- Email: [email protected]
The development of hemoglobin solutions as red cell substitutes: Hemoglobin solutions
Pergamon
0955-3886(94)
Transfus. Sci. Vol.16, No.I, pp. 5-17, 1995 Copyright © 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0955-3886/95 $9.50 + 0.00
The Development of Hemoglobin Solutions as Red Cell Substitutes: Hemoglobin Solutions Steven A. Gould, MD* Lakshman R. Sehgal, PhD Hansa L. Sehgal, BS Gerald S. Moss, MD
• Although the efficacy of hemoglobinbased oxygen carriers was established more than 60 years ago, all prior clinical trials have demonstrated significant toxicity characterized by renal dysfunction, gastrointestinal distress, and systemic vasoconstriction. The mechanisms of these toxicities now appear to be understood. Tetrameric forms of the hemoglobin molecule extravasate from the circulation and interact with endothelial derived relaxing factor, leading to unopposed vasoconstriction. Although numerous efforts are underway to chemically modify the native tetramer, it is likely that all tetrameric forms of the hemoglobin molecule will continue to extravasate. We have focused on developing a polymerized form of hemoglobin that is virtually free of unreacted tetramet. The development and characterization of this polymerized pyridoxylated hemoglobin solution (Poly SFH-P) is described. Clinical trials have been completed successfully in volunteers, and are now underway to assess the safety and efficacy of Poly SFH-P as a clinically useful red cell substitute in the treatment of acute blood loss in the setting of trauma and surgery. •
Department of Surgery, Michael Reese Hospital and Medical Center, Chicago University of Illinois, College of Medicine, Chicago Northfield Labs Inc., Evanston, IL ,60201, U.S.A. Author for correspondence: Steven A. Gould, Northfield Laboratories Inc., 1560 Sherman Avenue, Suite 1000, Evanston, IL 60201, U.S.A.
INTRODUCTION There continues to be great interest in developing a clinically useful 02 carrier to serve as a red cell substitute. The goal is to develop a safe and effective alternative to human blood. Although the current blood supply is safer than ever due to improved donor screening and testing, it is likely that disease transmission will always occur. ~ This is due to the inevitable occurrence of new viruses, and to the small but real incidence of false negative screening tests, often due to the window period of infectivity prior to conversion of the markers used for screening. In addition, the use of allogeneic blood will always involve the need for compatibility testing, and include a limited shelf life. The potential benefits of a red cell substitute include universal compatibility and the accompanying i m m e d i a t e availability, freedom from disease transmission, and long term storage. The two potential candidates to serve as clinically useful red cell substitutes include hemoglobin solutions and perfluorochemical emulsions? We have had considerable experience with the perfluorochemicals, including the clinical trial in patients with a 20% perfluorochemical emulsion, a-s The results showed particular preparation to be ineffective as an 02 carrier. The current status of the newer perfiuorochemical emulsions is reviewed elsewhere in this edition. The remainder of this paper will
6 Transfus. Sci. Vol. 16, No. 1
deal with the history and development of hemoglobin solutions.
tein and prepare it in a form that would be safe and useful in the clinical setting. 6
BACKGROUND
CONSIDERATIONS FOR A HEMOGLOBIN SOLUTION
The basic concept of developing a hemoglobin solution can be illustrated by examining the red blood cell {Fig. 1). The "active ingredient" in the red blood cell is the four-part hemoglobin molecule, the tetramer, that chemically binds and actually carries the oxygen. The hemoglobin molecule has several important characteristics. For example, 1 g of h e m o g l o b i n binds 1.39 mL of oxygen and is almost fully saturated with oxygen at ambient pressure. Few, if any, biologically acceptable substances have a greater oxygen-binding capacity. Oxygen is normally unloaded from hemoglobin in the capillaries at a pO2 of approx. 40 torr, allowing oxygen molecules to diffuse from hemoglobin to the intracellular mitochondria without producing interstitial hypoxia. The physiologic capability of the hemoglobin molecule is dear. The surrounding cell m e m b r a n e accounts for the need for compatibility testing and the limited shelf life. It has long been known that although the cell itself has a finite lifespan, the hemoglobin protein itself is rather durable and will survive and function, outside of the cell. The history of the development of hemoglobin solutions is the tale of the efforts to harvest the hemoglobin pro-
RED BLOOD CELL
Cell ane Mem~br
Hemogl MoleCmUl ) obiene
There are several properties that must be considered when describing the nature of a hemoglobin solution. The first is the source of the hemoglobin to be used as the starting material (Table 1). We have always felt that the human source is the best known and readily available, making it the preferred choice. The bovine source has been proposed as a way of providing easier access to an urdimited supply. The recombinant and transgenic approaches have received considerable interest because of their use of the m o d e r n molecular biology approaches. However, there is no evidence to suggest any physiologic or functional benefit of using any source other than human hemoglobin. In fact, the choice of source has not been a crucial issue in the development of a red cell substitute. The limiting factor in this field has been the safety of the hemoglobin solution. It is now evident that unmodified tetrameric h e m o g l o b i n is unsafe because of the associated toxicities of renal dysfunction, gastrointestinal distress, and vasoconstriction. 6 Therefore, the most important property of a hemoglobin solution is the nature of the hemoglobin preparation itself. Unmodified tetramer dissociates into dimers when removed from the red cell (Fig. 2). It is now believed that the observed toxicities are due in large part to this dissociation, an the presence of free tetramer in the circulation. Efforts to eliminate these toxicities have consisted of a variety of attempts to stabilize the simple unmodified tetramer. Table 2 lists the various preparations. Each effort Table 1. Hemoglobin Source •
Figure 1. A schematic representation of the red blood cell containing the hemoglobin tetramer.
Human
• Bovine • Recombinant •
Transgenic
Hemoglobin Solutions
UNMODIFIED TETRAMER
Tetramer
CROSSLINKED TETRAMER
Dimers
Figure 2. Representation of unmodified tetramer as it dissociates into dimers.
Figure 4. Cross-linked tetramer.
Table 2. Hemoglobin Preparation • Unmodified tetramer • Conjugated tetramer • Cross-linked tetramer • Polymer • Encapsulation involves the use of a different approach to try and stabilize the t e t r a m e r and t h e r e b y avoid toxicity. A conjugated tetramer [Fig. 3} involves the binding of a macromolecule such as polyethylene glycol (or PEG) to form a larger molecule. A cross-linked tetramer {Fig. 4} is an intramolecular chemical link to stabilize the native hemoglobin molecule. Polymerization involves intermolecular crosslinking of tetramers as s h o w n in Fig. 5. The polymerization results in a variety of m o l e c u l a r w e i g h t sizes. Finally, encapsulation as shown in Fig. 6, involves the " p a c k a g i n g " of the h e m o g l o b i n m o l e c u l e w i t h i n a lipid
CONJUGATED TETRAMER"
l
7
~-~;=~Macromolecule ~ (e.g.PEG)
Figure 3. Conjugated tetramer.
POLYMER
Figure 5. Polymers composed of 2, 3, and 4 tetramers linked together.
ENCAPSULATION
~
Liposome~
. . . . IODIR
Figure 6. Liposome encapsulated hemoglobin. membrane or liposome to form a "synthetic" red blood cell. Only testing in h u m a n s will e v e n t u a l l y resolve the issue of safety with these various preparations. However, it is clear that the nature of the actual hemoglobin preparation itself is far more important than the choice of source material.
8 Transfus. Sci.
Vol. 16, No. 1
The third characteristic of any h e m o g l o b i n s o l u t i o n is the i n t e n d e d clinical use. Table 3 lists some of the various potential applications for a safe and effective h e m o g l o b i n solution. Although they represent a diverse set of circumstances, it is likely that the major role of a hemoglobin solution will be as a red cell s u b s t i t u t e for use in acute blood loss. Table 3. Hemoglobin Clinical Use
• • • • • •
Acute blood loss Hemodilution Ischemia Cardioplegia Organ preservation Cancer therapy
There are numerous clinical trials u n d e r w a y in b o t h v o l u n t e e r s and patients using a variety of these approaches. U l t i m a t e l y , each effort involves a selection of h e m o g l o b i n source, preparation, and intended clinical use. Table 4 s h o w s a list of the source and preparation for each of the current approaches in clinical trials. Table 4. Current Approaches
• • • •
Human polymer, tetramer-free Bovine polymer, tetramer present Human cross-linked tetramer Recombinant cross-linked tetramer
DEMONSTRATION OF EFFICACY
The concept of using a hemoglobin solution as an oxygen carrier was first fully tested by A m b e r s o n e t al. in 1934. 7 A m b e r s o n prepared a crude red cell lysate, which was an unmodified tetrameric form of the hemoglobin solution. In a rather remarkable experiment, he gradually removed all of the blood from cats and replaced it with this primitive hemoglobin solution. The results demonstrated the proof of concept and paved the way for all future efforts. The cats survived on a short-term basis, and were able to walk normally. In an absolutely elegant maneuver, Amberson evaluated neurologic function by holding the cats
on their backs and dropping them. The cats were able to land upright, demonstrating a fully intact nervous s y s t e m that is required in order to perform this c o m p l i c a t e d neurologic event. The p o t e n t i a l u t i l i t y of h e m o g l o b i n solutions was firmly established. THE PROBLEM OF TOXICITY
Based on this early d e m o n s t r a t i o n by Amberson, the efficacy of a hemoglobin solution was uncontestable. However, 60 years later there is still no h e m o globin s o l u t i o n in clinical use. The reason for this relates to the results in clinical trials. 8-1' These trials have consistently demonstrated major toxicities: renal dysfunction, gastrointestinal distress, and s y s t e m i c v a s o c o n s t r i c t i o n characterized by a rise in blood pressure and a fall in heart rate. '2 The Savitsky et al. s t u d y in 1978 '1 clearly e s t a b l i s h e d the futility of further efforts to develop a safe t e t r a m e r i c form of h e m o g l o b i n solution. MECHANISMS OF TOXICITY
The mechanisms of toxicity of the tetramer n o w s e e m to be reasonably well understood. In general, these toxicities are related to the presence of the tetramer outside of the red cell and the dissociation into dimers. The renal dysfunction is thought to be related to filtration and excretion of the hemoglobin molecule, primarily in the form of dimers, although vasoconstriction is also likely to play a role. The vasoconstriction and other adverse effects are c u r r e n t l y thought to be related to extravasation of the tetrameric hemoglobin molecule. 13 Our understanding of the control of v a s o c o n s t r i c t i o n of blood vessels has been improved with the identification of e n d o t h e l i u m derived relaxing factor (EDRF), also known as nitric oxide. 14''s It is well known that EDRF (nitric oxide) avidly binds to hemoglobin. Figure 7 is a graphic representation of the control of vasoconstriction. It is a cross-section of a blood vessel representing the lumen,
Hemoglobin Solutions 9
CONTROL OF VASOCONSTRICTION Muscle Cells
Tetramer (T) - Extravasates and Binds EDRF Both Sides
!
