Search International and National Patent Collections
Some content of this application is unavailable at the moment.
If this situation persists, please contact us atFeedback&Contact
1. (WO1992009630) IMIDOESTER CROSS-LINKED HEMOGLOBIN COMPOSITIONS
Note: Text based on automatic Optical Character Recognition processes. Please use the PDF version for legal matters
IMIPOESTER CROSS-LINKED HEMOGLOBIN COMPOSITIONS
BACKGROUND OF THE INVENTION
The present invention is directed toward a novel process for producing a purified cross- linked hemoglobin product suitable as a blood replacement for transfusion in human beings and in animals or as an oxygen transport fluid. The subject cross-linked hemoglobin product is essentially free of contaminants in that it contains very low levels of bacterial endotoxin and phospholipid. The structure and function of hemoglobin have been reviewed (Buπn, H.F., and B.G. Forget, Hemoglobin: Molecular. Genetic and Clinical Aspects. W.B. Saunders Company, Philadelphia, 690 pp., (1986). Mammalian hemoglobin is a tetramer containing two each of two different types of subunits, alpha and beta. Each subunit is of approximate molecular weight 16,000 and contains a heme group with a central iron atom. The protein portion, or globin functions to keep the heme iron in the reduced or ferrous state, thus to allow the hemoglobin molecule to reversibly bind oxygen.
Mammalian hemoglobin should not be thought of as a static tetramer, that is a single protein species with a molecular weight of 64-68,000. It is instead composed of sub-units connected by noncovalent bonds, and existing in dynamic equilibria involving monomer-dimef and dimer-tetramer interactions. At any moment, the hemoglobin would consist of a certain proportion of monomers with a molecular weight of approximately 16,000, dimers of molecular weight 32,000, and tetramers. The distribution of monomer, dimer, and tetramer species in solution are dependent upon hemoglobin concentration, the ligand state of the hemoglobin, as well as pH and salt composition of the solution in which it is contained.
Mammalian hemoglobin is contained within red blood cells or erythrocytes. In the mammals these specialized cells have lost their nuclei during maturation, and are simply sacs of hemoglobin containing very low concentrations of other proteins. The hemoglobin is contained with circulating erythrocytes for several reasons. First of all, if the hemoglobin were not in cells but instead circulated free in solution, it would readily be cleared from circulation by passage through the capillary walls, the kidney and other sites. Certain invertebrate organisms have solved this problem by evolving polymeric hemoglobins with molecular weights in the millions which are too large to pass through capillaries and the nephron of the kidney. Other functions served by the red cell are to provide a sheltered environment to protect the hemoglobin molecule from proteolytic serum proteins, maintain a balance of ions proper for hemoglobin function, provide an enzyme system which maintains the heme iron in the reduced (functional) state, and control the presence of aliosteric effector molecules.
There have been numerous efforts to produce clarified stroma-free hemoglobin solutions for blood replacement transfusion in humans (Pennell, R.B., and W.E. Smith, Preparation of stabilized solutions of hemoglobin. Blood 4: 380-385, (1949); Rabiner, S.F., Helbert, J.R., Lopas, H., and L.H. Friedman, Evaluation of a strom-free hemoglobin solution for use as a plasma expander. J. EXP. Med. 126: 1127-1142 (1967); Christensen, S.M., Medina, F., Winslow, R.W., Snell, S.M., Zegna, A., and M.A. Marini, Preparation of human hemoglobin Ao for possible use as a blood substitute. J. Biochem. Biophvs. Meth. 17: 143-154 (1988); and Feola, M., Gonzalez, H., Canizaro, P.C., Bingham, D. and P. Periman, Development of a bovine stroma-free hemoglobin solution as a blood substitute. Surg. Gvn. Obst. 157: 399-408 (1983)). These solutions contain the hemoglobin in an unmodified state, and the hemoglobin is free to dissociate into its subunits. Infusion of unmodified hemoglobin leads to rapid clearance ofthe hemoglobin from the circulation. In addition, infusion of unmodified hemoglobin reportedly leads to a host of deleterious effects within the body including nephrotoxicity. The nephrotoxicity associated with unmodified hemoglobin has prompted the Center for Biologies Evaluation and Research ofthe U.S. Food and Drug Administration to state that unmodified hemoglobin should not be present in a hemoglobin-based oxygen carrier.
There have been numerous reports describing the stabilization of hemoglobin solutions by forming covalent chemical cross-links between the hemoglobin polypeptide chains (Rausch, C.W., and M. Feola, Extra pure semi-synthetic blood substitute, International patent application no. PCT/US87/02967, Int. publication no. WO/88/03408 19 May 1988; Bonhard, K., and N. Kothe ;

Cross-linked hemoglobin of extended shelf life and high oxygen transport capacity and process for preparing same, U.S. Patent No. 4,777,244; Bucci. E., Razynska, A., Urbaitis, B., and C. Fronticelli Pseudo cross-link of human hemoglobin with mono-(3,5-dibromosalicyl)fumarate. .

Biol. Che . 264: 6191-6195 (1989); and U.S. Patent No. 4,001,200.)
INFORMATION DISCLOSURE STATEMENT
Cross-linked hemoglobin compositions have been described in U.S. Patents 4,001,200; 4,011,401; 4,053,590; and 4,061,736. More elaborate forms of cross-linking hemoglobin for subsequent purification have also been disclosed in U.S. Patent 4,857,636.
The preparation and cross-linking of bovine hemoglobin has been described in Feola, M., Gonzalez, H., Canizaro, P.C., Bingham, D. and P. Periman, Development of a bovine stroma-free hemoglobin solution as a-blood substitute. Surg. Gvn. Obst. 157: 399-408 (1983) and International patent application no. PCT/US87/02967, Int. publication no. WO/88/03408 19 May, 1988.
