Rudolph, Alan S., Anthony Sulpizio, Paul Hieble, Victor Macdonald, Mark Chavez, and Giora Feuerstein. Liposome encapsulation attenuates hemoglobin-induced vasoconstriction in rabbit arterial segments. J. Appl. Physiol.82(6): 1826–1835, 1997.—Free hemoglobin (Hb) induces a potent vasoconstrictor response that may limit its therapeutic application as a red blood cell replacement. We have investigated whether encapsulation of stroma-free Hb (SFHb) or cross-linked Hb (αα-Hb) in liposomes modulates Hb vasoactivity in isolated blood vessels. Relaxation of rabbit thoracic vessels was measured before and after exposure to acellular SFHb, αα-Hb, and liposome-encapsulated SFHb or αα-Hb. SFHb and αα-Hb caused significant inhibition of carbachol-induced relaxation at 0.5 mg/dl, whereas encapsulation inhibited vessel relaxation at 30- to 60-fold higher Hb concentrations. The contractile response of rabbit ear arterial segments to electrical stimulation in the presence of acellular αα-Hb resulted in a 150% increase (EC150) in contractile amplitude at 0.23 mg/dl, whereas the EC150 for encapsulated αα-Hb was 13.7 mg/dl. Mechanistic studies of the vasoconstrictor activity of Hb demonstrated that acellular αα-Hb had no effect on norepinephrine release in the rabbit ear artery. In addition, neither acellular nor encapsulated αα-Hb preparations inhibited endothelial nitric oxide (NO) synthase activity isolated from bovine pulmonary artery. However, inhibition of vessel relaxation by acellular or encapsulated αα-Hb was reversed by the NO donorS-nitrosylpenacillamine, implicating Hb-NO binding as a possible mechanism for the vasoconstrictor response. In vitro stopped-flow kinetic studies of Hb-NO binding showed similar rates of reaction for conversion of oxyhemoglobin to methemoglobin (metHb; <2 ms), followed by rapid conversion of metHb to NO-Hb (300 ms) for both acellular and encapsulated αα-Hb, demonstrating that liposome encapsulation does not retard NO-Hb binding. The attenuated vasoactivity of encapsulated Hb may, therefore, result from the limited access of encapsulated Hb to NO imposed by the physical size of the liposome and reduced penetration of Hb across the vascular endothelium.
- nitric oxide
- blood substitute
the search for a red blood cell substitute remains driven by frequent shortages of blood in major urban centers, the logistics of providing blood for military medicine, and continuing public concern over the safety of the global blood supply (11). The development of chemically modified acellular hemoglobin (Hb) solutions as oxygen-carrying red blood cell substitutes has been hampered by a number of adverse reactions, including nephro- and neurotoxicity, mononuclear phagocytic system activation, and coagulation thrombosis after infusion in exchange-transfused and hemorrhagic-shock models (see Ref. 53 for review). Mild adverse reactions in human safety trials have slowed the progress toward use of a Hb-based red blood cell substitute in the clinic (27, 31, 45, 56).
There is a long and rich history of investigation into the vascular reactivity of Hb solutions. Early studies with transfusion of hemolysate preparations demonstrated in animals and humans a rebound hypertension that persisted for as long as 2–3 h after transfusion (2, 3, 30). This effect was also observed in purified Hb preparations in normovolemic, exchange-transfused, or hemorrhaged animals with increases in mean arterial pressure 20–40 mmHg over that expected by a simple volume replacement (18, 36, 50). The hypertensive response elicited by infusion of acellular Hb is associated with reduced cardiac output and increased systemic and pulmonary vascular resistance (3, 36,50). Although isolated vascular segments provide limited information wth regard to complex systemic vascular responses, these models have proved useful in discerning possible mechanisms of Hb-induced vascular reactivity. The vasoconstrictor activity of acellular Hb solutions has been demonstrated in isolated organs, vascular segments, and microvascular beds in situ at pharmacological doses (8,15, 24, 26, 52). Regional blood flow studies after stroma-free Hb (SFHb) administration have shown increased blood flow to the heart, brain, liver, gut, and kidneys in the dog (34). These findings were recently confirmed in the rat with the use of a chemically cross-linked acellular Hb (αα-Hb) preparation (47). It has been suggested by these investigators that the hypertensive response of acellular chemically modified Hb is thus manifested by a redistribution of blood flow. Administration of acellular Hb in humans has also resulted in increased mean arterial pressure. This hypertensive response may be related to observations of gastrointestinal vasoconstriction that have been suggested to cause the esophageal peristalsis, abdominal cramping, nausea, and vomiting observed in healthy normal volunteers (4, 33).
