Extensive cortical rewiring after brain injury

Abstract

Previously, we showed that the ventral premotor cortex (PMv) underwent neurophysiological remodeling after injury to the primary motor cortex (M1). In the present study, we examined cortical connections of PMv after such lesions. The neuroanatomical tract tracer biotinylated dextran amine was injected into the PMv hand area at least 5 months after ischemic injury to the M1 hand area. Comparison of labeling patterns between experimental and control animals demonstrated extensive proliferation of novel PMv terminal fields and the appearance of retrogradely labeled cell bodies within area 1/2 of the primary somatosensory cortex after M1 injury. Furthermore, evidence was found for alterations in the trajectory of PMv intracortical axons near the site of the lesion. The results suggest that M1 injury results in axonal sprouting near the ischemic injury and the establishment of novel connections within a distant target. These results support the hypothesis that, after a cortical injury, such as occurs after stroke, cortical areas distant from the injury undergo major neuroanatomical reorganization. Our results reveal an extraordinary anatomical rewiring capacity in the adult CNS after injury that may potentially play a role in recovery.

Figure 1.

BDA injection location. A, Left, Diagram illustrating the location of the M1 and PMv hand representations in cerebral cortex of a squirrel monkey. Movement representations as defined by microelectrode stimulation techniques are shown 3 months after an ischemic lesion in the M1 hand area. Sites whose stimulation evoked movements of the digits (red), wrist, and forearm (green) comprised the hand representation. Stimulation of surrounding sites evoked proximal (elbow, shoulder, and trunk; blue) or orofacial (yellow) movements or no response (gray hatching). Lesion location is shown in gray. Right, Demonstration of alignment of anatomical and physiological data. A1 shows the injection core in case 9406. Asterisks indicate locations of selected blood vessels, and cross indicates location of injection site. A2 shows the location of the PMv hand representation (outlined in red) in relation to the surface blood vessel pattern. Asterisks indicate locations of selected blood vessels at which they penetrate radially into the cortical gray matter. These reference markers provided a means to coregister the anatomical and physiological data. Because of shrinkage in the histological processing of the tissue, scaling was required to achieve optimal coregistration. A3 shows these two sets of data superimposed with an outline of the injection core. Finally, A4 shows the complete map of movement representations in PMv with respect to the dense core of the BDA injection. Scale bar, 1 mm. B, Location of injection core in each of the animals in the study (4 control cases on the left, 4 experimental cases on the right). BDA injections were well confined within the PMv hand representation in cases 9406, 21B, and 1662. In case 367E, the injection extended somewhat into nonresponsive and unmapped territory. However, because this case showed a similar pattern of connections when compared with other experimental cases, it was retained for additional analysis. Color code for movement representations is identical to A. Scale bar, 1 mm.

Figure 2.

Pattern of terminal labeling from PMv in control animals. Reconstruction of the typical distribution of terminals observed in flattened, tangential sections through the frontoparietal cortex in a control case (1934; large injection, 2 sections). Because this animal received the largest BDA injection of any in the study and displayed the most extensive and densest distribution of terminals, it provides a reasonable estimate of the limits of normal PMv connectivity. Contours delineate physiologically or histologically defined areas. A gray arrowhead indicates the location of the hand/face septum. A black downward arrow indicates the location of area 1/2 of S1. Terminal labeling in the frontal cortex was found in areas rostral to PMv, the anterior operculum, the rostrolateral portion of M1, the PMd, and the SMA. Terminal labeling in the parietal cortex was confined primarily to the posterior operculum/inferior parietal cortex (PO/IP; area 7b, S2, and PV; black arrowheads) and the posterior parietal cortex. Very sparse labeling can be seen in the primary somatosensory cortex. CS, Central sulcus; CMA, cingulated motor area. Scale bar, 5 mm.

Figure 3.

Pattern of terminal labeling from PMv in experimental animals (>5 months after injury). A, Distribution of PMv terminal labeling in experimental case 9406 (large injection; 2 sections). Terminal labeling was found in each of the regions noted in the control cases. Additionally, atypical labeling was observed caudal to the area 3b hand representation, i.e., in the area 1/2 hand representation (black downward arrow). This same result was replicated in each of the other experimental cases. Note that the atypical terminals were located rostral to the typical labeling found in posterior operculum/inferior parietal cortex (PO/IP; delimited by large black arrowheads). Gray arrowhead indicates the location of the hand/face septum. B, Distribution of PMv terminal labeling in experimental case 1662 (small injection; 2 sections). Scale bar, 5 mm. CS, Central sulcus.

Figure 4.

