THE USE OF CROSS-LINKED HEMOGLOBIN IN TREATING SUBARACHNOID HEMORRHAGE
Background of the Invention A release of blood into the subarachnoid space occurs following breach of a blood vessel, as for example, in the rupture of an aneurysm in the arterial blood supply to the brain. The pooling of blood in the subarachnoid space exposes the brain dura matter to blood contact. Over a period of two to three days a number of events are thought to occur: red blood cells (RBC) begin to lyse, liberating RBC components including free hemoglobin into the surrounding subarachnoid space, and the subsequent progressive conversion of oxyhemoglobin to methemoglobin with the possible production of superoxide anion radicals. Considerable evidence suggests that these RBC components mediate the pathogenesis of cerebral vasospasm, a condition associated with significant morbidity and mortality.
More recently there have been many attempts to identify and isolate the spasmogen contained in the RBC components. Many investigators have identified hemoglobin as the vasoactive agent. For a comprehensive review, see Macdonald & Weir, Stroke. 22:971 (1991) . A number of investigators have contributed significantly to this research including Miyaoka et al. , Neurol . Med. Chir.. 16:103 (1976), Sonobe et al . , Acta Neurochir.. 44:97 (1978), and Okamoto, Nippon Geka Hokan. 51:93 (1982) . In addition, the effect of hemoglobin in causing vasoconstriction of cerebral arteries is well documented (for example, see Cook et al. , Proc. West Pharmacol- Soc.. 22:429 (1979) . The prevailing paradigm in the field is that hemoglobin is the biochemical culprit in initiating vasospasm, although the mechanism of action is
-
incompletely understood. See generally, New Trends in Management of Cerebro-Vascular Malformations, Proceedings of the International Conference, Verona, Italy, eds. Pasqualin et al. , Springer-Verlag, New York (1994) . One theory is that hemoglobin acts by causing release of endothelin, a powerful vasoconstrictor, which in turn may mediate the vasospasm (Costentino et al., Stroke. 25:904 (1994) . Another theory is that vasoconstriction results when hemoglobin scavenges nitric oxide.
In subarachnoid hemorrhage, an artery bursts (typically an aneurysm) and floods the subarachnoid space to more or less uniformly coat the brain mass. In the period of about three to four days, the presence of blood components may cause regions of vasospasm, or severe vasoconstriction in which the neural tissue becomes ischemic resulting in neuronal injury and death. The effects may be sufficiently severe and cover a large enough portion of the total brain mass to result in serious neurological impairment or death.
Thus, therapies which prevent or minimize ischemia arising in subarachnoid hemorrhage have great benefit. Heretofore, several drug regimens have been proposed for treating vasospasm. For example, intravenous administration of nicardipine or nimodipine have been used with some success. A discussion of available therapeutic approaches is contained in E.C. Haley, "Principles of Pharmaceutical Therapy for Vasospasm Following Subarachnoid
Hemorrhage" , New Trends in Management of Cerebro- Vascular Malformations, supra. p. 85.
Summary of the Invention Investigators and clinicians have sought therapies, generally comprising a regimen of administering a drug or combination of vasoactive
drugs, to improve intracranial blood flow in order to prevent or reduce the amount of permanent injury to the brain tissue in the management of subarachnoid hemorrhage. Accordingly, it is an object of the present invention to provide a method for reducing neuronal damage arising from post-hemorrhage vasospasm.
In the present method, solutions of cross-linked hemoglobin are infused into a patient or other mammal after the onset or suspected onset of subarachnoid hemorrhage in a therapeutically effective dose, generally in the range of 1000 to 5500 mg/kg of body weight. It is found that contrary to the heretofore observed apparent adverse effects of free hemoglobin in the subarachnoid space, infused cross-linked hemoglobin reduces the area of hypoperfusion and regions of tissue at ischemic risk, prevents post- hemorrhage vasospasm, and reduces neuronal damage.
