Imaging and preventing spreading depression independent of cerebral blood flow

Abstract

A moving cortical inactivation termed spreading depression (SD) is considered the physiological event responsible for migraine aura. SD is a mass depolarization of neurons and glia lasting a minute or more. It arises focally and migrates as a wave across gray matter at 2–5 mm/min. SD is generated by a sudden increase in cell membrane permeability to small ions in neurons and glia. This neurogenic origin does not preclude SD being initiated by local vascular change. However, the brain slice preparation does permit the study of SD independent of changes to blood flow or the microvasculature. When arising under near-normoxic conditions, the SD of aura dissipates within 30 min and causes no neuronal damage. In contrast, during the first 3–4 h following stroke, the combined metabolic stress of recurring SD and energy deprivation exacerbates ischemic injury to neurons in the penumbra. Purpose: To image and pharmacologically block SD under normoxic and stroke-like conditions in slices of rodent cerebral hemisphere where blood flow is not a modulating factor. Methods: Coronal brain slices are submerged in flowing artificial cerebrospinal fluid (aCSF). Normoxic SD is induced by briefly raising [K +] o. An ischemic version of SD, the anoxic depolarization (AD), is induced by removing O 2/glucose from the aCSF for 10 min. Intrinsic optical signals (IOSs) represent change in the way tissue scatters light. Light transmittance (LT), essentially unscattered light, is imaged using a charge-coupled device (CCD) to measure second-by-second regional ΔLT during periods of 1 h or more. The front of the propagating SD or AD event is imaged as an elevated LT, caused by cell swelling. A negative voltage shift, the electrophysiological signature of SD, is simultaneously recorded extracellularly. In the wake of the AD, the LT increase is overridden by a reduction in LT. This is caused by dendritic beading which efficiently scatters light, thereby revealing neuronal damage not seen following classic SD. Results: Normoxic SD evoked by elevated bath K + arises multifocally in neocortical layer II/III and migrates as a wave at 2–5 mm/min. It can be elicited repeatedly with no dendritic beading or loss of evoked electrical activity. The AD also initiates multifocally in neocortical gray (and 1–3 min later, in striatum or hippocampus); in its wake, dendritic beading develops and evoked electrical activity is permanently lost. Potentially useful therapeutics can be assessed in terms of whether they block the SD or reduce damage following AD. Glutamate receptor antagonists do not block the AD or resultant damage. However, sigma 1 receptor (σ 1R) ligands such as 10–100 μM dextromethorphan, carbetapentane, or 4-IBP block the AD and resultant damage. Normoxic SD is also blocked. σ 1R antagonists oppose AD or SD blockade by these ligands. Conclusions: Recurring SD is innocuous in normoxic slices. However, the metabolic stress of O 2/glucose deprivation (OGD) for 10 min and a single ischemic SD event (the AD) induces acute neuronal damage in slices. We propose that damage in the ischemic core is difficult to prevent because the AD arises upon stroke onset and normally resists pharmacological blockade. By dissociating the mass depolarization of AD from the ischemia, σ 1R ligands offer a possible way to reduce early stroke damage.