One strategy that has been explored in the effort to deliver Hb as an oxygen-carrying resuscitative fluid is to encapsulate the protein in liposomes (7, 10, 12, 20). Encapsulation of Hb has been demonstrated to increase circulation persistence, reduce kidney distribution, and enable freeze-dried preservation for long-term storage (43, 44). Early preparations of encapsulated bovine Hb induced a transient hypertension and reduced cardiac output with increased heart rate and total peripheral resistance index after administration in normovolemic rats (37). These effects were shown to correlate with the soybean-based lipid components of the liposome vehicle that included contaminating lysolecithin and may have involved a direct vasoconstrictor response induced by increased thromboxane A2 and platelet-activating factor (38, 41). Modifying the lipid components of the liposome to include a synthetic distearoylphosphatidylcholine (instead of the soy-based lecithin) was shown to alleviate some of these untoward hemodynamic responses. Encapsulated-Hb preparations used in exchange transfusion in rats and rabbits showed a return to basal mean arterial pressure with no rebound hypertension observed (16, 39, 40). Administration of liposome-encapsulated Hb with hypertonic saline in a hemorrhagic shock model in the rat also decreased mortality and improved tissue oxygenation, with rapid recovery of basal mean arterial pressure (40).
In these studies, we present the first examination of the vasoactivity of encapsulated Hb in isolated vascular segments. Using the rabbit aorta preparation, we investigated the effect of free and encapsulated Hb (chemically modified by intramolecular cross-linking and native SFHb) on the vasorelaxation produced by endogenous or exogenous sources of nitric oxide (NO). Using the isolated rabbit ear artery preparation, we quantified the effect of free and encapsulated αα-Hb on electrical stimulation-induced vasoconstriction. Additionally, we correlated the vasoactivity of encapsulated and acellular Hb to the kinetics of NO-Hb binding in encapsulated and acellular forms determined by in vitro stopped-flow kinetic studies.
MATERIALS AND METHODS
SFHb, αα-Hb [Hb cross-linked at Lys99 between the α-subunits with bis(3,5-dibromosalicyl)fumarate] were obtained from pilot plant production by the Blood Detachment of the US Army (54). After production, the following αα-Hb characteristics were measured: methemoglobin (metHb) = 4.7%, oxygen-carrying capacity (P50) = 21.5 Torr, phospholipids <0.1 μg/ml, endotoxin = 0.125 endotoxin units/ml, and free iron = 1.93 × 10−6 mol Fe/mol heme. The material passed both US sterility and rabbit-pyrogen testing. SFHb characteristics were similar, with the exception of P50 (5.2 Torr) and metHb (0.5%). Liposome-encapsulated Hb preparations were prepared as previously described (44). Briefly, lipid components distearoylphosphatidylcholine, dimyristoyl phosphatidylglycerol (Avanti Polar Lipids, Birmingham, AL), cholesterol (Cal Biochem, San Diego, CA), and α-tocopherol (Sigma Chemical, St. Louis, MO), in a 10:9:0.9:0.1 mol ratio, were dissolved in chloroform, and the solvent was then evaporated under vacuum. The film was then hydrated with sterile water and lyophilized to a powder. The freeze-dried powder was hydrated with either SFHb or αα-Hb (14 g/dl). The multilamellar liposome-encapsulated Hb solutions were processed with a high-shear apparatus (Microfluidics, Boston, MA) to form unilamellar vesicles. Vesicle size was measured by photon-correlation spectroscopy (Coulter Electronics, Norwalk, CT). Unencapsulated Hb was removed by tangential-flow filtration, using 300-kDa cutoff polysulfone filters (Millipore, Boston, MA). The washed liposome-encapsulated Hb preparations were then suspended in modified Dulbecco’s phosphate-buffered saline to 35% by volume. Liposome vehicles without Hb were prepared in the same fashion, with the replacement of modified Dulbecco’s phosphate-buffered saline for Hb. Endotoxin levels of all acellular and encapsulated Hb preparations were measured by using a chromophore-based Limulus amebocyte lysate test that was read at 580 nm (Cape Cod Associates).
