Hand Motor Recovery Following Extensive Frontoparietal Cortical Injury Is Accompanied by Upregulated Corticoreticular Projections in Monkey

Motor testing.

All motor-control testing devices were attached to the animal's cage and the animal was free to move around the cage while performing the motor tests. Monkeys were food-restricted for 18–24 h before each motor-testing session. Initial motor testing established the preferred hand using a standard dexterity board task with 150 trials over a 3 d period and a hand-preference score was computed based on the number of times each hand was used to acquire small food pellets from wells of various diameters (Nudo et al., 1992). A modified dexterity board (mDB) was used to test fine motor function with testing of each hand controlled by access doors and positioning of the food targets without any restraint on the animal (Pizzimenti et al., 2007). Quantitative 3D video recordings were used to measure duration and accuracy of reaches to the food targets (small food pellets in wells of varying diameters), as well as duration of manipulation and number of times the digit lost contact with the target during manipulation. These measures were also used to compute total, reach (based on duration and accuracy of reach), and manipulation (based on success/failure, duration and number of times the index finger lost contact with the food target) performance score on each trial (Pizzimenti et al., 2007). A modified movement assessment panel (mMAP) was used to assess forces applied during attempts to grasp and manipulate a small food target (carrot chip with central hole for placement on rods attached to the force-measuring device) by each hand, again without any restraint on the animal (Darling et al., 2006). A performance score for each trial to acquire the carrot chip from a flat surface and from straight and curved rods (five trials by each hand for each of these tasks representing three levels of difficulty) was computed based on whether acquisition was successful and on manipulation duration and 3D forces (total absolute impulse) applied during manipulation (Darling et al., 2009).

Neurosurgical and neuroanatomical procedures.

The surgical procedures to induce brain lesions and to inject neurotracers were identical to those described previously for F2 lesions (McNeal et al., 2010) and F2P2 lesions (Morecraft et al., 2015a). Briefly, animals receiving lesions were anesthetized with isoflurane for exposure of cortex and then transferred to intravenous ketamine anesthetization for mapping of hand/arm representations of M1 and LPMC (for F2 lesions) or of M1, LPMC, and the anterior parietal cortex (for F2P2 lesions) using intracortical microstimulation (ICMS; Fig. 1). Following somatotopical mapping, each animal was then returned to isoflurane anesthesia and vessels within the hand/arm cortical field were carefully microcauterized. After a 5–10 min coagulation period, the lesion was made using subpial aspiration with the aid of modified Frazier suction tubes (sizes 4, 5, and 6). The craniotomy was closed using standard neurosurgical technique and, after surgery, each animal was administered buprenorphine and penicillin procaine as previously described (McNeal et al., 2010; Morecraft et al., 2015a, 2016). After postoperative recovery, each animal was tested (see above) for 5 or 11 months after the lesion to track motor recovery. Then 30–34 d before terminating the lesion/motor recovery experiment (i.e., transcardiac perfusion), a second surgery was performed to inject the tract tracer into the arm/hand representation of M2 (Fig. 2), which is described below. The controls did not receive cortical lesions, only tract tracer injection in the M2 arm/hand representation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Line drawings of the medial wall of the cerebral cortex in cases SDM62 (control), SDM48 (F2 lesion), and SDM87 (F2P2 lesion) showing the FD injection site located in the physiologically defined arm/hand representation of M2 in the monkey (Macaca mulatta). The injection site core is shown in solid green and the injection site halo is identified by the dotted line surrounding the core region. The curved arrow indicates the extension of the injection site halo onto the dorsolateral crown of the medial frontal cortex. The inset shows the physiological map of movement representation obtained with intracortical microstimulation to localize the M2 arm/hand representation before injecting FD. Notably, the cortex lining this part of the medial wall of the cerebral cortex is spared in most (∼97%) stroke patients. cf, Calcarine fissure; cgs, cingulate sulcus; D, digit; Ea, ear; El, elbow; Hp, hip; L, leg, NR, no response; ots, occipital temporal sulcus; poms, medial parieto-occipital sulcus; rs, rhinal sulcus; ros, rostral sulcus; SDM, South Dakota Monkey; Sh, shoulder; Ta, tail; Tr, trunk, Wr, wrist.