Casoconstriction
L .......
Cells
Figure 7. Control of vasoconstriction. the endothelial cell layer, and the interstitial space within which reside the vascular smooth muscle cells. EDRF is produced by the endothelial cells and is released on both the luminal and the abluminal sides. The physiologic activity of EDRF occurs on the vascular smooth muscle cells in the interstitial space. The EDRF released into the l u m e n is rapidly bound by the hemoglobin within the circulating red cells, and does not act on the smooth muscle cells. It is also evident that red cells do not normally move beyond the endothelial cell lining into the interstitial space, laving the EDRF available to exert its vasorelaxant effect. The vasoconstriction that occurs with the tetrameric form of the hemoglobin is considered to be due to extravasation of the hemoglobin beyond the endothelial cell barrier, and the immediate binding of the EDRF to the hemoglobin, leading to unopposed vasoconstriction. This is illustrated in Fig. 8. The presence of the tetramer in the interstitial space where EDRF normally acts thereby disrupts the normal control of vascular relaxation. Our working hypothesis has been that prevention of extravasation will prevent vasoconstriction, since EDRF would function normally. It has also been our contention that all forms of tetramer will extravasate and produce both vasoconstriction and the other toxicities historically associated with hemoglobin solutions. 13'16'17 The only forms of hemoglobin that
Figure 8. The effect of tetramer on extravasation and binding of EDRF. would appear to prevent extravasation include the polymer and encapsulation. The cross-linked tetramers have indeed been shown to extravasate, t3'~7 Since the encapsulation approach is still in the developmental stage, only the polymer has been able to be evaluated in the clinical setting. Figure 9 illustrates the likely scenario when the polymer is infused. As with the red cell, the polymer will exist only in the lumen, leaving the EDRF to exert its normal physiologic vasorelaxant role on the vascular smooth muscle cell in the interstitial space. Our approach, therefore, has been to specifically design a hemoglobin solution using a two step process to create a tetramer-free form of polymerized hemoglobin. ~8-2t The first step is the polymerization, which results in an array of different size polymers. The
P o l y m e r (P)
-
Binds EDRF in Lumen But No Extravasation
!
N0 Vasoconstriction
Figure 9. Representation of polymer and absence of extravasation.
10 Trans[us. ScL Vol. 16, No. 1
POLYMERIZATION AND PURIFICATION 0 0 0 0 oooo
61utaraldehyde
o0%%00
0 0 00(~3
&8
°&88coS°°
o o Oo
Polymer
Figure 10. Preparation of Poly SFH-P, including polymerization and removal of unreacted tetramer. second step, which is just as important, is the removal of all unreacted tetramer. Figure 10 is a representation of this process. Glutaraldehyde is used as the cross-linking (polymerizing) agent. The second step is to remove virtually all unreacted tetramer, leading to a pure polymeric solution. The remainder of this paper will deal with the historical development of our human polymerized h e m o g l o b i n solution, including the physiologic observations that led to the characteristics of our current preparation. UNMODIFIED HEMOGLOBIN
The basic, unmodified hemoglobin solution is currently prepared from outdated blood, beginning with the washing and lysis of the red cells (RBCs} with pyrogen-free water. A series of filtration steps permits the complete separation of the RBC membrane debris (stroma} from the hemoglobin molecules. The resultant solution is referred to as stroma-free
h e m o g l o b i n (SFH}. Because the RBC antigens are located on the cell membrane, SFH is universally compatible and can be infused w i t h o u t regard to specific blood type. The properties of this unmodified, tetrameric or "stripped" h e m o g l o b i n solution are given in Table 5. Although SFH can be prepared with a hemoglobin concentration of 14 g/dL, this solution has a colloid osmotic pressure (COP} greater than 6 0 m m H g , which renders it unacceptable for clinical use. 22 The hemoglobin concentration of 7 g/dL is iso-oncotic. The low Pso is due to the loss of the organic ligand, 2,3-diphosphoglycerate {2,3-DPG}, during preparation. The [02] curve of SFH is thus both anemic { ~ hemoglobin content} and shifted to the left { ~ pso}. Despite these limitations, SFH supports life in primates in the absence of RBCs. 2~ Animals survive a total exchange transfusion with SFH to zero hematocrit with maintenance of normal 02 c o n s u m p t i o n (Vo2], cardiac output, and arteriovenous oxygen content difference (Cao2-Cvo2], although a decline from baseline values occurs in some of these measures. In addition, a considerable decrease occurs in the pf¢o2 from approx. 50 to 20 torr. The pfZo2 is the partial pressure at w h i c h oxygen unloads from the hemoglobin molecule and is in equilibrium with the tissue po2. This decline indicates a marked increase in oxygen extraction and is the mechanism used to compensate for the fall in hemoglobin content and pso. This low pf¢'o2 is of some concern and led us to a t t e m p t to restore a more normal value. 23
Table 5. Properties and Parameters of Stroma-ffee Hemoglobin and Whole Blood Properties and parameters Stroma-free hemoglobin Whole blood Hemoglobin content {g/dL} 6-8 12-14 Oxygen-carrying capacity (vol%) 8.0-11.0 16-19 Binding coefficient (mL O2/g Hbl 1.30 1.30 Ps0 {torr} 12-14 26-28 Methemoglobin {%} <2 <1 Colloid osmotic pressure (torr) 18-25 18-25 Osmolarity {mOsm] 290-310 290-310
Hemoglobin Solutions
PYRIDOXYLATED HEMOGLOBIN A shift to the left in the [Oz] curve, with no change in Vo2, cardiac output, or Cao2-Cvo2, produces a decrease in the pfdoz (Fig. 11). Attempts to establish a n o r m a l ps0 by the simple addition of 2,3-DPG to the h e m o g l o b i n solution were u n s u c c e s s f u l because the ligand disappears rapidly from the circulation after infusion. However, a modification of the hemoglobin molecule by the addition of pyridoxal phosphate results in a pyridoxylated hemoglobin (SFH-P) with a Pso of 20-22 tort, which is considerably higher t h a n the pso of u n m o d i f i e d SFH34-26 We evaluated SFH-P in eight baboonsY Four received SFH and four received SFH-P, with a final hemoglobin content of 7 g/dL and zero hematocrit. The p(Zoz levels were significantly higher at the end of the exchange in the animals receiving SFH-P. A l t h o u g h hemodynamic parameters were normal, a decline from the baseline values occurred in both groups. These results illustrate three points. First, a shift to the right in the dissociation curve (1' Pso} results in an increased p~Zo2because all else remained constant. This observation is of physiologic i m p o r t a n c e because oxygen unloading can occur at a higher tissue po~. Second, although it was increased, the pV¢o2 level in the animals treated
11
with SFH-P was still substantially lower t h a n the n o r m a l value of 40-50 torr found in control animals. Third, hemodynamic function still showed a reduction from the baseline values. It became apparent that a nonanemic hemoglobin solution was requiredY NONANEMIC ISO-ONCOTIC HEMOGLOBIN SOLUTION The advantages of a nonanemic hemoglobin solution are self-evident. Such a solution w o u l d have the same oxygen capacity as w h o l e blood. In addition, according to our data, the infusion of a n o n a n e m i c solution should be associated with normal p~/o2 levels, even at zero h e m a t o c r i t ? 3 The principal obstacle to n o r m a l i z a t i o n of h e m o g l o b i n concentration is the effect of an elevation in protein concentration on oncotic pressure. The relationship b e t w e e n h e m o globin concentration and oncotic pressure is shown in Fig. 12. At hemoglobin c o n c e n t r a t i o n of 7 g/dL, the oncotic pressure is similar to that of plasma (20 torr). In contrast, at hemoglobin levels of 15 g/dL, oncotic pressure increases by more than 300%. The infusion of such a solution would theoretically produce large shifts of fluid from the extravascular space into the i n t r a v a s c u l a r space. These changes are likely to be harmful. One approach to producing a non-
| Pso SFH-P
02
Content
COP (torr)
4O 2O
P02 Figure 11. [02] curves showing a leftward shift in P~owith a constant Cao2 - Cvo2 leads to a lower pVo2. Curve B is shifted to the left compared with curve A. (From Ref. 33.)
0
4
8 12 [Hb] gm %
t
18
Figure 12. Relationship between COP and hemoglobin concentration for SFH-P. (From Ref. 34.)
