CA2074852A1 - Organ perfusion hemoglobins - Google Patents

Abstract

ABSTRACT

Novel modified hemoglobin products exhibiting properties rendering them particularly useful for organ perfusion are prepared by reacting hemoglobin with divinyl-sulfone under controlled conditions. The products are substantially free from intramolecular crosslinking and have low oxygen affinities and low viscosities, at the hypothermic temperatures normally used for organ perfusion, e.g. about 15°C.

Description

This invention relates to modified hemoglobins and uses thereof. More specifically, it relates to covalently modified and polymerized hemoglobins, processes for their preparation, and processes of preserving biologi-cal organs therewith.

Human inkernal organs for transplantation are normally preserved in cold saline or buffer solutions. ~he normal viability periods of these organs are from a few hours to vne day. This viability period can be critical, especially in the case of liver transplantation, since a removed liver can keep its viability for only a very short period of time (4-6 hours)~

One suggested method for increasing the preserva-tion time of organs is a perfusion of the organ with hemoglobin (Hb) - based oxygen carriers. These materials can also be used for effective cardioplegia in open-heart surgery.
However, whilst a wide variety of Hb-based oxygen carriers are known, they all have excessive oxygen affin-ities and inadequate oxygen delivering capacities at the hypothermic conditions needed for organ perfusion (10-15 C). Workers in the field have tended to concentratetheir efforts ~n developing Hb-~ased oxygen carriers for tran~fusion purposes, which have oxygen a~finity resembling that of whole blood at 37 C, in ~heir search for an adequate Hb-based blood substitute.
The oxygen affinity of Hb-based oxygen carriers is commonly and conveniently expressed as its P50 ~ the partial pressure of oxygen at which 50% of the hemoglobin siteæ are saturated with oxygen. The higher the P50 value, the lower the oxygen affinity. The P50 value is highly dependent on temperature, the oxygen affinity increasing as the temperature decreases. For example, red blood cells at 15 C have a P50 value of 2 torr, compared with 26.5 torr at 37C. Accordingly, for successful use in hypothermic organ . .

, ' '' :',"

perfusion, acellular oxygen~delivering red cell substitutes should have a P~O value much higher than 27 torr at 37 C, and ideally between 55 and 100 torr at 37~C~

~ a~ NoO 4 061r~36 Morris et al., dis-closes a pharmaceutically acceptable, intramolecularly crosslinked, stromal free hemoglobin or use as a blood substitute and blood plasma expander. A wide variety of difunctional crosslinking reagents are suggested in Morris lQ et al., including divinylsulfon~. In examples VI-IX, Morris st al. describe reacting hemoglobin with divinylsul-fone. In each case, an intramolecularly crosslinked hemoglobin is obtained, with P50 values at 37 C of from 7-55 torr. Morris et al. also suggest the use of their cross-linked hemo~lobin or the storage and preservation of viable isolatad perfused mammalian organs for their event-ual transplant into ~ recipient.

The present invention provides novel divinylsul-~one-treated hemoglobins which possess properties desired in oxygen carriers. One of these products is a non-cross-linked, intr~molecularly modified hemoglobin, which has a near normal oxygen-delivery capacity at 15C, a relatively low oxygen af~inity, a normal physioloyical half~lifet a low methemoglobin content, and relatively high Hill coefPi-cient. This combination of properties renders it suitable not only as a component of a potential oxygen carrier ~or trans~usional purposes, but also as an oxygen carrier in acellular perfusional fluids. Another of the products is divinylsulfone-modified polymerized hemoglobin. This has a long retention time, low oxygen affinity and lower methemoglob:Ln content. Itq combination of properties renders it primarily suitable as an oxygen carrier for transfusional purposes. The invention also provides mixture~ of these two materials in controlled relative proportions, in which the combination of their respective properties can be optimized. It also provides a process "` 207~2 whereby the individual products, or mixtures thereof in controlled, predetermined relative proportions can be obtained, in a single synthetic step using a single reagent, divinylsulfone (DVS).

Divinyl.sulfone reacts w.ikh primary groups on the peptide chains of hemoglobin, to form secondary amine linkages. In a first reaction step, a modified hemoglobin is formed:
O O

CH2ac~--s--CH=CH2 + H2N--Hb ~ CH2=CH--7--CH--CH--NH~Hb ~5 ~ 0 DVS Hb modified Hb ~Hh-~VS) In a second reaction step/ a crosslinked hemoglo-bin is formed:

O O
ll ll Hb-NH2 + CH2 = CH-S-CH.CH.NH.Hb ~ Hb.NH.CH.CH~S.CH.CH.NH.Hb Il 1~
C O
(Hb-DVS) poly (Hb-DVS) The crosslinking can theoretically take place between two globin chaine of the same hemoglobin tetrameric units, in which case the product is referr0d to as intramo-lecularly crosslinked, or between two globin chains of different hemoglobin tetrameric units, in which case the product is intermolecularly crosslinked, effectively polymeriz~d. However, in contradiction to the predicted theory, and contrary to the teachings of the aforementioned Mor~is et al. patent, it is found in accordance with the present invention that, under selected conditions, divinyl-sulfone reacts with hemoglobin to produce a product which . ...
,:.,., , , .:: ,, . , : : ~
'.~' '.

