buehler 2007 effects of endogenou

Effects of Endogenous Ascorbate on Oxidation, Oxygenation,and Toxicokinetics of Cell-Free Modified Hemoglobin afterExchange Transfusion in Rat and Guinea Pig

Paul W. Buehler, Felice D’Agnillo, Victoria Hoffman, and Abdu I. AlayashLaboratory of Biochemistry and Vascular Biology, Division of Hematology, Center for Biologics Evaluation and Research,Food and Drug Administration, Bethesda, Maryland (P.W.B., F.D., A.I.A.); and Division of Veterinary Services, Office of theCenter Director, National Institutes of Health, Bethesda, Maryland (V.H.)

Received May 30, 2007; accepted July 9, 2007

ABSTRACTChemically modified hemoglobin (Hb) solutions are promisingoxygen therapeutics; however, these agents are prone to intra-vascular oxidation. Using a 50% exchange transfusion (ET)model with bovine polymerized hemoglobin (PolyHbBv), weexamined heme oxidation, oxygenation markers, and toxicoki-netics in rats, an ascorbic acid (AA)-producing species, and inguinea pigs, a non-AA-producing species. Plasma AA de-creased by 50% in guinea pigs after ET, but it was unchangedin rats for the first 20 h post-ET. Both species cleared PolyHbBvfrom the circulation at similar rates. However, exposure to ferricPolyHbBv over time was 5-fold greater in the guinea pig. Massspectrometry analysis of plasma revealed oxidative modifica-tions within the tetrameric fraction of PolyHbBv in guinea pig.Oxygen equilibrium curves of PolyHbBv measured in plasma

after ET were more left-shifted in guinea pigs compared withrats, consistent with increased ferric PolyHbBv formation. Re-nal hypoxia-inducible factor (HIF)-1�, whose activity strictlydepends on the partial pressure of oxygen increased over time,and it correlated inversely with circulating ferrous PolyHbBv inboth species. Interestingly, HIF-1� activity was greater inguinea pigs compared with rats at 72 h post-ET. Mean arterialpressure increases were also greater in guinea pigs; however,minimal differences in cardiac and renal pathology were ob-served in either species. The present findings suggest theimportance of plasma AA in maintaining the stability of acellularHb susceptible to oxidation, and they may be relevant to hu-mans, which display a similar plasma/tissue antioxidant statusto guinea pig.

Hemoglobin (Hb)-based oxygen carriers (HBOCs) repre-sent a class of complex biological entities being developed asoxygen-bridging agents with volume-expanding properties.Despite their therapeutic promise, HBOCs demonstrate asignificant potential for toxicity based on administration oflarge quantities of Hb into the plasma compartment. Nor-mally Hb remains protected in the red blood cell, whereprocesses exist to reduce oxidized Hb and to modulate nitricoxide (NO) binding. It has become increasingly evident that

Hb oxidative toxicity can limit the safety and efficacy ofcurrent generation HBOCs (Alayash, 2004). This promptedthe design of new strategies aimed at reducing or controllingHb-oxidative side reactions. In vivo oxidation of cell-free Hbis driven spontaneously and/or chemically by variety of oxi-dants, including hydrogen peroxide (H2O2) and NO. NO-induced oxidation of heme iron has the added complication ofproducing an immediate elevation in blood pressure as aresult of removal of NO (a vasodilator) by Hb. Thus, twoprimary safety concerns with HBOCs in the extracellularspace include hypertension and oxidative stress (Riess, 2001;Alayash, 2004). The latter effect depends on plasma andtissue reductive capacity to maintain the HBOC in a reducedand functional state.

HBOCs have generally demonstrated promising safety andefficacy in animals, and, in many cases, favorable phase I

The findings and conclusions in this article have not been formally dissem-inated by the Food and Drug Administration, and they should not be construedto represent any agency determination or policy. This work was funded by aCenter for Biologics Evaluation and Research/Food and Drug AdministrationUnmet Needs Award.

Article, publication date, and citation information can be found athttp://jpet.aspetjournals.org.

doi:10.1124/jpet.107.126409.

ABBREVIATIONS: HBOC, hemoglobin-based oxygen carrier; AA, ascorbic acid; LGO, L-gulonolactone oxidase; DHA, dehydroascorbate;PolyHbBv, bovine polymerized hemoglobin; PE, polyethylene; MAP, mean arterial pressure; ET, exchange transfusion; Hct, hematocrit; HIF,hypoxia-inducible factor; MALDI-MS, matrix-assisted laser desorption ionization/mass spectrometry; RBC, red blood cell; PK, pharmacokinetic;AUC, area under the plasma-concentration time curve; CL, plasma clearance; MRT, mean residence time; Vss, apparent volume of distribution atsteady state; OEC, oxygen equilibrium curve; Clast, last measurable concentration; SEC, size exclusion chromatography; BL, baseline; PEG,polyethylene glycol.

0022-3565/07/3231-49–60THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 323, No. 1U.S. Government work not protected by U.S. copyright 126409/3254180JPET 323:49–60, 2007 Printed in U.S.A.

49

at Sem

melw

eis University on F

ebruary 2, 2013jpet.aspetjournals.org

Dow

nloaded from

http://jpet.aspetjournals.org/

clinical safety (Przybelski et al., 1996; Carmichael et al.,2000). However, recent well publicized late-phase clinicaltrial failures with certain HBOCs suggest that preclinicalanimal testing may not have been sufficiently predictive ofsafety in humans (Sloan et al., 1999; Kim and Greenburg,2004). This recognition has highlighted the need for im-proved animal models and biomarkers to better predict andmonitor the safety and efficacy of HBOCs in humans. In thisregard, the selection of animal species to evaluate HBOCs iscritical for predicting safety in both normal and disease-statehuman subjects.

In the present study, we compare the rat, recognizing itspopularity as a small animal species used in nonclinicalHBOC studies, and the guinea pig as a potentially morerelevant small animal species for predicting human safety.The rationale for evaluating the guinea pig is based on sev-eral factors that suggest a similarity with humans in terms ofoverall antioxidant status. Guinea pigs, like humans, areincapable of endogenous production of ascorbic acid (AA) dueto the evolutionary loss of functional hepatic L-gulonolactoneoxidase (LGO). In contrast, rats generate 38 �g AA/mg pro-tein/h under normal conditions and even greater amountswhen subjected to stress (Chatterjee, 1973). Interestingly,extensive investigations of several HBOCs in nonclinicalstudies often involving AA-producing species, such as rat andswine, have demonstrated acceptable safety profiles.

Guinea pigs and humans seem to have compensated for thelack of endogenous AA production by increasing tissue anti-oxidant enzyme content and efficiency. For example, copperand zinc superoxide dismutase enzymatic activity in kidneyand liver is approximately 2-fold higher in humans andguinea pigs compared with rats (Nandi et al., 1997). Morerecently human and guinea pig but not rat red blood cellshave been found to possess a similar oxidoreductase system,cytochrome b561, responsible for efficiently recycling dehy-droascorbate (DHA) to functional AA (Su et al., 2006). There-fore, the general antioxidant status of guinea pig and humanare quite similar so that the balance of reductive capacity istilted toward the tissue and away from the plasma in guineapigs and humans. Germane to these similarities is thatplasma antioxidant capacity is necessary for maintainingHBOCs in a reduced form so that they are able to carry anddeliver oxygen efficiently, maintain heme stability, and limittoxicity. This is especially important when tissue antioxidantstatus is diminished during hemorrhagic shock and ischemicconditions.

In the current study, rats and guinea pigs were subjectedto 50% exchange transfusion with bovine polymerized hemo-globin (oxyglobin; PolyHbBv), a Food and Drug Administra-tion-approved HBOC for veterinary use. We identified astrong correlation between plasma AA levels and the suscep-tibility of PolyHbBv to undergo oxidation, and we correlatedthese events for the first time with tissue oxygen-sensingmechanisms. Despite dramatic interspecies differences inplasma PolyHbBv oxidation and stability, minimal differ-ences were found comparing the overall effect on cardiac andrenal pathology, which may reflect differences in local tissueantioxidative mechanisms. These data are consistent withthe general safety of certain HBOCs in normal human sub-jects treated in early phase clinical trials. However, moreextensive work using the guinea pig in models of tissuecompromise (e.g., ischemia/reperfusion) when tissue oxida-

tive status is reduced may ultimately provide answers to whycertain HBOCs fail in pivotal clinical trials.

