JOURNAL OF CELLULAR PHYSIOLOGY 146:43-51 (1991)

Characterization of Atrial Natriuretic Peptide Receptors in Brain Microvessel Endothelial Cells PEGGY A. WHITSON,* M. HELEN HULS, AND CLARENCE F. SAMS

Biomedical Operations and Research Branch, NASA/)ohnson Space Center (PA.W., C.F.S.) and KRUC life Sciences (M.H.H.), Houston, Texas 77058 Atrial natriuretic peptide (ANP) binding and ANP-induced increases in cyclic guanosine monophosphate (cGMP)levels have been observed in brain microvessels (Chabrier et al., 1987; Steardo and Nathanson, 1987), suggesting that this fluid-regulatinghormone may play a role in the fluid homeostasis of the brain. This study was initiated to characterize the A N P receptors in primary cultures of brain microvessel endothelial cells (BMECs). The apparent equilibrium dissociation constant, Kd, for A N P increased from 0.25 nM to 2.5 nM, and the number of A N P binding sites as determined by Scatchard analysis increased from 7,100 to 170,000 sitesicell between 2 and 10 days of culture following monolayer formation. Time- and concentration-dependent studies on the stimulation of cCMP levels by A N P indicated that guanylate cyclase-linked A N P receptors were present in BMECs. The relative abilities of ANP, brain natriuretic peptide (BNP), and a truncated analog of A N P containing amino acids 5-27 (ANP 5-27) to modulate the accumulation of cCMP was found to be A N P > BNP >> ANP 5-27. Affinity cross-linking with disuccinimidyl suberate and radiolabeled A N P followed by gel electrophoresis under reducing conditions demonstrated a single band corresponding to the 60-70 kD receptor, indicating the presence of the nonguanylate cyclase-linked ANP receptor. Radiolabeled A N P binding was examined in the presence of various concentrationsof either ANP, BNP, or A N P 5-17 and suggested that a large proportion of the A N P receptors present in blood-brain barrier endothelial cells bind all of these ligands similarly. These data indicate both guanylate cyclase linked and nonguanylate cyclase linked receptors are present on BMECs and that a higher proportion of the nonguanylate cyclase linked receptors is expressed. This in vitro culture system may provide a valuable tool for the examination of ANP receptor expression and function in blood-brain barrier endothelial cells.

The concept of a blood-brain barrier (BBB) was established in the late 1960s. Reese and coworkers (Reese and Karnovsky, 1967; Brightman and Reese, 1969) used peroxidases to illustrate that brain endothelial cells were responsible for the permeability properties of the BBB. Unlike capillaries throughout the rest of the body, cerebral microvessels are characterized by epitheliallike tight junctions, minimal pinocytic activity, and a higher mitochondria1 content. These unique properties of the cerebral vasculature are largely responsible for regulating the chemical environment of the brain. Tissue transplants have demonstrated that trophic factors secreted by the brain, in particular the glial cells, modulate the expression of genes responsible for the BBB phenomenon within the brain capillary endothelia (Stewart and Wiley, 1981). Several studies using autoradiographic distributions have suggested the presence of ANP receptors in various regions of the brain, primarily in the subfornical organ or other regions not protected by the BBB (Quirion et al., 1984, 1986; Bianchi et al., 1986; Saavedra et al., 1986).However, the apparent binding of ANP 0 1991 WILEY-LISS, INC.

to brain microvessels, the resultant increase in cGMP (Charbrier et al., 1987, 1988; Steardo and Nathanson, 19871, and the 5,000-fold greater surface area of the brain microvessels as compared to regions not protected by the BBB (Pardridge et al., 1986) suggest a physiologically significant role for circulating ANP in the fluid homeostasis of the brain. Receptors for ANP have been identified in several cell and tissue types, including endothelial cells (Schenk et al., 1985; Leitman and Murad, 1986; Leitman et al., 1986), brain microvessel endothelial cells Received June 29, 1990; accepted October 2, 1990. *To whom reprint requestslcorrespondence should be addressed. Abbreviations used ANP, human atrial natriuretic peptide (28 amino acids); ANP 5-27, truncated rat atrial natriuretic peptide (amino acids 5-27 with a tyrosine substituted at position 8);BBB, blood-brain barrier; BM, binding medium; BMEC, brain microvessel endothelial cell; BNP, brain natriuretic peptide; cGMP, cyclic guanosine monophosphate; IBMX, 3-isobutyl-l-methylxanthine; Kd,equilibrium dissociation constant (MI; RIA, radioimmunoassay .

44

WHITSON ET AL.