12
Transfus. Sci. Vol. 16, No. 1
anemic hemoglobin solution w i t h normal COP values is polymerization of the hemoglobin. The COP of any solution is proportional to the n u m b e r of colloidal particles. If a 15 g/dL solution of hemoglobin could be polymerized, the result would be a reduction in COP, whereas no change would occur in hemoglobin concentration (Fig. 13). This idea was tested in our laboratories.IS The hemoglobin solution obtained was pyridoxylated by modification of previously described techniques. Briefly, pyridoxal-5'-phosphate was added to the solution at a 4:1 molar ratio. The solution is transferred to a sealed reservoir and deoxygenated with nitrogen (N2) by using a gas exchanger. Deoxygenation of the solution is continued until the po2 was below 10 torr and the [02] was below 1 vol%. Sodium borohydride was t h e n added to the solution, and the reaction c o n t i n u e d for 2--4 h. The goal of the polymerization is to normalize the oxygen capacity while m a i n t a i n i n g the COP w i t h i n normal limits (20-25 tort). Several parameters can be manipulated to obtain, in a reproducible manner, the polymerized solution of choice. 19 These are the hemoglobin concentration, the concentration of the polymerizing agent, the rate of addition of the latter, and the rate of mixing. Finally, the duration for which the above-mentioned parameters are controlled will in t u r n d e t e r m i n e the reaction yield, molecular weight distri-
bution of the polymeric species, oxygen affinity (pso), and finally the viscosity of the solution. The polymerization is conducted to a physiologic end point with respect to COP. The reaction is thus monitored by the decrease in COP. Figure 14 shows the changes in COP monitored during the polymerization process conducted over a 3-h period. As can be seen from the figure, the reproducibility of the process for five batches is acceptable. The polymerization and evolution of the polymeric species are also followed by s o d i u m dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE). Figure 15 shows the formation of new polypeptide bands during the process. Typically, four to six bands are seen, ranging in molecular weights from 16,000 to 96,000. The first four bands (16,000-64,000 Da) represent greater than 90% of all species. With the formation of polymers, one would expect changes in the viscosity of the solution. Figure 16 shows the changes in viscosity observed for five different batches. The viscosity at the start of polymerization ranges from 1.0 to 1.5 cp and increases to reach a range of 1.9-2.1 cp at the end. This represents a relatively modest increase but is consistent with the absence of high-molecular-weight species in the solution, as d e m o n s t r a t e d by SDSPAGE. The changes in ps0 during the poly-
POLYSFH-P 0 000 O0 O0
%%%00
O oo%
@
~P
(torr)
O 0 0 00 0 0 1 2 Polymerization Time (Hrs)
3
[Hb] - 15 gm/dL [Hb] - 15 gm/dL COP > 70 torr =-,- COP = 25 torr
Figure 13. Polymerization. (From Ref. 34.)
Figure 14. Decrease in COP measured during the course of polymerization for five different batches. (From Ref. 34.)
Hemoglobin Solutions 13
Figure 15. Evolution of new polypeptide bands over the course of the polymerization process, as monitored by SDS-PAGE. (From Ref. 34.)
POLYSFH-P
POLYSFH-P Viscosity20 (cP)
Pso 24 (torr) 2O Y
1.0
~ 1
~
,
2
3
Polymerization Time (Hrs)
/
i 1
I 2
I 3
PolymerizationTime (Hrs)
Figure 16. Changes in viscosity of five batches of hemoglobin solution during polymerization. (From Ref. 34.)
Figure 17. Alteration in Pso for five different batches of hemoglobin solution during polymerization. (From Ref. 34.)
merization were followed and are shown in Fig. 17. A modest increase in Pso was n o t e d in all five b a t c h e s shown. This increase reflects a decrease in buffering capacity of the solution with polymerization. It is consistent with the observation that a drop in pH is observed during the course of the polymerization. This is not surprising because it is some of the same amino groups that are involved
in the buffering as well as the cross-linking. The characteristics of the final product are listed in Table 6. The solution at 12-15 g/dL has a COP between 20 and 25 torr. The m o l e c u l a r w e i g h t range based on SDS-PAGE is between 16,000 and 96,000 Da. The n u m b e r average m o l e c u l a r w e i g h t calculated from the COP is approx. 150,000 Da. The average
IS 16:1-§
14
Transfus. Sci.
Vol.16, No. 1
Table 6. Properties of Polymerized Pyridoxyl-
ated Hemoglobin Property
S T A B I L I T Y OF P O L Y S F H - P Parameter range
Hemoglobin (g/dL) 12-14 Oxygen-carrying capacity 16-19 (vol%) Methemoglobin (%) <5 Molecular weight range 64,000-400,000 Number average molecular 150,000 weight Pso (tort) 18-22 Binding coefficient (mL O2/ 1.30 g Hb) COP (tort) 20-25 molecular weight obtained by ultracentrifugation is 120,000. The Ps0 of the solution under standard conditions {pH = 7.40, pco2 = 35 tort, temperature = 35°C) ranges from 18 to 22 tort. The viscosity of the solution {hemoglobin concentration = 8.0 g/dL} ranges from 1.9 to 2.2 cp. The phospholipid content is below the detection limits of the thinlayer chromatography technique used [<0.4 mg/dL). We have followed the stability of the polymerized solution when stored at 4-8°C as a sterile, pyrogen-free product. The stability with respect to rate of conversion to methemoglobin is shown in Fig. 18. the m e t h e m o g l o b i n remains below 2% for a 150-day period. The stability of the polymeric species was determined by monitoring the
S T A B I L I T Y OF P O L Y S F H - P (Met Hb)4"0I
% z
o
~ Time(Days)
Figure 18. Changes in methemoglobin concentration on storage at 4-8°C, as a sterile, pyrogen-free solution. Each point represents the mean + SEM. {From Ref. 34.]
T-150DaysA Batch- !
k Batch-2
Batch-3
Figure 19. High-pressure liquid chromatography scans for three batches at the time of manufacture and 150 days later. No changes are noted. {From Ref. 34.1
solution with high-pressure liquid chromatography, SDS-PAGE, and viscosity. No changes were observed for at least a 1-year period. Figure 19 shows the highpressure liquid chromatographic scans for three different batches at the time of manufacture and 150 days later. No visible changes occurred. Finally, the changes in viscosity during a 150-day storage period were monitored. Once again, no change in viscosity was observed. Thus the polymerized stromafree h e m o g l o b i n (Poly SFH-P), w h e n stored at 4-8°C, appears to be quite stable. EFFICACY OF POLY SFH-P
Seven adult baboons were anesthetized, paralyzed, intubated, and mechanically ventilated with room airY '29 The respiratory rate and tidal v o l u m e were adjusted to maintain a pco~ between 35 and 45 tort before the start of the study and were not changed during the study. The animals were surgically prepared with arterial and central venous catheters for infusion, blood sampling, and monitoring. A thermal dilution balloontipped catheter was floated into the pulmonary artery. A Foley catheter was inserted into the urinary bladder. Standard hemodynamic monitoring was performed for electrocardiogram, arterial pressures, p u l m o n a r y capillary wedge pressure, and central venous pressure.