Materials and MethodsSolutions. Oxyglobin (PolyHbBv) was purchased from Biopure

Corporation (Cambridge, MA). This solution consists of a heteroge-neous mixture of glutaraldehyde-polymerized bovine hemoglobin ata concentration of 13 g/dl in modified lactated Ringer’s. A detaileddescription of the physicochemical properties of the mixture as awhole and of each individual fraction has been described previously(Alayash et al., 2001; Buehler et al., 2005).

Animals and Surgical Preparation. Male Sprague-Dawley ratsand Hartley guinea pigs were purchased from Charles River Labo-ratories, Inc. (Wilmington, MA), and they were acclimated for 1 weekupon arrival to the Food and Drug Administration’s Center for Bio-logics Evaluation and Research animal care facility (Bethesda, MD).All animals were fed normal diets throughout the acclimation period,and they weighed 350 to 450 g at the time of study. Animal protocolsfor each species were approved by the Food and Drug Administra-tion’s Center for Biologics Evaluation and Research InstitutionalAnimal Care and Use Committee with all experimental proceduresperformed in adherence to the National Institutes of Health guide-lines on the use of experimental animals (Institute of LaboratoryAnimal Resources, 1996).

On days of surgery, rats and guinea pigs were anesthetized via thei.p. route with a cocktail of 100 mg/kg ketamine HCl and 5 mg/kgxylazine HCl (Phoenix Scientific Inc., St. Joseph, MO). Under asepticconditions, a midline incision was made around the neck region,allowing for blunt dissection and exposure of the right commoncarotid artery and the left external jugular vein. Saline-filled cath-eters containing 50 IU of heparin per ml prepared from sterile PE50tubing (Clay Adams, Parsippany, NJ) were placed in each vessel, andthey were tunneled under the skin to the back of the neck. Immedi-ately following surgeries, animals were administered a subcutane-ous dose of buprenorphine (0.1 mg/kg) (Reckitt and Coleman Corp.,Kingston, UK), and they were allowed 24 h of recovery before exper-imentation. For blood pressure measurements, the right femoralartery was also catheterized with PE10 tubing fused to PE50 tubingto allow for simultaneous administration of PolyHbBv and monitor-ing of arterial blood pressure. The right carotid artery catheter wasconnected to a Gould P23 XL pressure transducer (Gould InstrumentSystems Inc., Valley View, OH) for recording blood pressure. Arterialblood pressure was recorded continuously at 100 Hz using anMP100A-CE data acquisition system (Biopac Systems, Inc., SantaBarbara, CA). Data were analyzed off-line using AcqKnowledge soft-ware (Biopac Systems, Inc.) to determine mean arterial pressure(MAP) from the following formula: [diastolic � 1/3(systolic � dia-stolic)], with each being averaged over consecutive minutes of ac-quired data.

Blood/Tissue Collection. Blood (350 �l) was sampled (beforeinfusion), immediately after infusion, and at 4, 12, 24, 48, and 72 hafter end of exchange transfusion (ET) for the following analyses: 1)plasma AA, 2) hematocrit (Hct), and 3) plasma oxygen equilibriumvalues. Heart and kidneys were harvested at the same time pointsfor histopathology and renal tissue hypoxia-inducible factor (HIF)activity. Each time point represents an n � 3 to 5 animals.

In a separate group of animals (n � 5–6), blood samples (0.2 ml)were obtained from the arterial catheter before infusion (baseline)and at the end of ET (time 0), and at 0.25, 0.5, 1, 1.5, 2, 2.5, 3, 4, 8,12, 24, 48, and 72 h after the end of ET. Plasma was used forevaluation of 1) total PolyHbBv, 2) ferrous PolyHbBv, 3) ferric Poly-HbBv, and 4) PolyHbBv polymeric component distribution.

Plasma Ascorbic Acid Analysis. Plasma AA was analyzed by amodified reverse-phase high-performance liquid chromatographymethod. In brief, plasma samples (200 �l) were centrifuge-filtered at12,000 rpm for 45 min at 4°C using Microcon YM-10 filter tubes(Millipore Corporation, Bedford, MA) to remove PolyHbBv and other

50 Buehler et al.

at Sem

melw

eis University on F


Dow

nloaded from


proteins greater than 10,000 Da. Standard curves and controls wereprepared in normal saline using AA (Sigma-Aldrich, St. Louis, MO)and acetaminophen (Sigma-Aldrich) as the internal standard. Allstandards and controls were subjected to the same filtration steps.Total AA (oxidized and reduced AA) was evaluated by adding 5 �l ofa 100 mM solution of dithiothreitol (Sigma-Aldrich) to each filteredsample to reduce any DHA. Standards and samples (50 �l) were runon a PrimeSep D (4.6 � 100 mm; 5 �m) reverse-phase column witha PrimeSep D (4.6 � 10 mm) guard column (SIELC Technologies,Inc., Prospect Heights, IL) attached to a Dionex Summit P680 pumpwith a Dionex UVD 170S detector (Dionex Corp., Sunnyvale, CA).The mobile phase consisted of 40% acetonitrile /1% acetic acid, and itwas pumped at a rate of 1 ml/min. Absorbance was measured at 254nm for detection of AA and acetaminophen.

Pharmacokinetic/Toxicokinetic Analysis. Fully conscious andfreely moving rats (n � 6) and guinea pigs (n � 5) underwent a 50% ET,replacing blood with PolyHbBv. Arterial and venous catheters wereextended, tethered, and connected to separate syringe pumps (model11; Harvard Apparatus Inc., Holliston, MA) set on withdrawal (1 ml/min) and infuse (1 ml/min), respectively. The 50% ET volume in the ratwas calculated as 50% ET (milliliters) � [0.06 (milliliters per gram) �body weight (grams) � 0.77]/2 (Lee and Blaufox, 1985) and as 50% ET(milliliters) � [0.07 (milliliters per gram) � body weight (grams)]/2 inthe guinea pig (Ancill, 1956). Plasma from blood in the heparinizedwithdrawal syringe for each transfused animal was obtained to deter-mine the total PolyHbBv removed during the exchange transfusionperiod (approximately 12 min). Blood samples (0.2 ml) were obtainedfrom the arterial catheter before infusion (baseline) and at the end of ET(time 0), and at 0.25, 0.5, 1, 1.5, 2, 2.5, 3, 4, 8, 12, 24, 48 and 72 h afterthe end of ET.

Plasma concentrations of ferrous PolyHbBv and ferric PolyHbBvwere determined using a photodiode array spectrophotometer (model8453; Hewlet Packard, Palo Alto, CA). Plasma from the baseline(pre-ET) sample for each animal was used to correct for backgroundinterference and turbidity. Concentrations of ferrous PolyHbBv (oxy/deoxy), ferric PolyHbBv, and hemachrome were determined using amulticomponent analysis based the extinction coefficients for eachHb species (Winterbourn, 1985). Electron paramagnetic resonancewas also used to confirm that ferric PolyHbBv in the circulation wasprimarily in its nonliganded state as opposed to partially liganded orfully liganded (e.g., Fe3�-NO) forms, which are indistinguishable byUV-visible spectrophotometry (data not shown).

Pharmacokinetic Analysis. The dose (milligrams) of PolyHbBvreceived by each animal at the end of ET was determined by sub-tracting the total amount of PolyHbBv in the plasma from wholeblood collected in the ET syringe from the total amount of infusedPolyHbBv according to the following equation: dosereceived � ([Poly-HbBv]inf � Vinf) � ([PolyHbBv]totalET � VET, where dose[PolyHbBv]inf is the concentration of PolyHbBv (milligrams per mil-liliter) infused, Vinf is the PolyHbBv infusion volume (milliliters),[PolyHbBv]totalET is the concentration of PolyHbBv (milligrams permilliliter) from plasma sampled out of the withdrawal syringe, andVET is the volume (milliliters) collected in the withdrawal syringe.MALDI-MS was performed to confirm that no Hb originating fromred blood cells (RBCs) was present in the withdraw syringe sampledplasma. MALDI-MS was also performed on both rat and guinea pigplasma to explore stability of circulating PolyHbBv fractions asdescribed under Results (Fig. 4).

