CA1298552C - Purified hemoglobin solutions and method for making same - Google Patents

Abstract

PURIFIED HEMOGLOBIN SOLUTIONS
AND METHOD FOR MAKING SAME
ABSTRACT
Hemoglobin solutions are purified by heating in the presence of a reducing agent or other deoxygenating conditions such that nonhemoglobin proteins are selectively precipitated and/or virus is inactivated while minimizing hemoglobin loss.

Description

~ 2~55Z

PU~IFIE~ H~IOGLOBIN SOLUTIONS
AND METHOD FOR ~AKING SAME

T~CHNICAL FIELD

Thl~ invention relates to a method for purifying hemoglobin solution~. In par~icular it relates to ~ method for inactivating viruses ~nd selectively removing nonhemoglobin proteins from hemoglobin solutions while only minimally inacti~ating the blological activity o the desired hemo~lobin product.
In the current practice of medicine whole blood or red blood oell containing suspensions are the o~ly oxygen carryin~ fluids which msy be infused into ~patients or trsum~ victim~. Due to the neeessity for matching donor and recipient blood types the infu~ion of red blood cells in any form is restricted to ;~ ~ setting~ in whlch blood ty pin~ and cross-matching may be performed. The typing and cross-matching process :~ 25 may take as long as 45 minutes. As a result o f this . requirement tr~uma vic~ims suffering substan~ial blood 109~ ~U~t now be in~used with non-oxy~en tr~nsporting salt or colloid solution~ until such ~ime as properly ~ ' `

~29l3552 typed and cross-matched blood is availsble. ~lany trauma YiCti~ are therefore subjected to periods of oxygen deprivation which may be highly detrimental or eYen fatal. ~ven in a hospital ~etting patients suff~ring acute blood 109s may not receive blood in a timely fashion due to a shortage of the appropriate blood type.
Another problem associated with the infu~ion of blosd or products dsrived from blood is the ri~k of transmission of ~iral contamination. Various prospective studies have shown that the incidence of posttransfusion hepatitis in recipients of hepati~is ~ - -surface antigen ~e~ative blood collected from volunteer donorQ r3n8es from 4 to 14 percent ~lum and Vyas, Haematologia, (1982~, 15 153-173~o There is al~o the risk of transmis~ion oP the virus causing Acquired Immunodeflciency Syndrome (variously called HTLV-III, L~V or HIV), cytomegaloviru~, Epstein-Barr Yirus or IITLV-I, the putative causative agent for adult T cell lymphoma leukemia. Products derived from animal blood are also at risk since such blood may contsin a number of pathogenic agents including the viruse~ causing rabies, encephalitis, foot-and-mouth disease, etcO
A~ a result of these considerations a number of re~earcher~ have investi~ated the poqsibility o~
using oxygen carrying re~uscit~tion fluidq based on cell-free hemo~lobin ~olu~ions. The ba3ic premise of this work is that by the remo~al of the oxygen-carrying hemoglobin from the red blood cell and its subsequent purif$cation, one m~y elim~nate the blood type specific antigen~ ~nd, hopefully, the bacterial and viral contamlna~ion. While the ly~i~ of red blood cell~ to rel~ase hemoglobin snd the subsequent remcYal of the residual cell membranes (the stroma) have ,............... ..
.. :

;. ..
: : ' , 35~2 indeed been shown to result in the removal of type specific antigsns, there is little data available on the amount of residual viru~ present in the variou~
prep~rations which have been described in the literature. Experience with plasma protein~ uch as ~lbumln suggest3 that viral contamination i8 a problem even wlth blcod derived proteins which have been subiected to the elaborate fr~ctionation ~chemes ~hich are used to prepare theQe products commercially. For ex~mple, ~lbumin prepared b~ commercial fraction~tion procedures from pooled pl~ma sampleQ ha~ a ~igniflc~nt probability of cont~mination with hepatitls Yiru~ if the albu~in solutions are not heat treated (G~llis et ~1., JO Clin. Invest. (1948), 27:239-244;), One would expect a similar situation to hold for hemoglobin solutions. It i8 ther~fore a primar~ objective o~ the invention to inactivate viruses which may be present in hemoglobin solutions.
In U.S. Patents 3,864,478 and 4,439,357 Bonhard and coworkers claim the p~oduction of hepRtitis-sa~e h~moglobin solution~ evidently by ~irtue of th~ fact that the red cell starting materisl wa~ washed and then exposed to ~-propiolactoneO No data were cited, however, to indic~te whether thi~
proc~dure does in fact remove or inactivate viruses in he~globin ~olutions. While cell wa~hin~ may reduce the number of ~iru~0s pre~ent in s~lution, it does no~
remove viruse~ which may be adheren~ to or incorporated within the cell~. Moreovcr, while ~-proplol~ctone ~BPL) can induce Rome ~iralinactivation~ B~rker ~nd Murray (J. A~o Med. A5COC~ D
(1971~, 216:1970-1976) noted ~hs~ hepfltitis infec~ed pl~s~a which was treated with BPL ~lone w~s still able to trsn~mit the disease to human recipient Virus i~ac~ivation wlth BPL often exhibits a "t~iling-off"

:" ' ' ','.' .: . .. :
, ............. . .. ..

1~985;~2 phenomenon wherein a portion of the original virus popul~tion i~ much more resi~tant to inactivation by the agent employed than i9 the bulk of the viruses (Hartman, J. snd LaGrippo, G~A. Hep~titis Frontiers, (1~57) Little, Brown and Co., Burton, Chapt. 33).
Moreover, BPL is a known carcinogen (S~x, N.J., Cancer ~9~L~ Che~lc81s (19~1) 9 Van Nostrand Reinhold Co., New York, p. 404). In U.S. Patent 4,526,715 Kothe and Eichentopf discuss the prepsra~ion of a hepatitis-free hemoglobin ~olution by a method employing washing and filtraeion. While these a~thors demonstrated ~ha~
w~shing can reduce the concentration of viruses in ~olution, the method sugge~ted would not rem~ve white blood cells. Any Yiru~ incorporated into white cells, 6uch as HTLV-III, would not b~ eliminated by this proce~slng step. Such viru3es wou~d, however, be relea~ed into ~olution during cell lysis. Viruses resdily pa~s through misroporous filters, and ultrAfilter~ are known to contain pinholes which allow the p~s~age o particles greater in size than the nominal moleculcr weight cut-off of the ~embrane. The abillty of the described procedure to quantitatively remove viru~e~ associated with white blood cells is ther~fore question~ble. Until now, a procedure which demonst~ably reduo~ produce related virus ti~ers by f~ctor of 10 or more in hemoglobin solutiQns, and which c~n reliably insctiYate retroviruses which may be incorporated into the blood formed elemen~s, has not been disco~ered.
On the other hand, various l~terature and - p~tent ~ource~ disclo~e me~hods for inactivat~ng Ylru~es in blood pl~a protein fr~ctions. An effecti~e method employx dry heat inactiva~ion7 i.e., th@ lyophilized protein which is suspected to bear viral oontam~n~tion is simply heated in the dry state ~2985S~

at te~peratures aboYe ~bout 50 degrees C0 until the desired virsl lnactivseion i6 achie~ed. A
represPntAtive method of this ~ort is disclosed in PCT
publication W0 82t03871.
Another technique also employs heat, but the protein fraction is heated while in aqueou~ solution rather than dry. S~ab~lizer6 are included in the solutions in order to preserve the biological zctivi.ty of the deQired protein. For example, see U.S~ patents 4,297,344; 4,317~086; and European patent applications 53,338 and 52,827. The stabilizers that have been used for this purpose are glycine, alpha- or beta-alanine, hydroxyproline~ glutamine, alpha-, beta- or gamma-aminobutyric acid, monosaccharides, ol~gosaccharides, sugar alcohols, organic carboxylic acids, neutral amino acids~ chela~ing cgents and calcium ions.
These methods are both founded on the discovery that heat will inactivate viruses at a greater r~te thsn the protein~, pro~ided that an agent or stabilizer ~8 present or cond~tions are identified which stsbilize the degired protein but which do not ut the same time similarly stabilize the ~iral contsmlnants.
` Unfortunctely, proteins are known ~o exhibit widely v~rying su~ceptibility to denaturation during he~ting due to differences in their chemical ~nd physic~l s~ruetureO The biologically active form of a protein i~ maint~ined by complex interactions ~etween lt~ con~tituent amino ~cids. These interac~ion~
- i~clude hydrogen bond~ng; salt linkages bet~een charged grOUp8, dipole~dipole interactions, hydrophobic effects and di~persion forces. Although the factorq governing protein stability in general, 8~d hemoglobin Qtability in particular, haYe been '. ' , .,, ,:, .
.,.. : . :;,. .
, ' :. :"
.; '' ~98~52 studied fnr many decades, the thermal stability of a protein cannot be predicted even when the amino acid ~equence is known. Bull and Breise noted a 35 degree C. spread in the denaturation temperature of twenty proteins which they studied wlth no correlation being evident between this temperature and protein structur~l features (Arch. Biochem. Biophyq, (1972) 156:604-612).
Protein ~tability also ~aries as a functio~
of the composition of the medium in which the proteln i~ placed, being sensitiv~ to pHt ~alt concen~ration, the present of detergents or organic solvents, and the presence or ab~ence of ligands which may bind to the protein. For example, .~ome protein~ are easily denatured by acid pH while others are sctually ~tabilized under theqe condition~ (Tanford, Phv~ic~l Chemi~try of Macromolecules (1961) John Wiley and Sons, New York, p, 625; White et ~1., Principals of Biochemistrv ~1978) McGraw-Hill, New York, p.l64)~
The stabilization of proteins by lig~nd binding is a frequent (but not universal) occurence, and has been usad to preserve protei~s during purification. This strategy i~ exemplified by the u~e of a long chain fatty acid such as caprylic acid tc stabilize albumin during heatin8. However, ~ince different proteins bind di~fer~nt lig~nd~, the addition of a lig~nd which s~abili~ea one pro~in doe~ not ncce~sarily ~tabilize another.
It should be empha3ized ths~ protein~
derived from the same tissue (e.g. blood~ or even the ~sme c~ ay exhibit marked differences in thermal ~tabiliey. For example, plasma protein Factor VIII is : very r~pidly insctivated whan heated in solution at 60 degree3 C. while, as noted above, albumin msg survive such te~per~ure when ~tabili~ed with certsin fatty . . .
.... .
, .
... .. .
.,, ~