Dimethyl adipimidate reactions with normal human hemoglobin (HbA) and sickle cell hemoglobin (HbS) have been described in Plese, C.F., and E.L. Amma, Circular dichroism as a probe of d e allosteric R-T transformation in hemoglobins and modified hemoglobins (1977). U.S. Patent 3,925,344 (1975) discloses cross-linking hemoglobin with imidoesters to prepare a plasma protein material or plasma expander; however, this material is not suitable as an oxygen transfer fluid. As reported by Pennadiur-Das, R., Heath, R.H., Mentzer, W.C, and B.H Lubin "Modification of hemoglobin S with dimethyl adipimidate Contribution of individual reacted subunits to changes in properties Biochim. Biophvs. Acta 704: 389-397 (1982), " cross-linking wit the imidoesters (DMA) increased the oxygen affinity of the hemoglobin which makes it unsuitabl for transporting oxygen and releasing it. Pennathur-Das, R., Vickery, L.E., Mentzer, W., an B.H. Lubin, Modification of hemoglobin A with dimethyl adipimidate. Contribution of individu reacted subunits to changes in oxygen affinity. Biochim. Biophvs. Acta 580: 356-365 (1979).
Thus, it is desirable to find a method for cross-linking hemoglobin which makes it stabl and predominantly in a tetramer form or tetramer multiples but still has a p50 or oxygen affinit which is low enough to effectively transport oxygen and then release it to the cells of a livin organism. Also, it is important that the molecular weight is predominantly at least 64,000 whic is the tetramer form of hemoglobin with little to none below 64,000; i.e., dimer or monomer form and little to none greatly over 64,000 which can be molecules large enough to cause complimen activitation in mammalian organisms. Developing such a polymerization method is therefore ver desirable in order to successfully find an oxygen transport substitute.
SUMMARY OF THE INVENTION
In one aspect the subject invention is a stable, cross-linked hemoglobin composition capabl of transporting and releasing oxygen in living cells. The hemoglobin is essentially free o impurities and is cross-linked with an imidoester whereby the hemoglobin is present i predominantly tetramer and larger forms. The preferred imidoesters for cross-linking th hemoglobin are dimethyl adipimidate or dimethyl suberimidate. Hemoglobin predominantly in tetramer or larger form can be interpreted as where at least 80% of the cross-linked hemoglobi has a molecular weight of at least 64,000 Daltons, more preferably where at least 90 to 95% o the cross-linked hemoglobin has a molecular weight of at least 64,000 Daltons. In addition, die imidoester cross-linked hemoglobin is very stable to methemoglobin as compared to conventional means of cross-linking and has an oxygen affinity, p50, of at least 13mm Hg.
The cross-linked hemoglobin composition can also include a pharmaceutically acceptable carrier medium for facilitating transfusion or injection into a mammal being treated or for use as an oxygen carrying fluid for analytic, transplant or laboratory usage. Typical pharmaceutically acceptable carrier medium can include purified water or saline solution.
In another aspect the subject invention is a method for preparing a cross-linked hemoglobin composition comprising a) collecting mammalian blood, b). separating red blood cells from the collected blood and washing die red blood cells, c) lysing die washed red blood cells, d) centrifuging the lysed blood cells to obtain a lysate of hemoglobin, e) purifying the lysate by anion or cation exchange chromatography, f) performing an ultrafiltration or microfiltration of eluted lysate obtained from the chromatography whereby purified hemoglobin is obtained, and g) cross-linking the purified hemoglobin with an imidoester to obtain a hemoglobin present in predominantly tetramer form.

The lysing of the red blood cells can be conducted by hypo-osmotic shock. The ultrafiltration is typically conducted with a 300,000 molecular weight membrane and the microf-ltration is typically conducted with from about a 0.025 to about a 0.040 micron filter. Alternatively, cross-linking with imidoester step (g) can be performed after the lysate of hemoglobinis obtained in step (d) whereby step (e) begins with the purification of cross-linked hemoglobin. The method can additional include a step (h) where the purified, cross-linked hemoglobin is size excluded to remove low molecular weight hemoglobin (Jess than 64,000 Daltons). Typically, the size exclusion is by low pressure size exclusion chromatography.
The present invention therefore describes an imidoester, cross-linked hemoglobin composition essentially free of any impurities and being present in predominantly a tetramer or larger form which has improved stability to oxidation of the hemoglobin over more traditional forms of cross-linking hemoglobin such as with glutaraldehyde. The hemoglobin composition has a predominant molecular weight of at least 64,000 Daltons and has improved cross-link stability. The hemoglobin is suitable for use as a blood oxygen transport substitute for mammals or generally as an oxygen transport fluid. ""
DETAILED DESCRIPTION OF THE INVENTION
The present invention describes a method for purifying and cross-linking hemoglobin with certain imidoesters which have not been used to cross-link hemoglobin for use as a transfusion therapeutic. These bifunctioπal imidoester cross-linkers include dimethyl adipimidate and dimethyl suberimidate, among others. There are advantages in using these cross-linking agents over hemoglobin cross-linkers which have been used in the past. Particular reaction conditions have been discovered which provide a narrow molecular weight distribution and suitable transporting of oxygen.
Imidoesters such as dimethyl adipimidate and dimethyl suberimidate are bifunctioπal cross-linking reagents that react specifically and quickly with proteins under mild conditions. Stable, amidine adducts are formed with epsilon-amino groups of lysine residues and the amino terminal amine of polypeptide chains (Wold, F. (1972) Bifunctional reagents. Methods in Enz. 25: 623-651 ;

Peters, K., and F.M. Richards (1977) Chemical cross-linking: reagents and problems in studies of membrane structure. Ann. Rev. Biochem. 4£: 523-551. The cross-linking reagent typically reacts with one-to-two amino groups per hemoglobin subunit. Some cross-linking occurs between the two beta chain lysine residues (beta-82) that stabilize the tetrameric form of human hemoglobin.