(Smith et al., 1988), and brain microvessels (Chabrier et al., 1987, Steardo and Nathanson, 1987). However, the characterization of the receptors in BBB-type endothelial cells has not been performed in detail. The high-affinity binding sites for ANP consist of at least two distinct receptor subpopulations (for review see Baxter et al., 1988). One of the receptor populations has been found to contain particulate guanylate cyclase activity (Kuno et al., 1986; Paul et al., 1987; Takayanagi et al., 1987). ANP binding to this receptor results in an increase in intracellular cGMP concentrations. Truncated ANP molecules do not bind well to this receptor, nor do they stimulate guanylate cyclase activity to the same extent (Leitman and Murad, 1986; Scarborough et al., 1986). The intracellular portion of this receptor contains two domains with sequence homologies to a protein kinase catalytic domain and a guanylate cyclase catalytic domain (Chinkers and Garbers, 1989; Chinkers et al., 1989). A recent report by Schulz et al. (1989) indicates that within this class of guanylate cyclase-coupled ANP receptors, there are two distinct subtypes. Although the sequence homology of these receptors is high, the ability of various peptides to modulate cGMP accumulation is distinct. Another class of ANP receptors that has been characterized differs in that there is little distinction between truncated ANP analogs and ANP for binding, and cGMP levels are unaffected once ligand is bound to the receptor (Scarborough et al., 1986). In addition, affinity cross-linking followed by reducing and nonreducing SDS gels suggests that this receptor exists as a dimer linked by disulfide bond(s) (Takayanagi et al., 1987). Each monomer of this receptor has an extracellular amino-terminal ANP recognition site, a single hydrophobic transmembrane region, and a short cytoplasmic tail (Fuller et al., 1988). Maack and coworkers (1987) have suggested that these receptors function in the sequestration and metabolic clearance of circulating ANP. Recently, however, Hirata et al. (1989) have suggested that ANP mediates the accumulation of inositol phosphates through this nonguanylate cyclaselinked receptor. The studies described here were conducted to quantitate and characterize ANP receptors present in primary monolayer cultures of bovine BMECs. The presence of at least two types of ANP receptors was indicated by the ANP modulation of cGMP and the partition binding results. Binding studies using ANP and a truncated ANP and affinity cross-linking data suggest that a majority of the ANP receptors are the nonguanylate cyclase-coupled receptors. This is the first study to characterize the types of ANP receptors that exist in the specialized endothelial cells composing the blood-brain barrier. The physiological significance of multiple types of ANP receptors in the brain endothelium is not yet clear, but the presence of at least two ANP receptor types described by these studies suggests an active role for ANP in regulating brain endothelial cell function. This study suggests that the in vitro primary culture of brain microvessel endothelial cells is an important model system for the examination of ANP interactions with the blood-brain barrier. A portion of these results was presented in an abstract (Whitson et al., 1989a).

MATERIALS AND METHODS Materials Human ANP (prohormone sequence 99-126) and rat ANP 5-27 (prohormone sequence 103-125 with a tyrosine substitution at position 106, [Tyr81 ANP 5-27) were purchased from Peninsula Laboratories (Belmont, CA). BNP was purchased from Cambridge Research Biochemicals (Cambridge, England). Disuccinimidyl suberate was from Pierce Chemical Company (Rockford, IL) and donor-defined horse serum was from Hyclone (Logan, UT). [1251]-labeledANP and insulin were purchased from Amersham (specific activity 2000 Ci/mmole, Arlington Heights, IL). Dispase and collagenase/dispase were from Collaborative Research. All other chemicals and medium were from Sigma (St. Louis).

Isolation and culture of bovine BMECs BMECs were isolated from the gray matter of bovine brains following the procedures of Audus and Borchardt (1987). Large blood vessels and meninges were removed from two fresh bovine brains and gray matter was isolated and homogenized in a Stomacher Lab Blender 400 for 1.5 min. The homogenate was digested in dispase (1ml of a 0.5% solution/100 ml of homogenate) for 3 h at 37"C, with addition of Minimal Essential Medium, pH 9-10, after 30 min to raise the pH to 7.2-7.4. After centrifugation and se aration in a 13% dextran gradient, the pellets were igested further in collagenaseldispase (1mg/ml) for 4.5 h at 37°C. The cell material was layered onto a pre-established 50% Percoll gradient and centrifuged as described previously (Bowman et al., 1981). The second band was isolated and washed to remove Percoll. The final pellets were resuspended in culture medium (1:l ratio of F-12 and Minimal Essential Medium) containing 20% horse serum and 10%dimethylsulfoxide and frozen at - 70°C. BMECs were plated at a density of approximately 50,000 cells/cm2 in 12- or 6-well plates coated with freshly prepared rat tail collagen and fibronectin (Audus and Borchardt, 1987). The environment was controlled at 37°C in a water-saturated atmosphere of 95% air-5% cop.Cells were grown to confluence (7-10 days) before binding assays or cross-linking experiments were performed. During this culture period, culture medium was replaced after 3 days and every other day until the cells reached confluence, after which medium was replaced daily.

if

ANP binding to BMECs ANP binding was determined as previously described (Smith et al., 1989; Whitson et al., 1989b) with the following modifications. Cells were washed twice (approximately 1 min) with the binding medium (BM, 1:l ratio of F-12 and Minimal Essential Medium, pH 7.4, with 50 kg/ml of bovine serum albumin, 4°C).Triplicate wells contained a constant amount of radiolabeled ANP (0.01-0.04 nM), unlabeled ANP at final concentrations of 0 to 100 nM, and BM to a final volume of 0.4 ml. Radiolabeled ANP and unlabeled ANP, BNP, and ANP 5-27 stock solutions were prepared by dissolving the lyophilized peptides in 0.1% acetic acid, aliquoting in small volumes and storing frozen (-20°C). Freshly