New Zealand White rabbits (2–3 kg) were killed by pentobarbital sodium overdose administered intravenously in the ear vein (according to the Guide for Care and Use of Laboratory Animals, DHHS Publ. No. 86–23). The thoracic aorta was rapidly excised and dissected free of adhering connective tissue and fat. The isolated aortas were cut into ring segments ∼3- to 4-mm wide and suspended, via two tungsten wire (0.008-in. diameter) hangers positioned in the lumen of each segment, in 50-ml organ baths containing Krebs-Henseleit solution having the following composition (in mM): 119 NaCl, 4.7 KCl, 1.2 KH2PO4, 2.5 CaCl, 25 NaHCO3, and 11 glucose. The baths were maintained at 37°C and aerated continuously with 95% O2-5% CO2.
A resting tension of 2 g was applied to each segment, and the tissues were equilibrated under these conditions for 45–60 min before testing. Isometric tensions were measured by using a force-displacement transducer (Grass FT03c) and recorded on a physiograph (Narcotrace 80). A concentration-response curve to norepineprine was performed in all tissues to assist in the selection of a submaximal contractile concentration of norepinephrine to be used for the relaxation studies. It was determined that concentrations of 0.1–0.2 μM produced a submaximal (50–70%) contraction of these aortic tissues. The tissues were then exposed to the appropriate concentration of norepinephrine. After the contraction stabilized, the ability of the tissues to relax to NO was determined by exposing the tissues to increasing concentrations (10−7 to 10−5 M) of carbachol (a cholinergic agonist that promotes the endogenous release of NO) until no further relaxation could be obtained. At the completion of the carbachol concentration-response curve, the norepinephrine and carbachol were removed from the bath by rinsing several times with Krebs-Henseleit buffer, and the baseline tension was readjusted to 2 g. The tissues were then exposed to SFHb, αα-Hb, liposome-encapsulated SFHb, or αα-Hb over an equivalent milligram Hb dose range. Thirty minutes later, the carbachol-induced relaxation was repeated, as described above, in the presence of the Hb preparation. The relaxation response at each carbachol concentration is expressed as a percent reduction in the norepinephrine-induced contraction.
The isolated perfused rabbit ear artery preparation used in these studies was modified from a previously described method (49). Male New Zealand White rabbits (3–5 kg) were killed by pentobarbital sodium overdose administered intravenously via the marginal ear vein. The ears were removed, and a 3- to 5-cm portion of the central ear artery was dissected from the base of the ear and immediately placed in Krebs-Henseleit solution. With the aid of a dissecting microscope, the artery was cannulated at both ends with polyethylene tubing (Intramedic PE-50). The tubing was secured at each end of the artery with two ties of 4-0 surgical sutures. The cannulated artery was mounted in a tubular glass chamber and simultaneously perfused intraluminally and superfused extraluminally with Krebs-Henseleit buffer by using a Gilson Minipulse 2 multichannel pump. At the pump setting used in these studies, the resultant flow rate was 2.6 and 1.3 ml/min for the intraluminal and superfusion circuits, respectively. The intraluminal perfusion pressure was measured with a pressure transducer (Statham, P23A) and recorded on a physiograph (Narco Biosystems).