To accomplish the neural tracer injections in control and lesion cases, the animals were anesthetized with isoflurane, after which skin and bone flaps and a dural incision were made over the midline region to expose the medial aspect of the frontal lobe. Then, while under ketamine anesthesia, ICMS was used to localize of the arm/hand representation of M2. Once localized, the animal was switched back to isoflurane and the anterograde neural tracer fluorescein dextran (FD) or biotinylated dextran amine (BDA) was injected into the central region of the M2 arm/hand area as described previously (McNeal et al., 2010; Fig. 2). Three equally spaced injections totaling 0.9–1.2 μl were made in the central region of the M2 arm/hand representation (McNeal et al., 2010; Morecraft et al., 2015a, 2016). After a 30–34 d survival period to allow for dextran tracer axonal transport, each monkey was deeply anesthetized with an overdose of pentobarbital (≥50 mg/kg, i.p.) and perfused through the heart with 0.9% saline followed by 4% paraformaldehyde and sucrose as previously described (McNeal et al., 2010). The brainstem was blocked (separated) from the diencephalon at a level rostral to the inferior colliculus at a 90° angle to the long axis of the brainstem (Brodal, 1980; Keizer and Kuypers, 1989; Schmahmann et al., 2004). The cortex was frozen-sectioned in the coronal (frontal) plane and the brainstem frozen-sectioned in the transverse plane, both at a thickness of 50 μm and in cycles of 10. In all cases, one cortical and one brainstem tissue series was processed for cytoarchitectural analysis using Nissl staining procedures as described previously (Morecraft et al., 1992, 2012, 2015b). Additional series of tissue sections were used for immunohistochemical localization of each injected dextran tract tracer as previously described (e.g., BDA + FD; Morecraft et al., 2007a,b, 2015a; McNeal et al., 2010; Fig. 3). Cortical sections through the injection site were analyzed to determine the anatomical location of the injection site, as well as the injection site core and halo volumes as previously described (McNeal et al., 2010; Morecraft et al., 2015a, 2016) and transverse brainstem sections were microscopically evaluated for terminal bouton number (described below).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Photomicrographs showing representative examples of the FD injection site in the arm/hand region of M2 and terminal labeling in the Gi. A, Coronal section showing the FD injection site in control case SDM84. The blue arrow identifies a densely coalesced FD-labeled fiber bundle emerging from the injection site and descending toward the internal capsule. Anatomical orientation (bottom left) also applies to B and D. B, Horizontal brainstem section showing ipsilateral terminal labeling in the lGi in control case SDM84. The black arrowheads indicate labeled terminal boutons. The inset is a higher-power photomicrograph from the same tissue section (see asterisk in main panel for orientation) showing an intertwined terminal axon field with numerous well defined boutons. C, Coronal section showing the FD injection site in F2 lesion case SDM45. The blue arrow in the main panel identifies a lightly stained FD-labeled fiber bundle emerging from the injection site. A portion of the LPMC lesion can be seen in the upper right region of the tissue section. The inset is from an adjacent coronal section showing a darkly stained portion of the same bundle curving around the depth of the cortex lining the fundus of the cingulate sulcus (asterisk) toward the internal capsule. Anatomical orientation (bottom right) also applies to E and F. D, Horizontal brainstem section showing contralateral terminal labeling in the mGi in F2 lesion case SDM45. Black arrowheads indicate labeled terminal boutons. The inset is a higher-power photomicrograph from the same tissue section showing FD-labeled terminal boutons in the mGi subsector. The brown-labeled fibers and boutons (brown arrowheads) are from a BDA injection site placed into another cortical area of interest in case SDM45. E, Coronal section showing the FD injection site in F2P2 lesion case SDM71. The blue arrow identifies a densely coalesced FD-labeled fiber bundle emerging from the injection site and descending toward the internal capsule. The scale bar in the bottom right corner of the panel also applies to A and C. F, Horizontal brainstem section showing ipsilateral terminal labeling in the dGi in F2P2 lesion case SDM71. Black arrowheads indicate FD-labeled terminal boutons. The insets all are higher-power photomicrographic images from selected terminal fields of the main panel identified by a matching green, blue, or red asterisk. In A, C, and E, the white dashed line demarcates the external boundary of the FD injection site core and the white dotted line indicates the external limit of the injection site halo. as, Spur of the arcuate sulcus; cgs, cingulate sulcus; d, dorsal; l, lateral; m, medial; v, ventral.

