Pharmacodynamic characterization of hemoglobin-induced vasoactivity in isolated rat thoracic aorta

Pharmacodynamic characterization of hemoglobininduced vasoactivity in isolated rat thoracic aorta H. W. KIM and A. G. GREENBURG PROVIDENCE, RHODE ISLAND

The origin and mechanism of vasocontraction observed after vascular exposure to acellular Hbs remain controversial. To help resolve the underlying mechanism, we characterized Hb-induced vasoactivities in terms of Hb purity, heme iron oxidation state, and ligand and pharmacodynamic properties. Isolated rat thoracic aortic rings with intact endothelium were suspended in oxygenated Krebs buffer, and isometric tension responses to various test Hb preparations were measured. In norepinephrine tone–enhanced aortic rings, both crude and purified Hbs exhibited similar dose-response characteristics; stroma-free Hb and HbA0, two Hb preparations with disparate purity, were equally potent in inducing vessel ring contraction. Purified Hb preparations significantly attenuated vasodilatory potency of both acetylcholine, an endothelium-dependent NO generator, and glyceryl trinitrate, an endothelium-independent NO generator. With the exception of nitrosylated Hb, ferrous Hbs, oxy Hb, and carbon monoxy Hb elicited contraction, whereas ferric derivatives, met Hb, and cyanomet Hb did not. In addition, NEM-Hb, an Hb with blocked cysteine residues, did not notably attenuate Hb vasoactivity. These results indicate that Hb itself is directly responsible for inducing contraction in the rat thoracic aortic rings. A primary mechanism for the Hb-induced vasoactivity appears to be heme iron inactivation of endothelium-derived NO. Nonheme interaction with endothelial NO does not appear to play a prominent role in this vascular model. In conclusion, Hb elicits dose-dependent contraction in isolated rat thoracic aorta with intact endothelium. Vasoactivity of Hbs, however, could greatly vary with heme iron oxidation state, nature of heme ligand, and model vessels used in the evaluation. (J Lab Clin Med 2000;135:180-7)

Abbreviations: Ach = acetylcholine; GTN = glyceryl trinitrate; Hb = hemoglobin; HbA0 = hemoglobin A0; HbCN = cyanomet hemoglobin; HbNO = nitrosylated hemoglobin; HbO2 = oxy hemoglobin; logEC50 = log median effective dose; NE = norepinephrine; NEM = N-ethyl maleimide; NEM-HbCN = N-ethyl-maleimide–modified cyanomet hemoglobin; NEM-HbNO = N-ethyl maleimide–modified nitrosylated hemoglobin; NO = nitric oxide; RBC = red blood cell; SFH = stroma-free hemoglobin

F

ree human or animal Hbs in solution have been shown to elicit contraction in many types of mammalian blood vessels in vitro.1-3 Vascular exposure to free Hb in vivo may alter local and sysFrom the Department of Surgery, The Miriam Hospital and Brown University, Providence. Submitted for publication May 27, 1999; revision submitted October 4, 1999; accepted November 1, 1999. Reprint requests: Hae W. Kim, PhD, Department of Surgery, The Miriam Hospital, 164 Summit Ave, Providence, RI 02906. Copyright © 2000 by Mosby, Inc. 0022-2143/2000 $12.00 + 0 5/1/104463 180

temic hemodynamics with the potential for adverse pathophysiologic consequences. However, the Hbinduced vascular contractile responses appear to vary substantially with the properties of the model vascular system used (eg, endothelial integrity, vessel type, and species).4-6 In addition, even within the same vascular model, Hb-induced vasoactivity appears to vary greatly with the characteristics of Hb itself. There continues to be an ongoing debate regarding the origin and mechanism of the Hb-induced vascular contraction. Initially, vasoactivity observed after Hb infusion or perfusion to isolated organs was thought to be due to erythrocyte stromal lipids, bacterial and other impurities, or conta-

J Lab Clin Med Volume 135, Number 2

Kim and Greenburg 181

Table I. Basic characteristics of test hemoglobins SFH and HbA0

Hb concentration (g Hb/dL)* metHb content (% of total Hb) pH* p50 (mm Hg) n (Hill coefficient) Sterility Endotoxin (EU/mL) Stromal lipids (µg/g Hb)

4.5 ± 0.5 <5% 7.4 ± 0.5 10-13 2.3-2.5 Sterile <0.03 <2

SDS/PAGE 16-kd fraction† Others Non-Hb protein bands‡ Human serum albumin (mg/mL)

