AMPA and GABA_A/B Receptor Subunit Expression in the Cortex of Adult Squirrel Monkeys during Peripheral Nerve Regeneration

Abstract

The primate somatosensory neuroaxis provides a highly translational model system with which to investigate adult neural plasticity. Here, we report immunohistochemical staining data for AMPA and GABA_A/B receptor subunits in the area 3b cortex of adult squirrel monkeys one and five months after median nerve compression. This method of nerve injury was selected because it allows unique insight into how receptor expression changes during the regeneration of the peripheral nerve. One month after nerve compression, the pattern of subunit staining provides evidence that the cortex enters a state of reorganization. GABA 1 receptor subunits are significantly down-regulated in layer IV, V, and VI. Glur2/3 AMPA receptor subunits and postsynaptic GABA_BR1b receptor subunits are up and down regulated respectively across all layers of cortex. After five months of recovery from nerve compression, the pattern of AMPA and GABA_A/B receptor subunits remain significantly altered in a layer specific manner. In layer II/III, GluR1, GluR2/3, and GABA 1 subunit expression is significantly up-regulated while post synaptic GABA_BR1b receptor subunits are significantly down regulated. In layer VI, V, and VI the GluR2/3 and presynaptic GABA_BR1a receptor subunits are significantly up-regulated, while the postsynaptic GABA_BR1b receptor subunits remain significantly down-regulated. Taken together, these results suggest that following nerve injury the cortex enters a state of reorganization that has persistent effects on cortical plasticity even after partial or total reinnervation of the peripheral nerve.

Introduction

The receptive field representations of adult neocortex are relatively stable due to a balance between excitatory and inhibitory processes that limit network sensitivity to changes in the pattern of peripheral inputs (for review, see Feldman 2009). While manipulations of peripheral inputs have been known to alter developmental processes since the early 1960's (e.g., Wiesel and Hubel 1963), demonstrations of the effects of deprivation on the mature brain emerged substantially later (Merzenich et. al. 1983a/b). Using a peripheral nerve injury model of sensory deprivation in non-human primates (median nerve transection), two phases of somatosensory plasticity were identified, one immediate, and the other more protracted. The immediate, “unmasking” phase of the cortical response to the loss of sensory inputs, evident minutes to hours after nerve transection, was characterized by an expansion in the representations of skin surfaces on the hand with intact innervation. However, most of the deprived cortex could not be activated by cutaneous peripheral stimulation at this time. Over the ensuing days to weeks after injury, the remaining deprived cortex came to express new receptive fields conveyed by the intact nerves to the hand.

In a subsequent paper, Wall et al. (1983) examined the process of recovery from peripheral nerve injury in a nerve compression model. This approach permitted the introduction of sensory loss that was identical to that accomplished with nerve transection (Merzenich et al., 1983a/b), but with nerve compression, the injured median nerve regenerated and reinnervated skin surfaces that had been deafferented by the manipulation. Electrophysiological mapping of the cortex at various time points after the nerve compression revealed that reorganization of median nerve cortex proceeded during the period in which the nerve was regenerating. Subsequently, as the regenerating median nerve reinnervated the hand, reactivation of the median nerve cortex by medial nerve territory on the hand proceeded sequentially from proximal (i.e., palm) to distal (i.e., digit tips) skin sites. Following complete regeneration, the map of the median nerve territory in cortex was indistinguishable from the one that existed prior to nerve injury (Wall et al. 1983).

Over the last two decades, our laboratory has used the mature non-human primate somatosensory system to systematically investigate the contributions of glutamatergic and GABAergic mechanisms to the process of reorganization that follows deprivation (Garraghty et al. 1991, 2006; Garraghty and Muja 1996; Myers et. al. 2000; Wellman et. al. 2002; Mowery and Garraghty 2009; Mowery et. al. 2011; Sarin, Mowery, and Garraghty 2012). Our studies have provided evidence that AMPA and GABA receptors play key roles in the adult brain's response to sensory deprivation. In our most recent studies we have used the median nerve compression model of sensory deprivation to undertake a series of immunohistochemical studies designed to focus on the changes to AMPA and GABA receptor subunit expression following deprivation, reorganization, and, ultimately reinnervation. Nerve compression (or crush) is uniquely different from other forms of nerve injury in that the peripheral nerve regenerates along an intact neural sheath; essentially reinnervating original cutaneous skin receptors (Barbay et al., 2002, Kis et al., 1999 and Wall et al., 1983). Re-establishment of original topography is more complete than what has been reported for other nerve regeneration paradigms (e.g. transection and repair — Wall and Kaas, 1986).