Gigantocellular reticular nucleus of the medulla: anatomical subdivisions.

Several anatomical atlases of the rhesus monkey brain developed for the primary use of stereotaxic application were consulted to locate the general region of the gigantocellular reticular nucleus of the medulla (Gi) and possible subdivisions of the Gi (Snider and Lee, 1961; Martin and Bowden, 2000; Paxinos et al., 2000). However, these atlases do not parcellate the Gi into medial and lateral sectors (only a dorsal, ventral, and centrally located Gi region). In the current report, medial and lateral sectors of the centrally located Gi were identified because it is known that some descending cortical projections preferentially terminate in either the medial or lateral parts of the Gi. Also, the available atlases are prepared in the coronal plane and, as noted, our brainstem tissue sections were prepared in the transverse plane (at a 90° angle to the long axis of the brainstem). Thus, in the current report, we subdivided the medulla Gi into four anatomical subsectors in the transverse dimension, based upon major reoccurring anatomical landmarks [e.g., medial longitudinal fasciculus (MLF), medial lemniscus (ML), inferior olive (IO), hypoglossal nucleus, intermediate reticular nucleus (IRt), and nucleus ambiguus] and some basic cellular characteristics. Along with a basic description of each Gi subdivision, the general location and rostral-to-caudal boundaries of the medullary Gi are described below.

The rostral-most part of the medullary Gi was located immediately above (e.g., 0.5–1 mm) the rostral pole of the hypoglossal nucleus, where the inferior pole of the nucleus prepositus first appears (Fig. 4A,A′). Transverse sections immediately above this level, which included the inferior pole of the facial motor nucleus, were excluded from our quantitative analysis. For the current study, the inferior extent of the Gi coincided with the transverse level just above (0.5–1 mm) the obex of the fourth ventricle (Sakai et al., 2009; Fig. 4D,D′). Some of the reticular nucleus (Paxinos et al., 2000, their MRt and VRt) does extend caudally from this location, corresponding to transverse sections containing the pyramidal decussation. However, these caudal levels were not included in this study because these nuclei are not classified as the Gi proper (Paxinos et al., 2000).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

A–D′, Rostral (A, A′) to caudal (D, D′) series of matching NeuN-stained (A, B, C, D) and Nissl-stained (A′, B′, C′, D′) tissue sections through brainstem levels containing the Gi in Macaca mulatta. The abbreviations for the Gi subdivisions are highlighted in yellow in A and A′ and apply to all panels below. The black asterisk identifies the emerging root fibers of the hypoglossal nerve. The black arrowheads in C′ and D′ identify portions of the internal arcuate fibers. Scalebar, 1 mm. The schematic drawing of the brainstem in D shows the corresponding transverse level of each section. Anatomical orientation in the bottom right corner of D′ applies to all panels. d, Dorsal; d, dorsal subdivision of the paramedian reticular nucleus of Brodal; l, lateral; LP, lateral paragigantocellular reticular nucleus; m, medial; NA, nucleus ambiguus; NP, nucleus prepositus; NS, nucleus of the tractus solitarius; PT, pyramidal tract; pRt, parvicellular reticular nucleus; STT, spinothalamic tract; SV, spinal trigeminal complex (nucleus and tract); TS, tectospinal tract; v, ventral; v, ventral subdivision of the paramedian nucleus of Brodal; VC, spinal trigeminal complex. IRt, intermediate reticular nucleus.

As noted, before the stereological analysis, the Gi was parcellated into four anatomical subdivisions or subsectors based upon major anatomical landmarks and some cellular characteristics. These included the dorsal, medial, lateral, and ventral Gi subsectors (Fig. 4A,A′). Fusiform and small to medium-sized multipolar cells were located throughout the Gi (Fig. 4A–D′), with large multipolar neurons noted primarily in the dorsal and medial subsectors (Fig. 4B–C′).