~07~8~2 is substantially free from intramolecular crosslinks.

The products according to the invention are thus covalently modified, non-crosslinked nemoglobin and/or inter~olecularly crosslinked hemoylobin, and/or mixtures o~
the two modified hemoglobins in controlled relative propor-tions. The products are mod.i~ied and/or crosslinked with sulfone-secondary amino covalent bonds, and the products are substantially free from intramolecular crosslinks.
They show oxygen affinities at 12-15 C which render them suitable for use in organ perfusion.

Figure 1 is a graph of the ion exchange chroma-tography puri~`ication of the produ~t prepared according to Example l below, Figure 2 is the results of electrophoresis according to Example 3 below;

Figure 3 is a Hb-GPC analysis curved derived according to Example 4 below;

Figure 4 is a Hb-GPC of typical poly-hemoglobin derivatives ~rom the examples below;
Figure 5 is the SDS-PAGE of HbBv and derivatives from Example 5 below;

Figure 6 is a graphical pres~ntation of the dependence o~ oxygen affinity on pH o~ the products of the invention, according to ExamplP 6 below;

Figure 7 is a graphical presentation of the dependence of oxygen af~inity on temperature o~ the prod-ucts of the invention, according to Example 7 below;

~7~2 Figure 8 is a graphical presentation of the clearance of intravenously injected hemoglobin products, according to Example 8 below: and Figure g is a plot of oncotic pressure against hemoglobin concentration, derived according to Example 9 below.

The products of the present invention are pre-pared by a process o~ reacting bovine or human deoxygenated stromal free h~moglobin under anaerobic conditions with divinylsul~one. When relatively low concentrations of hemoglobin and DVS are used in anaerobic one step reaction, one can produce a modified hemoglobin having exactly the same elution time on gel-permeation HPLC as that of the native, unmodified Hb, indicating that initially only a modi~ied, non-crosslinked product is formed. In a two step reac~ion where low concentrations of DVS are used in the f irst step and higher concentrations in the second step, a product comprising 50% modified Hb and 50% modified polymeriæed Hb tpoly-Eb-DVS) can bQ obtained. Raising the DVS concentration still higher in the second step yields a product comprising 15% modified Hb and 85% poly-Hb-DVS.
Thus by varying the concentration o~ DVS in the reaction mixture, one can vary the relative proportions o~ modified Hb and poly-Hb produced, in a aontrolled manner to obtain a product of pre-selected composition. In all cases, the product produced is substantially free from intramolecular crosslinking, and has the desirable oxygen carrying prop-erties described above.

The modification reaction to produce Hb-DVS is carried out using deoxygenated Hb, under anaerobic condi-tions. Relatively low concentrations of deoxygenated Hb and DVS in solution are preferably used, suitably ~rom 5-10~ of deoxygenated Hb in the reac~ion solution, and 0.5-1.5 microlitres DVS per ml of solu~ion. The reaction is ~, :: , ... .
, . :. .
,, :: :
, ' ' ' , " ' ," ~, ~

~07~8~2 carried out under slightly basic conditions (pH 7-7.8), suitably buffered to a pH value in that range with sodium phosphate buffer. The temperature should be kept within the range 1-5 C, and t~e reaction takes place for a period of 18~30 hours.

The second stage reaction, to convert the Hb-DVS
to poly-Hb-DVS is conducted under essentially the same conditions, except that a higher concentration of Hb-DVS in solution is used and a higher concentration of DVS is present. These higher concentrations are, suitably, from 12-20% Hb-DVS in the reaction solution, and from 2-5 microlitres DVS per ml of reaction solution. Other reac-tion conditions, including pH, temperature, anaerobic reaction, and time of reaction can be the same in the second ~ep as in the first step. The reactions are appropriately quenched after the desired reaction time by addition of bu~fered lysine-HCL. The unreacted vinyl groups of DVS are deactivated by lysine-HCL.
Thus according to the present invention, it is found that under certain conditions the anaerobic reaction of Hb with the bifunctional reagent DVS can produce an intramolecularly-modi~ied derivative, Hb-DVS, without the introduction of an undesired non-speci~ic intramolecular crosslinkage. In order to direct the reaction to this path, one uses relatively low concentrations of Hb and low molar ratios of DVS:Hb-tetramsr. The Hb-DVS so formed is remarkably homogeneous with respect to its molecular mass and electrophoretic properties, as demonstrated in the specific examples below.

At the increased concentrations of Hb-DVS and relatively high molar ratio of DVS:Hb~tetramer, the mod-i~ied hemoglobin so formed can be anaerobicallypolymerized. From 50-85% of the starting material can be thus polymerized, to produce a poly Hb-DVS which is not ~7~2 homogeneous with respect to molecular weight, but which, on analysis, shows the presence of molecular masses of 16 kda and multiples thereof. No product corresponding to a 32 kda molecular weight is obæerved, demonstrating the lack of intramolecular crosslinkages, even at the high DVS concen-tration needed for the polymerization, thereby indicating that poly-Hb-DVS does not contain intramolecular crosslink-ages as part of its structure. This conclusion is consist-ent with the very low oxygen affinity and methemoglobin content of the material.