1~98552 acids. Thi~ illustrates the fact that the optimum conditions for protein stabilization cannot be predicted on the basis of protein source.
With regard to hemo~lobin stability the extant litera~ure i~ particul~rly confusing and often conflicting, ~lemoglobin has long been known to be susceptible to oxidation to ~he met form in which heme iron is in the ferric (+3) form rather than the normal ferrous (+2) ~tate. Methemo~lobin does not reversibly bind oxygen and is therefore non-functional as an oxygell carrier. It is also less stable in solution.
It is therefore universally aocepted that a useful hemoglobin based oxygen earryin~ solution ~hould contain low amount~ of methemoglobin~ but despite years of intense Atudy the precise mechanism by which hemoglobin oxidizes ~ B not yet known. In general, however, hemoglobin ~olutions which are etored cold or even frozen oxidize less rapidly than those stored at hi8her temperatures (Iorio, Methods in Enzymology (1981) 76:57-71). Thus, in general, researchers attempting to preserve hemoglobin struc~ure and functlon a~oid high temperatures.
The relationsllip between oxygen and hemoglobin 3tsbility i9 complex and the literature contradictorya Kikugawa et al., (Chem. Pharm. Bull.
(19~1) 29:1382-1389) claims that deoxyhemoglobin was more ~table thsn oxyhemoglobin during incubation at 37 degrees C. 9 and Rieder (J. Clin. Invest. (1970) 49:2369-2376~ snd Winterbourn and Carrell (J. Clin.
Invest~ (19743 54:67-689) have asserted that deoxyhemoglobin heated in an evacua~ed vessel is more heat stable than oxyhemoglobin heated under ambient oxygen p~rtial pressureO Muller and Schmid reported that deoxyhemoglo~in denatured at a higher tempera~ure than oxyhemo~lob1n when both were heated in a ,............. . ................................ .
... .
. ',: . . .

calorime~er. On the other hand, Mansouri and Winterhalter (Biocbemistry (1973) 12:4946-4949) have noted that in their experiments lowering the oxygen pressure increased the r~te of autoxidatlon. Banerjie and Stetzkowski (Biophys. Acta (1970) 22:636-639), Wallace et ~1., (J. Biol. Chem. (1982) 257:4966-4977), and Brooks (Proo. Royal SDC. London B (1935~ 118:56-577) have ~lso noted a similar phenomenon, leading several of ~hese researchers ~o propose ~ha~ it is ~ctuslly the deoxygenated hemoglobin which preferentially undergoes conYersion to the met form.
Eyer and coworker~ (Mol. Pharmacol. (1975) 11:326-334) found that methemoglobin formation by hydrogen peroxide wa~ much h$gher when the oxidant was infused 15 into solutlons of deoxyhemoglobin as uppo~ed to the oxygen~ted protein. Part of thi~ complexity stems from ~he fact th~ oxygen i~ both a ligand which can reversibly a~sociate with hemoglobin and a reactant which may oxidize the protein.
The complexity of hemoglobin stab~lity is further illustrated by the reported effects of antioxid~nts and reducing cgents. The antioxidant ~scorbic acid h~s been ~hown to both reduce ~ethemo~lobin (Gibson, Biochem. J. (1943) 37:615-618) and to oxidize oxyhemoglobin ~Har~ey and R~neko ~ Brit .
J. H~Dnatol. (1976~ 32:193-203). Reduced glut~thio~e is another antloxidane wh~ eh induce~ me~he~oglobin for~stlcn ~hen sdded to solution~ of purified hemoglsb~n, e~en though it i~ belie~ed to function ~8 a protecti~e ~gent fGr he~o~lobin in ivo (Sampath and C~ugh~y, J. Am. Chem. Soc. (1985) 107:4076-4078). In point o~ faet, many redueinB a~ent~ are known to enhance meehemoglobin ormation even though others exhibit the expected ability to reduce the oxidized ., :;,. ' ... .

5~
g protei~ (Eyer et al., in Biochemical and Clinical Aspects of Hemo~lobin Abnormalieies (1978) Academic Press 7 New York, pp. 495-503; Kawanishi and Caughey in Biochemical and Clin~cal Aspects of OxvRen (1979) Caughey W,S. ed, Academic Pre~s, New York~ pp~ 27-34).
This behavior apparently occurs because, in ~ddition to being ~ble to dirPctly reduce methemoglobin, reducing agents may alQo generate peroxide~ when they are oxidized in other reactions. Thus9 the net effect of addin8 a particular reducing agent depends on which other enzyme~ snd reartan~s are present ~s well as the o~dation reduction potential of the reducing agent.
~ntioxidants have been used in as90ciation with hemoglobins in the past. Hollocher~ "J. Biol.
Chem" 241:9 (1966) observe that thiocyanate decrea~es the heat stability of hemoglobin~
Europaan Patent Application 78961 teaches st~bilizing cro~slinked hemoglobin again.qt oxidation by the use of an antioxidant.
Daland et 81,, ~IJ. Lab. Clin. MedO" 33:1082-1088 (1948) ~mploys a reducing agent to reduce red blood cell hemoelobin ln order to a~say for sickle cell anemla.
Sodium a~corb~te W~5 di~clo~ed to be inefecti~e in protecting the hemoglobin molecule from det~riorstion during prolonged storage. Rabiner et al.~ "~nn. Surg~" 171:615 (1970~.
Hemoglobin solution~ ha~e been proposed for u8e a8 hlood ~ubstitutes, ei~her as a solution of cry~talline hemoglobin or as a polymer crosslinked to - other hemoglobin or other macromolecules such as polysaccharide~ See for example U.S~ patents 4,001,401; 4,OSl f 736; 4,0S3,590; and 4,001,200; and We~t Ge~man Off~nlugungsschriften 3029307 ~nd 2616086.
All of these products are obtained by proce-qses which .: , ,.............. '' .:' :. , :, `

: :

1~2 use human red blood cells from whole blood as a ~tar~ing ~aterial. The hemoglobln is sep~rated from the formed matter (including stroma) of the red cells by lysis and centrifugationD followed by processing in accordance with known techniques, including ~ubstikution with pyridoxal groups. These methods ar~
not concerned with assuring ~hat any viruses present in the whole blood are removed.
Taken as a whole, the prior art suggests only that hemoglobin stability is a complex function of solution composition, p~l and temper~ture with no indication flS to whether or how a solution of hemoglobin might be heated to 60 degrees C. or more for a prolonged period of time. This is evidently the reason why the successful heating of hemo~lobin solutions for the purpose of inactivatin~ viruses has ne~er be~n attempted, desplte the immense amount of research wh~ch has been performed on hemoglobin structure and function, and the intense lnterest in the u~e of the protein in the formulation of o~ygen earrying intravenolls solutions. Surprisingly, I have discovered a set of conditions ùnder whioh hemoglobin may be heated at temperatures of 60 degrees C. or more for 10 or more hours with li~tle loss of struc~ural integri~y or oxygen transport oapacity9 making possible the heat in2ctivation cf viru~ in hemoglobin, whether cro~ ked or otherwise.
Another problem in the development of a hemoglobin b~sed oxygen transport solution is the purification o the hemoglobin. Commonly used ~ethods for the obtaining of partially purified hemoglobin solution (so oalled "~troma-free hemoglobin") employ '' ' ' ' ' . ' `: , ... . .
.: ,................................ .

~29~3552 --11~

cell lysis with solvents or by exposure to hypotonic conditions~ followed by the re~o~al of membrane frsgments by filtration, centrifugation and/or precipitation under acidic condition~. See for example (Rabiner et al., Ann. Surg. (1970), 171:615-622; Feola et 81., Surg. Gyn. Obstet. (1983), 157:399-408; Bonsen et al., (1977) UOS. Patent 4,001,401; and Bonhard (1975) U~S. Patent 3,~64,478. While the~e procedurea re~ove sub~tantial a~ounts of the cell stroma the~ do not effectively remoYe many sf the ~ontRminating oluble proteins. If one wishes to modiy the he~oglobin chemically, especially wi~h nonspecific reagentq such ~ glutar~ldehyde, the pre~ence of intraoellular proteins results in a variety of byproducts which complicate sub~equent purification, reduoe yields snd incre~se the prob~bility of product toxicity. To miti8~te such problems, researchers have frequently purified hemoglobin by variou~ chromatographic techn~ques.
Although the~e techniques are oapable of ef~ective purification, they are often laborious and require the use of expensive chromotographic media which are dlfficult to sterilize and depyrogenate. Other purlfication techniques, such as electrophoresis or ultra-centri~ugation, are not amenable to l~rge scale production. In ~he present inYention~ a subs~antial purification is achieved by means of a simple heatin8 proces~ which can be readily performed in l~rge scale production with equipment which i8 easily ~terilized 3~ and depyrogenated. Therefore, by this invention one may purify hemoglobin solutions ~hrough ~elective removal of nonhemoglob~n prot~ina without denaturing &
substantial portion of ~he he~oglobin 80 that it beco~es insapa~le 4f performing i~ oxygen tran-cport function in ViYo.

