Effect of Early SANGUINATE^® (PEGylated carboxyhemoglobin bovine) Infusion on Cerebral Blood Flow to the Ischemic Core in Experimental Middle Cerebral Artery Occlusion

Abstract

Background:

Sanguinate, a bovine PEGylated carboxyhemoglobin-based oxygen carrier with vasodilatory, oncotic and anti-inflammatory properties designed to release oxygen in hypoxic tissue was tested to determine if it improves infarct volume, collateral recruitment and blood flow to the ischemic core in hyperacute MCA occlusion (MCAO).

Methods:

Under an IACUC approved protocol, fourteen mongrel canines underwent endovascular permanent MCA occlusion (MCAO). Seven received Sanguinate (8mL/kg) intravenously over 10 minutes starting 30 minutes following MCA occlusion while seven received a similar volume of normal saline. Relative cerebral blood flow (rCBF) was assessed using neutron-activated microspheres prior to MCAO, 30 minutes following MCAO and 30 minutes following intervention. Pial collateral recruitment was scored and measured by arterial arrival time (AAT) immediately prior to post-MCAO microsphere injection. Diffusion-weighted MRI (3T Achieva, Philips Healthcare, Best, Netherlands) was used to assess infarct volume approximately 2 hours following MCAO.

Results:

Mean infarct volumes for control and Sanguinate treated subjects were 4739 mm³ and 2585 mm³ (p =0.0443; r²=0.687), respectively. Following intervention, rCBF values were 0.340 for controls and 0.715 in the Sanguinate group (r²=0.536; p=0.0064). Pial collateral scores improved only in Sanguinate treated subjects and AAT decreased by a mean of 0.314 seconds in treated subjects and increased by a mean of 0.438 seconds in controls (p<0.0276).

Conclusion:

Preliminary results indicate that topload bolus administration of Sanguinate in hyperacute ischemic stroke significantly improves infarct volume, pial collateral recruitment and CBF in experimental MCAO immediately following its administration.

INTRODUCTION

Rapid blood flow restoration vial clot removal (i.e. embolectomy) has been shown to reduce debilitating effects of stroke whose success hinges on rapid treatment within the early hours of stroke onset. Emblematic of the necessity of rapid treatment, the Joint Commission on Accreditation of Healthcare Organizations has set standards requiring comprehensives stroke centers to rapidly triage and treat stroke victims. This pilot study assesses a new pharmaceutical agent designed to slow infarct progression in ischemic tissue and synergistically work with reperfusion strategies.

Sanguinate (PEGylated Bovine Carboxyhemoglobin) is an investigational biopharmaceutical oxygen carrier that facilitates the transfer of oxygen to oxygen-deprived cells and tissues using a unique oxygen-delivery system has the potential to treat disorders caused by anemia or hypoxia/ischemia (1). Sanguinate is purified and modified bovine hemoglobin that has been covalently conjugated with polyethylene glycol and combined with carbon monoxide to suppress vasoconstriction and provide anti-inflammatory effects (1–3). The bovine hemoglobin is modified to carry oxygen for targeted release to hypoxic tissue (i.e. with low partial pressure of oxygen). It has been shown to successfully address detrimental vasoconstrictive side effects associated with older generation artificial blood products and has been the only oxygen carrier to date which has demonstrated safety in humans in a phase I clinical trial (1–3). In the setting of acute ischemic stroke, Sanguinate has been shown to reduce infarct volume and improve pial collateral recruitment in an experimental small animal stroke model (4,5). The pathophysiology of stroke is complex with multiple concomitantly acting detrimental processes. Targeting a single process may not be adequate. It follows that potential for benefit is greater if multiple mechanisms for infarct progression can be simultaneously addressed. Through its vasodilatory effects at the pial arteriolar level, its ability to deliver oxygen to severely hypoxic tissues, its oncotic effect, and its anti-inflammatory effects, Sanguinate shows promise in slowing the progression of ischemic insult (5). Sanguinate’s mechanism of action differs significantly from most current neuroprotective strategies and represents a truly novel approach to neuroprotection during cerebral ischemia. We hypothesize that Sanguinate can help improve collateral blood flow, pial collateral recruitment and infarct volume during acute ischemic stroke due to middle cerebral artery occlusion (MCAO).