SFH

HbA0

92.6 7.4 Multiple 0.09

94.3 5.7 1 <0.01

Electrolytes were equivalent to standard Ringer lactate. *Adjusted to indicated values. †Percentage of dissociable Hb subunits by sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS-PAGE). ‡Determined by using isoelectric focusing and Western blot methods.

minants in the Hb preparation.7 On the basis of the recent discovery of NO as a vascular endotheliumderived relaxation factor8,9 and its unique property to avidly bind heme proteins,10 a new mechanism has been proposed: Hb scavenging of endothelium-derived NO may be primarily responsible for the observed vasoactivity.3,8 Alternatively, formation of nitrosothiol compounds, products of NO and nonheme interaction, has been proposed to play a key role in eliciting a vascular response.11,12 Hb may also stimulate production or release of endothelin and other vasoactive mediators.13 Finally, disparate oxygen affinities of different Hb preparations may contribute to the discrepancies in observed Hb vasoactivity.14 Clearly, more studies are needed to resolve controversies regarding the Hbassociated vasoactivity. To help clarify the underlying mechanism, we characterized Hb-induced vasoactivity in terms of Hb purity, dose-response characteristics, effects on endothelium-dependent and endotheliumindependent NO-producing vasodilators, and Hb molecular characteristics. METHODS Preparation of rat aortic rings and tension recording.

Male Sprague-Dawley rats (250-350 g body weight) were anesthetized with sodium pentobarbital (50-75 mg/kg administered intraperitoneally). Through a midline incision, the heart and lungs were removed en bloc. After a wash in Krebs buffer (NaCl, 118 mmol/L; KCl, 4.8 mmol/L; CaCl2, 2.5 mmol/L; MgSO4, 1.2 mmol/L; KH2PO4, 1.2 mmol/L; NaHCO3, 24 mmol/L; glucose, 11 mmol/L; and disodium ethylenediamine tetraacetic acid, 0.03 mmol/L; pH = 7.4), the thoracic aorta was carefully excised, avoiding endothelial damage, and placed in a petri dish containing fresh buffer solution. After removal of nonvascular tissues, the vessels were cut transversely with a sharp surgical scissors into 3- to 4-mm vessel ring segments and kept in Krebs buffer until use. In some vessel preparations the endothelium was removed by gently rubbing the intima with a cotton-tipped applicator. The vessel rings were mounted between two opposing stainless steel hooks; one hook was secured to a tissue holder, where-

as the other was connected to a tension transducer (Grass Model FT03) by means of a silk suture (3-0). The vessel preparation was placed in a 25-mL experimental tissue bath containing Krebs buffer maintained at 37°C. The buffer was oxygenated continuously by bubbling 95% O2-5% CO2 gas. The vessels were allowed to relax for 1 hour at 2 g of imposed tension before experimental procedures were initiated. Change in tension was recorded on a Grass Polygraph (Model 7). The responsiveness of vessels was first assessed by treating with 50 nmol/L NE. The integrity of the endothelium was then assessed by treating vessels with 33 µmol/L ACh. Functional endothelium was considered absent if vascular relaxation to Ach was less than 10% of pretreatment values. Whenever possible, control responses and responses to an agonist or an antagonist were performed on the same vessel ring after a buffer change and re-equilibration. Test agents were added cumulatively to the bathing buffer, and changes in tension were monitored. Unless otherwise noted, each set of experiments was repeated in at least 5 vessel rings. Test hemoglobin solutions. Two batches of human Hb solutions with different purity were provided by Hemosol, Inc (Etobicoke, Canada). SFH solution is a relatively pure Hb preparation with mixed Hb species (eg, HbA0, HbA1, and HbF) and contained minor non-Hb contaminants. HbA0 is a highly purified Hb solution containing only HbA0, with minimal non-Hb contaminants. Characteristics of these test Hb solutions are shown in Table I. After Hb concentration and pH adjustment, Hb solutions were placed in aliquots in 1-mL vials and stored frozen at –80°C until use. On the day of the experiment, Hb solution was thawed and diluted as needed with Krebs buffer. HbNO was obtained by adding 0.1 mL of deoxygenated SFH or HbA0 to a 0.9-mL mixture of sodium nitrite (0.1 mmol/L) and slightly excess sodium dithionite. MetHb was obtained by incubating SFH or HbA0 for 1 to 3 days at 37°C. HbCN was produced by using an HbCN reagent (DMA Inc, Arlington, Tex). The final Hb derivatives were dialyzed against 60 volumes of Krebs buffer. Total Hb content and percentage of HbO2 and metHb were evaluated with a IL282 Co-Oximeter (Instrument Laboratories, Lexington, Mass). HbNO and HbCN were evaluated by using a diode array spectrophotometer (Model 4892; Hewlett-Packard, Palo Alto, Calif). Cysteine-masked Hb (NEM-Hb) was prepared