Hemoglobin Solutions
Cardiac output was determined by the thermal dilution method. The study was conducted with the use of ketamine anesthesia. After stabilization of the animals, a set of baseline measurements was obtained. An isovolemic exchange transfusion with the Poly SFH-P was then performed. Whole blood was removed in 50mL aliquots and was replaced with approximately equal volumes of the infusate. Additional volume adjustments were made as required to maintain the pulmonary capillary wedge pressure at baseline values. The exchange was stopped at hematocrits of 20, 10, and 5% to obtain additional sets of measurements. The exchange transfusion was then carried out to obtain a complete washout of the RBCs. A hematocrit of less than 1% was achieved. These animals, at zero hematocrit, had a Poly SFH-P concentration of approx. 10g/dL. They were then exchanged transfused with dextran 70 to a hemoglobin concentration of 1 g/dL. 3° The data from the second half of the study were compared with a control group {n=6) that underwent an exchange transfusion with dextran 70 to a hemoglobin concentration of 1 g/dL. Arterial and mixed venous blood gases were measured by standard electrodes {IL 813}. Whole blood and plasma hemoglobin levels were determined on the IL 282 Co-oximeter. Oxygen carried by whole blood hemoglobin was measured by the IL 282 Co-oximeter. The physically dissolved oxygen in the aqueous phase was calculated from the po2 of the plasma separated by centrifugation. The Poly SFH-P [02] was measured by the IL 282 Co-oximeter. Hematocrits were determined by the microhematocrit method. Ps0 values of whole blood and plasma hemoglobin were determined with the hemoglobin oxygen dissociation curve analyzer (Hemo-O-Scan}. The efficacy of the Poly SFH-P was calculated as described previously.3~ At each hematocrit level, the arterial [02] {[O2]a}was determined for each compart-
15
ment by direct measurement or calculation. Total Vo2 was calculated as the product of the cardiac output and Cao2 Cvo2. The contribution of Poly SFH-P to oxygen delivery was calculated as the ratio of the Poly SFH-P to total arterial
[02}. Poly SFH-P 02 delivery = [02], Poly SFH-P [O2]a, total The contribution of Poly SFH-P to Vo2 was calculated as the ratio of Poly SFH-P to total Cao2 - Cvo2. Poly SFH-P Vo2 = Poly SFH-P {Cao2- Cvo2} Total {Cao2- Cvo2} All contributions were expressed as their percent values. All animals receiving Poly SFH-P survived the exchange transfusion, as did the previous animals receiving SFH-P. The final hematocrit was 0.8 + 0.4% {mean + SEMJ. The difference in the initial bag pso values of the two infusates was statistically significant (P<0.05). However, the mean in vivo plasma Pso for the Poly SFH-P was 17.0 _+ 0.5 torr, which was not significantly different from the mean value of 17.6 + 0.8 torr for the SFH-P. Both of these plasma Ps0 values are significantly below the mean baboon RBC Pso of 31.3 + 0.8 torr. The Poly SFH-P [O2]a, is significantly greater than the SFH-P value at all hematocrits {P<0.001). At a hematocrit of 5%, the [O2}a was 9.5 + 0.2 vol% for Poly SFH-P and 5.0 + 0.4 vol% for SFH-P. The percent contributions of Poly SFH-P to total oxygen delivery and total VO2 were compared with those of SFH-P. Poly SFH-P makes a significantly greater {P<0.2} contribution to total oxygen delivery than SFH-P at all hematocrits. The contribution to total Vo2 is greater by Poly SFH-P at all hematocrits, with the difference significant (P<0.005J at a hematocrit of 20%. These results document that Poly SFH-P is an effective 02
16 Trans[us. Sci. Vol. 16, No. 1 carrier, and provides benefit over the tetrameric form of SFH-P. The in vivo Pso in the Poly SFH-P animals undergoing the second exchange transfusion w i t h dextran 70 ranged from 18 to 11 torr. In contrast, the in v i v o pso of the control group ranged from 31.5 to 25.5 tort. The pVo2 was significantly lower in the Poly SFH-P group compared with the control group. Both groups of animals raised their cardiac o u t p u t in an identical manner in response to their anemia. The critical oxygen delivery in the control group was 6.6 m L / m i n / k g of body weight compared w i t h 5.7 m L / m i n / k g in the test group. These results indicate that the infusion of Poly SFH-P does not alter the normal physiologic response to progressive anemia, and therefore provides evidence of lack of vasoconstriction. CLINICAL TRIALS Based on these preclinical observations we have begun clinical trials in b o t h h e a l t h y v o l u n t e e r s and p a t i e n t s to assess the safety and efficacy of Poly SFH-P. For these trials a decision was made to prepare the Poly SFH-P in a fashion that w o u l d allow one u n i t of Poly SFH-P to deliver the e q u i v a l e n t a m o u n t of h e m o g l o b i n contained in a one-unit blood transfusion. Therefore, each unit contains 500 ml at a 10 g/dL concentration, thereby delivering 50 g of hemoglobin. In addition, c o n t i n u e d improvement in the process enabled the Pso to be increased to 30 torr, thereby delivering 50 g of hemoglobin. The characteristics of one u n i t of Poly SFH-P used for clinical trials is shown in Table 7. To date the Phase I clinical trials have included b o t h h e a l t h y volunteers, 32 stable patients, and patients being resusTable 7. Characteristics of Poly SFH-P for Clinical Trials • H b = lOg/dL • Pso = 30 tort • Met Hb <3% • Tetramer <1% • 1 unit = 500ml (50g)
citated from hemorrhagic shock following major trauma. The Poly SFH-P was successfully infused in doses up to the equivalent of a one-unit transfusion of blood (50 g) in these recipients without the undesirable effects historically associated w i t h h e m o g l o b i n solutions, including vasoconstriction, kidney dysfunction, or gastrointestinal distress. We have r e c e n t l y been approved to begin Phase II trials in patients. The protocol will involve increasing the dose to three units {150 g) in patients suffering acute blood loss following major trauma and during major surgery. A l t h o u g h the work is still in progress, the preliminary results appear to confirm the safety and efficacy observations from the preclinical studies. The lack of observed toxicity to date suggests that the proposed mechanisms of toxicity discussed earlier in this paper, as well as the attempts to prevent these by polymerization, appear to be accurate. Further clinical testing will be required to fully assess the ultimate clinical utility of Poly SFH-P. REFERENCES 1. Dodd RY: The risk of transfusiontransmitted infection. N e w Engl 1 Med 1992; 327(61:419-421. 2. Gould SA, Rosen AL, Sehgal LR et al.: Red cell substitutes: Hemoglobin solution or fluorocarbon? J Trauma 1982; 22: 736. 3. Gould SA, Rosen AL, Sehgal LR et al.: How good are fluorocarbon emulsions as 02 carriers? Surg Forum 1981; 32:299. 4. Gould SA, Sehgal LR, Rosen AL et al.: Assessment of 35% Fluorocarbon emulsion. 1 Trauma 1983; 23(81:72-24. 5. Gould SA, Rosen AL, Sehgal LR et al.: Fluosol-DA as a red cell substitute in acute anemia. N e w Engl I Med 1986; 315:1653. 6. Winslow RM: Hemoglobin-based Red Cell Substitutes. Baltimore, Johns Hopkins University Press, 1992. 7. Amberson WR, Flexner l, Steggerda FR et al.: On the use of Ringer-Locke solution containing hemoglobin as a substitute for normal blood in mammals. ] Cell Comp Physiol 1934; 5:59-82. 8. Amberson WR, Jennings JJ, Rhode CM: Clinical experience with hemoglobinsaline solutions. J Appl Physiol 1949; 1:469-489.
Hemoglobin Solutions
9. Brand JL, Frank R, Lichtman HC: The effect of hemoglobin solutions on renal functions in man. Blood 1951; 6:11521158. 10. Miller JH, McDonald RK: The effect of hemoglobin on renal function in the human. J Clin Invest 1951; 30:10331040. 11. Savitsky JP, Doczi J, Black J et al.: A clinical safety trial of stroma-free hemoglobin. Clin Pharmacol Ther 1978; 23: 73-80. 12. Center for Biologics Evaluation and Research: Points to consider in the safety of hemoglobin-based oxygen carriers. Transfusion 1991; 31:369-371. 13. Keipert PE, Gonzales A, Gomez CL et al.: Acute changes in systemic blood pressure and urine output of conscious rats following exchange transfusion with diaspirin-crosslinked hemoglobin solution. Transfusion 1993; 33:701-708. 14. Martin W, Villani GM, Jothianandan D et al.: Selective blockade of endothelium-dependent and glycerol trinitrateinduced relaxation by hemoglobin and by methylene blue in the rabbit aorta. J Pharmacol Exp Ther 1985; 232:708-716. 15. Moncada S, Palmer RMJ, Higgs EA: Nitric Oxide: physiology, pathophysiology, and pharmacology. Pharmocol Rev 1991; 40:109-142. 16. Hess JR, MacDonald VW, Brinkley WW: Systemic and pulmonary hypertension after resuscitation with cell-flee hemoglobin, l Appl Physiol 1993; 74(4):17691778. 17. Keipert PE, Gomez CL, Gonzales A e t al.: Diaspirin-crosslinked hemoglobin: tissue distribution and long term excretion following exchange transfusion. J Lab Clin Med (in press). 18. Sehgal LR, Rosen AL, Gould SA et al.: Preparation and in vitro characteristics of polymerized pyridoxylated hemoglobin. Transfusion 1983; 23:148. 19. Sehgal LR, Gould SA, Rosen AL eta]. Polymerized pyridoxylated hemoglobin: a red cell substitute with normal 02 capacity. Surgery 1984; 95(4):433-438. 20. Gould SA, Sehgal LR, Sehgal HL et al.: Artificial blood, in Carlson RW, Geheb MA (eds): Principles & Practice of Medical Intensive Care. Philadelphia, WB Saunders Co, 1993. 21. Moss GS, Gould SA, Sehgal LR et al.:
22.
23.
24.
25.
26.
27.
28. 29.
30.
31.
32.
33.
34.
17
Hemoglobin solution: From tetramer to polymer. Surgery 1984; 95:249. Moss GS, DeWoskin R, Rosen AL et al.: Transport of oxygen and carbon dioxide by hemoglobin-saline solution in the red cell-flee primate. Surg Gynecol Obstet 1976; 142:357. Gould SA, Sehgal LR, Rosen AL et al.: Hemoglobin solution: Is a normal [Hb] or Pso more important? ] Surg Res 1982; 33:189. Benesch RE, Benesch R, Renthal RD et al. : Affinity labeling of the polyphosphate binding site of hemoglobin. Biochemistry 1972; 11:3576. Greenberg AG, Hayashi R, Siefert Iet al.: Intravascular persistence and oxygen delivery of pyridoxylated, stroma-free hemoglobin during gradations of hypotension. Surgery 1979; 86:13. Sehgal LR, Rosen AL, Noud G e t al.: Large volume preparation of pyridoxylated hemoglobin with high in vivo P5o. J Surg Res 1981; 30:14-20. Gould SA, Rosen AL, 8ehgal LR et al.: The effect of altered hemoglobin-oxygen affinity on oxygen transport by hemoglobin solution. J Surg Res 1980; 28:246. Gould SA, Sehgal LR, Rosen AL et al.: Polyhemoglobin: An improved red cell substitute. Surg Forum 1985; 36:30-32. Gould SA, Sehgal LR, Rosen AL et al.: The efficacy of polymerized pyridoxylated hemoglobin solution as an 02 carrier. Ann Surg 1990; 211:394-398. Rosen AL, Gould SA, Sehgal LR et al.: Effect of hemoglobin solution on compensation to anemia in erythrocyte-ffee primate. JAppl Physiol 1990; 68(3):938943. Rosen AL, Gould SA, Sehgal LR et al.: Evaluation of efficacy of stroma-free hemoglobin solutions, in Bolin RB, Geyer RP [eds), Blood Substitutes. New York, Alan Liss, 1983, p. 79. Gould SA, Sehgal LR, Sehgal HL et al.: Clinical experience with human polymerized hemoglobin. Transfusion (suppl) 1993; 33:60S. Gould SA, Moss GS, Rosen AL et al.: Red cell substitutes, in Civetta J, Taylor R, Kirby R [eds), Critical Care. Philadelphia, JB Lippincott, 1988, p. 1495. Moss GS, Sehgal LR, Gould SA et al.: Alternatives to transfusion therapy. Anethesiol Clin N A m 1990; 8:569.
