Hemoglobin-based oxygen carriers (HBOCs) containing stroma-free, ultra-purified bovine hemoglobin (Hb) were developed to overcome challenges associated with the transfusion of allogenic blood, such as transmission of diseases, immunologic incompatibility, transportation, storage difficulties, short shelf life, and supply shortages (1). Hemoglobin-glutamer-200 (HBOC-200) (Oxyglobin®, Biopure, Cambridge, MA) was the first HBOC approved by the United States Federal Drug Administration for treatment of anemia in dogs. HBOCs are also considered good candidates for abuse by human and equine athletes based on proposed capability to increase oxygen delivery to tissues (2,3).
HBOC-200 share properties with HBOC-250, which is currently under clinical investigation in human patients and has been recently approved for use in humans in South Africa. In both these solutions, the bovine stroma-free Hb is highly purified and subsequently reacted with glutaraldehyde as an intra- and intermolecular cross-linker that targets surface amino groups and links adjacent molecules. As a result, Hb polymers of various sizes are formed. The reaction mixture is subjected to further fractionation to reduce the number of unreacted Hb and potentially toxic tetramers and dimers (4). The molecular weight (MW) of ∼50% of Hb aggregates in HBOC-200 are between 65 and 130 kD, and only ∼10% are >500 kD in size (Robert A. Houtchens, PhD, Biopure Corp., Cambridge, MA, 2004, personal communication). Significant heterogeneity in aggregate size is a common feature of this first generation HBOC and should affect the pharmacokinetic profiles of HBOC solutions.
The disposition of this protein-based drug is not species dependent. All mammalian species dispose of Hb using similar pathways, and there is no unique metabolic process in the horse that would not allow a generalized interpretation of this model. The specific aim of this study was to determine the pharmacokinetic profile of HBOCs after an IV administration and model the elimination from circulation based on aggregate size.
Seven thoroughbred horses, one male and six female, were used in this study, with a median age and weight of 6 yr (4–11 yr) and 541 kg (427–623 kg). Horses were maintained on pasture and brought into stalls 2 days before the experiment. Each experiment started at 7 am, and all horses remained housed for the duration of the study. They were fed grass hay and water ad libitum. All horses in the study were no longer actively racing but were otherwise in good health, dewormed, and vaccinated. The University of Pennsylvania Institutional Animal Care and Use Committee approved the study protocol.
Catheters (14F; Angiocath, Becton Dickinson, Sandy, UT) were placed in both jugular veins for blood collection and HBOC-200 administration. Before placement, the areas over the veins were clipped, washed with surgical soap (Chlorhexidine®), and rinsed with a viricide (Nolvasan®) and 70% isopropyl alcohol. Both areas of catheter placement were infiltrated with 1.5 mL of 2% solution of a local anesthetic (Lidocaine®). The 250 mL of HBOC-200 solution (13 g/dL) was administered in 5 min, timed with a stop watch using an IV drip system.
Control blood samples were collected from the contralateral vein before IV infusion and at 1, 2, and 5 min during the infusion. After the 5-min infusion period, blood samples were collected at 1, 5, 15, 30, and 45 min and at 1, 1.5, 2, 4, 6, 8, 10, 12, 16, 20, 24, 36, 48, and 72 h. Heparin was used as anticoagulant. Blood samples were centrifuged (2500 ×g for 15 min) to obtain plasma. Aliquots of plasma were stored at 4°C and at −70°C as backup samples. Samples were analyzed within 24 h of collection.
Quantification of HBOC-200 in equine plasma was accomplished by methods previously described by this laboratory (5,6). Briefly, HBOC-200 was recovered from plasma by solid-phase extraction. The extracted HBOC-200 in elution solvent was dried at 80°C under a stream of nitrogen. The dried HBOC-200 was digested with a proteolytic enzyme, trypsin, to yield tryptic peptides for analysis by liquid chromatography coupled with mass spectrometry. The proteolytic digestion was required to specifically differentiate bovine Hb from native equine Hb. The tryptic peptides were separated by liquid chromatography on a Zorbax 300 SB-C8 column (2.1 × 50 mm, 3.5 μm; Agilent, Wilmington, DE), and a unique peptide specific for HBOC-200 was identified and quantified. Limit of quantification in plasma was 50 μg/mL. The intra-day and inter-day coefficients of variation were <17% for quantification of HBOC-200 in equine plasma. Standard operating procedure for the quantification of the analyte by this laboratory meets requirements for accreditation by the American Association for Laboratory Accreditation and ISO 17025 International Guidelines.