182

J Lab Clin Med February 2000

Kim and Greenburg

A

C

B

D

Fig 1. Typical responses of rat thoracic aortic rings to selected blood components. Aortic rings in basal state were treated, and none of the test blood components elicited notable contraction (A). When aortic rings were precontracted with NE, blood components that contained Hb elicited contraction (B), whereas the plasma fraction did not (C). Similar contractile response was seen with purified Hb (D).

by reacting HbO2 with equimolar NEM reagent. NEM-HbCN was prepared similarly by reacting NEM reagent with HbCN. Preparation of rat blood components. Heparinized rat whole blood was obtained from healthy rats. RBCs were prepared from rat whole blood by centrifugation at 2500g for 15 minutes and removing plasma and buffy coat. After 3× wash with normal saline, the RBCs were resuspended in normal saline solution (washed RBC). RBC lysate (hemolysate) was prepared by adding an equal volume of cold sterile water to a volume of packed washed RBCs. One batch of RBC lysate was filtered through a 0.2-µm membrane filter to remove RBC stroma. Finally, Hb concentrations of washed RBCs, RBC lysate, and filtered RBC lysate were adjusted to 7 ± 0.5 g Hb/dL by adding appropriate amounts of normal saline solution as necessary. Drugs. Commercial preparations of NE (levarterenol bitartrate; Winthrop Laboratories, New York, NY) and GTN (Baxter Health Care Corp, Deerfield, IL) were used. Other chemicals were obtained from Sigma Chemical Co (Saint Louis, MO). Statistical analysis. Unless otherwise indicated, data are presented as means ± 1 SD. Statistical significance was determined by using the Student paired (before and after treatment comparisons) or unpaired t tests (between-group comparisons) at a P level of .05. For multiple comparison, analysis of variance and Neuman-Keuls tests were used. The relative effectiveness of vasoactive agents is represented as EC50 (concentration of a drug at which it is half maximally effective). The EC50 values were obtained by using nonlinear regression software (GraphPad Prism software, San Diego,

CA). RESULTS Vasoactivities of selected blood components. When aortic rings in basal state were treated, none of the test blood components elicited notable contractile responses (Fig 1, A). Vessel rings showed contractile responses only after submaximal tone enhancement with NE. In NE tone-enhanced aortic rings with intact endothelium, all blood components that contained Hb elicited contraction, whereas the plasma fraction did not (Fig 1, B-D). At 4 µmol/L Hb (in tissue bath), washed resuspended RBCs, RBC hemolysate, micropore-filtered RBC hemolysate, and chromatographically purified Hb (SFH) caused maximal tension increases of 0.8 ± 0.08 g, 0.88 ± 0.13 g, 0.78 ± 0.15 g, and 0.81 ± 0.11 g, respectively. These tension increases were significant when compared with that of plasma (P < .01). However, the maximal tension increases were not significantly different among Hb-containing groups. In addition, when response rates over a 2-minute period were plotted, all blood components that contain Hb showed a similar pattern (Fig 2). The first-order response parameters, rate constant and time constant, were not significantly different. In endothelium-removed aortic rings, treatment with 2 to 4 µmol/L Hb did not elicit a significant contraction; the mean tension changes were

J Lab Clin Med Volume 135, Number 2

Kim and Greenburg 183

Fig 2. Response rate relationships of selected blood components. Values are expressed as changes in aortic ring tension in grams (means ± SD, n = 4-6 each). *Trace amount of Hb present because of hemolysis during preparation.

Fig 4. Hb inhibition of Ach-induced relaxation of rat thoracic aortic rings precontracted with 50 nmol/L NE. Vessel rings were preincubated with or without 2 µmol/L Hb before Ach dose-response measurements. Values are expressed as percentage changes in tension over pretreatment values (means ± SD, n = 6 each).

Fig 3. Dose-dependent contractile responses of aortic rings to HbA0 and SFH, two human Hb preparations differing in purity but with comparable Hb concentrations. Values are expressed as percentage changes in tension (means ± SD, n = 6 each) over pretreatment values. See Table I for comparison of characteristics.