Dimethyl adipimidate has been studied as an antisickling agent for hemoglobin S (Lubin, B.H.,

Pena, V., Mentzer, W.C, Bymun, E., Bradley, T.B., and L. Packer (1975) Dimediyl adipimidate: a new antisickling agent. Proc. Natl. Acad. Sci. U.S.A. 72: 43-46). Modification of the beta-82 lysines alters the conformation of the deoxygenated form of HbS and so prevents its crystallization or gelation. Crystallization is the molecular basis for the sickling phenomenon seen with the afflicted red cells. Although modification of die beta-82 lysine by die cross-linking agent i required to prevent sickling, the actual cross-linking ofthe two beta-82 lysines within a hemoglobi tetramer is not. The present invention uses the dimethyl adipimidate to bovine hemoglobin with th intent of forming both stabilized tetramers, caused by introducing intersubunit cross-links betwee the beta-82 lysines, and high molecular weight aggregates of stabilized tetramers, caused b introducing intermolecular cross-links between epsilon-amino groups of separate, stabilized, hemoglobin tetramers.
The advantages for cross-linking hemoglobin with imidoesters, and specifically dimethyl adipimidate is twofold: first, cross-linking will retard d e dissociation of tetrameric hemoglobin o molecular weight 64,000, into 32,000 molecular weight dimers. In unmodified hemoglobin tetrameric and dimeric species are in a dynamic equilibrium. Stabilization of the tetrameric hemoglobin by cross-linking with imidoesters will substantially reduce d e passage of hemoglobin into d e renal tubules by preventing filtration across the membrane of the glomerular capsule.
Secondly, cross-linking will result in the formation of high molecular weight oligomers of the hemoglobin tetramer which would be extremely resistant to renal filtration and more stable to oxidation at physiological temperature than either unmodified hemoglobin or hemoglobin modified with d e cross-linking agent glutaraldehyde.
Dimethyl adipimidate is a relatively inexpensive reagent and is available in bulk quantities. The cross-linking process utilizing dimed yl adipimidate lends itself well to scale up.
The oxygen binding properties of the various forms of bovine hemoglobin were measured and appear in Table I. The P50 is the partial pressure of oxygen at which 50% of the available sites have bound oxygen. The Hill coefficient or n value is an expression of the cooperativity among the oxygen binding sites with high n values denoting high cooperativity. The P50 values were measured during both the oxygenation and deoxygenation of the hemoglobin solution. The mean P50 values were: 13.5 mM Hg for dimediyladipimidate cross-linked hemoglobin, 14.25 mM Hg for the glutaraldehyde cross-linked hemoglobin and 23.75 mM Hg for unmodified bovine hemoglobin. If extreme cross-linking was carried out with imidoesters or glutaraldehyde, the p50 values for these hemoglobins was greatiy decreased. Also, wid both the imidoesters and glutaraldehyde, cross-linking ofthe hemoglobin reduced its cooperativity with the greater reduction being seen in the glutaraldehyde treated sample.


The stability of the various cross-linked hemoglobins was compared and showed the imidoester version to have superior stability. Therefore, the imidoester under proper cross-linking procedures, described below, yielded a better oxygen transport hemoglobin when considering stability and oxygen affinity.
In the 36-37°C incubation, the dimediyladipimidate cross-linked, glutaraldehyde cross- linked, and unmodified hemoglobin samples contained 2.5%, 8.0% and 3.5% med emoglobin, respectively, at the start of the incubation. After 5 hours at 36-37°C the percentages of methemoglobin were 8.9%, 44.4%, and 8% for the dimethyladipimidate cross-linked, glutaraldehyde cross-linked, and unmodified bovine hemoglobin samples, respectively. After 24 hours, die methemoglobin concentrations were 24.1, 87.9, and 42% for die dimethyladipimidate cross-linked, glutaraldehyde cross-linked, and unmodified bovine hemoglobin samples, respectively. After 48 hours the methemoglobin concentrations in dimediyladipimidate cross-linked and uncross-linked bovine hemoglobin samples were 46.5% and 68%, respectively. The dimediyladipimidate cross-linked hemoglobin was more stable than unmodified hemoglobin to mediemoglobin formation. The literature reports that glutaraldehyde cross-linked human hemoglobin autooxidizes at a rate four times higher than unmodified human hemoglobin. Therefore, it was unexpected that the DMA cross-linked hemoglobin would be more stable than unmodified hemoglobin.
The stability of unmodified bovine hemoglobin and of dimediyladipimidate cross-linked and glutaraldehyde cross-linked bovine hemoglobin to autoxidation also was measured after incubation for 52 days at 4°C in die same phosphate buffer as described above. After 28 days die dimediyladipimidate cross-linked bovine hemoglobin increased from 2.5% to 9.6% methemoglobin.

Over the same time period d e glutaraldehyde cross-linked hemoglobin increased from 8.0% t 23.5% methemoglobin. When measured after 52 days, d e dimethyladipimidate cross-linke hemoglobin contained 14.3% mediemoglobin while after 37 days d e unmodified bovin hemoglobin contained 16.3% mediemoglobin. As in the 37 °C incubation, die dimediyladipimidat cross-linked hemoglobin was significandy more stable than glutaraldehyde cross-linked hemoglobi or unmodified bovine hemoglobin.