The conditions of reaction outlined above are important for obtaining the products of the present inven~
tion. When too high a temperature is used, very high quantities of methemoglobin are produced, which is undesir-able. Similarly, when too high a molar ratio of DVS:Hb-tetramer i~ used, even at low tempera~ures, the methemoglo-bin level is too high, implying the possibility of non-specific intramolecular crosslinkage. The intramolecular modification obtained with the relatively low molar ratio of DVS:Hb-tetramer described above appears to change the molecular conformation in such a way that subsequent non-speci~ic intramolecular crosslinking can be virtually avoided.
Both Hb-DVS and poly-Hb-DV~ possess properties desired in oxygen carriers. The long retention time of poly-Hb-DVS in the mammalian circulatory system, apparent from the specific examples below, its low oxygen a~finity and low methemoglobin content make this protein a potential oxygen carrier for transfusional purposes. The modified derivative, Hb-DVS with it6 normal physiological half life, low oxygen affinity, low methemoglobin content and rela-tively high Hill coe~ficienk is suitable for use as a potential oxygen carrier in acellular per~usional fluids.
The findings herein are based upon experimental work conducted on animal hemoglobin, namely bovine hemoglobin ::

2~7~

(BvHb), but are helieved equally applicable to other hemoglobins, including human hemoglobin.

T~e invention is further described and illus-trated with reference to the following ~pecific examples:

BOVINE NEMOGLOBIN

Stroma~ree bovine hemoglobin (HbBv) was prepared by t~luene extraction and crystallization~ essentially accordinq to a procedure previously described for human hemoglobin b~ De Venuto et al. [1]. The washed crystals were dissolved in distilled water and dialy~ed twice at 2 C
(1:25, v/v) against water and then three more times (1:15, v/v) again~t cold 0.0S M sodium phosphate buffer (pH 7.4~.
Sterilization was achieved by passing the final solution through a sterile 0.22 ~m Millipore Sterivex-GS filter.
Total hemoglobin and methemoglobin concentrations were ~easured on an IL 482 CO-Oxymeter (Instrumentation Labora-tory) calibrated for Hb v extinction coefficients. Differ-ent batches of the final solution contained 10-15% (w/v) hemoglobin and 1.5-3% methemoglobin (~ of total hemoglo-bin). These were kept at 2-4 C and used within 48 hours ~rom the time they were prepared.

EXAMPLE 1 - P:REPARATION OF HbBv--DVS

A 25 ml aliquot of 8% (w/v~ solution of HbBv in 0.05 ~ sodium phosphate buffer (pH 7.4) (reaction buffer), containing 0.25 ml of caprylic alcohol to prevent foaming, was deoxygenated by flushing with nitrogen in a 100 ml Pyrex reactor for 3 hours. Next, a 25 ~l aliquot of DVS
was added to the reaction vessel, and the reagents were stirred gently under a continuous flow of nitrogen for 24 hours. ~he unreacted vinyl groups were deactivated by the addition of 2.0 ml of deoxygenated 2.0 M lysine-HC1 in 0.05 M sodium phosphate buffer (pH 8.0). The quenching reaction ,. . . , : : :
. . .
' , ~: :, , .

207~8~

was carried out under a nitrogen atmosphere for an addi-tional 18 hours. After quenching, the reaction solution was clarified by centrifugation at 15000 xg for 30 minutes.
The solution was dialyzed thrice (1:40l v/v) against 0.02 M Tris-HC1 buffer (pH 7.4) in order to remove unbound DVS
and excess lysine. The protein was then converted to the carbonyl form, and the HbBv-DVS was separated from the unmodified hemoglobin on a DEAE-Sephadex column (~-50, 2.5 x 30 ~m) which was equilibrated with the same CO saturated buffer. The column was loaded with 5 ml batches and eluted with a linear gradient of pH and ionic strength using 0.05 M Tris malaate (pH 6.5) as the second buffer. The second ~ands from all batches were pooled together and concen-trated to ~15% (w/v) by ultrafiltration (YMT-30 membrane, Amicon). All work was done at 2-4 C. The final HbBv-DVS
solution was sterilized as before and storecl at the same temperature. ~hese solutions were used not later than 2 weeks after preparation. Just before use, the carbonyl derivative of HbBv-DV5 was converted to the oxy form as previously described by Shih et al. The amount of methemo-globin in the final oxy derivative was 3-4% of the total hemoglobin.

Fig. 1 o~ the accompanying drawings presents the ion-exchange chromatography purification of the product, HbBv-DVS, on a DEAE-Sephadex A-50 column. The figure is discussed ln the "Results" below.

