeidel ML. Characteristics of ANP-sensitive guanylate cyclase in inner medullary collecting duct cells. Am J Physiol.. 1989;256:F766-F775.
Winquist RJ, Faison EF, Waldman SA, Schwartz K, Murad F, Rapoport RM. Atrial natriuretic factor elicits an endothelium-independent relaxation and activates particulate guanylate cyclase in vascular smooth muscle. Proc Natl Acad Sci U S A.. 1984;81:7661-7664.
Clowes AW, Schwartz SM. Significance of quiescent smooth muscle migration in the injured rat carotid artery. Circ Res.. 1985;56:139-145.
Cahill PA, Redmond EM, Keenan AK. Vascular atrial natriuretic factor receptor subtypes are not independently regulated by atrial peptides. J Biol Chem.. 1990;265:21896-21906.
Chabrier PE, Roubert P, Lonchampt M-O, Plas P, Braquet P. Regulation of atrial natriuretic factor receptors by angiotensin II in rat vascular smooth muscle cells. J Biol Chem.. 1988;263:13199-13202.
Brandt RR, Heublein DM, Mattingly MT, Pittelkow MR, Burnett JC. Presence and secretion of atrial natriuretic peptide from cultured human aortic endothelial cells. Am J Physiol.. 1995;268:H921-H925.
Kato H, Iwai N, Inui H, Kimoto K, Uchwiyama Y, Inagami T. Regulation of vascular angiotensin release. Hypertension.. 1993;21:446-454. ]CG