In one aspect the invention relates to the enhancement of stability for mammalia hemoglobins including bovine hemoglobin, after cross-linking with imidoesters including dimethy adipimidate. It is disclosed herein mat the stability of mammalian hemoglobins can be enhance by reaction wid imidoesters, specifically dimediyl adipimidate under specific reaction conditions In addition, a method for producing purified mammalian hemoglobin, specifically bovin hemoglobin is also disclosed. The hemoglobin would be essentially free of impurities which mean that it would not contain any materials in a quantity which would cause any deleterious effects t the mammal treated. The hemoglobin would be sufficiendy free of endotoxin (for example, les than 0.5 EU/ml at least a 40 mg/ml concentration of hemoglobin), phospholipid, viruses, and othe nonhemoglobin proteins prior to cross-linking. However, it is also possible to first cross-link th hemoglobin after collection of die lysated hemoglobin and then purify die cross-linked hemoglobin, in fact, this method may be more efficient for large scale production. The purified and cross linked hemoglobin would be suitable as an oxygen transporting blood extender for transfusion into humans and animals.
Typically, the hemoglobin of die subject invention would be prepared by collecting die mammalian blood under aseptic conditions, washing the red cells, lysing d e red cells, centrifuging die lysate, either ultrafiltrating with at least a 300,000 molecular weight membrane or microfiltrating with a 0.025 to about a 0.040 micron filter, purifying by anion or cation exchange chromatography to obtain a purified hemoglobin, then cross-linking with d e imidoester and optionally performing a size exclusion chromatography or ultrafiltration to remove any low molecular weight hemoglobin. Alternatively, the hemoglobin may have been cross-linked after centrifuging and collecting the lysate. The purification steps could then be carried out including die optional size exclusion step.
The hemoglobin thus prepared would be essentially free of impurities and consist predominandy of a stabilized tetramer. Another way to characterize the cross-linked hemoglobin would be to state that it has a molecular weight distribution of predominantly at least 64,000. Predominantly or predominant as used herein means at least 80% of die cross-linked hemoglobin, preferably 90 to 95% of the cross-linked hemoglobin.
In order to prepare the subject specialized oxygen transport hemoglobin various cross-linking reaction parameters are specified. First, the concentration of chloride ion, and/or sodium chloride has a profound effect on the extent of cross-linking of (bovine) hemoglobin with DMA over the range of from 1 mM Tris-HCl to 50 mM Tris-HCl plus 2.0 M NaCl. In 1 and 10 mM Tris-HCl, after 10 additions of DMA, 90.3% and 72.2% (respectively) of the hemoglobin remained unpolymerized. In contrast to this result, in the presence of 50 mM Tris-HCl plus 0.5, 0, and 2.0 M NaCl, only 25.8, 24.4, and 20.7% remained uncross-Iinked (respectively). A steady decrease in die unpolymerized hemoglobin, and an increase in the high molecular weight (> 400,000 daltons) was observed as the salt concentration was increased. The highest percentage of hemoglobin in the 64-400,000 range was observed when die cross-linking was done either in 50 mM Tris-HCl or in 50 mM Tris-HCl plus 100 mM NaCl.
Secondly, buffer type was shown to have a strong influence on cross-linking. For example, the DMA cross-linking reaction proceeded poorly in glycine-HCl, ethanolamine-HCl, and in a solution containing sodium phosphate plus lysine. In these buffer types, it is assumed mat the primary amino function inhibited the cross-linking reaction, which is thought to occur primarily to amino groups on the N-terminus of the protein and its lysine residues. The greatest amount of cross-linking was seen in die presence of 2-amino-2-methyl-l-propanol buffer with only 23.4% of the hemoglobin remaining uncross-Iinked. This buffer, obviously, also has a primary amino functionality. The next four buffers in terms of DMA cross-linking efficiency were Tris (Tris[hydroxymethyl]aminomethane), sodium carbonate, CHES (2-[N- Cyclohexylaminojedianesulfonic acid), and CAPSO (3-[Cyclohexyl-amino]-2-hydroxy-l- propanesulfonic acid); three of which have primary amino functionalities. In contrast, CAPS (3- [Cyclohexylamino]-l-propanesulfonic acid) was not a good buffer for die cross-linking reaction, and die resulting product was only 27.2% cross-linked. Sodium phosphate also was not a good choice for die cross-linking reaction.
Thirdly, pH played an important role in effectively polymerizing hemoglobin in an imidoester. For example, attemps to cross-link hemoglobin with DMA at pH values lower than 8.0 were generally unsuccessful. At pH 8.0 after 10 additions of DMA and a final stoichiometric ratio of DMA to hemoglobin equal to 10, the hemoglobin was only 12.2% cross-linked. An increase in the cross-linking efficiency was witnessed at H 9.0, 10.0, and 10.5, with only 58%, 48%, and 50%, respectively, remaining uncross-Iinked at the end of die DMA additions. It was apparent from the data, however, that die reaction of DMA with hemoglobin requires a pH of 9 or greater in order to be successful.
Finally, it was observed that deoxygenated hemoglobin was more readily cross-linked than oxygenated hemoglobin. After providing die above described polymerization conditions for optimum cross-linking, the uncross-Iinked hemoglobin can be filtered or chromatographed out to yield the 64,000 or greater molecular weight and then further filtering or chromatograph to yield predominandy 64,000 weight. Therefore, optimizing the cross-linking of the hemoglobin gready facilitates the economical and practical preparation of d e subject oxygen transfer material whi is predominandy 64,000 to 300,000 molecular weight.
After the cross-linked hemoglobin is prepared it is typically placed in a pharmaceutical acceptable carrier medium appropriate for injecting or infusion into a mammal being treate Typical pharmaceutically acceptable carrier medium can include physiologically acceptable s solutions appropriate for injection or infusion such as saline solutions which are commercial available. A pharmaceutical acceptable carrier would be any of a variety of liquids which wou be relatively inert to the cross-linked hemoglobin and would not have deleterious effect on t mammal being infused or injected. Suitable fluids can also include other blood products eidi natural or synthetic or fluorocarbon fluids. Generally die subject cross-linked hemoglobin wou be admixed in a suitable pharmaceutical carrier whereby die hemoglobin concentration could administered in adequate amount of from about 40 to about 140 mg/ml.