~A~IE-2 - PREPARATION OF POLY HbBv-DVS
To 25 ml of 15% (w/v) solution of HbBv DVS in reaction bu~fer, 0.25 ml of caprylic alcohol was added.
The HbBv-DVS solution was then deoxygenated by flushing with nitrogen for 3 hours. Next, 75 ~l of DVS was added and the reaction was allowed to proceed, with gentle stirring, for 24 to 72 hours. Every 24 hours, the pH of a small aliquot of the reaction solution was detarmined and ;.
,,,~ , , ' :: ", . .

21~ 8~
... ,i the progress of the reaction was monitored by HP-GPC. If necessary, the pH was adjustad to 7~4 with 0.5 M NaOH.
Once polymerization had produced the desired molecular weight distribution for Poly Hb~v DVS, the reaction was quenched by adding 2.0 ml of deoxygenated 2.0 M lysine-HCl in reaction bu~fer. The reaction mixture was kept anaerobîc and stirred for an additional 18 hours, and was then clarified by centrifugation at 15000 xg for 30 min-utes. The solution was dialyzed thrice (~:40, v/v) against a reaction bufPer containing 0.1 M NaCl, an~ the proteln wa~ converted to the carbonyl form. The reaction solution was then applied, 5 ml batch at a time, to a 4x50 cm Biogel P-100 column (Bio-Rad), pre-equilibrated with the same CO
saturated bu~fer. In this gel-filtratlon medium, all materials with molecular weights higher than ~100,000 were immediately excluded from the columnl enabling a good separation of the polymerized material (M~ 2 130,000~ from the tetrameric HbBv-DVS fraction,. The pooled polymerized fractions were then concentrated by ultrafiltration to ~10%
(w/v). All work was done at 2-4 C~ The final Poly-HbBv-DVS s~lution was sterilized as before, and stored at the same temperature. These solutions were used within 2 weeks ~rom the time o~ preparation. Just before use, the carbonyl derivative of Poly-HbBv DVS was converted to the ~5 oxy form. The amount of met~emoglobin in the later sol-ution was less than 6~.

~A~oe~ ELLULOSE ACETATE ELECTROPHORESIS

Electrophoretic follow up of the reaction of the synthesis of HbBv-DVS was carried out using cellulose acetate electrophoresis. Super Sepraphore (Gelman) strips were employed in a Sepratek ~Gelman) cham~er according to the procedure of Gebott and Peck [3]. The strips were stained with Ponceau S svlution and destained in 5% acetic acid. Figs. Z (a), (b) and (c) show the results from, respectively, HbBv, HbBv-DVS from a 24 hour reaction :, ' . :, ' ' . ; ' ' : ' 2~7~

mixture (Example 1), and HbBv-DVS Prom a 48 hour reaction mixture, and are discussed in the "Results" section below.

~ P-GPC analyses were per~ormed on a TSK-G 3000 SW
column (7.5 x 600 mm, LKB) and TSK~GWSP pre-column (7.5 x 75 mm, LKB) using a Waters HPLC apparatus equipped with a model 720 proqrammable system controller, data module, two pumps and an automatic sample (WISP model 710B). Sample solutions (100 ~1), wikh a total protein content of 0.5-5 mg, were eluted at 0.8 ml/min with 0.1 M sodium phosphate buffer (pH 6.8) containing 0.1 M KC1. The column outflow was monitored at 576 nm, using a variable wavelength detector (model 450, Waters).

The column was calibrated with highly purified~
well~ch~racterized, non-heme globular protein standards.
The void volume (vO = 11.6 ml, 14.5 min) was d~termined from the elution of Blue Dextran 20~0. There was a linear relatiohship between log k~ and the elution time. However, the apparent molecular weights of the monomeric (M~ =
64,000) and the polymerized h~moglobin components were shifted with respect to the non-heme protein standards.
These hemoglobin derivatives showed lower apparent molecu-lar weights than expected. To correct this problem the las'c three peaks of these components (M~ = 192,000; 128,000 and 64,000) were employed as standards ~or the molecular weight estimation of the intermolecularly-crosslinked hemoglobins.

Figs. 3 (a) and (b) are the analysis curves from/
respectively, HbBv, HbBv-DVS isolatPd ~rom a 20 hour reaction mixture by ionic exchange chromatography (solid line) and HbBv-DVS 48 hour reaction mixture (dotted line).
Figure 4 show~ the HP-GPC of typical poly HbBv-DVS prepara-tions (Example 2) namely: (a) poly-HbBv-DVS, 24 hour : .. :: ,..
.,: .

;, ~7~

reaction mixture (dotted line) and isolated poly-HbBv-DVS
obtained from the same reaction mixture (æolid line); (b) poly HbBv-DVS 72 hour reaction mixture (dotted line) and isolated poly HbBv~DVS obtained from the same reaction mixture (solid line). They are further discussed in the "Result~ section below.