.. ,. : .
. "" ,.:: ' : ' :, :' ' , 1~9~55~

The term "hemoylobin" as used herein is generic for oxy-, carboxy-, and deoxyhemoglobin, as well as pyridoxalated (covalently bonded ~o pyridoxal groups by reaction with pyridoxal-5'-phosphate) andtor crosslinked derivatives thereof unless otherwise stated. Crosslinked derivatives include glutaraldehyde, ring-opened diol, and 3,5-dibromosalicyl-bis-~umarate (DBBF) crosslinked hemoglobin, among others~ The~e hemoglobin derivatives are well-known.
SUMMARY OF INVENTION
I have now discovered a me~hod for reducing the risk of ~iologically infectious virus in hemoglobin-containing compositions~ and removing heat precipitable nonhemoglobin proteins, which comprises: heating a substantially cell-free hemoglobin solution ~t a temperature of 45 degrees to 85 degrees C., while maintaining said hemoglobin in substantially its deoxyhemoglobin form, to inactivate virus present without substantially inactivating said hemoglobin. Additionally or alternatively, the same method can cause certain nonhemoglobin proteins to be selactively precipita~ed, also without substantially biologically inactivating the hemoglobin. This method may be used to accomplish either or both of the above purposes of viral inactivation and 2S precipitation of nonhemoglobin protein. Hemoglobin may be deoxygenated by any desired method.
In accordance with an a~pect of the invention, the method of inactivating virus present in substantially cell-~ree hemoglobin solution comprises heating said hemoglobin at a temperature o essentially 45 degrees to 85 degrees C., while maintaining said hemoglohin in - substantially its deoxyhemoglobin form, whereby substantial inactivation of said hemoglobin is avoidedO
In accordance with anokher aspect of the invention, the method of passing a solution of substantially cell-free hemoglobin through diffusion cell m~ans, said diffusion call means having membrane wall m~ans along 85S~2 12a which said hPmoglobin solution flows, said membrane wall means being capa~le of passing oxygen but not hemoglobin through said membrane wall means while circulating inert gas along the side of said membrane wall means opposed to said hemoglobin solution, to cause removal of oxygen from the hemoglobin solution and conversion o~ other Eorms of hemoglobin to deoxyhemoglobin; and thereafter heating the resulting deoxyhemoglobin solution at essentially 45 degrees to 85 degrees C. in an oxygen-free en~ironment to inactivate virus present and to precipitate heat-precipitatable nonhemoglobin protains without substantially inactivating said hemoglobin.
In accordance with another aspect of the invention, the method of precipitating nonhemoglobin proteins from a hemoglohin solution without substantially precipitating or deactivating said hemoglobin which comprises heating substantially cell free hemoglobin solution at a temperature of 45 degrees to 85 degrees C., while maintaining said hemoglobin in substantially its deoxyhemoglobin form, whereby heat precipitable nonhemoglobin proteins present are precipitated.
In accordance with another aspect o~ the inven'-ion, a method for reducing the viral infectivity nf the hemoglobin composition suspected to contain such infectivity, compris~s heating the hemoglobin composition and a reducing agent together under a temperature and time such that the virus is rendere~ inactive but the hemoglobin is not substantially inactivated, the amount of reducing agent present being effective to maintain hemoglobin in the deoxy form during heating.
In accordance with another aspect of the invention, a method for reducing the viral infectivity of a hemoglobin composition suspected to contain such in~ectivity comprises heating the hemoglobin composition and a reducing agent together under a temperature and time such that the virus is rendered substantially inactive, but the hemoglobin is not substantially ., l2b inactivated, said reducing agent havi~g a reducing potential which is greater than ascorbate and being present in an amount effective to maintain hemoglobin in the deoxy form durinq heating.
In accordance with another aspect of the invention, a method for reducing the viral infectivity of a hemoglobin composition suspected to contain such infectivity, comprises the heating o~ the hemoglobin composition and a reducing agent together under a temperature and time such that the virus is rendered substantially inactive but the hemo~lobin is not sub~tantially inactivated, said reducing agent heing selected from the group consisting o~ redox dyes, ~isulfoxy compounds and sulfhydryl compounds, said reducing agent being present in an amount effective to maintain hemoglobin in the deoxy form during heating.
In accordance with a further aspect of the invention, a method for reducing the viral infectivity of a hemoglobin composition suspecked to contain such infectivity, comprises heating the hemoglobin composition under conditions which cause hemoglobin to assume and maintain its natural deoxygenated form, at an elevated temperature and time such that harmful virus present is rendered inactive but the hemoglobin i5 not substantially inactivated.
In accordance with another aspect of the invention, an aqueous solution of cross-linked hemoglobin iR free of act~ve virus, and has a p50 under physiologic conditions of at least 26 mm. H~.
In one embodiment, the hemoglvbin solution may be deoxygenated by admixture with a ~hemical reducing agent which causes the hemoglobin to be converted and maintained in its substantially deoxyhemoglobin form and he~ted in the presence of , ~
,, lZ~8552 thi~ reaBent. Alternati~ely, the hemoglobin may be converted into ~nd maintained in its substantially deoxyhemoglobin form and heated in the presence of such ~ reducing ~8ent. Preferably, the hemoglobin may b~ converted into and maintained in it~ ubstantially deoxyhemogl~bin for~ by exposure to inert, essentially oxygen free ga~ or vacuum~ to cause removal of oxygen from the hemoglobin and conversion of other forms of hemoglobin to deoxyhemoglobin. One ma~ntainR the deoxyhemoglobin in an oxygen-free environment during the ab~Ye-described heating, for accomplishing either or both of the above purpose~. Specifically~ the h~mo~lobi~ ~ay b~ exposed to such gac or vacuum through an oxy~en permeable, he~oglobin~retaining membrane, as described for example in the article by Robert Schmukler et ~1~ Biorheolo~~ (1985) 22:21-29.
Mor~ spec~fically, one may pass a solution of the substant$ally cell-free hemoglobin through di~fuslon cell means, the diffusion cell mean.~ having membr~ne wall means along which the hemoglobin solution flows, such membr~ne wall means being cspable of pa~sing oxy~n but not hc~oglobin through the membran~ wall mean~, while cireulating inert gas ~long the side of the membrane wall means oppo~ed to the hemoglobin solution, to c~use removal of oxygen from the hemo~lobin solution snd conversion of o~her forms of he~oglobin to deoxyhemoglobin. One ~hen heat~ the resulting deoxyhemo~lobin solution a~ es~entially 45 degrees to 85 deBre~s C. i~ ~ oxygen-free environment to inacti~ate virus pre~ent and/or to precipitate heat-precipit~table nonhe~oglobin protei~s without substantially in20~ivating the hemoglobin.
Prefer~bly, the flow volume of circulating, inert gas is at least 5 times th0 flow volume of ~he - . 35 hemoglobin solutio~ passin~ ~hrough the diffusion cell S5~

meanB~ and, most preferably, from about 10 to 50 timeQ
the flow volume thereof, ~lthough there really is no significant upper limit to the flow volume of circulating, inert gas that may be used apsrt from economic considerations. Typically, the circulating, inert gas may be nitrogen or argon, and the heating temperature m~y be from about 45-50 degrees C. to 85 degrees C. For example, a time of heating of about lU
hour~ at about a temperature of ~bout 60 degrees C, can provide excellent results both in ~he precipitation of ~onhemoglobin proteins and in the inactiYation of ~irus in hemoglobin solutions in ~ccordance with the proce~s of thiQ invention.
Prior to heating, the pH of the solution is prefer~bly Ad~usted to between 6.0 and 9.0 to inhibit methemoglobin formation and hydrolysis. One then heats the resulting deoxyhemoglobin solution at preferably essentially 55 degrees to 80 degrees C. in an oxygen-free environment to in~otivate virus present and/or to precipitate nonhemoglobin proteina without sub~tantially inactlvating the desired hemoglobin derivative. More specific~lly, a time of heating of about 8 to 12 hours, for ex~mple 10 hours, ~t Abou~ a temperature of about 60 to 75 degree~ C. can provide excellent results both in the prec~pitat~on of nonh~moglobin proteinQ and in the inac~iv~tion of virus in hemo~lobin solutions in ~ccordance with the proce 8 of ~hi~ invention.

DETAILED DESCRIPTION OF THE INYENTION

Biologically active hemQglobin is hemoglobin which 1~ c~pable of performing in vivo or in vitro the oxygen transport function of nati~e hemoglobin.
However, it is not necessary for the hemoglobln to ' ,.

.

