Correspondence between Altered Functional and Structural Connectivity in the Contralesional Sensorimotor Cortex after Unilateral Stroke in Rats: A Combined Resting-State Functional MRI and Manganese-Enhanced MRI Study

Stroke Model

All animal procedures were approved by the Animal Experiments Committee of the University Medical Center Utrecht and Utrecht University.

A total of 18 male Sprague Dawley rats, weighing 280 to 320 g, were included. Eleven rats underwent transient unilateral focal cerebral ischemia by 90 minutes occlusion of the right middle cerebral artery with an intraluminal filament, as described previously (van Meer et al, 2010). In brief, a 4.0 polypropylene suture (Ethicon, Piscataway, NJ, USA) with a silicon-coated tip was introduced into the right external carotid artery and advanced through the internal carotid artery until a slight resistance was felt, indicating that the middle cerebral artery was occluded. After 90minutes, the filament was withdrawn from the internal carotid artery to allow reperfusion.

Seven healthy rats served as controls.

Magnetic Resonance Imaging

Magnetic resonance imaging was performed at 70 days after stroke, when the recovery of sensorimotor function had reached a plateau (van Meer et al, 2010; van der Zijden et al, 2008). During MRI, rats were anesthetized and mechanically ventilated with 1 to 2% isoflurane in air/O₂ (2:1). End-tidal CO₂ levels were kept within normal range by adjusting ventilation volume and/or rate. Magnetic resonance imaging measurements were conducted on a 4.7-T horizontal bore MR system (Varian, Palo Alto, CA, USA). First, multiecho multislice T₂-weighted MRI (repetition time (TR)/TE (echo time) = 3600/15 milliseconds; echo train length = 12; voxel resolution =0.25 × 0.25 × 1.0 mm³) and gradient echo three-dimensional MRI (TR/TE = 6/2.58 milliseconds; flip angle = 40°; voxel resolution = 0.23 × 0.31 × 0.31 mm³) were conducted for assessment of ischemic lesion size and location, and for registration purposes, respectively. Thereafter, for at least 10 minutes, end-tidal isoflurane was reduced to 1%, followed by rs-fMRI with a gradient echo, echo planar imaging sequence (TR/TE = 500/19 milliseconds; flip angle = 35°; voxel resolution = 0.5 × 0.5 × 1.5 mm³; 1200 BOLD images). Subsequently, T₁-weighted MRI (Look–Locker gradient echo; TR/TE = 5000/3.4milliseconds; inversion time = 10milliseconds; image TR = 24 × 8 milliseconds; flip angle = 10°; voxel resolution = 0.5 × 0.5 × 0.5 mm³) was conducted for baseline measurement of premanganese tissue R₁.

Two days after the initial MRI experiments, 90 nL of 0.1 mol/L MnCl²was injected into the left (contralesional) primary motor cortex (M1) (Canals et al, 2008). Postmanganese T₁-weighted and gradient echo three-dimensional MRI were conducted 1 day after MnCl₂ injection.

Experimental Groups

Stroke group assignments were based on the size and location of ischemic lesions on T₂-weighted images at 70 days after stroke. One animal died beforehand because of stroke-induced cachexia. Animals with mostly subcortical tissue damage and intact primary somatosensory and motor cortices were assigned to group S_I (n = 5), whereas animals with a lesion involving both subcortical and primary somatosensory cortical tissues were assigned to group S_II (n = 5). One animal from group S_I and one animal from group S_II were excluded from further analysis because of unsuccessful MnCl₂ injection. Age-matched healthy rats were assigned to control group C (n = 7).

Data Analysis and Statistics

Gradient echo three-dimensional images were registered nonrigidly to a reference image matched to a three-dimensional model of a rat brain atlas (Paxinos and Watson, 2005). Bilateral sensorimotor cortical regions of interest (ROIs), i.e., the primary and secondary motor cortices (M1, M2), forelimb region of the primary somatosensory cortex (S1fl) and secondary somatosensory cortex (S2), were projected from the atlas onto the rs-fMRI time series and R₁ maps.

Preprocessing of rs-fMRI data included motion correction, spatial smoothing, and linear regression, as described previously (van Meer et al, 2010). Subsequently, low-frequency BOLD fluctuations were extracted by applying a band-pass filter with 0.01 < f < 0.1 Hz. Functional connectivity was calculated as the Fisher-transformed correlation coefficient z’ (ln((1 + r)/(1 – r))/2) between the mean low-frequency BOLD time-series signal in the left M1 (seed region; 22mm³) and other cortical sensorimotor ROIs. Group mean whole-brain functional connectivity maps were obtained by voxel-wise calculation of z’ with the mean time-series signal from the left M1 as reference, averaged across subjects.