0955-3886(94)
Transfus. Sci. Vol.16, No.I, pp. 5-17, 1995 Copyright © 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0955-3886/95 $9.50 + 0.00
The Development of Hemoglobin Solutions as Red Cell Substitutes: Hemoglobin Solutions Steven A. Gould, MD* Lakshman R. Sehgal, PhD Hansa L. Sehgal, BS Gerald S. Moss, MD
• Although the efficacy of hemoglobinbased oxygen carriers was established more than 60 years ago, all prior clinical trials have demonstrated significant toxicity characterized by renal dysfunction, gastrointestinal distress, and systemic vasoconstriction. The mechanisms of these toxicities now appear to be understood. Tetrameric forms of the hemoglobin molecule extravasate from the circulation and interact with endothelial derived relaxing factor, leading to unopposed vasoconstriction. Although numerous efforts are underway to chemically modify the native tetramer, it is likely that all tetrameric forms of the hemoglobin molecule will continue to extravasate. We have focused on developing a polymerized form of hemoglobin that is virtually free of unreacted tetramet. The development and characterization of this polymerized pyridoxylated hemoglobin solution (Poly SFH-P) is described. Clinical trials have been completed successfully in volunteers, and are now underway to assess the safety and efficacy of Poly SFH-P as a clinically useful red cell substitute in the treatment of acute blood loss in the setting of trauma and surgery. •
Department of Surgery, Michael Reese Hospital and Medical Center, Chicago University of Illinois, College of Medicine, Chicago Northfield Labs Inc., Evanston, IL ,60201, U.S.A. Author for correspondence: Steven A. Gould, Northfield Laboratories Inc., 1560 Sherman Avenue, Suite 1000, Evanston, IL 60201, U.S.A.
INTRODUCTION There continues to be great interest in developing a clinically useful 02 carrier to serve as a red cell substitute. The goal is to develop a safe and effective alternative to human blood. Although the current blood supply is safer than ever due to improved donor screening and testing, it is likely that disease transmission will always occur. ~ This is due to the inevitable occurrence of new viruses, and to the small but real incidence of false negative screening tests, often due to the window period of infectivity prior to conversion of the markers used for screening. In addition, the use of allogeneic blood will always involve the need for compatibility testing, and include a limited shelf life. The potential benefits of a red cell substitute include universal compatibility and the accompanying i m m e d i a t e availability, freedom from disease transmission, and long term storage. The two potential candidates to serve as clinically useful red cell substitutes include hemoglobin solutions and perfluorochemical emulsions? We have had considerable experience with the perfluorochemicals, including the clinical trial in patients with a 20% perfluorochemical emulsion, a-s The results showed particular preparation to be ineffective as an 02 carrier. The current status of the newer perfiuorochemical emulsions is reviewed elsewhere in this edition. The remainder of this paper will
6 Transfus. Sci. Vol. 16, No. 1
deal with the history and development of hemoglobin solutions.
tein and prepare it in a form that would be safe and useful in the clinical setting. 6
BACKGROUND
CONSIDERATIONS FOR A HEMOGLOBIN SOLUTION
The basic concept of developing a hemoglobin solution can be illustrated by examining the red blood cell {Fig. 1). The "active ingredient" in the red blood cell is the four-part hemoglobin molecule, the tetramer, that chemically binds and actually carries the oxygen. The hemoglobin molecule has several important characteristics. For example, 1 g of h e m o g l o b i n binds 1.39 mL of oxygen and is almost fully saturated with oxygen at ambient pressure. Few, if any, biologically acceptable substances have a greater oxygen-binding capacity. Oxygen is normally unloaded from hemoglobin in the capillaries at a pO2 of approx. 40 torr, allowing oxygen molecules to diffuse from hemoglobin to the intracellular mitochondria without producing interstitial hypoxia. The physiologic capability of the hemoglobin molecule is dear. The surrounding cell m e m b r a n e accounts for the need for compatibility testing and the limited shelf life. It has long been known that although the cell itself has a finite lifespan, the hemoglobin protein itself is rather durable and will survive and function, outside of the cell. The history of the development of hemoglobin solutions is the tale of the efforts to harvest the hemoglobin pro-
RED BLOOD CELL
Cell ane Mem~br
Hemogl MoleCmUl ) obiene
There are several properties that must be considered when describing the nature of a hemoglobin solution. The first is the source of the hemoglobin to be used as the starting material (Table 1). We have always felt that the human source is the best known and readily available, making it the preferred choice. The bovine source has been proposed as a way of providing easier access to an urdimited supply. The recombinant and transgenic approaches have received considerable interest because of their use of the m o d e r n molecular biology approaches. However, there is no evidence to suggest any physiologic or functional benefit of using any source other than human hemoglobin. In fact, the choice of source has not been a crucial issue in the development of a red cell substitute. The limiting factor in this field has been the safety of the hemoglobin solution. It is now evident that unmodified tetrameric h e m o g l o b i n is unsafe because of the associated toxicities of renal dysfunction, gastrointestinal distress, and vasoconstriction. 6 Therefore, the most important property of a hemoglobin solution is the nature of the hemoglobin preparation itself. Unmodified tetramer dissociates into dimers when removed from the red cell (Fig. 2). It is now believed that the observed toxicities are due in large part to this dissociation, an the presence of free tetramer in the circulation. Efforts to eliminate these toxicities have consisted of a variety of attempts to stabilize the simple unmodified tetramer. Table 2 lists the various preparations. Each effort Table 1. Hemoglobin Source •
Figure 1. A schematic representation of the red blood cell containing the hemoglobin tetramer.
Human
• Bovine • Recombinant •
Transgenic
Hemoglobin Solutions
UNMODIFIED TETRAMER
Tetramer
CROSSLINKED TETRAMER
Dimers
Figure 2. Representation of unmodified tetramer as it dissociates into dimers.
Figure 4. Cross-linked tetramer.
Table 2. Hemoglobin Preparation • Unmodified tetramer • Conjugated tetramer • Cross-linked tetramer • Polymer • Encapsulation involves the use of a different approach to try and stabilize the t e t r a m e r and t h e r e b y avoid toxicity. A conjugated tetramer [Fig. 3} involves the binding of a macromolecule such as polyethylene glycol (or PEG) to form a larger molecule. A cross-linked tetramer {Fig. 4} is an intramolecular chemical link to stabilize the native hemoglobin molecule. Polymerization involves intermolecular crosslinking of tetramers as s h o w n in Fig. 5. The polymerization results in a variety of m o l e c u l a r w e i g h t sizes. Finally, encapsulation as shown in Fig. 6, involves the " p a c k a g i n g " of the h e m o g l o b i n m o l e c u l e w i t h i n a lipid
CONJUGATED TETRAMER"
l
7
~-~;=~Macromolecule ~ (e.g.PEG)
Figure 3. Conjugated tetramer.
POLYMER
Figure 5. Polymers composed of 2, 3, and 4 tetramers linked together.
ENCAPSULATION
~
Liposome~
. . . . IODIR
Figure 6. Liposome encapsulated hemoglobin. membrane or liposome to form a "synthetic" red blood cell. Only testing in h u m a n s will e v e n t u a l l y resolve the issue of safety with these various preparations. However, it is clear that the nature of the actual hemoglobin preparation itself is far more important than the choice of source material.
8 Transfus. Sci.
Vol. 16, No. 1
The third characteristic of any h e m o g l o b i n s o l u t i o n is the i n t e n d e d clinical use. Table 3 lists some of the various potential applications for a safe and effective h e m o g l o b i n solution. Although they represent a diverse set of circumstances, it is likely that the major role of a hemoglobin solution will be as a red cell s u b s t i t u t e for use in acute blood loss. Table 3. Hemoglobin Clinical Use
• • • • • •
Acute blood loss Hemodilution Ischemia Cardioplegia Organ preservation Cancer therapy
There are numerous clinical trials u n d e r w a y in b o t h v o l u n t e e r s and patients using a variety of these approaches. U l t i m a t e l y , each effort involves a selection of h e m o g l o b i n source, preparation, and intended clinical use. Table 4 s h o w s a list of the source and preparation for each of the current approaches in clinical trials. Table 4. Current Approaches
• • • •
Human polymer, tetramer-free Bovine polymer, tetramer present Human cross-linked tetramer Recombinant cross-linked tetramer
DEMONSTRATION OF EFFICACY
The concept of using a hemoglobin solution as an oxygen carrier was first fully tested by A m b e r s o n e t al. in 1934. 7 A m b e r s o n prepared a crude red cell lysate, which was an unmodified tetrameric form of the hemoglobin solution. In a rather remarkable experiment, he gradually removed all of the blood from cats and replaced it with this primitive hemoglobin solution. The results demonstrated the proof of concept and paved the way for all future efforts. The cats survived on a short-term basis, and were able to walk normally. In an absolutely elegant maneuver, Amberson evaluated neurologic function by holding the cats
on their backs and dropping them. The cats were able to land upright, demonstrating a fully intact nervous s y s t e m that is required in order to perform this c o m p l i c a t e d neurologic event. The p o t e n t i a l u t i l i t y of h e m o g l o b i n solutions was firmly established. THE PROBLEM OF TOXICITY
Based on this early d e m o n s t r a t i o n by Amberson, the efficacy of a hemoglobin solution was uncontestable. However, 60 years later there is still no h e m o globin s o l u t i o n in clinical use. The reason for this relates to the results in clinical trials. 8-1' These trials have consistently demonstrated major toxicities: renal dysfunction, gastrointestinal distress, and s y s t e m i c v a s o c o n s t r i c t i o n characterized by a rise in blood pressure and a fall in heart rate. '2 The Savitsky et al. s t u d y in 1978 '1 clearly e s t a b l i s h e d the futility of further efforts to develop a safe t e t r a m e r i c form of h e m o g l o b i n solution. MECHANISMS OF TOXICITY
The mechanisms of toxicity of the tetramer n o w s e e m to be reasonably well understood. In general, these toxicities are related to the presence of the tetramer outside of the red cell and the dissociation into dimers. The renal dysfunction is thought to be related to filtration and excretion of the hemoglobin molecule, primarily in the form of dimers, although vasoconstriction is also likely to play a role. The vasoconstriction and other adverse effects are c u r r e n t l y thought to be related to extravasation of the tetrameric hemoglobin molecule. 13 Our understanding of the control of v a s o c o n s t r i c t i o n of blood vessels has been improved with the identification of e n d o t h e l i u m derived relaxing factor (EDRF), also known as nitric oxide. 14''s It is well known that EDRF (nitric oxide) avidly binds to hemoglobin. Figure 7 is a graphic representation of the control of vasoconstriction. It is a cross-section of a blood vessel representing the lumen,
Hemoglobin Solutions 9
CONTROL OF VASOCONSTRICTION Muscle Cells
Tetramer (T) - Extravasates and Binds EDRF Both Sides
!
