ARTIFICIAL CELLS, BLOOD SUBSTITUTES, AND BIOTECHNOLOGY Vol. 32, No. 3, pp.

339351, 2004

Active Myogenic Tone: A Requisite for Hemoglobin Mediated Vascular Contraction?

H. W. Kim,* J. Tai, and A. G. Greenburg
Brown University School of Medicine and The Miriam Hospital, Providence, Rhode Island, USA

ABSTRACT
Acellular free hemoglobin (Hb), when intravenously administered to animals and humans, elicits vascular contraction. A primary mechanism for the Hb mediated vasoconstriction is Hb scavenging of nitric oxide (NO), a potent relaxation factor, constitutively secreted by the vascular endothelium. However, in the isolated rat thoracic aorta in basal state, Hb does not elicit contraction. To investigate this apparent paradox, we assessed isolated rat aortic ring isometric contraction responses to Hb under different myogenic tone states: (1) following equilibration at a submaximal tension, (2) following agonist induced contraction, or (3) following a passive mechanical stretch. In vessel rings at basal state, Hb as high as 4 mM did not elicit any measurable contractions. In contrast, in vessel rings tone enhanced with norepinephrine, Hb as low as 0.1 mM Hb elicited a

*Correspondence: H. W. Kim, Brown University School of Medicine and The Miriam Hospital Providence, RI 02906, USA; E-mail: hae_kim@brown.edu. 339
DOI: 10.1081/LABB-200027425 Copyright & 2004 by Marcel Dekker, Inc. 1073-1199 (Print); 1532-4184 (Online) www.dekker.com

ORDER

REPRINTS

340

Kim, Tai, and Greenburg significant additional contraction. In vessel rings with passively induced tone, 4 mM Hb did not elicit a notable contraction. Similarly, in vessel rings in basal state, 0.171 mM acetylcholine, a NO dependent vasodilator, did not elicit relaxation. In these vessel rings, exogenous 8-Br-cGMP, a membrane permeable cGMP analog, did not elicit relaxation. In conclusion, in the isolated rat thoracic aorta, Hb mediated contraction may be contingent upon the state of myogenic tone.

INTRODUCTION Intravenous administration of cell-free hemoglobin (Hb) preparations as potential red cell substitutes have been shown to elicit a transient hypertension in animals and human subjects (Dunlap et al., 1995; Feola et al., 1992; Hess et al., 1993; Poli et al., 1997; Shoemaker et al., 1993). A principal cause for the hypertension is believed to be Hb scavenging of vascular endothelium derived nitric oxide (NO), a potent relaxation factor (Collins et al., 1993; Kim and Greenburg, 2000; Martin et al., 1984, 1986). NO has a high affinity for ferrous Hb (Gibson and Roughton, 1957). However, in the isolated rat thoracic aorta in the basal state, Hb does not elicit contraction. Hb elicits contraction only in vessel rings prior tone enhanced with a contractile agonist such as norepinephrine or phenylephrine (Martin et al., 1984, 1986). Endothelial NO synthase is presumably constitutively expressed and functionally active in many mammalian blood vessels to maintain normal vascular tone (Kubes and McCafferty, 2000). If so, why doesnt Hb elicit contraction in the basal state without prior tone enhancement? To investigate this apparent paradox, we explored the relationship between the state of vascular myogenic tone and Hb mediated contraction.

MATERIALS AND METHODS Preparation of Rat Aortic Rings and Tension Recording Male Sprague-Dawley rats (250350 g body weight) were anesthetized with sodium pentobarbital (5075 mg/kg, IP). Through a midline incision, the heart and lungs were removed en bloc. After a wash in Krebs buffer (in mM, NaCl, 118; KCl, 4.8; CaCl2, 2.5; MgSO4, 1.2; KH2PO4, 1.2; NaHCO3, 24; glucose, 11; and disodium EDTA, 0.03; pH 7.4), the thoracic aorta was carefully excised avoiding endothelial damage and