Manganese enhancement after MnCl₂ injection in the left M1 was measured in all ROIs and calculated as the difference in premanganese and postmanganese R₁ (ΔR₁) (van der Zijden et al, 2008). Group mean ΔR₁ maps were obtained by averaging across subjects.

Two-way repeated measures ANOVA (analysis of variance) with factors group and ROI was used to statistically compare intrahemispheric and interhemispheric functional connectivities (z′), as well as ipsilateral and contralateral ΔR₁s between groups, followed by post hoc Bonferroni's testing. Pearson's correlation analysis was used to test for correlation between functional (z0) and neuroanatomical connectivity (ΔR₁) of the left M1 with other sensorimotor cortical ROIs. Values are shown as mean ± s.d. P < 0.05 was considered significant.

Functional Connectivity

Figure 1A shows averaged brain T₂ maps for the different groups. Prolonged T₂, indicative of ischemic tissue damage, is evident in mainly subcortical regions in S_I rats, and in both subcortical and primary somatosensory cortical areas in S_II rats. Figure 1B shows averaged maps of resting-state functional connectivity with the left (contralesional) M1. In control rats, intrahemispheric and interhemispheric connectivity between M1 and ipsilateral and contralateral sensorimotor cortices was clearly apparent. At 10 weeks after stroke, functional connectivity of contralesional M1 was noticeably altered depending on the lesion extent in the opposite hemisphere. A significant increase in intrahemispheric functional connectivity of contralesional sensorimotor cortical ROIs with contralesional M1 was found in SII rats as compared with controls and SI animals (Figure 2A). Conversely, in the same group, a significant reduction in interhemispheric functional connectivity was detected between contralesional M1 and ipsilesional sensorimotor cortical ROIs as compared with groups C and S_I (Figure 2A).

OPEN IN VIEWER

Figure 1

(A) Color-coded maps of thresholded mean T₂, (B) z′ (with the left (contralesional) M1 as seed region), and (C) ΔR₁ (1 day after MnCl₂ injection into the left (contralesional) M1), for groups C, S_I, and S_II, overlaid on a T₂-weighted MRI template of consecutive coronal rat brain slices. The left ROIs M1 (primary motor cortex) (white), M2 (secondary motor cortex) (green), S1fl (forelimb region of the primary somatosensory cortex) (blue), and S2 (secondary somatosensory cortex) (purple) are delineated on the rat brain template. MRI, magnetic resonance imaging; ROI, region of interest.

OPEN IN VIEWER

Figure 2

(A) Intrahemispheric and interhemispheric z′ of the left (contralesional) M1 with left (contralesional) M2, S1fl, and S2, and right (ipsilesional) M1, M2, S1fl, and S2, respectively, for groups C, S_I, and S_II. (B) Intrahemispheric and interhemispheric manganese-induced δR₁ (per second) in left (contralesional) M2, S1fl, and S2, and right (ipsilesional) M1, M2, S1fl, and S2, respectively, at 1 day after MnCl₂ injection in the left (contralesional) M1, for groups C, S_I, and S_II. There was a significant group effect (F(2) = 4.45; P = 0.036) of increased intrahemispheric z′ in contralesional ROIs for group S_II as compared with groups C and S_I. Besides, a significant group effect (F(2) = 10.53; P = 0.002) of decreased interhemispheric z′ in ipsilesional ROIs was also found for group S_II as compared with groups C and S_I. A significant group x ROI effect (F(2) = 5.02; P = 0.004) of increased ΔR₁ in contralesional ROIs was found for groups S_I and S_II. A significant group x ROI effect (F(2) = 5.23; P < 0.001) of decreased ΔR₁ in ipsilesional ROIs was found for group S_II. *P < 0.05 versus equivalent ROI in group C. ROI, region of interest.

Structural Connectivity

Manganese-enhanced MRI showed clear manganese accumulation in the ipsilateral sensorimotor cortex, caudate putamen, and thalamus in all groups at 1 day after injection in the left (contralesional) M1 (Figure 1C). There were no significant group differences in total brain uptake of manganese, as measured from mean ΔR₁ in the whole rat brain tissue (ΔR₁ = 0.038 ± 0.020 per second, ΔR₁ = 0.032 ± 0.021 per second, and ΔR₁ = 0.040 ± 0.008 per second, for groups C, S_I, and S_II, respectively). At 10 weeks after stroke, manganese enhancement was increased in the contralesional sensorimotor cortex. A group x ROI interaction effect of significantly elevated manganese build-up in contralesional sensorimotor cortical ROIs was found for S_I and S_II rats (Figure 2B). Increased manganese enhancement was particularly observed in contralesional M2 of S_II animals.