Thus, in accordance with the present invention, a method of reducing the area of hypoperfusion and ischemia, preventing or limiting vasospasm, and reducing neuronal damage in patients and other mammals undergoing or suspected of undergoing subarachnoid hemorrhage comprises administering, by infusion or otherwise, a therapeutically effective amount of cross-linked hemoglobin. The dose range is generally 1000 to 5500 mg/kg of body weight, and is administered within 72 hours after the onset of hemorrhage. Administration of hemoglobin up to and during the first 72 hours post-hemorrhage may be in a single bolus, or in a series of infusions of individual doses in the 1000 to 5500 mg/kg range.
The hemoglobin of the present invention is cross-linked intramolecularly or intermolecularly to prevent dissociation of the individual protein subunits thereof, and to maintain the physiological
oxygen release properties of hemoglobin in the intact RBC.
Brief Description of the Drawings Figure 1 is a schematic of a longitudinal cross- section of the rat cranium showing the anatomical relation of the major structures of the brain to the catheter position for subarachnoid infusion of blood in the experimental model. Figure 2a-e depicts in cross-section the tissue slices obtained for histological analysis, derived from typical test animals undergoing experimental subarachnoid hemorrhage.
Detailed Description of the Preferred Embodiment
It has long been known that free hemoglobin released from disrupted RBCs has a significantly higher binding affinity for oxygen than in its natural counterpart in the red cell. This high affinity binding makes the hemoglobin less useful as an oxygen carrying molecule because of its poor release properties in the tissues. It was subsequently discovered that molecular cross-linking forces the hemoglobin tetramer into a conformation in which the binding affinity of oxygen approximates that of intact red cells. The acceptable P50 values for the cross- linked hemoglobins of the present invention is between 20 and 40 inclusive. The cross-linking also stabilizes the tetrameric hemoglobin which otherwise tends to dissociate into dimers.
Also within the scope of this invention are the therapeutic uses of cross-linked hemoglobins which have been further polymerized to produce macromolecules ranging from 120,000 to 600,000 molecular weight, or which are decorated with molecular strands which retard degradation and increase half-life of the hemoglobin. Decorating
molecular strands may be polymers of polyalkylene oxide, chondrointins, or polyamides.
The cross-linking may be carried out by any one or a combination of methods known in the art. For example, U.S. Patent Nos. 4,001,401 and 4,053,590 disclose intramolecular cross-linking between an alpha and beta subunit of a hemoglobin tetramer utilizing halogenated aromatic compounds such as cycloalkanes, diepoxides, and diazobenzidines. U.S. Patent No. 5,248,766 discloses a cross-linking polymerizing strategy covalently interconnecting tetrameric units with oxiranes to form polyglobins with molecular weight in excess of 120,000 daltons. Some of these compounds have a P50 in the range of 35 to 40. U.S. Patent No. 4,777,244 discloses a method for cross- linking with aliphatic dialdehydes.
In the present therapeutic embodiments, the preferred tetrameric hemoglobin is cross-linked with bis (3, 5-dibromosalicyl) fumarate to create a diaspirin cross-link between the two alpha subunits, as described in U.S. Patent Nos. 4,598,064, 4,600,531, and RE 34,271. In general any method of cross-linking which yields a free tetramer having a P50 in the operative range of 20 to 40 will have efficacy in the present therapies. Conditions may be adjusted for each such cross-linked tetramer or polymer derived therefrom without undue experimentation. Other requirements for a therapeutic hemoglobin (i.e. endotoxin levels below about 0.25 EU/ml, being virus- free, having uncross-linked hemoglobin residuals below 0.2%, etc.) are all well-known to those skilled in the art. A complete list of hemoglobin specifications for one particularly preferred form of hemoglobin useful in the present invention is set forth in Table 1 of the Example.
Administration of cross-linked hemoglobin is preferably by intravenous infusion, although arterial
- .
cannulization or other drug delivery method may be efficacious. It is apparent that the surprising reduction of tissue ischemia, hypoperfusion, and associated vasospasm results from an action upon the affected tissue beds, so that introduction of the therapeutic hemoglobin at the site of the lesion or directly into the subarachnoid space would not be expected to demonstrate a benefit, and may even precipitate a vasospastic episode. Therefore, a mode of administration involving perfusion of the brain tissue beds, preferably through systemic infusion, is important.
