Systemic administration of dendrimer N‐acetyl cysteine improves outcomes and survival following cardiac arrest

Abstract Cardiac arrest (CA), the sudden cessation of effective cardiac pumping function, is still a major clinical problem with a high rate of early and long‐term mortality. Post‐cardiac arrest syndrome (PCAS) may be related to an early systemic inflammatory response leading to exaggerated and sustained neuroinflammation. Therefore, early intervention with targeted drug delivery to attenuate neuroinflammation may greatly improve therapeutic outcomes. Using a clinically relevant asphyxia CA model, we demonstrate that a single (i.p.) dose of dendrimer‐N‐acetylcysteine conjugate (D‐NAC), can target “activated” microglial cells following CA, leading to an improvement in post‐CA survival rate compared to saline (86% vs. 45%). D‐NAC treatment also significantly improved gross neurological score within 4 h of treatment (p < 0.05) and continued to show improvement at 48 h (p < 0.05). Specifically, there was a substantial impairment in motor responses after CA, which was subsequently improved with D‐NAC treatment (p < 0.05). D‐NAC also mitigated hippocampal cell density loss seen post‐CA in the CA1 and CA3 subregions (p < 0.001). These results demonstrate that early therapeutic intervention even with a single D‐NAC bolus results in a robust sustainable improvement in long‐term survival, short‐term motor deficits, and neurological recovery. Our current work lays the groundwork for a clinically relevant therapeutic approach to treating post‐CA syndrome.