PK parameters were determined for total PolyHbBv, ferrous Poly-HbBv (oxy/deoxy), and ferric PolyHbBv. noncompartmental methodsin WinNonlin, version 4.1 (Pharsight, Mountain View, CA) were usedto calculate PK parameter estimates. The area under the plasma-concentration time curve (AUC)0–� was estimated using the lineartrapezoidal rule to the last measurable concentration (AUC0–C last).Extrapolation to infinity (AUCC last–�) was accomplished by dividinglast measurable concentration (C last) by the negative value of theterminal slope (k) of the log-linear plasma concentration-time curve.Thus, AUC0–� is equal to the sum of AUC0–C last and AUCC last–�.

Additional parameters were calculated as follows: plasma clearance(CL) as dose divided by AUC0–�, mean residence time (MRT) as k�1,apparent volume of distribution (Vss) as the product of CL and MRT,and half-life (t1/2) as the product of ln2 and MRT.

Plasma Oxygen Equilibrium Curve Measurements. Oxygenequilibrium curves (OECs) were obtained using a Hemox analyzer(TCS Scientific, New Hope, PA). Experiments were carried out in 0.1M phosphate buffer/0.1 M NaCl at pH 7.4, and the temperature wasmaintained at 37°C. To prevent formation of ferric PolyHbBv, 4 �l ofthe Hayashi enzymatic reduction system was added to the 4-mlsolution (Hayashi et al., 1973). Oxygen equilibrium parameters werederived by fitting the Adair equations to each OEC by the nonlinearleast-squares procedure included in the Hemox analyzer software(p50 PLUS, version 1.2; TCS Scientific) (Alayash et al., 2001). TheAdair constants were then used to generate an OEC curve to gener-ate the P50 and n50 (Hill coefficient) for oxygen binding. Therefore,the procedure made it possible to measure the oxygen binding pa-rameters of PolyHbBv, which is not fully saturated at atmosphericoxygen partial pressures (Nagababu E et al., 2002).

PolyHbBv Polymer Distribution and Globin Chain Dissoci-ation in Plasma. Plasma samples (50 �l) were evaluated by sizeexclusion chromatography (SEC) to compare distribution of HbGpolymeric components at time points post-ET. Samples were run ona BioSep-SEC-S3000 (600 � 7.5 mm) SEC column (Phenomenex,Torrance, CA) attached to a Waters 626 pump and Waters 2487dual-wavelength detector, controlled by a Waters 600s controllerusing Millenium32 software (Waters, Milford, MA). The runningbuffer consisted of 0.1 M NaH2PO4, pH 6.5, pumped at rate of 0.5ml/min, and the absorbance was monitored at 405 nm.

Plasma samples (10 �l) were desalted using C18 ZipTips (Milli-pore) according to the manufacturer’s instructions. A 1-�l aliquotwas pipetted onto a stainless steel MALDI-MS sample plate andmixed with 1 �l of 3,5-dimethoxy-4-hydroxycinnamic acid (sinapinicacid) (Sigma-Aldrich) saturated in 50% acetonitrile/0.1% trifluoro-acetic acid. The sample/matrix was air-dried and analyzed on aPerSeptive Biosystems DERP MALDI-time of flight mass spectrom-eter calibrated manually with purified human serum albumin (Sig-ma-Aldrich) and using Voyager 5.1 software with Data Explorer(Applied Biosystems, Framingham, MA) operated in linear mode.

Blood Pressure. Fully conscious and freely moving rats (n �5–6) and guinea pigs (n � 5–6) underwent a 50% ET replacing bloodwith PolyHbBv (13 g/dl). Withdrawal of blood and infusion of Poly-HbBv took place simultaneously at a rate of 1 ml/min from the rightfemoral artery and the left external jugular vein, respectively. Theright carotid artery catheter was connected to a Gould P23 XLpressure transducer (Gould Instrument Systems Inc.) for recordingblood pressure. Arterial blood pressure was recorded continuously at100 Hz using an MP100A-CE data acquisition system (Biopac Sys-tems, Inc.). Data were analyzed off-line using AcqKnowledge soft-ware (Biopac Systems, Inc.) to determine MAP from the followingformula: [diastolic � 1/3(systolic � diastolic)], with each being aver-aged over consecutive minutes of acquired data.

Hypoxia-Inducible Factor-1� DNA Binding Assay. HIF-1�

DNA binding activity of animal tissues was determined by using anenzyme-linked immunosorbent assay-based method and kit (Trans-binding HIF-1� Assay kit; Panomics, Redwood City, CA) according tomanufacturer’s specifications. HIF-1� transcriptional factor was as-sessed by detection of its binding to an oligonucleotide containing thehypoxia-responsive element (5�-TACGTGCT-3�) after lysis, and nu-clear extraction was achieved with the use of a nuclear extraction kitprovided by the manufacturer (Panomics). HIF-1� binding was de-tected by a mouse antibody directed against HIF-1� region availableafter DNA binding and revealed by anti-mouse IgG coupled to horse-radish peroxidases, which provides sensitive colorimetric detectionby a spectrophotometric microplate reader at 450 nm.

Histopathology. Hearts and kidneys were fixed in 10% formalin,embedded in paraffin, and then 5-�m sections were cut and stainedby standard hematoxylin and eosin procedures. Tissues were scored

Plasma Reductive Capacity and Hb Oxygen and Redox Kinetics 51

at Sem

melw

eis University on F


Dow

nloaded from


by a veterinary pathologist using a semiquantitative grading systemas follows: 0, minimal; 1, mild; 2, moderate; and 3, severe (minimalindicates a detectable process but barely present, mild indicatessmall aggregates of inflammatory or necrotic cells, moderate indi-cates large aggregated inflammatory or necrotic cells at 20�, andsevere indicates large multifocal aggregates making up greater than50% of the tissue area).

Statistical Analysis. Arterial blood pressures, heart rate, non-compartmental analyses of PK parameter estimates, hematocrit,and plasma ascorbate concentrations are expressed as means �S.E.M. Statistical comparison for arterial blood pressures, heartrate, PK parameter estimates, hematocrit, and plasma AA concen-trations were performed by analysis of variance with an a priori testfor planned comparisons between the rat and the guinea pig. In allanalyses, a value p � 0.05 was taken as the level of statisticalsignificance.

ResultsPlasma AA Levels following Exchange Transfusion.

Plasma reductants such as AA have been shown to effectivelyreduce oxidized Hb in humans (Ames et al., 1981). Figure 1Ashows plasma AA concentrations in rats and guinea pigsfollowing 50% ET. Baseline (BL) plasma AA concentrations

in the rat and guinea pig were 100.03 � 16.2 and 51 � 15.8�M, respectively. When subjected to 50% ET, rats exhibitedplasma AA concentrations similar to BL at 4 h (97.2 � 5.2�M), 12 h (104.6 � 7.0 �M), and 24 h (109.4 � 21 �M)post-ET. At 48 h post-ET, plasma AA concentration de-creased below BL (i.e., 51.6 � 13 �M), and this AA concen-tration was maintained at 72 h post-ET (50.8 � 17 �M). Anopposite situation occurred in guinea pigs following 50% ETsuch that AA plasma concentrations were approximately50% BL at 4 h (22 � 5.7 �M), 30% BL at 12 h (15.9 � 3.4 �M),36% BL at 24 h (18.5 � 5.8 �M), 41% BL at 48 h (20.9 � 3.1�M), and 96% BL at 72 h (48.1 � 16 �M) post-ET. These datasuggest that the rat rapidly up-regulates AA hepatic produc-tion to make up for losses during ET, whereas in the guineapig, a 50% loss of AA plasma concentration is observed 4 hafter 50% ET, and it remained low, which is consistent witha 50% exchange transfusion in a species incapable of endog-enous AA production.

Efficient AA functioning is dependent on the presence ofnormal circulating RBCs that chemically recycle DHA backto functional AA (May et al., 2004). Given this intimaterelationship between AA and RBCs, Hct levels in the twospecies were also measured (Fig. 1B). Baseline Hct levelswere 44.4 � 1.5 and 37.8 � 0.85% in the rat and guinea pig,respectively, whereas ET reduced Hct levels to 21.9 � 0.92and 18.3 � 0.47 in each species. Over the 72 h of evaluation,neither species demonstrated changes in Hct from end ETlevels. These results indicate that the exchange transfusionresulted in similar reductions in red cell volume in bothspecies; thus, red cell volume cannot account for the inter-species differences in plasma AA.

