Hypoxic–ischemic stroke in adult C57BL/6J mice results in a variable stroke size
We observed a wide variety of stroke sizes at 24 h, with three typical types of stroke. Some mice exhibited ischemia in the vast majority of the hemisphere (a, left), some had an intermediate-sized stroke with dense ischemia in the cortex and hippocampus and more diffuse injury in the striatum (a, middle), and others had either no visible ischemia or a small cortical stroke (a, right). Of 13 mice that underwent hypoxic–ischemic stroke, the five with the smallest strokes averaged a stroke size of 11.4 ± 2.4% of the contralateral hemisphere. Five mice had large strokes that measured 38.2 ± 2.7% of the hemisphere, and the remaining three mice had very large strokes measuring 56.4 ± 4.4% of the hemisphere. These three categories were all statistically different in size (small vs. large, P < 0.0001; large vs. very large, P < 0.001; ANOVA with Tukey's post hoc.)
Figure 2 Hypoxic–ischemic stroke results in variable stroke size. (a) Typical TTC stains from 24 h after hypoxic–ischemic stroke. Left, large and likely fatal stroke; middle, survivable large stroke; right, small stroke. (b) Stroke size in surviving (more ...)
In a larger cohort that we followed for 6 weeks, all-cause mortality was approximately one-third, and the majority of this was during hypoxia or during the first 3 days after stroke. In the surviving 28 mice, there was no difference in the mean or distribution of stroke sizes between surgeons (b). A dichotomization in stroke size was evident in the surviving mice, with one group of mice exhibiting loss of about 50% of the hemisphere and the other exhibiting much smaller stroke sizes. Although we did not autopsy mouse brains, the largest stroke size that we observed at 24 h in the smaller cohort (a, left) did not appear in the surviving mice and so was likely fatal.
Horizontal ladder test performance on day 1 predicts stroke size at 6 weeks
Given the dichotomization of stroke sizes in this model, we hypothesized that the smaller stroke sizes would result in either a quickly recovering deficit or no deficit and would introduce increased behavioral variability. In order to be able to study long-term functional recovery, our goal was to identify the subset of mice with survivable larger strokes during the first week after stroke. We examined the linear correlation between stroke size and performance in the Stroke group on the horizontal ladder test (a) on day 1 after stroke, rotarod on day 2, EBST on day 4, and from automated gait analysis, stride length and swing speed on day 5. Of these measures, only EBST and ladder correlated significantly among the stroked mice. Horizontal ladder performance on day 1 correlated highly with stroke size (P < 0.0001, R2 = 0.7652; b). This was reproducible in a second cohort of mice (P < 0.0001, R2 = 0.7551; c).
Figure 3 Mouse performance on the horizontal ladder test 1 day after stroke correlates with stroke size at 6 weeks after stroke. (a) Single frame shot from a video of a mouse traversing the horizontal ladder. The arrow identifies a left front paw error. (b and (more ...)
EBST on day 4 also correlated with stroke size, but not as tightly as horizontal ladder testing (P = 0.0061, R2 = 0.4785). Rotarod on day 2 correlated significantly only when the sham mice were added to the correlation (P = 0.0352, R2 = 0.2237).
We also examined interrater reliability on horizontal ladder test scoring. Two blinded raters (KD and LM) examined videos from 32 mice that were tested on day 1 after hypoxic–ischemic stroke. Interrater reliability was excellent, with Spearman's coefficient 0.873 (P = 7.5 × 10−11).
Based on the linear correlation between stroke size and day 1 horizontal ladder performance, we chose a cutoff of >18% error with the left front foot to assign mice to a “Large Stroke” group (b and c, gray box). In comparison to all stroked mice (“All Stroke”), this resulted in groupings of mice where the remaining right hemisphere volume, expressed as a percent of left hemisphere volume, was 52.3 ± 3.3%, n
= 6 in “Large Stroke”; 77.6 ± 6.6%, n
= 14 in “All Stroke”; and 103.6 ± 1.8%, n
= 6 in “Sham” (d). The “Large Stroke” group had less variability and also was more significantly different from sham mice than the “All Stroke” group. Left hemisphere size was not different in stroked mice than in sham mice (data not shown), supporting others' data that the hypoxic–ischemic stroke model does not cause significant ischemic damage to the contralateral hemisphere in C57BL/6J mice (Kuan et al. 2003
; Olson et al. 2004
; Adhami et al. 2006
Stroke-induced spontaneous gait and gait accuracy deficits demonstrated recovery over the course of the study
We evaluated two measures of spontaneous gait after hypoxic–ischemic stroke. The first was horizontal ladder test performance, which measures limb placement errors on a ladder and which we used for identification of the “Large Stroke” group (). The second measure was automated gait analysis using a Catwalk (Noldus) apparatus. In both cases the mouse is walking toward its home cage at a normal speed.
