|Home | About | Journals | Submit | Contact Us | Français|
Recent clinical evaluation of commercial glutaraldehyde-polymerized hemoglobins (PolyHbs) as transfusion solutions has demonstrated several adverse side effects. Chief among these is the hypertensive effect. Fortunately, previous studies have shown that the hypertensive effect can be attenuated by removing free hemoglobin (Hb) and low-molecular-weight (low-MW) PolyHbs from the PolyHb mixture. In this work, polymerized human Hb (PolyhHb) solutions were synthesized in two distinct quaternary states with high MW and subjected to extensive diafiltration to remove free Hb and low-MW PolyhHb components (<500kDa). The resultant PolyhHb solutions possessed high MW, distinct quaternary state, distinct reactivities with O2 and CO, similar NO deoxygenating rate constants, distinct autoxidation rate constants, high viscosity, and low colloid osmotic pressure. To preliminarily assess the ability of PolyhHb solutions to oxygenate surrounding tissues fed by a blood vessel, we evaluated the ability of PolyhHbs to transport O2 to cultured hepatocytes in a mathematical model of a hollow fiber bioreactor. The structure of individual hollow fibers in the bioreactor is similar to that of a blood vessel and provides an easy way to assess the oxygenation potential of PolyhHbs without the need for expensive and time-consuming animal studies. It was observed that PolyhHbs with low O2 affinities were more effective in oxygenating cultured hepatocytes inside the bioreactor than high O2 affinity PolyhHbs. Taken together, our results show that it is possible to synthesize high-MW PolyhHbs with no free Hb and low-MW PolyhHb components that are capable of transporting O2 to cultured cells/tissues.
Hemoglobin-based oxygen carriers (HBOCs) are being developed as red blood cell substitutes. However, vasoconstriction and hypertension still remain significant side effects.1–4 Although the exact mechanism of vasoconstriction upon HBOC transfusion is not known, there are two major hypotheses in the literature, namely, nitric oxide (NO) scavenging (by far the most popular one of the two) and oxygen oversupply.5–7 To limit/prevent vasoconstriction by either mechanism, the size (i.e., molecular weight, MW) of the HBOC should be increased and free hemoglobin (Hb) and low-MW components should be removed from solution to reduce the interaction of the HBOC with the endothelium.6,8,9 Therefore, strategies targeted toward increasing the size of HBOCs should be able to reduce these unwanted side effects. Polymerization of Hb represents such an approach.
Two commercial polymerized Hb (PolyHb)-based HBOCs were developed for clinical use, Hemopure® (OPK Biotech, Cambridge, MA) and PolyHeme® (Northfield Laboratories Inc., Evanston, IL). Hemopure® had an average MW of 250kDa10,11 and was composed of <5% of free Hb,12 whereas PolyHeme® had an average MW of 150kDa13 and was composed of <1% of free Hb.14 Unfortunately, both PolyHbs elicited substantial hypertension in vivo upon transfusion.4,11,15
Fortunately, a path exists to improve the safety of PolyHbs. Sakai et al. was the first to demonstrate that the extent of vasoconstriction and hypertension was inversely proportional to the size of the HBOC.16 Unfortunately, this study was performed on a wide variety of HBOCs synthesized with different chemistries ranging from acellular Hbs to cellular Hbs. Hence, it was not clear if this principle would be applicable to HBOCs synthesized with the same chemistry. However recently, Cabrales et al. demonstrated the same heuristic on a small library of tense-state PolyHbs (>500kDa) of varying size synthesized with the same polymerization chemistry, thereby verifying Sakai's observations.9 The results of Cabrales et al.'s study identified two tense-state PolyHbs (50:1 and 40:1) with MWs of 1.36–23.71 MDa,9 whereas another study identified a zero-link cross-linked bovine Hb (ZL-HbBv) with an average MW of 42 MDa17; both were reported to elicit no vasoconstriction in vivo. Therefore, polymerization of high-MW Hb seems like an effective and simple strategy to limit/prevent vasoconstriction in vivo.
This current work will expand on the study by Cabrales et al.,9 which focused on bovine Hb, and focus solely on the synthesis and in vitro biophysical characterization of polymerized human Hb (PolyhHb) solutions in two distinct quaternary states, with both high MW and no free Hb or low-MW PolyhHbs in solution. Glutaraldehyde will be used to nonsite specifically cross-link/polymerize hHb to yield high-MW PolyhHb solutions that are frozen in either the tense (T) or relaxed (R) quaternary states. The PolyhHb solutions will be subjected to extensive diafiltration on a 500kDa filter to remove free Hb and low-MW PolyhHbs. The biophysical properties of the PolyhHbs will then be extensively characterized and used in a mathematical model that describes O2 transport in a hollow fiber (HF) bioreactor. The HF bioreactor is composed of individual HFs that mimics the structure of a blood vessel. Therefore, the mathematical model will be used to preliminarily assess the ability of PolyhHbs' to oxygenate cultured cells in the device and by extension tissues fed by a blood vessel. The results of these simulations will inform the proper selection of PolyhHbs for applications ranging from tissue engineering to transfusion medicine.