All Hb preparations were delivered to the tissue via the intraluminal flow. The Krebs-Henseleit reservoir was aerated continuously with a 95% O2-5% CO2 mixture and was maintained at 37°C. The artery was equilibrated under these conditions for ∼20 min, during which time perfusion pressure stabilized at 15–25 mmHg. The artery was then challenged with a 1 μM concentration of norepinephrine via the intraluminal flow to determine the responsiveness of the artery to a contractile agent and to replenish the sympathetic nerve terminals. After this challenge, the norepinephrine was removed from the perfusion flow. When perfusion pressure returned to baseline, cocaine (final concentration at the artery = 1 μM) was added to the Krebs-Henseleit reservoir to augment the contractile response to electrical stimulation by inhibiting the uptake of neuronally released norepinephrine.
The periarterial sympathetic nerves were stimulated electrically through platinum electrodes located at both ends of the glass chamber to elicit a contractile response. The stimulation consisted of a 0.5-s train of 0.7-ms-duration square-wave pulses at a frequency of 10 Hz and supramaximal voltage (90–100 V). This stimulus was applied at 4-min intervals, as we have found that this cycle does not fatigue the tissues, provides a consistent contractile amplitude that can be maintained for several hours, and there is a 1-min lag time before the test solutions reach the tissue. The amplitude of the contractile response to electrical stimulation stabilized within 60 min and typically produced a peak increase of 30–50 mmHg over baseline perfusion pressure. When the contractile response to electrical stimulation stabilized, increasing concentrations of the Hb preparations were delivered to the artery via the intraluminal flow. The artery was exposed to each concentration of the acellular- or encapsulated-Hb preparation for two cycles of stimulation, because it was found that the peak effect of Hb exposure occurred within 8 min of exposure (2 stimulation cycles), and no additional effect was obtained by using a 12-min (3 stimulation cycles) exposure. Thus, all changes from baseline produced by individual Hb concentrations were calculated by using the response produced at the second stimulation. The change in contractile amplitude produced by each Hb preparation was expressed as a percent increase over baseline (i.e., response immediately preceding the introduction of Hb) amplitude.
Statistical significance of the vasoactivity of Hb and liposome-encapsulated Hb in the two models was evaluated with the Student-Newman-Keuls test. Statistical significance of varying concentrations of Hb or encapsulated Hb were calculated at maximal relaxation in the rabbit aortic segments. The 150% effective concentration (EC150) values in the electrically stimulated ear artery model for the two preparations were also evaluated and are noted in the text.
To determine the effect of acellular Hb on norepinephrine release, ear arteries were incubated for 30 min in Krebs-Henseleit buffer maintained at 37°C that contained 20 μCi of [3H]norepinephrine (sp. act., 55.5 Ci/mmol; New England Nuclear, Boston, MA). After the incubation period, the vessels were mounted in the perfusion chambers, as described above, and perfused/superfused for 2 h to remove any [3H]norepinephrine that might have bound to the surface of the blood vessel. [3H]norepinephrine release was induced at 20-min intervals by exciting the periarterial adrenergic nerves via a 2-min electrical stimulation consisting of square-wave pulses of 0.7- ms duration at 2-Hz frequency and supramaximal voltage. The superfusate was collected for 2 min before each stimulation to determine basal [3H]norepinephrine release and again during the stimulation period to determine stimulated [3H]norepinephrine release. Rauwolscine (100 nM, Carl Roth, Karlsruag, Germany), an α-adrenoceptor antagonist that enhances norepinephrine release by blocking the inhibitory α2-adrenoceptor located on the nerve terminal, was examined as a positive control for [3H]norepinephrine release. Each superfusate sample was mixed with 16 ml of Ready Safe scintillation cocktail (Beckman Instruments, Fullerton, CA) and counted on a beta-scintillation counter. Release was expressed as the ratio between 3H release during stimulation divided by 3H release during the preceding basal sample.
Endothelial NO synthase (eNOS) activity was determined by using enzyme isolated from the solubilized microsomal fraction from bovine pulmonary artery endothelial cells grown to confluence in culture (6). The assay measured the formation of [3H]citrulline, a by-product of NO synthesis, from [3H]arginine with the use of ion-exchange chromatography to separate the negatively charged arginine from the uncharged citrulline molecules. The effect of each concentration of the Hb preparations on eNOS activity was examined as a percent inhibition of [3H]citrulline production.
