Polynitroxyl Albumin and Albumin Therapy after Pediatric Asphyxial Cardiac Arrest: Effects on Cerebral Blood Flow and Neurologic Outcome

Introduction

Pediatric cardiac arrest (CA) results from asphyxia in 80% of cases. There are no therapies for improving neurologic outcome in children after asphyxial CA, with the possible exception of hypothermia, although it remains unproven. Outcome from pediatric asphyxial CA is poor: of children who sustain out-of-hospital CA, 30% have return of spontaneous circulation (ROSC), 12% survive to hospital discharge, and only 4% have favorable neurologic outcome (Donoghue et al, 2005).

Reperfusion and reoxygenation, indispensable to restore viability during cardiopulmonary resuscitation and after CA, may also have undesirable consequences. Although it is indisputable that reperfusion is essential for neuronal survival, the ideal reperfusion pattern after CA remains undefined. After 9 minutes of asphyxial CA in immature rats, hyperemia is seen from 5 to 15 minutes after ROSC in subcortical structures, whereas cortical hypoperfusion appears early and is sustained for up to 3 hours (Manole et al, 2009). It is postulated that cerebral hyperemia is beneficial, and hypoperfusion is detrimental for neuronal survival after CA (Snyder et al, 1975); however, a recent study in experimental cerebral ischemia—reperfusion has suggested that early postresuscitation hyperemia may be deleterious (Pignataro et al, 2008).

Reactive oxygen and nitrogen species generated during CA and reperfusion influence postresuscitation cerebral blood flow (CBF). Reactive oxygen and nitrogen species produce vascular damage and loss of autoregulation after brain injury (Nelson et al, 1992). Superoxide production increases with reperfusion after global ischemia in the brain (Kofler et al, 2005). Superoxide itself has limited reactivity; however, it reacts with nitric oxide to form peroxynitrite in a diffusion-limited reaction, leading to decreased nitric oxide availability, which can alter CBF after CA (Bayir, 2005). Treatment with a superoxide dismutase (SOD) mimetic early after CA has been shown to improve neurologic outcome (Cerchiari et al, 1987).

Polynitroxyl albumin (PNA) is an intravascular antioxidant that is synthesized by covalent addition of high molar ratio nitroxide (55:1 on average) to albumin. Polynitroxyl albumin has been shown to exert SOD-mimetic properties and to reduce infarct size after focal ischemia in rats (Beaulieu et al, 1998; Kuppusamy et al, 1996; Sugawara et al, 2001). In our pediatric asphyxial CA model, PNA decreased oxidative and nitrative stress induced by resuscitation with 100% oxygen (Walson et al, 2011). Albumin, the parent compound of PNA, has also been shown to improve perfusion and functional outcome in models of focal cerebral ischemia (Belayev et al, 1997, 1998, 2002; Liu et al, 2001). Besides colloidal properties, albumin binds redox-active transition metals, fatty acids, and heme and has some antioxidant properties due to a free cysteine residue (Gutteridge et al, 1984; Rowley et al, 1984). Given their large size, PNA and albumin are generally restricted to the intravascular space if the blood—brain barrier is intact, which is the case early after injury in our CA model (Manole et al, 2009). Accordingly, we hypothesized that PNA will ameliorate ischemia-induced CBF dysfunction and improve neurologic outcome versus treatment with albumin or normal saline (NS) after pediatric asphyxial CA in immature rats.

Materials and methods

All experiments were approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh and were performed in accordance with these guidelines and regulations. We used postnatal day 16 to 18 Sprague—Dawley rats of both sexes for CBF analysis (n=37, 6 to 7 per group) and for neurologic outcome (n=42, 10 to 12 per group). Rats were randomized to asphyxial CA (9-minute asphyxia) or sham, and further randomized to receive one of the following solutions at resuscitation or after sham surgery: PNA (20 mL/kg, 10%, SynZyme Technologies, Irvine, CA, USA), albumin (20 mL/kg, 10%, SynZyme Technologies), or NS (20 mL/kg). Ten percent of human serum albumin was prepared by dilution of 25% human albumin (Baxter, Deerfield, IL, USA; USP) with NS, and was supplied by SynZyme Technologies.

