Non-Anticoagulant Heparin Increases Endothelial Nitric Oxide Synthase Activity: Role of Inhibitory Guanine Nucleotide Proteins

J Mol Cell Cardiol 30, 2669–2682 (1998) Article No. mc980831

Non-Anticoagulant Heparin Increases Endothelial Nitric Oxide Synthase Activity: Role of Inhibitory Guanine Nucleotide Proteins Peter C. Kouretas1, Robert L. Hannan1, Navin K. Kapur2, Richard Hendrickson2, Eileen M. Redmond2, Adam K. Myers1, Young D. Kim1, Paul A. Cahill1,2 and James V. Sitzmann2 Departments of Physiology and Biophysics1 and Surgery2, Georgetown University Medical Center, Washington DC, USA (Received 13 July 1998, accepted in revised form 14 September 1998) P. C. K, R. L. H, N. K. K, R. H, E. M. R, A. K. M, Y. D. K, P. A. C  J. V. S. Non-Anticoagulant Heparin Increases Endothelial Nitric Oxide Synthase Activity: Role of Inhibitory Guanine Nucleotide Proteins. Journal of Molecular and Cellular Cardiology (1998) 30, 2669–2682. Heparin, which is widely used clinically, has recently been shown to have specific properties affecting the vascular endothelium. We hypothesized that heparin stimulates endothelial nitric oxide synthase (eNOS) activity by a mechanism independent of its anticoagulant properties and dependent on an inhibitory guanine nucleotide regulatory protein (Gi). We determined the effect of both heparin and N-acetyl heparin (Non-Hep), a heparin derivative without anticoagulant properties, on eNOS activity in cultured bovine aortic endothelial cells and on endothelium-dependent relaxation in isolated vascular rings. The eNOS activity was determined by measuring both citrulline and nitric oxide (NO) metabolite formation. Heparin and Non-Hep dose-dependently increased basal eNOS activity (ED50 1.0 lg/ml or 0·15 U/ml), an effect that was significantly inhibited by pertussis toxin (100 ng/ml), a Gi-protein inhibitor. Agonist-stimulated (acetylcholine, 10 l) eNOS activity was potentiated following pre-treatment with both heparin and Non-Hep and reversed by pertussis toxin. Heparin and Non-Hep induced a dose-dependent relaxation in preconstricted thoracic aortic rings, an effect that was significantly inhibited by pertussis toxin, endothelial inactivation (following treatment with sodium deoxycholate) and NGnitro--arginine-methyl ester (L-NAME). We conclude that heparin and non-anticoagulant heparin induce endothelium-dependent relaxation following activation of eNOS by a mechanism involving a Gi-protein. Administration of heparin derivatives without anticoagulant properties may have therapeutic implications for the  1998 Academic Press preservation of eNOS in conditions characterized by endothelial dysfunction. K W: Heparin; Non-anticoagulant heparin; Nitric oxide; Bovine endothelial cells; Gi-proteins.

Introduction Heparin, which is widely used in the clinical setting based on its ability to bind to antithrombin III and achieve anticoagulation (Abildgard, 1968), has recently been shown to have specific properties affecting the vascular endothelium. Studies have

reported that heparin modulates the production and release of the endothelial-derived vasoconstrictor endothelin in both animal (Yokokawa et al., 1992) and human models (Piatti et al., 1996), an effect that is dependent on nitric oxide (NO) production (Yokokawa et al., 1993). Yokokawa et al. have shown that heparin increases

Please address all correspondence to: P. A. Cahill, Department of Surgery, Georgetown University Medical Center, Building D Room 206C, Washington, D.C. 20007, USA.

0022–2828/98/122669+14 $30.00/0

 1998 Academic Press

2670

P. C. Kouretas et al.

NO levels in humans on hemodialysis (Yokokawa et al., 1995) and in cultured endothelial cells from spontaneously hypertensive rats (Minami et al., 1995). In contrast, a recent study suggests that high doses of porcine heparin inhibit NO production by decreasing eNOS protein and mRNA expression (Upchurch et al., 1997). The mechanism whereby heparin modulates these vasoactive mediators, however, has not been determined. Heparin has significant vasoprotective effects during conditions such as hemorrhagic shock (Wang et al., 1990; Rana et al., 1992) and ischemia-reperfusion injury (Hobson et al., 1998; Wang et al., 1998). The endothelial dysfunction observed within the femoral artery following ischemia-reperfusion injury is associated with a decrease in NO production (Ma et al., 1993; Summers et al., 1993; Kurose et al., 1994) and is protected against by the administration of NO donors (Lefer et al., 1993; Gauthier et al., 1994). The vasculopathy associated with ischemia-reperfusion injury within the coronary artery is also linked to dysfunction of an inhibitory guanine nucleotide regulatory protein (Gi-protein) within the endothelium (Evora et al., 1994; Niroomand et al., 1995). Moreover, the important role of pertussis toxin-sensitive Gi proteins in mediating the activation of endothelial nitric oxide synthase (eNOS) by endothelial agonists has been previously demonstrated (Flavahan et al., 1989; Flavahan and Vanhoutte, 1990; Liao and Homcy, 1992). Heparin derivatives exhibit effects independent of the anticoagulant properties of heparin, including inhibition of vascular smooth muscle proliferation (Guyton et al., 1980; Hoover et al., 1980), inhibition of complement (Kazatchkine et al., 1981; Ekre et al., 1992), protection of cardiac muscle during ischemia-reperfusion injury (Friedrichs et al., 1994; Black et al., 1995) and protection of endothelial function during ischemia reperfusion injury (Sternbergh et al., 1993, 1995; Hannan et al., 1997). However, whether the effects of heparin on vascular endothelial function, in particular any effects on NO production, are independent of its anticoagulant properties is unresolved. The mechanism(s) by which heparin increases the production of NO metabolites is still unknown. Furthermore, the ability of non-anticoagulant heparin derivatives to mimic the effects of heparin on NO metabolism has not been investigated. Therefore, the aim of the present study was to investigate the effect of heparin and non-anticoagulant heparin derivatives on eNOS activity.