ANP RECEPTORS IN THE BLOOD-BRAIN BARRIER

prepared dilutions were made just prior to the binding assays. The plates were incubated for 60-75 min at 1616°C to minimize internalization of bound receptorANP complexes because l-h incubations at 37°C result in 100%of total bound counts in the interior of the cell (Smith et al., 1989). After incubation, cells were washed twice with the BM and solubilized in 0.5 ml of 1N NaOH. A fraction of the solubilized cells was counted on a Packard Auto-Gamma 500 counter, and protein concentration was determined from 0.05 ml from 4 wells on each plate using the method of Lowry (1951) as modified by Peterson (1977). Nonspecific counts, i.e., counts bound in the presence of 100 nM ANP, were subtracted from total bound counts at each point, resulting in specifically bound counts. Scatchard (1949) analysis was used to determine the apparent binding constant and the number of ANP binding sites per cell. Data were also analyzed in terms of the mg of total protein. Stability of the radiolabeled ANP in binding medium was monitored by HPLC. Binding medium containing 1%bovine serum albumin caused significant degradation of ANP in 1 h. Decreasing the bovine serum albumin to 50 wg/ml resulted in a stable ligand for the duration of the binding assay. Insulin binding to BMECs Insulin binding was performed in the same medium, under the same conditions described above. [ 1251]-labeled insulin was added to a final concentration of 0.3 nM. Nonspecific binding was determined in the presence of 1,000 nM unlabeled insulin. Affinity cross linking, electrophoresis, and autoradiography Before cross-linking ANP to the specific receptors, the cells were incubated with radiolabeled ANP (0.5-1 nM) in either the presence or absence of unlabeled ANP for 60-75 min at 14-16°C. After reaching equilibrium, the cells were washed twice with ice-cold phosphatebuffered saline. The cross-linking procedure was performed as described by Whitson et al. (1989b) with the following modifications. A 100 mM solution of disuccinimidyl suberate was freshly prepared in dimethyl sulfoxide and diluted into 1 ml of phosphate-buffered saline for a final concentration of 1mM. The intact cells were incubated for 30 min at 16°C before aspirating the reaction solution and stopping further reaction by washing the cells three times with TE (0.01 M TrisHCl, pH 7.5, 0.001 M ethylenediamine tetraacetic acid). Samples were electrophoresed on a 7.5% polyacrylamide gel, with a 4.5% stacking gel, as described by Laemmli (1970). Polyacrylamide gels were dried and placed in exposure envelopes with X-ray film (Kodak X-OMAT AR) and intensifying screens. The exposure duration was 5-10 days at -70°C. Molecular weights of the bands visualized by autoradiography were assigned by comparison to the log plot of standard marker proteins and their mobilities in the gel (Fairbanks et al., 1971). Radioimmunoassay of ANP Sep-Pak C 18s from Waters (Milford, MS) were activated with 10 ml of methanol and 10 ml of deionized water. Extraction procedures were performed as de-

45

scribed previously (Sugawara et al., 1985; Chen et al., 1990) using 4 ml of BMEC-conditioned medium (24 h). The eluates from the Sep-Pak were dried under nitogen flow at 37°C. The dried extracts were reconstituted in 0.5 ml of radioimmunoassay (RIA) buffer (0.1 M phosphate buffer, pH 7.4, 0.05 M NaCl, 0.1% bovine serum albumin, 0.1% Triton X-100, 0.01% NaN3). RIA procedures were performed according to the methods described by Peninsula Laboratories (Belmont, CA). Radioimmunoassay of cGMP Monolayers of BMECs were preincubated for 10 min at 37°C in BM with 0.5 mM 3-isobutyl-l-methylxanthine (IBMX). ANP, BNP, or ANP 5-27 was added to the culture medium in duplicate wells at the concentrations indicated in the figure. After incubation at 37°C for the times indicated in the figure, the medium was aspirated and an acid/ethanol solution (1 ml 1N HCl in 99 ml95% ethanol, 4°C) was added to the cells. Cells were scraped from the culture dishes, placed into centrifuge tubes, and vacuum dried. Samples were stored at -70°C until RIA. RIA kits were purchased from Biomedical Technologies Inc. (Stoughton, MA) and performed as described for acetylated samples. RESULTS Characterization of BMECs Bovine BMECs isolated by the methods of Audus and Borchardt (1987) have been characterized by transmission electron microscopic examination, histochemical alkaline phosphatase staining, specific activity determinations of gamma-glutamyl transpeptidase in isolated vs. homogenate fractions, Factor VIII positive staining and transendothelial transport properties. We used Factor VIII, a marker for endothelial cells, for immunofluorescent staining (data not shown) to confirm the endothelial origin of the BMECs isolated in our laboratory. Greater than 95% of the cells we isolated were judged to be Factor VIII positive. ANP binding to BMECs as a function of culture time Primary cultures of BMECs were established as described in Materials and Methods. ANP synthesized by these cells or found in the medium was determined M per well. Thus no by RIA to be less than 3 x long washing periods were necessary to dissociate bound ANP before binding assays were performed. ANP binding to these cells after monolayer formation was determined as a function of the culture time. Cell number and total protein were also measured. Figure 1 illustrates a gradual increase and plateau in cell number and a progressive increase in total rotein per well during this time. Scatchard analysis o the ANP binding data from a single BMEC isolation are presented in Figure 2. Binding assays were performed 2,4,6, and 10 days after monolayer formation. Figure 3 depicts the increase in ANP binding sites as a function of culture days after monolayer formation; data were obtained using cells from six separate BMEC isolations. Between 2 and 10 days after the monolayer was formed, the number of ANP binding sites increased 24-fold, from 7,100 to 170,000 ANP binding sites per cell, and the apparent Kd increased 10-fold,from 0.25 to 2.5 nM.

P

WHITSON ET AL.

46

r

I

I

I

I

0.6 (D

0 7

x U

w

0.4

Elz j

0.2

W

0

0 I

I

I

100000

150000

200000

DAYS AFTER MONOLAYER FORMATION 0

50000

BOUND

(siteslcell)

Fig. 2. Scatchard analysis of ANP binding to BMECs. ANP binding was performed 2, 4, 6,and 10 days following monolayer formation using the techniques described in Materials and Methods. The radiolabeled ANP concentration was 0.01-0.04 nM for each binding assay. Scatchard (1949) analysis of the competition binding data are illustrated. Least squares analysis of the points resulted in R values between 0.90-0.94 for each set of data. Increases in both the number of binding sites (x-intercept) and the apparent K, (lislope) were observed with increased culture time beyond monolayer formation. 2 days post monolayer W, 4 days post-monolayer (W, 6 days postmonolayer ( A), and 10 days post-monolayer ( + ).