Ferric PolyHbBv Formation following ExchangeTransfusion. To examine the possible correlation amongplasma AA levels, oxidation, and oxidative stability ofPolyHbBv, we measured the postexchange levels of ferricPolyHbBv (Fig. 2). The Cmax for ferric PolyHbBv occurred attime 0 (end of ET) in the rat and at 12 h post-ET in the guineapig, accounting for 2.2 and 23.2% of the total PolyHbBvconcentration in each species at each Tmax, respectively. TheAUC0–� for ferric PolyHbBv in the rat and guinea pig was 6.6and 34.5% of the total PolyHbBv AUC0–�, respectively.Plasma CL of ferrous PolyHbBv occurred faster in the guineapig compared with the rat; however, no such difference wasfound in total PolyHbBv plasma CL.

The critical period of ferric PolyHbBv exposure in theguinea pig seems to exist between the 4- and 48-h post-ETtime points when total PolyHbBv plasma concentrations are19.5 and 8.0 g/dl, respectively. Between these time points,the percentage of ferric PolyHbBv in the plasma ranged from10.7% (4 h) to 40% (36 h). By 72 h post-ET, the plasmaconcentrations of total PolyHbBv were less than 5% of theCmax; thus, the percentage of total PolyHbBv as ferric Poly-HbBv becomes small and of minimal relevance. In the rat,percentage of ferric PolyHbBv is �5% of the total PolyHbBvconcentration between the end of ET until 24 h post-ET. At48 h post-ET, total PolyHbBv is approximately 0.4 g/dl, with18% of this concentration as ferric PolyHbBv. By 72 h post-ET, the plasma concentration of total PolyHbBv is less than5% of the Cmax; thus, the percentage of total PolyHbBv as theoxidized (ferric) form is negligible. These data suggest thatoxidative processes are probably responsible for the greaterformation of plasma ferric PolyHbBv in guinea pigs rather

Fig. 1. Plasma ascorbic acid and hematocrit in rat and guinea pig follow-ing PolyHbBv exchange transfusion. A, Pre and postexchange transfu-sion plasma ascorbic acid concentrations in the rat (E) and guinea pig (F).B, pre and postexchange transfusion hematocrit in the rat (filled bars)and the guinea pig (open bars). Values are reported as micromolar �S.E.M. Significant differences (p � 0.05) between rat and guinea pig aredenoted by ‡, whereas significant differences from baseline values aredenoted by * (n � 5–6 animals/species/time point).

52 Buehler et al.

at Sem

melw

eis University on F


Dow

nloaded from


than increased whole body CL of ferrous PolyHbBv. More-over, these findings identify the rat as a species that canprevent the physiological accumulation of ferric PolyHbBvwithin the plasma compartment, whereas the guinea pig maymore accurately reflect plasma accumulation of ferric Poly-HbBv in other non-AA producing species (i.e., humans).

Pharmacokinetic Analysis. Table 1 shows the pharma-cokinetic parameter estimates for total PolyHbBv, ferrousPolyHbBv, and ferric PolyHbBv following a 50% PolyHbBvfor blood ET. Rats and guinea pigs received 3122 � 55 and3272 � 106 mg/kg doses of PolyHbBv, respectively. Thisdosing level is representative of a high single administrationof PolyHbBv equal to 7 U in a 70-kg adult human or 220 g ofcell-free PolyHbBv. At time 0 (end of ET), plasma concentra-tions of PolyHbBv in both species were made up of 94%ferrous PolyHbBv and 6% ferric PolyHbBv. Figure 2 shows

the data used to define noncompartmental parameter esti-mates plotted as log mean � S.E.M. (�5% mean values)plasma concentration versus time. The Vss for PolyHbBv wasapproximately 2 to 2.5 times greater in each species than theexpected circulating blood volume following a 50% ET withan oncotically matched solution (i.e., colloid osmotic pressureequal to blood). Although not directly measured in this study,the data suggest an increase in circulating vascular volumeand a rapid saturating nonspecific tissue distribution. Inboth species, increased urination and hematuria were ob-served during the 50% ET and in the initial 4 h of observationpost-ET. The total PolyHbBv estimates for t1/2 and the MRTwere similar between species. The primary exposure param-eters (Cmax and AUC0–�) and total CL demonstrated speciessimilarity for total PolyHbBv following dosing in the rat andguinea pig. However, when the oxidation state of PolyHbBvwas accounted for in the PK analysis, ferrous PolyHbBv andferric PolyHbBv exposure parameter differences between therat and guinea pig became evident.

Plasma PolyHbBv Component Plasma Distribution/Globin Chain Dissociation. The SEC chromatographs ofplasma from rat and guinea pig obtained from blood samplesat the end of ET until 72 h post-ET were separated on aBioSep-SEC-S3000 (600 � 7.5 mm) SEC column (Phenome-nex) are shown in Fig. 3, A and B. As expected the chromato-grams show a more rapid loss of tetramer in each species anda slower disappearance of multitetrameric PolyHbBv compo-nents. PolyHbBv tetramer and multitetrameric species seemto be similarly eliminated from the plasma at time pointspost-ET. Alternatively, larger PolyHbBv multitetramers maybreak down to smaller mol. wt. components before elimina-tion. However, upon closer inspection of PolyHbBv elimina-tion curves and SEC chromatograms obtained from guineapig plasma, both PolyHbBv tetrameric and multitetramericfractions seem to be eliminated more rapidly in the first 4 hpost-ET. This observation may be due to the acceleratedoxidation of PolyHbBv in guinea pigs; thus, increased elimi-nation may serve as an early protective mechanism removingoxidized PolyHbBv.

Structural and Oxidative Stability of PolyHbBv. Al-though tetramer stability is not evident from the SEC chro-matograms, MALDI-MS analysis of plasma collected at 4 and24 h suggests that tetramers of PolyHbBv remain stable inthe rat, but they become unstable in the guinea pig circula-tion (representative spectra shown in Fig. 4, A and B). In Fig.4A, the MALDI-MS spectra of rat RBC Hb demonstrates an� globin chain [M � H]1� ion at m/z 15201.60 and a � globinchain [M � H]1� ion at m/z 15851.51. These ions predomi-nate even though rat Hb exists as multiple � and � globinchain variants (Garrick et al., 1975). At 4 and 24 h post-ET,the cross-linked PolyHbBv derived � globin chains are seenas an �*-�* [M � H]1� ion with m/z 30940.17, whereasresidual �* [M � H]1� (limited percentage of intensity) andintense �* [M � H]1� globin chain ions are observed at m/z15285.00 and 16108.70, respectively. The cross-linked � glo-bin chain ions suggest nonsite-specific modification by glu-taraldehyde (Buehler et al., 2005). The �* [M � H]1� ionregion of the 24-h plasma sample indicates no peak intensitychange from the 4-h sample, suggesting limited tetramerdestabilization in vivo.

In Fig. 4B, the MALDI-MS spectra of guinea pig RBC Hbdemonstrates an � globin chain [M � H]1� ion at m/z

Fig. 2. Log-linear plasma concentration versus time profiles of PolyHbBvin rats and guinea pigs. Plasma concentrations of total PolyHbBv (f),ferrous PolyHbBv (F), and percentage of plasma ferric PolyHbBv (E) attime points from the end of exchange transfusion until 72 h in the rat (A)and guinea pig (B). Convergence of dotted and dashed lines indicates theperiod of greatest ferric PolyHbBv exposure. The pharmacokinetic esti-mates derived from these data are shown in Table 1.


at Sem

melw

eis University on F


Dow

nloaded from


15228.59 and � globin chain [M � H]1� ion at m/z 15921.00.The � and � globin chain ions are consistent with thosereported in the literature (Day et al., 1996). At 4 and 24 h

post-ET, the cross-linked PolyHbBv-derived � globin chainsare seen as an �*-�* [M � H]1� ion at m/z 31025.19, whereasresidual �* [M � H]1� (limited percentage of intensity) andintense �* [M � H]1� globin chain ions are observed at m/z15250.81 and 16125.87, respectively. The �* [M � H]1� ion atm/z 15250.81 in the spectra of the 24-h plasma sample indi-cates increased peak intensity of 35% from the 4-h sample,suggesting � globin chain cross-link destabilization in vivo.Rat and guinea pig plasma samples did not demonstratemass ions for � or � globin chains of Hb originating from rats’or guinea pigs’ own RBCs. Thus, destabilized � globin couldonly have come from PolyHbBv.