Materials and Methods Materials Heparin (bovine lung), N-acetyl heparin (porcine mucosal), sodium nitrite, potassium nitrate, sodium phosphate, phosphoric acid, acetylcholine (ACh), calcium chloride, NG-nitro--arginine-methyl ester (L-NAME), b-nicotinamide adenine dinucleotide phosphate, reduced form (NAPDH), ethylenediamine tetraacetic acid (EDTA), Dowex AG50WX8 (Na+ form), Tris HCl, HEPES, adenosine triphosphate (ATP), guanidine triphosphate (GTP), dithiothreitol (DTT), magnesium chloride, thymidine, sodium dodecyl sulfate (SDS) and bovine serum albumin (BSA) were purchased from Sigma Chemical (St. Louis, MO). The analytical column (4.6 mm×25 cm) was a strong-anion-exchange column containing Whatman Partisil-10 SAX10 lm particle size (Whatman Inc., Clifton, NJ). Pertussis toxin (Ptx) and was obtained from List Biological Laboratory (Campbell, CA). Antisera against endothelial nitric oxide synthase (eNOS), Gia3 Gia1–2 were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Nitrocellulose membranes (HYBOND-C) and the ECL detection system were obtained from Amersham (Arlington Heights, IL). [14C]-L-arginine (specific activity 320 lCi/mmol) and [32P]-NAD (specific activity 800 lCi/mmol) were purchased from New England Nuclear (Boston, MA). All other chemicals were of the highest purity commercially available.

Cell culture Bovine aortic endothelial cells (BAEC), repository No. AG07680B, were purchased from the N.I.A. Cell Culture Repository, Coriell Institute for Medical Research (Camden, NJ). Endothelial cells tested positive for the endothelial cell specific von Willebrand factor and angiotensin-1 converting enzyme activity and negative for a-smooth muscle actin. Endothelial cells were seeded into conventional plastic tissue culture flasks (Corning) and grown in RPMI1640 media supplemented with 10% heat-inactivated fetal bovine serum (FBS) plus 100 U/ ml penicillin, 100 lg/ml streptomycin, 0.01 lg/ml gentamicin and 0.25 lg/ml fungizone. Cells were grown to 75–80% confluence in a humidified atmosphere of 95% air, 5% CO2 at 37°C and were routinely subcultured after treatment for 5 min with 0.125% trypsin–EDTA at 37°C. Cells between passage 12–16 were routinely used for these studies.

Heparin stimulates eNOS activity in cultured endothelial cells

Treatment of cells Confluent cells were washed three times in RPMI1640 and placed in serum-depleted media containing RPMI-1640 medium supplemented with 0.2% heat-inactivated fetal bovine serum (FBS), plus 100 U/ml penicillin, 100 lg/ml streptomycin, 0.01 lg/ml gentamicin and 0.25 lg/ml fungizone for 24 h. Cells were subsequently treated with heparin or the non-anticoagulant heparin derivative N-acetyl heparin over a range of doses (0.01– 1,000 lg/ml or 0·001–140 U/ml) and membranes prepared for analysis of NO synthase (NOS) activity and expression. After treatment with heparin and N-acetyl heparin, the overlying media was removed and analyzed for nitrite/nitrate levels. Cells were also treated with the endothelial-dependent agonist acetylcholine (10−5) for 20 min after pretreatment with either heparin or N-acetyl heparin (1.0 lg/ml for 2 h) to assess agonist-stimulated NOS activity. In the final series of experiments, cells were pretreated for 24 h with the guanine inhibitory (Gi) nucleotide protein inhibitor pertussis toxin (100 ng/ ml) and basal and agonist-stimulated eNOS activity was measured. All drugs were prepared in RPMI1640 medium supplemented with 0.2% heat-inactivated FBS plus 100 U/ml penicillin, 100 lg/ml streptomycin, 0.01 lg/ml gentamicin and 0.25 lg/ ml fungizone. After treatment, cells were washed twice with Hanks Balanced Salt Solution (HBSS) prior to the preparation of membranes.

Preparation of membranes Cultured endothelial cells were washed twice with HBSS following exposure to drugs. Ice cold homogenization buffer containing 10 m Tris-HCl and 1 m EDTA (pH 7.5) was then added and the cells detached using a cell scraper. The cell suspension was then sonicated for 4×5 s and centrifuged at 30 000 ×g for 30 min at 4°C. The supernatant was discarded and the resulting pellet resuspended in 50 m Tris HCl (pH 7.4) containing 1 m EDTA and 0.01% bacitracin and stored at −70°C. Membrane protein was measured by the method of Bradford (1976) with bovine serum albumin as a standard, and subsequently used in NOS activity assays, Western blot analysis and ADP ribosylation.

Endothelial nitric oxide synthase (eNOS) activity Endothelial nitric oxide synthase (eNOS) activity was measured by determining the conversion of

2671

L-[14C]arginine to L-[14C]citrulline based on the method of Bredt and Snyder (1990) and modified by Cahill et al. (1995). The 30 000×g particulate fraction (25 lg) was incubated in a total volume of 135 ll in 50 m Tris-HCl containing 0.1 m EDTA 1 m NADPH, 1 m CaCl2, 1 lCi/ml of L[14C]arginine (approximately 100 000 cpm). The reaction was initiated by the addition of 25 ll of protein extract and carried out at 25°C for 1 h. The reaction was terminated by the addition of 2 ml of ice cold stopping buffer 30 m HEPES and 3 m EDTA pH 5.5. The reaction mixture was then passed over a chromatography cation exchange column (120 micron polystyrene filter) containing 500 ll DOWEX AG50WX-8 (Na+ form). The L-[14C]citrulline was eluted with 2×0.5 ml of distilled water. The eluent was collected, 4 ml of scintillation fluid added and counted by liquid scintillation spectrometry.