ORDER

REPRINTS

Active Myogenic Tone

341

placed in a petri-dish containing fresh buffer solution. Following removal of nonvascular tissues, the vessels were cut transversely with a sharp surgical scissors into approximately 2 mm ring segments and kept in Krebs buffer until use. The vessel rings were mounted between two opposing stainless hooks; one hook secured to a tissue holder while the other connected to a tension transducer (Grass Model FT03) via a silk suture (3-0). The vessel ring preparation was placed in a 25 mL experimental tissue bath containing a Krebs buffer maintained at 37 C. The buffer was oxygenated continuously by bubbling 95% O25% CO2 gas. The vessels were initially set at imposed tension of 2 g and allowed to relax for 1 h before experimental treatments were initiated. Change in tension was recorded on a polygraph (Grass Instruments, Quincy, MA). The responsiveness of vessel rings was first assessed by treating with 50 nM norepinephrine (NE); vessel rings that developed less than 0.5 g of contraction were considered nonviable and excluded. The integrity of the endothelium was assessed by treating vessels with 1050 mM acetylcholine (Ach). If vessel rings relaxed to Ach less than 10% of pretreatment values, the functional endothelium was considered absent and discarded.

Test Hemoglobin Solution Highly purified stroma-free Hb (mainly Hb fraction A0) resconstituted in a Ringers lactate solution (6.4 0.2 gHb/dL) (Hemosol, Inc., Etobicoke, Canada) was used. The Hb solution was aliquoted in 1mL vials and stored frozen 80 C until use. On the day of experiment Hb solution was thawed and diluted as needed with the Krebs buffer before use. Total Hb content and percent of oxyHb and metHb were evaluated with a IL282 Co-Oximeter (Instrument Laboratories, Lexington, MA). Only Hb solution with less than 5% metHb content was used.

Drugs and Chemicals Commercial preparation of norepinephrine for intravenous injection (levarterenol bitartarate, Winthrop Laboratories, New York, NY) was used. Acetylcholine chloride and other chemicals were purchased from Sigma Chemical Co. (Saint Louis, MO).

ORDER

REPRINTS

342

Kim, Tai, and Greenburg

Statistical Analysis Data are presented as mean 1 standard deviation (SD). Statistical significance was determined by Students paired (before and after treatment comparisons) or unpaired t-tests (between group comparisons) at p 0.05 level. For multiple comparisons, ANOVA and Neuman Keuls tests were used.

EXPERIMENTAL PROTOCOLS Mechanical Stretch (Passive Tone Enhancement) and Hb Mediated Vasoactivity A group of aortic rings (N 6 each) were mechanically stretched and their responses to Hb were compared with those of pharmacologically tone enhanced vessel rings. Each experimental run was conducted with a pair of vascular rings prepared from adjacent regions of the same vessel; one was tone enhanced with 50 nM norepinephrine and the other mechanically stretched. The tension developed by the mechanically stretched ring was manually adjusted to match that of norepinephrine treated vessel ring. Once vessel rings were stabilized at the desired tension level, 1 mM Hb was added to both vessel rings and contractile responses compared.

Effect of Agonist or Passively Induced Tone on Hb Mediated Vasoactivity Vessel ring tension responses to incremental doses of NE (0.1 nM 1 mM) were assessed in the absence or presence of 2 mM Hb. To further evaluate effect of passive tension on Hb mediated vascular contraction, vessel rings were subjected to NE (1 and 10 nM) or mechanical stretch that match the NE induced tone. A pair of vessel rings from one animal was used for each experimental run; one vessel ring was treated with a standard NE dose and the other mechanically stretched to match the tone of the NE treated vessel ring. When the steady state tensions reached an approximately equal level, both vessel rings were treated with 1 mM Hb (Figs. 3a,b) and changes in vessel ring tensions were compared.

ORDER

REPRINTS

Active Myogenic Tone

343

Effect of Exogenous cGMP on Unstimulated Vessel Ring Whether basal treatment with Ach elicit increase in endothelial NO level was assessed by treating the vessel rings up to 100 mM Ach and monitoring resultant tension changes. In addition, whether increase in basal cytosolic cyclic GMP would elicit vessel ring relaxation and allow contraction to subsequent Hb treatment was investigated by pretreating the vessel rings with 8-bromo-cGMP (Br-cGMP; 18 mM). Changes in vessel tensions were compared with NE pretreated vessels.

