Two-photon imaging of stroke onset in vivo reveals that NMDA-receptor independent ischemic depolarization is the major cause of rapid reversible damage to dendrites and spines

Abstract

We adapt a mouse global ischemia model to permit rapid induction of ischemia and reperfusion in conjunction with two-photon imaging to monitor the initial ionic, structural, and functional implications of brief interruptions of blood flow (6-8 min) in vivo. After only 2-3 min of global ischemia, a wide spread loss of mouse somatosensory cortex apical dendritic structure is initiated during the passage of a propagating wave (3.3 mm/min) of ischemic depolarization. Increases in intracellular calcium levels occurred during the wave of ischemic depolarization and were coincident with the loss of dendritic structure, but were not triggered by reperfusion. To assess the role of NMDA receptors, we locally applied the antagonist MK-801 [(+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine maleate] at concentrations sufficient to fully block local NMDA agonist-evoked changes in intracellular calcium levels in vivo. Changes in dendritic structure and intracellular calcium levels were independent of NMDA receptor activation. Local application of the non-NMDA glutamate receptor antagonist CNQX also failed to block ischemic depolarization or rapid changes in dendrite structure. Within 3-5 min of reperfusion, damage ceased and restoration of synaptic structure occurred over 10-60 min. In contrast to a reperfusion promoting damage, over this time scale, the majority of spines and dendrites regained their original structure during reperfusion. Intrinsic optical signal imaging of sensory evoked maps indicated that reversible alteration in dendritic structure during reperfusion was accompanied by restored functional maps. Our results identify glutamate receptor-independent ischemic depolarization as the major ionic event associated with disruption of synaptic structure during the first few minutes of ischemia in vivo.

Figure 1.

Common carotid artery occlusion reversibly blocks local cortical blood flow. A, Experimental schematic; 6–8 min of bilateral CCA occlusion was produced by tensioning sutures placed around each artery while imaging, reperfusion was induced for up to 2 h by removing the tension, and in some animals a second round of CCA occlusion was performed. B, Examples of two-photon imaging of blood flow using Texas Red-labeled plasma. Moving red blood cells create dark streaks across vessels as they are scanned. The angle of streaks in vessels that are oriented horizontally was used to determine velocity. Under control conditions two vessels are shown, the smaller of which is a capillary. Blood-flow velocity was calculated in millimeters per second, and negative values correspond to right-to-left flow. The inset shows a close up of streaking along the capillary used for blood flow velocity measurements. In the larger vessel, blood-flow direction has reversed and is now left to right, hence, positive values. The bottom image is a maximal intensity projection over 100 μm of cortex showing vessels that are deeper. C, Group data from animals (n = 7) showing average red blood cell velocity for small vessels (4.3 ± 0.2 μm diameter) before, during CCA occlusion, and at various time points after. A one-way ANOVA indicated a significant reduction in velocity during occlusion (*p < 0.05).

Figure 2.

Reversible loss of spiny dendritic structure during bilateral carotid artery occlusion. A, Maximal intensity projection images of layer V neuron tuft dendrites (maximal intensity projection image) before and 4 min after ischemia and various time points during reperfusion. The white-boxed area in the top left corresponds to a region in close up images shown in the close-up panel below. Bottom, Close-up views of top panels; white arrowheads show the location of dendritic spines present throughout the ischemia/reperfusion; larger gray arrows indicate transiently lost spines. B, Average percentage of blebbed dendrites during CCA occlusion and the indicated average times of reperfusion. All changes in blebbed dendrite percentages were significantly different from control except for 106 min of reperfusion (by one-way ANOVA, *p < 0.05; data from n = 16 different animals). C, Group data showing changes in spine number relative to control preischemia conditions in 10 different animals during occlusion (n = 711 spines), ∼5 min after reperfusion (n = 801 spines), ∼15 min after reperfusion (n = 310 spines), and ∼80 min reperfusion (n = 688 spines). A significant difference (*p < 0.05) was observed between control and occlusion, control and 5 min reperfusion, and control and 14 min reperfusion (by one-way ANOVA).

Figure 3.