Western blot analysis Membrane proteins (20–50 lg/lane) were separated on 10% SDS-polyacrylamide gel, as described previously (Cahill et al., 1994). After SDS-PAGE, the separated proteins were electrophoretically transferred to nitrocellulose membranes (HYBONDC, Amersham) using a Transphor electroblotter unit (Hoefer Scientific Instruments, San Francisco, CA) at 100 V for 2 h. After transfer, the membranes were incubated for 90 min in blocking solution containing 24 m Tris base (pH 7.6), 0.05% (v/v) Tween-20 and 15 m NaCl (TTBS) supplemented with 5% nonfat dry milk. The membranes were washed once for 5 min with TTBS. The membranes were incubated with the specific anti-sera (1:1000 dilution) against both eNOS and Gia1–2, Gia3 in TTBS for 60 min at room temperature with gentle rocking. After washing the blots three times for 10 min each in TTBS, they were incubated with secondary antibody solution (horseradish peroxidase-conjugated IgG diluted in TTBS (1:5000) for 1 h at room temperature with gentle rocking. The blots were washed three times for 10 min each with TTBS and then incubated with a mixture of equal volume of ECL detection solution A and B (Amersham, Arlington Heights, IL) for 1 min at room temperature. The blots were then covered in plastic wrap and placed in a film cassette to expose to Hyperfilm (Amersham, Arlington Heights, IL) film for 5–15 s. The signal intensity (integral volume) of the appropriate bands on the autoradiogram was analyzed using the Imagequant software package (Biosoft, Indianapolis, IN).

2672

P. C. Kouretas et al.

Pertussis-toxin catalyzed ADP-ribosylation G-protein substrates of pertussin toxin were assayed using pertussis toxin-catalyzed incorporation of [32P]ADP-ribose from [32P]NAD, as previously described (Cahill et al., 1994). Membrane (20–30 lg) were resuspended in 100 ll of 100 mm Tris-HCl (pH 8.0), containing 5 m DTT, 10 m thymidine, 6 m MgCl2 2 m GTP, 2.5 m ATP and 10 l [32P]NAD. After addition of activated pertussis toxin (20 lg/ml), the membrane preparations were incubated for 90 min at 30°C. The ADP-ribosylation reactions were stopped by centrifugation at 15 000×g for 3 min and the pellet was resuspended in SDS-sample buffer containing 62.5 m Tris-HCl (pH 6.8), 2% SDS, 10% (v/v) glycerol and 5% (v/ v) b-mercaptoethanol. The samples were boiled for 10 min and the proteins separated on a 10% SDSpolyacrylamide gel. Gels were dried on cellophane and exposed to Hyperfilm-MP-5 film (Amersham, Arlington Heights, IL) with an intensifying screen at −70°C for 1–3 days. The signal intensity (integral volume) of the appropriate bands on the autoradiogram was analyzed using a Personal Densitometer and the Imagequant software package (Biosoft, Indianapolis, IN).

of functional endothelium was determined by addition of acetylcholine (10 l). Drugs were than washed out and the tissue allowed to equilibrate for a further 30 min. The aortic rings were first precontracted with a submaximal dose of KCl (30 m) before cumulative dose response curves for heparin and Non-Hep (1000–16000 lg/ml) were determined. The viability of each ring was again assessed at the end of each experiment with 30 m KCl.

Protocol II In parallel experiments, cumulative dose-response curves were determined for heparin and Non-Hep in preconstricted vessels before they were washed extensively and L-NAME (final concentration 100 l) added for 20 min or pertussis toxin (2000 ng/ml) added for 4 h prior to the addition of KCl (30 m). The acetylcholine (10 l) response was significantly blocked following both these treatments. Dose-response curves for heparin and NonHep were then repeated. The tissue viability was again checked with KCl (30 m) at the end of the experiment.

Isolated vascular ring studies

Protocol III The investigation conforms with the Guide for the Care and Use of Laboratory Animals. The thoracic aorta of male Sprague Dawley rats (200–250 g) was cut into 5 mm rings and mounted for isometric recording in 20-ml jacketed organ chamber containing modified Krebs solution (0.12  NaCl, 4.7 m KCl, 2.4 m MgSO4, 1.2 m KH2PO4, 11 m glucose, 25 m NaHCO)3, 1.25 n CaCl2 and bubbled with 95% O2/5% CO2 at 37°C). After equilibration, the aortic rings were placed at the optimal point of their length–tension curve, usually at 2 g resting tension. The rings were then equilibrated for 90 min. Responses were measured with a force-displacement transducer (model FT 03; Grass Instrument Co, Quincy, MA) and recorded with a multichannel paper recorder (model 7 and 7E, Grass Instrument Co).

Protocol I Eight aortic rings were examined in parallel. After equilibration, for every aortic ring examined, KCl was added into the tissue bath (final concentration 30 m) to check the tissue viability. The presence

Aorta were randomly divided into two groups. One group was treated with 0.3% sodium deoxycholate for 1 min to inactivate the endothelium (CusmaPelogia et al., 1993). Endothelial inactivation was confirmed by a lack of response to acetylcholine (10 l) in precontracted vessels. The parallel group retained an intact functional endothelium, verified using acetylcholine. The aortic rings were first precontracted with KCl (30 m) before cumulative dose response curves for heparin and Non-Hep were determined. The tissue viability was again checked with KCl (60 m) at the end of the experiment.

Statistics Statistical analyses were performed using analysis of variance (ANOVA) followed by the Dunnets test for multiple comparisons and a two-tailed unpaired Student’s t-test for comparison of two groups. A probability value of P<0.05 was considered significant.

Heparin stimulates eNOS activity in cultured endothelial cells

Figure 1 Time course for bovine heparin (1.0 lg/ml) stimulation of endothelial nitric oxide activity (eNOS). Data are expressed as [14C]citrulline (pmol/mg/h) and are the mean ±..., ∗PΖ0.05 v control, n=3.