The subject invention is described in greater detail as follows for each of the various ste and molecular composition described above. The methods described apply to any mammali hemoglobin even though for purposes of demonstration d e specific examples which describe~ti purification, cross-linking and modification use bovine hemoglobin. Bovine hemoglobin doe however, have an advantage for having a very high p50 value suitable for oxygen transport. Example 1. Collection of blood and preparation of red cell hemolysate.
Blood was drawn in an aseptic manner from Holstein steers at the farm belonging to T Upjohn Company Agricultural Division. First, die hair was shaved from the neck of each ste with electrical clippers. The shaved area then was wiped with a clotii towel soaked in an iodi solution, followed by washing with 95% ethanol from a squirt botde. The ethanol was allowe to air dry prior to blood drawing. The blood was drawn into 0.5 - 1.0 liter evacuated bloo collection botdes which were sterile and pyrogen ree. An appropriate amount of a sterile sodiu heparin solution was added to each botde prior to blood collection. The botdes were placed on ic immediately after filling, and were transported to the laboratory on ice. For initial studies, 2.5- 5. liters of blood were drawn.
In die laboratory, all procedures were conducted eidier on ice or at 4° C to minimize th growth of microorganisms and die concomitant pyrogen formation. In addition, the purificatio steps should be conducted in a clean environment with filtered air that has a low particulat content. The first steps in the purification ofthe hemoglobin involve die separation ofthe red bloo cells from the blood plasma, followed by repeated washings of the red cells with a salt solutio and centrifugation to pellet the red cells. The whole blood first was transferred to pyrogen-free 0. liter centrifuge botdes and centrifuged for 30 - 40 minutes at 4000 RPM and 4β C The red cell were pelleted during d e centrifugation. The blood plasma in the supernatant was removed b aspiration. The red blood cells or erythrocytes were resuspended in an ice cold solution containin 16 grams of sodium chloride (NaCl) per liter of purified water. This solution is referred to as 1.6% saline. The red cells were resuspended by stirring with a sterile and pyrogen-free glass rod, or by gende repeated inversion of d e centrifuge botde. After d e red cells were resuspended, they were recentrifuged at 10,000 RPM for 30-40 minutes at 4° C. The red cells were resuspended in 1.6% saline and centrifuged again at 10,000 RPM for 30-40 minutes at 4° C. The red cells were washed witii 1.6% saline a total of three times prior to red cell rupture or lysing.
After washing die cells, die red cells were lysed by hypo-osmotic shock in the following manner. To one volume of washed, pelleted red cells were added four volumes of ice cold 0.0025

M sodium phosphate buffer, pH 7.4. The red cells were suspended in this dilute phosphate buffer by gende inversion of the centrifuge botdes. After the cells were completely suspended, the red cell suspensions were incubated at 0-4° C for one hour, followed by centrifugation at 10,000 RPM for 30-90 minutes at 4° C After centrifugation, die hemoglobin in d e supernatant was recovered by aspiration. In certain experiments, the protocol was modified in die following manner. After d e one hour incubation with 0.0025 M sodium phosphate buffer at 0-4° C, a 2.0 M solution of sodium chloride was added to a final concentration of 0.2 M prior to centrifugation. Addition of die sodium chloride to the red cell lysate provided a clearer supernatant after centrifugation.
A polishing step may be included after d e centrifugation to produce die hemolysate. A filter aid which is eidier a cellulosic, diatomaceous earth, polymer, or silica derivative would be added to the centrifuged hemolysate with agitation. The filter aid would then be removed by filtration or centrifugation. Addition of a polishing step with a filter aid would remove additional stroma fragments, thus phospholipid, which would not be entirely removed by centrifugation.
An alternative to hypo-osmotic shock for lysing the red cells is mechanical disruption such as with a French press or a larger scale cell homogenizer.
The pyrogen content of the four hemolysates processed varied from 0.05 to 0.2 endotoxin units (EU) per ml by the Limulus Amoebocyte Lysate (LAL) pyrogen test. Downstream processing by anion exchange or cation exchange chromatography would remove die last traces of endotoxin and phospholipid.
Example 2 - Reduction of virus content
If die bovine blood is contaminated with viruses, a certain amount of viral load would be removed during washing of die red cells with low speed centrifugation. Another percentage would be removed with the high speed centrifugation of die hemolysate, and would reside in die pellet. Another aliquot would be removed in a polishing step.
Further reduction of viral load would be by microfiltration and/or ultrafiltration.

Microfiltration with a 0.025-0.040 micron filter will gready reduce die titer of most viruses. An Ultipor N66 nylon 66 microfiltration cartridge from Pall Corporation has a 0.04 micron porosity, and would be used to filter die hemolysate prior to further downstream processing. In addition, if needed, a 300,000 molecular weight cutoff ultrafiltration membrane from Filtron Corporatio is available which inserts into a Pellicon production scale ultrafiltration unit from Amico Corporation. The hemoglobin would pass freely through all of these ultra- and microfiltratio membranes, whereas a proportion of the viruses would be retained. Therefore, a combination o these two filtration membranes would gready reduce d e viral load. Each of these filter types shows low protein retention.
In addition to filtration, the viral content ofthe hemolysate or filtered hemolysate could b reduced by die addition of appropriate detergents, ethylenedinitrilotetraacetic acid, the FDA- approved reagent trinitrobutyl phosphate, or a combination of these additives.