Glutaraldehyde (70%), synapinic acid, NaCl (sodium chloride), NaOH (sodium hydroxide), Na2S2O4 (sodium dithionite), NaCNBH3 (sodium cyanoborohydride), NaBH4 (sodium borohydride), KCl (potassium chloride), CaCl2 (calcium chloride), sodium lactate, N-acetyl-L-cysteine (NALC), Na2HPO4 (sodium phosphate dibasic), and NaH2PO4 (sodium phosphate monobasic) were purchased from Sigma-Aldrich (St. Louis, MO). KCN (potassium cyanide), KFe(CN)6 (potassium ferricyanide), and all other chemicals were purchased from Fisher Scientific (Pittsburgh, PA). HF cartridges were obtained from Spectrum Labs (Rancho Dominguez, CA).
T-state PolyhHb was synthesized as previously described in the literature.9,21–23 hHb was diluted with 20mM phosphate buffer (PB; pH 8.0) to yield 1.5 L of 0.3mM hHb that was stored in a 2 L glass bottle submerged in an ice bath. The contents of the glass bottle were then subjected to vacuum for 1min, followed by argon purging for 1min. This gas exchange step was repeated a total of 10 times. The partially deoxygenated hHb solution was then purged with argon for 40min. After another two short gas exchange steps and long argon purging step, Na2S2O4 was used to remove any residual O2 in the hHb solution. A stock solution of Na2S2O4 was prepared by dissolving 300mg Na2S2O4 in 200mL of cold PB, which was kept on ice in the dark. The Na2S2O4 stock solution was injected into the hHb solution under constant stirring, while the pO2 of the hHb solution was monitored after each 30mL infusion of Na2S2O4 using a RapidLab 248 Blood Gas Analyzer (Siemens USA, Malvern, PA). Once the pO2 was out of range (<0mm Hg), the infusion of Na2S2O4 was terminated. Next, deoxygenated glutaraldehyde (70%) was added to the deoxygenated hHb solution to polymerize hHb at the following glutaraldehyde to hHb tetramer molar ratios: 50:1 and 40:1. The polymerization reaction was continued for 2h in the dark at 37°C, and then quenched with 22.5mL of 2M NaBH4 in 20mM PB to yield T-state PolyhHb.
R-state PolyhHb was synthesized as previously described in the literature.21–23 About 1.5 L of 0.3mM hHb in 20mM PB was purged with pure O2 for 2h in an ice bath, and its pO2 was continuously monitored with a RapidLab 248 Blood Gas Analyzer. After the pO2 was out of range (>749mm Hg), glutaraldehyde was added at the following glutaraldehyde to hHb molar ratios—30:1 and 20:1. At the end of the 2h polymerization period in the dark at 37°C, 3.5mL of 8M NaCNBH3 in 20mM PB was added to the reaction mixture to reduce the resultant Schiff bases and methemoglobin (metHb; i.e., oxidized form of Hb). The reaction mixture was then stirred on ice for another 30min, and then 13.5mL of 2M NaBH4 was added to quench the reaction to yield R-state PolyhHb.
Both T-state PolyhHb and R-state PolyhHb were first passed through a column packed with glass wool to remove any large particles from solution. The clarified PolyhHb solution was then subjected to 4 cycles of diafiltration on a 500kDa MW cutoff HF cartridge (Spectrum Labs) with a modified lactated Ringer's buffer (NaCl 115mM, KCl 4mM, CaCl2 1.4mM, NaOH 13mM, sodium lactate 27mM, and NALC 2g/L). The concentrated PolyhHb solution was then stored at −80°C.
The total protein concentration of hHb/PolyhHb was measured by the Bradford method24 using the Coomassie Plus protein assay kit (Pierce Biotechnology, Rockford, IL).
The metHb level of hHb/PolyhHb was measured by the cyanomethemoglobin method.25
Twenty-five micrograms of hHb or PolyhHb was mixed with an equal volume of denaturation sample buffer (Bio-Rad, Hercules, CA), and boiled for 5min before being loaded on the sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel. Electrophoresis was conducted in a Mini Format 1-D Electrophoresis System (Bio-Rad) using a 12% SDS-PAGE gel. After electrophoresis, the gel was fixed and stained in Staining Buffer (Bio-Rad), and destained before image analysis. Raw photos were processed by Quantity One (Bio-Rad) software.
Twenty micrograms of hHb or PolyhHb was mixed with a 1/4 volume of native-PAGE Sample Buffer (Invitrogen, Carlsbad, CA), and loaded on a Novex® 4%–12% Tris-Glycine gel (Invitrogen). After electrophoresis, the gel was fixed and stained in Staining Buffer (Bio-Rad), and destained before image analysis. Photos were taken and processed by Quantity One (Bio-Rad) software.
The absolute MW distribution of hHb and PolyhHb solutions was characterized by size exclusion chromatography coupled with multi-angle static light scattering as described previously in the literature.9,26
Protein samples were diluted with deionized water to a concentration of 0.2mg/mL, and mixed in 1:1:1 (v/v/v) ratio with 0.1M HCl and saturated sinapinic acid in 50% (v/v) acetonitrile solution on a matrix-assisted laser desorption/ionization (MALDI) plate (Bruker Inc., Billerica, MA). Results were obtained on a Bruker Microflex machine (Bruker Inc.).