Using this technique we previously reported that there are significant changes in AMPA and GABA_A/B receptor subunits one week after median nerve compression, and that the resulting patterns of subunit immunostaining closely resemble ones found early in development (Mowery and Garraghty 2009; Mowery et. al. 2011; Sarin, Mowery, and Garraghty 2012). The current studies extend those observations to one month and five months after median nerve compression. At one month after the nerve injury, the cortex shows evidence of having undergone reorganization. Five months after median nerve compression, a time when any ongoing nerve regeneration should be complete, the pattern of AMPA and GABA receptors suggests that the affected region of cortex has regained primary inhibitory tone (GABA_A) and yet retains increased potential for NMDA mediated plasticity.

Methods

Nerve Compression

We report data from 5 adult squirrel monkeys (Saimiri sciureus) that underwent median nerve compressions. Animals survived for either 1 month (N=2) or 5 months (N=3). Animals were maintained on isoflurane anesthesia (2–4%) throughout the surgery. The ventral forearm was shaved between the wrist and elbow, and an incision was made along its midline 70 mm from the wrist. Under microscopic view, the median nerve was located and isolated by blunt dissection and then elevated. The nerve was compressed with hemostats at a 45° angle and held in place for 30 seconds. Compression was repeated three times distally over the span of a few mm to ensure complete axonal compression. The hemostat was rotated 90°, and the process was repeated 3 more times to create a crossing pattern (adapted from Wall et al., 1983). The nerve was repositioned, the skin was sutured over the incision site, and the animal was allowed to recover. All protocols were approved by the Institutional Animal Care and Use Committee (IACUC).

Immunohistochemistry

After one month or five months of recovery, animals were anesthetized with isoflurane gas and transcardially perfused with cold 0.9% saline solution followed by 400 ml of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Following perfusion the brain was extracted and the left and right somatosensory cortices were dissected out and post-fixed for 2 h in cold fixative (4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Tissue was cryoprotected overnight in 30% sucrose in phosphate buffer (pH 7.4).

Sectioning

Frozen 40 m sections of the pre and post-central gyri were cut in a coronal plane and collected in staining nets in phosphate buffer. Sections that clearly contained tissue including the central sulcus were kept for immunohistochemistry. To decrease staining inconsistencies within animals, left and right hemispheres were stained simultaneously in 4 in² wells that contained 8 (1 × 2 in.) wells. These were placed 5 across and 5 down in staining vessels (20 1/2 in²). During primary antibody staining, tissue was incubated in and washed in separate staining vessels (4 1/2 × 20 in.).

Staining

After sectioning, tissue was washed in immuno-phosphate buffer (IPB) and incubated in blocking solution for 30 min followed by hydrogen peroxide .01% for 15 min at room temperature to reduce endogenous peroxidase activity. Tissue sections were incubated overnight at 4 °C in antisera containing GluR1 (1:1000 Chemicon), GluR2/3 (1:1000 Thermo Scientific), GABA_A 1 (1: 1000 Thermo Scientific), GABA_B1a (1:1000 Alpha Diagnostics International), or GABA_B1b (1:1000 Alpha Diagnostics International), and washed three times for ten minutes in IPB before being incubated in goat anti-rabbit biotinylated antibodies for 1 h at room temperature. Sections were again washed three times for ten minutes in PBS, incubated in ABC solution for one hour (ABC Elite Kit, Vector), washed again three times for ten minutes in Acetate-Imidazole buffer, and incubated in Acetate-Imidazole buffer containing Nickel Sulfate, 0.5 mg/ml 3.3 -diaminobenzidine-4HCI (DAB, Vector) and 0.01% H2O2 for approximately 8 min. Sections were washed three times in PBS for ten minutes, mounted on gelatin coated glass slides and dried overnight. Once dry, the sections were dehydrated in ascending ethanols, cleared with xylenes, and cover-slipped. Positive and negative controls were generated by omitting the primary or secondary antibody. Light microscopy of tissue did not reveal the presence of any non-specific binding. Separate lots of antibodies were used between animals, but the quantified difference in the pattern of staining was the same in each animal for each subunit. See Table 1 for a list of antibodies.