Results Heparin and non-anticoagulant heparin stimulates NOS activity in cultured endothelial cells Similar to previous studies (Yokokawa et al., 1993, 1995; Li et al., 1996), we confirmed increased NOx levels in endothelial cells treated with heparin and N-acetyl heparin for 2 h (data not shown). In order to determine whether the increase in NOx levels was secondary to a direct stimulation of eNOS, we measured the ability of both heparin and N-acetyl heparin (Non-Hep) to stimulate eNOS activity using the arginine to citrulline conversion assay. Heparin at a dose of 1.0 lg/ml (0.15 U/ml) caused a timedependent increase in basal eNOS activity, with significant increases observed after 1 min, reaching a peak following 2 h exposure (Fig. 1). Non-Hep caused a similar time-dependent increase in eNOS activity (data not shown). Based on these experiments, the time point of 2 h was chosen for the remainder of experiments for both heparin and Non-Hep. Heparin and Non-Hep (2 h treatment) caused a dose-dependent increase in basal eNOS activity with an ED50 of approximately 1.0 lg/ml which was chosen for the remainder of the experiments (Fig. 2). Dextran sulfate (1 lg/ml), a randomly sulfated polyanion with similar size and charge characteristics to heparin, also stimulated eNOS activity following a 2 h treatment. However, higher concentrations of dextran sulfate (>100 lg/ ml) significantly inhibited eNOS activity in these cells (data not shown). The effect of heparin and Non-Hep on agoniststimulated NOS was also determined. As outlined

2673

in Figure 3(a) both acetylcholine (ACh) and heparin caused a significant increase in basal eNOS activity. In addition a 2 h pre-treatment of endothelial cells with heparin caused a significant potentiation of ACh-stimulated eNOS activity when compared to ACh alone [Fig 3(a)]. Similarly, Non-Hep treatment increased basal eNOS activity and potentiated AChstimulated eNOS activity in these cells [Fig. 3(b)]. To determine whether the increase in eNOS activity following heparin exposure was secondary to an increase in the expression of eNOS protein levels, Western blot analyses were performed using specific antisera against eNOS. Western blots prepared from particulate fractions of endothelial cells exposed to both heparin and N-acetyl heparin (1.0 lg/ml) for 2 h and 24 h revealed a major protein band with a molecular weight of 150 kDa corresponding to eNOS [Fig. 4(a) and (b)]. There was no significant increase in eNOS protein expression in endothelial cells treated for 2 or 24 h with heparin or N-acetyl heparin.

Inactivation of Gia proteins inhibits heparin- and nonanticoagulant heparin-mediated increases in eNOS activity Since several endothelial agonists stimulate eNOS activity through activation of an inhibitory Giprotein on endothelial cells (Flavahan et al., 1989; Flavahan and Vanhoutte, 1990; Liao and Homcy, 1992), we determined the effect of inactivating Gia proteins with pertussis toxin (Ptx) on heparin and Non-Hep stimulated eNOS activity. Endothelial cells were pre-treated with Ptx (100 ng/ml) 24 h prior to heparin treatment before eNOS activity was determined. As outlined in Figure 5(a) inactivation of the Gia proteins with Ptx resulted in a significant decrease in heparin-stimulated eNOS activity. In addition, ACh-stimulated eNOS activity, and potentiation of the ACh response following pre-treatment with heparin was also significantly inhibited by Ptx treatment [Fig. 5(a)]. Similarly, Ptx treatment significantly inhibited Non-Hep stimulated eNOS activity and potentiation of the ACh-response in these cells [Fig. 5(b)].

Heparin and non-anticoagulant heparin increases endothelial Gia protein activity To determine whether heparin and Non-Hep induced increases in eNOS activity correlated with changes in G-protein functionality, we measured Ptx catalyzed NAD-dependent ADP ribosylation of

2674

P. C. Kouretas et al.

Figure 2 Dose response relationships for (a) bovine heparin and (b) N-acetyl heparin stimulation of eNOS activity. Cells were treated in the absence or presence of heparin or N-acetyl heparin for 2 h before eNOS activity was determined. Data are expressed as [14C]citrulline (pmol/mg/h) and are the mean ±..., ∗P±0.05 v control, n=4.

Figure 3 The effect of (a) heparin and (b) N-acetyl heparin on ACh stimulated eNOS activity. Cells were treated in the with or without heparin (1 lg/ml or 0.15 U/ml) or N-acetyl heparin (1.0 lg/ml) for 2 h in the absence or presence of ACh (10−5 ) for 20 min before NOS activity was determined. Data are expressed as [14C]citrulline (pmol/mg/h), eNOS activity was completely inhibited by treatment with L-NAME (100 l) and EDTA (5 m). Results are mean ±..., ∗P±0.05 v control, n=5.

Gia substrates in endothelial cells. Pertussis toxin catalyzed the incorporation of 32P ADP-ribose into one major peptide band (approx. 40 kDa) in cultured endothelial cells, a protein that co-migrated with Gia proteins immunodetected using specific antibodies (data not shown). The ribosolyation was linear over a range of 10–100 lg of membrane

protein. Pertussis toxin-catalyzed ribosylation of Gia substrates was significantly increased in heparin and Non-Hep treated cells, compared to control [Fig. 6(a)]. Given that Gia3 proteins have been shown to be linked to agonist- and shear stress-stimulated eNOS activity (Flavahan et al., 1989; Flavahan and Vanhoutte, 1990; Liao and Homcy, 1992; Ohno et

Heparin stimulates eNOS activity in cultured endothelial cells

2675

Figure 4 The effect of heparin and N-acetyl heparin on eNOS protein expression after (a) 2 h and (b) 24 h. Cells were treated in the absence or presence of heparin (1 lg/ml) or N-acetyl heparin (1 lg/ml) for 2 or 24 h before eNOS protein expression was determined by Western blot. Western blots reveal a protein band with a molecular weight of >150 kDa corresponding to eNOS. Densitometric analysis reveals no difference in eNOS expression following exposure to heparin or N-acetyl heparin. Results are mean ±..., n=3.