The equilibrium O2 binding properties of hHb/PolyhHb were measured using a Hemox-Analyzer (TCS Scientific Corporation, New Hope, PA) at 37°C. Samples were prepared by thoroughly mixing 50–100μL of hHb/PolyhHb with 5mL of Hemox buffer (pH 7.4), 20μL of Additive-A, 10μL of Additive-B, and 10μL of anti-foaming agent (all purchased from TCS Scientific Corporation). The samples were allowed to saturate to a pO2 (i.e., partial pressure of O2) of 145±2mm Hg using compressed air. After giving the sample enough time to equilibrate, the gas stream was switched to pure N2 to deoxygenate the sample. The absorbance of oxygenated and deoxygenated Hb in solution was recorded as a function of pO2 via dual wavelength spectroscopy. O2-hHb/PolyhHb equilibrium curves were fit to a four-parameter (A0, A∞, P50, and n) Hill model (Equation 1). Where Y is the fractional saturation of hHb/PolyhHb (Y has no units and ranges from 0 to 1, a value of 0 corresponds to fully deoxygenated hHb/PolyhHb, while a value of 1 corresponds to fully deoxygenated hHb/PolyhHb), while A0 and A∞ represent the absorbance at 0mm Hg and full saturation, respectively. P50 represents the pO2 at which the heme binding sites in the Hb sample are 50% saturated with O2, while n is the Hill coefficient, a measurement of cooperativity in O2 binding.
The goodness of fit was verified by comparing the experimental data with the curve fitted data
The circular dichroism (CD) spectra of hHb/PolyhHb was measured on an AVIV CD spectrometer (Lakewood, NJ). hHb/PolyhHb was first diluted to 1.7mg/mL in 20mM PB (pH 8.0), and scanned from 500 to 250nm in a 0.1cm path length quartz cell. Samples were also diluted to 0.085mg/mL in 20mM PB (pH 8.0), and scanned from 250 to 200nm in a 0.1cm path length quartz cell. Data were collected and analyzed by AVIV data collection software. Each curve was an average of 3 scans.
Fast kinetic measurements of gaseous ligand reactions with hHb/PolyhHb were carried out in an Applied Photophysics SF-17 micro-volume stopped-flow instrument as previously described.27 O2 dissociation kinetics were measured by rapidly mixing 30μM of O2-hHb/PolyhHb (heme concentration) with 1.5mg/mL of sodium dithionite (British Drug House, Poole, England), and the deoxygenation process was monitored by absorbance changes at 437.5nm in 0.05M Tris buffer, pH 7.4, at 25°C. The averaged kinetic traces were fit to exponential equations using Marquardt-Levenberg fitting routines in the Applied Photophysics program. The kinetics of CO association with deoxygenated hHb/PolyhHb were measured in the stopped-flow instrument following absorbance changes at 437.5nm in 0.05M Tris buffer, pH 7.4, at 25°C in the presence of 1.5mg/mL of Na2S2O4. The CO saturated stock solution (about 1mM) was prepared by flowing prewashed CO gas through degassed buffer solution.
The fast kinetics of NO oxidation with O2-hHb/PolyhHb solutions were measured in the stopped-flow instrument as previously described.28 NO gas prewashed in deoxygenated 1M NaOH and buffer solutions was bubbled through deoxygenated 0.05M Tris buffer, pH 7.4, to make an NO-saturated stock solution (~2mM) in a gas-tight serum bottle. Specific amounts of the NO stock solution were then transferred with a Hamilton syringe into a gastight syringe containing deoxygenated buffer solution to make appropriate NO concentrations for the reaction with O2-hHb/PolyhHb solutions. O2-hHb/PolyhHb solutions (1μM heme) were mixed with NO solutions (0–50μM), and the absorbance changes of the reaction were followed at 420nm. Multiple traces were taken and averaged for each reaction. The averaged trace was fit to exponential equations to obtain the reaction rate constant.
All hHb/PolyhHb solutions were reduced to O2-hHb/PolyhHb immediately before autoxidation experiments. Experiments were carried out with 20–25μM hHb/PolyhHb (heme concentration) in sealed cuvettes with room air-equilibrated 50mM Chelex-treated PB at 37°C. Absorbance changes in the range of 450–700nm due to spontaneous oxidation of O2-hHb/PolyhHb were recorded up to 24h in a temperature-controlled photodiode array spectrophotometer (Hewlet Packard 8453, Palo Alto, CA). Similar oxidation assays were also performed in the presence of superoxide dismutase (4.6U/mL) and catalase (414U/mL). A multicomponent analysis was performed to calculate the oxy and ferric species based on their individual extinction coefficients.29 Autoxidation rates were obtained from plots of the loss of O2-hHb/PolyhHb versus time using nonlinear least squares curve fitting with a double exponential equation in Sigma-Plot (SPSS, Chicago, IL).
hHb/PolyhHb viscosity was measured in a cone and plate viscometer DV-II plus with a cone spindle CPE-40 (Brookfield Engineering Laboratories, Middleboro, MA), at a shear rate of 160/second. The colloid osmotic pressure (COP) of hHb/PolyhHb was measured using a Wescor 4420 Colloid Osmometer (Wescor, Logan, UT).
All data were analyzed by t-test, and a p-value of 0.05 or less was considered significant.
The main function of these materials is to store and transport O2. Therefore, the purpose of the mathematical model is to preliminarily determine the degree to which these polymerized human Hb (PolyhHb) molecules are capable of transporting O2 to surrounding cells cultured in the HF bioreactor and by extension tissues fed by a blood vessel without performing expensive and time-consuming in vitro (bioreactor) or in vivo (animal) studies. In reality, the HF bioreactor is used to facilitate cell culture; however, it can also be used to simulate the structure of a blood vessel surrounded by tissue. The main advantage of the O2 transport model is that the model parameters have been experimentally measured for all PolyhHbs.