The quantity and timing of hemoglobin administration may vary according to the circumstances of the individual case. The effective dose range is between 1000 and 5500 mg/kg of body weight, with a preferred dose of 2000 to 4000 mg/kg. It may be advantageous to administer an initial dose of 1500 to 3000 mg/kg with a later supporting dose of a like amount. Since vasospasm may occur over a 3 to 4 day period post-hemorrhage, and because unpolymerized cross-linked hemoglobins have a relatively rapid clearance rate, a multi-dose regimen may be required to maintain hemoglobin concentrations at a therapeutically effective level over a sustained period.
The level of free circulating hemoglobin may be monitored conventionally, and treatment levels may be adjusted for parameters known to those skilled in the art. It is also advantageous to administer equivalent doses of cross-linked hemoglobin in situations where ischemic stroke is suspected from hemorrhagic or thrombolitic origins, and before definitive diagnosis has been obtained. Further advantages of present invention will be apparent from the Example which follows.
Example
In the following experiments, the effect of a preferred cross-linked hemoglobin, alpha-alpha diaspirin cross-linked hemoglobin (DCLHb) , on cerebral blood flow and brain injury was assessed, following experimentally induced subarachnoid hemorrhage in rats. The DCLHb solution, obtained from Baxter Healthcare Corporation (Deerfield, IL) was prepared according to Chateerjee et al. , J. Biol . Chem. 261:9929 (1986) . Outdated human red blood cells were lysed by exposure to hypertonic buffer. The hemolysate was centrifuged to separate and remove stroma lipids. After ultrafiltration, molecular hemoglobin was cross- linked at the α chain by reaction with the diaspirin compound, bis (3, 5-dibromosalicyl) fumarate.
Elimination of viral contamination and protein purification was achieved by heat pasteurization. Estep et al. , "Virus inactivation in hemoglobin solutions by heat", Blood Substitutes. Edited by Change TMS, Geyer RP, New York, Marcel Dekker, pp. 129-134 (1989) . The final DCLHb solution had a concentration of 10.2 g/dl-1 (see Table 1). The DCLHb solution was stored at -70*C until needed for the current study at which time it was thawed to 5'C, and on the day of the study passively warmed to room temperature. Oxygen transport of DCLHb is similar to whole blood with a slight right shift in the oxygen dissociation curve. The α-OC cross-linking with bis (3, 5-dibromosalicyl) fumarate prolongs the intravascular half-life to about 24 hours. The viscosity of DCLHb (1.3 centistokes) is comparable to serum albumin and considerably less than whole blood (>4.0 centistokes).
Table 1
Chemical Assay of 10% Diaspirin Cross-linked
Hemoglobin Solution
Hemoglobin content 10.2 g/dl
Methemoglobin 0.7 g/dl p50 ( 37 ' C ) 32.0 mmHg
Osmolality 290 mOsm/kg
Oncotic Pressure 42.7 mmHg
Viscosity 1.3 centistokes
PH 7.50
Na+ 140 mEq/L
K+ 5.0 mEq/L
Ca++ 2.2 mEq/L
Mg++ 1.0 mEq/L ci- 115 mEq/L
Lactate 30 mEq/L
The subarachnoid catheter was prepared by tying a small knot at one end of a 3 cm length of polethylene tubing (PE-10) . The catheter was trimmed to allow 5 mm between the knot and the proximal end of the catheter. A small amount of cyanoacrylate glue was placed on the knot. The catheter was flushed with saline to assure free passage of fluid and sealed with HematoSeal at the distal end. Male Spontaneously Hypertensive Rats (weight, 350-400 grams; age, 16-20 weeks) were anesthetized with 1.2 MAC isoflurane (1.44% end-tidal) via a face mask as described in Cole et al., Lab Anim. Sci.. 40:506 (1990). The animals were placed prone on a sterotaxic head holder. Temperature was servo- controlled at 37'C. The model for experimental subarachnoid hemorrhage has previously been described by Solomon et al. , "Decrease in cerebral blood flow in