Blood Pressure and Heart Rate Response to Poly-HbBv. HBOC-mediated blood pressure elevation is a com-mon finding in animals and humans, and it can result inreduced tissue oxygenation. To investigate the possiblerelationship between PolyHbBv oxidative stability and thesusceptibility and/or extent of hemodynamic alterations,we monitored blood pressure and heart rate in fully con-scious and freely moving rats and guinea pigs (n � 5 forboth species) during the course of the 50% ET. Figure 5Ashows the changes in MAP between rat and guinea pig.Baseline blood pressure values monitored over 20 min inrats ranged as follows: systolic (140.4 � 6.44 –143.8 � 6.40mm Hg), diastolic (109.6 � 4.93–125.2 � 6.22 mm Hg),MAP (119.0 � 5.81–125.2 � 6.22 mm Hg), and pulse pres-sure (23.8 � 4.41–27.4 � 2.11 mm Hg). In guinea pigs,baseline blood pressure values ranged as follows: systolic(73.4 � 3.60 – 83.2 � 5.35 mm Hg), diastolic (48.8 � 4.85–56.2 � 5.76 mm Hg), MAP (60.2 � 4.14 – 63.0 � 4.10 mmHg), and pulse pressure (23.8 � 4.41–27.4 � 2.11 mm Hg).Immediately after the start of ET, a concomitant drop inheart rate occurred (Fig. 5B).

The onset of ET resulted in immediate responses in allhemodynamic parameters measured in both species. The endof the 50% ET (approximately 10 min) resulted in the follow-ing blood pressure changes in the rat: systolic, 157.4 � 6.83mm Hg (12% BL); diastolic, 115.6 � 4.3 mm Hg (6% BL);MAP, 136.4 � 7.10 mm Hg (11% BL); and pulse pressure,41.8 � 3.3 mm Hg (75% BL). The overall arterial bloodpressure elevation (systolic, diastolic, and MAP) was greaterin the guinea pig versus the rat. Interestingly the increase inMAP (diastolic � [(1/3) systolic � diastolic]) in the rat was

TABLE 1Pharmacokinetic parameter estimates following PolyHbBv transfusion

Estimate Total PolyHbBv Ferrous PolyHbBv Ferric PolyHbBv

RatDose (mg) 1197 � 104 (3122 � 55.0 mg/kg) 1112 � 97.9 (2900 � 55.0mg/kg) 84.58 � 6.03 (222 � 0.00 mg/kg)Cmax (mg � ml�1) 42.4 � 1.65 41.5 � 1.55 0.94 � 0.4Tmax (h) 0.20 � 0.014 0.20 � 0.014 0.20 � 0.014AUC0–� (mg � h � ml�1) 824.3 � 32.6 774.2 � 42.7 54.01 � 10.8CL (ml � h�1) 1.45 � 0.0994 1.55 � 0.115Vss (ml) 31.7 � 7.97 37.9 � 16.0t1/2 (h) 15.6 � 1.92 13.4 � 2.14MRTi.v. (h) 21.1 � 4.29 18.9 � 4.69

Guinea pigDose (mg) 1186 � 49.5 (3272 � 106 mg/kg) 1117 � 38.2 (3143 � 70.1mg/kg) 63.32 � 5.19 (188.8 � 15.6mg/kg)Cmax (mg � ml�1) 40.0 � 2.22 39.4 � 2.22 4.10 � 1.0Tmax (h) 0.19 � 0.01 0.19 � 0.01 12AUC0–� (mg � h � ml�1) 788.1 � 90.6 512.8 � 67.5 272.0 � 53.0CL (ml � h�1) 1.86 � 0.31 2.43 � 0.31Vss (ml) 45.3 � 6.1 49.8 � 6.1t1/2 (h) 15.7 � 0.64 12.9 � 0.65MRTi.v. (h) 25.1 � 1.87 20.7 � 1.49

Fig. 3. Size-exclusion chromatography patterns of plasma PolyHbBv overtime in the rat and guinea pig. A, representative (single animal) patternof tetrameric and multitetrameric fractions in PolyHbBv as they exist inthe plasma of the rat over time. B, representative (single animal) patternof tetrameric and multitetrameric fractions in PolyHbBv as they exist inthe plasma of the guinea pig over time. In each case, plasma samples (50�l) were run on BioSep-SEC-S3000 (600 � 7.5 mm) SEC column. Therunning buffer consisted of 0.1 M NaH2PO4, pH 6.5, pumped at rate of 0.5ml/min and the absorbance was monitored at 405 nm.

54 Buehler et al.

at Sem

melw

eis University on F


Dow

nloaded from


predominantly driven by an elevation in systolic pressure,whereas diastolic pressure remained similar to baseline.

Plasma Oxygen Equilibrium Curve Analysis. The im-pact of increased plasma PolyHbBv oxidation on PolyHbBvoxygen-carrying capacity in both species was examined.OECs in plasma samples obtained over the course of 72 h areshown in Fig. 6, A and B, for rat and guinea pig, respectively.From plasma OECs, we were able to calculate a P50 � 36.4 �0.989 mm Hg immediately after the end of ET in the guineapig, compared with 41.7 � 0.73 mm Hg in the rat. At 48 hpost-ET, a greater shift toward increased oxygen affinity(reduced oxygen off-loading) was observed in the guinea pig(P50 � 25.1 � 0.441 mm Hg) compared with the rat (38.9 �

2.72 mm Hg). These observations are consistent with theincreased ferric PolyHbBv accumulation over time in guineapigs.

Modulation of HIF-1� Activity by PolyHbBv. As anindirect estimate of tissue oxygenation, we also analyzedHIF-1� activity in nuclear extracts obtained from kidneys ofeach species, and we examined its relationship with plasmaoxygen content as reflected by the presence of circulatingferrous PolyHbBv (oxy) as a function of time (Fig. 7). Therewas a clear inverse relationship between the levels of ferrousPolyHbBv (oxy) and HIF activity in both rat (Fig. 7A; r2 �098) and guinea pig (Fig. 7B; r2 � 0.87) in the first 50 h afterET with PolyHbBv. It is noteworthy that maximal suppres-

Fig. 4. A and B, MALDI-MS evaluation of HbG globin chain destabilization in plasma sampled from rat and guinea pig. Aa, MALDI-MS spectra ofrat red blood cell hemoglobin obtained from rats whole blood sampled at baseline. The [M � H]1� ion at m/z 15201.6 and m/z 15851.51 represent theprimary � and � globin chain ions in rat red cells and serves as a control for the presence of red cell hemoglobin, which could be present in plasmasamples (a and b). Ab, MALDI-MS spectra of rat plasma sampled at 4-h postexchange transfusion. The [M � H]1� ions denoted �* at m/z 15328.46,�* m/z 16110.63, and �*-�* m/z 30940.17, represent the tetrameric fraction of �, �, and �-� cross-linked species of PolyHbBv. Ac, MALDI-MS spectraof rat plasma sampled at 24-h postexchange transfusion. The [M � H]1� ions denoted �* at m/z 15285.00, �* m/z 16108.70, and �*-�* m/z 30932.26represent the tetrameric fraction of �, �, and �-� cross-linked species of PolyHbBv. The arrow indicates the [M � H]1� ion for albumin; this ion becomesmore abundant (less suppressed) as PolyHbBv is cleared from the plasma. Ba, MALDI-MS spectra of guinea pig red blood cell hemoglobin obtainedfrom guinea pig whole blood sampled at baseline. The [M � H]1� ions at m/z 15228.59 and m/z 15921.00 represent the primary � and � globin chainions in guinea pig red cell hemoglobin and serve as a control for the presence red cell hemoglobin, which could be present in plasma samples (a andb). Bb, MALDI-MS spectra of guinea pig plasma sampled at 4-h postexchange transfusion. The [M � H]1� ions denoted �* at m/z 15160.21, �* at m/z16093.84, and �*-�* at m/z 31025.1 represent the tetrameric fraction of PolyHbBv �, � (shown in blue in the full spectra and inset) and �-� cross-linkedspecies of PolyHbBv. Bc, MALDI-MS spectra of guinea pig plasma sampled at 24-h postexchange transfusion. The [M � H]1� ions denoted �* at m/z15250.81, �* at m/z 16125.87, and �*-�* at m/z 31009.68 represent the tetrameric fraction of �, � (shown in red in the full spectra and inset) and �-�cross-linked species of PolyHbBv. The increase in intensity of the �* ion could not be assigned to the guinea pigs own red cell Hb (m/z 15228.59); thus,the intensity increase is consistent with �-� cross-link destabilization. The arrow indicates the [M � H]1� ion for albumin, this ion becomes moreabundant (less suppressed) as PolyHbBv is cleared from the plasma.