The PolyhHb selection criteria for transfusion versus bioreactor applications will be different and will mainly depend on a variety of factors such as inlet pO2, flow rate of solution, viscosity of solution, concentration of PolyhHb, size of PolyhHb, O2 affinity of PolyhHb, O2 dissociation rate constant of PolyhHb, and cell/tissue O2 consumption rate. A convenient feature of this model is that all of these parameters can be varied in the simulation and its effect on oxygenation of the surrounding cells or tissues can be examined, to select PolyhHbs for either of these two applications.
Therefore, the aim of these simulations is to explore the oxygenation potential of PolyhHbs in oxygenating a tissue-engineered construct (i.e., a bioartificial liver assist device) and by extension tissue fed by a cylindrical blood vessel. A mathematical model previously developed by Chen and Palmer was used to assess the ability of PolyhHbs in oxygenating mammalian cells (C3A hepatocytes) cultured in a HF bioreactor.30 These results should yield some preliminary insight into how PolyhHbs would oxygenate tissues when transfused in vivo.
The O2 transport model is based on the geometry of a single HF in a HF bioreactor, which mimics the in vivo capillary/sinusoid structure (Fig. 1). In this geometry, cell culture medium supplemented with HBOC enters the HF bioreactor at the entrance of the lumen to provide nutrients to cultured cells, which reside in the extra capillary space (ECS), and leave the HF bioreactor through the exit port of the lumen carrying away waste products. The cell culture medium recirculates throughout the entire HF bioreactor system. The HF membrane has a MW cut-off of 35kDa and therefore confines HBOCs (MW>64kDa) within the lumenal space of the HF without contacting the cells residing in the ECS.
The velocity profile in each of the three subdomains (lumen, membrane, and ECS) can be calculated from a set of momentum transport partial differential equations, shown in Equation 2.
Navier-Stokes equation (lumen):
Brinkman's Equation (membrane and ECS):
Where is the dimensionless velocity vector and P′ is the dimensionless pressure. l0, v0, ρ, and μ, represent the reference length, reference velocity, fluid density, and viscosity, respectively.
Equation 3 shows the mass conservation equations, which describe the transport of dissolved O2, total HBOC, and O2-HBOC in dimensionless form. C′ can either represent the dimensionless O2 partial pressure (pO2), total HBOC concentration, or O2-HBOC concentration. C0 can either represent the reference O2 partial pressure, total HBOC concentration, or O2-HBOC concentration. D can either represent the diffusivity of O2 or HBOC. R represents the rate of formation of O2/O2-HBOC. The HBOC diffusivity is estimated from the PolyhHb MW using Equation 4.
The reaction between O2 and HBOC is described by Equation 5, where m is the number of O2 binding sites (i.e., heme groups) on a single HBOC molecule. The number of O2 binding sites on PolyhHb can be calculated by dividing the weight average MW of PolyhHb by the MW of tetrameric Hb multiplied by a factor of 4, which represents the number of heme groups per Hb tetramer. Given the thermodynamic relationship describing the equilibrium between HBOC and O2 (Equation 6, where a1–a4 are the Adair constants), the rate of formation of O2 ( [mm Hg/s]) or O2-HBOC (RoxyHBOC [mol/m3/s]) in the lumen is shown in Equation 7. S is defined as the HBOC saturation; that is, the molar fraction of HBOC that is saturated with O2 and Seq is the HBOC saturation at equilibrium. [O2] is the concentration of dissolved O2, [HBOC] is the concentration of total HBOC and α is the solubility of O2 in the aqueous medium. The O2 consumption rate in the ECS is described by Michaelis-Menten (M-M) kinetics.
Further details about the mathematical model and parameters/constants used in the simulations can be found in the literature.30 The experimentally measured biophysical properties of the hHb/PolyhHb solutions were used in these simulations (i.e., MW, P50, n, and k−). The coupled set of partial differential equations was solved by the finite element method in Comsol Multiphysics (COMSOL, Inc., Burlington, MA) yielding numerical solutions.
The major chemical reaction in the polymerization of hHb with glutaraldehyde involves Michael addition between α,β-unsaturated oligomeric aldehydes and primary amine groups on lysine residues that are present on the surface of hHb (Fig. 2A). Some Schiff bases form during the polymerization step, but they are all reduced via the addition of NaBH4 or NaCNBH3 at the end of the polymerization reaction. Therefore, polymerization of hHb is based on stable C-N bonds that will not hydrolyze in solution.
Figure 2B shows the SDS-PAGE of hHb and all PolyhHb solutions under denaturing conditions. After sample denaturation, hHb yields two bands in the gel. The lower band corresponds to individual α/β subunits, whereas the upper band corresponds to α-α/β-α/α-β dimers that do not dissociate into individual subunits. In contrast, all PolyhHb solutions under denaturing conditions display strong bands >250kDa and weak bands close to 15, 30, 45, and 60kDa in MW corresponding to individual α/β subunits, α-α/β-α/α-β dimers, α/β trimers, and α/β tetramers. In fact, PolyhHb's with higher cross-link density locked in the same quaternary state exhibit less individual α/β subunits, dimers, trimers, and tetramers compared to lower cross-link density PolyhHb.