rats after experimental subarachnoid hemorrhage: a new animal model", Stroke. 16:58-64. After a betadine prep, the occipital crest was located and a small midline incision made. The occipital bone was cleared of muscular tissue and the atlanto-occipital membrane identified and cleaned of extraneous connective tissue. Hemostasis was achieved with gelfoam as needed. Careful dissection prevented opening of the atlanto-occipital membrane as this lies directly over the cisterna magna and perforation would cause leakage of cerebrospinal fluid (CSF) . The catheter was positioned over the membrane and a 22-g needle advanced through the membrane to a depth of <1 mm. A small amount of CSF drainage occurred, confirming correct location, and the catheter was carefully placed in the cisterna magna (see Figure 1) . The cannula was advanced along the inner table of the occipital bone until the knot rested against the membrane (care taken to insure superficial placement) . After placement of the catheter a small amount of dental acrylic was applied over the knot to secure it in place. The deep muscle layer was closed and the catheter sutured in place. The superficial muscle layers and skin were closed, and the skin infiltrated with 0.375% bupivicaine. The rat was returned to an incubator and allowed to recover for 72 hours, with a 12 hour light-dark cycle. Rats exhibiting any neurologic sequelae were excluded from further study (rare) . Animal Preparation: Following the 72 hour recovery period, the animals were anesthetized as above, orotracheally intubated and ventilated with a Harvard Rodent Respirator (Boston, MA) . The femoral vessels were cannulated for continuous blood pressure monitoring (Micro-Med Analyzer, Louisville, KY) , blood sampling, and fluid administration. Temperature was servo-controlled at 37'C with a heating blanket.
Arterial blood (125 μl) was collected at 30 minute increments and analyzed for pH, PaCθ2 Paθ2, glucose, and hematocrit (IL-1306 pH Blood Gas Analyzer, Instrumentation Laboratory, Lexington, MA; YSI Model 23-A Glucose Analyzer, Yellow Springs Instruments,
Yellow Springs, OH; IEC MB Centrifuge Microhematocrit, DAMON/IEC Division, Needham Heights, MA) . Prior to administration of subarachnoid blood, each rat was placed in a prone, 20" head-down, position to insure entrance of blood into the basal cistern. The sealed catheter was cut at the distal end and 20-50 μl of CSF aspirated. Fresh autologous blood (0.3 ml) was infused over 10 minutes into the cisterna magna, with an accompanying average increase in mean arterial pressure of 20-30 mmHg for 10 minutes. The rat was maintained in the head-down position for a total of 20 minutes.
Part A: Each rat was randomized to one of the following hypervolemic-hemodilution groups: Control (n=10) : 7.5 ml of fresh donor blood was given (no hematocrit manipulation [45%] ) .
30/DCLHb (n=10) : blood volume and hemocrit (30%) were manipulated by a 3.0 ml exchange transfusion with 10% DCLHb (Baxter Healthcare Corporation, Deerfield, IL, Lot 94D01AD11) followed by an additional 7.5 ml infusion of DCLHb.
30/Albumin (n=10) : blood volume and hematocrit (30%) were manipulated by a 3.0 ml exchange transfusion, and a 7.5 ml infusion, of oncotically- matched (7.5%) human albumin solution (Baxter Hyland, Glendale, CA, U.S.A.).
Each exchange transfusion was accomplished by simultaneously withdrawing and infusing the appropriate solution at a rate of 1.0 ml/min; and each 7.5 ml hypervolemic infusion was administered over 15 minutes. When given as a bolus, hemoglobin substitutes increase blood pressure. Rabinovici et
al . , "Characterization of hemodynamic, hematologic, and biochemical responses to administration of liposome-encapsulated hemoglobin in the conscious, freely moving rat", Circ. Shock. 29:115-132 (1989) . However, in this species, if DCLHb is given initially as an exchange transfusion, normotension is maintained. Cole et al. , "Focal cerebral ischemia in rats: effect of hypervole ic hemodilution with diaspirin cross-linked hemoglobin versus albumin on brain injury and edema", Anesthesiologv. 78:335-342 (1993) .