Table 1.

List of antibodies used in this study. Antibodies are listed by name, subunit specific target, manufacturer, and dilution used in this study.

Antibody	Immunogen	Manufacturer	Dilution
GluR1	106 kDa (monomer) and 200 kDa (dimer) of rat GluR1 receptor	Chemicon (Billerica), Rabbit monoclonal, # 05-855R	1:1000
GluR2/3	Peptide corresponding to amino acid residues from the C-terminus of rat GluR2/3 receptor	Thermo Scientific (Rockford), Rabbit polyclonal # PA1 -4660	1:1000
GABAAα1	51 kDa C-Terminus of rat GABAAα 1 subunit	Thermo Scientific (Rockford), Rabbit polyclonal, # PA1-4658	1:1000
GABAB1a	18aa peptide from N-Terminus of rat GABAB1a subunit	Alpha Diagnostic (San Antonio), Rabbit polyclonal, #GBR1A11-A	1:1000
G ABAB 1b	16aa peptide from N-Terminus of rat GABAB1b subunit	Alpha Diagnostic (San Antonio), Rabbit polyclonal, #GBR1B11-A	1:1000

Luminance/Densitometry

Immunohistochemical quantification of tissue sections was carried out as previously described (Mowery and Garraghty, 2009) at 1480× under brightfield illumination using a microscope (Nikon Eclipse 80i; Nikon Instruments; Melville, NY, USA) and the Stereo Investigator software (MBF Bioscience; Williston, VT, USA). The Stereo Investigator software's luminance function (formerly densitometry/optical density) was used to quantify staining intensity of the traced contours (representative soma, neuropil, and white matter). The luminance function measures the amount of light passing through the tissue per μm² of contour (0 = black and 256 = white), and in this study was used as an indirect measurement of stained subunit protein expression (based on the principles of densitometry/absorptiometry).

Great effort was made to maintain light parameters within the Stereo Investigator software between users and user sessions. Experimental and control data were collected on the same day under the same light parameters whenever possible. Video settings on the camera were saved once at the beginning of data collection and reloaded every user session. Each day at the beginning of data quantification, light settings were set to approximately 122 (gain) and 55.119 ms (exposure). The camera feed was set between “binocular” and “camera” settings. The condenser was set to the “up” position, and the diaphragm was fully open.

Tissue was focused at 4× and light settings were set to reflect a normal curve on the video histogram. The objective was switched to 40×, and the ND8 and ND32 filters were set to the “out” position. A normal light curve was re-established so that white matter was below the threshold of the histogram (256). After this, no adjustments were made to the settings throughout the duration of the session. When feasible, in an effort to limit variability, quantification was carried out per antibody within a single user session per animal. By switching the filters on and off, we were able to go between 4× and 40× without adjusting the light parameters, which reduced section to section variability even further.

Area 3b

Electrophysiological receptive field mapping was not used in this study as craniotomy and electrode penetration would introduce unknown and uncontrolled variables into the experimental design. The somatotopic representations referred to throughout this study are based on receptive field mapping observations we and others have consistently reported and are therefore approximations of the cortical regions corresponding to the injured nerves. The median nerve compression created a deprived state within a large region of area 3b cortical tissue corresponding to digits 1–3 (Garraghty et al., 2006, Sur et al., 1982 and Wall, Fellman, and Kaas 1983). Therefore and area 1.5 mm posterior and 1.5 mm anterior to the central sulcus was dissected out. Along the anterior/posterior axis this section of tissue was comprised primarily of area 3b as indicated by extensive electrophysiological mapping (see Fig. 9 Sur et. al. 1982). Along the mediolateral axis this section spanned the natural anatomical boundaries imposed by the longitudinal fissure (medial) and the lateral fissure (lateral).

Figure 9.