Figure 2C shows the absolute MW distribution of hHb and PolyhHb solutions under nondenaturing conditions. Tetrameric hHb displays a weight averaged MW of ~62kDa, whereas all PolyhHbs display weight averaged MWs >62kDa. The weight averaged MW of hHb and all PolyhHbs is displayed in Table 1. Light scattering results clearly show that both T- and R-state PolyhHbs possess no individual α/β subunits, dimers, trimers, and tetrameric hHb in solution.
Figure 2D shows the native PAGE analysis of hHb and PolyhHb solutions under nondenaturing conditions. The position of the hHb tetramer in the gel is labeled by an arrow. None of the PolyhHb solutions display bands at or lower than the position of the hHb tetramer. This demonstrates that PolyhHb is free of Hb tetramers, trimers, dimers, or individual α/β subunits in solution monomers.
Figure 2E shows the MALDI mass spectral analysis of hHb and PolyhHb solutions. It is evident that hHb is composed of 2 peaks corresponding to individual α and β subunits. There is no peak corresponding to dimers as indicated in the SDS-PAGE analysis of hHb (Fig. 2B). Also, there are no peaks corresponding to α/β subunits, dimers, trimers, or tetramers for all PolyhHb solutions, in stark contrast to the SDS-PAGE results (Fig. 2B).
To estimate the stability of PolyhHb during long-term storage, the MW distribution of 40:1 T-state PolyhHb was analyzed immediately after synthesis, after 8 months of storage at −80°C, and after 8 months of storage at −80°C coupled with incubation at 37°C for 5 days (Supplementary Figure S1; Supplementary Data are available online at www.liebertonline.com/ten). Eight months of storage at −80°C had no influence on the MW distribution of the 40:1 T-state PolyhHb, whereas incubation of the 40:1 T-state PolyhHb at 37°C for 5 days significantly increased the MW of the PolyhHb. Analysis of the MW of the 40:1 T-state PolyhHb under all conditions yielded no free Hb tetramers, trimers, dimers, or α/β subunits in solution. For this study the PolyhHb was in the oxygenated form and not the deoxygenated form. Therefore, it is more subjective to oxidative reactions for prolonged exposures to high temperatures unlike the deoxygenated version of the same compound.
Figure 3A shows the O2-hHb/PolyhHb equilibrium curves. T-state PolyhHb solutions are right-shifted compared to the control of hHb, whereas R-state PolyhHb solutions are left-shifted compared to the control of hHb. Figure 3B shows the regressed O2 affinity (P50) of T- and R-state PolyhHbs. Compared to the control of hHb (13.27±0.55mm Hg), T-state PolyhHbs display much higher P50s. In contrast, R-state PolyhHbs has extremely low P50s compared to both hHb and T-state PolyhHbs. Both T- and R-state PolyhHbs display cooperativity coefficients <1, whereas unmodified hHb has a cooperativity coefficient of 2.59±0.12 (Fig. 3C). The O2 affinity and cooperativity coefficient of hHb and all PolyhHbs are listed in Table 1.
Figure 4 displays the metHb levels of hHb, T-, and R-state PolyhHb solutions. Both T- and R-state PolyhHbs display metHb levels <10%, but higher than the metHb level of hHb (0.88%±0.33%). Table 1 summarizes the biophysical properties of hHb and PolyhHb solutions.
The influence of glutaraldehyde polymerization on hHb structure is determined by CD spectroscopy in the far-ultraviolet and near-ultraviolet regions. The influence of hHb polymerization on the secondary structure of hHb is determined by far-ultraviolet CD spectroscopy in the wavelength range 200 to 250nm (Fig. 5A). In the far-ultraviolet region, all hHb/PolyhHb spectra share a common peak at ~220nm with similar spectral intensity. In contrast, the influence of hHb polymerization on the heme environment is determined by near-ultraviolet CD spectroscopy in the wavelength range 250 to 500nm (Fig. 5B). In the near-ultraviolet region, all hHb/PolyhHb spectra share common peaks at ~260 and 420nm, with slight differences in intensity among all hHb and PolyhHb solutions.
Figure 6 shows a representative time course of CO binding to deoxygenated Hb and the CO concentration dependence of the apparent reaction rate constants for deoxygenated native hHb and PolyhHb solutions. Unlike the slightly lower CO binding rate constants of T-state PolyhHb solutions compared to that of hHb, R-state PolyhHb solutions exhibits much higher CO reactivities characterized by the large increase (>20-fold) in bimolecular rate constants. Table 2 summarizes the CO binding (k'on,CO), O2 dissociation (koff), and NO binding/oxidation (k'ox,NO) rate constants as the mean of a minimum of three separate measurements. The O2 dissociation rate constants of T- and R-state PolyhHb solutions differ from that of native hHb by either a ~30% increase or reduction, respectively, whereas the NO reactivities are very similar among all hHb/PolyhHb solutions in our study.
The autoxidation rate constants of oxygenated hHb and PolyhHb solutions are summarized in Table 3. The high O2 affinity PolyhHb solutions exhibit similar spontaneous oxidation rate constants compared to that of unmodified hHb. Conversely, the low O2 affinity PolyhHb solutions show much elevated spontaneous oxidation rates under the same experimental conditions. Additionally, PolyhHb solutions show much less of a response compared to hHb in the presence of the antioxidant enzymes superoxide dismutase (SOD) and catalase.