SDS-PAGE was carried out in slabs (1.5 mm thick), according to the method of Weber and co-workers ~4], using polyacrylamide concentration of 12% (w/v). Protein samples originating from the HP-GPC procedure were concentrated and desalted (Centriprep-10, Amicon) with 0.01 M sodium phos-phate buffer (pH 7.0), as part of their preparation for the SDS-PAGE. This step was necessary in order to obtain the small sample volumes needed for the continuous buffer system utilized here and to eliminate SDS precipitation by the potassium ions present in the HP-GPC elution buffer.
Gels were stained with Coomassie blue R-250 and destained by di~usion in 12.5% isopropanol and 10% acetic acid.
Relative mobilities, calculated as ratios of distances of migration of protein bands and dye, were used in conjunc-tion with a calibration curve obtained from five protein markers in order to estimate molecular masses. the markers were bovine serum albumin (66 kDa), bovine erythrocyte carbonic anhydrase ~29 kDa), bovine pancrea~ trypsinogen (24 kDa), soybean trypsin inhibitor (20.1 kDa) and bovine hemoglobin ~16 kDa).
The results are presented in Fig. 5 which shows SDS PAGE of HbBv and derivatives obtained from its anaerobic reaction with DVS: (a) mixture of protein marker~; (b) HbBv-DVS 48 hour reaction mixture7 (c) pure HbBv-DVS; (d) Poly-HbBv-DVS isolated from a 24 hour reac-tion mixture; (e) the slowest Hp-GPC peak isolated from poly-HbBv-DVS 24 hour reaction mixture (dotted peak of E'ig.

2~4~2 4a); (f) 5 ~g of pure HbBv: (g) 5 ~g of pure HbBv-DVS. The figures are further discussed in the "Results" section below.

EX~MPLE 6 - oxyGEN EOUILIBRIUM DETE~MINATIoNs Oxygen equilibrium curves were determined using a TCS modql B Hemox~analyzer at 37 C. In these experi-ments, HbBv, HbBv-DVS and Poly-HbBv~DVS (isolated from a 24 hour reaction mixture) were employed at a protein concen tration of 2 mg/ml. The pH of the samples was measured at 37 C with a Fisher Accumet 750 instrument. If not other-wise statedl buffers used were 0.15 M Tris-HC1 or Bistris-HC1 containing 0.15 M C1- ions. ~hese buffer~ were prepared by titrating the reagents with HC1 to pH 5.5 b~fore read-justing with NaOH to the desired pH. This was done so that the molarity of the buffer would also represent the concen-tration of C1- in solution. The sample6 were equilibrated with pure nitrogen (les~ than 5 ppm oxygen) and reoxygen-a~ed wikh 35.2% oxygen in nitrogen. In some cases, whenthe effect of C02 on the oxygenation properties had to be evaluated, gas mixtures containing 5.0% C0z were employed.
Methemoglobin content after oxygen equilibrium measurements was 3-8%. oxygenation data were ~ormulated in terms of the Hill plot (log Y/(1-Y) versus log P) where Y is fractional saturation of t~e hemoglobin with oxygen and P is the oxygen pressure in mm Hg. Hill plots were derived using digitized data points from the oxygen equilibrium curves.
Overall oxygen af~inity and cooperativity were character-ized in terms of oxygen pressure at half-saturation (P50) and the Hill coefficient (n50) given by the Hill plot at half-eaturation.

Data in the present specification were based on at lea~t two independent determination~ for each point.
Values for individual PgO determinations were within + 1.0 mm Hg of the mean value.

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

~7~8~2 - ~4 ~-Fig. 6 illustrates the dependence of the oxyyen affinity on pH for HbBv ~o), HbBv-DVS (-), and poly HbBv-DVS (~). This figure is further discussed in the "Results"
seçtion below.

~X~MPLE 7 ~ EFFECT ON TEMPERATURE ON THE OXYGEN-~1 DING P~OPERTIES

The effect of temperature on the oxygen affinity of HbBv, HbBv-DVS and Poly HbBv-DVS is shown in Fig. 7. A
range of temperatures extending from 15 to 37~C was e~plored. In this range, at pH 7.4 and in the presence of 0.15 M Cl- ions linear plots of Log P50 VS. 1/T were obtained, indicating constant heats of oxygenation. one can see that ~or all three oxygen carriers the P50 values are decreasing at lowsr temperatures. At any one tempera-; ture, however, the PgO values of HbBv-DVS and Poly HhBv-DVS
are much high~r than that of HbBv.

The amounts of oxygen that would be unloaded at 37C and 15C by these five human and ~ovine hemoglobin-based oxygen carriers upon decreasing the PO2 from an arterial value o~ 100 to a venous level of 40 mm Hg were calculated from the origina] oxygenation curves used for the generation of Fig. 7. ~hese values are compared in Table I, îmmediately below. The Hill coefficient values for H~Bv-~VS and Poly HbBv-DVS, at all temperatures, were n50 ~2.0 and n50 - 1.6~ respectively.