Figure 5 The effect of pertussis toxin (Ptx) on (a) heparin and (b) N-acetyl heparin-induced stimulation of eNOS activity and potentiation of ACh-stimulated eNOS activity. Data are expressed as [14C]citrulline (pmol/mg/h). Endothelial cells were pre-treated with Ptx for 24 h and then exposed to either ACh (10−5 ) for 20 min alone or ACh after pretreatment with heparin (1.0 lg/ml) or N-acetyl heparin (1.0 lg/ml) for 2 h. NOS activity was inhibited by treatment with L-NAME (100 l) and EDTA (5 m). Results are mean ±..., ∗PΖ0.05 v control in the absence of Ptx, + PΖ0.05 v within each group in the presence of Ptx, ‡PΖ0.05 compared to ACh alone, n=3.

al., 1993), we determined whether a heparin or Non-Hep-mediated increase in Gia protein activity was secondary to an increase in Gia3 protein expression. Western blot analysis of the particulate

fraction of endothelial cells exposed to heparin and Non-hep for 2 h revealed that there was no significant change in Gia3 protein expression compared to control cells [Fig. 6(b)]. There was also no sig-

2676

P. C. Kouretas et al.

Figure 6 The effect of heparin or N-acetyl heparin on (a) Ptx-catalyzed ADP ribosylation of Gia substrates and (b) Gia3 protein expression. Endothelial cells were treated in the absence or presence of heparin (1 lg/ml) or N-acetyl heparin (1 lg/ml) for 2 h before Ptx-catalyzed ADP ribosylation of Gia substrates and Gia3 protein expression was determined. (a) Representative autoradiograph and cumulative densitometric analysis of Ptx-catalyzed ADP ribosylation Gia substrates by both heparin (1.0 lg/ml) and N-acetyl heparin (1.0 lg/ml). (b) Representative Western blot and cumulative denistometric of Gia3 protein expression. Results are mean ±..., −PΖ0.05 v control, n=3.

nificant change in Gia1–2 expression in cells exposed to heparin and Non-Hep [Fig. 6(b)].

Heparin and non-anticoagulant heparin-induced endothelial-dependent relaxation in preconstricted vessels: role of NO and Gia proteins Following pre-contraction of aortic rings with KCl (30 m), treatment with heparin or Non-Hep induced a dose-dependent relaxation in these rings [Fig. 7(a) and (b)]. The role of endothelial NO in mediating the heparin-induced relaxation was examined following treatment of the aortic rings with a specific NOS inhibitor, L-NAME. Following 100 l L-NAME treatment for 20 min prior to the addition of KCl, the heparin and Non-Hep-induced relaxation was significantly inhibited [Fig 7(a) (b) and (c)]. However, endothelial-independent relaxation using SNAP (100 l) remained intact. The role of the endothelium in mediating the heparin and Non-Hep-induced relaxation was examined following inactivation of the endothelium with sodium deoxycholate (Cusma-Pelogia et al., 1993). In the presence of an intact endothelium

(defined as responsive to acetylcholine, 10 l), heparin caused a significant relaxation in preconstricted vessels [Fig. 8(a) and (b)]. In contrast, in endothelial deactivated vessels, the acetylcholine response was inhibited concomitant with a significant inhibition of the heparin-induced relaxation response [Fig. 8(a) and (b)]. However the endothelial-independent relaxation using the NO-donor, SNAP, was unaffected. Similar results were observed following endothelial denudation (data not shown). Endothelial inactivation with sodium deoxycholate also significantly inhibited the Non-Hep-induced relaxation (data not shown). The role of Gia proteins in mediating the endothelial dependent heparin and Non-Hep-induced relaxation was also examined following inactivation of Gia proteins with pertussis toxin. Following treatment with pertussin-toxin (200 ng/min) for 4 h, the heparin and Non-Hepinduced relaxation of vessels pre-contracted with KCl was significantly attenuated [Fig. 9(a) and (b)].

Discussion The current studies demonstrate for the first time that heparin stimulates eNOS activity in cultured

Heparin stimulates eNOS activity in cultured endothelial cells

2677

Figure 7 (a) and (b). Representative tracings demonstrate the effect of L-NAME (100 l) on heparin- and N-acetyl heparin (Non-Hep)-induced relaxation of precontracted rat thoracic aortic rings. (c) Cumulative heparin- and NonHep-induced relaxation of pre-constricted aortic rings following treatment with L-NAME. Data are expressed as percent KCl (30 m) contraction and represent the mean ±..., ∗P<0.05 v heparin alone or Non-Hep alone, n=3. SNAP, S-nitroso-n-acetylpenicillamine (100 l).

endothelial cells and isolated vascular rings independent of its anticoagulant properties, via a mechanism involving a pertussis toxin-sensitive inhibitory G-protein. Moreover, heparin and N-acetyl heparin potentiated agonist-stimulated eNOS activity in these cells. Heparin, which is widely used clinically based on its ability to bind to antithrombin III and achieve anticoagulation (Lindahl et al., 1979; Rosenberg and Lam, 1979), has also been recently shown to have specific properties that modulate the function of the vascular endothelium. Several studies have demonstrated that heparin modulates the production and release of specific endothelial-derived vasoactive mediators including endothelin (Imai et al., 1993; Reantrogoon et al., 1994; Piatti et al., 1996) and NO (Yokokawa et al., 1993, 1995; Li et al., 1996). Whether the increased nitrite/nitrate levels observed in previous studies were indicative of a direct effect of heparin on eNOS activity or sec-

ondary to activation of eNOS by some other substance modulated by heparin was unclear. Similar to previous studies using heparin (Minami et al., 1995; Li et al., 1996), we demonstrated a significant increase in nitrate and nitrite levels (the stable metabolites of NO) in bovine aortic endothelial cells. In addition we have shown that a heparin-derivative devoid of anticoagulant properties mimics this effect (data not shown). We further demonstrated that heparin and N-acetyl heparin dose-dependently increased eNOS activity in these cells without affecting eNOS protein levels. Activation of eNOS by heparin and its derivatives was functionally relevant since in preconstricted thoracic aortic rings, heparin and N-acetyl heparin induced a dose-dependent relaxation in pre-constricted vessels, an effect that was significantly inhibited by endothelial inactivation and L-NAME. Despite the fact that heparin and N-acetyl heparininduced relaxation of preconstricted aortic rings