Example 3 - Ion exchange chromatography of hemoglobin
Anion exchange chromatography has been utilized extensively in die purification o mammalian hemoglobins (Williams, R.C, K. Tsay (1973) Anal. Biochem. 54: 137-45). Under die proper pH and ionic strength conditions, hemoglobin can be applied to an anion exchanger so that it binds to die exchanger. Proteinaceous and non-proteiπaceous impurities would be removed from hemoglobin during development of the column. Conditions also can be selected under wh*ich hemoglobin has little or no affinity for the selected anion exchange resin. Under tiiese conditions, d e impurities would be left behind on die column. The principle contaminants present in freshly prepared hemolysate are phospholipids, potentially residual virus which was not removed by upstream processing steps, low levels of bacterial endotoxin, and proteins other than hemoglobin. Endotoxin, phospholipid, and plasma proteins are more negatively charged than hemoglobin and should exhibit a higher affinity for the anion exchanger under the conditions selected. Under die proper loading and eluting conditions, anion exchange chromatography should be useful for resolving hemoglobin from these contaminants. In addition, anion exchange chromatography should further reduce die viral load.
We have utilized three anion exchange chromatographic protocols for the purification of hemoglobin: binding at elevated pH and elution with a descending pH gradient, binding at elevated pH and elution with a step gradient of lower pH, and loading under pH conditions in which die hemoglobin does not bind to die anion exchanger, but passes through die column un-retained. All of our experiments have been carried out in a column mode, but it would be expected that similar batch conditions could be developed that would give nearly identical results. The anion exchanger utilized in these experiments was Q-Sepharose Fast Flow from Pharmacia, but it would be expected mat these procedures would work witii a number of commercially available low to mid-pressure anion exchangers that have been developed for protein purification including those with quaternary amine and diethylaminoethyl functionalities.
Cation exchange chromatography also has been utilized extensively for the purification of mammalian hemoglobins (Bucci, E. (1981) Preparation of isolated chains of human hemoglobin, Medi. in Enzvmol. 7fi: 97-106). Under the proper pH and ionic strength conditions, hemoglobin can be applied to a cation exchanger so that it binds to the exchanger. Phospholipid and endotoxin, as well as plasma proteins should have little affinity for the cation exchanger under conditions with which hemoglobin would bind. Under d e proper loading and elution conditions, cation exchange chromatography should be useful for resolving hemoglobin from these contaminants. In addition, cation exchange chromatography would give an additional reduction in viral load.
Two cation exchange chromatographic protocols were designed for the purification of hemoglobin. In die first protocol, the hemoglobin was loaded wider conditions in which it bound to die cation exchanger and was then eluted with a linear pH gradient. In the second protocol, die hemoglobin was loaded under conditions in which it binds to die resin, and then eluted with a step pH gradient. In both of these protocols the phospholipid and endotoxin would pass unretarded through the resin, thus resolving these contaminants from the hemoglobin. A number of commercially available cation exchangers would be adequate. Preferably, a S-Sepharose Fast Flow from Pharmacia was used for its excellent protein binding and flow characteristics. A host of other low-to-mid-pressure chromatographic media with an Sulfopropyl- or Carboxymethyl- functionality would be expected also to perform the needed function. Scaling of these resins is economical. Example 3A - Anion exchange chromatography with a linear pH gradient
This experiment provided a demonstration of anion exchange chromatography conditions in which bovine hemoglobin was loaded onto the exchanger and then eluted witii a linear pH gradient. A 2.5 X 5.5 cm column of Q-Sepharose Fast Flow was prepared and cleaned by overnight exposure to 0.5 M NaOH. The column was washed with 100 mM Tris, pH 8.5 until die pH of the column effluent was 8.5. The column tiien was equilibrated witii 50 M Tris, pH 8.5. Five ml of a sample of freshly prepared bovine red cell hemolysate containing approximately 350 mg of hemoglobin were diluted to 10 ml with 100 mM Tris, pH 8.5 and loaded onto the Q-Sepharose column at 2.5 ml per minute. Following loading, die column was washed with 50 mM Tris, pH 8.5 until die absorbance of the column effluent returned to baseline. The column then was eluted with a linear gradient consisting of the starting buffer versus 50 mM Tris, pH 6.5 over ten column volumes. The hemoglobin eluted essentially as a single major peak with several minor peaks eluting before and after it. The single major peak accounted for 95% of the hemoglobin which was loaded onto d e column. All ofthe chromatographic procedures were carried, out at 5° C
Example 3B - Anion exchange chromatography with a step pH gradient
This experiment provided a demonstration of anion exchange chromatography conditions in which bovine hemoglobin was loaded onto the exchanger and then eluted widi a single pH step gradient. A 2.5 X 5.5 cm column of Q-Sepharose Fast Flow was prepared and cleaned as in

Example 3A. Eight ml of a sample of bovine red cell hemolysate containing approximately 400 mg of hemoglobin were diluted to 16 ml with 50 M Tris, pH 8.5 and loaded onto the column at 2.7 ml per min. Following loading die column was washed widi 50 mM Tris, pH 8.5 until the absorbance of the effluent returned to baseline. The column men was developed widi 50 mM Tris, pH 7.4 and die hemoglobin eluted from the column essentially as a single peak. The amount o hemoglobin present in die Q-Sepharose rich fraction was nearly identical to that loaded onto die column, indicating a nearly quantitative recovery across this step.
Example 3C - Anion exchange chromatography in which the hemoglobin is not retained
This experiment was a demonstration of anion exchange chromatography conditions in which die bovine hemoglobin did not bind to the anion exchanger, but passed dirough the column un-retained. a 2.5 X 5.5 cm column of Q-Sepharose Fast Flow was prepared and cleaned by overnight exposure to 0.5 M NaOH. die column was washed widi 100 mM Tris, pH 7.4 until die pH of die column effluent measured 7.4. Ten ml of a freshly prepared bovine red cell hemolysate containing 880 mg of hemoglobin were diluted to 40 ml with 50 mM Tris, pH 7.4, and 35 ml were loaded onto die column at 2.5 ml per min. Following loading, die column was washed with 50 mM Tris, pH 7.4 until die hemoglobin had finished eluting from d e column. The hemoglobin passed through die column without binding, and die recovery was 88%.