Plasma concentration versus time curves after IV administration of HBOC-200 were analyzed by standard linear compartmental analysis (WinSaam; www.WinSaam.com; Kennett Square, PA) (7). The use of the classic interchanging compartmental or exponential model was not considered appropriate for this protein-based drug because the drug is known to be heterogeneous in that it consists of small and large polymers. A noninterchanging compartmental model was developed to define the pharmacokinetics and estimate the elimination of infused HBOC-200 based on the size distribution of the bovine Hb aggregates (Fig. 1). The number of compartments required to define the pharmacokinetics was based on appearance of the curve, the reduction in the sums of squares, and the minimization of fractional standard deviation (FSD) of all the variables.
In developing a compartmental model to describe the pharmacokinetics of HBOC-200, it was assumed, based on the reported composition of bovine Hb, that the administered HBOC-200 was composed of a range of aggregate sizes. Large aggregates would be confined to plasma, designated as compartment-2 (C2). The smaller aggregates in compartment-1 (C1) would be expected to have a volume of distribution larger than the plasma volume. This volume would be expected to include extravascular (interstitial) fluid spaces. It was further assumed that the smaller aggregates were moving into the extravascular space and not returning. Therefore, when compared with smaller nonprotein drugs, no interchange was occurring between intravascular and extravascular (interstitial) fluid spaces. The model was based on the following equation:
where Cp t was the plasma concentration of HBOC-200 at time t, SA was the proportion (by mass) of the HBOC dose made up of small aggregates in C1, 1-SA was the proportion of the dose made up of large aggregates in C2, DIV was the dose in grams per kilograms, VS and VL were the estimated volumes (milliliters per kilograms) of the small and large aggregates, respectively, and k1,0 and k2,0 were the fractional elimination rate constants of the small and large aggregates from C1 and C2.
In the model fitting process, a value of ∼64.5 mL/kg was used as an initial estimate of the plasma volume of a thoroughbred horse. A Bayesian statistical constraint (sd set equal to 4) was used to help identify this variable (7). The weights W(K), applied in the fitting process, used the FSD of the data and were in the form of W(K) = 1/(C*QO(K)2, where QO(K) is the kth observed datum and C is its FSD. The fitting process (iterations) ceases when the improvement in the sums of squares of the last iteration is <1%.
Half-lives (t1/2) were calculated as the natural log (base2) divided by k1,0 and k2,0. Clearance (Cl) from each compartment was calculated as:
Area under the plasma concentration curve from 0 to 24 h (AUC0 24) was calculated by integration of the predicted concentration of HBOC-200. The maximal plasma concentration Cmax was obtained from the observed data.
Total Hb in blood, plasma Hb (Hbp), was also measured using a hemoximeter (OSM3 Hemoximeter, Radiometer USA, Westlake, OH). Total protein (TP) was measured using refractometry (Leica Microsystems, Bannockburn, IL), and hematocrit (Hct) determined by centrifugation using a high-speed microhematocrit (Micro-MB centrifuge, Fisher, Pittsburgh, PA).
Pharmacokinetic variable estimates of HBOC-200 were expressed as median and range, and the nonparametric Wilcoxon and Kruskal-Wallis rank-sum tests were used for statistical comparisons of variables (JMP Version 4.0; SAS Institute Inc, Cary, NC). Analysis of variance was used for variable analysis. The plasma concentrations of HBOC-200 were expressed as mean () and sd. Significance was designated at P < 0.05.
Pharmacokinetic variable estimates are shown in Table 1. There was a significant difference (P < 0.038) between the volumes of VS and VL. Cmax was 907.1 ± 297.2 μg/mL and was attained at 5 min. There was a biexponential decrease to 98.1 ± 26.9 μg/mL at 24 h (Fig. 2). HBOC-200 was still quantifiable in three horses at 36 h (79.3 ± 15.3 μg/mL) after the administration. HBOC-200 was not detected in urine. The median Cl from C1 of the smaller HBOC-200 aggregates was 42.2 mL · kg−1 · h−1 (28.6–82.7 mL · kg−1 · h−1), which represented 47.0% (37.2%–72.9%) of the material infused. The median Cl from C2 of the larger HBOC-200 aggregates was 3.9 mL · kg−1 · h−1 (2.6–6.0 mL · kg−1 · h−1), which represented 53.0% (27.1%–62.8%) of the material infused. Infusion of this dose of HBOC-200 did not produce significant changes (P < 0.228) in Hct, TP, and total Hbtot, and the values were 38.6% ± 2.1%, 6.3 ± 0.33 g/dL, and 12.7 ± 0.84 g/dL, respectively. Discoloration of plasma was seen after the administration of HBOC-200: the Hbp changed from a control background of 0.04 g/dL to 0.2 g/dL for 6 h after the infusion and 0.1 g/dL at 24 h after the infusion. The mean FSD for all HBOC-200 estimated variables was 0.052.