~29~5S2 function with the efficacy found in its red blood cell environment. Rather, a comparison is made between the material without the heat treatment her~in and a comparable lot sfter such heat treatment~ Thi~
oomparison can be made with in ViYo or in vitro assays already known in the art, for example measurement of arter$ovenous oxygen differences in ~he rat after exchange transfusion with the test composition, by changes in ~he absorption spectrum of the hemoglobin before and after treatment9 or by direct determination of ~he oxygen binding charact~ristics of heated and unheated hemoglobin. Hemo~lobin that is biologically inacti~e, for example, may have been converted ~o methemoglobin~ had its protein component denatured, or has been otherwlse adver~ely impacted by heat or other means.
Hemoglobln compositions include the hemoglobin derlva~ives discussed sbove, nativ~, ~ub~tantially purified heMoglobln, or crude red blood cell llemolysates. Ordinarily one will not be interested in methemoglobin or it~ derivatives because they are not biologically efficacious.
Suitable hemoglobin compositions may contain at lease 99% hemoglobin by weight of total protein, a totsl pho~phol$pid content of less than about 3 ug/ml~
le~a-thsn aboue 1 ug/ml of ei~her phospllatidylserine or phosphatidyleth~nolamine, an inactive heme pigmen~
of les~ thsn 6%~ an oxy~n affinity (P50) of about from 24 to 28 mm. Hg (37 degree~ C., pH 7.4, pC02 of 40 mm. Hg, and 1 m~l hemo~lobin) with a Hill~s constant of Qt le~st abou~ 1.8 and an oxygen combining capacity of ae leas~ about 17 ml. 02/dl. of hemoglobin solution. The~e sp~cif$ca~ions are not critical;
others may be employ~d.

.. .. .
't.' ' ',"' ""' ,~ ,.' ' ' . ' ' ,. ' .. ....

~2~;iS2 A preferred hemoglobin composition for processing in accordance with this inventlon may be an aqueous ~olution cont~inin~ 5 to 15 g./dl. of hemo-globin which is cross-linked predominantlly between the alpha chains by reaction with the diaspirin reagent 3,5-dibromosalicyl-bis-fumarate, wi~h an inactive heme pigment content of less than 6 per cent, a P50 under physiologie conditions of at least 26mm~
Hg, and containing electrolytes at concentrations of 100-160 mmoles/L ~odlum chloride, 3 to 5 mmoles/L
potassium chloride, O to 30 mmoles/L sodiu~ lactate, O to 25 mmoles/L aodium pyruvate, O to 30 mmoles/L
sodium bicarbona~e, and O to 2 mmoles/L magnesium chloride, at a pl! of 7~25 to 7.45 at 37 degrees C., One ~uch preferred solution o$ ~he above described hemo~lobin i8 present in solution at a concentration o 14 g~/dlo ~ having less than 6 percent inactive heme pigment, and exhibiting a P50 of 32 ~m.
HB under physiologic conditions. Such a preferred solution contains about 100 mmo1.es/L sodium chloride, 4 mmolestL potassium ohloride, 10 mmoles/L sodium lactate~ 20 mmoles~L sodium pyruvate, 0.5 mmole/L
calcium ohloride, ~nd O . 25 mmole/L ~agnesiuim ehloride. The pH of the solution at 37 degreeq C. may be 7.4.
The hemoglobin compo~i~ion generally will be di~sol~ed in watcr or buffer solution at a conc~ntration of about from 1 to 40 g/dl, preferably sbout from l to 14.5 g/dl prior to heat treat~en~.
The concentration selcceed will depend upon whether ~he solution is intended to be used as such for ~herapeutic use or to be furth~r processed by ultrafiltration and the like, or lyophilized. In the latter situations the concentr~tion can be any ~hat is conveniently handled in the ~ubsequent proceæsing : .:

55~

step~. Where the product is to be infused it may have a COncentratiOD of about from 13.5 to 14.5 gram~ of hemoglobin composition per dl.
Stro~a-free hemoglobin solutions which ~re useful in this in~ention ean be prepared using conventional techniques. Such techniques include, but sre not limited to, those disclosed in U.S. Patent No.
4,401,652 to Simmond~ et al~, European Patent Application No. 8210684g.1 to Bonhard et al., Cheung 10et al., Anal. Biochem. (1984) 137:481-484 and De Venu~o et 81., J. Lab. Cli~v Med~ (1977) 89:509-516.
Other methods of preparing Ruch ~olutions will be appsrent to those skilled in the art.
The hea~ treatmen~ step can be performed before or after chemical modific~tion of hemoglobin, as long a3 the hemoglobin i9 in the deoxy form.
The hemoglobin composition generally will be di~solved in water at a concentration of about from 0~001 to 4Q g~dl, preferably about from 0.03 to 3 g/dl or 1 to 14 g/dl prior to heat treatment. The concenerstion ~elccted will depend upon the ability to deoxygen~te the solution while maintaining adequate pH
control ~g well as the available or desired hemoglobin concentration for previou~ or ~ub~eguent process 2~ stepc, respectively.
~ s noted above, deoxy~enation may be effected by chemicsl or physicfll means. If a reducing 88ent 1B used it should be capable of fully con~erting hemoglobin to the deoxy form either before or during, 3Q but preferably before, hea~ing without inducing substantial methe~oglobin for~ation. I have found that ~scorbate is relatively ineffective in heat st~bilizing hemoglo~in for ~he purposes herein. Thus the redueing agent ~hould have a greater or more effecti~e reducing po~enti~l than ascorbate. Reduced ., :
" ' ."' -'.' ... ..
. . .
.. ...
. ,:': ,.:

~98~52 redox dyes snd ~ulfhydryl or sulfoxy compounds include many 2cceptable agents. Suitable reducing agents are alkali metal dithionite, bisulfite, metabi~ulfite or sulfite, reduced ~lutathione and dithiothreitol.
Dithionite is preferred, Other preferred reducing a~ents which give an intermediate levcl of protection are compounds which induce hemoglobin deoxygenation during~ but not prior to, heating. The~e include, but are not limited to, reduced glutathione, N-acetyl-L-cys~eine and N-2-mercap~o-propionyl glycine, ~ther appropriate agen~s will be easily determined by routine experiments as described in Example 1 below.
The quantity of reducing agent to be included in the aqueous solution of the hemoglobin compo~ition will vary depending upon the reducing strength of the ~gent, the ~usntity o hemoglobin, the estimated ~iral burden snd or quantity of nonhemoglobin proteins (and~ as a consequence, the i~t~n~ity of the heat treatment), the presence of oxidizing solutes and oxy~en, the`neces~ity for proper pH control, and other factors AS will be apparent to the skilled ar~isan. Accordin~ly, the optimal concentrstion will be determined by routine exp~riments, This can be done by following the in ~itro changes in the hemoglobin U,V.-vislble spectrum a~ described below in Example 2 and in Figure l, to n3sure that only sufficient reducing agent is included to pre~erve A sub6tan~ial proportion of the biological activity of the hemoglobin under the viral inactivation conditions or the like, but no more ~han that ~mount. The amount of ditllionite which can be added i~ limited by the prop*nsity of this agent ~o genera~e acid equivalents upon re~c~ion wi~h oxygen, The ~olution must be adequately buffered to prevent the pH from dropping below 6~0. Since dithioni~e must .,............. .: .;
... , :

~2~8SS2 be added in excess of the amount of oxygen, and thus hemoglobin, in the system, there is ~ complex relationship between the concentration of hemoglobin, buffer~ ~nd dithionite. A useful combination of these parameter~ i6 a hemoglobin conoentration of 1-9 g/dl., dithionite concentration of 10-100 m~, and a sodium phosphate buffer concentration of lOOmM
Various additives may be present in the compositlon in addition to the reducing ~gent, for example, buffers ~uch as phosphate, bi~arbonate or tris (to pH of about 7-B), inorgsnic lons in concentrations generally less than or equal to tllat found in plasma ~e~g., sodlum, chloride~ potassium, ma~nesium, and c~lclum chloride at concentrations of typically no more than about 150 meq/l each and lyophilization stabilizer~ such a~ amino acids or sacch~rides. One may use non-reduclng sugars 3uch 8S
manno~e or su~ar alcohols when lyophilized hemoglobin compositions are heat treated. The concentration of additives ln the hemoglobin solution can vary, depending upon the effect upon hemoglobia ~tability.
For exa~ple, when sodium phosphate (pll 7.4) i9 utllized as a buffer, concentrations above 70mM result in a decrease in hemoglobin stability. This would sugge3t thst hemoglobin stability i reduced in hyper~on~c medi~. The pH of the solution can also vary depending upon the ldentity of the reducing agent, additives snd heat treaement conditions. The pH can range from 6.0 to 9Ø Preferred ranges are from abou~ 7.0 to ~bout 8.5. The most preferred pH is from about 7.4 to about 7.6.
He~oglobin can ~lso be maintained in the deoxy form using var~ou~ solution de~assing procedure~, Thesc inc~udé, bu~ are not limited to, deoxygenation by me~ns of clrculation of the . .
... .