2678

P. C. Kouretas et al.

Figure 8 (a) Representative tracing demonstrating the effect of endothelial inactivation on heparin-induced relaxation of precontracted rat thoracic aortic rings. The endothelium was inactivated with 0.3% sodium deoxycholate treatment for 1 min prior to the addition of KCl. Endothelial inactivation was confirmed by loss of the ACh response. (b) Cumulative heparin-induced relaxation of pre-constricted aortic rings following endothelial inactivation. Data are expressed as percent KCl (30 m) contraction and represent the mean ±..., ∗P<0.05 v heparin in the presence of endothelium, n=3. SNAP, S-nitroso-n-acetylpenicillamine (100 l), an NO donor.

was immediate, their stimulation of eNOS activity was maximal after 2 h. However, a significant increase in eNOS activity was observed within the first 1 min. These data are consistent with heparininduced stimulation of eNOS activity and subsequent NO production independent of its anticoagulant properties. Moreover, activation of eNOS by heparin and its derivatives leads to, in part, relaxation of preconstricted aortic rings. A previous study highlighting an effect of heparin that was independent of its anticoagulant properties has recently been reported. Black et al. (1995) demonstrated a cardioprotective effect of heparin and N-acetyl heparin in an in vivo model of myocardial ischaemia and reperfusion injury and concluded that the mechanism of cytoprotection was unrelated to alterations in the coagulation cascade. Similarly, heparin derivatives exhibit effects independent of heparin’s anticoagulant properties, including inhibition of vascular smooth muscle proliferation (Guyton et al., 1980; Hoover et al., 1980), inhibition of complement (Cofrancesco et al., 1979; Friedrichs et al., 1994) and protection of endothelial

function during ischemia reperfusion injury (Sternberg et al., 1995). Whether stimulation of NO production by heparin derivatives plays a role in the cytoprotection effects of these substances is unclear. Nevertheless it is noteworthy that the endothelial dysfunction characteristic of such conditions is associated with a decrease in NO production (Ma et al., 1993; Summers et al., 1993) and protected against by the administration of NO donors (Lefer et al., 1993; Siegfried et al., 1992). The mechanism of heparin-mediated activation of Gi-proteins is unclear at this point. One possible explanation for the diverse biological properties exhibited by heparin and heparinoids is the propensity of these polyanions to complex with plasma proteins and enzymes and modify their inherent activity (Jaques, 1967; Elbein, 1974). To date, several studies have shown that heparin, like many other polyanionic compounds, modulates the coupling of adrenergic and muscarinic receptors to Giproteins that regulate adenylyl cyclase and K+ channels in several cell types (Willuweit and Aktories, 1988; Ito et al., 1990; Dasso et al., 1991).

Heparin stimulates eNOS activity in cultured endothelial cells

2679

Figure 9 (a) Representative tracings demonstrating the effect of pertussis toxin (200 ng/ml) on heparin and Non-Hepinduced relaxation of precontracted thoracic aortic rings. (b) Cumulative heparin- and Non-Hep-induced relaxation of pre-constricted rings following Gi-protein inactivation with pertussis toxin. Data are expressed as percent relaxation and represent the mean ±..., ∗P<0.05 v pre-Ptx, n=3. SNAP, S-nitroso-n-acetylpenicillamine (100 l).

This effect of heparin was blocked by pertussis toxin (Wang et al., 1996). Indeed, several previous studies have shown that heparin uncouples agonist-receptors from inhibitory Gi-proteins through a direct interaction with Gi-proteins in broken cell preparations (Willuweit and Aktories, 1988; Ito et al., 1990; Dasso et al., 1991; Wang et al., 1996). In the present study, using intact cells, we demonstrate that heparin interacts directly with Gia proteins inasmuch as inactivation of these proteins with pertussis toxin resulted in attenuation of heparin and Non-Hep-mediated relaxation of preconstricted aortic rings and stimulation of eNOS activity in cultured endothelial cells. Furthermore, pertussin toxin-catalyzed ADP-dependent ribosylation of Gia substrates was significantly enhanced in endothelial cells treated with heparinoids. Both heparin and its derivative significantly increased Gia protein labeling independent of a change in the expression of Gia protein levels in these cells. In contrast to the current study, a recent study reported that high doses of porcine heparin decreased NO production and expression in cultured

endothelial cells (Upchurch et al., 1997). Despite several important differences between these studies, including the concentrations of heparin used, the measurement of NO metabolites v eNOS activity, the source of heparin (bovine v porcine) and the time course for NO activation, a decrease in NO production after heparin treatment may play some role in mediating unique thrombotic events in cardiovascular settings such as coronary angioplasty. In contrast, in the current study, lower doses of heparin (0.01–0.15 U/ml) stimulated eNOS activity and subsequent NO production, an effect that may contribute, in part, to the well-documented beneficial effects of heparin in reducing both thrombotic complications of angioplasty and in reducing graft occlusion following surgical revascularization (Guyton et al., 1980; Hoover et al., 1980). Of note, a randomly sulfated polyanion with similar size and change characteristics to heparin mimicked the effects of heparin treatment at low doses suggesting that heparin’s action is secondary to its negative charge rather than to a specific polysaccharide. However, higher concentrations of