Reaction kinetics of αα-Hb and liposome-encapsulated αα-Hb binding with NO were initiated by combining 0.1 M phosphate buffer (pH 7.2, 24°C, purged with N2, and equilibrated with NO) with the same buffer equilibrated with room air containing either αα-Hb or liposome-encapsulated αα-Hb. Rapid mixing (2 ms) into a spectral flow cell was accomplished with a dual-syringe stopped-flow apparatus connected to a rapid-scanning spectrophotometer (RSM-1000 Stopped-Flow; On-Line Instrument Services, Bogart, GA) (35). Spectral scans from 485.3 to 715.3 nm in 1.15-nm steps were collected at a rate of 1,000/s. The final concentration of reactants in the flow cell was as follows: NO concentration ([NO]) = 0.96 mM; αα-Hb heme concentration ([heme]) = 42 μM; liposome-encapsulated αα-Hb [heme] = 6.9 μM (17.8 M intraliposomal). Optical scattering with liposome-encapsulated αα-Hb required averaging over 16 ms to increase the signal-to-noise ratio of the spectral data set. All data were analyzed with the GFIT global analysis program (28).
Vasorelaxation of precontracted rabbit thoracic aorta segments.
A typical physiograph recording of the vasodilatory response to carbachol in aortic segments precontracted with norepinephrine is shown in Fig.1 A. The relaxation to carbachol is markedly inhibited in the presence of a pharmacological dose (2.5 mg/dl) of acellular αα-Hb. The same concentration of encapsulated αα-Hb had essentially no effect on the relaxation response. The inhibition of relaxation by acellular αα-Hb was concentration dependent, with significant inhibitions of 27 (P < 0.001), 67 (P < 0.001), and 89% (P < 0.001) compared with the maximum response to carbachol produced at concentrations of 0.5, 2.5, and 5.0 mg/dl, respectively (Fig. 2). From this data, the concentration of acellular αα-Hb required to produce a 50% effective concentration (EC50) of the carbachol maximum was calculated to be 1.6 mg/dl. Acellular SFHb possessed a profile of inhibition similar to acellular αα-Hb. Concentrations of 0.5, 1.0, 2.5, and 5.0 mg/dl of acellular SFHb produced statistically significant inhibitions of 15 (P < 0.05), 32 (P < 0.05), 51 (P < 0.001), and 84% (P < 0.001), respectively, compared with the maximum control carbachol relaxation (Fig.3). The EC50 calculated from these data for acellular SFHb was 2.4 mg/dl. These results indicate that the two acellular Hb preparations possess very similar potency in inhibiting NO-mediated vascular relaxation in this model.
Substantially higher concentrations of liposome-encapsulated αα-Hb were required to inhibit carbachol-induced relaxation. Concentrations of 15, 30, and 60 mg/dl of encapsulated αα-Hb produced statistically significant inhibition of 22 (P < 0.05), 29 (P < 0.005), and 60% (P < 0.005), respectively, compared with the maximum response to carbachol (Fig.4). The EC50 calculated from these data for encapsulated αα-Hb was 50 mg/dl. These results demonstrate that liposome encapsulation produced an attentuation of ∼30-fold of the vasoconstrictor activity of αα-Hb to inhibit NO-mediated relaxation compared with the acellular SFHB or αα-Hb. In vessels exposed to both acellular and encapsulated αα-Hb, addition of a supramaximal concentration (10−5 M) ofS-nitrosylpenacillamine (SNAP), an exogenous NO donor, overcame the Hb-induced inhibition of relaxation (Fig. 5). This suggests that the inhibition of vasorelaxation in this model is mediated by interactions of acellular or encapsulated Hb with NO.
Contraction of electrically stimulated perfused rabbit ear arterial segments.