'':': '; .

hemoglob1n solution through a membr~ne gas exch3nge device whlch i~ concurrently flushed with an inert ga~
such as nitrogen as de~cribed, for example, by Schmukler et al. (Biorheolo~y (1985) 22:21-29,), expo~ure of solution to vacuum, and/or ~parging inert gas through ~he solution using, for example, known designs of blood bubble oxygenators as described in U.S. Patents 3,892,534 or 3,792,377. The suita~ility of such procedures will be limited by the extent they promote de~radation of hemo~lobin through foaming, acidifiction~ etc. Foaming may be controlled by adding compatible defoaming agents to the solution, such as caprylic alcohol, if ~uch agent3 do not adversely effect he~t stability. Alternatively, mechanical defoamin~ devices can be used ~o mitigate this problem. Mechanical deoxygenation may al80 be used in con~unction with chemic~l reductants such thst the concentr~tion o~ the latter required to effect complete deoxygenation i~ reduced.
The time and temperature of treatment will depend on a number of f~ctors such a~ Yiral burden, protein concentr~tion, nature of hemoglobin (i.e. cros~linked or not), ~nd the desirability of precipitating unmodified hemoglobin. The first nonhemoglobin proteln~ typically precipitste within 30 minutes to l hour at about 60 de8rees C, As heat tre~tment continue3~ more nonhemoglobin proteins precipitate.
In a preferred embodimen~ wherein viral burden is also reduced, heat treatment i R continued ~o about 10 or 15 hours. Purific~tion of the solu~ion may typically proceed until a reduction of at least 20 percent and prefer~bly at least 50 percent by weight of nonhemoglobin proteins has been achieved. This can be accomplished by the method hereln without a substantial lo~s of hemo~lobin biological activity;

,' , ' ~9i~355~

i.e. only about from 1 to 15 mole percent of hemoglobin is rendered inactive in the ordinary case.
The temperature of heat treatment will range typically about from 45 degrees C. to 85 degrees C., typically 50 to 80 degrees C., preferably about 60-66 degrees C., if the inactivation is to occur over a reasonably brief period of time. The time typically will range about from 1 to 30 hours, but optionally up to 150 hours, preferably 2-10 hours for solutions. The shorter incubations will be used with higher temperatures. The heat treatment of the deoxy~enated hemoglobin 801ution may be effectad by any method for heating such as microwave or infrared radia~ion, or thermal contact by such device~ as resi~tance heaters or water baths.
The temperature of the compositlon is typically lncreased in a manner to avoid localized overheating~
up to a vir~l inactivating and precipitating temperature. The heating time will range typically from about 20 to 96 hours for dry compos~tions. The time and temperature of inactivation will depend upon a number of factors such a~ the viral bioburden, the protein concentration, the na~ure of the hemoglobin (cros~linked or not) and the reducing ~gent consentration ~when present).
The efficacy of the tre~tme~ process for viral kill i~ best assayed by seeding an aliquot of the oomposition to be treated with 8 candidate virus such as sindbis 9 cytome~alo~irus or T4, Suitable methods for such a s~ys are disclosed in PCT
public~tion W0 82/03871~ The redu ing agent concen-tration (when used) and the time and ~emperature of candidate virus inactivation are balanced agsins~ the loss of hemoglobin biolo&ical activi~y, ~o arrive at ... ...
,"'. :' .,~,. .. :.
, ;............. ,: .
.. ,: , .

~2~i~35~2 the optimal condition~ for heat inacti~ation. The inacti~ation of candidate virus should proceed until a reduction of at least 3, and preferably 6, logs of viral actiYity has been achieved. This can be accomplished by the method herein without ~
substAntial loss in hemoglobin biological activity, i.e., only about from 1 to 15 mole percent of hemoglobin is biologically inactive.
The resulting product will contain biologically ~ctive hemoglobin; will be ~ubstantially free of biologically inacti~e hemoglobin; snd will be free of biologically infectiou~ virus. The residues of biologically noninfectiou~ Yirus may be detected by immune sssays for viral antigens9 since these antigens may not be immunologically destroyed by the process.
However, ~iral infectivity assays will demonstrate that virus inactiva~ion has occurred. The presence of viral antigens coupled with a loss in or sub~tantial lack of viral infectivity i~ an indicia of product~
treated in accord with this proce~s, where viral inactivation is desired.
Heat treated solutlon~ may be proceqsed in order to make them convenient for therapeutic use.
Dilute hemoglobin solutions may be concentrated by ultrafiltration and/or lyophili7a~ion.
Ultrafiltration i~ useial also if necessary to remo~e excess reduc~ng ~gent when present, i.e. to reduce the concentration of reducing agent to 8 physiologically acceptable level. This will ordinarily be on the order of less ~han abou~ SmM, but the exact amount will depend on the esti~ated rate of infusion and the character of the reducing agent. For example, reduced glutathione i~ relati~ely innocuous and ~ay remain in the compos~tion in rel~tively hi~h proportions.

: '"' : ............. .... :,.
... . . .
.. ':, .. ::
:... :~
.,.:
,. r ` '.~ .

The heat treatment step can be performed before or after the pyridoxalation or cross-linking referred to above. Prefer~bly the hemoglobin is pyridoxalated 3nd crosslinked before heat treatment.
This helps to ensure that any viral contamination which may occur during manufacturing is also dealt with. If the amount of reducing a8ent used d~ring heat ~r~a~ment is physiologically acceptable 3 then the hea~ing can ocour in final fill~d containers such as b~gs or Yials.
The hemoglobin composition is advantageously in aqueouq ~olution when heat treated, but dry composition also can be heat treated. For example, if the hemoglobin composition i~ intended for long-term storage it m~y be lyophilized or dried from a solution containing the reducing agent, and then heated.
St~oma-free hemoglobln solutions which are use~ul in this invention can be prepared using conventional techn~gues. Such techniques include~ but are not limited to, those disclosed in U.S. Pstent No.
4,401,652 to Simmonds et al., European Patent Application No. 82106~49,1 to Bonhard et al., the Cheung et 81. article, "The Preparat~on of Stroma-free Hemoglobin by Selective DEAE-Cellulose Ab~orption,l'Ana~ytical Biochemistry 137 pp. 481-484 (1984) ~nd De Venu~o et al., "Characteristics of Stroma-Free HemoglDbin Prepared b~ Crystallization,"
J~ Lab. Clin~ MedA 89:3, p, 509-516 ~1977. Other methQds of preparin~ such solution~ will be apparent to tho~e skilled in the art.
For ~he purposes of thi6 invention the reducing agent, when used, may be a subst~nce or chemical or phy~ical inter~ention that prevents hemoglobin denaturation by main~aining the hemoglobin in the deoxy form during heating. ~educing agents .
.'' '., 3S5;~

compri~es chemireductants which convert hemoglobin to the deoxygenated form. Preerred agents convert oxyhemoglobin to the deoxy form without consistent methemoglobin formation.
As st~ted above, hemoglob$n can also be maintained in the deoxy form using various solution degassing procedures. These also inclu~e but are not limited to bubbling with nitrogen gas, sparging with inert gases, and exposing solutions ~o a vacuumO The 1~ &uitability of such procedures will be limited by the extent that they promote degradation of hemoglobin, e.g. through foaming, acidificatlon9 etc.
The concentration o hemoglobin preferably present in solution will vary dependent upon the identity of the reducing agent utilized and subsequent processing steps. Where the reducing agen~ i6 a chemireductant, the concentration of hemoglobin will genera~ly vary from about O.OOl ~o about 40 g/dl. The preferred concentration ranges from sbout 0~03 to 3 and up to about 14 g/dl~ For example~ if sodium dithionite is the reducing agent a~d greater concentrations of hemoglobln are used, the amount of dithionlte which must be sdded to sustain the unoxidized state may caus~ the solution to become too acid. In such cases, the probabllity of nonspeciic precipitation may be increa3ed. If the reducin~ agent i8 a phy31c~1 tnterven~ion, e.g. spsrging or diffusing with inert gas, the ~cidity problem is eliminated.
Under such circum~tance~ hemo~lobin concentrations c~n ra~ge from about O~OOl to about 30 g/dlo ~

BRIEF DESCRIPTION OF DRAWINGS

Fi~ a graph which discloses the . .
.: :
:. .
: .; , . ::
..
. . .

~9~35S2 atabilizing effect of the deoxy form of hemoglobin deoxygenated by dithionite reducing agent in hemoglobin after heating in solution at S6 degrees C.
for lO hours.

Fig. 2 is a achematic view of appar~tus for deoxygenating hemoglobin solu~ions4 Fig. 3 i9 an elevational view of spparatus for gas-sparging hemoglQbin solutions.

The following example~ are in~ended to be illustra~ive, and should not be construed as limiting the scope of the invention.

Thi.q contemplated procedure i5 illustrative of the manner in which hemoglobin compositions may be treated in accord with this invention. A solution is prepared which contains 1g% ~troma-free hemo~lobin, 30 mM of sodlum dithlonite and sufficient sodium bic~rbonate buffer (.1 to ~3M) ~Q maintain the pH at 7.5.~ One hundred ml. of this solution i8 sealed in 3 ZO gla~s ~i~l 80 as to leave no gag head space~ and then heated at ~0 degreçs C. for lO hours by im~er~ion in a water bath. Af~er heating~ the solu~ion is removed from the bath. After hesting, the solution is removed from the vial, diafil~ered over a 30,000 MW cutoff membr~ne to removs excess dithioni~e and to adjust the ionic contant of ~he medium, concen~rated by ultrafiltration to a hemoglobin eon~ent of 14 g/dl, and pa~sed through a 0.2 micron filter to remo~e any :' ..

~9~3S52 particulate ma~ter ~nd to remove bacteria~

Two sliquots of stroma-free hemoglobln ~olu~ion prepared as above were diluted to a S concentration of 0~04 g/dl in 0.1 M sodium phosphate buffer solution, pH 7.4. One aliquot was admixed with sufficlent sodium dithionite to give a final concentration of 92 mM and quickly sealed into a ~las~
vial with no headspace. The other aliquot was sealed into a similar vial but without the added dithioniteO
Absorption spectra over the range of 400-700 nanometera were ~aken of both samples directly from the vials. These spectra revealed that the sample contsining dithionite was completely deoxygenated (as ~hown in Figure 1) whereas the other sample exhibited a typical oxyhemoglobin spectrum. Both sample~ were incuhated at 56 de8ree~ C. for 10 hours and, after cooling to room temperature, absorption spectra were agaln taken. These spectra revealed ~hat 2D the hemoglobin in the deoxygenated sample W89 ~irtually unchanged, as shown in Fi8. 1, whereas ~he absorption spectrum of the oxygenated sample was i~dlcative of a highly degraded sample. When the ample heated in the deoxy sta~e W89 dialyzed to remo~e ehe dithioni~e a normal oxyhemoglobin ~pectrum wa3 ob~ained. Thus, biological activity of the hemoglobin was re~ained during heating in the deoxy, but not the oxy~ statc.
.