Results

Physiologic Data

The duration of CA was similar in the CA groups. Table 1 presents physiologic data (mean arterial pressure (MAP), PaCO₂, arterial pH, and PaO₂). Both NS- and albumin-treated groups had increased MAP versus shams at 10 minutes after resuscitation. Shams, NS-, and PNA-treated rats had MAP comparable with baseline, whereas albumin-treated rats had increased MAP versus baseline at 10 minutes after resuscitation. pH was lower in shams 120 minutes after resuscitation. This may result from the CA groups receiving sodium bicarbonate at resuscitation. PaO₂ was statistically lower in shams versus CA; however, this difference does not seem physiologically meaningful.

OPEN IN VIEWER

Table 1

MAP, PaO₂, PaCO₂, and pH at baseline and after asphyxial CA

		Baseline	5 minutes	10 minutes	30 minutes	120 minutes	150 minutes
MAP	CA-PNA	69±4	71±7	77±8	56±5	67±13	69±6
	CA-albumin	69±4	77±5	90±4∗	70±4	67±11	70±4
	CA-NS	80±3	82±4	93±6∗	75±4	72±14	75±4
	Sham	72±4	75±4	72±4	71±4	73±15	73±4
PCO 2	CA-PNA	33±2			43±6	39±3	41±3
	CA-albumin	35±2			51±5	36±3	32±2
	CA-NS	38±3			44±6	36±3	36±6
	Sham	36±1			39±2	41±3	42±2
pH	CA-PNA	7.23±0.02			7.20±0.05	7.33±0.03∗	7.31±0.03∗
	CA-albumin	7.24±0.03			7.15±0.04	7.37±0.02∗	7.37±0.03∗
	CA-NS	7.31±0.04			7.26±0.07	7.40±0.03∗	7.38±0.04∗
	Sham	7.24±0.02			7.26±0.02	7.21±0.02 †	7.20±0.02 †
PaO 2	CA-PNA	203±11			176±20	266±18∗	275±40∗
	CA-albumin	175±19			178±29	258±17∗	256±14
	CA-NS	216±9			217±27	249±17∗	269±25
	Sham	212±11			198±12	189±10 †	182±12

CA, cardiac arrest; NS, normal saline; PNA, polynitroxyl albumin.

∗P<0.05 versus sham, ^†P<0.05 versus CA-NS.

Cerebral Blood Flow

We report postresuscitation CBF for each rat normalized to its respective baseline (percentage baseline of CBF, Figure 1). Postresuscitation CBF displayed marked regional and temporal differences in the three treatment groups.

OPEN IN VIEWER

Figure 1

Regional postresuscitation CBF, percentage change from baseline. Color flow map of CBF at baseline, 5 and 150 minutes after CA. (∗P<0.05 CA-albumin versus shams; ^#P<0.05 CA-NS versus shams; ^†P<0.05 CA-PNA versus shams). CA, cardiac arrest; CBF, cerebral blood flow; NS, normal saline; PNA, polynitroxyl albumin.

Functional Outcome

Beam Balance and Inclined Plane Test

There were no differences between treatments after CA in either beam balance or inclined plane performance (Figures 2A and 2B). All CA groups had decreased beam balance performance versus shams from days 1 to 4 (P<0.05, Figure 2A). On the final day of beam balance testing, albumin and NS-treated rats performed similar to shams, whereas PNA-treated rats performed worse than shams (P<0.05). On the inclined plane, the CA groups performed worse than shams on the first 2 days of testing (P<0.05). There were no differences between groups on days 3 to 5 (Figure 2B). There were no sex-dependent differences in beam balance or inclined plane test performance between treatment groups (n=5 to 6 per sex per group).

Morris Water Maze

Visible Platform. On the first and third days of acclimatization in the water maze, there was no difference in performance between groups. On the second day, PNA-treated rats had reduced latency to finding the platform versus NS-treated rats (P<0.05). All groups except for albumin-treated rats improved the ability to find the platform from the first to the third day of acclimatization (P<0.05 for shams, PNA- and NS-treated rats; P=0.3 for albumin-treated rats). There was an overall effect of time (P<0.001). The overall performance of the four groups was similar (P=0.2). There were no sex-dependent differences between groups, except for the first day of acclimatization: NS-treated female rats performed better than both PNA- and albumin-treated female rats (n=5 to 6 per sex per group).