RESULTS Passive Vascular Tone (Mechanical Stretch) and Hb Mediated Contraction In a typical rat thoracic aortic ring preparation equilibrated for 1 h at 0.5 g resting tension, treatment with 30 nM norepinephrine (NE) increased isometric tension (tone) by approximately 0.5 g (Fig. 1, upper). In this vessel ring, subsequent treatment with 2 mM Hb elicited a notable additional contraction. A second vessel ring prepared from the same animal was mechanically stretched to a comparable tension level produced by 30 nM NE (Fig. 1, lower). In these passively tone induced vessel rings, subsequent treatment with 2 mM Hb did not elicit a notable contraction. To test whether the level of passively imposed tension (PT) had any effect on subsequent Hb mediated contraction, the vessel rings were mechanically stretched stepwise up to twice the basal tension (0.52.0 g). At each tension level, a 15-min stabilization period were allowed. The vessel rings were then treated with 2 mM Hb and changes in isometric tension (DT) followed. In these vessel rings, the total isometric tension levels increased proportionally to adjusted PT levels (Fig. 2). However, when the DT values were corrected for respective PT values, there was no notable tension increases elicited by Hb treatment at any PT level (Fig. 2).

NE Induced Tone Enhancement vs. Tone Matched Mechanical Stretch To further investigate the role of state of vascular tone in the Hb mediated contraction, a group of vessel ring pairs were pretreated.

ORDER

REPRINTS

344

Kim, Tai, and Greenburg

0.5 g 0.5g 1 min

30 nM NE

2M Hb

0.5 g Stretch 2 M Hb

Figure 1. Typical isometric tension responses of a pair of isolated rat thoracic aortic rings to hemoglobin (Hb). In a vessel ring with 30 nM norepinephrineinduced tone enhancement, 2 mM Hb elicited a notable additional contraction (upper tracing). In contrast, in a tone matched vessel ring with a passively induced tension by mechanical stretching, the same dose of Hb did not elicit a notable contraction (lower tracing).

Developed Tension (DT) after Hb (g)

DT DT-PT

0 0 1 2 3

Passive Tension (PT, g)

Figure 2. Responses of rat aortic rings with different passive tension (PT) levels to Hb. The vessel rings were mechanically stretched to different tension levels. After stabilized at a desired tension level, the vessel rings were treated with 2 mM Hb and resultant total developed tension (DT) were recorded. When adjusted for the imposed tension (DTPT), Hb treatment did not elicit any additional tension increase at all the PT values tested (0.22.5 g).

ORDER

REPRINTS

Active Myogenic Tone

0.75

345

P<0.01

Hb mediated tension increase (g)

0.50

0.25

NS

0.00

NE-Low

NE-High

PT-Low

PT-High

Figure 3. Responses of agonist or passively tone enhanced aortic rings to 2 mM Hb. In norepinephrine (NE) tone induced vessel rings, the Hb mediated additional tension increase was significantly higher in 10 nM NE (NE-High) compared with 1 nM NE (NE-Low) (P < 0.01, N 46). In vessel rings passively tone enhanced to 1 nM (PT-High) or 10 nM NE (PT-Low) levels, Hb did not elicit additional tension increases (P > 0.05, N 46).

One of the pair was treated with NE (0.1 or 1.0 nM) and the other mechanically stretched to match the tension of the NE treated vessel ring. After stabilization, both vessel rings were treated with 2 mM Hb. In 1 and 10 nM NE treated vessel rings, Hb caused significant additional tension increases in; 9.4 3.8 and 9.2 3.5%, respectively (P < 0.05) (Fig. 3). In contrast, in mechanically stretched vessel rings tone matched to either 1 or 10 nM NE, the same Hb dose did not elicit notable tension increases.

Effect of NE Doses on Subsequent Hb Mediated Contraction To assess the effect of pretreatment NE dose on subsequent magnitude of Hb mediated contraction, NE dose-responses of vessel rings were assessed in the absence or presence of 2 mM Hb. For NE dose of 0.5 nM6 mM, vessel ring contractile response to NE exhibited a sigmoidal dose-response relationship. The median effective dose (EC50) was 6.8 nM with a maximal tension of 1.4 g. For NE doses of 1 nM to 0.1 mM, vascular tension increased sharply with increasing NE doses. At NE doses higher than 50 nM, vascular tension appeared to have

ORDER

REPRINTS

346
0.4

Kim, Tai, and Greenburg

Tension Increase(g)

0.3 0.2 0.1 0.0 0.1

10

Log[NE] (M)