2680

P. C. Kouretas et al.

dextran sulfate (>100 lg/ml), similar to previous studies, significantly inhibited eNOS activity in these cells (Upchurch et al., 1997). Heparin’s interaction with various proteins may also explain the discrepancy in the ED50 of heparin’s stimulation of eNOS activity in cell culture studies when compared to the dose required to elicit vasorelaxation in the isolated vascular ring preparations. Heparin, and presumably other heparinoids such as N-acetyl heparin, avidly bind to and regulate the metabolism of several growth factors such as basic fibroblast growth factor (Folkman et al., 1988; Bashkin et al., 1989), acidic fibroblast growth factor (Barzu et al., 1989) and endothelial growth factor (Rosengart et al., 1989). These growth factors are associated with the basement membrane and subendothelial extracellular matrix. Therefore, it is likely that the abundance of these heparin binding sites in isolated vascular rings sequester both heparin and N-acetyl heparin to a greater extent than binding sites in a homogenous endothelial culture. An alternative explanation for the discrepancy may be that the aortic rings are insensitive to the vasorelaxant effects of heparin because of aortic banding due to the high resting force of 2 g. However, several previous studies by us and others have used this same resting tension when working with rat aortic rings (Hou et al., 1997). It was found that the maximum contraction of the aorta to KCl was unaffected by changing the resting tension over a range of 1.5–5 g which would rule out aortic banding as the reason for this discrepancy. Furthermore, the dose of heparin required to activate eNOS in vitro may not be the same as that required to induce vasodilation in vitro inasmuch as the regulation of vascular reactivity in vitro is dependent on several factors including endothelial derived relaxing and contracting substances any one of which heparin or N-acetyl heparin may modulate. Indeed the concentrations required in the isolated vascular ring studies (150–2400 U) are consistent with the active doses of heparin used in humans (700–1400 U) and in an in vivo canine model of ischemia-reperfusion (Black et al., 1995; Hannan et al., 1997). Finally, since heparin has previously been shown to bind to glass as well as other synthetic surfaces (Hersh et al., 1969; Tunbridge et al., 1981), the possibility that both heparin and N-acetyl heparin bind to the glass of the organ bath chamber cannot be ruled out. In conclusion, these data suggest that heparin, independent of its anticoagulant properties, activates eNOS activity in endothelial cells and isolated vascular rings via a mechanism involving

a pertussis toxin sensitive inhibitory G-protein. It is tempting to speculate that heparin’s significant vasoprotective effects during conditions such as hemorrhagic shock and ischemia-reperfusion injury are due, in part, to its effects on basal and agoniststimulated eNOS activity through activation of Giproteins.

Acknowledgments This work was supported by grants from the American Heart Association and the National Institute of Health DK47067 and HL08978.

References A U, 1968. Highly purified antithrombin 3 with heparin cofactor activity prepared by disc electrophoresis. Scand J Clin Lab Invest 21: 89–91. B T, L JC, P M, M S, C J, 1989. Heparin-derived oligosaccharides: affinity for acidic fibroblast growth factor and effect on its growthpromoting activity for human endothelial cells. J Cell Physiol 140: 538–548. B P, D S, K M, S CM, F J, V I, 1989. Basic fibroblast growth factor binds to subendothelial extracellular matrix and is released by heparitinase and heparin-like molecules. Biochemistry 28: 1737–1743. B SC, G MR, F GS, K KS, D EM, L BR, 1995. Cardioprotective effects of heparin or N-acetylheparin in an in vivo model of myocardial ischaemic and reperfusion injury. Cardiovasc Res 29: 629–636. B MM, 1976. A rapid and sensitive method for quantification of microgram quantities of protein utilizing the principle of protein dye binding. Ann Biochem 72: 248–254. B DS, S SH. 1990. Isolation of nitric oxide synthase: a calcium-requiring enzyme Proc Natl Acad Sci USA 87: 682–685. C PA, R EM, F C, S JV. 1995. Nitric oxide regulates angiotensin II receptors in vascular smooth muscle cells. Eur J Pharmacol 288: 219– 229. C PA, W Y, S JV. 1994. Altered adenylyl cyclase activities and G-protein abnormalities in portal hypertensive rabbits. J Clin Invest 93: 2691–2700.44. C E, R F, P E. 1979. Correlation of sulfate content and degree of carboxylation of heparin and related glycosaminoglycans with anticomplement activity. Relationships to the anticoagulant and platelet-aggregating activities. Thromb Res 14: 179–187. C-P N, O SF, N D,  C MH, S R, F ZB. 1993. Endothelium inactivation in in vitro perfused vascular beds. Comparison of methods. J Pharmacol Toxicol Methods 29: 157–163.

Heparin stimulates eNOS activity in cultured endothelial cells D LL, T CW. 1991. Heparin and other polyanions uncouple alpha 1-adrenoceptors from G-proteins. Biochem J 280: 791–795. E HP, N Y, L O, H P, H O, N T, V I, C I. 1992. Antiinflammatory effects of heparin and its derivatives: inhibition of complement and lymphocyte migration. Adv Exp Med Biol 313: 329–340. E AD, 1974. Interactions of polynucleotides and other polyelectrolytes with enzymes and other proteins. Advan Enzymol 40: 29–64. E PR, P PJ, S HV. 1994. Impaired endothelium-dependent relaxation after coronary reperfusion injury: evidence for G-protein dysfunction. Ann Thoracic Surg 57: 1550–1556. F NA, S H, V PM, 1989. Pertussis toxin inhibits endothelium-dependent relaxations to certain agonists in porcine coronary arteries. J Physiol (Lond) 408: 549–560. F NA, V PM, 1990. G-proteins and endothelial responses. Blood Vessels 27: 218–229. F J, K M, S J, W M, I D, V I, 1988. A heparin-binding angiogenic protein—basic fibroblast growth factor—is stored within basement membrane. Am J Pathol 130: 393– 400. F GS, K KS, M PJ, G MR, L BR. 1994. Effects of heparin and N-acetyl heparin on ischemia/reperfusion-induced alterations in myocardial function in the rabbit isolated heart. Circ Res 75: 701–710. G TW, D KL, L AM. 1994. Nitric oxide attenuates leukocyte-endothelial interaction via P-selectin in splanchnic ischemia-reperfusion. Am J Physiol 267: G562–G568. G JR, R RD, C AW, K MJ. 1980. Inhibition of rat arterial smooth muscle cell proliferation by heparin. In vivo studies with anticoagulant and nonanticoagulant heparin. Circ Res 46: 625–634. H RL, K PC, K YD, M JL, T LN, W YN, W RB. 1997. Heparin preserves coronary endothelial function after ischemiareperfusion injury independent of its anticoagulant properties. FASEB J 11: A283. H LS, W HH, B IW, J. 1969. Heparin ionically linked to glass. J Biomed Mater Res 3: 471– 474. H RW, W JG, F D, K JC. 1998. Heparinization reduces endothelial permeability and hydrogen ion accumulation in a canine skeletal muscle in ischemia-reperfusion model. J Vasc Surg 7: 585–591. H RL, R W, H W, K MJ. 1980. Inhibition of rat smooth muscle proliferation by heparin. Circ Res 47: 578–583. H M-C, C PA, Z S, R EM, S JV. 1997. Enhanced G-protein-induced relaxation in portal hypertensive rats: role of nitric oxide. Hepatology 26: 27–33. I T, H Y, E T, M F. 1993. Heparin has an inhibitory effect on endothelin-1 synthesis and release by endothelial cells. 21: 353–358. I H, T R, I M, H E, S T, K Y, 1990. Heparin uncouples the muscarinic receptors from GK protein in the atrial cell membrane of the guinea-pig heart. Pflu¨gers Arch 417: 126–128.