The dorsal subsector (dGi) is located directly ventral to the hypoglossal nucleus throughout most of the evaluated part of the medulla (Fig. 4B–D′), with the exception of the superior medulla where the dGi is located ventral to the nucleus prepositus (Fig. 4A,A′). Its medial border coincided with the lateral edge of the MLF and its lateral border coincided with the IRt. Encompassed within the dGi is an aggregate of darkly stained cells identified as the dorsal subdivision of the paramedian reticular nucleus localized by Brodal in the cat (Brodal, 1953; Brodal and Torvik, 1954) and by Somana and Walberg in the monkey (Somana and Walberg, 1978; Fig. 4B–D′). In the monkey, Somana and Walberg describe the cells of the dorsal subdivision as being continuous with the ventral subdivision of the paramedian reticular nucleus (discussed below), located medial to the hypoglossal nerve rootlets (Fig. 4B′–D′, asterisk), and being primarily composed of large neurons with fewer small and medium-sized neurons dispersed between the large cells.

The medial subsector of the Gi (mGi) is located lateral to the MLF and tectospinal tract, ventral to the dGi and IRt, medial to the lateral subsector (lGi), and dorsal to the IO (Fig. 4A–D′). This subsector contained small and medium-size neurons with some large multipolar neurons dispersed in the medial and central region of the mGi (Fig. 4B–C′). The medial part of the mGi contains the ventral subdivision of the paramedian reticular nucleus (Fig. 4B–D′) identified by Brodal in the cat (Brodal, 1953; Brodal and Torvik, 1954) and by Somana and Walberg in the monkey (Somana and Walberg, 1978). According to Somana and Walberg, the ventral subdivision is formed by large, medium-size, and small neurons and lies medial to hypoglossal rootlets (Fig. 4B′–D′, asterisks indicate hypoglossal fibers). Notably, both the dorsal subdivision of the paramedian reticular nucleus and the ventral subdivision project to the cerebellum (Brodal, 1953; Brodal and Torvik, 1954; Somana and Walberg, 1978).

The lGi is located dorsal to the IO (Fig. 4A–D′) and, at lower medullary levels, dorsal to the IO, epiolivary subdivision of the lateral reticular nucleus, and lateral paragigantocellular nucleus (Fig. 4B–D′). It is located ventral to the IRt and the darkly stained cluster of cells forming the nucleus ambiguus (Fig. 4B–D). Its lateral border included the spinothalamic tract and, at inferior levels of the medulla, the lateral reticular nucleus (Fig. 4D,D′). The medial border of the lGi is adjacent to the lateral edge of the mGi. Compared with mGi, fusiform and multipolar cells are smaller and less dense and this characteristic can be used to differentiate the mGi from the lGi, at a location just medial to the dorsal midpoint of the IO (Fig. 4A–D′). At lower levels, cells in the lGi occasionally appear as curvilinear columns due to the presence of internal arcuate fibers (Fig. 4C–D′, arrowheads) and olivocerebellar fibers, which pass through the lGi en route to their respective target structures.

The ventral subsector (vGi), which occupies the remainder of the medullary Gi, is bounded medially by the paramedian portion of the ML, ventrally by the ML, laterally by the IO, and dorsally by the mGi (Fig. 4A–D′). The vGi is cell sparse with most cells occurring in the dorsal region of the subsector. We extended the ventral boundary of the vGi to include a small portion of the ML because NeuN tissue sections demonstrated the presence of scattered neurons in this location flattened between ascending ML fiber bundles, and we occasionally noted terminal boutons in this vicinity.

Stereological analysis.

Using stereological techniques, transverse brainstem sections spanning the Gi in the medulla were analyzed to estimate the number of terminal boutons from the M2 CRP in controls and from iM2 CRP in lesioned cases. Terminal bouton number in the ipsilateral and contralateral medullary Gi subsectors were estimated using brightfield microscopy and unbiased stereological counting methods (Glaser and Glaser, 2000; West, 2012). The methods used to calculate terminal bouton numbers are the same as those described in detail in our previous papers for quantifying bouton number in the spinal cord and amygdala (Morecraft et al., 2007a; McNeal et al., 2010). Specifically, estimates of the total number of terminal boutons within each Gi subsector were determined using a 100× oil objective, the Optical Fractionator Probe, and our Neurolucida-equipped (RRID:SCR_001775) microscope workstations (MBF Bioscience; RRID:SCR_004314). For the current project, the counting frame measured 100 × 70 μm and the spacing interval between counting frames was 300 × 150 μm. Estimated bouton numbers were calculated using StereoInvestigator software. Every tissue section through the medulla (spaced 500 μm apart) containing the Gi was used for stereological evaluation. Thus, in all subjects six sections were evaluated with the exception of case SDM91, in which five sections were evaluated. Using cortical tissue sections, unbiased estimates of the total injection site volume and core injection site volume were determined using the Cavalieri probe and calculated using StereoInvestigator software as described in our previous reports (Pizzimenti et al., 2007; McNeal et al., 2010).