.:.': , .,,:, :''. :
'' ; : , ... ..

The procedure of Example 1 i8 repeated in this contemplated exa~ple with the heated te~t co~position coa~aining reductant. This compo~ition w~ dl~ided 5 into 3 ~liquots which re~pectively were seeded with sinbl~, enceph~lo~yoc~rdities (EMC)~ ~nd ~deno type 5 viru~ 80 that ~he concentrstion of ~iru~ was, reQpecti~ly~ 6-7 log 10 plaque forming unit~
(PFU)/ml, 4 log 10 PFU/ml and 4.5 1O8 10 ti9au2 culture 50~ infecti~e dose (TClD-S0)/ml.
The TCID designatlon may be explained a~
follows: In biologic~l qu~nt~t~ion, th~ end po$nt i3 u3ually taken 8~ the dilut~on ~t which a certain propor~ion of the t~st sys~em cell~ reac~ or die. The 100% end point i~ frequently u~ed. However, it~
accuracy i8 gr~atly affect~d by s~all chance vsriation~ ~ desir~ble end point i8 one representing a situation i~ which one-half of the test system react~ whil.e the other one-half doe~ not. The be~t ~ethod i8 to use lnrge ~umb~rs of test ~ysteos at clo8ely spaced dilutions near the ~alue for 50~
re~ction and then interpolate a correct v~lue. The negative logarithm of the TCID end polnt titer i8:

. 5~

25 [Neg~tive log.~ ¦~uo of % ~ortality ~ lo~arithm of hi8hest _ st each diluti~n -0.51x of : virus ~00 / D~lution :~ .concentr~tio~l _ _ \ used :.............. . :.. : .
.: . ". ...

:.,,.; ^:: . :.
., .- .: .
. :; ..
:,., ::

i52 The ti~sue culture 50% end point represents a Yiral titer thae 8i~es rise to cytopathic cha~ges in 50~ of the cell~ in an inoculated culture. In applying the sbo~e technique for determination of concentra~ion, lo~arithmlc dilution~ are prepared ln minimum e3sential ~ediu~ plus 2~ fetal calf ~erum. 0.2 ml of e~ch dilution i~ added to repllc~te cultures of BGM~
(Buffalo Green Monkey Kidney) cellq in mierotiter pl~es. The inoculated ~ultures are incub~ted at 36 de8ree~ C. undsr 52 carbo~ dioxide and ohserqed microscopi ~ oYer a period of 7 ~o 8 days. The percent ~ortality of cells in a culture at a glYen dilution i~ determ~ned b~ observing for cellul~r degener~ti4n, as evidenced by refractile cell3. The TCID~50 can then be calculated as ~hown above~
The EMC and ~indbis ~iruq infeeti~e tieer is obtain*d by pr~pari~g dilution~ of viral suspension as deseribed abo~e. BGMK cell monolnyers were prepared in 35mm petri diches~ Viral ud~orp~ion to the cells 20 W~3 initisted by adding 0.2 ml of suspe~ion to the monolayer. After 1 hour~ the monolayer was overlaid with 2 ml o~ nutrient agar mediu~ ~nd incubated for 24-72 hours at 37 degree~ C. The plaques which formed were then ~ade Yi~ible by staini~g the cells with neutr~l red at 1:2000 by weight in ~aline.
Th~ resul~s with ~iru~ vere Rub~ected to r~gre~sion 8~al~ with the method of lea~t ~quares to allow the fltting of ~ linear lin~ to th~ data ~nd plotted. Similar re~ults were obt~ined vith all 30 Yiru~es- The viral ~nfecti~e ti~er in all three aliquot~ reduced ~ig~ificantly by the ~e~hod of heat treatment~ thereby reduein~ the risk of pa~ient i~fec~ion b~ hep~tl~is or other Yiruses~

.: ,.
.. .. : ,.
.

The method of Ex~mple l was repeated in this contemplated example except that the stroma-free he~oglobin had been crosslinked b7 3~5-dibromo~alicyl-bl~-fumar~te ~nd ~ubsequenely pyridoxalated in accord with the ~ethod disclosed by Tye et al~, ln Bolin et al., edltor~, ~dv~nce~ in Blood Substi_ute Resf rch, New York, Alan ~. Lis~, (1983) and litera~ure ci~ed therein~

EXAMPLE 5 - ~

An ~liquot of stro~-free hemoglobin (SFH) containing 8 g/dl SFH and prep~red by standflrd techniqu~ Wa8 diluted with seven volume~ of i80t~nic ~odium pho~phate buffer solution, pH 7~4, to 8ive a ~olution (1 g/dl) in SFHo Sodium dithionite wa~ added to this solution to give a fin~l ooncen~ration of 8.7 mg/ml and the p~ ad~uste~ to 7.5 with ~odium hydroxide. This 801ution W~9 then sealed into airtigh~ cont~lner~ which were he~ted ~t 60 degrees C~
20 for 10 hours. After coolin~ to room teoperature, ~he solutlon~ were ce~trlfuged ut 5000 x g. for 5 m~. And the ~upern~tant recovered ~nd respun to remo~e any rs~idual part~cul~te ma~er. The pellet re~ulting fro~ the origin~l e~ntrifugation ~a~ wssh~d five time~
25 in i90to~10 sodium pho~ph2te buffer, pH 7~4t ~d fin~ly resu~pend~d in a ~inimum ~olu~e of the s~me buff~r 801ution, Aliquots of the SFH 801ution before heating~ the supernatant obtained after heating and centrlfugatio~, ~nd ~b~ wa~hed pre~ipitate obtained 30 ~fter he~tl~g were solub~lized i~ 1.5% SDS containing I ~g/ol di~hiothreltol and ~nalyzed by pol~acrylamide gel eleetrophore~t~. The results of this a~alr8i5 demonstrated th~t the level of impurities was redue~d .' '~ "'', ' . `
. ~
., .

5~2 in heated SFH solution~ as compared to the originsl Ullhe8~ed 501ution and th~t the pellet consiYts pr~dominately of impurity proteins.

In thi~ example, ~rom~-fr~e ox~hemoglobin solu~ion i5 treated phy~ically, rather tha~
chemic~lly, to ~xch~nge di~ol~ed oxygen from the ~olution with physiologically inactive gas to remove oxy~n from the o~yhemoglobin moleule, prior ~o h~tin8 in a manner previou51~ de.qcribed to inacti~ate viru~ and to precipitate nonhemoglobin proteins as desired. The pre~ent approsch provides a gentl~ and biocompa~ible proce~ ~or relstively rapid ~nd complete deoxygenation of hemoglobin ~ith conservation of its biologlcal ~ctl~ity (~.e. for~ation of littls or no methemoglobin)~
Referring to Fig. 2, a typical apparatus for deoxygenating hemoglobin is shown in sch~matic form.
~emo~lobi~ solution i~ placed in dispen~lng v~s~el 10.
A convcntion~l roller pu~p 12 pump~ the h~oglobin ~olution through line 14 to one end of a ~e~brane oxy~enator 16, for example a Model No. 08-2A of Sci-Med Life SystQ~s, IncO of Minneapolis, Minnssota~
Ater p~ing throu~h the membr~ne o~ygenstor (vhich i~ uAed a~ a diffuaio~ de~loe herei~7 ~ot ~8 an ox~genator~ the hemoglobin ~olueio~ retur~s to di~pensin~ ~o~l 10 thrDugh lin~ 18.
A te~pera~ure probe 20, pr~3qure g~uge 22, . and ~acuu~-ga0 lln2 24 ~ay CORneCt to ~e~gel 10, wi~h line 24 belng co~trolled by relief ~lv~ 26. This permlt~ the evacu~tio~ of disp~nsing ves~el 10 through l~e 24 ~o remo~e oxyge~ fro~ the ve~sel.
Ox~gen-free ga~, for example, ~itrogen or argon, ~ay be delivered rom gas ~ouro~ 28 through a .,'', ':.
. ' con~entionRl oxygen trsp 30 lnto multipl~ w~y ~alve 32 by line 34. V~cuu~ line 36 connects to any convention~l vacuum ~ource, tD proYide suction to line 24, and al80 to Yacuum llne 38 which co~muni~tes wi~h th¢ outlet of the ~s side of membrane oxygen~tor 16~
Ga6 line 34~ in turn, may communicate throu~h multlple way valve 32 to line 40, which paBse~ through flow er 42 a~d line 44 to the ga~ en~ry port 46 of oxygen~tor 16. Thu~, by ~ppropr~ate control of multiple way valve 32, oxygen ~ ay be removed from ~essel 10~ and ~hen g~ from qource 28 ~ay be directed through oxygenator 16 in its ~a~ side i~ couoter current ~an~er to the flow of he~oglobin solution on the l~quid ~ide of oxygenator 16. ~ence, by a dif~u8ion proce~, the hemoglobiR in the ~olu~lon i8 deoxygen~ted to deoxyhe~oglobin.
Following the deox~xenatioD prOCe~B, dispen3in~ vessel 10 may be disconnected from the rest of the sys~em without per~tting the entrance of oxygen, and heated in accordance with coDditio~s described aboYe, typically in the sbsence of any added chemical reduc~n8 ~gent~, to inactiva~e qirus in the hemoglobl~ solution, and to precipit&te nonhe~oglobin proteins.
~x~gen trap 30 ~ay b~ O~iolear Model No.
DPG-250 of LabCleur of Oakl~nd, California. The ~ultiple wa7 Yalve 32 may be one sold by Kont~s Glass Co.
of Vineldnd, New Jerse~, with added stopcocks 48 bein8 from the Rotaflo Comp~n~ of ~a~landO Oth~r eo~ponent~
~a~ be of conventional desig~ and ~re com~ercially aYall~ble.
One msy deoxy~e~ate and pre~surize dispensi~g ~es~el 10 by pumpi~g 8a8 from ~ourc~ 28 into the dispensing ve-~sel. This can be accompli~hed by appropri~te control of ~ultiple w~y valve 32 and roller pu~p 12 9 follnwed by eYacu~tion of container . .
.'.''. ':'', :.. '. "
. . ..