Casoconstriction
L .......
Cells
Figure 7. Control of vasoconstriction. the endothelial cell layer, and the interstitial space within which reside the vascular smooth muscle cells. EDRF is produced by the endothelial cells and is released on both the luminal and the abluminal sides. The physiologic activity of EDRF occurs on the vascular smooth muscle cells in the interstitial space. The EDRF released into the l u m e n is rapidly bound by the hemoglobin within the circulating red cells, and does not act on the smooth muscle cells. It is also evident that red cells do not normally move beyond the endothelial cell lining into the interstitial space, laving the EDRF available to exert its vasorelaxant effect. The vasoconstriction that occurs with the tetrameric form of the hemoglobin is considered to be due to extravasation of the hemoglobin beyond the endothelial cell barrier, and the immediate binding of the EDRF to the hemoglobin, leading to unopposed vasoconstriction. This is illustrated in Fig. 8. The presence of the tetramer in the interstitial space where EDRF normally acts thereby disrupts the normal control of vascular relaxation. Our working hypothesis has been that prevention of extravasation will prevent vasoconstriction, since EDRF would function normally. It has also been our contention that all forms of tetramer will extravasate and produce both vasoconstriction and the other toxicities historically associated with hemoglobin solutions. 13'16'17 The only forms of hemoglobin that
Figure 8. The effect of tetramer on extravasation and binding of EDRF. would appear to prevent extravasation include the polymer and encapsulation. The cross-linked tetramers have indeed been shown to extravasate, t3'~7 Since the encapsulation approach is still in the developmental stage, only the polymer has been able to be evaluated in the clinical setting. Figure 9 illustrates the likely scenario when the polymer is infused. As with the red cell, the polymer will exist only in the lumen, leaving the EDRF to exert its normal physiologic vasorelaxant role on the vascular smooth muscle cell in the interstitial space. Our approach, therefore, has been to specifically design a hemoglobin solution using a two step process to create a tetramer-free form of polymerized hemoglobin. ~8-2t The first step is the polymerization, which results in an array of different size polymers. The
P o l y m e r (P)
-
Binds EDRF in Lumen But No Extravasation
!
N0 Vasoconstriction
Figure 9. Representation of polymer and absence of extravasation.
10 Trans[us. ScL Vol. 16, No. 1
POLYMERIZATION AND PURIFICATION 0 0 0 0 oooo
61utaraldehyde
o0%%00
0 0 00(~3
&8
°&88coS°°
o o Oo
Polymer
Figure 10. Preparation of Poly SFH-P, including polymerization and removal of unreacted tetramer. second step, which is just as important, is the removal of all unreacted tetramer. Figure 10 is a representation of this process. Glutaraldehyde is used as the cross-linking (polymerizing) agent. The second step is to remove virtually all unreacted tetramer, leading to a pure polymeric solution. The remainder of this paper will deal with the historical development of our human polymerized h e m o g l o b i n solution, including the physiologic observations that led to the characteristics of our current preparation. UNMODIFIED HEMOGLOBIN
The basic, unmodified hemoglobin solution is currently prepared from outdated blood, beginning with the washing and lysis of the red cells (RBCs} with pyrogen-free water. A series of filtration steps permits the complete separation of the RBC membrane debris (stroma} from the hemoglobin molecules. The resultant solution is referred to as stroma-free
h e m o g l o b i n (SFH}. Because the RBC antigens are located on the cell membrane, SFH is universally compatible and can be infused w i t h o u t regard to specific blood type. The properties of this unmodified, tetrameric or "stripped" h e m o g l o b i n solution are given in Table 5. Although SFH can be prepared with a hemoglobin concentration of 14 g/dL, this solution has a colloid osmotic pressure (COP} greater than 6 0 m m H g , which renders it unacceptable for clinical use. 22 The hemoglobin concentration of 7 g/dL is iso-oncotic. The low Pso is due to the loss of the organic ligand, 2,3-diphosphoglycerate {2,3-DPG}, during preparation. The [02] curve of SFH is thus both anemic { ~ hemoglobin content} and shifted to the left { ~ pso}. Despite these limitations, SFH supports life in primates in the absence of RBCs. 2~ Animals survive a total exchange transfusion with SFH to zero hematocrit with maintenance of normal 02 c o n s u m p t i o n (Vo2], cardiac output, and arteriovenous oxygen content difference (Cao2-Cvo2], although a decline from baseline values occurs in some of these measures. In addition, a considerable decrease occurs in the pf¢o2 from approx. 50 to 20 torr. The pfZo2 is the partial pressure at w h i c h oxygen unloads from the hemoglobin molecule and is in equilibrium with the tissue po2. This decline indicates a marked increase in oxygen extraction and is the mechanism used to compensate for the fall in hemoglobin content and pso. This low pf¢'o2 is of some concern and led us to a t t e m p t to restore a more normal value. 23
Table 5. Properties and Parameters of Stroma-ffee Hemoglobin and Whole Blood Properties and parameters Stroma-free hemoglobin Whole blood Hemoglobin content {g/dL} 6-8 12-14 Oxygen-carrying capacity (vol%) 8.0-11.0 16-19 Binding coefficient (mL O2/g Hbl 1.30 1.30 Ps0 {torr} 12-14 26-28 Methemoglobin {%} <2 <1 Colloid osmotic pressure (torr) 18-25 18-25 Osmolarity {mOsm] 290-310 290-310
Hemoglobin Solutions
PYRIDOXYLATED HEMOGLOBIN A shift to the left in the [Oz] curve, with no change in Vo2, cardiac output, or Cao2-Cvo2, produces a decrease in the pfdoz (Fig. 11). Attempts to establish a n o r m a l ps0 by the simple addition of 2,3-DPG to the h e m o g l o b i n solution were u n s u c c e s s f u l because the ligand disappears rapidly from the circulation after infusion. However, a modification of the hemoglobin molecule by the addition of pyridoxal phosphate results in a pyridoxylated hemoglobin (SFH-P) with a Pso of 20-22 tort, which is considerably higher t h a n the pso of u n m o d i f i e d SFH34-26 We evaluated SFH-P in eight baboonsY Four received SFH and four received SFH-P, with a final hemoglobin content of 7 g/dL and zero hematocrit. The p(Zoz levels were significantly higher at the end of the exchange in the animals receiving SFH-P. A l t h o u g h hemodynamic parameters were normal, a decline from the baseline values occurred in both groups. These results illustrate three points. First, a shift to the right in the dissociation curve (1' Pso} results in an increased p~Zo2because all else remained constant. This observation is of physiologic i m p o r t a n c e because oxygen unloading can occur at a higher tissue po~. Second, although it was increased, the pV¢o2 level in the animals treated
11
with SFH-P was still substantially lower t h a n the n o r m a l value of 40-50 torr found in control animals. Third, hemodynamic function still showed a reduction from the baseline values. It became apparent that a nonanemic hemoglobin solution was requiredY NONANEMIC ISO-ONCOTIC HEMOGLOBIN SOLUTION The advantages of a nonanemic hemoglobin solution are self-evident. Such a solution w o u l d have the same oxygen capacity as w h o l e blood. In addition, according to our data, the infusion of a n o n a n e m i c solution should be associated with normal p~/o2 levels, even at zero h e m a t o c r i t ? 3 The principal obstacle to n o r m a l i z a t i o n of h e m o g l o b i n concentration is the effect of an elevation in protein concentration on oncotic pressure. The relationship b e t w e e n h e m o globin concentration and oncotic pressure is shown in Fig. 12. At hemoglobin c o n c e n t r a t i o n of 7 g/dL, the oncotic pressure is similar to that of plasma (20 torr). In contrast, at hemoglobin levels of 15 g/dL, oncotic pressure increases by more than 300%. The infusion of such a solution would theoretically produce large shifts of fluid from the extravascular space into the i n t r a v a s c u l a r space. These changes are likely to be harmful. One approach to producing a non-
| Pso SFH-P
02
Content
COP (torr)
4O 2O
P02 Figure 11. [02] curves showing a leftward shift in P~owith a constant Cao2 - Cvo2 leads to a lower pVo2. Curve B is shifted to the left compared with curve A. (From Ref. 33.)
0
4
8 12 [Hb] gm %
t
18
Figure 12. Relationship between COP and hemoglobin concentration for SFH-P. (From Ref. 34.)