OPEN IN VIEWER

Figure 2

(A) Beam balance, (B) inclined plane test, (C) hidden platform test (∗P<0.05 CA-albumin versus shams, ^#P<0.05 CA-NS versus shams, ^†P<0.05 CA-PNA versus shams). (D) Probe trial. Triangles represent individual rats in each group, males (▴) or females (Δ), (∗P<0.05 versus CA-NS). CA, cardiac arrest; NS, normal saline; PNA, polynitroxyl albumin.

Hidden Platform. There was an overall effect of time (P<0.001). On the first day of testing, PNA-treated rats had reduced latency to finding the hidden platform versus NS-treated rats (P<0.05), performing similar to shams (Figure 2C). During the subsequent days, all groups except for PNA-treated rats improved the ability to find the hidden platform from the first to the fourth day, showing the ability to learn the task (P<0.05 for shams, albumin- and NS-treated rats; P=0.1 for PNA-treated rats). There was no difference among groups on the final day of the hidden platform test. The overall performance of the four groups was similar (P=0.3). There were no sex-dependent differences between groups (n=5 to 6 per sex per group).

Probe Trial. After hidden platform testing, we measured the time spent in the quadrant where the platform was previously located. Random swimming in all four quadrants would yield a time in the target quadrant of 15 seconds of the total 60-second testing period. Albumin- and PNA-treated rats spent more time in the target quadrant versus NS-treated rats (P<0.05). On post hoc analysis, this difference was attributable to benefit in males (n=5 to 6 per sex per group, Figure 2D). The power to detect a difference for groups and genders in the probe trial was 0.145. This gender difference in performance could not be explained by different durations of CA in males and females. The durations of CA in minutes were 5.5±0.4 versus 6.2±0.3 for PNA, 5.6±0.5 versus 5.6±0.5 for albumin, and 5.5±0.6 versus 4.7±0.6 for NS in male and female rats, respectively.

Neuropathological Outcome

The CA groups had decreased CA₁ hippocampal neuronal survival versus shams (P<0.05). There were no differences in cell counts among CA treatment groups and no difference between males and females (n=5 to 6 per sex per group). Hippocampal neuronal counts in dorsal CA₁ were 46,160±2,933 for shams, and 18,989±3,951, 22,053±2,853, and 21,848±2,803 for the CA groups treated with PNA, albumin, and NS, respectively (P<0.05 CA groups versus shams) (Figure 3). There were no differences in CA₃ hippocampal neuronal survival between CA groups and shams. Hippocampal neuronal counts in dorsal CA₃ were 25,635±1,869 for shams, and 26,929±3,479, 24,822±3,078, and 25,635±1,869 for CA groups treated with PNA, albumin, and NS, respectively (P=0.8, 0.6, and 0.3 versus shams).

OPEN IN VIEWER

Figure 3

(a) Neuronal counts in the CA₁ hippocampus. Triangles represent individual rats in each group, males (▴) or females (Δ). (∗P<0.05 versus shams). (b) Representative photomicrographs of the CA₁ hippocampus stained with Neu-N from a representative shams (A), CA-NS (B), CA-PNA (C), and CA-albumin (D) rat. Size bar=60 μm. CA, cardiac arrest; NS, normal saline; PNA, polynitroxyl albumin.

Discussion

To our knowledge, this is the first study exploring the effects of the intravascular antioxidant colloid PNA or albumin on either CBF or functional outcome after asphyxial CA in immature animals. Polynitroxyl albumin and albumin produced divergent CBF changes after asphyxial CA. Polynitroxyl albumin prevented early hyperemia, whereas albumin exacerbated hyperemia in all brain regions, including the cortex, a region where hypoperfusion is seen. Remarkably, despite divergent CBF changes, both PNA and albumin improved functional outcome assessed by the probe trial test compared with NS. The decrease in postresuscitation CBF after PNA suggests that reactive oxygen and nitrogen species contribute to hyperemia after CA. The improved functional outcome after PNA and albumin despite divergent early CBF changes suggests four possibilities: (1) either blunting or exacerbating early hyperemia is beneficial, (2) alterations of CBF do not influence functional outcome, (3) PNA improves functional outcome independent of effects on CBF, or (4) albumin improves functional outcome independent of effects on CBF in our pediatric asphyxial CA model.