I 0

2

4

6

8

1

0

1

2

DAYS AFTER MONOLAYER FORMATION

Fig. 1. Cell number and total protein content of BMECs after monolayer formation. Cell number (A) was monitored by counting nuclei from duplicate trypsinized wells as a function of days after the monolayer was formed. B illustrates the total protein determined from quadruplicate wells using the method of Lowry et al. (1951) as modified by Peterson (1977).Data indicate that the cells are no longer proliferating 6 days after the monolayer was formed even though the total protein continued to increase.

The equivalent number of ANP binding sites calculated by Scatchard analysis relative to the total protein increased from 0.022 to 0.70 pmole sites per mg total protein. Insulin binding to BMECs as a function of culture time Insulin binding was also examined in these cells to determine whether insulin receptors were equally affected by the culture time beyond confluence. The fraction of radiolabeled insulin specifically bound per mg total protein was 3.0, 3.2, and 4.0% of the total radiolabeled insulin after 2, 6, and 10 days of culture beyond monolayer formation. These data compare to 11.5, 17.7, and 41.0%of the radiolabeled ANP specifically bound per mg of total protein in identically cultured cells after 2, 6, and 10 days of culture. Peptide-modulated cGMP levels in BMECs ANP binding to the guanylate cyclase-coupledreceptor (120-140 kD) is characterized by increased cGMP levels. A time- and dose-dependent cGMP response to

ANP was observed in the primary cultures of BMECs (Figs. 4,5). Addition of ANP (10 nM) resulted in a 4-t o 8-fold increase in cGMP levels above the control within 20 sec after addition. Culture time slightly increased the ANP-inducible cGMP response when values were normalized per million cells; ANP (10 nM) increased cGMP concentrations above control to 1.30 0.20 pmol/million cells/5 min after 2 to 4 days of culture beyond monolayer formation (n = 3) and to 1.52 0.19 pmol/million celld5 min after 8-10 days of culture (n = 3). In order to determine whether cGMP levels were associated with particulate vs. soluble guanylate cyclase, cGMP levels induced by 250 nM ANP were measured in the presence of 100 p.M methylene blue, a known inhibitor of soluble guanylate cyclase (Ignarro et al., 1986). The presence of 100 pM methylene blue did not alter cGMP accumulation. The relative ability of ANP, BNP or ANP 5-27 to modulate cGMP accumulation was also determined (Fig. 5). These ligands were selected to distinguish the two subtypes of guanylate cyclase-coupled receptors based on their differential abilities to stimulate cGMP as described by Schulz et al. (1989). In our studies the relative abilities of these ligands to modulate cGMP accumulation was found to be ANP > BNP >> ANP 5-27. Saturation of cGMP accumulation was not appar-ent even at 1 kM. ANP 5-27 at 250 nM did not stimulate cGMP levels above those observed for ANP at 1 nM (less than 0.40 % 0.02 pmol/million cells/5 min). To distinguish between a lack of ANP 5-27 binding to the guanylate cyclase-coupled receptor or binding to this receptor with no cGMP accumulation, cGMP levels were quantitated for 10 nM ANP in the absence or presence of 100

*

ANP RECEFTORS IN THE BLOOD-BRAIN BARRIER 2.8I

106

47 I

I

I

2.4 2.0 1.6

105

1.2

0.8 I

o4

0.4 0 LOG [peptide]

DAYS AFTER MONOLAYER FORMATION

Fig. 3. ANP bindine siteslcell as a function of culture time bevond monolayer formation, The number of ANP binding siteskell, d c u lated from Scatchard (1949)analysis of the binding data, are plotted on a log scale as a function of the days after monolayer formation. Each point represents the mean number of sitesicell from the indicated number of separate experiments (n). Error bars depict the standard deviation. A 24-fold increase in the number of ANP receptorsicell was monitored over 8 days. 1 .o

I

I

I

I

T

-.-

t

3.6r

3.2

S

I

I

I

I

I

UI

1 "

-10

-9

-8

-7

-6

LOG [peptide]

0

I

I

I

I

I

2

4

6

8

10

Fig. 5. Accumulation of cGMP as a function of ANP, BNP and ANP 5-27 concentrations. The dose-dependenceof cGMP levels upon addition of the indicated concentration of ANP (o,.), BNP (u,.) or ANP 5-27 ( A , A) for 5 min is illustrated. Data in A represent the mean values 3 4 days after monolayer formation (n = 3) and data in B represent the mean values 7-10 days after monolayer formation (n = 2). Values represent accumulation in excess of the cGMP levels of an unstimulated IBMX-containingcontrol.

pre-equilibrated with radiolabeled ANP are illustrated in Figure 6. A band corresponding to 64 kD was observed. Specificity of cross-linking was demonstrated Fig. 4. ANP stimulation of cGMP levels: time-dependence. Concen- in the control lane (+) in which intact cells were tration of cGMP in BMECs was assayed by RIA as described in preincubated with a 300-fold excess of unlabeled ANP Materials and Methods. All incubations were performed in the presence of IBMX, a phosphodiesterase inhibitor. Using 10 nM ANP, a in addition to the same concentration of labeled ANP. rapid time-dependent increase in cGMP was noted. Cyclic GMP values The appearance of specifically cross-linked radioactive illustrated are the difference between ANP-induced levels and com- material at the top of the gel suggests that secondary parable controls exposed to IBMX for the same time. Error bars cross-linking of the ANP-receptor complex to other represent the standard deviation of duplicate points from a single proteins was occurring. No cross-linking was observed experiment. in the region that would correspond to a molecular weight of 120-140 kD,the molecular weight of the nM ANP 5-27. ANP 5-27 did not effectively alter ANP guanylate cyclase-coupled receptor. stimulation of cGMP even though its concentration was Partition effects of radiolabeled ANP by 10 times that of ANP (data not shown). unlabeled ANP, BNP,and ANP 5-27 Affinity-cross-linking of radiolabeled ANP Binding curves were generated using various concenThe results of cross-linking, denaturing SDS-gel trations of unlabeled ANP, BNP, and ANP 5-27 in the electrophoresis, and autoradiography of intact BMECs presence of a constant radiolabeled ANP concentration. TIME (min)

WHITSON ET AL.