2~7~5~
~ 15 -TABLE I

OXYGEN UNLOADING BY SOM~ HUMAN AND BOVINE
HEMOGLOBIN-BASED OXYGEN CARR~ERS AT 15 AND 37 C

oxygenTemp. P50 ~ % oxyHb Carrier t ~)(mm Hg)(P02100-49 mm Hg) Human blood 37 26.5 25 -- __ Human Hb 37 14 6 HbBv 37 27 15 HbBv-DVS 37 52 37 Poly HbBv-DVS 37 61 32 As in the oxygen equilibrium curves of Fig. 6, the data points in Fig. 7 are for HbBv (o), HbBv~DVS ~-) and Poly HbBv-DVS (~). This figure and Table I are ~urther discuæ~ed in the "Results" section bælow.

ÆXA~PLE 8 - INTRA~A$CULAR R~T~ ION ~IME IN ~

Plasma retention pro~iles of HbB~, HbBv-DVS and Poly HbBv-DVS (isola~ed from a 24 hour reaction mixture) were obtained from experiments performed on Sprague-Dawley male rats (300-350 g). Hemoglobin derivatives were exchanged by difiltration into phosphate-buffered saline (pH 7.4) (0.01 M sodium phosphate in 0.15 M NaCl) and their concentrations regulated to 6-8~. These solutions were ~iltered through sterile 0.22 ~m filters just before injection. ~he rats were anaesthetized with Somnotol. The right ~emor~l vein was cannulated and a sample of blo~ was ' : ,; ,, :
:: . ,, : , ~.,:, ~ .., .. . ... .

20748~2 withdrawn, ag a control, prior to the bolus injection of 2.0 ml of the tested material. Blood samples (250 ~.l) were drawn via the vein at 2, 15 and 30 min after the injection~
Subsequently, the surgery site was closed and samples were continued to be taken at 1, 2, 3, 4 and 6 hours following the injection time by cutting a minor slice at the end of khe tail. The samples were centrifllged and analy~ed for the hemoglobin content o~ their plasma.

Figure 8 graphically presents the clearance of intravenously injected HbBv (O), HbBv-DVS (-~ (upper panel) and two different initial plasma levels of poly HbBv-DV5 (lower panel). The dashed lines indicate values obtained for pla~ma half life. The ~igure is further discussed in the l'Resultsll section below.

EXAMP~E 9 - COLLOID OSMOTIC PRESSURE

The colloid osmotic (oncotic) pressure for HbBv, HbBv-DVS and Poly HbBv-DVS was determined on solutions of the products at various concentrations, at 20 C, with a W~SCOR 4400 oncometer calibrated with WESCO~ AC-010 Waker Manometer and 5% (w/v) albumin solution. The ef~ect of the hemoglobin concentration on the oncotic pressure is pres-ented in Fig. 9. Virtually the sam2 behaviour is o~tainedfor HbBv and HbBv-DVS. To attain physiological oncotic pressures, (COP approximately 25 mm Hg3 both the~e deriva-tives had to be diluted to 7 g/dl, compared to the inter-molecularly-crosslinked Poly HbBv-DVS which was iso-oncotic with plasma at 14 g/dl. The relationship between oncotic pressure and hemoglobin concentration was non linear and COP values for HbBv and HbBv-DVS were increased approxi-mately three-fold with respeck to those of Poly HbBv-~VS at similar concentration~.

:' ' ' ' .

, ~7~2 ~ESUI.~S

PURI~ICATION AND CHEMICAL CHh~ACTERI~TION OF HbBv-DVS

Fig. 1 presents the ion exchange chromatography on a DEAE-Sephadex column of HbBv reacted anaerobically with ~VS. The two ~ractions are nearly equal, each repre-senting ~50% of the total protein. The fast peak, with identical retention time to that of pure HbBv, represents the unmodified hemoglobin. The slower peak is pure HbBv-DVS. These results are consistent with an electrophoretic analysis of the reaction mixture. The analysis on cellu-lose acetate strips indicates that one of the fractions gave a sharp band with a mobility identical to that of untreated HbBv, while the protein eluting in the other peak, HbBv-DVS, gave a sharp band with a higher mobility (Fig. 2ajib~. HP-GPC of the slower ion-exchange chromatog-raphy peak shows identical elution time with that of native HbBv proving that HbBv-DVS is an intramolecularly-modified monomeric (64 kDa) hemoglobin derivatives (Fig~ 3a, b).

Increasing the reaction t.ime from 24 to 48 hour~, can result in up to 80-gO~ of HbBv-DVS~ A 50% increase in the mol.ar ratio between DVS and HbBv can also provide similar yields. The problem, however, is that under these conditions there is a concomitant synthesis of small amounts o~ a dimeric (128 kDa) derivative (Fig. 3b). This situation is also illustrated in the electrophoretic analysis of a ~8 hour reaction mixture (Fig. 2c). While the band corresponding to the native HbBv virtually disap-peared showing that most of the hemoglobin was modified, the high mobility HbBv-DVS band is quite di~Eused, which is probably due to the polymerization.