A physiograph recording of the potentiation produced by acellular αα-Hb of electrically induced contraction of the perfused rabbit ear artery is shown in Fig. 6. The contractile amplitude increases dramatically over the concentration of 0.025–0.5 mg/dl compared with vessels perfused with vehicle buffer. Perfusion of αα-Hb at concentrations >0.5 mg/dl caused a direct contraction of three of six vessels without electrical stimulation. Encapsulated αα-Hb also increased the electrically stimulated contractile amplitude in a similar fashion but at a concentration nearly two orders of magnitude higher than acellular αα-Hb (0.5–62.5 mg/dl). Direct contraction of the vessel without electrical stimulation by encapsulated αα-Hb was observed in one of four vessels at a concentration of 62.5 mg/dl. The potentiation of the electrically stimulated contractile response produced with acellular and encapsulated αα-Hb is summarized in Fig. 7. Both preparations produced dose-dependent potentiation of contractile-response amplitude in this model. For purposes of comparison, the concentration required to produce a 150% increase over baseline contractile amplitude (EC150) was calculated for both Hb preparations. The EC150 for acellular and encapsulated αα-Hb was 0.23 and 13.7 mg/dl (P < 0.005), respectively, demonstrating a statistically significant 60-fold reduction in potency. These data demonstrate that encapsulation of αα-Hb attenuates the potentiation of the contractile responses in arterial vessels when compared with the acellular form.
The effect of acellular and liposome-encapsulated αα-Hb on eNOS activity was determined by measuring the conversion of [3H]arginine to [3H]citruline by enzyme isolated from bovine pulmonary artery endothelial cells. Concentrations of 0.05, 0.1, 0.5, 1.0, and 5.0 mg/dl of acellular Hb and concentrations of 2.5, 5, 10, 20, and 40 mg/dl of encapsulated αα-Hb had no effect on eNOS activity (data not shown).
[3H]Norepinephrine release from perfused rabbit ear artery.
Acellular αα-Hb also did not affect the release of norepinephrine from the perfused electrically stimulated rabbit ear artery (Fig. 8). Control stimulated-to-basal [3H]norepinephrine release ratios were 3.0 ± 0.07, 2.9 ± 0.09, and 3.0 ± 0.1 for the three stimulations preceding the introduction of acellular αα-Hb. Stimulated-to-basal [3H]norepinephrine release ratios after concentrations of 0.1, 0.5, and 2.0 mg/dl of acellular αα-Hb were 3.2 ± 0.08, 3.3 ± 0.1, and 3.0 ± 0.3, respectively. Rauwolscine increased the stimulated to basal release ratio to 9.9 ± 2.1 at a concentration of 100 nM.
Kinetics of αα-Hb-NO and encapsulated αα-Hb-NO interactions.
Stopped-flow kinetic studies with NO and Hb demonstrated that NO conversion of αα-oxyHb to αα-metHb (indicated by an absorbance peak at 630 nm) was complete within the 2-ms mixing time of the apparatus (Fig. 9,A andC). The NO-induced oxidation occurred with both αα-Hb and encapsulated αα-Hb solutions. Subsequent binding of NO to ferric heme (indicated by an increase in absorbance at 563 nm) was complete in 300 ms with both preparations (Fig. 9, B andD).
The vasoconstrictor activity of acellular Hb has been investigated in a number of in vitro and in vivo animal models. Early studies suggested that the presence of lipophilic and nonlipophilic components in Hb preparations were responsible for vasoconstriction (13, 15). Improvements in Hb-manufacturing processes have eliminated contaminants from Hb solutions and have allowed for a more direct investigation into the vasoconstrictor activity of the Hb molecule. Hb-induced vasoconstriction in vitro has been shown to be dose dependent and to vary with species and vessel origin (51). SFHb-induced vasoconstriction was demonstrated in cerebral arteries of the monkey and dog in models of subarachnoid hemorrhage and delayed cerebral vasospasm. Vasoconstrictor activity of SFHb was markedly attenuated by denuding the endothelium and treatment with indomethacin or aspirin (cyclooxygenase inhibitors), diphloretin phosphate (a prostaglandin-receptor antagonist), and imidazolyl(methyl)phenyl-2-propenoic acid (a thromboxane-synthesis inhibitor) (51). On the basis of these findings, it was suggested that the mechanism of the vasoconstrictor response of oxyHb was related to the release of arachidonic acid metabolites and prostaglandins from the endothelium and not solely to an endothelium-derived relaxation factor-related mechanism. In similar studies, the vasoconstrictor activity of SFHb in isolated dog basilar arteries was not altered by denuding the vessels (9).