48

+

I

0 2

3

8

200

-

100

80

60

116 97

-

66-

40

20

0 -10

43-

-9

-a

-7

-6

LOG [peptide]

Fig. 7. Radiolabeled ANP binding in the presence of various concentrations of unlabeled ANP, BNP, or ANP 5-27. This figure illustrates the competition of radiolabeled ANP with ANP, BNP, or truncated ANP 5-27. A constant concentration of radiolabeled ANP (0.03 nM) was equilibrated with each of the indicated concentrations of ANP (a), BNP (W, or ANP 5-27 (A). The percent of the total radiolabeled ANP bound at each concentration of the unlabeled ligand is indicated. Data represent the mean percent change at each concentration from three separate experiments for each peptide. Fig. 6. Affinity cross-linking of radiolabeled ANP to intact BMECs. Affinity cross-linking procedures were performed using disuccinimidyl suberate and radiolabeled ANP(-) as described in Materials and Methods. The molecular weights corresponding to the mobility of marker proteins are indicated in kD.An equal volume of the control (+I, which was incubated with radiolabeled ANP in the presence of a 300-fold excess of unlabeled ANP, was also eledrophoresed to demonstrate the specificity of the cross-linking. A single distinct band with a mobility that corresponded to a molecular weight of 64 kD was observed.

A right-ward shift in the curve using a truncated ANP analog relative to the curve using ANP would indicate a lower affinity for the truncated ligand. These characteristics would be consistent with a large proportion of guanylate cyclase-linked ANP receptors (Agui et al., 1989; Fethiere et al., 1989; Uchida et al., 1989). In contrast, the presence of similar curves would indicate near equivalent affinities for the ligands, as previously described for the nonguanylate cyclase linked receptors. Similar partitioning effects of radiolabeled ANP by either ANP, BNP or ANP 5-27 are depicted in Figure 7. DISCUSSION The specialized endothelial cells of the blood-brain barrier control the exchange of nutrients and fluids between the brain and the circulatory system. Defective re ulation of BBB permeability can result in cerebra edema and cause brain injury by limiting tissue perfusion or increasing intracranial pressure. The culture of-brain microvessel endothelial cells as a

P

model system of the blood-brain barrier has been used previously to initiate studies to precisely define BBB permeability and metabolic mechanisms as well as the factors regulating these characteristics (Bowman et al., 1983; Audus and Borchardt, 1987). This work reports the characterization of ANP receptors in the bloodbrain barrier using primary cultures of brain microvessel endothelial cells. A better understanding of the ANP receptors present in these s ecialized cells will provide insight into the role of t is fluid-regulating hormone as it relates to the fluid homeostasis of the brain. The number of ANP binding sites and the apparent Kd shortly after monolayer formation in cultured bovine aortic endothelial cells have been reported as 16,000 sites per cell and 0.1 nM, respectively (Leitman and Murad, 1986). Primary bovine BMECs were found to have a binding capacity of 0.052 pmol per mg total cell protein and a Kd of 0.4 nM (Smith et al., 1988). Scatchard (1949) analysis of ANP binding to BMECs resulted in a linear plot, suggesting a single class of binding sites, multiple receptors with similar affinities, or an abundance of one receptor over another (Tran et al., 1985;Leitman and Murad, 1986). Primary bovine brain microvessel endothelial cells isolated in our laboratory expressed 7,100 ANP binding sites per cell (0.022 pmole per mg total protein) 2 days after monolayer formation. However, the number of sites increased approximately 24-fold over a period of 8 days. The Kd we observed ranged from 0.25 nM 2 days after monolayer formation to 2.5 nM 10 days after monolayer

R

ANP RECEPTORS IN THE BLOOD-BRAIN BARRIER

formation. The increase in the Kd as a function of culture time may be representative of different proportions of multiple receptor types, or due to changes in receptor affinities with differentiation of the cells. Our binding data in the presence of various concentrations and types of ligands would suggest that the proportions of the receptor types were not dramatically altered as a function of culture time. However, it is currently not possible to determine whether the observed increase in Kd with culture time results from a small change in the ratio of the receptor populations or altered affinities of one and/or both receptors during the differentiation of the cells. Even though our values for the number of ANP binding sites and the Kd increase dramatically with culture time, they are comparable to previous reports in BMECs or aortic endothelial cells shortly after monolayer formation. Raub et al. (1989) observed alterations in transendothelial electrical resistance in bovine BMECs cultured on polycarbonate membranes; resistance initially increased, reaching a maximum approximately 6 days after monolayer formation, and then decreased. Their data suggests that a differentiation process or a process in which tight junctions are formed is occurring immediately following formation of the monolayer. Our data indicated that ANP receptors were being modulated during culture, possibly as a response to this differentiation or maturation process. The lack of change in the fraction of insulin bound during this same period in our experiments indicates that the increase in ANP receptor number is specific and not the result of a general up-regulation of all receptors. Thus the increase in receptor number as a function of culture time may be linked to the cellular differentiation response. The partitioning effects of radiolabeled ANP with various concentrations of unlabeled ANP or a truncated ANP analog have been used previously to indicate the relative proportions of the guanylate cyclaselinked and nonguanylate cyclase-linked ANP receptors (Agui et al., 1989; Fethiere et al., 1989; Uchida et al., 1989). Cells containing only guanylate cyclase-linked ANP receptors exhibit a shift in binding data with different ANP ligands reflecting the significantly lower affinity of this receptor for the truncated ANPs. Therefore, higher concentrations of the unlabeled truncated ANP are required to decrease the saturation of radiolabeled ANP relative to partitioning with unlabeled ANP, resulting in a shift of the partitioning curves. In contrast, in cells with only the nonguanylate cyclase-linked ANP receptors, the curves are very similar due to the lack of discrimination of this receptor for the various ANP ligands. Results from our studies examined the ability of ANP, BNP, and ANP 5-27 t o partition with radiolabeled ANP for specific binding to BMECs throughout the 10-day culture period after monolayer formation. All three peptides exhibited similar partition effects throughout the culture period. These data indicate that the majority of the receptors present in BMECs are not selective in their ability to bind these three peptides, suggesting an abundance of the low molecular weight nonguanylate cyclase-linked receptor. The cGMP accumulation, however, provides direct evidence for the presence of the guanylate cyclase-linked receptors. Taken together, the binding