2681

J LB. 1967. The pharmacology of heparin and heparinoids. Progr Med Chem 5: 139–198. K MD, F DT, M DD, R RD, A KF, 1981. Structural determinants of the capacity of heparin to inhibit the formation of the human amplification C3 convertase. J Clin Invest 67: 223–228. K I, W R, G MB, G DN. 1994. Modulation of ischemia/reperfusion-induced microvascular dysfunction by nitric oxide. Circ Res 74: 376– 382. L AM, S MR, M XL. 1993. Protection of ischemia-reperfusion injury by sydnonimine NO donors via inhibition of neutrophil-endothelium interaction. J Cardiovasc Pharm 22 Suppl 7: S27–S33. L JM, H H, L R CA, R MJ, V S TJ, C BS, 1996. Heparin and protamine stimulate the production of nitric oxide. J Cardiovasc Surg 37: 445–452. L JK, H CJ. 1992. Specific receptor-guanine nucleotide binding protein interaction mediates the release of endothelium derived relaxing factor. Circ Res 70: 1018–1026. L U, B G, H M, T L, F LA, L A, 1979. Structure of the antithrombinbinding site in heparin. Proc Natl Acad Sci USA 76: 3198–3202. M XL, W AS, L DJ, L AM. 1993. Diminished basal nitric oxide release after myocardial ischemia and reperfusion promotes neutrophil adherence to coronary endothelium. Circ Res 72: 403–412. M M, Y K, K M, I M, H T, K H, H T, Y K, T T. 1995. Promotion of nitric oxide formation by heparin in cultured aortic endothelial cells from spontaneously hypertensive rats. Clin Exp Pharmacol Physiol 22 Suppl 1: S146–S147. N F, W C, W A, B M, S C, M R, B T, S RH, K W, R B, 1995. Impaired function of inhibitory G proteins during acute myocardial ischemia of canine hearts and its reversal during reperfusion and a second period of ischemia. Possible implications for the protective mechanism of ischemic preconditioning. Circ Res 76: 861–870. O M, G GH, D VJ, C JD. 1993. Shear stress elevated endothelial cGMP. Role of a potassium channel and G-protein coupling. Circulation 88: 193– 197. P PM, M LD, V G, C M, N R, G B, F E, P CV, P AE, P G. 1996. Effects of low-dose heparin infusion on arterial endothelin-1 release in humans. Circulation 94: 2703–2707. R MW, S G, W P, A A, Z M, C IH. 1992. Protective effects of preheparinization on the microvasculature during and after hemorrhagic shock. J Trauma 32: 420–426. R S, A LM, S MM, D HS, R L, 1994. Specific heparin fractions suppress endothelin-1 production in cultured human umbilical vein endothelial cells. Arch Biochem Biophys 314: 315– 322. R RD, L L. 1979. Correlation between structure and function of heparin. Proc Natl Acad Sci USA 76: 1218–1222.

2682

P. C. Kouretas et al.

R TK, K JP, M T, C RE. 1989. Pharmacokinetics and distribution of heparinbinding growth factor I (endothelial cell growth factor) in the rat. Circ Res 64: 227–234. S MR, E J, R T, M XL, L AM. 1992. Cardioprotection and attenuation of endothelial dysfunction by organic nitric oxide donors in myocardial ischemia-reperfusion. J Pharm Exper Ther 260: 668–675. S WC, III, M RG, A B, 1993. Heparin prevents postischemic endothelial cell dysfunction by a mechanism independent of its anticoagulant activity. J Vasc Surg 17: 318–327. S WC, III, S M, M RG. 1995. Heparinoids with low anticoagulant potency attenuate postischemic endothelial cell dysfunction. J Vasc Surg 21: 477–483. S ST, Z MJ, F JA. 1993. Production of endothelium-derived relaxing factor (EDRF) is compromised after ischemia and reperfusion. Am J Surg 166: 216–220. T LJ, L JV, P RK, W AL, M T. 1981. Stability of diluted heparin sodium stored in plastic syringes. Am J Hosp Pharm 38: 1001–1004. U GR, W GN, F JE, F AJ, P A, S AM, A CS, K JF, L J, 1997. High dose of heparin decreases nitric oxide production by cultured bovine endothelial cells. Circulation 95: 2115–2121.

W P, S G, R MW, B ZF, C IH, 1990. Preheparinization improves organ function after hemorrhage and resuscitation. Am J Physiol 259: R645– R650. W SZ, E R, Z SZ, -F EE. 1996. Selective enhancement of antagonist ligand binding at muscarinic M2 receptors by heparin due to receptor uncoupling. Eur J Pharmacol 296: 113–118. W B, A K. 1998. Heparin uncouples alpha 2-adrenoceptors from the Gi-protein membranes of human platelets. Biochem J 249: 857–863. W JG, K JC, V CR, H RW, 1998. Heparin decreases ischemia-reperfusion injury in isolated canine gracilis model. Arch Surg 123: 470–472. Y K, M AK, K M, H T, M K, Y K, T T, 1992. Heparin suppresses endothelin-1 action and production in spontaneously hypertensive rats. Am J Physiol 263: R1035–R1041. Y K, T H, K M, M AK, Y M, T T, 1993. Heparin regulates endothelin production through endothelium-derived nitric oxide in human endothelial cells. J Clin Invest 92: 2080–2085. Y K, M R, S MG, K M, Y K, M M, K H, H T, T T, M AK. 1995. Increased nitric oxide production in patients with hypotension during hemodialysis [see comments]. Ann Intern Med 123: 35–37.