~85S2 10. One may then iMpose a vacuum on the g8S side of membrane oxygenator 16 up to about minus ~0 inches of mercury, then adjusting the flow of gas from source 28 th~ough the oxygenator ~o the desired selected flow rate. The flow of hemoglobin solution fro~ dispensing vessel may be pressurized up to about +5 psi. with this condition remaining throughout the deoxygenation procedur The course of deoxygenation may be monitored by sampling hemoglobin solution through a flow cuve~te~
~ A) In this particular procedure, making use of the described apparatus of Fig. 2, one liter of stroma-free oxyhemoglobin ~olution containing one gram of oxyhemoglobin per deciliter, a pH of 7.0, and buffered wlth 10 m~llimolar sodium phosphate ~olution was allowed to flow through membrane deviee 16, having a membrane area of 0.8 Aquare meter, and into dispensing vessel 10, which had a capacity of 10 liters~ Vessel 10 was deoxygenated in the manner described above and slightly pressurized with nitrogen (+35 kPa).
The solution Wa8 then circulated through thè
system by per~staltic pump 12 at a rate of 150 ml./min, whlle oxygen-free nitrogen was passed through the gas ~ide of membrane oxygenator 16 at a ratc of 2,000 ml./min., with ~ vacuum of -500 mm Hg being applied at gas outlet 50 of membrane device 16~
This ratio of ga~ to hemo~lobin solution ~low W8S kept relatively constant st 13 to 1 throu~hou~ the prooedure.
During the process,:samples were taken from the system under nitrogen blanket, and the absorbances and spectra were recorded using a flow-through cu~ette having ~ 0.2 cm. path leng~h. The results ~re presented in Table I below, and indicate ~hat complete deoxygenation (99.7%) was achieved after 60 minutes cf .... .
' ;'. ' .:.:.

S5~

eirculation:
T~BLE 1 D~oxy~n~t~on% Oxy- ~ Deo~y Z Mes % Hb Ti~e (Min.) hemoglobin Hb Hb Ssturatio~

0 95.~ 2,~ 1.8 97.5 40.~ 58.3 1.3 4~.9 ~0 16.6 82.~ ~8 16.8 4~5 9~ 0.6 4~6 4~ 0.9 9808 ~4 0.8 10 60 0 99.7 0.4 o ~ote: Hb . heooglobin (B) The procedure of Exa~ple ~ (a) was repeated, exoept the r~tio of g~9 to liquid flow rste was retuc~d to 2 to 1 by uaing a nitrog@n flow rate a~
1,000 ml./min. through the gas ~ide of membrane de~ce 16 and ~ he~oglobin solutlon flow r~te o 500 ~ in~
through the othe~ ~ite o~ m~br~ne oxy~enator 16. The resultn of this procedure are prese~ted in Table II
below, showin~ th~t the deoxygenation i3 less offectlYe uader these conditions:

TABLE II

Deox~gen~tion 2 Oxy- 2 Deoxr ~ Met 2 Hb Ti~e (Min.)he~oglobln Hb Hb % S~uratlon : ~5 0 95.7 3.1 1.2 96~9 43~ 55.7 1.1 43.7 ~708 70.g 1.4 2a.2 ~0 21.~ 76.9 1.5 22~1 15~) 14~9 8 ~2 lo9 15~;~

. .
. ., , . . .

:, .

(C) The procedure of ExAmple 6 (A) w~s repeated, except that high purity argon g~s (Union Carbide Linde Division~ was used. The oxygen was purified to below 1 ppm with the use of oxygen trap 3G. All the system parametera were unchanged except that the ratio of argo~ gas to hemoglobin solution flow rate was 40 to 1, with the argon flow rate of the gas side of membr~ne device 16 being 4 liters per minute and the hemoglobin solution flow rate through the other side of membrane device 16 being 0.1 liter per minute. The results of this experiment are as shown in Table III below:

TABLE III

Deoxy~enation ~ Oxy- ~ Deoxy % Met % llb lS Time (Min.) hemoglobln Hb Hb % Saturation 0 96.1 2.7 1.1 97.2 30.5 69.5 ~ 30,5 18.2 82.1 0 18.1 - 70 l~ol 85~9 0 14~1 110 10~8 8~9~6 0 10~7 130 9~8 90~4 0 9 Aocordingly, when the above yrocessed olutions of deoxyhemoglobin are heated at Q ~emper&ture of essentiælly 45 ~o 85 degree~ ~. in accordance with this i~vention snd maintained thereat, one can in~ctiva~e sub~tantial smounts of ~irus present and precipitate ~ubstan ial amounts of nonhemoglobin proteins. The deoxyhemoglobin present exh~ bits : ,, ., .,:
.... .
`~ '~" , , , :; ::.

~8S52 improved heat stability, reducing los~es of hemoglobin during the process.

EXA~IPLE 7 Referring to Fig. 3, apparatus is provided for ~parging he~oglobin solutions with oxygen-free inert gas. Basic~lly, known designs of bubble oxygenators for blood may be used for this new purposep for example designs as disclosed in U.S.
Patent NO.~r 3,892,534 or 3,729,377.
Apparatu& 50 defines a gas exchange rolumn 52 which has a hemoglobin solution inlet line 54 which lead~ from solution reservoir 56~ Roller pump 58 or the like ls provided to circulate the hemoglobin solution from reservoir 56 to column 52 snd beyond.
Oxygen-$ree, inert gas is bubbled into the bottom of column 52 rom a ga~ source 60 ~hrough line 62 and porous sparger 64, to cause gas bubbles to rise through the column 52 while filled with hemoglobin solutlon. At the top of column 52, gas and solution pass throu~h horizontal column 66, containing conventional silicone-coated wire antifoam sponges 68, with the ga~ be~ng vented through vent 70, and flowing hemoglobin solution passing downwardly through the cur~ed debubblin~ channel 72. From there, the hemoglobin solution runs through outlet line 74 back 26 to reservoir 56.
By this process b the hemoglobin in solution can be deoxygena~ed, and thereafter heated as previously described to inactivate virus and to precipitate nonhemoglobin proteins.

',' '' ', .
. , .
. . :., ,:, . . . .

1~8SS2 In order to measur~ the efficacy of hemo~lobin 301ution heat treatment as a viral inactivation procedure, Sindbis, polio, and pseudorsbris viruses were sPeded into separate ~olutions of 1 g/dl hemoglobin containing 50 mM sodium dithionite. The solutions were sealed into vials, heated at 60 degrees C., and aliquot~ removed at various time intervals and stored for subsequent analysi~ of viral activity~ Viral acti~ities were determined by standard plaque a~says. The results of thi~ study are shown in Table III below:

TABLE III

Virus Titer (LoglO
15 Sample Plaque Forming Units/ml) Sindbis Polio P~eudorsbies VlruQ Stock Soln. 7.45 7050 6.0~

Unheated Hb Conerol Soln. 5.42 6.19 4.98 Hb Soln. he~ted 60 C, 0.00 0.00 0.00 30 minutes Hb Soln. he~ted 60 C, 0.00 0.00 0.00 60 min .::' .' -. .
,',':` ",' : ' ".:. i::
. .
. .
.: . ." -:.. ; . .
. . .

~X~ ;52 Hb Soln. heated 60 C, 0.00 0.00 0.00 90 min Hb Soln. heated 60 C, 0.00 0.00 0.00 120 min These results demonstrate that all three viruses were rapidly inactiva~ed in the hemoglobin solution under these conditions.

: In thi~ exsmple, troma-free oxyhemoglobin 10 solution i9 treated phy~ically as in Example 6, rather ehan chemically, to remove oxygen from the solution prior to heating.
To effect deoxygenation, two liters of a 0.54 g/dl hemoglobin solution was circulated in a ~S closed system throu~h a SciMed Life Sy~tem 0.8 square meter membrane oxygenator which was concurrently flushed with nitrogen. After 70 minutes of circulation, the hemoglobin was 96% deoxy~enated as s~se#~ed spectrophotometrically. The solution wa~
then heated at 60 degrees C, for 5 hours~ with a 93%
reco~ery of total hemoglobin content. These results demonstra e that hemoglobin solutions may be succes~fully he~t treated after deoxygenstion by passage through a membrane de~ice.

EXAMPLE lO

Approximately 45 ml of solution containing l g/dl hemoglobin w~ sparged as in Example 7 with oxygen-free argon and then heated at 60 degrees C.

.
" , :' ' ,;.
.:............................................... .
. :.
. :.

~8552 -3~-Absorption spectra were recorded from thi~ solution using a flow through cell before, during and after heating and the relative concentration of oxy, deoxy, and methemoglobin calculated. In this experiment the degree of deoxy~enation after sparging was approximately 95%, with further deoxygenation occurring during heating, as shown in Table IV.