49

data and the cGMP results indicate that both guanylate and nonguanylate receptors are present, but are expressed on the cell surface in different proportions. Affinity cross-linking of radiolabeled ANP to intact BMECs, followed by denaturing gel electrophoresis, resulted in one band that corresponded to a molecular weight of 64 kD. This value is consistent with the molecular weight of the nonguanylate cyclase-linked receptor (Takayanagi et al., 1987). No band was observed at 120-140 kD, indicating that either there are fewer of the guanylate cyclase-coupled receptors or the efficiency of cross-linking to these receptors is diminished. Since other investigators have previously crosslinked both receptor types simultaneously using similar methodologies (Leitman et al., 1986; Rathinavelu and Isom, 19881, these data again indicate that there are a low number of guanylate cyclase-coupled receptors relative to nonguanylate cyclase-coupled ANP receptors in BMECs. Recently Schulz et al. (1989) have identified two subtypes of guanylate cyclase-coupledreceptors, GC-A and GC-B, with distinct binding and cGMP accumulation characteristics. A typical sigmoidal dose response was observed for the GC-A receptor with an EC50 of 3 nM for ANP, 25 nM for BNP, and 65 nM for a truncated ANP (ANP 5-25). For the GC-B receptors, the dose response of cGMP did not appear to be saturable even at 1 pM. In addition, BNP exhibited greater cGMP accumulation than ANP, and ANP 5-25 only minimally stimulated the accumulation of cGMP. As discussed by Schulz et al. (1989),the apparently nonsaturable stimulation of cGMP with these peptides suggests that the natural ligand may not have been identified for these receptors, or that this response may be secondary in the action of this receptor for signal transduction, i.e., the protein kinaselike domain may be involved as well. In our studies, ANP induced dose- and time-dependent increases in cGMP levels in BMECs, providing evidence for the existence of the guanylate cyclase-coupled ANP receptors in these primary cultures. The dose response of cGMP accumulation for ANP, BNP, and ANP 5-27 appeared similar to that of the GC-B receptors described by Schulz et al. (1989) in that the response did not appear to be saturable at 1 kM. However, in the BMECs, ANP stimulated cGMP accumulation to a reater degree than BNP, in contrast to the results of chulz et al. (1989). A number of explanations could be provided for these data; a mixture of GC-A and GC-B receptors could exist in this cell type, a third type of guanylate cyclase-coupledreceptor may exist or species variation of the GC-B receptor in bovine BMECs may alter the relative peptide-stimulated cGMP accumulation ability in comparison to that isolated from rat brain as described by Schulz et al. (1989). The lack of significant cGMP accumulation in response to ANP 5-27 in our cells is consistent with the observations of Schulz et al. (1989) for the GC-B receptor. These data would suggest that a mixture of GC-A and GC-B receptors is unlikely since the GC-A receptors exhibited an EC50 of 65 nM for a similar truncated analog. The molecular characterization of the receptor will be necessary to distinguish between the latter two possibilities. In order to confirm that this cGMP accumulation was

8

50

WHITSON ET AL.

the result of particulate guanylate cyclase instead of an indirect effect on soluble guanylate cyclase, cGMP accumulation was measured in the presence of methylene blue, a known inhibitor of soluble guanylate cyclase (Ignarro et al., 1986). Fethiere et al. (1989) demonstrated in two cell lines, each containing either the guanylate cyclase-coupled or the nonguanylate cyclase-coupled receptors, that methylene blue altered cGMP accumulation in the cell line containing only the nonguanylate cyclase-coupled receptors. Since we observed no significant difference in cGMP accumulation in the presence of methylene blue, we conclude that membrane-bound guanylate cyclase-linked receptors are responsible for the ANP-inducible accumulation of cGMP. To summarize, our data indicate that at least two ANP receptor types exist in primary brain microvessel endothelial cells. The majority of the receptors in this specialized endothelial tissue appear by affinity crosslinking and binding studies to be the low molecular weight nonguanylate cyclase-coupled receptor. The pattern of cGMP accumulation in response to ANP, BNP and ANP 5-27 indicates the presence of a guanylate cyclase-linked ANP receptor; however, the cGMP responses did not correspond directly to either the GC-A or GC-B receptors described by Schulz et al. (1989). The presence of at least two types of ANP receptors suggests dual or multiple mechanisms by which ANP may be involved in the fluid homeostasis of the BBB. The fact that ANP receptors appear to be up-regulated after monolayer formation, which corresponds to the increased transendothelial electrical resistance measured by other investigators (Raub et al., 1989), provides evidence for the possibility that ANP receptors are modulated during the differentiation/ development of a BBB-like structure. Up-regulation of the receptors may be the result of transcriptional or translational regulation, or possibly the translocation of receptors from the interior to the surface of the cell. Further experiments will be necessary to determine the mechanism by which differentiation or development of brain endothelial cells modulates ANP receptor expression. The results of this study suggest that primary cultures of brain microvessel endothelial cells may provide an excellent model for the investigation of ANP involvement in blood-brain barrier function.

