Research Article: Invasion of Peripheral Immune Cells into Brain Parenchyma after Cardiac Arrest and Resuscitation

Date Published: June 1, 2018

Publisher: JKL International LLC

Author(s): Can Zhang, Nicole R. Brandon, Kerryann Koper, Pei Tang, Yan Xu, Huanyu Dou.


Although a direct link has long been suspected between systemic immune responses and neuronal injuries after stroke, it is unclear which immune cells play an important role. A question remains as to whether the blood brain barrier (BBB) is transiently disrupted after circulatory arrest to allow peripheral immune cells to enter brain parenchyma. Here, we developed a clinically relevant cardiac arrest and resuscitation model in mice to investigate the BBB integrity using noninvasive magnetic resonance imaging. Changes in immune signals in the brain and periphery were assayed by immunohistochemistry and flow cytometry. Quantitative variance maps from T1-weighted difference images before and after blood-pool contrast clearance revealed BBB disruptions immediately after resuscitation and one day after reperfusion. Time profiles of hippocampal CA1 neuronal injuries correlated with the morphological changes of microglia activation. Cytotoxic T cells, CD11b+CD11c+ dendritic cells, and CD11b+CD45+hi monocytes and macrophages were significantly increased in the brain three days after cardiac arrest and resuscitation, suggesting direct infiltration of these cells following the BBB disruption. Importantly, these immune cell changes were coupled with a parallel increase in the same subset of immune cell populations in the bone marrow and blood. We conclude that neurovascular breakdown during the initial reperfusion phase contributes to the systemic immune cell invasion and subsequent neuropathogenesis affecting the long-term outcome after cardiac arrest and resuscitation.

Partial Text

We present here a new cardiac arrest and resuscitation model in mice. Our model is different from the previous KCl-overdose models in that it is highly reproducible and clinically relevant, without the need for chest compressions at an unmanageable rate of 5 compressions per second continuously for up to 120 s [13]. The esmolol dose used in our model is comparable to that used in heart transplantations in humans. The resuscitation by oxygenated blood infusion mimics the clinical setup of cardiopulmonary support. We showed that this model can produce characteristic ischemic injuries to the selectively vulnerable neurons in the CA1 region of the hippocampus. The spatial injury pattern resembles what has been reported previously in rats [8, 9] and mice [13]. However, the temporal profile of injury severity in mice differs from that in rats. Specifically, when the injuries are measured by the unhealthy neurons in the hippocampal CA1 region using conventional H&E staining, an apparent histological recovery was observed on Day 10 as compared to Day 3 after cardiac arrest and resuscitation. This finding is rather unexpected and challenges the traditional viewpoint that neuronal injury is irreversible after global ischemia. Although some very minor artifacts from the automatic tissue stainer are visible in the H&E sections, the neuron counting by 5 investigators, who were blinded to tissue identities, all showed a decrease in the number of unhealthy pyramidal neurons in the hippocampal CA1 regions on Day 10 relative to Day 3. A similar recovery based on CA1 neuronal counting was also reported in a different cardiac arrest model using neonatal mice [13]. In that study, the recovery was attributed to the disappearance of dead neurons on the longer recovery time, leading to an apparent decrease in the percentage of damaged neurons among the remaining neurons observable by histology. In our studies using adult mice, the histology damage, when measured by the unhealthy neurons (as opposed to dead neurons), seems to be more severe on Day 3 than on Day 10. Among many possible interpretations, the simplest is that the unhealthy neurons on Day 3 might have, in part, recovered by Day 10. Strongly associated with the reduction of unhealthy neurons on Day 10 is the recovery of neuroinflammation as measured by the activation of astrocytes and microglia. The role of microglia in various models of neuronal injuries has been extensively studied [19]. In global ischemia, the morphological changes from ramified to amoeboid shape, as seen in Figures 2M-2P, indicates a transition from an ATP-dependent activation to a hypoxic or anaerobic energy metabolic state with concomitant switching from activated microglia to macrophage function [19]. Using a similar cardiac arrest and resuscitation model in rats and non-invasive MRI cerebral perfusion measurements, we showed previously that cerebral perfusion was severely reduced within hours after resuscitation, and this hypoperfusion was independent of cardiac output and could last up to 48 hours [9, 10]. The transient anoxic condition during circulatory arrest and the protracted hypoxic condition during the critical phase of reperfusion can trigger environment-dependent extracellular signaling through the production of inflammatory cytokines and the commitment of pro-inflammatory and anti-inflammatory polarization of microglia and macrophages. The parallel time dependence of neuronal recovery and the resolution of neuroinflammatory changes in reactive microglia and astrocytes from Day 3 to Day 10 (compare Fig. 1 and Fig. 2) implicate an inflammation-related neuronal injury and repair process.




Leave a Reply

Your email address will not be published.