TABLE IV

Sample Temp- % oxy % deoxy % met er~ture Hb Hb Hb Hb Soln~-Air Equili- 25 C 90 9 bra~ed Hb Soln.-Argon Sparged 25 C 6 95 0 Hb Soln~-Heated for 60 C 4 96 0 80 minutes Hb Soln.-Heated for 60 C 2 98 0 140 minutes The hemoglobin formation during heating was minim~l and a precipitste of nonhemoglobin proteins observcd. The total hemoglobin concentrations were not messurably diminished during ~he heating period.
These data demonQtrate that hemoglobin heat treatment ~ay be performed after solution deoxygenation by 24 sp~rging with inert gas.

' ' , .: .
.

~2~8.~52 Stroma frse hemo810bin was cross lqnked by reacting with bis(3,5 dibromosalicyl) fumarate (DBBF), and the resulting product purified by column chromatography. The cross-linked hemoglobin was diafiltered against isotonic sodium phosphate buffer solution, pH 7.4, the concentration adjusted to one g/dL, and aliquots sealed into glass vials. A portion of this solution was admixed with sufficient sodium dithionite to 13 give a final concentration of 50 mM, the pH of the solution adjusted to 7.5 with sodium hydroxide, and aliquots of this solution sealed into glass vials.
Aliquots of the hemoglobin mixed with dithionite were heated at 60 degrees C. for 10 hours and the hemoglobin compared to unheated samples after the removal of dithionite by dialysis. The absorption spectra of both heated and unheated samples were virtually identical, as were the oxygen binding characteris$ics as determined by means of an Aminco Hem-0-Scan~ Analyzer. These data demonstrate that crosslinked hemoglobin may be heated at 60 degrees C. for 10 hours to inactivate viruses and precipitate contaminating proteins without a significant loss of hemoglobin function.

Claims (50)

1. The method of inactivating virus present in substantially cell-free hemoglobin solution comprising heating said hemoglobin at a temperature of essentially 45 degrees to 85 degrees C., while maintaining said hemoglobin in substantially its deoxyhemoglobin form, whereby substantial inactivation of said hemoglobin is avoided.

2. The method of Claim 1 in which said hemoglobin is maintained in substantially deoxyhemoglobin form by the presence of a chemical reducing agent.

3. The method of Claim 1 in which said temperature is 60 degrees to 75 degrees C.

4. The method of Claim 1 in which heat precipitable nonhemoglobin proteins present are precipitated by said heating.

5. The method of Claim 1 in which said hemoglobin is maintained in substantially deoxyhemoglobin form by exposing said hemoglobin to inert, essentially oxygen-free gas or vacuum to cause removal of oxygen from the hemoglobin and conversion of other forms of hemoglobin to deoxyhemoglobin, and by maintaining said deoxyhemoglobin in an oxygen-free environment during said heating.

6. The method of Claim 5 in which said hemoglobin is exposed to nitrogen.

7. The method of Claim 5 in which said hemoglobin is exposed to argon.

8. The method of Claim 5 in which said hemoglobin is exposed to said gas or vacuum through an oxygen-permeable, hemoglobin-retaining membrane.

9. The method of Claim 5 in which said hemoglobin is exposed to inert, oxygen-free gas by bubbling said gas through a solution of said hemoglobin.

10. The method of passing a solution of substantially cell-free hemoglobin through diffusion cell means, said diffusion cell means having membrane wall means along which said hemoglobin solution flows, said membrane wall means being capable of passing oxygen but not hemoglobin through said membrane wall means while circulating inert gas along the side of said membrane wall means opposed to said hemoglobin solution, to cause removal of oxygen from the hemoglobin solution and conversion of other forms of hemoglobin to deoxyhemoglobin; and thereafter heating the resulting deoxyhemoglobin solution at essentially 45 degrees to 85 degrees C. in an oxygen-free environment to inactivate virus present and to precipitate heat-precipitatable nonhemoglobin proteins without substantially inactivating said hemoglobin.

11. The method of Claim 10 in which the flow volume of circulating, inert gas is at least 5 times the flow volume of said solution passing through the diffusion cell means.

12. The method of Claim 11 in which said inert gas is mitrogen.

13. The method of Claim 11 in which said inert gas is argon.

14. The method of Claim 11 in which said temperature is essentially 60 degrees to 75 degrees C.

15. The method of Claim 11 in which aid flow volume of said circulating inert gas is from 10 to 50 times the flow volume of said solution passing through the diffusion cell means.

16. The method of precipitating nonhemoglobin proteins from a hemoglobin solution without substantially precipitating or deactivating said hemoglobin which comprises heating substantially cell-free hemoglobin solution at a temperature of 45 degrees to 85 degrees C., while maintaining said hemoglobin in substantially its deoxyhemoglobin form, whereby heat precipitable nonhemoglobin proteins present are precipitated.

17. The method of Claim 16 in which said temperature is 60 degrees to 75 degrees C.

18. The method of Claim 17 in which said hemoglobin is maintained in substantially its deoxyhemoglobin form by the presence of a chemical reducing agent under conditions such as the nonhemoglobin protein selectively precipitates but the hemoglobin is not substantially biologically inactivated.

19. The method of claim 18 in which said reducing agent has a reducing potential which is greater than ascorbate.

20. The Method of claim 16 in which said reducing agent is a reduced dye, sulfoxy, or sulfhydryl compound.

21. The method of claim 20 wherein the agent is dithionite, bisulphite, metabisulphite, sulphite, reduced glutathione or dithiothreitol.

22. The method of claim 16 in which said time of heating is about 1-15 hours.

23. A method for reducing the viral infectivity of the hemoglobin composition suspected to contain such infectivity, comprising heating the hemoglobin composition and a reducing agent together under a temperature and time such that the virus is rendered inactive but the hemoglobin is not substantially inactivated, the amount of reducing agent present being effective to maintain hemoglobin in the deoxy form during heating.

24. The method of claim 23 wherein the time of heating is from about 1 15 hours for aqueous solutions of the composition and from 20-96 hours for dry compositions.

25. The method of claim 23 in which the heating temperature is from about 50 degrees C. to 80 degrees C.

26. The method of claim 23 in which said hemoglobin composition is crosslinked.

27. A hemoglobin composition carrying inactivated virus, made in accordance with the method of claim 1.

28. A method for reducing the viral infectivity of a hemoglobin composition suspected to contain such infectivity comprising heating the hemoglobin composition and a reducing agent together under a temperature and time such that the virus is rendered substantially inactive, but the hemoglobin is not substantially inactivated, said reducing agent having a reducing potential which is greater than ascorbate and being present in an amount effective to maintain hemoglobin in the deoxy form during heating.

29. The method of claim 28 in which the heating temperature is from about 50 degrees C. to 80 degrees C.

30. The method of claim 29 in which the time of heating is about 1-15 hours for aqueous solution of the hemoglobin composition and from about 20-96 hours for dry hemoglobin composition.

31. The method of claim 29 in which the reducing agent is a reduced redox dye, sulfoxy, or sulhydryl compound.

32. The method of claim 31 in which the reducing agent is dithionite, bisulphite, metabisulphite, sulphite, reduced glutathione, or dithiothreitol.

33. The method of claim 29 in which the composition is crosslinked.

34. The method of claim 28 in which said reducing agent is an alkali metal dithionite.

35, The hemoglobin composition containing inactivated virus treated in accordance with the method of claim 29.

36. A method for reducing the viral infectivity of a hemoglobin composition suspected to contain such infectivity, comprising the heating of the hemoglobin composition and a reducing agent together under a temperature and time such that the virus is rendered substantially inactive but the hemoglobin is not substantially inactivated, said reducing agent being selected from the group consisting of redox dye, disulfoxy compounds and sulfhydryl compounds, said reducing agent being present in an amount effective to maintain hemoglobin in the deoxy form during heating.

37. The method of claim 36 in which the heating temperature is from about 50 degrees C. to 80 degrees C.

38. The method of claim 37 in which the time of heating is from about 1-15 hours for aqueous solutions of hemoglobin solution.

39. The method of claim 37 in which said reducing agent is dithionite, bisulphite, metabisulphite, sulphite, reduced glutathione, or dithiothreitol.

40. The method of claim 37 in which said hemoglobin composition is crosslinked.

41. The method of claim 37 in which said reducing agent is an alkali metal dithionite.

42. The method of claim 37 in which the reducing agent has a reducing strength which is greater than ascorbate.

43, The method of claim 37 in which said hemoglobin composition is pyridoxalated.

44. The method of claim 37 in which non-hemoglobin protein materials are precipitated during said heating step.

45. A crosslinked hemoglobin composition which is free of active virus and which has been treated in accordance with the method of claim 37.

46, A method for reducing the viral infectivity of a hemoglobin composition suspected to contain such infectivity, comprising heating the hemoglobin composition under conditions which cause hemoglobin to assume and maintain its natural deoxygenated form, at an elevated temperature and time: such that harmful virus present is rendered inactive but the hemoglobin is not substantially inactivated.

47, The method of claim 46 where the elevated temperature is from about 50 degrees C. to 80 degrees C.

48, The method of claim 47 in which the time of said heating is about 1-15 hours for aqueous hemoglobin solutions and about 20-96 hours for dry compositions.

49. An aqueous solution of cross-linked hemoglobin, free of active virus, and having a P50 under physiologic conditions of at least 26 mm. Hg.

50. A hemoglobin composition carrying inactivated virus, made in accordance with the method of claim 10.