Fig. 5 presents typical results of SDS-PAGE. The electrophoretic pattern of pure HbBv-DVS shows a single band with a mobility corresponding to molecular mass of 16 , . .
; .
: ~ :
~, ~7~

kDa (FigO 5C)~ For a 48 hour reaction mixture at relative-ly high protein loading, an additional faink band corre-sponding to a molecular mass twice that of the major band component can be observed (Fig. 5b). When loading small amounts of proteins on the gels it is sometimes possible to obtain separation of the ~ and B chains. In Fig. 5E there can be seen the ~ and B chain bands for the native HbBv.
In Fig. 5g there is a similar pattern for the modified ~bBv-DVS.
PURIFICATION AND CHEMICAL CHARACTERIZATION OF POLY HbBv-DVS

HP-GPC o~ typical Poly HbBv~DVS preparations is pr~sented in Fig. 4. For a 24 hour reaction mixture the amounts of the modified monomeric hemoglobin (HbBv-DVS), and the modi~ied polymerized hemoglobin (poly HbBv-DVS), are nearly equal. The isolated Poly HbBv-DVS produced under these ~onditions had a molecular mass range ~rom 130 to about 500 k~a with a weight-average molecular weight, M~
~200,000 (Fig. 6a). For a 72 hour reactisn mixture the yield of Poly Hb~v-DVS in this case is from 130 to well above 500 kDa (Fig. 4b).

Cellulose acetate electrophoresis of isolated Poly Hb~v~DVS, obtained from a 24 hour reaction mixture, is shown in Fig. 2d. There is only one very di~ferent band with a mean mobility much higher than that of HbBv or even HbBv-DVS.

Fig. 5d shows the SDS-PAGE pattern of Poly HbBv-DVS isolated from a 24 hour reaction mixture. One sees bands corresponding to a molecular mass of 16 kDa and multiples thereo~, thus indicating the production of a mixture o~ modi~ied polymerized hemoglobins. llhe slowest HP-GPC peak isolated ~rom the same reaction mixture exhibits one band corresponding to a molecular mass of 16 kDa. This ~and is identified with the nonpolymerized HbBv-- , , , ~, . . .
: , :
,;. :-,, 2~7~g52 - lg -DVS (Fig. 5e and 6a).

OXYGEN-BINDING PROPERTIES

The ~xygen equilibria of Hb~v, HbBv-DVS and Poly ~bBv-DVS were studied in the pH range 5.8-9~0. The effect of pH on the oxygen a~finity is presented in Fig. 6. one can see that while near p~ ~.0 the P50 values for all three materials are similar (~75 mm Hb)~ they vary at higher pH
values. At pH 9.0, for example, the P50 value for native HbBv is 12 mm Hg while those for HbBv-DVS and Poly HbBv-DVS
are 31 and 41 respectively.

Figure 7 presents the effect of temperature on the oxygell affinity of HbBv, HbBv-DVS and Poly HbBv-DVS
solutions at pH 7.40 and shows that both HbBv-DV5 and Poly HbBv-DVS still exhibited P5~ values of about 20 mm Hg at the low temperature of 15 C, by contrast with only 7 mm Hg for native ~ovine hemoglobin.
2~
~ comparison of the oxygen unloading properties at 15 C and 37 C of products acaording to the present invention to other human and bovine oxygen carriers is provided in Table I. From this data it i8 seen that at 15 C the oxygen-unloading of HbBv-DVS (~ = 20%) and Poly HbBv-DVS (~ = 22%) are comparable to that of human blood at 37C (~ = 25%). Thus under hypothermic conditions Hb derivatives according to the pres~nt invention display very desirable oxygen-delivering capacities.
INTRAVASCULA~ R~T~NTION IN T~

Clearance of both HbBv and HbBv-DVS from the circulation is presented in the top portion of Fig. 8.
With dos~s yielding initial hemoglobin plasma contents of ~10 mg/ml, HbBv-DVS showed a vascular half-life of 100 ~ 10 min (n=3 rats), which is not significantly dif~erent from .. . ..... .
- , . . .. . . .
., ::: . , : :, ;. , ' ., ' ~:

2~7~852 that of HbBv (~0 + l0 min, n=3 rats). Poly HbBv-DVS, on the other hand, displayed a much longer half-life of ~4.5 hours. A lower dose of the same material exhibited a similar curve but resulted in a slightly shorter half-life of 4.0 hours (Fig. 8/ bottom).

VISCOSITY STUDIES

Absolute viscosities at 15 C and 37 C for concen-trations of 7.0 and 14.0 g/dl were determined for HbBv,HbBY-D~S and Poly HbBv DVS. The viscosities of HbBv-~VS
solutions were found to be identical to those of HbBv solutiGns at the same temperatures and concentrations. The viscosities of HbBv-DVS solutions, at any measured set of temperatures and concentrations, are much lower than that of human whole blood at 37 C. At 15 c, the viscosity of a 13 g/dl Poly HbBv-~VS solution i5 lower than that of human whole blood. The viscosity of a 14.0 g/dl Poly HbBv-~VS
solution at 15C is only slightly higher than that of human whole blood at 37 C. These viscosity properties indicate that these hemoglobin derivatives are likely to have beneficial effects when used to preserve isolated organs -a situ~tion that requires perfusion of a constricted microvasculature under hypothermic conditions.
One symmetrical peak in ion-exchange chromatogra-phy and one sharp band in cellulose acetate electrophoresis indicate a well-deEined molecule. The absence of any 32 kDa band from the SDS-PAGE pattern proves the lack of intramolecular crosslinkage, while the single-peak HP-&PC
demonstrates the absence of intermolecular crosslinkage.