12
Transfus. Sci. Vol. 16, No. 1
anemic hemoglobin solution w i t h normal COP values is polymerization of the hemoglobin. The COP of any solution is proportional to the n u m b e r of colloidal particles. If a 15 g/dL solution of hemoglobin could be polymerized, the result would be a reduction in COP, whereas no change would occur in hemoglobin concentration (Fig. 13). This idea was tested in our laboratories.IS The hemoglobin solution obtained was pyridoxylated by modification of previously described techniques. Briefly, pyridoxal-5'-phosphate was added to the solution at a 4:1 molar ratio. The solution is transferred to a sealed reservoir and deoxygenated with nitrogen (N2) by using a gas exchanger. Deoxygenation of the solution is continued until the po2 was below 10 torr and the [02] was below 1 vol%. Sodium borohydride was t h e n added to the solution, and the reaction c o n t i n u e d for 2--4 h. The goal of the polymerization is to normalize the oxygen capacity while m a i n t a i n i n g the COP w i t h i n normal limits (20-25 tort). Several parameters can be manipulated to obtain, in a reproducible manner, the polymerized solution of choice. 19 These are the hemoglobin concentration, the concentration of the polymerizing agent, the rate of addition of the latter, and the rate of mixing. Finally, the duration for which the above-mentioned parameters are controlled will in t u r n d e t e r m i n e the reaction yield, molecular weight distri-
bution of the polymeric species, oxygen affinity (pso), and finally the viscosity of the solution. The polymerization is conducted to a physiologic end point with respect to COP. The reaction is thus monitored by the decrease in COP. Figure 14 shows the changes in COP monitored during the polymerization process conducted over a 3-h period. As can be seen from the figure, the reproducibility of the process for five batches is acceptable. The polymerization and evolution of the polymeric species are also followed by s o d i u m dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE). Figure 15 shows the formation of new polypeptide bands during the process. Typically, four to six bands are seen, ranging in molecular weights from 16,000 to 96,000. The first four bands (16,000-64,000 Da) represent greater than 90% of all species. With the formation of polymers, one would expect changes in the viscosity of the solution. Figure 16 shows the changes in viscosity observed for five different batches. The viscosity at the start of polymerization ranges from 1.0 to 1.5 cp and increases to reach a range of 1.9-2.1 cp at the end. This represents a relatively modest increase but is consistent with the absence of high-molecular-weight species in the solution, as d e m o n s t r a t e d by SDSPAGE. The changes in ps0 during the poly-
POLYSFH-P 0 000 O0 O0
%%%00
O oo%
@
~P
(torr)
O 0 0 00 0 0 1 2 Polymerization Time (Hrs)
3
[Hb] - 15 gm/dL [Hb] - 15 gm/dL COP > 70 torr =-,- COP = 25 torr
Figure 13. Polymerization. (From Ref. 34.)
Figure 14. Decrease in COP measured during the course of polymerization for five different batches. (From Ref. 34.)
Hemoglobin Solutions 13
Figure 15. Evolution of new polypeptide bands over the course of the polymerization process, as monitored by SDS-PAGE. (From Ref. 34.)
POLYSFH-P
POLYSFH-P Viscosity20 (cP)
Pso 24 (torr) 2O Y
1.0
~ 1
~
,
2
3
Polymerization Time (Hrs)
/
i 1
I 2
I 3
PolymerizationTime (Hrs)
Figure 16. Changes in viscosity of five batches of hemoglobin solution during polymerization. (From Ref. 34.)
Figure 17. Alteration in Pso for five different batches of hemoglobin solution during polymerization. (From Ref. 34.)
merization were followed and are shown in Fig. 17. A modest increase in Pso was n o t e d in all five b a t c h e s shown. This increase reflects a decrease in buffering capacity of the solution with polymerization. It is consistent with the observation that a drop in pH is observed during the course of the polymerization. This is not surprising because it is some of the same amino groups that are involved
in the buffering as well as the cross-linking. The characteristics of the final product are listed in Table 6. The solution at 12-15 g/dL has a COP between 20 and 25 torr. The m o l e c u l a r w e i g h t range based on SDS-PAGE is between 16,000 and 96,000 Da. The n u m b e r average m o l e c u l a r w e i g h t calculated from the COP is approx. 150,000 Da. The average
IS 16:1-§
14
Transfus. Sci.
Vol.16, No. 1
Table 6. Properties of Polymerized Pyridoxyl-
ated Hemoglobin Property
S T A B I L I T Y OF P O L Y S F H - P Parameter range
Hemoglobin (g/dL) 12-14 Oxygen-carrying capacity 16-19 (vol%) Methemoglobin (%) <5 Molecular weight range 64,000-400,000 Number average molecular 150,000 weight Pso (tort) 18-22 Binding coefficient (mL O2/ 1.30 g Hb) COP (tort) 20-25 molecular weight obtained by ultracentrifugation is 120,000. The Ps0 of the solution under standard conditions {pH = 7.40, pco2 = 35 tort, temperature = 35°C) ranges from 18 to 22 tort. The viscosity of the solution {hemoglobin concentration = 8.0 g/dL} ranges from 1.9 to 2.2 cp. The phospholipid content is below the detection limits of the thinlayer chromatography technique used [<0.4 mg/dL). We have followed the stability of the polymerized solution when stored at 4-8°C as a sterile, pyrogen-free product. The stability with respect to rate of conversion to methemoglobin is shown in Fig. 18. the m e t h e m o g l o b i n remains below 2% for a 150-day period. The stability of the polymeric species was determined by monitoring the
S T A B I L I T Y OF P O L Y S F H - P (Met Hb)4"0I
% z
o
~ Time(Days)
Figure 18. Changes in methemoglobin concentration on storage at 4-8°C, as a sterile, pyrogen-free solution. Each point represents the mean + SEM. {From Ref. 34.]
T-150DaysA Batch- !
k Batch-2
Batch-3
Figure 19. High-pressure liquid chromatography scans for three batches at the time of manufacture and 150 days later. No changes are noted. {From Ref. 34.1
solution with high-pressure liquid chromatography, SDS-PAGE, and viscosity. No changes were observed for at least a 1-year period. Figure 19 shows the highpressure liquid chromatographic scans for three different batches at the time of manufacture and 150 days later. No visible changes occurred. Finally, the changes in viscosity during a 150-day storage period were monitored. Once again, no change in viscosity was observed. Thus the polymerized stromafree h e m o g l o b i n (Poly SFH-P), w h e n stored at 4-8°C, appears to be quite stable. EFFICACY OF POLY SFH-P
Seven adult baboons were anesthetized, paralyzed, intubated, and mechanically ventilated with room airY '29 The respiratory rate and tidal v o l u m e were adjusted to maintain a pco~ between 35 and 45 tort before the start of the study and were not changed during the study. The animals were surgically prepared with arterial and central venous catheters for infusion, blood sampling, and monitoring. A thermal dilution balloontipped catheter was floated into the pulmonary artery. A Foley catheter was inserted into the urinary bladder. Standard hemodynamic monitoring was performed for electrocardiogram, arterial pressures, p u l m o n a r y capillary wedge pressure, and central venous pressure.