Relationship of atypical labeling in the parietal cortex with anatomical and physiological boundaries. Example of alignment of myelin-stained section with a somatosensory map. A, Myelin-stained section in experimental case 21B. The hand/face septum is indicated by a thick black line. The boundary between area 3b and area 1/2 is demarcated by a sharp transition from dense to light myelin staining (thin black line). Caudal border of the ischemic lesion in M1 is indicated by a white triangle. As in Figure 1, asterisks represent fiducial markers indicating the locations of selected large blood vessels observable in the same location throughout the depths of the cortical gray matter. CS, Central sulcus. Scale bar, 1 mm. B, Photograph of cortical surface vasculature in the same case (21B). Somatosensory mapping data are superimposed on the vascular pattern. In the collection of physiological data, microelectrode penetration sites (red, white, and yellow dots) are located with reference to the surface vasculature. Asterisks indicate the location of surface blood vessels at which they penetrate radially into the cortex. Locations correspond to those of large blood vessels identified in the myelin-stained sections. C, Alignment of blood vessel pattern in both anatomical and physiological datasets allowed confirmation of the border between areas 3b and 1/2. These data were then superimposed on the reconstructions of BDA labeling using the same approach. Gray rectangle indicates the location of higher-power photomicrographs shown in D. D, Left, Photomicrograph from area 1/2 in same experimental case (21B). Numerous BDA-labeled terminal arbors and fibers can be observed. Asterisk corresponds to largest asterisk in A. Gray box shows location of higher-magnification photomicrograph on the right. Scale bar, 100 μm. Right, Small fiber with varicosities (arrowheads). Scale bar, 1 μm. E, Photomicrograph of area 1/2 in a control case with a large injection (1934). Area 1/2 is virtually devoid of BDA labeling. Scale bar, 100 μm. F, Photomicrograph of the area 1/2 of an experimental case with a large injection (9406). Here, in addition to the numerous terminal arbors and fibers, a labeled somata can be seen. Scale bar, 10 μm.

Figure 5.

Coregistration of terminal labeling with neurophysiological maps of hand representations in S1. Left, Typical pattern of terminal labeling observed in parietal areas of control cases (case 1934; large injection). Middle and Right, Pattern of terminal labeling in two experimental cases (21B and 367E, respectively). Outlines of area 3b functional representations of digits and palm are shown. Arrowheads indicate the caudal border of the posterior operculum/inferior parietal cortex (black) and hand/face septum (gray), respectively. Black downward arrows point to the area 1/2 hand representation. Dotted line indicates the approximate location of the lateral fissure. Scale bar, 1 mm.

Figure 6.

Distribution of labeling in the ipsilateral hemisphere. A, Cell body distribution (excluding S1) in experimental group compared with control group. B, Distribution of terminal labeling (excluding S1) in experimental group compared with control group. Proportions of extrinsic PMv inputs and outputs are shown.

Figure 7.

Distribution of labeled cell bodies and terminals in S1. Distribution of labeled cell bodies and (voxels with labeled) terminals in areas 3a, 3b, and 1/2 of control and experimental cases. Proportions of extrinsic PMv inputs and outputs are shown. ^*p < 0.05, statistically significant differences.

Figure 8.

Abrupt changes in fiber trajectory at the site of lesion. In each of the four experimental cases, abrupt changes in fiber trajectory were observed at the rostral border of the lesion in M1. A, Diagram indicating the locations of the photomicrographs and reconstructions. B, Photomicrographs of BDA-labeled fibers at the rostral border of the M1 lesion in experimental cases 1662 (top) and 9406 (middle) showing fibers that abruptly changed trajectory by 90° or more (black arrowheads). A photomicrograph taken at a corresponding location in a control case 9409 (bottom) is shown for comparison. Note that the large-diameter fibers coursing over long distances are less abundant in the control case and most of the labeling consisted of small terminal arbors. Also note that the orientation of fibers is more variable. Scale bar, 500 μm. C, Reconstruction of small BDA-labeled fibers in experimental case (9406; 1 section). Fibers appear to loop around area 3b to terminate in area 1/2 of S1. Scale bar, 1 mm. CS, Central sulcus.

Figure 9.

Quantitative analysis of fiber trajectories. A, Reconstruction of sections in control case 9409 (small injection). Hand representations of M1 and PMv are outlined. Locations of labeled terminals are shown as blue dots. For quantitative analysis, each section was aligned vertically, using a line drawn between the middle of the injection site and the cluster of BDA labeling found in SMA. A rectangular analysis window of constant dimensions and orientation was then positioned at the rostral border of the M1 hand representation, within the zone of termination. Orientations of every axonal segment within the analysis window were plotted and analyzed for each section of each case. Scale bar, 5 mm. B, Top, Reconstruction of large fibers in control case 9409 (4 sections superposed). In general, axons were relatively short and had scattered orientations. Rectangle indicates 1.2 × 5.0 mm analysis window. Bottom, Reconstruction of large fibers at the border of the lesion in experimental case 9406 (2 sections superposed). In each case in the experimental group, many axons coursed along the rostral border of the M1 lesion for several millimeters. Small gray rectangle indicates the location of a sample axonal segment that is enlarged and displayed on a polar plot in the inner panel. The beginning of the segment is indicated by 1 and the end by 2. This particular axon segment had an orientation of 163°. In the quantitative analysis, orientations were grouped into 10° bins. Thus, this axon segment was tallied in the 160-170° bin (also see Fig. 10 and Materials and Methods). Scale bar, 1 mm.

Figure 10.

Alteration of axonal orientation at the border of M1 lesion. A, Polar histogram illustrating the distribution of large fiber (axonal) orientations at the rostral border of the M1 lesion. For additional details regarding methodology, see Figure 9 and Materials and Methods. Note that control group excludes case 3024 for this analysis (n = 3) because the M1 hand representation was somewhat distorted in this case. After the M1 injury, fiber orientation was more focused and was directed in a more caudolateral direction. B, Diagram illustrating the presumed alterations in PMv intracortical connections after M1 injury. After PMv targets in M1 are destroyed, PMv intracortical fibers are thought to seek new targets; axons abruptly change orientation near the lesion border and begin to course more caudolaterally, sweep around area 3b, and finally terminate in area 1/2.