Table 4 shows the viscosity and COP of hHb as well as T- and R-state PolyhHb solutions. The viscosity of both PolyhHbs in a distinct quaternary state increases with increasing cross-link density, whereas the COP is virtually unaffected by cross-link density. However, T-state PolyhHbs exhibit a much lower COP versus R-state PolyhHbs.
Figure 7A shows the normalized O2 consumption rate of C3A hepatocytes as a function of the inlet pO2 (pO2,in). The value of the normalized O2 flux starts at 15–20 at low pO2,ins (~5mm Hg) and then decreases as the pO2,in increases for all HBOCs. At higher pO2,ins (~150mm Hg), O2 transport is still at least 2–3 times greater versus the case with no HBOC supplementation.
The pO2 profiles within the HF bioreactor (including lumen, membrane, and ECS) are shown in Figure 7B for all HBOCs at varying concentrations. Each unit represents the cross-sectional view of a single HF (Fig. 1C). The top horizontal boundary represents the HF centerline, whereas the left boundary represents the inlet of the lumen and the right boundary represents the lumen exit. The maximum heme concentration ([Heme]) in the HF bioreactor is normalized by the average heme concentration in human blood (8800μM). Without any HBOC supplementation in the HF bioreactor, most of the ECS is hypoxic, with pO2 levels below 20mm Hg. As the HBOC concentration increases, O2 transport in the HF improves and the hypoxic region in the ECS gradually reduces in size. Among all the HBOCs simulated, R-state PolyhHbs show the least capacity to oxygenate the ECS versus hHb and T-state PolyhHbs. However, T-state PolyhHbs are better at oxygenating the ECS versus native hHb and R-state PolyhHb.
Similar oxygenation results are observed for the ECS zonation breakdown plots (Fig. 7C). As mentioned previously, in the absence of HBOCs, the hypoxic region (<25mm Hg) dominates the majority (~95%) of the ECS volume, and only a small percentage of hepatocytes (~5%) are subjected to in vivo pO2 levels (25–70mm Hg). The hypoxic region gets smaller in size when HBOCs are present in the cell culture media. However, under the simulation conditions, only T-state PolyhHbs successfully improve O2 transport and recapitulated the in vivo pO2 environment in the ECS with virtually no part of the ECS experiencing hypoxic oxygenation (<5%). In contrast, a significant volume of the ECS is hypoxic with supplementation of hHb (~55%) and R-state PolyhHbs (65%–80%).
Light scattering, native-PAGE, and MALDI analysis of both T- and R-state PolyhHb solutions show that they exhibit large weight averaged MWs, without the presence of hHb tetramers, trimers, αβ dimers, and individual α/β subunits in solution in approximately physiological conditions. On the other hand, SDS-PAGE shows the presence of a small percentage of individual α/β subunits and αβ/α-α/β-β dimers, trimers, and tetramers in the PolyhHb solutions under denaturing conditions. This apparent contradiction is caused by the different environmental conditions to which the sample is exposed to during these measurements.
In SDS-PAGE, hHb/PolyhHb solutions are subjected to harsh denaturing conditions, which should dissociate α/β monomers and dimers that are not cross-linked to the PolyhHb superstructure. However, these harsh sample processing conditions actually chemically cross-link α/β subunits into higher order structures, that is, dimers, trimers, and tetramers. Evidence of this phenomenon is found in the SDS-PAGE of hHb, where only one band corresponding to α/β subunits should be present in the gel. Instead, the SDS-PAGE of hHb has an additional band corresponding to dimers. There are no chemical cross-links between the individual subunits of unmodified hHb. Therefore, SDS-PAGE of hHb should only yield 1 band corresponding to α/β subunits. The fact that there are 2 bands indicates that the sample preparation condition for SDS-PAGE induces chemical cross-links between subunits. This calls into question the actual presence of cross-linked dimers, trimers, and tetramers in PolyhHb solutions as determined by SDS-PAGE. These cross-linked species are not present in solution when PolyhHbs are analyzed by light scattering, native-PAGE, and MALDI mass spectral analysis.
Therefore, it is safe to conclude that the final PolyhHb products are free of α/β subunits, dimers, trimers, and tetramers in solution. However, the PolyhHb products are composed of some uncross-linked α/β subunits that are not chemically integrated into the PolyhHb superstructure. At a fixed quaternary state, SDS-PAGE, native-PAGE, light scattering, and MALDI all show an increase in the MW of PolyhHb with increased cross-link density. This result is expected since the number of potential chemical cross-links increases with increasing cross-link density.
Compared to other HBOCs reported in the literature, T- and R-state PolyhHb solutions possess smaller MWs (1.10–18.44 MDa) than 50:1 T- and 40:1 R-state PolybHb (16.59–26.33 MDa)21,22 and ZL-HbBv (42 MDa),17 but larger MWs than O-raffinose cross-linked hHb (O-R-polyHbA0, 64–600kDa)31 and Oxyglobin® (87.2–502.3kDa).32 The two commercially manufactured PolyHb products, Hemopure® and PolyHeme®, possess average MWs of 250 and 150kDa, respectively.11,13 Hence, PolyhHb products are up to 70 times larger than Hemopure®, and 120 times larger than PolyHeme®. In addition, Hemopure® and PolyHeme® both have <5% unreacted hHb tetramers and αβ dimers in solution. Free hHb tetramers, αβ dimers, and low-MW PolyHbs are believed to elicit vasoconstriction and systemic hypertension in vivo. All PolyhHbs are free of hHb, αβ dimers, and low-MW PolyHbs and should generate limited/no vasoconstriction compared to current commercial PolyHbs.