Analysis of cortical function during reversible alteration in synaptic structure during ischemia. A, Composite video and two-photon images; background shows a low power view of the cortical surface vasculature produced by illuminating the cortex with green light. The inset shows a branched large vessel (area with white box around it) where two-photon imaging of both blood flow and dendrite structure were performed. The color composite image (red and green) on the right is of both dendrites (green) and vessels (red). Dendrites were found to undergo a reversible change in structure within this region (see E). B, Grayscale map derived from stimulating the contralateral hindlimb (HL); image represents the percentage change in 635 nm reflected light observed 0–1.5 s after limb stimulation. C, C', Color IOS response maps produced using a threshold based on 50% of the smoothed maximal response value. Areas responding to contralateral hindlimb stimulation are coded in green, whereas contralateral (FL)-responsive areas are coded in red. Before the ischemic onset, discrete hindlimb and territories can be observed. At an average time of 50 min after ischemia, both the hindlimb and maps have nearly recovered. An EEG recording electrode is also visible below the hindlimb area. D, Analysis of group data indicate that a significant contralateral limb stimulated response returns for trials taken on average 0.75–2 h after ischemia (n = 6 animals). No significant IOS responses were evoked for trials taken on average 15–45 min after ischemia with either contralateral limb stimulus (n = 4 animals). No significant response was observed when the contralateral forelimb was stimulated and the IOS response quantified in hindlimb territory, or vice versa, indicating specificity. Responses were normalized to control (prestroke) contralateral forelimb and hindlimb responses in their respective territories. (*p < 0.05, one-way ANOVA). E, Close-up view of dendrites and vessels from the area outlined in panel A at the border of the hindlimb area showing intact dendritic and vessel structure during the first 2 min of ischemia (before ischemic depolarization). The next set of images was taken after ischemic depolarization (5 min after occlusion) and show extensive dendritic damage. The right-most panel shows structural recovery 1 h after reperfusion.

Figure 4.

Changes in dendritic morphology occur synchronously over a period of seconds during global ischemia. A, Images shown are sequential 1 μm sections of dendrites taken during z-sectioning through layer I. The sequence shown begins 92.5 s after occlusion of both CCAs and was taken before ischemic depolarization (C). Because of relatively poor z-resolution, oversampling results and some processes are viewed in multiple images taken 5.5 s apart. B, Same region as in A; a change in dendritic morphology occurs at 217.5 s, just after the onset of a DC potential shift observed at ∼200 s after occlusion (C). Because the EEG electrodes sample a large area and the depolarization moves slowly across the brain (∼3 mm/min), the EEG signal can occur before a local change in morphology is observed. C, Record of DC cortical surface potential from the same animal as imaged in A and B. Left, A DC shift was observed ∼200 s after bilateral CCA occlusion. Right, AC filtered EEG record (0.1–10 Hz bandpass), indicating rapid suppression of the EEG during ischemia and return of spontaneous activity over a 20–30 min period.

Figure 5.

Changes in intracellular calcium in vivo during transient ischemia coincide with slow EEG activity and dendritic beading. A, Change in normalized OGB-1 fluorescence taken from the neuropil in layer I cortex of an animal subjected to bilateral CCA ligation. To control for changes in fluorescence intensity with depth, the OGB-1 fluorescence during ischemia was divided by that obtained just before the onset of ischemia using identical settings and laser power. On the same time scale, we plot the cortical DC EEG indicating that the change in OGB-1 fluorescence was coincident with ischemic depolarization. The ischemic episodes were within a single animal taken ∼2 h apart; both episodes showed an OGB-1 fluorescence increase and ischemic depolarization after CCA occlusion. B, Composite image showing GFP-labeled dendrites surrounded by diffuse OGB-1-labeled layer I neuropil from the same animal as the one shown in A. Maximal intensity projections were made over 4 μm. Note that no OGB-1-labeled glial or neuronal cell bodies were present here. At 206 s after ischemic onset and after ischemic depolarization (estimated to have occurred at 117 s), the layer I neuropil shows a large increase in OGB-1 fluorescence indicating a sudden rise in [Ca²⁺]_i. At this time, the dendrites have become severely blebbed, seen over a uniform change in background neuropil OGB-1 fluorescence. Note that because z-stacks were only taken every 2 min, 206 s was the earliest time point after ischemic depolarization with images of these laterally projecting dendrites. The images are of lower contrast because OGB-1 labels the background neuropil. Two minutes after reperfusion (Reper; 514 s after occlusion), neuropil OGB-1 fluorescence levels are already beginning to decline and 13 min after ischemia the dendritic structure is nearly fully restored (738 s). C, Composite image showing GFP-labeled dendrites surrounded by diffuse OGB-1-labeled layer I neuropil from the same animal as the one shown in A and B after the second period of ischemia. Sequential 1 μm sections through the tissue are shown as in Figure 4A. At 153 s after CCA occlusion and after ischemic depolarization (occurs at ∼142 s), the OGB-1 fluorescence markedly increased in the layer I neuropil and the dendrites became suddenly blebbed indicating near coincidence between Ca²⁺ elevation and dendritic structural changes.

Figure 6.

In vivo confirmation that MK-801 blocks layer I NMDA receptors. A, Experimental schematic for in vivo confirmation of MK-801 effects. The OGB-1 fluorescence responses to NMDA application were tested before and after locally applying MK-801 (300 μm preincubation and 30 μm in agarose) to the cortex. B, Group data showing that OGB-1 fluorescence changes (df/F_o) to NMDA application before and after MK-801 incubation. NMDA was applied by local pressure ejection (4 psi) for ∼2 s after 6 s from the start of imaging. There was a significant reduction in NMDA-evoked OGB-1 fluorescence responses after MK-801 treatment (p < 0.05; n = 4 animals). The data show that this concentration of MK-801 applied locally to cortex can block layer I NMDA receptors.