We recently developed a method for specific identification, confirmation, and quantification of various types of HBOC solutions (5,6). The method identifies unique peptide fragments specific for HBOC-200, bovine, equine, canine, and human Hb. The amino acid sequences of these unique tryptic peptides were identified and correlated with their characteristic mass spectral data. The proteolytic digestion of HBOC allows the distinction of bovine Hb from native hemolyzed equine, canine, or human Hb. One of the characteristics of equine blood is the ease of hemolysis after collection and transport. Thus, specific identification of the different types of Hb, native versus foreign (i.e., HBOC), is essential for an accurate description of the pharmacokinetic profile of HBOC solutions and is essential in determining its illegal administration to animal or human athletes in competition.
The analysis technique used for this study differs from other recently described methods in its sensitivity and specificity (8,9). The use of the Q-TOF mass spectrometer in the quantification improved the limit of quantification to 50 μg/mL and allowed a more complete description of the pharmacokinetic profile of HBOC-200. Some discoloration of plasma was noted after the infusion of this dose of HBOC-200, but the changes in HbP were minimal. The changes in HbP when using the more traditional clinical methods of measurement, as compared with the analytical methods described above, were not sensitive enough to determine accurate changes in HBOC-200.
Most drugs administered by IV exhibit compartmental distribution and elimination phases, and drugs in plasma distribute into and attain equilibrium with extracellular/interstitial spaces and, in some cases, intracellular spaces. The estimated rate constants between compartments for traditional drugs are based on tissue perfusion, protein and tissue binding, solubility in various tissues, and metabolism. However, large MW heterogeneous HBOCs such as HBOC-200 used in this study cannot be considered as conventional drugs, and models used to estimate distribution and elimination variables of drug pharmacokinetics do not fully apply. Based on this assumption, the standard interchanging compartmental analysis to determine pharmacokinetic variable estimates was not considered a totally appropriate approach. It was assumed that the rate of transfer of HBOC-200 molecules from the intravascular to the extravascular compartment would be largely based on aggregate size, and smaller aggregates would move more rapidly from plasma into the extravascular compartment. The larger aggregates would remain in plasma for a prolonged period of time and would slowly be removed by the reticulo-endothelial system.
A number of previous pharmacokinetic studies involved a combination of blood removal and volume replacement (10–15). This study was performed in normal horses without any previous removal of blood. The HBOC-200 was infused over a short period of time, the dose administered (∼0.060 g/kg) in this study was small compared to a therapeutic dose (1.3–3.9 g/kg), and changes in hemodynamic and oncotic pressure were not expected and, therefore, not measured. This small dose eliminated any changes in the plasma volume or oncotic pressure that normally would have resulted from the strong colloid oncotic effects of a large dose of the HBOC promoting influx of interstitial water into the circulation (2). No changes in Hct or TP were observed supporting this contention. This administration also closely mimics an IV bolus administration generally used to determine the initial pharmacokinetics of drugs. The method of analysis was sensitive enough that a complete description of the kinetics was possible despite the subtherapeutic dose.
Previous pharmacokinetic studies described a monoexponential elimination process after the continuous infusion of HBOC-250, which has a slightly higher average MW (250 versus 200 kD). The t1/2 ranged from 8.5 to 16 hours for the smallest and largest dose and 20 hours after the administration of ∼48 g (11,15). Renal excretion was also not noted in this study. We speculate that the finding of a mono-exponential elimination in previous studies was related to the lack of early sampling times, as compared with our study where sampling occurred from the contralateral vein during and immediately after the infusion. The intent of the present study was to determine the kinetics of HBOC with a range of aggregate sizes.