| INTRODUCTION
Cardiac arrest (CA) accounts for nearly 500,000 deaths annually in the United States and Europe. [1][2][3][4][5] Survival in patients has continued to increase with improvements in resuscitation care but overall is still low with 50% of cardiopulmonary resuscitation (CPR) attempts restoring spontaneous circulation. 3,5,6 Despite these advances, outcomes remain bleak in those that survive CA, with irreversible neurologic injury being one of the biggest challenges. Approximately 55% of patients subsequently die after resuscitation, 7 most commonly related to severe neurological injury. 8 Neurological injury from CA is multifactorial starting with the primary injury caused by ischemia due to the immediate cessation of cerebral blood flow following CA. 9 CA-induced ischemia also secondarily triggers both excitotoxicity and neuroinflammation synergistically to produce microglial activation during the acute period and neuronal cell death. [10][11][12][13] An increase in inflammatory cytokines and oxidative stress markers is seen as soon as 4 h post-CA. 14 Thereafter, reperfusion initiates a secondary injury resulting in a nonspecific peripheral inflammatory response along with inflammation in the brain, perpetuating the microglial activation. This combination of both the primary insult and the secondary reperfusion injury can lead to sustained brain injury resulting in neurocognitive dysfunction and neurofunctional deficits. 9 Post-cardiac arrest syndrome (PCAS) is used to describe this unique and complex pathophysiological processes following resuscitation. 15 We have developed a rodent model of CA that recapitulates many of the key features of human CA and the period following return of spontaneous circulation (ROSC) including the neuroinflammatory cascade seen post-ROSC in patients. 14,[16][17][18] Currently, targeted temperature management by therapeutic hypothermia (TH) is the standard of care post-ROSC and has nonspecific effects on inflammation. 7 A prolonged and sustained elevation of cytokines such as IL-6 has been associated with worse outcomes and increased severity of PCAS, and TH has not been shown to consistently modify the systemic inflammatory response. 19,20 Moreover, rewarming after hypothermia has been associated with worsening neuroinflammation and injury in preclinical models, 21,22 and with impaired cerebral autoregulation in traumatic brain injury and cardiopulmonary bypass. 23,24 In adults, the rate of death decreases from 55% to 41% with 24 h of TH, however, 30%-40% of survivors continue to have poor neurological outcomes. 7 While TH is effective, there is a need to adjunct interventions to further improve both survival and neurologic deficits. Notably, elevation of the glial activation marker, 18 kDa translocator protein (TSPO) is observed from 5 to 180 days post-CA in a rat model of CA. 25 We hypothesize that more specific mitigation of neuroinflammation by targeting activated microglia will lead to both improvements in survival, neurologic function, and extent of neurologic injury.
Thus, we propose a glia-targeted approach to mitigate neuroinflammation for the reduction of mortality and neurological impairment post-CA. Hydroxyl-terminated poly(amidoamine) (PAMAM) dendrimers are emerging as potential drug delivery platform for neurological indications requiring targeted treatment of glia. [26][27][28] Dendrimers are taken up selectively by activated microglia and astrocytes in several models of brain injury. 26,27,29,30 Further, N-acetyl cysteine (NAC) conjugated to hydroxylterminated PAMAM dendrimers (D-NAC) has been shown to be effective in reducing neuroinflammation and oxidative injury in both large and small animal models at a fraction of the dose of free NAC needed to achieve a therapeutic effect. 27,31,32 A precursor to glutathione, NAC has both antioxidant and anti-inflammatory properties and is widely used clinically in children and adults. [33][34][35][36] NAC is a free radical scavenger as well as an NF-κB inhibitor. 37 However, NAC has clinical drawbacks as large, repeated doses are needed in order to achieve the bioavailability necessary for therapeutic effect and high doses of NAC have been shown to increase extracellular glutamate release and can result in neurotoxicity. 28,38 Here, we explore D-NAC as a treatment for PCAS. Rats resuscitated after asphyxial-CA were treated with D-NAC, or saline, and survival, neurological deficits, and hippocampal damage were evaluated. We demonstrate that a single dose of D-NAC after ROSC in a rodent CA model improves survival rate and neurological outcomes compared with saline group. Our results also indicate that D-NAC decreases mortality and improves both brain and behavioral outcomes.
2 | RESULTS 2.1 | Systemically administered dendrimer colocalizes with activated microglia following CA Experimental procedures for CA are as described in Figure 1(a). To demonstrate dendrimer uptake in the brain following CA, dendrimer (a) (a) Timeline for cardiac arrest procedure in rats optimized to consistently generate clinically relevant cardiac arrest phenotypes. (b) Sequential outcome measurements collected at different time intervals for comparative effectiveness of therapeutic interventions labeled with Cy5 fluorophore (D-Cy5) was injected intravenously after 30 min following CA. Rats were sacrificed 4 h after dendrimer administration and brains perfused and fixed for immunohistochemistry.
Fixed, cryoprotected and sectioned brains were colabeled with Iba1 and DAPI for microglia and nuclear stain, respectively. D-Cy5 was found to colocalize with activated microglia in the hippocampus, striatum and primary motor cortex at 4 h following CA. Based on the Pearson correlation and plot profile analysis, the D-Cy5 signal showed good colocalization with the microglia marker, Iba1, in the motor cortex, striatum, and hippocampus (Pearson correlation coefficient, Figure 2(d), (e), (f)). These results suggest that dendrimers are taken up in the brain post-CA and selectively localize in the activated microglia ( Figure 2).

| Systemic D-NAC improves survival rate
A survival analysis was conducted to evaluate D-NAC effect over a period of 2 weeks. Mortality after resuscitation from CA following initial ROSC was high at 45% (a percentage that is similar to what is seen clinically) with the highest rate of death occurring in the first 48 h after ROSC ( Figure 3). In the present study, we observed a significantly improved survival rate after 10 mg/kg of D-NAC treatment compared to saline-treated CA rats (86% survival vs. 45%, Figure 3). This suggests that D-NAC treatment significantly improves post-ROSC survival.