Audus, K.L., and Borchart, R.T. (1987) Bovine brain microvessel endothelial cell monolayers as a model system for the blood-brain barrier. Annals of N.Y. Acad. Sci., 507:9-18. Baxter, J.D., Lewicki, J.A., and Gardner, D.G. (1988) Atrial natriuretic peptide. BioiTechnology, 6:529-546. Bianchi, C., Gutkowska, J., Ballak, M., Thibault, G., Garcia, R., Genest, J., and Cantin, M. (1986) Radioautographic localization of '*%atrial natriuretic factor binding sites in brain. Neuroendocrinology, 44:365-372. Bowman, P.D., Betz, A.L., and Goldstein, G.W. (1982) Primary culture of microvascular endothelial cells from bovine retina: selective growth using fibronectin coated substrate and plasma derived serum. In Vitro, 18:626-632. Bowman, P.D., Ennis, S.R., Rarey, K.E., Betz, A.L., and Goldstein, G.W. (1983) Brain microvessel endothelial cells in tissue culture: a model for study of blood-brain barrier permeability. Ann. Neurol., 14:396-402. Brightman, M.W., and Reese, T.S. (1969) Junctions between intimately apposed cell membranes in the vertebrate brain. J. Cell Biol., 40:64&677. Budzik, G.P., Firestone, S.L., Bush, E.N., Connolly, P.J., Rockway, T.W., Sarin, V.K., and Holleman, W.H. (1987) Divergence of ANF analogs in smooth muscle cell cGMP response and aorta vasorelaxation: evidence for receptor subtypes. Biochem. Biophys. Res. Comm., 144:422-431. Chabrier, P.E., Roubert, P., and Braquet, P. (1987) Specific binding of atrial natriuretic factor in brain microvessels. Proc. Natl. Acad. Sci. USA, 84:207a2081. Chabrier, P.E., Roubert, P., Plas, P., and Braquet, P. (1988) Bloodbrain barrier and atrial natriuretic factor. Can. J. Physiol. Pharmacol., 66:276-279. Chen, Y.-M., Whitson, P.A., and Cintron, N.M. (1990) Immunoreactive prohormone atrial natriuretic peptides 1-30 and 31-67: existence of a single circulating amino-terminal peptide. Biochem. Biophys. Res. Comm., 166:794-800. Chinkers, M., and Garbers, D.L. (1989) The protein kinase domain of the ANP receptor is required for signaling. Science, 245:1392-1394. Chinkers, M., Garbers, D.L., Chang, M.-S., Lowe, D.G., Chin, H., Goeddel, D.V., and Schulz, S. (1989)A membrane form of guanylate cyclase is a n atrial natriuretic peptide receptor. Nature, 338:78-83. Fairbanks, G., Steck, T.L., and Wallach, D.F.H. (1971) Electrophoretic analysis of the major polypeptides of the human erythrocyte membrane. Biochemistry, I0:26062617. Fethiere, J., Meloche, S., Nguyen, T.T., Ong, H., and De Lean, A. (1989) Distinct properties of atrial natriuretic factor receptor subpopulations in epithelial and fibroblast cell lines. Molecular Pharmacology, 35586592. Fuller, F., Porter, G.J., Arfsten, A.E., Miller, J., Schilling, J.W., Scarborough, R.M., Lewicki, J.A., and Schenk, D.B. (1988). Atrial natriuretic peptide clearance receptor: complete sequence and functional expression of cDNA clones. J. Biol. Chem., 263:9395-9401. Hirata, M., Chan, C.-H., and Murad, F. (1989) Stimulatory effects of atrial natriuretic factor on phosphoinositide hydrolysis in cultured bovine aortic smooth muscle cells. Biochem. Biophys. Acta, 1010:34&351. Ignarro, L.J., Harbison, R.G., Wood, K.S., and Kadowitz, P.J. (1986) Dissimilarities between methylene blue and cyanide on relaxation and cyclic GMP formation in endothelium-intact intrapulmonary artery caused by nitrogen oxide-containing vasodilators and acetylACKNOWLEDGMENTS choline. J. Pharmacol. Exp. Ther., 236:30-36. T., Andresen, J.W., Kamisaki, Y., Waldman, S.A., Chang, L.Y., The authors gratefully acknowledge the excellent Kuno, Saheki, S., Leitman, D.C., Nakane, M., and Murad, F. (1986) technical assistance of Dr. Y.-M. Chen and J.V. FesCo-purification of an atrial natriuretic factor receptor and particulate guanylate cyclase from rat lung. J . Biol. Chem., 261:5817perman of KRUG Life Sciences for the RIA of ANP and 5823. cGMP. The authors also express gratitude to Dr. KenU.K. (1970) Cleavage of structural proteins during the neth L. Audus of the University of Kansas for his Laemmli, assembly of the head of bacteriophage T4.Nature, 277:680-685. assistance in establishing the BMEC isolation proce- Leitman, D.C., and Murad, F. (1986) Comparison of binding and cyclic dure in our laboratory, and to Dr. Nitza M. Cintron of GMP accumulation by atrial natriuretic peptides in endothelial cells. Biochem. Biophys. Acta, 885:74-79. NASAiJohnson Space Center for her support and inD.C., Andresen, J.W., Kuno, T., Kamisaki, Y., Chang, J.K., terest in this project. This work was supported by a Leitman, and Murad, F. (1986) Identification of multiple binding sites for grant from NASA-Regulatory Physiology 199-18-11-15 atrial natriuretic factor by affinity cross-linking in cultured endoto P.A.W. thelial cells. J . Biol. Chem., 261:1165&11655. Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J. (1951) Protein measurement with the folin phenol reagent. J . Biol. Chem., LITERATURE CITED 193:265-275. Agui, T., Kurihara, M., and Saavaedra, J.M. (1989) Multiple types of Maack, T., Suzuki, M., Almeida, F.A., Nussenzveig, D., Scarborough, receptors for atrial natriuretic peptide. Eur. J. Pharmacol., R.M., McEnroe, G.A., and Lewicki, J.A. (1987) Physiological role of silent receptors of atrial natriuretic factor. Science, 238:675-677. 162:301307.