Fig 5. Hb inhibition of GTN-induced relaxation of rat thoracic aortic rings precontracted with 50 nmol/L NE. Vessel rings were preincubated with or without 2 µmol/L Hb before GTN dose-response measurements. Values are expressed as percentage changes in tension over pretreatment values (means ± SD, n = 6 each).

1.1% ± 2.6% (n = 9, P > .05) over the pretreatment values. Vasoactivities of Hbs with disparate purity. To assess the possible role of minor Hb components (eg, HbA1 and HbF) and non-Hb contaminants in Hb-mediated vasoactivity, contractile responses of rat aortic rings to SFH and HbA0, two Hb preparations with different purity, were compared. The main difference between the two Hb solutions is the presence of non-Hb proteins in SFH (Table I). Despite notable differences in purity, both SFH and HbA0 elicited similar contractile responses, including almost identical dose-response curves in rat aortic rings (Fig 3). The threshold Hb concentration for induction of vasoconstriction was approximately 4 nmol/L for both Hb preparations. At 4 µmol/L, there was no significant difference (P > .05) between the maximal contractions elicited by SFH and HbA0; SFH increased vascular tension 43.8% ± 17.1% over the pretreatment values, whereas HbA0 increased vascular tension 41.0% ± 26.4%. The logEC50s for SFH and HbA0 were –6.82 ± 0.19 mol/L and –6.86 ± 0.10 mol/L, respectively and were not significantly different (P > .05). Of note, the SFH- and HbA0-induced contractile responses were completely reversible after a buffer

washout. Effects of Hb on NO-producing vasodilators. In NE tone-enhanced endothelium intact vessel rings, Ach treatment elicits relaxation. In these vessel rings SFH pretreatment attenuated the Ach-induced relaxation. Pretreatment with 2 µmol/L SFH significantly attenuated the vasodilatory effectiveness of Ach (0.06-63 µmol/L). In addition, 2 µmol/L SFH pretreatment shifted the Ach dose-response curve to the right, as indicated by the increase in logEC50 values (from –7.37 ± 0.73 mol/L without SFH to –6.05 ± 0.30 mol/L with SFH; P < .02; Fig 4). The maximal relaxation was reduced to 16.2% ± 4.9% with SFH compared with 43.0% ± 7.7% without SFH (P < .001, Fig 4). Similarly, 2 µmol/L SFH pretreatment also reduced the vessel ring relaxation response to 0.6 to 600 nmol/L GTN. The GTN dose-response curve significantly shifted toward the right with SFH pretreatment; the logEC50 values with and without 2 µmol/L SFH were –7.25 ± 0.14 mol/L and –8.27 ± 0.09 mol/L, respectively (P < .001). However, unlike Ach, Hb inhibition of GTN-induced relaxation did not accompany a significant reduction in the maximal relaxation; the maximal relaxation values

184

J Lab Clin Med February 2000

Kim and Greenburg

A

C

B

D

Fig 6. Effects of different heme iron oxidation states and ligands on Hb-induced contractile responses of isolated rat thoracic aortic rings. Although ferrous Hb (HbO2) at 0.2 µmol/L elicits a notable contraction in Ach prerelaxed aortic rings (A), ferric Hb (metHb) does not, even at 10 times higher Hb concentrations (B). Similarly, HbCN does not cause notable contraction (C). Prenitrosylated HbNO at 2 µmol/L does not cause contraction; subsequent treatment with ferrous Hb elicits immediate contraction (D).

with and without 2 µmol/L SFH were 52.7% ± 14.1% and 57.1% ± 2.5%, respectively (P > .05, Fig 5). Hb molecular characteristics and vasoactivity. The nature of Hb ligand and heme iron oxidation state profoundly influenced Hb-mediated vessel ring responses. With the exception of HbNO, all ferrous Hb tested caused contraction in rat aortic rings with or without prior relaxation. HbO2 elicited contraction at relatively low Hb concentrations (threshold ~2 nmol/L; Fig 6, A). HbCO exhibited similar potency in eliciting vessel ring contraction. Ferric (Fe+3) Hb derivatives, metHb and HbCN, at 0.2 to 2.0 µmol/L elicited reduced or no contraction (Fig 6, B and C). In contrast, HbNO, an Hb ligand with NO, did not cause a notable contraction; subsequent treatment with 2 µmol/L HbO2 produced an immediate contraction (Fig 6, D). Treatment with human serum albumin solution (2.8 µmol/L), a nonheme control protein, did not elicit a notable vasoactive responses. Endothelial NO could also interact with nonheme residues of globin chain. For example, NO could interact with reactive amino acid residues of Hb, such as cysteine exposed on molecular surface. To investigate this possibility, cysteine residues on Hb (Cysβ93) were blocked with NEM, a cysteine-specific reagent, to produce NEM-Hb. Treatment of aortic rings with NEMHb did not notably attenuate contractile responses; in fact, NEM-Hb–induced contraction response closely