Figure 7.

No effect of NMDA receptor blockade on ischemic depolarization, changes in [Ca²⁺], or dendritic damage. A, Maximal intensity projection images (10 μm) from an NMDAR-antagonist (MK-801, applied locally) treated mouse showing GFP-labeled spiny dendrites and OGB-1 fluorescence in layer I. The left image was taken before bilateral CCA occlusion. After occlusion (center image), an increase in neuropil fluorescence and disruption of dendritic structure was observed. After reperfusion (right image), a reduction of neuropil OGB-1 fluorescence and recovery of dendritic structure was observed. B, C, Quantification of neuropil OGB-1 fluorescence from the MK-801-treated animal in A over short (B) and long (C) time scales shows a rapid increase in neuropil fluorescence 140 s after occlusion of the CCAs. The ratio values in B were calculated for individual frames of a stack for neuropil regions of interest and divided by values for the same frames taken before ischemia (as in Fig. 5B). The graph in C (slower time scale) was produced by averaging all neuropil fluorescence over each projected stack and dividing by a corresponding preischemia value. D, E, Cortical DC potential recording from the same MK-801-treated animal on short (D) and long (E) time scales indicate that an ischemic depolarization occurs after occlusion of the CCAs. Recovery of the cortical DC potential was observed after reperfusion at 400 s. Because of handling the animal during occlusion and reperfusion, sharp artifactual transients (near 0 and 300 s) and shifts in baseline are observed during reperfusion. F, Group data on percentage of blebbed dendrites (n = 5) in MK-801-treated animals. A significant difference (*p < 0.05 by one-way ANOVA) was observed between control and occlusion, and control and 2 min reperfusion. G, Group data showing changes in spine number relative to control preischemia condition in 5 MK-801-treated animals during occlusion (n = 625 spines), ∼3 min after reperfusion (n = 620 spines), ∼15 min after reperfusion (n = 461 spines), and ∼80 min after reperfusion (n = 469 spines). A significant difference (*p < 0.05 by one-way ANOVA) was observed between control and occlusion, and control and 3 min reperfusion.

Figure 8.

Antagonism of AMPA/kainate receptor activity fails to block ischemic depolarization or changes in morphology of dendrites. A, Ischemic depolarization assayed by a surface electrode occurs despite inclusion of 1 mm CNQX in the overlying agarose. B, Reversible suppression of EEG activity in a CNQX-treated animal during bilateral CCA occlusion (DC record filtered at 0.1–10 Hz bandpass). C, Dendritic morphology preischemia and after 6 min of bilateral CCA occlusion (D). E, Group data on the percentage of blebbed dendrites preischemia and at the indicated times of reperfusion (data from n = 4 animals; *p < 0.05 one-way ANOVA). F, Intrinsic optical signal map; for simplicity we have divided the hindlimb (HL) map by the forelimb (FL) map so activation in the hindlimb is coded by darker gray, whereas hindlimb activation is coded by lighter tones. The maps are the average of at least 30 stimulus trials. G, Application of CNQX completely blocked the maps in this and a total of three separate mice.

Figure 9.

Waves of light scattering from the cortical surface signal ischemic depolarization during bilateral common carotid artery ligation. A, The brain was illuminated with light from 635 nm LEDs and ratio images (images were divided by a preischemic baseline image) were created depicting changes in reflected light signal −34, 73, and 83 s relative to ischemic onset. At the −34 s time point, a uniformly gray image is shown before ischemia indicating little regional difference in light scattering. The brightening of the image on the top right corner is a wave of increased light scattering that propagates across the cortex (from anterior lateral regions to medial posterior) during ischemic depolarization. B, Quantification of changes in light scattering in the animal shown in A (made from an area of interest in the image center). The disruption in signal just before the start of ischemia is from handling the animal (blocking LEDs, −30–0 s). After carotid ligation and over the first 75 s, there is a gradual darkening of the cortex and a reduction in reflected light signal. At 75 s, the signal changes from a darkening to brightening and a wave of increased reflection spreads across the brain (upward deflection). In the bottom graph, the cortical EEG measured at the same time is plotted and indicates rapid EEG suppression during CCA ligation followed by a slow AC EEG ripple (AC-filtered version of DC shift) that is coincident with the wave of light scattering.

Figure 10.

Global ischemia structural and ionic mechanisms (in vivo timeline). Summary timeline for the first 2 h after 6–8 min of global ischemia followed by reperfusion in layer I cortex.