MATERIALS AND METHODS

The experimental protocol was approved by the University of Chicago Institutional Animal Care and Use Committee and is reported in compliance with Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines. Fourteen mongrel canines (8 female, median weight 28.9 kg (24.2–33.2 interquartile), median age 11 mo (9–13 mo interquartile) underwent baseline MRI under general anesthesia (see below), and 24 hours later underwent arteriography, and baseline cerebral blood flow (CBF) assessment by microspheres under general anesthesia (see below). Subjects subsequently underwent endovascular MCAO using embolic occlusion coils deployed at the M1 segment and carotid terminus (alternating left and right MCA between animals) using a previously described methodology (6). Vertebral and bilateral internal carotid arteriograms confirmed MCAO without involvement of other vessels and assessed pial collateral recruitment (see below). Thirty minutes following occlusion subjects received either Sanguinate (Prolong South Plainfield, NJ) (8mL/kg; n=7)) or normal saline (8mL/kg; n=7) as a bolus over 10 minutes via a 5F 23 cm femoral vein sheath whose tip was in the inferior vena cava. All control animals preceded the experimental group. Microspheres (see below) were used to assess CBF, prior to MCAO, 30 minutes following MCAO and 30 minutes following intervention. The subjects were then taken to MRI for quantification of infarct volume (see below). Subjects were euthanized either the same evening or the following day after which the brain was harvested (see below). Subjects were to be excluded for unanticipated events such as abnormal baseline MRI, intraprocedural vessel perforation or excessive deviation in physiologic parameters during MCAO.

Choice of anesthesia:

Anesthesia was induced with 6mg/kg IV propofol. A combination of 1.5% Isoflurane end tidal concentration, remifentanil, rocuronium and propofol was used thereafter. Because isoflurane impairs CBF autoregulation in doses over 1.5%, isofluorane was maintained at the same rate throughout the experiments.

Physiologic parameters:

Small arterial blood aliquots were used to measure pH, PaO2, PaCO2, glucose, hematocrit, and hemoglobin after intra-arterial access and prior to each microsphere injection using a blood analysis system (EPOC, Integrated MedCraft, Colchester CT). Blood pressure was measured continuously via a 5-French intra-arterial sheath and maintained within physiologic limits. ETCO2 was measured continuously and kept within normal range targeted between 35 and 40 mmHg. Central venous pressure was monitored via a femoral vein sheath. Temperature was monitored via rectal probe and maintained using heating pads. Blood glucose was maintained < 120 mg/dL.

Microsphere injection:

For each CBF assessment, approximately 10 million (4 mL) nonradioactive neutron activated microspheres (STERIspheres, BioPAL, Medford, MA) were injected into the left ventricle through a 4 French pigtail catheter over 20 seconds with 10mL saline flush while a 20mL reference blood sample was collected over 2 minutes from the abdominal aorta via a femoral sheath using a withdrawal pump (World Precision Instruments, Sarasota). Microsphere injection and blood withdrawal was repeated with different isotopes prior to MCAO, 30 minutes following MCAO and 30 minutes following Sanguinate infusion. Differences in microsphere isotope allowed for multiple CBF measurements. Excised brains were sectioned using a calibrated brain cutting matrix (Canine Brain Matrix, Stoelting Co., Wood Dale, IL). Three contiguous 6 mm thick sections at and posterior to the MCA were used. Each section was divided into 8 regions (inferior cortical, middle cortical, superior cortical, and deep for the left and right hemispheres corresponding to the core, and penumbra for MCAO (figure 1) and individually weighed wet and placed in a drying oven (Model 20 Lab Oven, Quincy lab, Inc, Burr Ridge, IL) prior to microsphere analysis at a blinded independent laboratory. Relative CBF (rCBF) was derived by comparing normalized CBF from the central portion of the middle cerebral artery territory (figure 1) to derive a ratio of ischemic to non-ischemic territories. Baseline CBF measures (prior to MCAO) were used to normalize rCBF measures in order to account for differences in gray versus white matter content from the sectioned tissue samples.

Figure 1:

Single sectioned slice (A) divided into four zones (B, superior and inferior penumbra zones, middle core zone and deep gray nuceli) in each hemisphere. Arrows (B) point to the central portion of the MCA territory considered to be the core infarct zone on the ischemic side.