T- and R-state PolyhHbs exhibit vastly different O2 affinities. T-state PolyhHbs possess P50s>35mm Hg, which are considerably higher than that of unmodified hHb (13.27±0.55mm Hg), hHb inside human red blood cells (26–28mm Hg), PolyHeme® (29mm Hg),12 and ZL-HbBv (6.4mm Hg),17 but comparable to Hemopure® (38mm Hg),12 O-R-polyHbA0 (50.9mm Hg),33 and Oxyglobin® (35.1mm Hg).32 Since hHb is totally deoxygenated before, during and after polymerization (but not during storage), the resultant T-state PolyhHbs are conformationally frozen in the T-state after polymerization. Hence, T-state PolyhHbs exhibit higher P50s compared to hHb. In contrast, R-state PolyhHbs exhibit P50s<2mm Hg, indicating their extremely high O2 affinity. This is because hHb is fully saturated with pure O2 before, during and after polymerization. Therefore, the resultant R-state PolyhHbs are conformationally frozen in the R-state after polymerization. Hence, R-state PolyhHbs exhibit lower P50s compared to hHb. This pattern is consistent with PolybHb (the P50 is 41mm Hg for 50:1 T-state PolybHb and 0.66mm Hg for 40:1 R-state PolybHb).21,22
Both PolyhHbs show no cooperativity. Polymerization of hHb freezes the structure of hHb in a well-defined quaternary state. Since, the individual subunits are cross-linked to each other, quaternary structure changes, which would occur during normal O2 binding/offloading are hindered by the presence of chemical cross-links. This results in the loss of cooperative O2 binding to the hHb tetramer for all PolyhHbs. The two commercially produced PolyHbs, Hemopure® (n=1.4) and PolyHeme® (n=1.7),12 and other reported HBOCs, PolybHb (n<1),21 ZL-HbBv (n=1.2),17 O-R-polyHbA0 (n=1),33 and Oxyglobin® (n=1.4),32 also show reduced cooperativity compared to native hHb (n=2.59±0.12) or bHb (n=2.5).21 These results indicate that the extent of polymerization of PolyhHb is significantly greater than that of commercial PolyHbs.
The metHb level of T- and R-state PolyhHbs is higher than that of hHb. However, R-state PolyhHbs have a higher metHb level than T-state PolyhHbs. This observation is consistent with the fact that R-state PolyhHbs are maintained in an oxygenated environment, which enhances autoxidation of the heme prosthetic group. On the other hand, current HBOCs in phase III trials possess similar metHb levels. For example, Hemopure® has a metHb level <10%, and PolyHeme® has a metHb level <8%.12 Therefore, T- and R-state PolyhHbs exhibit metHb levels that are acceptable for clinical evaluation. Nonetheless, the metHb level of PolyhHb solutions could be further reduced after polymerization by incubating these solutions with additional NaCNBH3, a mild reducing agent.
The CD spectra of hHb and PolyhHb solutions show no significant differences either in the far-ultraviolet region or in near-ultraviolet region. This indicates that the influence of polymerization on the secondary structure and heme environment of hHb is negligible.
Fast kinetic analysis of gaseous ligand binding with PolyhHb solutions shows that the O2 dissociation rate constant of PolyhHbs (49.7–51.7s−1 for T-state PolyhHbs and 31.4–25.5s−1 for R-state PolyhHbs) are comparable to PolybHb (53s−1 for 50:1 T-state PolybHb, and 22s−1 for 40:1 R-state PolybHb)21 and ZL-HbBv (27.4s−1),17 but smaller than the rate constants for O-R-polyHbA0 (130s−1)33 and Oxyglobin® (61.8s−1).32 The CO association rate constants of T-state PolyhHbs (0.181–0.184μM−1s−1) are comparable to 50:1 T-state PolybHb (0.18μM−1s−1),21 ZL-HbBv (0.28μM−1s−1),17 and Oxyglobin® (0.19μM−1s−1).32 In contrast, the CO association rate constant of R-state PolyhHbs (4.76–4.88s−1) are comparable to 40:1 R-state PolybHb (4.84s−1)21 and O-R-polyHbA0 (1.2s−1).33 Both T-state PolyhHbs possess O2 dissociation rate constants that are larger than that of hHb (40.4s−1) and CO binding rates smaller than that of hHb (0.214s−1). On the contrary, both R-state PolyhHbs possess O2 dissociation rate constants that are smaller than that of hHb, and CO association rate constants that are much larger than hHb. These results are consistent with the O2 affinities of PolyhHbs, which are in turn reflected by their P50 values obtained under equilibrium conditions. NO binding/oxidation of oxygenated PolyhHb solutions occurs very rapidly on the order of 107 M−1s−1, and remains largely unchanged after glutaraldehyde polymerization. However, the kinetic parameters for O2 dissociation and CO binding of PolyhHb solutions indicate the reduced or increased O2 affinities after T-state or R-state PolyhHb modification, respectively. More specifically, the large reduction in the P50 of R-state PolyhHb solutions is mostly reflected in their increased CO binding rate constants, whereas the increase in the P50 values of T-state PolyhHb solutions are in agreement with their O2 dissociation rate constants. These physicochemically characteristic changes are determined by the nature of the glutaraldehyde cross-linking reactions under oxygenated or deoxygenated conditions.