Example 3D - Cation exchange chromatography of hemoglobin
This experiment was a demonstration of cation exchange chromatography conditions in which bovine hemoglobin was loaded onto die exchanger and then eluted widi a linear pH gradient. A 2.5 X 5.5 cm column of S-Sepharose Fast Flow was prepared and cleaned by overnight exposure to 0.5 M NaOH. The column was washed widi 100 mM Bis-Tris, pH 6.0 until d e pH of die column effluent reached 6.0. The column then was equilibrated widi 50 mM Bis-Tris, pH 6.0. Five ml of a sample of freshly prepared bovine red cell hemolysate were diluted to 10 ml with 50 mM Bis-Tris, pH 6.0 followed by loading onto the S-Sepharose column at 2.5 minutes per hour. Following loading, the column was washed with 50 mM Bis-Tris, pH 6.0 until die absorbance of the column effluent reached baseline. The column then was eluted with a linear gradient between the column buffer and 50 mM Bis-Tris, pH 8.0 over ten column volumes. All of die chromatographic procedures were conducted at 5° C
Example 4 - Cross-linking bovine hemoglobin with dimethyl adipimidate
A sample containing 170 mg/ml of purified bovine hemoglobin was treated as follows. The cross-linking reaction was carried out at a hemoglobin tetramer concentration of 1.75 mM. or 112 mg/ml. High tetramer concentrations promote die intermolecular cross-linking of hemoglobin tetramers. The hemoglobin is modified in die presence of 50 mM Tris buffer, pH 8.8 at 4° C. The hemoglobin is modified by die addition of dimediyl adipimidate in equimolar amounts (i.e. 1.75 mM dimediyl adipimidate and 1.75 mM hemoglobin tetramer). The dimediyl adipimidate treatment is repeated eight times at 30 minute intervals at 4°C. For each application a 50 mg/ml solution of dimethyl adipimidate in 0.025 M sodium carbonate, pH 9.25 is added to the hemoglobin solution to a final concentration of 1.75 mM. The hemoglobin solution is vortexed during addition ofthe cross-linker reagent, and tiien incubated on ice for 30 min. After 30 minutes a 0.4 ml sample was withdrawn and added to 4 ml of 0.25 M sodium phosphate, pH 7.0. The phosphate buffer quenches die cross-linker reaction by hydrolysis of die imidoester. Quenching is carried out over two hours at room temperature. The remaining hemoglobin solution is again reacted with a fresh solution of the cross-linking agent. Eight cycles of reaction are carried out over 3.5 hours. In this experiment the specific imidoester dimethyl adipimidate was used. Other imidoesters could also be used.
Example 5 - Size Exclusion HPLC of hemoglobin cross-linked with dimethyl adipimidate
Hemoglobin concentration was determined spectrophotometrically using published extinction coefficients (Benesch, R.E., Benesch R., S. Yung, (1973) Equations for the spectrophotometric analysis of hemoglobin mixtures. Analvt. Biochem. 55: 245-48). Reaction products were monitored after each addition of cross-linker using size exclusion HPLC (SEC- HPLC). Quenched reaction product was diluted to 1 mg hemoglobin tetramer per ml in 0.Z~M sodium phosphate, pH 7.0. The diluted product was analyzed by SEC-HPLC in two separate modes. The first was a Zorbax GF-250 SEC column (Dupont) equilibrated with 0.2 M sodium phosphate, pH 7.0 and run at 1 ml/min. The second was a Superose 12 FPLC column (Pharmacia) equilibrated in 0.1 M sodium phosphate, pH 7.0. Both columns had been calibrated with proteins of known molecular weight.
Successive iterations of cross-linking with dimethyl adipimidate resulted in a progressive increase in the amount of polymerized, tetramer stabilized, hemoglobin. The molecular weight ranges could be defined on Superose 12 SEC-HPLC as follows: 1) % > 400,000 daltons, 2) % > 64,000 daltons and < 400,000 daltons, and 3) % < 64,000 daltons. After four iterations of cross-linking the relative percent values were 1) 0%, 2) 40.64%, 3) 59.36%. After seven iterations die percent values were 1) 4.1%, 2) 50.4%, and 3) 45.5%. After eight iterations die percent values were 1) 7.5%, 2) 50.6%, and 3) 41.9%. The class 3 peak with a molecular weight value less than 64,000 daltons consisted of a single symmetrical peak with a molecular weight of 30,000 daltons. This fraction coeluted with unmodified bovine hemoglobin and is thought to consist mainly of hemoglobin dimer. The methemoglobin content of the hemoglobin after seven iterations of cross-linking was <2.5%. The visible spectrum of die cross-linked hemoglobin resembled that of unmodified bovine hemoglobin.
Example 6 - SDS polvacrvlamide gel electrophoresis of cross-linked hemoglobin
Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was conducted on cross-linked and non-cross-linked hemoglobin samples as described. Hemoglobin samples were diluted to 2 mg/ml in 0.005 M sodium phosphate, pH 7.4, 0.15 M NaCl, and die diluted samples were mixed widi 3 parts of a standard SDS-gel buffer followed by incubation at 37° C for 3 hours. Samples were run on 10-20% ISS mini-gels in the non-reduced form. SDS-PAGE of the cross- linked hemoglobin after 7 iterations of cross-Unking indicated diat 80% of die sample was cross- linked among subunits. At least eight multimers ofthe subunit molecular weight were present, all in approximately equal proportion.