Hemoglobin Solutions
Cardiac output was determined by the thermal dilution method. The study was conducted with the use of ketamine anesthesia. After stabilization of the animals, a set of baseline measurements was obtained. An isovolemic exchange transfusion with the Poly SFH-P was then performed. Whole blood was removed in 50mL aliquots and was replaced with approximately equal volumes of the infusate. Additional volume adjustments were made as required to maintain the pulmonary capillary wedge pressure at baseline values. The exchange was stopped at hematocrits of 20, 10, and 5% to obtain additional sets of measurements. The exchange transfusion was then carried out to obtain a complete washout of the RBCs. A hematocrit of less than 1% was achieved. These animals, at zero hematocrit, had a Poly SFH-P concentration of approx. 10g/dL. They were then exchanged transfused with dextran 70 to a hemoglobin concentration of 1 g/dL. 3° The data from the second half of the study were compared with a control group {n=6) that underwent an exchange transfusion with dextran 70 to a hemoglobin concentration of 1 g/dL. Arterial and mixed venous blood gases were measured by standard electrodes {IL 813}. Whole blood and plasma hemoglobin levels were determined on the IL 282 Co-oximeter. Oxygen carried by whole blood hemoglobin was measured by the IL 282 Co-oximeter. The physically dissolved oxygen in the aqueous phase was calculated from the po2 of the plasma separated by centrifugation. The Poly SFH-P [02] was measured by the IL 282 Co-oximeter. Hematocrits were determined by the microhematocrit method. Ps0 values of whole blood and plasma hemoglobin were determined with the hemoglobin oxygen dissociation curve analyzer (Hemo-O-Scan}. The efficacy of the Poly SFH-P was calculated as described previously.3~ At each hematocrit level, the arterial [02] {[O2]a}was determined for each compart-
15
ment by direct measurement or calculation. Total Vo2 was calculated as the product of the cardiac output and Cao2 Cvo2. The contribution of Poly SFH-P to oxygen delivery was calculated as the ratio of the Poly SFH-P to total arterial
[02}. Poly SFH-P 02 delivery = [02], Poly SFH-P [O2]a, total The contribution of Poly SFH-P to Vo2 was calculated as the ratio of Poly SFH-P to total Cao2 - Cvo2. Poly SFH-P Vo2 = Poly SFH-P {Cao2- Cvo2} Total {Cao2- Cvo2} All contributions were expressed as their percent values. All animals receiving Poly SFH-P survived the exchange transfusion, as did the previous animals receiving SFH-P. The final hematocrit was 0.8 + 0.4% {mean + SEMJ. The difference in the initial bag pso values of the two infusates was statistically significant (P<0.05). However, the mean in vivo plasma Pso for the Poly SFH-P was 17.0 _+ 0.5 torr, which was not significantly different from the mean value of 17.6 + 0.8 torr for the SFH-P. Both of these plasma Ps0 values are significantly below the mean baboon RBC Pso of 31.3 + 0.8 torr. The Poly SFH-P [O2]a, is significantly greater than the SFH-P value at all hematocrits {P<0.001). At a hematocrit of 5%, the [O2}a was 9.5 + 0.2 vol% for Poly SFH-P and 5.0 + 0.4 vol% for SFH-P. The percent contributions of Poly SFH-P to total oxygen delivery and total VO2 were compared with those of SFH-P. Poly SFH-P makes a significantly greater {P<0.2} contribution to total oxygen delivery than SFH-P at all hematocrits. The contribution to total Vo2 is greater by Poly SFH-P at all hematocrits, with the difference significant (P<0.005J at a hematocrit of 20%. These results document that Poly SFH-P is an effective 02
16 Trans[us. Sci. Vol. 16, No. 1 carrier, and provides benefit over the tetrameric form of SFH-P. The in vivo Pso in the Poly SFH-P animals undergoing the second exchange transfusion w i t h dextran 70 ranged from 18 to 11 torr. In contrast, the in v i v o pso of the control group ranged from 31.5 to 25.5 tort. The pVo2 was significantly lower in the Poly SFH-P group compared with the control group. Both groups of animals raised their cardiac o u t p u t in an identical manner in response to their anemia. The critical oxygen delivery in the control group was 6.6 m L / m i n / k g of body weight compared w i t h 5.7 m L / m i n / k g in the test group. These results indicate that the infusion of Poly SFH-P does not alter the normal physiologic response to progressive anemia, and therefore provides evidence of lack of vasoconstriction. CLINICAL TRIALS Based on these preclinical observations we have begun clinical trials in b o t h h e a l t h y v o l u n t e e r s and p a t i e n t s to assess the safety and efficacy of Poly SFH-P. For these trials a decision was made to prepare the Poly SFH-P in a fashion that w o u l d allow one u n i t of Poly SFH-P to deliver the e q u i v a l e n t a m o u n t of h e m o g l o b i n contained in a one-unit blood transfusion. Therefore, each unit contains 500 ml at a 10 g/dL concentration, thereby delivering 50 g of hemoglobin. In addition, c o n t i n u e d improvement in the process enabled the Pso to be increased to 30 torr, thereby delivering 50 g of hemoglobin. The characteristics of one u n i t of Poly SFH-P used for clinical trials is shown in Table 7. To date the Phase I clinical trials have included b o t h h e a l t h y volunteers, 32 stable patients, and patients being resusTable 7. Characteristics of Poly SFH-P for Clinical Trials • H b = lOg/dL • Pso = 30 tort • Met Hb <3% • Tetramer <1% • 1 unit = 500ml (50g)
citated from hemorrhagic shock following major trauma. The Poly SFH-P was successfully infused in doses up to the equivalent of a one-unit transfusion of blood (50 g) in these recipients without the undesirable effects historically associated w i t h h e m o g l o b i n solutions, including vasoconstriction, kidney dysfunction, or gastrointestinal distress. We have r e c e n t l y been approved to begin Phase II trials in patients. The protocol will involve increasing the dose to three units {150 g) in patients suffering acute blood loss following major trauma and during major surgery. A l t h o u g h the work is still in progress, the preliminary results appear to confirm the safety and efficacy observations from the preclinical studies. The lack of observed toxicity to date suggests that the proposed mechanisms of toxicity discussed earlier in this paper, as well as the attempts to prevent these by polymerization, appear to be accurate. Further clinical testing will be required to fully assess the ultimate clinical utility of Poly SFH-P. REFERENCES 1. Dodd RY: The risk of transfusiontransmitted infection. N e w Engl 1 Med 1992; 327(61:419-421. 2. Gould SA, Rosen AL, Sehgal LR et al.: Red cell substitutes: Hemoglobin solution or fluorocarbon? J Trauma 1982; 22: 736. 3. Gould SA, Rosen AL, Sehgal LR et al.: How good are fluorocarbon emulsions as 02 carriers? Surg Forum 1981; 32:299. 4. Gould SA, Sehgal LR, Rosen AL et al.: Assessment of 35% Fluorocarbon emulsion. 1 Trauma 1983; 23(81:72-24. 5. Gould SA, Rosen AL, Sehgal LR et al.: Fluosol-DA as a red cell substitute in acute anemia. N e w Engl I Med 1986; 315:1653. 6. Winslow RM: Hemoglobin-based Red Cell Substitutes. Baltimore, Johns Hopkins University Press, 1992. 7. Amberson WR, Flexner l, Steggerda FR et al.: On the use of Ringer-Locke solution containing hemoglobin as a substitute for normal blood in mammals. ] Cell Comp Physiol 1934; 5:59-82. 8. Amberson WR, Jennings JJ, Rhode CM: Clinical experience with hemoglobinsaline solutions. J Appl Physiol 1949; 1:469-489.
Hemoglobin Solutions
9. Brand JL, Frank R, Lichtman HC: The effect of hemoglobin solutions on renal functions in man. Blood 1951; 6:11521158. 10. Miller JH, McDonald RK: The effect of hemoglobin on renal function in the human. J Clin Invest 1951; 30:10331040. 11. Savitsky JP, Doczi J, Black J et al.: A clinical safety trial of stroma-free hemoglobin. Clin Pharmacol Ther 1978; 23: 73-80. 12. Center for Biologics Evaluation and Research: Points to consider in the safety of hemoglobin-based oxygen carriers. Transfusion 1991; 31:369-371. 13. Keipert PE, Gonzales A, Gomez CL et al.: Acute changes in systemic blood pressure and urine output of conscious rats following exchange transfusion with diaspirin-crosslinked hemoglobin solution. Transfusion 1993; 33:701-708. 14. Martin W, Villani GM, Jothianandan D et al.: Selective blockade of endothelium-dependent and glycerol trinitrateinduced relaxation by hemoglobin and by methylene blue in the rabbit aorta. J Pharmacol Exp Ther 1985; 232:708-716. 15. Moncada S, Palmer RMJ, Higgs EA: Nitric Oxide: physiology, pathophysiology, and pharmacology. Pharmocol Rev 1991; 40:109-142. 16. Hess JR, MacDonald VW, Brinkley WW: Systemic and pulmonary hypertension after resuscitation with cell-flee hemoglobin, l Appl Physiol 1993; 74(4):17691778. 17. Keipert PE, Gomez CL, Gonzales A e t al.: Diaspirin-crosslinked hemoglobin: tissue distribution and long term excretion following exchange transfusion. J Lab Clin Med (in press). 18. Sehgal LR, Rosen AL, Gould SA et al.: Preparation and in vitro characteristics of polymerized pyridoxylated hemoglobin. Transfusion 1983; 23:148. 19. Sehgal LR, Gould SA, Rosen AL eta]. Polymerized pyridoxylated hemoglobin: a red cell substitute with normal 02 capacity. Surgery 1984; 95(4):433-438. 20. Gould SA, Sehgal LR, Sehgal HL et al.: Artificial blood, in Carlson RW, Geheb MA (eds): Principles & Practice of Medical Intensive Care. Philadelphia, WB Saunders Co, 1993. 21. Moss GS, Gould SA, Sehgal LR et al.:
22.
23.
24.
25.
26.
27.
28. 29.
30.
31.
32.
33.
34.
17
Hemoglobin solution: From tetramer to polymer. Surgery 1984; 95:249. Moss GS, DeWoskin R, Rosen AL et al.: Transport of oxygen and carbon dioxide by hemoglobin-saline solution in the red cell-flee primate. Surg Gynecol Obstet 1976; 142:357. Gould SA, Sehgal LR, Rosen AL et al.: Hemoglobin solution: Is a normal [Hb] or Pso more important? ] Surg Res 1982; 33:189. Benesch RE, Benesch R, Renthal RD et al. : Affinity labeling of the polyphosphate binding site of hemoglobin. Biochemistry 1972; 11:3576. Greenberg AG, Hayashi R, Siefert Iet al.: Intravascular persistence and oxygen delivery of pyridoxylated, stroma-free hemoglobin during gradations of hypotension. Surgery 1979; 86:13. Sehgal LR, Rosen AL, Noud G e t al.: Large volume preparation of pyridoxylated hemoglobin with high in vivo P5o. J Surg Res 1981; 30:14-20. Gould SA, Rosen AL, 8ehgal LR et al.: The effect of altered hemoglobin-oxygen affinity on oxygen transport by hemoglobin solution. J Surg Res 1980; 28:246. Gould SA, Sehgal LR, Rosen AL et al.: Polyhemoglobin: An improved red cell substitute. Surg Forum 1985; 36:30-32. Gould SA, Sehgal LR, Rosen AL et al.: The efficacy of polymerized pyridoxylated hemoglobin solution as an 02 carrier. Ann Surg 1990; 211:394-398. Rosen AL, Gould SA, Sehgal LR et al.: Effect of hemoglobin solution on compensation to anemia in erythrocyte-ffee primate. JAppl Physiol 1990; 68(3):938943. Rosen AL, Gould SA, Sehgal LR et al.: Evaluation of efficacy of stroma-free hemoglobin solutions, in Bolin RB, Geyer RP [eds), Blood Substitutes. New York, Alan Liss, 1983, p. 79. Gould SA, Sehgal LR, Sehgal HL et al.: Clinical experience with human polymerized hemoglobin. Transfusion (suppl) 1993; 33:60S. Gould SA, Moss GS, Rosen AL et al.: Red cell substitutes, in Civetta J, Taylor R, Kirby R [eds), Critical Care. Philadelphia, JB Lippincott, 1988, p. 1495. Moss GS, Sehgal LR, Gould SA et al.: Alternatives to transfusion therapy. Anethesiol Clin N A m 1990; 8:569.