Manganese accumulation was relatively low in the right (ipsilesional) hemisphere, contralateral to the injection site. Still, S_II rats showed a significant interaction effect of reduced manganese build-up in ipsilesional sensorimotor cortical ROIs (Figure 2B).

Correlation Between Functional and Structural Connectivity

Functional connectivity (z′) and manganese build-up (ΔR₁) from the left (contralesional) M1 to intrahemispheric cortical sensorimotor ROIs (M2, S1fl, and S2) correlated significantly for all groups together (r = 0.31, P = 0.02). We found a trend for a correlation between z′ and ΔR₁ in right (ipsilesional) cortical sensorimotor ROIs (M1, M2, S1fl, and S2) for the pooled groups (r = 0.18, P = 0.08).

Intrahemispheric Connectivity

In line with our previous study (van Meer et al, 2010), we found a significant increase in intrahemispheric functional connectivity within the contralesional sensorimotor cortex in rats with a relatively large unilateral infarction in the opposite (ipsilesional) sensorimotor network. In addition, we detected increased accumulation of the neuronal tracer manganese within the contralesional sensorimotor cortex of rats with large unilateral ischemic lesions, which suggests that the increase in functional connectivity is related to stronger local neuroanatomical association. The ipsilateral concentration of manganese after injection in the contralesional sensorimotor cortex may be explained by enhanced local neuronal connectivity and diminished axonal transport to remote (contralateral) regions (see below). An apparent regional linkage between functional and structural connectivity is supported by the observed significant correlation between synchronized low-frequency BOLD signals (z′) and manganese-induced ΔR₁ in intrahemispheric sensorimotor regions when data from all experimental groups were combined.

The poststroke increase in neuroanatomical and functional connectivity in the contralesional cortex may be brought about by remodeling of neuronal elements, i.e., axonal sprouting, synaptogenesis, and dendritic growth, which has been detected in various animal stroke models (reviewed in Nudo, 1999; Dancause, 2006). Similarly, in chronic human stroke patients, elevated activation responses to sensory stimulation of the unaffected hand have been found in an area of the contralesional somatosensory cortex with increased cortical thickness (Schaechter et al, 2006). Although not the subject of this study, we speculate that reorganization of the contralesional cortical tissue may be induced by increased compensatory use of the unaffected limbs and/or altered input from the affected ipsilesional cortex.

Interhemispheric Connectivity

In agreement with our previous study (van Meer et al, 2010), interhemispheric functional connectivity between contralesional M1 and ipsilesional sensorimotor cortical ROIs in rats with large subcortical and cortical infarctions was significantly decreased as compared with controls and rats with mainly subcortical tissue injury. This reflects desynchronization of signaling between the bilaterally homologous sensorimotor fields, which may be a direct result of extensive damage to the ipsilesional sensorimotor cortex. The observed negative z′ values in S_II animals even suggest anticorrelations, but more likely point toward minimal correlation between interhemispheric signals, as discussed in our previous paper (van Meer et al, 2010). In line with the reduction in interhemispheric functional connectivity, we detected a significant decline in transcallosal manganese transport toward ipsilesional sensorimotor ROIs in SII rats. This seems to be in contrast with an earlier study in which we found an increase in transhemispheric manganese transfer at 10 weeks after stroke in the same rat model (van der Zijden et al, 2008). However, in that study, MnCl₂ was injected into the perilesional sensorimotor cortex, and elevated levels of manganese were measured in subcortical areas of the contralesional hemisphere. Furthermore, in this study, we used a much smaller MnCl₂ dose to prevent potential manganese-induced neurotoxicity (Canals et al, 2008), but which could have reduced the sensitivity to detect tracer accumulation in remote regions, opposite to the injection site. Evidently, the choice of an optimal dose for MEMRI experiments is not straightforward, as it depends on multiple factors and may vary contingent on the research question.

Correlation Between Functional and Structural Connectivity

Our study shows parallel changes in functional and structural connectivity in the ipsilesional and contralesional cortical sensorimotor network after experimental stroke in rats. Furthermore, we detected a significant overall correlation between rs-fMRI-based functional connectivity and MEMRI-based neuroanatomical association in the contralesional (left) hemisphere for all ROIs and animal groups pooled together. Our findings in the anesthetized rat brain corroborate the concept that synchronization of low-frequency BOLD fluctuations are closely associated with structural connectivity, as previously reported for monkey (Vincent et al, 2007) and human brains (Damoiseaux and Greicius, 2009).