at Sem

melw

eis University on F


Dow

nloaded from


sion of HIF activity was achieved over the first 10 h whenmaximum ferrous PolyHbBv (oxy) was observed in the circu-lation of both species. Interestingly, by 72 h post-ET, whenPolyHbBv (oxy) was cleared from the circulation, there wasapproximately a 2-fold increase in renal HIF activity in theguinea pig compared with the rat. Our data are consistentwith a recent report that showed a dramatic effect of in-creased oxygen affinity delivery by RBCs on angiogenesisand HIF-1�. In this report, RBCs were treated with myo-inositol trispyrophosphate (allosteric modifier of oxygen af-finity), which caused a 7 mm Hg drop in the RBC P50. Addi-tion of treated RBCs to hypoxia endothelial cells led to aconsiderable suppression in HIF-1� activity (5-fold) as wellas vascular endothelial growth factor (1.29-fold) (Kieda et al.,2006).

Heart and Kidney Histopathology. To investigatewhether the interspecies differences in PolyHbBv oxidation/oxygenation translated into differences in organ pathology,we performed hematoxylin and eosin staining and his-topathological analysis of heart (Fig. 8) and kidneys (Fig. 9)

obtained at baseline (24 h after surgery) and 24 and 72 hpost-ET. Tissues from each species (n � 3–5/treatment pertime point) were scored by a veterinary pathologist using asemiquantitative grading system such that mean � S.E.M.severity scores provide a general evaluation of tissue inflam-mation and necrosis. Rat and guinea pig heart (left ventricle)and kidneys exhibited minimal (0) to mild (1) pathologicalresponse to PolyHbBv at 24 h post-ET in rat and guinea pigheart and kidneys with regard to inflammation and necrosis.Figure 8, B and D, shows mild and focal myofiber necrosis inthe left ventricle and focal inflammation (inset). Myofiberinvolvement tended to resolve in both species within 72 h.Figure 9, B and D, shows glomeruli in kidneys harvested 24 hafter dosing revealing no glomerular injury in either species.This was also observed in the 72-h group post-ET harvested

Fig. 5. Cardiovascular effects of PolyHbBv in the rat and guinea pig. A,percentage of change from baseline in MAP in the guinea pig (f) and rat(F). MAP increase compared with baseline was 20% in rats comparedwith 50% in guinea pigs. Values represent mean � S.E.M. B, percentageof change from baseline in heart rate response in rat (f) and guinea pig(F). Measurements were recorded continuously and plotted at 2-minintervals. Significant differences in MAP (p � 0.05) were observed in boththe rat and guinea pig over the entire 2-h period post-transfusion com-pared with baseline (n � 5 for each species).

Fig. 6. OECs for PolyHbBv in rat and guinea pig plasma. OECs forsamples obtained post-transfusion from plasma of rats (A) and guineapigs (B) at specific time intervals are shown. OECs were determined byautomated Hemox analyzer at 37°C. All samples contained approxi-mately 60 �M heme in a 0.01 M phosphate buffer containing 0.1 M NaCl,pH 7.4. Antifoaming agent and Hayashi reduction reagents (see Materi-als and Methods) were added to the buffer. Results are expressed aspercentages of oxygen saturation versus partial pressure of oxygen. P50values were obtained from these curves after correction for the incom-plete saturation OECs at 100 mm Hg (see Materials and Methods).

56 Buehler et al.

at Sem

melw

eis University on F


Dow

nloaded from


kidney tissue. No abnormal histopathology was noted in anyof the heart or kidneys tissues of animals subjected to ETwith albumin (data not shown).

DiscussionInfusion of HBOCs can lead to deleterious effects due to the

spontaneous and uncontrolled oxidation of its heme iron.This process can be enhanced by endogenous oxidants, suchas H2O2 and NO and/or their metabolites in circulation. Therapid conversion of the ferrous (oxygen carrying iron) to theferric (nonoxygen carrying iron) may compromise the effec-tiveness and safety of HBOC for at least two reasons: 1) ferricHb does not bind oxygen, and its accumulation can reducethe oxygen-carrying capacity of HBOCs; and 2) ferric Hb inthe presence of oxidants can promote and sustain a viciousoxidative cycle that ultimately leads to the loss of highlycytotoxic heme. We previously reported that endogenousand/or exogenous reducing agents such as AA (Dunne et al.,2006) or selenium (Baldwin et al., 2004) can control Hboxidation reactions in vitro and in vivo and that these reduc-tants can potentially suppress the untoward consequences ofthese reactions. In the current work, we tested the hypothe-sis that oxidation of polymerized bovine Hb would be modu-

lated differentially in the circulations of two species withdifferent reductive capabilities, i.e., with and without theability to synthesize AA, because these differences may affectHb oxygen-carrying capabilities.

Plasma antioxidative status was clearly compromised inguinea pigs following ET as evidenced by the 50% reductionin plasma AA. In contrast, rat plasma AA levels remainedslightly above baseline levels until PolyHbBv was clearedfrom the circulation of the animal. This probably reflects theability of rats to generate sufficient AA to compensate for theloss of AA incurred from the ET and from its use during thereduction of PolyHbBv. In addition to AA, urate has beenshown to play a significant role in maintaining the ferrousstate of Hb in plasma and RBCs (Ames et al., 1981). How-ever, both rat and guinea pig metabolize urate similarly, andthey maintain mean plasma levels of 9.3 and 7.0 �g/ml,respectively (Mudge et al., 1968; Habu et al., 2003). There-fore, the influence of urate in these studies is probably notsignificant. Moreover, we recently estimated that in rabbits,which have approximate plasma concentrations of 50 �M AAand 30 �M urate, respectively, AA was kinetically morecompetent in reducing the oxidized forms (ferric/ferryl) of theinfused cross-linked Hb than urate (Dunne et al., 2006).

When taking the oxidative status of PolyHbBv heme ironinto account, a disconnect in the pharmacokinetics of ferricand ferrous PolyHbBv becomes apparent in the guinea pig.The in vivo oxidation of PolyHbBv in the guinea pig resultedin a 5-fold higher overall exposure (AUC0–�) to ferric PolyH-bBv. Furthermore, in the guinea pig, approximately 20 and40% of the circulating PolyHbBv was oxidized to ferric Poly-HbBv by 4 and 36 h post-ET, respectively. These valuescorrespond to earlier in vitro autoxidation experiments car-ried out on PolyHbBv in the absence of added antioxidants(Alayash et al., 2001). Conversely, in the rat, ferric PolyHbBvdid not increase until plasma concentrations of PolyHbBvwere 20-fold less than the Cmax. Thus, increased levels offerric PolyHbBv at 48 and 72 h occur at low total PolyHbBvconcentrations. Our data are in agreement with several stud-ies using large-volume HBOC administration that docu-mented significant ferric Hb formation in sheep (Lee et al.,1995), rabbits (Dunne et al., 2006), and humans (O’Hara etal., 2001). In addition, administration of various percentagesof oxidized PEGylated-Hb to rats as a 30% ET demonstratedthat excess of 10% PEGylated-Hb in the oxidized form (ferricPEGylated-Hb) reduced both kidney and liver oxygen tension(Linberg et al., 1998). These observations suggest a potentialfor increased toxicity as a result of diminished oxygen deliv-ery and oxidative side reactions caused by excessive ferricHBOC exposure.