Pyridoxilated-hemoglobin-polyoxyethylene conjugate of human Hb with a larger MW of ∼90 kD or more eliminates many untoward physiological aspects of the HBOC infusion, including the rapid elimination from circulation of the smaller aggregates. A reported monoexponential decline, with a t1/2 of ∼40 hours, in dogs is consistent with its larger MW (13).
Organ distribution and elimination of individual MW components was studied in rats after the transfusion of human Hb polymerized with periodate-oxidized, ring-opened raffinose (10). The t1/2 of individual MW components were approximately 4 hours for the smaller monomeric fraction, 9 hours for the dimer, and 15 hours for the larger fractions representing trimers to nanomers (10). These studies also suggested elimination based on MW or aggregate size. Using 14C-labeled α Hb, the largest concentration of labeled breakdown components was found in the kidney and in tissues of the reticulo-endothelial system, spleen, bone marrow, and liver (10). In male Sprague-Dawley rats administered 400 mg/kg of diaspirin cross-linked human Hb (DCLHb), a lower MW compound, a terminal t1/2 of 3.8 hours was reported. The production of liver cirrhosis did not alter its disposition, probably because of extrahepatic breakdown by the reticulo-endothelial system (14).
The passage of molecules through the capillary walls is also regulated by size and charge. The two-pore concept suggests that most microvascular walls have a large number of functional small pores, restricting the passage of proteins, and a small number of non–size-selective large pore pathways (>240 Å in diameter), permitting the passage of macromolecules from blood to tissue (16). The smaller Hb aggregates will leave the vascular compartment and are rapidly removed via the lymphatic system. DCLHb manufactured from human blood has been shown to filter into the lung and soft tissue lymph of sheep and the vascular septa of perfused rat lungs (17,18). Eliminating the low MW components of DCLHb resulted in a longer t1/2 and undetectable levels of polymer in hilar lymph fluid (19). Also HBOC-200 enters the lymphatics in sheep within two to three hours (Robert A. Gunther, PhD, University of California Davis, personal communication, 2001), indicating that indeed smaller sized Hb polymers in HBOC-200 may exit the circulation as rapidly as those in DCLHb. An increase in MW of Hb aggregates to 90 kD seems to eliminate many untoward physiological aspects of HBOC infusion, including the rapid removal from circulation of smaller Hb polymers and vasoconstrictive properties (13,15,20). One may, therefore, speculate that newer generation HBOCs composed primarily of larger size Hb aggregates may follow a mono-rather than biexponential elimination pattern, leading to longer plasma retention times.
This pharmacokinetic model also identified the percentage of HBOC-200 eliminated from C1 and C2. Of the HBOC-200 infused, 47% was eliminated from plasma at a t1/2 of less than two hours. The median VS was 86.9 mL/kg, which was larger than the expected plasma volume and represents the rapid diffusion of the smaller bovine Hb aggregates out of the circulation into the extravascular compartments. The median VL was 63.9 mL/kg, which was compatible with published plasma volumes of the resting thoroughbred horse of 63.3–66.2 mL/kg (21,22). This plasma volume represented the larger Hb aggregates still contained with the plasma volume. It is not surprising that, in a canine hemorrhagic shock model, the circulating blood volume diminishes rapidly after resuscitation with the first generation HBOC solutions (2). In this regard, HBOC 200-type Hb solutions behave differently from transfused blood and are more similar to traditional colloids, such as hydroxyethyl starches, that are also characterized by heterogeneity of MW distribution (23)
In summary, elimination of first generation HBOCs such as HBOC-200 is more complex than previously assumed and follows biexponential kinetics because of their heterogeneous nature. The increase in the total aggregate size in the production of newer HBOC preparations or the removal of smaller MW substances from current generation HBOCs would prevent the rapid elimination of a large proportion of the infused substance and increase the intravascular half-life of the total concentration infused.
Future studies evaluating the pharmacokinetic profile of individual HBOC solutions in hypovolemic subjects using animal models of hemorrhagic shock will yield further clinically relevant information. Such studies will help in developing a rational volume replacement therapy for the potentially large group of trauma patients who may benefit from treatment with HBOCs in the future.
We are grateful to A. Hess, D. Telies, W. Chaula, D. Tsang, and F. Hao for their excellent technical assistance.
1. Winslow RM. New transfusion strategies: red cell substitutes. Annu Rev Med 1999;50:337–53.
2. Driessen B, Jahr JS, Lurie F, et al. Arterial oxygenation and oxygen delivery after hemoglobin-based oxygen carrier infusion in canine hypovolemic shock: a dose-response study. Crit Care Med 2003;31:1771–9.