Figure 4. Pretreatment NE doses and 2 mM Hb mediated additional contraction. Hb elicited vascular tension increases were calculated (vascular tension developed after Hb NE tension after NE alone) and plotted against NE dose. A notable Hb mediated contraction occurred only at NE doses greater than 1 nM. The dotted line represents a regression line fitted with a variable slope sigmoidal curve.

reached a plateau. In vessel rings pretreated with 2 mM Hb, notably higher vascular tension values were observed than respective values without Hb at most NE doses. The NE dose response curve was shifted to the left. In addition, the maximal tension value was higher with Hb pretreatment. The EC50 and the maximal developed tension were 4.4 nM and 1.6 g, respectively. Interestingly, when the tension increases contributed by Hb (vascular tension with Hb NE tension with NE alone) was calculated and plotted against the NE dose, the Hb mediated tension increases appeared rather abruptly around the NE doses of 15 nM (Fig. 4).

Vascular Tone and GC-cGMP Mediated Relaxation In vessel rings in the basal state without agonist induced tone enhancement, Ach doses up to 1 mM did not elicit any notable relaxation. In these vessel rings, subsequent 2 mM Hb treatment did not result in an additional contraction. Pretreatment of vessel rings with 8-Br-cGMP, a membrane diffusable cGMP analog, did not elicit relaxation. In these vessel rings, 2 mM Hb still did not elicit a notable additional contraction (Fig. 5).

ORDER

REPRINTS

Active Myogenic Tone

2

347
# *
1

Tension (g)

NE Ach Br-cGMP

Pre

PostTx

PostHb

Figure 5. Hb mediated additional tension responses in 30 nM NE, 65 mM actylcholine (Ach) or 18 mM 8-brom-cyclic guanosine 3,5-mono-phosphate (BrcGMP) pretreated rat aortic rings. As shown previously, in 30 nM NE pretreated vessel rings, 2 mM Hb elicited a significant additional tension increase (P < 0.05, N 5). When Ach as high as 1 mM were treated to vessel rings at basal state without prior tone enhancement, no additional contraction occurred with subsequent Hb treatment. Similarly, treatment with Br-cGMP as high as 20 mM did not allow Hb mediated additional contraction.

DISCUSSION A primary goal of this study was to investigate the nature of relationship between the state of vascular tone and subsequent Hb mediated contraction. Abrupt increase in vascular tension by mechanical stretching caused vessel rings to gradually relax before reaching a stable tension level. This stretch induced relaxation may involve endothelial NO release. However, Hb treatment to passively stretched vessel rings did not elicit contraction. In addition, in vessel rings under the basal state (without tone enhancement) or following a passive stretch, Ach did not elicit relaxation. Acetylcholine is known to mediation vascular relaxation through NO-cGMP mediated mechanism (Furchgott, 1983; McDaniel et al., 1993). In these vessel rings, subsequent Hb treatment did not result in contraction. These results indicate that no or minimal endothelial NO production/release under such conditions. Similarly, treatment with Br-cGMP, a membrane diffusable form of cGMP, did not elicit relaxation in vessel rings in the basal state. Only in vessel rings tone enhanced with an agonist, did Ach elicit relaxation. In these vessel rings, subsequent Hb treatment elicited contraction (or inhibition of relaxation). Hb did not elicit contraction in vessel rings with passively induced tone (mechanical stretch). Hb mediated contraction occurred only after vessel rings were tone enhanced with an agonist such as norepinephrine. Interestingly, increasing the pretreatment norepinephrine dose did not result in proportionally higher vascular tension when treated with the