Horizontal ladder testing was done on days 1, 4, and 7 and then weekly until day 35 after stroke (a). We examined foot faults with all limbs and found that the front limb contralateral to the stroke, the left front, was the most reliable to measure. Right front foot faults and bilateral hind limb faults did not change after stroke. Ladder performance in the “All Stroke” group did worsen after stroke, but was only significant on days 4 and 21 (a). The “Large Stroke” group displayed significantly worse function on all days except 14 and 35.
Figure 4 Gait measures demonstrate stroke-induced deficits that recover during the weeks after hypoxic–ischemic stroke. (a) Horizontal ladder testing and statistics. Left front swing speed (b) and stride length (c) from automated gait analysis. Bars, SEM; (more ...)
Automated gait analysis yielded many measures, most of which demonstrated some changes after stroke. No measure was different between groups before surgery. We chose to focus on stride length and swing speed because they displayed statistically significant changes after stroke, and both measures are relevant to clinical functional deficits. Segregation of mice into “Large Stroke” versus “All Stroke” groups was not necessary to see differences on day 12 – both groups were significantly different from Sham on day 12 in both measures (b and c). Swing speed was also impaired on day 26 in the “Large Stroke” group, as was stride length on days 26 and 33.
Rotarod reveals poststroke deficits that do not recover after 1 month
We next evaluated function on the rotarod, which tests how long a mouse can remain on a rotating rod. In this study we trained mice extensively and only included mice that had learned the task before surgery (latency to fall >250 sec), so we were testing motor recovery and not motor learning. No significant differences were detectable among groups before surgery. We observed a nonstatistically significant decrease in rotarod performance in the “All Stroke” group, but segregating out the “Large Stroke” group yielded significance for all days (a). Mice in the “Large Stroke” group did not demonstrate significant recovery of rotarod ability over the course of the study.
Figure 5 Rotarod and EBST testing deficits persist after 1 month. (a) Rotarod testing demonstrated clear deficits after stroke in the Large Stroke group, but nonsignificant deficits in the entire group. (b) Baseline EBST results reflected side preferences in mice. (more ...)
EBST testing was variable, but revealed poststroke deficits out to 5 weeks
The EBST is a measure of postural asymmetry that measures the direction animals turn toward when they are held by the tail. Interestingly, many mice exhibited a side preference on baseline testing, with the average mouse preferring to twist to the right, but all types were seen (b). No significant differences in side preference were detectable between surgical groups at baseline. After surgery, the “Large Stroke” group demonstrated a clear effect of stroke by preferring to swing to the contralateral side (c), while large variability in the “Sham” group limits the usefulness of this test. Subtracting each mouse's baseline preference did not alter the results in terms of trends or statistical significance and did decrease the variability in the shams while increasing the variability in the stroked mice (data not shown).
There were no stroke-induced changes in spontaneous activity
To assess spontaneous activity, mice were evaluated in an activity chamber before and 8 and 22 days after stroke or sham surgery. Neither “All Stroke” nor “Large Stroke” groups exhibited differences from “Sham” mice in total distance traveled or number of vertical rears (a and b). The apparatus also recorded revolutions, or which way the mice turned as they explored the chamber. Despite the asymmetry observed in the Large Stroke group on EBST, there was no difference between groups in the number or direction of spontaneous revolutions (c and d). Finally, mice in each group spent equal proportions of their time in the periphery compared with the center of the chamber, implying that stroke did not affect anxiety levels. At baseline (day −4), the percent of time spent in the periphery of the chamber was Sham 54.9 ± 4.8% versus Large Stroke 65.4 ± 5.7%; on day 8, Sham 65.1 ± 4.6% versus Large Stroke 56.4 ± 5.8%; and on day 22, Sham 56.9 ± 6.0% versus Large Stroke 60.1 ± 5.3%.
Figure 6 Activity chamber demonstrated no significant stroke-induced deficits. There were no differences between groups in (a) total distance traveled, (b) vertical rears, (c) total revolutions, or (d) direction of revolutions, as shown here by % counterclockwise (more ...)