Digital subtraction angiographic evaluation of pial collateral recruitment:

Arteriograms (OEC 9800, GE Healthcare, Milwaukee) acquired 30 minutes following occlusion and 30 minutes following intervention were evaluated in a blinded fashion for pial collateral recruitment by quantifying arterial arrival time (AAT, figure 2) and scoring pial collateral recruitment. AAT was defined as the time interval between contrast arrival at the normal M1 segment and contrast arrival at the reconstituted M3/4 junction on the hemisphere distal to the permanent MCAO (6). Pial collateral recruitment was assessed using a previously published 11-point scoring system which compares postocclusion and preocclusion arteriographic images to assess the extent of reconstitution of the occluded MCA territory and incorporates transit time relative to jugular vein opacification (6).

Figure 2:

Mean diffusivity images acquired approximately 2 hours post MCAO selected from two subjects with the same pial collateral score of 9. The regions of diffusion restriction (representing infarcted tissue) are depicted by arrowheads. The subject in figure A was a control with a larger infarct area compared to the subject in figure B which received Sanguinate.

MRI:

Subjects underwent MRI under general anesthesia prior to MCAO and in a separate session 2 hours following MCAO using a 32-channel sensitivity encoding coil on a 3T MRI (Ingenia, Philips Healthcare, Best, Netherlands) system. Diffusion tensor imaging (DTI) [field of view (FOV)=140×140mm, matrix = 128×128, number of excitations (NEX)=1, repetition time(TR)/echo time (TE) 192–2131/71, Slice thickness (ST)=3mm, b value= 1000, 32 directions scan time ~4 minutes] was acquired approximately 2 hours following MCAO. Susceptibility weighted imaging (SWI) was subsequently acquired [TR=17ms, TE=24ms, FA = 10° ms, BW = 192 Hz/pixel, matrix size = 336 × 336 × 200, FOV=140×140mm, voxel size = 0.44 × 0.44 × 0.5 mm³, slice thickness= 2 mm, voxel]. SWI images were visually inspected for hemorrhage.

Infarct volume determination:

Infarct volumes were calculated in a blinded fashion from DTI derived mean diffusivity (MD) maps on Image J (National Institutes of Health) using a recently published semi-automated threshold method. Briefly, the mean signal from the unaffected hemisphere was measured on each cross section of the MD map. Voxels in the affected hemisphere that exhibited MD values less the 1.5*s of the contralateral normal hemisphere were defined as “diffusion positive” and their volume summed to determine the total volume of infarcted tissue. Bland-Altman statistic for infarct volume calculated at 2 hours was 15.9% (6).

Infarct volume varies with pial collateral recruitment and time interval between MCAO and imaging; to account for this, we also compared differences measured infarct volume relative to the predicted infarct volume (estimate) between the two groups. The equation for predicted infarct volume was V(t)=V_f (1 - e ^(-Gt)) as previously described and validated (7,8) where G =−0.0013*PCS+0.0179, Vf =−2472.7*PCS+342971, t = time interval between imaging onset and MCAO (in minutes) and PCS = pial collateral score.

Statistical analysis:

JMP5 software (SAS Institute, Cary, NC) was used for statistical analysis. Power analysis using a type I error of 5% and power of 80% was used to estimate the required number of subjects based on anticipated mean change in rCBF (primary outcome measure) and standard deviation from control subjects in prior studies. Shapiro-Wilk Test was used to assess all dependent variables for normality. Normally distributed variables were expressed in mean values and assessed for significance using one-way analysis of variance (assuming unequal variances) whereas non-parametric variables were expressed as median and evaluated using Wilcoxan Rank Sums (one-way test, chi-square approximation). One-way analysis of variance was used to compare: 1) change in rCBF before and after intervention (Sanguinate/saline infusion); 2) decrease in AAT before and after intervention and 3) the difference in predicted versus measured infarct volume approximately 2 hours following MCAO based on pial collateral recruitment score and time elapsed between MCAO and start of imaging. To account for the impact of pial collateral recruitment on infarct volume and rCBF following intervention, least mean squares regression analysis was used.

RESULTS

All 14 experiments were successfully completed with no significant deviation in physiologic parameters and no hemorrhagic events identified on SWI or on autopsy. Demographic distribution is displayed in table 1. Pial collateral scores measured immediately prior to intervention ranged from 8 to 11. Mean arterial pressure 30 minutes following MCAO and 30 minutes following intervention increased by a mean of 4.14 (σ=7.45) mmHg in the Sanguinate group and decreased by a mean of 0.857 (σ=6.23) mmHg in the control group. Diffusion imaging in the experimental group was acquired at a mean of 145 minutes following MCAO versus 119 minutes in the control group. This significant difference was due to differences in transport time to the MRI suite between the two groups. Pial collateral score, rCBF and arterial arrival were measured at consistent timepoints between control and experimental groups and did not require subject transport. Accounting for pial collateral recruitment prior to intervention, infarct volumes were lower in the experimental group (figure 2, table 1). Similarly, Pial collateral score, rCBF and arterial arrival time improved following intervention (figure 3, table 2).