Hb autoxidation is an intrinsic characteristic of the heme group, and is usually affected by chemical modification of the Hb molecule. The autoxidation rate constant of low O2 affinity T-state PolyhHbs (0.00145–0.00134min−1) is higher than that of R-state PolyhHbs (0.00050–0.00069min−1). The elevated autoxidation rates, especially for low O2 affinity PolyhHb solutions, are consistent with that of previously reported cross-linked and PolyHbs. For example, the autoxidation rate constant of ZL-HbBv is three times greater than that of bHb, whereas the rate of Oxyglobin® autoxidation is 1.3 times greater than that of bHb. Low O2 affinity 50:1 T-state PolybHb has an increased rate of autoxidation compared to high O2 affinity 40:1 R-state PolybHb and native bHb.21 This undesirable change is one of the biggest challenges facing HBOC development. The increased Hb oxidation not only lowers its capacity to carry O2, but also introduces enhanced toxicity that could elicit cell and tissue damage. The weakened protection by SOD and catalase against autoxidation of PolyhHb solutions underlines the complexity of the issues faced by this chemical modification.
The viscosity of both types of PolyhHb solutions increases with increasing cross-linking density and is higher than that of whole blood (~3cp), Hemopure® (1.3cp at 13g/dL), and PolyHeme® (2.1cp at 10g/dL),12 but similar to that of PolybHb (11.4cp for 50:1 T-state PolybHb and 7.8cp for 40:1 R-state PolybHb at a concentration of 10g/dL).21 High-viscosity HBOCs are preferred for transfusion, since they can elicit the generation of endothelial-derived relaxing factors, through flow-mediated endothelial mechanotransduction.34 This can potentially neutralize the vasoconstrictive effect caused by NO scavenging or O2 oversupply upon PolyHb infusion, if it still exists. The COP of both PolyhHb solutions is less than the COP of whole blood (27mm Hg).12 However, R-state PolyhHbs exhibit COPs comparable to Hemopure® (25mm Hg) and PolyHeme® (23mm Hg).12 T-state PolyhHbs exhibit COPs comparable to PolybHb (1mm Hg for 50:1 T-state PolybHb and 7mm Hg for 40:1 R-state PolybHb).21 Low COP fluids can facilitate the outward flow of intravascular fluid across blood vessels into the tissue space, thereby reducing the blood volume. However, this problem could be remediated by supplementing human serum albumin to the low COP PolyhHb solution, thereby increasing the PolyhHb solution COP.35
Among all the HBOCs we studied via simulation, the O2 equilibrium curves of R-state PolyhHbs are left-shifted corresponding to low P50 values, whereas T-state PolyhHbs are right-shifted with correspondingly high P50 values, compared to native hHb. It is apparent from Figure 7A that all PolyhHb solutions have slightly less capacity for improving O2 transport at lower inlet pO2s (<40mm Hg) compared to native hHb. This is probably due to the low cooperativities and large diffusion coefficients of these molecules that are conferred upon polymerization. However, there is not much difference between all PolyhHb solutions and native hHb in the normalized O2 consumption rate at higher inlet pO2s, especially for T-state PolyhHbs, which possess higher P50s. This becomes more evident in the pO2 profile and ECS zonation plots (Fig. 7B and C). Generally, T-state PolyhHbs (high P50) enhance O2 transport to the ECS in the HF bioreactor. Under the simulated conditions (inlet pO2=80mm Hg), T-state PolyhHbs can release a considerable amount of O2 to the HF ECS especially as the local pO2 drops below the P50 of the HBOC. However, for R-state PolyhHbs, which have very high O2 affinities, the pO2 needs to be below 1mm Hg to make these HBOCs offload their store of O2. At a pO2 below 1mm Hg, hepatocytes will suffer from extensive hypoxia. In this case, R-state PolyhHbs perform like myoglobin which primarily stores O2 and only releases it under extreme hypoxia in the tissue. Therefore, R-state PolyhHbs are not suitable for tissue engineering applications where it is important to recapitulate in vivo O2 levels and gradients.
We have demonstrated that high-MW PolyhHbs can be synthesized in a defined quaternary state without the presence of free hHb and low-MW PolyhHb species in solution. Our results show that the PolyhHb MW and O2 affinity can be easily regulated by controlling the glutaraldehyde cross-link density and hHb quaternary state. O2 transport simulations indicate that T-state PolyhHbs oxygenated the ECS of a HF bioreactor to a greater extent versus R-state PolyhHbs. Taken together, these results support the use of PolyhHbs as efficacious O2 carrying solutions with possible applications in transfusion medicine and tissue engineering.
This work was supported by National Institutes of Health Grants R01HL078840 and R01DK070862 to A.F.P. The authors would like to thank David R. Harris for purifying the hHb for these studies.
The findings and conclusions in this article have not been formally disseminated by the Food and Drug Administration and should not be construed to represent any agency determination or policy.
No competing financial interests exist.