| D-NAC treatment improves neurologic deficit score
Neurological Deficit Scale (NDS) testing is designed to reflect a full range of behavioral repertoire from arousal to coma. 14,17,18,39 The effect of D-NAC treatment on NDS scores and subscores were analyzed ( Figure 4 . According to our previous study, an NDS score of 60 was utilized as a standard for separating favorable and unfavorable neurologic outcomes after resuscitation. 39 Based on this prior work, animals with scores above 60 (favorable outcomes) were classified as mild injury phenotype. As shown in Table 1, in CA + D-NAC group 79% animals were categorized as mild injury phenotype when measured at 24 h, which was significantly higher than the 40% in We also observed that the survival rate is markedly improved in rats receiving dendrimer-delivered NAC (D-NAC) (10 mg/kg on a NAC basis) after ROSC. We used an asphyxia time of 7 min to ensure moderate-to-severe injury following ROSC. 16  Neuroprotection was also observed within the hippocampus following D-NAC treatment. Significant histological changes were detected in the CA1 and CA3 subregions of the hippocampus of CA rats. Specifically, we detected decreased hippocampal cell density in CA rats compared to healthy controls and found that D-NAC treatment increased cell density compared to saline-treated CA rats. These findings align with recent data indicating a role for neuroinflammation and specifically glia in neuronal death post-ischemia. 51 This time delay has proven to be critical in animal models. We have previously shown that hypothermia immediately after CA in this model is more effective in rescuing the electrophysiological function (as measured by the qEEG technique) than conventional hypothermia initiated 6 h post-CA. 17 Thus, we conclude that time to deliver treatment is an important factor in rescuing the brain function, but that TH  14,[16][17][18]39,45 Our CA protocol has been described in detail previously. 14,[16][17][18]39,45 Briefly, rats were endotracheally intubated by direct laryngoscopy and mechanically ventilated with 2% isoflurane in 50% O 2 + 50% N 2 gas, after which the femoral artery and vein were cannulated with polyethylene 50 tubing catheters to monitor blood pressure and sample arterial blood gases and to administer intravenous medications. We carried out 10 min of baseline recording for mean arterial pressure (MAP) under isoflurane. Following baseline recording, anesthesia was washed out for a total 5 min, starting with 2 min of 100% O 2 without isoflurane to capture nonanesthetized MAP (Figure 1(a)). Then for 3 min, FiO 2 was decreased to 20% with 80% of N 2 gas (room air).
Rocuronium bromide 2 mg/kg IV was administered for muscle paralysis at 2 min of washout time. Following 5 min of gas washout, global asphyxia was induced by stopping and disconnecting the ventilator and clamping the tracheal tube. Global asphyxia was accompanied by transient hypertension, followed by progressive bradycardia, hypotension and eventual CA (defined by MAP<10 mmHg and nonpulsatile-pressure wave) are observed (Figure 1(a). 18

| Resuscitation
After the period of 7 min of asphyxia, CPR was initiated by unclamping the tracheal tube, restarting mechanical ventilation with 100% O2, administering epinephrine (5 μg/kg, i.v.), and applying sternal chest compressions with two fingers (~200 compressions/min), to generate MAP > 50 mmHg within 2 min was defined as a successful ROSC and NaHCO3 (1 mmol/kg, i.v.) to normalize the arterial pH. After undertaking the aforementioned measures ROSC were achieved the time was noted.
Anesthesia was not provided post-resuscitation to minimize drug effects on recorded electrical (EEG) signals and to facilitate the return of arousal.  8,9,12,38 (Figure 1(a)).
Rats were monitored and evaluated for 4 h after injection, and were subsequently euthanized, perfused transcardially with saline and then brains were fixed for cryosectioning and staining.

| Treatment with D-NAC
D-NAC was synthesized using generation 4 hydroxyl-terminated PAMAM dendrimers (Dendritech) as previously described. 66 We have previously demonstrated reproducible and scalable D-NAC synthesis that results in a payload of about 22-23 NAC molecules per dendrimer. 66 At 30 min post-ROSC, the rats were randomized into two groups to receive (1) saline treatment (CA + Saline control) at 0.2 mL, or (2) D-NAC in 0.2 mL saline (CA + D-NAC containing 10 mg/kg on a NAC basis) treatment intra-peritoneal (i.p.) (Figure 1(b)). The sham group (i.e., healthy control) animals were not subjected to CA or any other manipulations.

| Neurologic deficit scale score
The NDS score, which was previously developed and validated extensively, ranges from 0 to 80 and serves as a surrogate quantitative rodent neuro-deficit and coma score, analogous to human coma scores. 14,17,18,39 See supplementary information for details of NDS sub-scores components; Table S1). The NDS was determined at 4, 24, and 48 h post-ROSC ( Figure 4). For the differential between mild and severe injury, score 60 was used as a cut-off value following our previous publication. 39 Trained personnel who were blinded to the control and treatment groups conducted NDS examinations. 14

| Immunofluorescence
To evaluate the colocalization of D-Cy5 and microglia, 30 μm coronal brain sections were incubated overnight at 4 C with goat anti-IBA1

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.