Polynitroxyl albumin is a colloid that has nitroxide moieties with potent antioxidant properties (Kuppusamy et al, 1996). Nitroxides act as electron scavengers and have SOD-mimetic activity. Polynitroxyl albumin has been shown to attenuate superoxide-induced vascular adhesion of leukocytes to endothelial cells versus albumin (Russell et al, 1998). The dismutation of superoxide yields hydrogen peroxide (Yost and Fridovich, 1976). Hydrogen peroxide is a relatively stable molecule; however, it can produce hydroxyl radical, which is a potent reactive oxidant, in the presence of transition metals in their reduced states (Rowley and Halliwell, 1983). Hydrogen peroxide is removed primarily by glutathione peroxidases. We have shown that CA leads to depletion of glutathione, protein nitration, and lipid peroxidation, which were prevented by PNA (Walson et al, 2011). Polynitroxyl albumin has beneficial effects in models of focal ischemia—reperfusion (Steinbauer et al, 2000; Zhang et al, 2000) and focal brain ischemia (Beaulieu et al, 1998; Sugawara et al, 2001).

Albumin is a colloid that increases MAP in hypovolemia. A beneficial effect of albumin in neurologic pathologies is suggested by studies in stroke and traumatic brain injury in which lower serum albumin levels are associated with poor outcome (Bernard et al, 2008; Idicula et al, 2009). Albumin has beneficial effects in animal models of focal ischemia (Belayev et al, 1997, 2002; Liu et al, 2001); however, it was not found to be neuroprotective either in a neonatal model of stroke or after clinical traumatic brain injury in adults (Myburgh et al, 2007; Wang et al, 2007). Albumin's beneficial effects for focal ischemia—reperfusion are attributed to antioxidant properties, inhibition of copper-induced peroxidation, fatty acid oxidation (Gutteridge et al, 1984; Rowley et al, 1984; Yao et al, 2010), binding of fatty acids and heme, and early upregulation of vascular endothelial growth factor (Yao et al, 2010). Albumin contains one free cysteine moiety, which is believed to be the key scavenger of reactive oxygen and nitrogen species generated in the intravascular compartment (O’Neill et al, 1993). However, it has been shown that human serum albumin preparations commonly used in the clinic are predominantly oxidized at the free cysteine, which was resistant to reduction (Lang et al, 2004).

Postresuscitation CBF was divergently affected by administration of PNA and albumin versus NS in our model. The blood—brain barrier is intact to gadoteridol (a molecule smaller than albumin) in the first 3 hours after resuscitation in our model (Manole et al, 2009), and consequently, we expect that PNA and albumin exert their effects in the intravascular space. Similar to our findings with PNA, SOD administration reduced early postresuscitation hyperemia in an asphyxial CA model in dogs (Cerchiari et al, 1987), and administration of SOD-albumin complex improved blood flow and neurologic outcome in a global ischemia—reperfusion model (Takeda et al, 1993). Similar to our study, albumin was shown to improve cortical perfusion after permanent focal cerebral ischemia (Liu et al, 2001) and improved perfusion in the microcirculation during the postischemic reperfusion period (Belayev et al, 2002). One possible explanation for our findings is mitigation of superoxide-induced vascular dysfunction by PNA versus albumin. Superoxide causes microvascular abnormalities such as loss of autoregulation after ischemia, and treatment with scavengers of superoxide can prevent the loss of autoregulation (Nelson et al, 1992). Intravascular antioxidants such as PNA may mitigate the effect of superoxide, attenuate postischemic vascular injury, and prevent the loss of blood pressure autoregulation of CBF after CA. By preventing the loss of autoregulation, PNA may prevent hyperemia early after ROSC, whereas albumin, by increasing MAP and/or cardiac output in the absence of such vascular protection, further promotes cerebral hyperemia. MAP increases early postresuscitation in our rats, but would be expected to remain in the autoregulatory range (Pryds et al, 2005). This possibility is supported by data showing that PNA, but not albumin, inhibits the superoxide-induced adherence of polymorphonuclear cells to the endothelium (Russell et al, 1998). Another mechanism by which albumin produces hyperemia may be attributed to formation of S-nitroso-albumin. It was shown that albumin binds nitric oxide at the free cysteine and forms S-nitroso-albumin, a compound with endothelium-derived relaxing factor-like properties, vasodilatory effects on coronary vessels, and a half-live substantially longer than nitric oxide (Ignarro, 1989; Keaney et al, 1993). Additional mechanistic studies would be required to test these hypotheses.