Additional studies of purified αα-Hb preparations have focused attention on the release of vasoactive peptides, such as endothelin (ET) or other eicosanoids, and on adrenergic mechanisms of action (17,46). Pretreatment with phosphoramidon (a metalloproteinase inhibitor of Big ET conversion to biologically active ET) was shown to attenuate the pressor response in rats after the infusion of a preparation of αα-Hb similar to that used in the present studies. Pressor responses to α-adrenoceptor agonists were observed to be enhanced in the rat after the administration of αα-Hb (47, 48). It has been suggested that αα-Hb may sensitize α1- and α2-adrenoceptors because the enhancement of norepinephrine-, phenylephrine-, or clonidine-induced pressor responses was inhibited by standard α-adrenoceptor antagonists (48).
Although several mechanisms may contribute to the vasoconstrictor activity of Hb, previous reports implicate the binding of NO as the major contributor to this action. The vasoconstrictor activity of Hb in isolated vessel segments is endothelium dependent and can be blocked by inhibiting NO synthase with agents such as nitro-l-arginine methyl ester or by inhibiting guanylyl cyclase with methylene blue (9,14, 26, 32). Vasoconstrictor activity of Hb is also blocked by inhibitors of NO synthase and can be reversed by the administration ofl-arginine, the substrate for NO synthase (32, 46). The administration of cyanometHb, which is unable to interact with NO, does not evoke a hypertensive response (18, 24). Thus, the binding of NO by Hb effectively removes a major vasodilatory component of vascular tone and may be responsible for the observed Hb-induced constriction.
In our studies, encapsulation in liposomes was found to significantly reduce the vasoconstrictor activity of Hb, ∼30-fold in the rabbit aortic segment and 50-fold in the electrically stimulated ear artery. Although other mechanisms cannot be definitively ruled out, the mechanism of vasoconstrictor action for both preparations most likely involves NO binding. Complete vessel relaxation was obtained in aortic segments by the addition of SNAP, an NO donor, indicating that guanosine 3′,5′-cyclic monophosphate (cGMP)-mediated vasorelaxation within the vascular smooth muscle cell was not affected by either acellular or liposome-encapsulated Hb preparation. Previous studies with this model also demonstrate that vasoconstriction is abolished by tetrodotoxin and is inhibited in a concentration-dependent manner by the α1-adrenoceptor antagonist prazosin and is enhanced by rauwolscine (1, 55). This indicates that the contractile response is mediated by α1-adrenoceptors activated by norepinephrine released from sympathetic nerve terminals. We demonstrated that the Hb vasoactivity of acellular and encapsulated Hb in this model is not due to differences in norepinephrine release, because αα-Hb did not affect the release of [3H]norepinephrine. In addition, we can rule out the hypothesis that the differences in the vasoconstrictor activity of acellular and encapsulated Hb are due to a direct effect on NOS activity, as neither acellular nor encapsulated Hb was found to have an inhibitory effect on eNOS activity.
Because Hb is a potent ligand for NO, we have explored whether differences in the kinetics of NO binding in aqueous solutions of acellular and encapsulated Hb could account for the differences in vascular reactivity. Previous spectroscopic studies of Hb-NO ligand interactions have shown that the heme group acts as a NO sink in a diffusion-limited binding reaction that occurs at equal rates to Hb subunits and favors the T quaternary structure (19, 29). Stopped-flow kinetic measurements of Hb-NO binding reveal that the NO-induced conversion of oxyHb to metHb for both acellular αα-Hb and encapsulated αα-Hb is rapid (<2 ms). Subsequent reaction kinetics of NO-metHb for either acellular or liposome-encapsulated met-αα-Hb are similar, indicating that the liposome bilayer does not present a significant diffusion barrier to NO. This suggests that the attenuated vasoconstrictor activities of encapsulated Hb are not related to inherent differences in NO-binding kinetics between acellular and encapsulated forms.