Table 1:

Subject demographics, MRI time, baseline pial collateral scores and baseline rCBF in the MCA ischemic zone

Subject Group	Female gender	Weight mean (σ)	Occlusion side	MCAO-MRI time-mean (σ)	Pial collateral score mean (σ)	Baseline rCBF mean (σ)
Sanguinate	3 of 7	30.4 (4.19) kg	4 right MCA	145(13.3) min.	9.43 (0.976)	0.441 (0.257)
Control	5 of 7	26.3 (4.73) kg	4 right MCA	119 (10.2) min.	9.57 (1.51)	0.300 (0.171)
P value	0.592*	0.101†	1.00*	0.0014†	0.838†	0.250†

^*2-Tail Fisher’s Exact Test

^†t-Test

Figure 3:

Composite of DSA arteriograms obtained by combing the full arteriographic series in two different DSA acquisitions from the same subject. One series was obtained 30 minutes following MCAO and immediately prior to Sanguinate infusion (A) and the other was obtained 30 minutes following Sanguinate infusion (B). Note the difference in pial collateral recruitment (oval regions) and the difference in caliber of the internal carotid artery before (A) and after (B) Sanguinate infusion.

Table 2:

Comparison of rCBF, increase in rCBF, pial collateral score and arterial arrival time, infarct volume and difference of predicted (estimate based on pial collateral score and time to imaging relative to MCAO) versus measured infarct volume between control and treatment groups.

Variable	Control (σ)	Sanguinate (σ)	Standard error	P	r2/χ2
Mean rCBF following intervention*	0.34 (0.194)	0.715 (0.322)	0.10	0.0064*	0.536
Mean increase in rCBF†	0.0405 (0.150)	0.274 (0.134)	0.0537	0.0096†	0.442
Median (quartiles) increase in pial score‡	0 (1/0)	+1 (0/+1)	NA	0.0057‡	7.63
Mean increase in arterial arrival time (s)†	0.438 (0.561)	−0.314 (0.561)	0.212	0.0276†	0.344
Measured infarct volume (mm3) *	4739 (4400)	2585 (2680)	777	0.0443*	0.687
Predicted - measured infarct volume (mm3)†	616 (1994)	4336 (2050)	765	0.0049†	0.496

^*Least Squares Regression – r2 adjusted

^†ANOVA

^‡Wilcoxan Rank Sums

DISCUSSION

This work demonstrates that early administration of Sanguinate following MCAO in a mongrel canine model improved pial collateral recruitment and cerebral blood flow to the ischemic territory and led to reduced infarct volumes. This evidence corroborates similar work in a spontaneous hypertensive rat model and a Wistar rat model where Sanguinate was shown to improve collateral perfusion to the ischemic territory and reduce infarct volume following MCAO (4,5).

The effect of Sanguinate on collateral blood flow during the 2 hours MCAO was previously investigated after 30 and 90 minutes of ischemia in spontaneously hypertensive rats using multi-site laser doppler probes at the core and the collateral zone (4). Similar to this study, collateral flow increased in Sanguinate-treated animals whereas controls demonstrated a decline in cerebral blood flow. Brain injury was smaller compared to controls in animals treated at 30 minutes but not 90 minutes. Because spontaneously hypertensive rats have poor collateral recruitment and faster infarct evolution (9), the infarct was likely already completed in animals treated at 90 minutes. We therefore suggest that in subjects with good collateral recruitment and slower infarct evolution, a difference in ischemic tissue damage in Sanguinate treated subjects versus controls would have been detected even if Sanguinate delivery was delayed. The current study only focuses on early CBF assessment. Most subjects had pre-intervention collateral scores of 9 or greater (good collateral recruitment). In theory, Sanguinate is expected to slow down the progression of tissue at risk to ischemic damage; as a result, Sanguinate is expected to lead to a greater impact greater impact in the early hours of MCAO where infarct volume growth is faster (7).