Using MALDI-MS, the current study evaluated the in vivooxidative modification as reflected in the structural stabilityof the tetrameric component of PolyHbBv, which constitutes37% in the preinfused final polymer mixture. The mass spec-tra from rat and guinea pig plasma at 4 and 24 h post-ETdemonstrated stability of the PolyHbBv tetramer in plasmasampled from rats at both early and later time points. How-ever, PolyHbBv tetramer instability occurred in guinea pigsat later sampling time points. This observation is based onthe detection of PolyHbBv derived � and � globin chain massions in plasma collected from guinea pigs. These ions couldnot be assigned to RBC derived Hb � and � globin chains.This suggests that PolyHbBv globin chains sustained oxida-

Fig. 7. Correlation between the levels of PolyHbBv (oxy) and HIF-1�DNA binding activity in kidneys from rats and guinea pigs. Ferrous formof PolyHbBv (f) measured in plasma samples taken from rats (A) andguinea pigs (B) plotted as a function of HIF-1� activity as measured byenzyme-linked immunosorbent assay-based method (see Materials andMethods) and time (hours) post-transfusion with PolyHbBv.


at Sem

melw

eis University on F


Dow

nloaded from


tive changes, but they probably do not dissociate in theguinea pig. This may be a result of oxidative stress impartedon the protein, because tetramer destabilization does notoccur in PolyHbBv before infusion based on previous exten-sive mass spectrometry analysis (Buehler et al., 2005). Con-versely, rat Hb (Fig. 4Aa), guinea pig Hb (Fig. 4Ba), humanHbA0, and non cross-linked PEGylated Hb have been shownto readily dissociate in vitro (Iafelice et al., 2007). The overallimpact of Hb monomer formation in vivo could result in acuterenal failure based on globin chain deposition in the glomer-uli of the kidneys (Savitsky et al., 1978).

In spite of some of the reported functional limitations ofPolyHbBv, such as incomplete saturation of its OEC at highpartial pressure of oxygen (i.e., 100 mm Hg) (Fig. 6), insen-sitivity to some allosteric modifiers of Hb function, and re-duced cooperativity (Alayash et al., 2001), we show in thisstudy that this Hb retained its oxygen-carrying capabilities,particularly when the oxidation of its iron was controlled,

such as the case in the rat presumably by endogenous AA.Conversely, uncontrolled oxidation of this Hb in the guineapig, as evidenced by increased levels of its ferric form andleft-shifted OECs, may have compromised its ability to carryoxygen. Whether the dramatic interspecies differences in theoxidation/oxygenation profiles documented herein translateinto different tissue oxygenation is not yet known. However,analysis of HIF-1� activity in kidney tissues obtained fromthe rats and guinea pigs may offer some clues. HIF-1�, amajor molecular transducer of hypoxia, is normally degradedin the presence of oxygen, a substrate for prolyl hydroxylasethat targets HIF-1� to ubiquitination and destruction by theproteasome. By contrast, HIF-1� is stabilized during hyp-oxia, enabling control of the expression of hundreds of genes.We show for the first time a clear correlation between oxy-genation state of Hb during the course of transfusion withrenal HIF-1� activity in both species. Moreover, the decreasein oxygen-releasing capacity of PolyHbBv in the guinea pig

Fig. 8. Heart histopathological evaluation of rat andguinea pig tissue after PolyHbBv exchange transfusion.Hematoxylin and eosin-stained heart harvested at baseline(24 h after surgery) and 24 h postexchange with PolyHbBvin rat (A and B, respectively) and guinea pig (C and D,respectively). The black arrows indicate areas of pathology.In the rat heart (B), single necrotic myofibers are observedin the left ventricle. B, inset shows a small inflammatorylesion in the wall of the left ventricle. In the guinea pigheart (D), areas of necrotic myofibers are seen in the apicalregion of the left ventricle. These findings were mild and insome cases not observed in animals. Resolution was gen-erally achieved by 72 h postexchange.

Fig. 9. Kidney histopathological evaluation of rat andguinea pig tissue after PolyHbBv exchange transfusion.Hematoxylin and eosin-stained kidney harvested at base-line (24 h after surgery) and 24 h postexchange with Poly-HbBv in rat (A and B, respectively) and guinea pig (C andD, respectively) demonstrate no glomerular damage in ei-ther the rat or the guinea pig.

58 Buehler et al.

at Sem

melw

eis University on F


Dow

nloaded from


due to its enhanced oxidation probably contributes to theincrease in HIF-1� activity observed after 50 h post-ET. Weare currently assessing HIF-1� protein levels and the expres-sion of several HIF-1 target genes such as erythropoietin,vascular endothelial growth factor, and heme oxygenase inkidney, heart, and brain of these animals to better under-stand the interplay between Hb oxygenation, redox reac-tions, and overall oxygen hemostasis.

The MAP response in both rats and guinea pigs was sim-ilar in that elevation occurred immediately after the start ofET. However, the two species differed dramatically in theextent of MAP elevation over baseline values. It is temptingto speculate that AA in rats may have also played a part incontrolling Hb reaction with NO, a critical molecule in bloodpressure control. In addition to the well established role ofAA in reducing Hb oxidation, AA may also contribute to Hbvia other important reactions, i.e., nitrosative reactions.There has been in recent years a growing interest in endog-enous NO storage compounds and the role of AA in inducingNO release from these compounds, particularly low-molecu-lar-weight S-nitrosothiols, such as S-nitrosothiol glutathione(Foster and Stamler, 2004), or S-nitrosated proteins, suchalbumin (Grandley et al., 2005). Recent mechanistic studies,using an N-nitrosated tryptophan derivative, showed thatthe primary oxidized product of AA, DHA, efficiently con-sumes NO and subsequent release of ascorbyl radical (Kytziaet al., 2006). Along with oxidative status of the circulatingHBOC, arterial blood pressure response can contribute todecreased efficacy and possible toxicity via reduced bloodflow to vital tissues, increased workload on the heart, andacute renal injury.

Interestingly, neither the enhanced oxidation nor exagger-ated blood pressure responses observed in guinea pigs com-pared with rats translated into increased tissue pathology.Both rats and guinea pigs demonstrated minimal histopa-thology in heart and kidney at 24 h post-ET, and most find-ings were resolved by 72 h post-ET. In both species, myocar-dial toxicity was limited to single myofiber necrosis of the leftventricle at 24 h. Although the absence of histopathologicaldifferences between rat and guinea pig was a surprisingfinding, it is likely that these observations may in fact pro-vide useful information relative to predicting human safetyof HBOCs based on animal data. In addition, this worksupports the idea of early biomarker identification detectablein the absence of histopathological evaluation. Even thoughguinea pigs were unable to control the oxidation of PolyHbBvin the circulation, they, like humans, have increased tissueantioxidant mechanisms (Nandi et al., 1997), which mayhave prevented tissue injury in these otherwise normaltransfused animals. Once protective antioxidant mecha-nisms in the tissue are overcome by events such as ischemiafollowed by reperfusion leading to H2O2 and neutrophil re-leased hypochlorous acid accumulation, the combination ofan oxidized HBOC and tissue oxidant accumulation maybecome overwhelming.

In summary, we show that rats, unlike guinea pigs, con-trolled oxidation of the infused cell-free Hb after 50% ET andthat they maintained adequate plasma oxygenation with lit-tle or no change in the P50 of Hb, due in large part to thepresence of adequate circulating AA levels, during the first20 h post-ET. By contrast, in guinea pigs, Hb oxidation pre-ceded in the plasma at much higher rates probably due to the

lack of endogenous sources of AA. The decreased ability ofthis Hb to carry oxygen due to the increased oxidation of itsiron was consistent with the increased HIF-1� activity, par-ticularly at later times post-ET. Therefore, this study has thefollowing implications for the use of HBOCs in clinical devel-opment: 1) the guinea pig may represent an appropriatespecies for the early evaluation of HBOC susceptibility tooxidation and destabilization in circulation, and it may bemore relevant to simulate the use of HBOCs in clinical set-tings associated with antioxidant depletion, such as traumaand diabetes; 2) endogenous antioxidant mechanisms inplasma, RBCs, and tissue should be taken into account whendesigning animal studies aimed at understanding HBOCoxygenation/oxidation reactions; and 3) coadministration ofreducing agents, such as AA, may offer a simple strategy tocontrol Hb oxidative side reactions.