resembled that of HbO2 (Fig 7, A and B). In addition, NEM-HbCN, an Hb heme and nonheme NO-binding sites blocked Hb, did not elicit contractile responses at all (Fig 7, C). Similarly, no notable contraction was seen with NEM-HbNO, a NEM-Hb ligand with NO (Fig 7, D). DISCUSSION

The purpose of this study was to characterize acellular Hb-induced vasoactivity in isolated rat thoracic aortic rings in terms of Hb purity, dose-response characteristics, effects on NO-producing vasodilators, and the molecular characteristics of Hbs. The results of this study show that blood components that contain ferrous Hb elicit concentration-dependent contraction in isolated rat thoracic aortic rings. Furthermore, crude and purified Hbs appear to cause comparable contractions if Hb concentrations are equivalent independent of purity. In NE tone-enhanced endothelium intact rat thoracic aortic rings, blood components that contain Hb (eg, washed red cells, hemolysate, and filtered hemolysate) elicited contraction with similar response rate characteristics. Furthermore, two Hb solutions, prepared from the same source but by different purification procedures, which resulted in different Hb purity, elicited similar dose-response characteristics (eg, EC50 and threshold concentrations) in rat thoracic aortic rings. If the Hb-mediated vasoactivities were due to contaminants, such as stromal lipids, plas-

J Lab Clin Med Volume 135, Number 2

Kim and Greenburg 185

A

C

B

D

Fig 7. Rat thoracic aortic ring responses to cysteine-modified, heme site–modified, or both types of Hbs. Compared with SFH (A), NEM-Hb, an Hb with cysteine residues blocked with cysteine-specific reagent NEM, does not notably attenuate Hb vasoactivity (B). Hb with both cysteine and heme site blocked (NEM-HbCN) does not elicit contraction (C); subsequent treatment with SFH caused contraction. Similarly, NEM-Hb preliganded with NO does not cause contraction (D).

ma proteins/peptides, or environmental contaminants (eg, bacteria and chemicals used in purification), there should have been a reduction in the contraction with HbA0, an ultrapure Hb preparation. These results reinforce the assertion that Hb itself, rather than environmental contaminants or residuals from incomplete purification (eg, stromal lipids, erythrocyte enzymes, and plasma proteins), is a primary causative agent for the observed Hb-mediated contraction in rat thoracic aorta. With the recent discovery that NO is an endothelium-derived vascular relaxation factor,8,9 the theory of Hb as a principal causative agent fits well because Hb has a high affinity for NO.10 Because cell-free Hb allows closer contact with the vascular endothelium, Hb could then more readily interact with endothelial NO, inactivating its vasodilatory property on the vascular smooth muscle. That Hb-induced contraction occurs only in vessels with intact endothelium further supports the view that Hb induces vascular contraction through interaction with endothelium-derived NO. Both Ach and GTN are believed to mediate vasodilation through NO generation, but Ach requires the presence of an endothelium, and GTN does not.15-17 That Hb inhibits both types of NO-dependent vasodilators also supports the theory of Hb inactivation of NO as a primary mechanism in Hb-mediated vascular smooth muscle contraction. Interestingly, although Hb pretreatment decreased the maximal response to Ach, the maximal response of GTN was not affected by Hb pretreatment. This suggests that these agents elicit vascular relaxation through NO generation, but the mode