3. Schumacher YO, Schmid A, Dinkelmann S, et al. Artificial oxygen carriers: the new doping threat in endurance sport? Int J Sports Med 2001;22:566–71.
4. Haney CR, Buehler PW, Gulati A. Purification and chemical modifications of hemoglobin in developing hemoglobin based oxygen carriers. Adv Drug Deliv Rev 2000;40:153–69.
5. Guan F, Uboh CE, Soma LR, et al. Confirmation and quantification of hemoglobin-based, oxygen carriers in equine and human plasma by liquid chromatography tandem mass spectrometry. Anal Chem 2004;76:5127–35.
6. Guan F, Uboh CE, Soma LR, et al. Unique tryptic peptides specific for bovine and human hemoglobin in the detection and confirmation of hemoglobin-based oxygen carrier. Anal Chem 2004;76:5118–26.
7. Stefanovski D, Moate PJ, Boston C. WinSAAM: a windows based compartmental modeling system. J Metab 2003;52:1153–66.
8. Varlet-Marie E, Ashenden M, Lasne F, et al. Detection of hemoglobin-based oxygen carriers in human serum for doping analysis: confirmation by size-exclusion HPLC. Clin Chem 2004;50:723–31.
9. Thevis M, Ogorzalek Loo RR, Loo JA, Schanzer W. Doping control analysis of bovine hemoglobin-based oxygen therapeutics in human plasma by LC-electrospray ionization-MS/MS. Anal Chem 2003;75:2955–61.
10. Hsia JC, Song DL, Er SS, et al. Pharmacokinetic studies in the rat on a o-raffinose polymerized human hemoglobin. Biomater Artif Cells Immobilization Biotechnol 1992;20:587–95.
11. Hughes GS Jr, Antal EJ, Locker PK, et al. Physiology and pharmacokinetics of a novel hemoglobin-based oxygen carrier in humans. Crit Care Med 1996;24:756–64.
12. Keipert PE, Gomez CL, Gonzales A, et al. The role of the kidneys in the excretion of chemically modified hemoglobins. Biomater Artif Cells Immobilization Biotechnol 1992;20:737–45.
13. Malchesky PS, Takahashi T, Iwasaki K, et al. Conjugated human hemoglobin as a physiological oxygen carrier: pyridoxalated hemoglobin polyoxyethylene conjugate (PHP). Int J Artif Organs 1990;13:442–50.
14. Palaparthy R, Kastrissios H, Gulati A. Pharmacokinetics of diaspirin cross-linked haemoglobin in a rat model of hepatic cirrhosis. J Pharm Pharmacol 2001;53:179–85.
15. Vandegriff KD, Malavalli A, Wooldridge J, et al. MP4, a new nonvasoactive PEG-Hb conjugate. Transfusion 2003;43:509–16.
16. Rippe B, Haraldsson B. Transport of macromolecules across microvascular walls: the two-pore theory. Physiol Rev 1994;74:163–219.
17. Conhaim RL, Cooler SD, McGrath AM, et al. Filtration of diaspirin crosslinked hemoglobin into lung and soft tissue lymph. Am J Respir Crit Care Med 1998;158:1204–12.
18. Conhaim RL, Rodenkirch LA, Watson KE, Harms BA. Acellular hemoglobin solution enters compressed lung capillaries more readily than red blood cells. J Appl Physiol 2000;89:1198–204.
19. Matheson B, Kwansa HE, Bucci E, et al. Vascular response to infusions of a nonextravasating hemoglobin polymer. J Appl Physiol 2002;93:1479–86.
20. Keipert PE, Gomez CL, Gonzales A, et al. Diaspirin cross-linked hemoglobin: tissue distribution and long-term excretion after exchange transfusion. J Lab Clin Med 1994;123:701–11.
21. Naylor JR, Bayly WM, Schott HC 2nd, et al. Equine plasma and blood volumes decrease with dehydration but subsequently increase with exercise. J Appl Physiol 1993;75:1002–8.
22. Muir WW, Kohn CW, Sams R. Effects of furosemide on plasma volume and extracellular fluid volume in horses. Am J Vet Res 1978;39:1688–91.
© 2005 International Anesthesia Research Society
23. Treib J, Baron JF, Grauer MT, Strauss RG. An international view of hydroxyethyl starches. Intensive Care Med 1999;25:258–68.