Example 7 - Oxygen equilibrium of cross-linked and non-cross-linked bovine hemoglobin
An oxygen equilibrium measurement was determined for cross-linked and non-cross-linked bovine hemoglobin using a Hemoxanalyzer instrument from TCS Medical Industries. After seven iterations of reaction with dimethyl adipimidate a sample of cross-linked hemoglobin was diluted to 1.5 mg/ml in phosphate buffered saline, pH 7.4, and subjected to analysis. Also analyzed under identical conditions were a sample of unmodified bovine hemoglobin, and a sample of bovine hemoglobin which had been cross-linked in our laboratory with glutaraldehyde using methods as described (Guillochon, D. et al., (1986) Effect of glutaraldehyde on hemoglobin, Biochem. Pharm. 2 : 317-23). The P50 values (oxygen partial pressure at half saturation) for the hemoglobin species were: 13.5 mm Hg for the dimethyl adipimidate cross-linked hemoglobin, 14.25 mm Hg for the glutaraldehyde cross-linked hemoglobin, and 23.75 mm Hg for unmodified bovine hemoglobin. Example 8 - Stability studies on cross-linked and non-cross-linked bovine hemoglobin
Bovine hemoglobin which had been subjected to seven iterations of cross-linking with dimediyl adipimidate, unmodified bovine hemoglobin, and glutaraldehyde cross-linked hemoglobin were subjected to incubation at 36-37° C in 0.125 M sodium phosphate, pH 7.1. The hemoglobin concentrations were 40 mg/ml for dimediyl adipimidate cross-linked bovine hemoglobin and unmodified bovine hemoglobin, and 28 mg/ml for glutaraldehyde cross-linked bovine hemoglobin. At certain time intervals samples of each hemoglobin species were removed, diluted 1/100 in 0.25 M sodium phosphate buffer, pH 7.3 and scanned between 680 nm and 420 nm in a Shimadzu Model 160 UV-visible scanning spectrophotometer. Methemoglobin content was determined by analysis of each spectrum using published equations (Benesch, R.E. et al., (1973) Equations for die spectrophotometric analysis of hemoglobin mixtures, Anal. Biochem. 55: 245-48). Spectra were taken at 22° C - Dimethyl adipimidate cross-linked bovine hemoglobin and glutaraldehyde cross-linked hemoglobin also were incubated at 4° C over a more extended time period. Samples were taken at time intervals, and die mediemoglobin content was determined as described above.
In the 36-37° C incubation, the dimediyl adipimidate cross-linked, glutaraldehyde cross-linked, and unmodified hemoglobin samples contained 2.5%, 8.0%, and 3.5% mediemoglobin, respectively, at the start of the incubation. After 5 hours at 36-37° C d e percentages of mediemoglobin were 8.9%, 44.4%, and 8%, respectively, for the dimediyl adipimidate cross-linked, glutaraldehyde cross-linked, and unmodified bovine hemoglobin samples, respectively.

After 24 hours, die mediemoglobin concentrations were 24.1, 87.9, and 42% for the dimediyl adipimidate cross-linked, glutaraldehyde cross-linked, and unmodified bovine hemoglobin samples, respectively. After 48 hours d e mediemoglobin concentrations in dimethyl adipimidate cross- linked and uncross-Iinked bovine hemoglobin samples were 46.5% and 68%, respectively. The dimethyl adipimidate cross-linked hemoglobin is more stable to methemoglobin formation than unmodified bovine hemoglobin at 36-37* C. The literature teaches that glutaraldehyde cross-linked human hemoglobin autoxidizes at a rate four times that of unmodified human hemoglobin (Guillochon, D. et al., (1986) Effect of glutaraldehyde on hemoglobin, Biochem. Pharm. 25: 317- 23). It was unexpected mat the dimethyl adipimidate cross-linked bovine hemoglobin would be more stable than unmodified bovine hemoglobin with respect to autoxidation.
The stability of dimethyl adipimidate cross-linked and glutaraldehyde cross-linked bovine hemoglobin to autoxidation also was measured after incubation for 28 days at 4° C in die same phosphate buffer as described above. After 28 days d e dimethyl adipimidate cross-linked bovine hemoglobin increased from 2.5% to 9.6% methemoglobin. Over the same time period the glutaraldehyde cross-linked hemoglobin increased from 8.0% to 23.5% methemoglobin.
Example 9 - Preparative size exclusion chromatography of dimethyl adipimidate cross-linked bovine hemoglobin
Preparative size exclusion chromatography was utilized to remove low molecular weight species, i.e. < 64,000 daltons, from dimethyl adipimidate cross-linked bovine hemoglobin. Dimethyl adipimidate cross-linked hemoglobin, 40 mg/ml, was applied to a 2.5 X 90 cm column of Sephacryl S-200 HR (Pharmacia) equilibrated in 0.025 M sodium phosphate, 0.15 M NaCl, pH 7.4, at 4° C The loading volume was 1% of the column bed volume, and the flow rate was 0.45 ml/min. Peaks from chromatography were pooled, and die pooled samples were analyzed by SEC-HPLC on a Superose 12 column (as described above) to determine d e molecular weight distribution of hemoglobin species within diem. The pre-column or load sample also was analyzed on die Superose 12 column.
Preparative size exclusion chromatography separated die DMA cross-linked hemoglobin into two major- fractions. Fraction 1 contained die higher molecular weight cross-linked hemoglobin species, and represented 51.3% of the total recovered hemoglobin. Fraction 2 contained cross-linked hemoglobin of molecular weight less man 64,000. SEC-HPLC of the pre-column cross-linked hemoglobin fraction indicated it to contain 2.6% as > 400,000 daltons, 47.2% > 4,000 daltons, and 50.2% < 64,000 daltons. From analytical SEC-HPLC, purified fraction 1 from preparative size exclusion chromatography contained 4.7% as > 400,000 daltons, 92.7% as > 64,000 daltons, and 2.6% as < 64,000 daltons. This method was successful at removing d e lower molecular weight hemoglobin species.