ORDER

REPRINTS

348

Kim, Tai, and Greenburg

same Hb dose. Rather, the norepinephrine dose and Hb mediated subsequent contraction had a sharp sigmoidal relationship indicating that Hb mediated tension development is independent of norepinephrine dose at doses above or below the activation dose of approximately 1 nM. Similar effects were also seen in aortic ring preparations tone enhanced with several other contractile agonists (Kim and Greenburg, 2001). This suggests that receptor mediated mechanisms may not be involved (Kim et al., 2002). Results from this study suggest Hb mediated contraction appears to require an active myogenic tone. Why is an active tone required for the Hb mediated contraction in the isolated rat aorta? One possible answer is that, in the isolated rat thoracic aorta, basal NO synthase is quiescent or only minimally active to produce sufficiently high levels of NO for extra-cellular Hb to be effective. Hb mediated contraction would not occur if little or no NO is present; no change in vascular tone because there is no NO to be scavenged by Hb. In many species, vascular endothelial nitric oxide synthase is believed to be constitutively active to produce low levels of NO in the basal state, but there is no definitive data about NO levels in unstimulated rat thoracic aorta. Isolated porcine pulmonary vessels have been reported to contract with Hb without prior tone enhancement (Muldoon et al., 1996). Higher NO and cGMP levels were observed in these vessels. However, in this study, vessel rings in the basal state pretreated Br-cGMP did not allow subsequent Hb mediated contraction. There may be a cGMP independent vascular relaxation pathway (Sand et al., 2002). In the absence of active myogenic tone, NO-cGMP or other relaxation mechanisms are locked effectively preventing further vascular dilation. Treatment with an contractile agonist induces development of an active myogenic tone that may switch-on the NO synthesis/release mechanism thus allowing subsequent Hb mediated contraction. Such a mechanism would prevent further vascular collapse when vascular tone is lost due to injury or pathologic conditions in the cardiovascular system and/or central nervous system. In a recent study, rat aortic ring responses to acetylcholine were reported to vary with the initial state of resting tension (Franchi-Michelli et al., 2000); relaxation at a low resting tension (0.5 g), contraction at an intermediate level, and relaxation again at a higher resting tension (1.3 g). Isometric tension responses to vasoactive agents may vary with pretreatment resting tension (or basal tone) level. Different mechanical stress levels on the endothelial and/or smooth muscle cells may differentially activate calcium and other ionic channels altering membrane potential. Alternatively, mechanical stress may activate a variety of intra- and

ORDER

REPRINTS

Active Myogenic Tone

349

intercellular signals that may alter vascular tone. For example, in cultured vascular endothelial or smooth muscle cells, cyclic stretches have shown to activate ERK, AP-1, RhoA, Ras, and other mitogen activated protein kinases (Ingram et al., 2000; Li et al., 1999; Sauzeau et al., 2000). Some of these signals, alone or in concert with other factors, may alter the state of myosin light chain phosphorylation/dephosphorylation which is an ultimate determinant for vascular relaxation or contraction. Stretch induced relaxation/contraction may be evolutionary adaptation to provide constant blood flow to critical organs. Dysfunction of this mechanism may be susceptible to vascular damages and blood pressure abnormalities as seen in some pathologic conditions such as atherosclerosis. In summary, the results of this study suggest that an active vascular myogenic tone may be a requisite to Hb mediated contraction of the isolated rat thoracic aorta. Under the basal conditions, the endothelial NO synthesis/release may be minimally active in this vessel type but upregulated upon induction of proper myogenic tone (contraction). In the absence of myogenic tone the endothelial NO synthesis/release mechanism or other relaxation mechanisms may be locked to prevent further vascular dilation. Upon restoration of adequate myogenic tone, the endothelium derived NO synthesis/release mechanism may become unlocked. This conditional relaxation mechanism may be a part of complex vascular homeostatic mechanisms.

REFERENCES Collins, P., Burman, J., Chung, H., Fox, K. (1993). Hemoglobin inhibits endothelium-dependent relaxation to acetylcholine in human coronary arteries in vivo. Circulation 87:8085. Dunlap, E., Farrell, L., Nigro, C., Estep, T., Marchand, G., Burhop, K. (1995). Resuscitation with diaspirin crosslinked hemoglobin in a pig model of hemorrhagic shock. Art. Cells, Blood Subs. and Immob. Biotech. 23:3961. Feola, M., Simoni, J., Angelillo, R., Luhruma, Z., Kabakela, M., Manzombi, M., Kaluila, M. (1992). Clinical trial of a hemoglobin based blood substitute in patients with sickle cell anemia. Surgery, Gynecology & Obstetrics 174:379386. Franchi-Michelli, S., Failli, P., Mazzetti, L., Bani, D., Ciuffi, M., Zilletti, L. (2000). Mechanical stretch reveals different components of endothelial-mediated vascular tone in rat aortic strip. Br. J. Pharmacol. 131:13551362.