The reduced vasoconstrictor activity of encapsulated Hb in these isolated vessel models may correlate with the lack of hypertensive response observed after the in vivo application of encapsulated Hb in exchange-transfused and hemorrhagic-shock models (16, 38-40). The mechanisms of vasoconstriction of acellular and encapsulated αα-Hb in these models are not clear. We suggest that the differences in the vasoconstrictor activity between acellular and encapsulated Hb are a result of differences in the spatial distribution of Hb in the vasculature. Differences in the flow properties of the two preparations based on their size (and effective molecular weight) would result in the acellular Hb transitting closer to the surface of the endothelium on the luminal side, shortening the diffusion length of Hb-NO binding, and creating a larger concentration gradient of NO away from the smooth muscle junction. This contrasts with the larger liposome-encapsulated Hb (0.2 μm) that would transit farther away from the luminal endothelial surface, increasing the diffusion length for Hb-NO binding and decreasing the concentration gradient of NO away from the smooth muscle junction and cGMP-induced vasorelaxation.
Alternatively, the differences in acellular and encapsulated Hb vasoactivity could be related to differences in spatial distribution in the subendothelial compartment where Hb would encounter and bind NO before it reached the vascular smooth muscle. In support of this, it was recently reported that acellular Hb, but not encapsulated Hb, could pass through endothelial junctions in an in vitro model of the endothelial cell layer (22). Biodistribution studies of αα-Hb in rats are also consistent with the distribution of free Hb into compartments not accessible to the encapsulated form. The vascular persistence of αα-Hb shows a 4- to 6-h half-life at clinically relevant doses (30–50% exchange transfusion) with a considerable amount (40%) unaccounted for in the vascular pool or associated organs. This was attributed to tissue extravasation (21). In contrast, encapsulated Hb has shown 18- to 20-h vascular persistence half-life with 97% of the encapsulated Hb accounted for in the vascular pool or organs of the reticuloendothelial system (44). These data strongly suggest that acellular Hb may have access to NO at both the luminal and subendothelial sites, whereas the encapsulated form, because of its large size, may encounter NO only at the luminal site. This does not rule out other mechanisms that may explain differences in vasoconstrictor activity between acellular and encapsulated Hb, and definitive studies that examine extravasation of Hb are warranted. Other possible mechanisms of acellular- and encapsulated-Hb vasoactivity include differences in interactions with adrenergic receptors or the amount of catalytically active iron or heme available to the vascular endothelium that has been shown to cause cellular damage by upregulating heme oxygenase and ferritin (5).
Vasoconstriction has emerged as a significant concern in the clinical use of Hb-based blood substitutes. A preparation shown to have reduced vasoconstrictor liability may possess therapeutic advantages for certain clinical indications. The present study evaluates the fundamental differences in the vasoconstrictor activity of Hb and suggests that Hb presentation (by encapsulation) to the vasculature may alter the vasoconstrictor response. We have determined the effects of these preparations on vascular tone in two models:1) an in vitro model of NO-mediated vasorelaxation (precontracted rabbit aorta) and2) a model of vasoconstriction (perfused rabbit ear artery). The demonstrated attenuation of vasoconstrictor activity by the encapsulation of Hb may extend the clinical utility of Hb-based red blood cell substitutes as well as provide a useful model for pharmacological interventions in the control of vascular tone.
We thank Richard O. Cliff, Victoria Kwasiborski, Chris Driscoll, and Lloyd Lippert for technical assistance.
Address for reprint requests: A. S. Rudolph, Center for Bio/Molecular Science and Engineering, Code 6910, Naval Research Laboratory, Washington, DC 20375-5348 (E-mail:).
We gratefully acknowledge the financial support of the US Naval Medical Research and Development Command, the Office of Naval Research, and the US Army Medical Research and Development Command.
The opinions and assertions herein are the private ones of the authors and are not to be construed as official policy, or reflecting the views of the US Department of Defense.
- Copyright © 1997 the American Physiological Society