We serially quantified CBF using baseline measurement for α and T1obs because repeated measurement of these parameters would preclude serial measurement of perfusion maps. We recognize that injection of nitroxides can reduce blood and tissue T₁ (Hyodo et al, 2008). To validate our approach, we serially measured these parameters after injection of PNA or albumin in separate rats. We noted a small decrease in T1obs with both agents. However, both PNA and albumin are large molecules that are restricted to the blood pool, and we attribute this effect to a change in the blood pool T₁. We believe that the baseline measurement provides a better representation of the tissue T₁ during the study. The PNA used in the study has a free tempol concentration of ∼1 mmol/L. A global effect of T₁ is seen in rapid T₁-weighted images; however, the free nitroxide is reduced quickly and does not affect serial CBF. This was confirmed by repeating experiments with tempol-free PNA in several rats. After albumin administration, α decreased slightly. This can be attributed to an additional magnetization transfer effect in the arterial blood as a result of the larger concentration of macromolecules. We did not correct for this magnetization transfer effect, but the result leads to an underestimate of CBF by ∼8% for the albumin-treated group. No difference in α is seen after the administration of PNA. The bound nitroxides efficiently relax the macromolecular nuclear spins blunting the additional magnetization transfer effect in the arterial blood.

Morris water maze performance was improved by both PNA and albumin after CA in male rats. Similarly, PNA or SOD confer neuroprotection after global ischemia, and thus, amelioration of vascular injury by PNA may improve outcome (Cerchiari et al, 1987; Kofler et al, 2005; Schleien et al, 1994). The improved functional outcome after albumin administration in our study was not surprising because albumin is also beneficial in animal models of focal ischemia: decreased infarct area, reduced brain edema, and improved functional outcome (Belayev et al, 1997, 1998). Albumin improved neurologic outcome in a pilot study in humans after stroke (Ginsberg et al, 2006) and a multicenter, randomized controlled trial of albumin in acute ischemic stroke is underway.

Neither PNA nor albumin attenuated CA₁ hippocampal neuronal death versus NS. Thus, the degree of hippocampal neuronal death cannot be directly linked to the degree of early postresuscitation CBF derangements in our model. Further studies are required before a definitive conclusion can be drawn regarding the influence of blunting or augmenting early hyperemia on neuronal death. We selected the CA₁ hippocampus given its vulnerability in this model and the capacity to quantify neuronal counts through stereology (Fink et al, 2005). Quantification of neuronal death in other anatomic areas after CA is warranted. Thalamic areas in particular have early hyperemia that intensified after treatment with albumin. Systematic characterization of thalamic pathology in different thalamic nuclei, along with quantification of gliosis and axonal injury in our model are important and the object of further study (Shoykhet et al, 2011).

Our data suggest that males treated with PNA or albumin have improved spatial memory retention after CA. We found no significant sex-dependent difference in postresuscitation CBF, gross vestibulomotor function, or neuronal loss. Our findings might relate to the evidence of a greater role for oxidative stress in males versus females (Du et al, 2004). At the time of testing, rats were 31-days old; thus, in addition to possible effects of innate gender differences at the time of injury, hormonal influences in males may contribute to the gender effects seen on functional outcome. However, because our study was not initially designed for sex-dependent analysis, the power to detect a sex difference was low and prospective evaluation of sex-dependent effect is warranted. Our work does support evaluation of sex-dependent treatment effects, even in prepubertal neuronal injury models.

Administration of PNA or albumin at resuscitation is mimicking the clinical scenarios of in-hospital CA or administration at resuscitation by first responders (emergency medical personnel) in an out-of-hospital CA. Assessment of delayed administration of PNA or albumin on neurologic outcome and the time window of administration of PNA and albumin to optimize cerebrovascular-targeted therapies is also valuable and will be evaluated in future studies.

Summary

Despite divergent effects on hyperemia early after CA in postnatal day 16 to 18 rats, both PNA and albumin modestly improved neurologic outcome in our model in a sex-dependent manner.