and level of NO production may differ. Mechanisms of NO production by these agents do appear to be distinct. Ach acts as an agonist for NO production through stimulation of endothelial NO synthase, whereas GTN produces NO directly through a yet unknown endothelium independent pathway. Taken together, Hb inactivation of endotheliumderived NO appears to be a primary cause of the Hbmediated vasocontraction in the isolated rat thoracic aorta and possibly in other mammalian blood vessels. Ferric (Fe+2) Hb affinity for NO is extremely high (dissociation constant ~10–12 mol/L).18 Ferric (Fe+3) Hb is also known to react with NO and could cause vasocontraction.19 In contrast to ferrous Hb, however, ferric Hb reaction with NO is much slower and reversible.20 The dissociation constant for metHb and NO reaction is reported to be approximately 5 × 10–4 mol/L.18 This would appear to explain the minimal or absent vasoactivity of ferric Hb, making it less likely to cause any physiologically significant hemodynamic effects in vivo. The molecular characteristics of Hb profoundly affected Hb vasoactivity in rat thoracic aortic rings. Results from this study and others2,6 suggest that ferrous Hb derivatives with heme sites that are available for reaction with NO (eg, HbO2, deoxy Hb, and HbCO) are generally vasoactive. Actually, NO induces oxidative conversion of HbO2 to metHb and nitrite/nitrates rather than direct replacement of O2 with NO, which would produce HbNO.18,21 Only deoxyHb binds NO directly at the heme iron forming HbNO. This may explain why only relatively low levels of HbNO were

186

Kim and Greenburg

detected after exposure of HbO2 to gaseous NO or in animals with sepsis/endotoxemia, a condition known to produce elevated NO levels. Additionally, partially deoxygenated Hbs or valency hybrid Hbs (α+3β+2 or α+2β+3) could produce Hb derivatives (such as a-only or β-only nitrosylated Hbs) with variable vasoactivities.22 Reactive groups of nonheme Hb sites (eg, amines and sulfhydril groups) could also interact with endothelial NO. For example, cysteine residues of the Hb β-chain (Cysβ93) have been suggested as sites of NO interaction. This may play an important role in blood pressure regulation.11,12 In this study, however, blocking cysteine residues with NEM, a cysteine-specific reagent, did not notably attenuate Hb vasoactivity; NEM-Hb was as potent as unmodified Hbs (eg, HbA0 or SFH) in eliciting contraction in rat aortic rings. In contrast, blocking heme sites with a cyano group in native or NEM-modified Hbs completely abolished the ability of Hbs to induce vessel ring contraction. These results indicate that, under the current experimental conditions, the contribution of these nonheme sites to overall Hbinduced vasoactivity appears to be negligible. Which of these NO-binding sites contributes more significantly to vasoactivity in vivo remains to be tested; the role of nonheme site interactions with endothelial NO in vivo is not known. However, when the 4 high-affinity heme sites are available for NO interaction, the nonheme sites are not likely to be principal players in overall Hb-mediated vasoactivity. The rate-limiting step in NO-Hb interaction appears to be NO diffusion into the heme pocket.18,23 In ferrous Hb and myoglobin, amino acid residues around the heme ligand–binding site, such as histidine, leucine, and valine, form a selective “gateway” for O2, NO, and other ligands. In a recent study with sperm whale myoglobin, substitution of the leucine and valine residues of the heme pocket with more space-filling aromatic residues (eg, phenylalanine or tryptophan) caused a selective steric hindrance for NO entry, creating an Mb with lower NO-binding affinity but normal oxygenbinding characteristics.18 This strategy may lead to production of a mutant human Hb with low NO-binding affinity and yet with reasonable oxygen-binding characteristics. A recent study examined blood pressure responses to various modified Hbs after 50% exchange transfusion in conscious rats.14 Interestingly, the investigators found an inverse correlation between NO-binding affinity and mean arterial pressure response. Hb solutions that exhibited either transient or no significant blood pressure increase showed tighter NO binding than Hb solutions that exhibited significant or sustained increases in blood pressure. On the basis of these observations, the authors assert that Hb scavenging could not have