ANP RECEPTORS IN THE BLOOD-BRAIN BARRIER Pardridge, W.M., Oldendorf, W.H., Cancilla, P., and Frank, H.J.L. (1986)Blood-brain barrier: interface between internal medicine and the brain. Ann. Internal Med., 105:82-95. Paul., A.K., Marala, R.B., Jaiswal, R.K., and Sharma, R.K. (1987) Coexistence of guanylate cyclase and atrial natriuretic factor receptor in a 180-kD protein. Science, 23511224-1226. Peterson, G.L. (1977) A simplification of the protein assay method of Lowry et al. which is more generally applicable. Anal. Biochem., 83:346-355. Quirion, R., Dalpe, M.,and Dam, T.-V. (1986) Characterization and distribution of receptors for the atrial natriuretic peptides in mammalian brain. Proc. Natl. Acad. Sci. USA, 831176178. Quirion, R., Dalpe, M., De Lean, A,, Gutkowska, J., Cantin, M., and Genest, J. (1984) Atrial natriuretic factor (ANF) binding sites in brain and related structures. Peptides, 5:1167-1172. Rathinavelu, A., and Isom, G.E. (1988). High affinity receptors for atrial natriuretic factor in PC12 cells. Biochem. Biophys. Res. Comm., 156:78-85. Raub, T.J., Kuentzel, S.L., and Sawada, G.A. (1989) Characterization of primary bovine brain microvessel endothelial cell monolayers on membrane filters: permeability of glial cell induced changes. J. Cell Biol., 109:315a. Reese, T.S., and Karnovsky, M.J. (1967)Fine structural localization of a blood-brain barrier to exogenous peroxidase. J. Cell Biol., 34:207217. Saavedra, J.M., Correa, F.M.A., Plunkett, L.M., Israel, A., Kurihara, M., and Shigematsu, K. (1986) Binding of angiotensin and atrial natriuretic peptide in brain of hypertensive rats. Nature, 320:758760. Scarborough, R.M., Schenk, D.B., McEnroe, G.A., Arfsten, A., Kang, L., Schwartz, K., and Lewicki, J.A. (1986) Truncated atrial natriuretic peptide analogs: Comparison between receptor binding and stimulation of cyclic GMP accumulation in cultured vascular smooth muscle cells. J. Biol. Chem., 261:12960-12964. Scatchard, G. (1949) The attractions of proteins for small molecules and ions. Ann. N.Y. Acad. Sci., 51:660-672. Schenk, D.B., Johnson, L.K., Schwartz, K., Sista, H., Scarborough, R.M., and Lewicki, J.A. (1985) Distinct atrial natriuretic factor receptor sites on cultured bovine aortic smooth muscle and endothelial cells. Biochem. Biophys. Res. Comm., 127:433-442.

51

Schulz, S., Singh, S., Bellet, R.A., Singh, G., Tubb, D.J., Chin, H., and Garbers, D.L. (1989) The primary structure of a plasma membrane guanylate cyclase demonstrates diversity within this new receptor family. Cell, 58:1155-1162. Smith, K.P., Kato, A., and Borchardt, R.T. (1988) Characterization of specific receptors for atrial natriuretic factor on cultured bovine brain capillary endothelial cells. Biochem. Biophys. Res. Comm., 157:30%314. Steardo, L., and Nathanson, J.A. (1987) Brain barrier tissues: end organs for atriopeptins. Science, 235470-483. Stewart, P.A., and Wiley, M.J. (1981) Developing nervous tissue induces formation of blood-brain barrier characteristics in invading endothelial cells: A study using quail-chick transplantation chimeras. Dev. Biol., 841183-192. Sugawara, A., Nakao, K., Morii, N., Sakamoto, M., Suda, M., Shimokura, M.,Kiso, Y., Kihara, M.,Yamori, Y., Nishimura, K., Soneda, J., Ban, T., and Imura, H. (1985) Alpha-human atrial natriuretic polypeptide is released from the heart and circulates in the body. Biochem. Biophys. Res. Comm., 129:439-446. Takayanagi, R., Snajdar, R.M., Imada, T., Tamura, M., Pandey, K.N., Misono, K.S., and Inagami, T. (1987)Purification and characterization of two types of atrial natriuretic factor receptors from bovine adrenal cortex: guanylate cyclase-linked and cyclase-free receptors. Biochem. Biophys. Res. Comm., 144:244-250. Tran, V.T., Beal, M.F., and Martin, J.B. (1985) Two types of somatostatin receptors differentiated by cyclic somatostatin analogs. Science, 228:492-495. Uchida, K., Mizuno, T., Shimonaka, M., Sugiura, N., Hagiwara, H., and Hirose, S. (1989) Subtype switching of ANP receptors during in vitro culture of vascular cells. Am. J . Physiol., 256:H311-H314. Whitson, P.A., Huls, M.H., and Sams, C.F. (1989a) Culture time and glial cell-conditioned medium increase atrial natriuretic peptide (ANP) binding and ANP-induced cGMP levels in brain microvessel endothelial cells. J. Cell Biol., 109:331a. Whitson, P.A., Stuart, C.A., Huls, M.H., Sams, C.F., and Cintron, N.M. (1989b). Dexamethasone effects on creatine kinase activity and insulin-like growth factor receptors in cultured muscle cells. J. Cell. Physiol., 140%-17.