The introduction of a molecule with multiple negative charges into the B-cleft of human or bovine hemoglobins could generate strong electrostatic interac-tions, resulting in a pseudo-crosslinkage, that confers stability to the tetrameric structure of hemoglobin. Such - , , ,.:
' : ' : , :

2~7~8~2 ~ 21 -tetrameric pseudo-crosslinkage is characterized by half~
life intravascular retention time increased four to five-fold with respect to that o~ normal hemoglobin [5,6]. The similar rekention times o~ the native and modified hemoglo-bins in the pre~ent investigation rule out ~he possi~ilityfor such pseudo-crosslinkage in HbBv-DVS.

At a 15% concentration of HbBv-DVS and relatively high molar ratio of ~VS/HbBv-tetramer it was possible to anaerobically polymerize the modified hemoglobin. From HP-GPC studies it was determined that 50-80% of the starting material was polymerized and that the product, Poly HbBv-DVS was not homogeneous with respect to the molecular weight. Consistent with these ohservations, the SDS-PAGE
analysis gave bands with mobilities corresponding to molecular masses of 16 kDa and multiples thereof. The absence of a 32 k~a band in the slowest HP-GPC peak iso-lated from a 24 hour reaction mixture (Fig. 5e) demon-strates the lack o~ intramolecular arosslinkage, even at the high DVS concentration needed for the polymerization, and strongly suggests that Poly HbBv-DVS does not contain intramolecular crosslinkages as part of its structure.
This conclusion is con~istent with the very low oXygen a~inity and methemoglobin content of this material.
REFERENC~S

1. De Venuto, F., Zuck, T.F., Zegna, A.I. and Moores, W.Y. (1977) J. Lab. Clin. Med. 89, 509-516.

2. Shih, T.B., Jones, R.T. and Johnson, C.5. (1982) Hemoglobin 6, 153-167.

3. Gebott, M.D. and Peck, J.M. (1978) Beckman Micro-zone Electrophoresis Manual, Chapter 8A, Beckman Instruments Inc., Fullerton, U.5.A.

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

, ~7~2 4. Weber, K., Pringle, J.R. and Oshorn, M. (1972) in MethQds in Enzymology (Hirs, C.H.W. and Timashef~, S.N., eds.~, Vol. 26, pp, 3-27, Aca-demic Press, London.

5. Bucci, Eo~ Razynska, A., Urbaitis, B. ~nd Fronticelli, C. (1989) J. Biol. Chem. 264, 6191-~O Fronticellil C., Bucci, E. ~azyn ka, A., Sznajder, J., Urbaitis, B. and Gryczynski, Z.
(1990) Eur. J. Biochem. 193, 331-336.

, , ~ ,, " ~ , . :.,,

Claims (7)

1. A process for producing low oxygen affinity hemoglobin products, which comprises:

in a first step, reacting deoxygenated hemoglobin in solution at a concentration of from 5-10% w/v with divinylsulfone under anaerobic conditions, to obtain a modified Hb-DVS which has a molecular weight similar to that of the reactant Hb and which is substantially free from intramolecular crosslinking;

and, in a subsequent step, reacting the modified Hb-DVS product of the first step, in solution at a concen-tration of from 12-20% w/v, with divinylsulfone under anaerobic conditions, to obtain a product comprising a mixture of said modified Hb-DVS and polymers thereof with divinylsulfone-derived crosslinks, both said products being substantially free from intramolecular crosslinking.

2. The process of claim 1 wherein, in the first step, the divinylsulfone is added to the reaction mixture in an amount of from 0.5 - 1.5 microlitres per ml.

3. The process of claim 2 wherein, in the subsequent step, the divinylsulfone added to the reaction solution is Prom 2 - 5 microlitres per ml.

4. The process of claim 1 wherein, both steps are conducted at pH's in the range 7-7.8, and at temperatures from 1-5°C.

5. A low oxygen affinity modified hemoglobin product comprising:

a predominant amount of a polymerized hemoglobin in which the hemoglobin units are linked by divinylsulfone residues, each said polymerized hemoglobin having a molecu-lar mass which is a multiple of 16 kDa and being substan-tially free from intramolecular crosslinking;

and a lesser amount of a divinylsulfone modified hemoglobin which is substantially free from intramolecular crosslinking and intermolecular crosslinking.

6. The modified hemoglobin product of claim 5 comprising from 50-80% of polymerized hemoglobin.

7. A polymerized hemoglobin consisting essentially of hemoglobin units intermolecularly linked by divinylsul-fone residues, and being substantially free from intramole-cular crosslinking.