J Lab Clin Med February 2000

been the cause of the hypertensive responses, but rather Hb oxygen affinities and Hb solution properties are more important factors. This assertion appears to be in conflict with the results of the present study. The reason for the discrepancies is not apparent. A possible explanation includes differences in ferrous heme concentrations in test Hb solutions; PEG-Hb and β82-Hb, two Hb preparations that exhibited the lowest blood pressure increases, also had the lowest heme concentrations. It is thus possible that the lower blood pressure increases observed with PEG-Hb and β82-Hb might have been due to less available NO-binding sites. Hb interaction with NO may still play a major role in Hb-induced changes in blood pressure responses. The finding that HbNO formation coincides with an increase in blood pressure after Hb infusion in endotoxemic rats24,25 and prenitrosylated Hbs elicit little or no aortic ring contraction observed from this study supports such a possibility. On the basis of the current findings, it appears that acellular human or animal Hbs can cause vasoconstriction in vivo. Indeed, Hb infusion can and does lead to varying degrees of hypertension, ranging from a mild transient change to a more sustained hypertension depending on factors such as Hb characteristics, administration protocol, and conditions of experimental subjects.26-32 However, it should be noted that in vivo hemodynamic responses are influenced by many factors, including volume effect, viscosity, Hb interactions with blood cells, various mediators, and drugs. Differences in cNOS enzyme and GC/cGMP levels in basal or agonist-stimulated states may also alter overall hemodynamic responses to Hb infusion. Furthermore, local and central hemodynamic control mechanisms and responses and effects of anesthesia33 may all mask the true Hb effects confounding interpretation. When evaluating Hb-associated hemodynamic changes, one should carefully consider various factors involved. Whatever the underlying cause or mechanism, Hbinduced vasoconstriction is reversible and can be modulated with a variety of agents, including nitrovasodilators and phosphodiesterase inhibitors.3 In conclusion, free Hbs elicit dose-dependent contraction in isolated rat thoracic aorta with intact endothelium. A primary mechanism for the Hb-induced vasoactivity appears to be ferrous Hb interaction with endothelium-derived NO. The vasoactivity of Hbs, however, can vary greatly with the molecular characteristics of Hbs, such as heme iron oxidation state and nature of heme ligand. Therefore when assessing Hbmediated vasoactivity, one should carefully define the characteristics of the Hb tested and the models used in the evaluation.

J Lab Clin Med Volume 135, Number 2

REFERENCES

1. Martin W, Villani GM, Jothianandan D, Furchgott RF. Selective blockade of endothelium-dependent and glyceryl trinitrate-induced relaxation by hemoglobin and by methylene blue in the rabbit aorta. J Pharmacol Exp Ther 1984;232:70816. 2. Muldoon SM, Ledvina MA, Hart JL, Macdonald VW. Hemoglobin-induced contraction of pig pulmonary veins. J Lab Clin Med 1996;128:579-84. 3. Kim HW, Greenburg AG. Ferrous hemoglobin scavenging of endothelium derived nitric oxide is a principal mechanism for hemoglobin mediated vasoactivities in isolated rat thoracic aorta. Artif Cells Blood Substit Immobil Biotechnol 1997;25:121-33. 4. Martin W, Smith JA, White DG. The mechanisms by which hemoglobin inhibits the relaxation of rabbit aorta induced by nitrovasodilators, nitric oxide, or bovine retractor penis inhibitory factor. J Pharmacol 1986;89:563-71. 5. Jing, M, Panico FG, Panico JL, Ledvina MA, Bina S, Muldoon SM. Diaspirin cross-linked hemoglobin does not alter isolated human umbilical artery or vein tone. Artif Cells Blood Substit Immobil Biotechnol 1996;24:621-8. 6. Hart JL, Ledvina MA, Muldoon SM. Actions of diaspirin cross-linked hemoglobin on isolated rat and dog vessels. J Lab Clin Med 1997;129:356-63. 7. Feola M, Simoni J, Fishman D, Tran R, Canizaro PC. Biocomaptibility of hemoglobin solutions I. Reactions of vascular endothelial cells to pure and impure hemoglobins. Artif Organs 1989;13:209-15. 8. Furchgott RF, Kahn MT, Jothianandan D. Evidence supporting the proposal that endothelium-derived relaxation factor is nitric oxide. Thromb Res 1987;7(Suppl):5. 9. Ignarro LJ, Buga GM, Wood KS, Byrns RE, Chaudhuri G. Endothelium-derived relaxing factor produced and released from artery and vein is nitric oxide. Proc Natl Acad Sci USA 1987;84:9265-9. 10. Gibson QH, Roughton FJW. The kinetics and equilibria of the reactions of nitric oxide with sheep hemoglobin. J Physiol 1957;136:507-26. 11. Jia L, Bonaventura C, Bonaventura J, Stamler JS. S-nitrosohemoglobin: a dynamic activity of blood involved in vascular control. Nature 1996;380:221-6. 12. Stamler JS, Jia L, Eu JP, et al. Blood flow regulation by Snitrosohemoglobin in the physiological oxygen gradient. Science 1997;276:2034-7. 13. Gulati A, Singh G, Rebello S, Sharma AC. Effects of diaspirin crosslinked and stroma-reduced hemoglobin on mean arterial pressure and endothelin-1 concentration in rats. Life Sci 1995;56:1433-42. 14. Rohlfs RJ, Bruner E, Chiu A, et al. Arterial blood pressure responses to cell-free hemoglobin solutions and the reaction with nitric oxide. J Biol Chem 1998;273:12128-34. 15. Bolotina VM, Najibi S, Palacino JJ, Pagano PJ, and Cohen RA. Nitric oxide directly activates calcium-dependent potassium channels in vascular smooth muscle. Nature 1994;368:850-3. 16. Ignarro LJ. Endothelium-derived nitric oxide: pharmacology and relationship to the actions of organic nitrate esters. Pharm Res 1989;6:651-9.