We did not attempt to normalize the total bouton number estimates to total number of labeled axons in the medullary pyramid as done by others (Fregosi et al., 2017) because axon sprouting was noted in the pyramid of the lesion cases. To control for our terminal bouton estimates, we provide a detailed assessment of the injection site size (total volume) and break this down into the core region volume (effective uptake/transport area) and halo region volume (potential uptake/transport area; Fig. 3). Thus, our injection site analysis estimates the amount of cortex containing potential CRP neurons across all experimental cases. This is particularly important in our experimental design since the controls were injected with slightly more tracer volume (Tables 2, 3, 4) so that any upregulated neuroplastic change in the M2 corticofugal projection following lateral cortical brain injury could not be attributed to the injection site involving a greater amount of cortex harboring CRP neurons (Morecraft et al., 2015a, 2016).

The design of our stereological application (e.g., choice of dissector probe for quantifying bouton numbers and injection site volumes, counting frame size, guard zone depth, etc.) and technical issues related to our use of dextran tracers and experimental strategy have been discussed in great detail in our previous papers (McNeal et al., 2010; Morecraft et al., 2013, 2015a, 2016).

Experimental design and statistical analyses.

Estimated numbers of terminal boutons were compared among the three groups (controls: four males and one female; F2 lesion: three males, 1 female; F2P2 lesion: four females, one male) using a mixed three-way (group × side × region) repeated-measures ANOVA (side and region as repeated measures). SDM68 was excluded from all analyses because of the smaller volume of neurotracer in this case. Specifically, the estimated number of boutons in the contralateral and ipsilateral projection to each of the four Gi subsectors was entered into this analysis. Huynh–Feldt ε values were examined to determine whether the assumption of sphericity was met when there were ≥3 levels in a repeated-measures factor (i.e., region) and corrected p values were reported for these effects. Post hoc testing for statistically significant effects identified by the ANOVA was performed using Tukey's HSD procedure. Effect sizes (Cohen's d) were computed for selected post hoc comparisons based on an uncorrected t-test comparison of estimated number of boutons in the two groups. Adjustments in degrees of freedom for F tests were made on the basis of the Huynh–Feldt ε values, resulting in adjusted P values, which are reported in Results. Statistical tests were accomplished using Statistica software.

Exploratory single linear regression analyses were also performed to determine whether the ratio of estimated number of boutons in individual lesioned animals (in the bilateral iM2 CRP to the medial subsector and in the total bilateral iM2 CRP) to average estimated number of boutons in controls (independent variable) were closely associated with recovery of reach and grasp/manipulation after the lesion (dependent variable). We assessed association of recovery measures with the strength of the bilateral projection rather than unilateral projection because the M2 projection to the Gi was bilateral (see Results) and the reticulospinal projection is also bilateral (Davidson and Buford, 2006; Schepens and Drew, 2006). Associations of recovery measures with the medial subsector and total projections from M2 to the Gi were considered because the medial subsector was the strongest projection and the total projection would include possible contributions from the other subsectors to recovery. The recovery measures used in these analyses were as follows: (1) postlesion weeks until first success and (2) consistent success (on all five trials) in the mDB and mMAP (curved-rod task); the ratio of highest postlesion skill to prelesion skill in (3) reaching (mDB testing: reach performance scores based on duration and accuracy of reaches) and (4) grasp/manipulation (mDB: manipulation performance scores based on success/failure, manipulation duration, and number of times contact between target and digits is lost); and (5) mMAP (manipulation performance scores based on duration and total applied absolute impulse), as described previously (McNeal et al., 2010; Morecraft et al., 2015a). Briefly, prelesion skill is defined as the mean performance score divided by the SD of performance scores over the last five consecutive motor testing sessions before the lesion. Postlesion skill is defined similarly except that the highest postlesion skill is that observed during five consecutive postlesion testing sessions. We computed p values for one-sided t tests of correlation coefficients because we hypothesized that postlesion duration until first and consistent success would be inversely related to number of boutons and that recovery of skill would be positively correlated with number of boutons. We did not correct p values for multiple regression analyses performed due to the small number of subjects in these exploratory analyses.