The procedure for measuring cerebral blood flow (CBF) is given in Sakurada et al . , "Measurement of local cerebral blood flow with iodo-C-14-antipyrine" , Am. J. Phvsiol.. 234:H59-H66 (1978) and Cole et al . ,
"Focal cerebral ischemia in rats: effect of hemodilution with α-α cross-linked hemoglobin on CBF",
J. Cereb. Blood Flow Metab.. 12:971-976 (1992) .
Immediately before CBF determination, physiologic parameters were evaluated as above. 100 mCikg~l of 14c-iodoantipyrine (New England Nuclear, Boston, MA) was given at a constantly increasing rate over 46 seconds. Twenty-one arterial blood samples were collected for determination of l^C activity with a quench correction (Bechman 8000 Liquid Scintillation Spectrometer [Beckman, Brea, CA, U.S.A.]) . After the ^c was infused, the brains were removed in <60 seconds and placed in 2-methylbutane (-35'C) . The brains were sectioned in 20 μm increments, and ten sections surrounding each of five anatomically predetermined coronal planes were placed on x-ray film (Kodak OM-1, Rochester, NY) for 21 days. The five anatomical planes were in 2.0 mm sequential increments. Section 1 was at the anterior midline extent of the corpus callosum, and Section 5 was 1.0 mm posterior to the posterior midline extent of the corpus callosum.
After film processing (21 days), assessment of CBF was done with a computer program based on the equation of Sakurada et al. , "Measurement of local cerebral blood flow with iodo-C-14-antipyrine" , Am. J. Phvsiol.. 234:H59-H66 (1978). A tissue-blood partition coefficient of 0.80 was used, and each autoradiograph calibrated to nine ^-^C standards (Amersham, Arlington Heights, IL) . By use of a Drexel/DUMAS Image Analysis System (Drexel University, Philadelphia, PA) each anatomical section was analyzed to define areas with a CBF of <40ml-100g~l-min-1. All image analysis was performed by an independent observer who was blinded to study protocol. The data was evaluated by analysis of variance with Scheffe'ε test for multiple comparisons, and mean values compared by t-tests as appropriate. The results are shown in Table 2.
Table 2
Control DCLHb Alb
Area of Hypoperfusion (%) 49 ± 13 13 + 5' 6 ± 3
Table area of hypoperfusion (CBF<40 ml lOOg-1 min-1) in a coronal brain section (% of total area, mean +.
SD) . * p<0.05 versus the Control group. t p < 0.05 versus the other two groups.
The area of hypoperfusion was less in the DCLHb group versus control, and was less in the Alb group versus the other two groups.
Part B: Different rats were prepared identically to Part A (n=6 for each group) . After subarachnoid hemorrhage they were allowed a 96 hour recovery period, after which they were anesthetized, a thoracotomy performed, and the brains perfused with formalin for standard hematoxylin and eosin staining as described in Sheehan et al. , "Theory and Practice of Histotechnology" , Columbus. Battelle Press (1980) . Six coronal sections (7 μm) coronal sections (from the parietal cortex anteriorly to the cerebellum posteriorly), were evaluated microscopically (X400) for the presence of dead neurons, which were individually counted and reported as a total for each animal (see Figure 2) . Dead neurons were defined as exhibiting pyknosis, karyorrhexis, karyolysis, cytoplasmic eosinophilia or loss of hematoxylin affinity, or dark, scalloped and swollen neurons. The criteria described in Garcia et al. , "Neuronal necrosis after middle cerebral artery occlusion in Wistar rats progresses at different time intervals in the caudoputamen and the cortex", Stroke. 26:636-643 (1995) were followed.
All microscopy was performed by an independent observer who was blinded to study protocol. The data
was evaluated by analysis of variance with Scheffe's test for multiple comparisons, and mean values compared by t-tests as appropriate. P<0.05 was considered significant. The results are shown in Table 3.
Table 3
Control DCLHb Alb Dead Neurons 510 ± 203 166 + 42+ 284 +
67 *
Table number of dead neurons in sixteen brain areas (mean ± SD) . * p<0.05 versus the Control group. t p<0.05 versus the other two groups.
The number of dead neurons was less in the Alb group versus the Control group. Significantly, fewer dead neurons were counted in the DCLHb group than in the other two groups.