Acknowledgments

We thank Dr. Paul Yeh for performing DNA trans-binding assess-ment of HIF-1� and Francine Wood for performing oxygen equilib-rium measurements.

ReferencesAlayash AI (2004) Oxygen therapeutics: can we tame haemoglobin? Nat Rev Drug

Discov 3:152–159.Alayash AI, Summers AG, Wood F, and Jia Y (2001) Effects of glutaraldehyde

polymerization on oxygen transport and redox properties of bovine hemoglobin.Arch Biochem Biophys 391:225–234.

Ames BN, Cathcart R, Schwiers E, and Hochstein P (1981) Uric acid provides anantioxidant defense in humans against oxidant- and radical-caused aging andcancer: a hypothesis. Proc Natl Acad Sci U S A 78:6858–6862.

Ancill RJ (1956) The blood volume of the normal guinea-pig. J Physiol 132:469–475.Baldwin AL, Wiley EB, and Alayash AI (2004) Differential effects of sodium selenite

in reducing tissue damage caused by three hemoglobin-based oxygen carriers.J Appl Physiol 96:893–903.

Buehler PW, Boykins RA, Norris S, Freedberg DI, and Alayash AI (2005) Structuraland functional characterization of glutaraldehyde-polymerized bovine hemoglobinand its isolated fractions. Anal Chem 77:3466–3478.

Carmichael FJ, Ali ACY, Campbell JA, Langlois SF, Biro GP, Willan AR, Pierce CH,and Greenburg AG (2000) A phase I study of oxidized raffinose cross-linked humanhemoglobin. Crit Care Med 28:2283–2292.

Chatterjee IB (1973) Evolution and the biosynthesis of ascorbic acid. Science 182:1271–1272.

Day BW, Jin R, and Karol MH (1996) In vivo and in vitro reactions of toluenediisocyanate isomers with guinea pig hemoglobin. Chem Res Toxicol 9:568–573

Dunne J, Caron A, Menu P, Alayash AI, Buehler PW, Wilson MT, Silaghi-Dumitrescu R, Faivre B, and Cooper CE (2006) Ascorbate removes key precursorsto oxidative damage by cell-free haemoglobin in vitro and in vivo. Biochem J399:513–524.

Foster MW and Stamler JS (2004) New insights into protein S-nitrosylation. Mito-chondria as a model system. J Biol Chem 279:25891–25897.

Garrick LM, Sharma VS, McDonald MJ, and Ranney HM (1975) Rat haemoglobinheterogeneity. Two structurally distinct alpha chains and functional behaviour ofselected components. Biochem J 149:245–258.

Grandley RE, Tyurin VA, Huang W, Huang Arroyo A, Daftary A, Harger G, Jiang J,Pitt B, Taylor RN, Hubel CA, et al. (2005) S-Nitrosoalbumin-mediated relaxationis enhanced by ascorbate and copper: effects in pregnancy and preeclampsiaplasma. Hypertension 45:21–27.

Habu Y, Yano I, Takeuchi A, Saito H, Okuda M, Fukatsu A, and Inui KI (2003)Decreased activity of basolateral organic ion transports in hyperuricemic ratkidney: roles of organic ion transporters, rOAT1, rOAT3 and rOCT2. BiochemPharmacol 66:1107–1114.

Hayashi A, Suzuki T, and Shin M (1973) An enzymic reduction system for metmyo-globin and methemoglobin, and its application to functional studies of oxygencarriers. Biochim Biophys Acta 310:309–316.

Iafelice R, Cristoni S, Caccia D, Russo R, Rossi-Bernardi L, Lowe KC, and PerrellaM (2007) Identification of the sites of deoxy haemaglobin pegylation. Biochem J403:189–196.

Institute of Laboratory Animal Resources (1996) Guide for the Care and Use ofLaboratory Animals 7th ed. Institute of Laboratory Animal Resources, Commis-sion on Life Sciences, National Research Council, Washington DC.

Kieda C, Greferath R, Crola Da Silva C, Fylaktakidou KC, Lehn JM, and Nicolau C(2006) Suppression of hypoxia-induced HIF-1alpha and of angiogenesis in endo-thelial cells by myo-inositol trispyrophosphate-treated erythrocytes. Proc NatlAcad Sci U S A 103:15576–15581.

Kim HW and Greenburg AG (2004) Artificial oxygen carriers as red blood cellsubstitutes: a selected review and current status. Artif Organs 28:813–828.

Kytzia A, Korth, H-G., Sustmann R, De Groot H, and Kirsch M (2006) On themechanism of the ascorbic acid-induced release of nitric oxide from N-nitrosatedtryptophan derivatives: scavenging of NO by ascorbyl radicals. Chemistry 12:8786–8797.


at Sem

melw

eis University on F


Dow

nloaded from


Lee HB and Blaufox MD (1985) Blood volume in the rat. J Nucl Med 26:72–76.Lee R, Neya K, Svizzero TA, and Vlahakes GJ (1995) Limitations of the efficacy of

hemoglobin-based oxygen-carrying solutions. J Appl Physiol 79:236–242.Linberg R, Conover CD, Shum KL, and Shorr RGL (1998) Hemoglobin based oxygen

carriers: how much methemoglobin is too much? Artif Cells Blood Substit ImmobilBiotechnol 26:133–148.

May JM, Qu ZC, and Cobb CE (2004) Human erythrocyte recycling of ascorbic acid:relative contributions from the ascorbate free radical and dehydroascorbic acid.J Biol Chem 279:14975–14982.

Mudge GH, Cucchi J, Platts M, O’Connell JM, and Berndt WO (1968) Renal excre-tion of uric acid in the dog. Am J Physiol 215:404–410.

Nagababu E, Ramasamy S, Rifkind JM, Jia Y, and Alayash AI (2002) Site-specificcross-linking of human and bovine hemoglobins differentially alters oxygen bind-ing and redox side reactions producing rhombic heme and heme degradation.Biochemistry 41:7407–7415.

Nandi A, Mukhopadhyay CK, Ghosh MK, Chattopadhyay DJ, and Chatterjee IB(1997) Evolutionary significance of vitamin C biosynthesis in terrestrial verte-brates. Free Radic Biol Med 22:1047–1054.

O’Hara JF, Colburn WA, Tetzlaff JE, Novick AC, Angermeier KW, and Schubert A(2001) Hemoglobin and methemoglobin concentrations after large-dose infusionsof diaspirin cross-linked hemoglobin. Anesth Analg 92:44–48.

Przybelski RJ, Daily EK, Kisicki JC, Mattia-Goldberg C, Bounds M, and Colburn WA(1996) Phase I study of the safety and pharmacologic effects of diaspirin cross-linked hemoglobin solution. Crit Care Med 24:1993–2000.

Riess JG (2001) Oxygen carriers (“blood substitutes”)–raison d’etre, chemistry, andsome physiology. Chem Rev 101:2797–2920.

Savitsky JP, Doczi J, Black J, and Arnold JD (1978) A clinical safety trial of stromafree hemoglobin. Clin Pharmacol Ther 23:73–80.

Sloan EP, Koenigsberg M, Gens D, Cipolle M, Runge J, Mallory MN, and Rodman GJr (1999) Diaspirin cross-linked hemoglobin (DCLHb) in the treatment of severetraumatic hemorrhagic shock: a randomized controlled efficacy trial. J Am MedAssoc 282:1857–1864.

Su D, May JM, Koury MJ, and Asard H (2006) Human erythrocyte membranescontain a cytochrome b561 that may be involved in extracellular ascorbate recy-cling. J Biol Chem 281:39852–39859.

Winterbourn CC (1985) Reactions of superoxide with hemoglobin, in CRC Handbookof Methods of Oxygen Radical Research (Greenwald RA ed), pp 137–141, CRCPress, Boca Raton, FL.

Address correspondence to: Dr. Paul W. Buehler, Center for BiologicsEvaluation and Research, Food and Drug Administration, 8800 Rockville Pike,Bldg. 29, Rm. 129, Bethesda, MD 20892. E-mail: [email protected]

60 Buehler et al.

at Sem

melw

eis University on F


Dow

nloaded from


buehler 2007 effects of endogenou

Documents