This work evaluated topload infusion of 8mL/kg Sanguinate administered over 10 minutes relative to controls 30 minutes following MCAO. It is not clear how well the dose and rate of delivery would be tolerated in humans. Sanguinate is generally prepared at 40mg/mL and has been evaluated 2,4 and 8mL/kg across 3 different species and delivered in a single dose over 5–15 minutes (1–3). Total blood volume in canines is estimated to be 77–78 mL/kg. A topload of 8 mL/kg did not appear to a have a substantial impact on mean arterial pressure and did not results in any significant physiologic events in this study. In human phase I studies, Sanguinate was delivered at 2,3 and 4 mL/kg over a 2-hour time period. As a result, the delivery method used in this study varied with what has been approved in current clinical studies. Additionally, Sanguinate releases CO in an exponential fashion with the majority released within 30 minutes enabling binding to oxygen and off-loading it in hypoxic conditions (pO2 of 5%). RBCs have a p50 of 26 mmHg while hypoxic tissues have values of 5 mmHg. Sanguinate has an average p50 of 7–16 mmHg. Sanguinate is presumed to be removed by the RES of RBCs in rats where is has an intravascular half-life of 12 hours. The intravascular half-life of Sanguinate is dose dependent. In humans, the intravascular half-life when delivered at 2, 3 or 4mL/kg over 2 hrs, was 6.6, 10.2 and 13.8 hours respectively. Sanguinate is associated with transient elevation of MAP and CVP attributable to the colloid osmotic pressure generated by the large and hygroscopic PEGylated protein (1–3). Additional studies will be needed to determine the best delivery method for Sanguinate.

Although encouraging, the scope of this work is limited and requires further exploration. The relatively small number of subjects could result in measurement bias, despite the well-controlled nature of this study. Anesthetics used in this study were chosen to minimize any influence on cerebrovascular reactivity and were kept at equivalent dosing between animals. The vasodilator reflex helps augment collateral supply during cerebral ischemia. For these experiments we limited isofluorane to levels which do not influence cerebrovascular reactivity and a skeletal muscle relaxant with the use of propofol and remifentanil to supplement anesthesia for ethical reasons. Although propofol and remifentanil reduce CBF and cerebral metabolic rate of oxygen they have no influence on cerebrovascular reactivity (10,11). Despite careful selection of anesthetics, the experimental outcomes may still be vulnerable to anesthesia by unknown mechanisms affecting collateral perfusion to an ischemic territory. It is also unclear whether the effect of Sanguinate on rCBF is sustained over time and over a range of pial collateral recruitment and what effect of Sanguinate has on reperfusion injury. Translation of the current results to humans also remains to be investigated.

As mentioned, current results corroborate findings in small animal studies (4,5). The endovascular canine model chosen for this experiment allows for precise and reproducible MCAO and adds value to research conducted on rodents. Canines have a long history of use in stroke research. Larger species have advantages over rodents due to gyrencephalic cortex and gray to white matter ratios that are closer to humans (12) as well as more sophisticated physiological monitoring and imaging methods. Comparative anatomic studies across vertebrate species indicates that the pial network organization along the forebrain of canines is more comparable to primates than that of felines or rabbits (13). Nonetheless, neuroprotection in young healthy animals does not necessarily translate to a human population with multiple comorbidities including older age, smoking, cardiac disease, diabetes, hypercholesterolemia and hypertension. Numerous preclinical studies demonstrating effective neuroprotection in acute ischemic stroke have failed to translate to the clinical setting (14). Sanguinate differs from many previously failed neuroprotectants in that it potentially addresses multiple targets (hypoxia, perfusion and inflammation).

CONCLUSIONS

Preliminary data presented here suggest that Sanguinate administered in the early phases of acute ischemic stroke improves CBF and pial collateral recruitment and reduces infarct volume. As such, Sanguinate administered prior to embolectomy in acute ischemic stroke has the potential slow the progression of ischemia to cerebral infarction. Further studies are needed to confirm these results and optimize its delivery and dosage in the setting of acute ischemic stroke.

ABBREVIATIONS:

MCAO

Middle Cerebral Artery Occlusion

PCS

pial collateral score

CBF

cerebral blood flow

rCBF

relative CBF

MAP

mean arterial pressure

AAT

Arterial Arrival Time

PEG

PEGylated - covalent conjugation with polyethylene glycol

MRI

magnetic resonance imaging

DTI

diffusion tensor imaging

MD

mean diffusivity

SWI

susceptibility weighted imaging

FOV

field of view

σ

standard deviation