Kim and Greenburg 187

17. Feelisch M. Biotransformation to nitric oxide of organic nitrates in comparison to other nitrovasodilators. Eur Heart J 1993;14:123-32. 18. Eich RF, Li T, Lemon DD, et al. Mechanism of NO-induced oxidation of myoglobin and hemoglobin. Biochemistry 1996;35:6976-83. 19. Alayashi AI, Brockner Ryan RE, Frantantoni JC, Cashon RE. Redox reactivity of modified hemoglobins with hydrogen peroxide and nitric oxide: toxicological implications. Artif Cells Blood Substit Immobil Biotechnol 1994;22:373-86. 20. Sharma VS, Traylor TG, R Gardiner. Reaction of nitric oxide with heme proteins and model compounds of hemoglobin. Biochemistry 1987;3837-43. 21. Doyle MP, Hoeskstra JW. Oxidation of nitrogen oxides by bound dioxygen in hemoproteins. J Inorg Biochem 1981;14:351-8. 22. Kruszyna R, Kruszyna H, Smith RP, Thron CD, Wilcox DE. Nitrite conversion to nitric oxide in red cells and its stabilization as a nitrosylated valency hybrid of hemoglobin. J Pharmacol Exp Ther 1987;241:307-13. 23. Liu X, Miller MJS, Joshi MS, Sadowaska-Krowicka H, Clark DA, Lancaster JR. Diffusion-limited reaction of free nitric oxide with erythrocytes. J Biol Chem 1998;273:18709-13. 24. Greenburg, AG, Kim HW. Nitrosyl hemoglobin formation in-vivo after intravenous administration of a hemoglobinbased oxygen carrier in endotoxemic rats. Artif Cells Blood Substit Immobil Biotechnol 1995;23:271-6. 25. Kim HW, Breiding P, Greenburg AG. Enhanced modulation of hypotension in endotoxemia by concomitant nitric oxide synthesis inhibition and nitric oxide scavenging. Artif Cells Blood Substit Immobil Biotechnol 1997;25:153-62. 26. Ning, J, Peterson LMN, Anderson PJ, Biro GP. Systemic hemodynamics and renal effects of unmodified human SFHS in dogs. Biomaterials Artif Cells Immobil Biotechnol 1992;20:723-7. 27. Hess JR, Macdonald VW, Brinkley WW. Systemic and pulmonary hypertension after resuscitation with cell-free hemoglobin. Am J Physiol 1993;74:1769-78. 28. Dunlap E, Farrell L, Nigro C, Estep T, Marchand G, Burhop K. Resuscitation with diaspirin crosslinked hemoglobin in a pig model of hemorrhagic shock. Artif Cells Blood Substit Immobil Biotechnol 1995;23:39-61. 29. Collins P, Burman J, Chung H, Fox K. Hemoglobin Inhibits Endothelium-dependent relaxation to acetylcholine in human coronary arteries in vivo. Circulation 1993;87:80-5. 30. DeAngeles DA, Scott AM, McGrath AM, et al. Resuscitation from hemorrhagic shock with diaspirin crosslinked hemoglobin, blood, or hetastarch. J Trauma 1997;42:406-14. 31. Kasper S, Walter M, Grune F, Bishoff A, Erasmi H, Buzello W. Effects of a hemoglobin-based oxygen carrier (HBOC201) on hemodynamics and oxygen transport in patients undergoing preoperative hemodilution for elective abdominal aortic surgery. Anesth Analg 1996;83:921-7. 32. Johnson JL, Moore EE, Offner PJ, Haenel JB. Resuscitation of the injured patient with polymerized stroma-free hemoglobin does not produce systemic or pulmonary hypertension. Am J Surg 1998;176:612-7. 33. Grissom TE, Bina S, Hart J, Muldoon SM. Effect of halothane on phenylephrine-induced vascular smooth muscle contractions in endotoxin-exposed rat aortic rings. Crit Care Med 1996;24:287-93