ORDER

REPRINTS

350

Kim, Tai, and Greenburg

Furchgott, R. F. (1983). Role of endothelium in responses of vascular smooth muscle. Cir. Resch. 53:557573. Gibson, Q. H., Roughton, F. J. W. (1957). The kinetics and equilibria of the reactions of nitric oxide with sheep hemoglobin. J. Physiol. 136:507526. Hess, J. R., Macdonald, V. W., Brinkley, W. W. (1993). Systemic and pulmonary hypertension after resuscitation with cell-free hemoglobin. Am. J. of Physiol. 74:17691778. Ingram, A. J., James, L., Cai, L., Thai, K., Ly, H. (2000). NO inhibits stretch-induced MAPK activity by cytoskeletal disruption. J. Biol. Chem. 275:4030140306. Kim, H. W., Greenburg, A. G. (1997). Ferrous hemoglobin scavenging of endothelium derived nitric oxide is a principal mechanism for hemoglobin mediated vasoactivities in isolated rat thoracic aorta. Art. Cells Blood Subs. Immob. Biotech. 25:121133. Kim, H. W., Greenburg, A. G. (2000). Pharmacodynamic characterization of hemoglobin-induced vasoactivity in isolated rat thoracic aorta. J. Lab. Clin. Med. 135:180. Kim, H. W., Greenburg, A. G. (2001). Contraction coupled endothelial nitric oxide release: a new paradigm for a local vascular control? J. Surg. Resch. 100:9398. Kim, H. W., Tai, J., Greenburg, A. G. (2001). Alpha adrenergic activation and hemoglobin mediated contraction in the isolated rat thoracic aorta. Art. Cells, Blood Subs. Immob. Biotech. 29(5):367380. Kubes, P, McCafferty, D. M. (2000). Nitric oxide and intestinal inflammation. Am. J. Med. 109:150158. Li, C., Hu, Y., Mayr, M., Xu, Q. (1999). Cyclic strain stress-induced mitogen activated protein kinase (MAPK) phosphotase 1 expression in vascular smooth muscle cells is regulated by Ras/RacMAPK pathways. J. Biol. Chem. 274:2527325280. Martin, W., Villani, G. M., Jothianandan, D., Furchgott, R. F. (1984). Selective blockade of endothelium-dependent and glyceryl trinitrate-induced relaxation by hemoglobin and by methylene blue in the rabbit aorta. J. Pharmacol. Exp. Ther. 232:708716. Martin, W., Smith, J. A., White, D. G. (1986). The mechanisms by which hemoglobin inhibits the relaxation of rabbit aorta induced by nitrovasodilators, nitric oxide, or bovine retractor penis inhibitory factor. J. Pharmacol. 89:563571. McDaniel, N. L., Rembold, C. M., Murphy, R. A. (1993). Cyclic nucleotide dependent relaxation invascular smooth muscle. Can. J. Physiol. Pharmacol. 72:13801385.

ORDER

REPRINTS

Active Myogenic Tone

351

Muldoon, S. M., Ledvina, M. A., Hart, J. L., Macdonald, V. W (1996). Hemoglobin-induced contraction of pig pulmonary veins. J. Lab. Clin. Med. 128:579584. Ning, J., Peterson, L. M. N., Anderson, P. J., Biro, G. P (1992). Systemic hemodynamics and renal effects of unmodified human SFHS in dogs. Biomat. Art. Cells Immob. Biotech. 20:723727. Poli de Figueriedo, L. F., Mathru, M., Solanki, D., Macdonald, V., Hess, J., Kramer, G. C. (1997). Pulmonary hypertension and systemic vasoconstriction may offset the benefits of acellular hemoglobin blood substitutes. J. Trauma 42:847856. Sand, A. E., Anderson, E., Freid, G. (2002). Effects of nitric oxide donors and inhibitors of nitric oxide signaling on endothelin- and serotonin-induced contractions in human placental arteries. Acta Physiol. Scand. 174:217223. Sauzeau, V., Jeune, H. L., Cario-Toumaniantz, C., Smolenski, A., Lohmann, S. M., Bertoglio, J., Chardin, P., Pacaud, P. (2000). Cyclic GMP-dependent protein kinase signalling pathway inhibits RhoA-induced Ca2 sensitization of contraction in vascular smooth muscle. J. Biol. Chem. 275:2172221729. Shoemaker, S., Gerber, M., Evans, G., Paik, L., Scoggin, C. (1993). Initial clinical experience with recombinant human hemoglobin. In: The Vth International Symposium for Blood Substitute. San Diego, CA, March, pp. 1720.