Changes in GABA_BR1a receptor subunit staining 5 months after nerve injury. Figure 9A: Photomicrographs showing qualitative examples of GABA_BR1a soma staining between control and experimental II/III, IV, and V/VI cortical layers five months after nerve injury: scale bar 5 μm. Figure 9B: Qualitative scatter-plot showing compared GABA_BR1a staining intensity data for all animals across layer and location five months after nerve injury. Figure 9C: Bar histogram showing the quantified difference in GABA_BR1a receptor subunit staining between soma and neuropil five months after nerve injury. Figure 9D: Bar histogram showing the quantified difference in GABA_BR1a receptor subunit staining between cortical layers II/III, IV, and V/VI five months after nerve injury. * = p < .05; ** = p < .01

In our previous studies the central dimple of the central sulcus provides a reliable landmark for locating the hand representation in area 3b (Churchill et al., 1998, Garraghty et al., 1994, Garraghty and Muja, 1995, Schroeder et al., 1995 and Schroeder et al., 1997). Median nerve inputs (the region of interest) have been consistently recorded 1 mm posterior and 1 mm anterior to the central sulcus in a cortical region surrounding the central dimple (e.g., Gingold et al., 1991, Myers et al., 2000; Sur et al., 1982). Tissue was mounted with the anterior face of the tissue section down. In this way, tissue was sliced posterior to anterior starting with presumed area 1 and ending in area 3a. As tissue was sliced in a coronal plane along the A/P axis, only sections that contained the central dimple of the central sulcus were kept for staining. Sections prior to the central dimple (~ area 1) were discarded and slicing stopped once the dimple was no longer seen (~ area 3a). Within these sections, areas approximate to digit representations one through three were identified according to the medial lateral axis where they are consistently represented (see Fig. 1 - Garraghty et. al. 2006). Sampling was carried out in the center of this region that we posit contains neurons dominantly driven by digit two median nerve input.

Figure 1.

Changes in GluR1 receptor subunit staining 1 month after nerve injury. Figure 1A: Top: Cartoon of the squirrel monkey cortex identifying area 3b (gray bar), CD; central dimple, R; rostral, C; caudal, M; medial, L; lateral. Bottom: Coronal section through area 3b identifying region of interest (black box), CD; central dimple, L; lateral, M; medial. Figure 1B: Photo-micrograph of a control section indicating the sampling region for immunohistochemical quantification of staining intensity (Median Input): scale bar 0.5 mm; CD central dimple, ROI region of interest. Figure 1C: Photomicrographs showing qualitative examples of GluR1 soma staining between control and experimental II/III, IV, and V/VI cortical layers 1 month after nerve injury: scale bar 5 μm. Figure 1D: Qualitative scatter-plot showing compared GluR1 staining intensity data for all animals across layer and location one month after nerve injury. Figure 1E: Bar histogram showing the quantified difference in GluR1 receptor subunit staining between soma and neuropil one month after nerve injury. Figure 1F: Bar histogram showing the quantified difference in GluR1 receptor subunit staining between cortical layers II/III, IV, and V/VI one month after nerve injury.

Bounding Box

A 1 mm wide tracing grid was aligned approximately within the area 3b median nerve representation (~ digit 2), which is ~ 1 mm lateral to the central sulcus dimple. There are no clear histological boundaries between digits, so it is conceivable that the tracing grid sometimes included regions that receive thalamocortical inputs from digit 1 or 3; however, these remain within the experimental input zone. Once the tracing grid was placed, the user moved throughout the predefined cortical layers (II–IV approximate) tracing the somata of user defined neurons within a centrally placed bounding box (100 mm²). Layers were discernable to the trained observer based on differences in packing density (4×) and neural phenotype (40×). Quantification was carried out in the middle of each layer, as there are no abrupt laminar boundaries.

Data Collection

Tissue was originally sliced at 40 μm, but following staining procedures and cover-slipping had a final tissue thickness of around 25 to 30 μm. The microscope was focused down through the z-plane to a depth of around 15 μm where uniformly stained somas within the plane of focus were randomly selected and traced within the bounding box (~ 30 per section). Additionally, within each cortical layer, neuropil measurements were taken (~ 10 per section). These were operationally defined as regions of tissue in which z-axis scrolling from the top to the bottom of the tissue revealed no somata. Finally white matter measurements were collected (~ 5 per section), that consisted of regions in which antibody binding was absent. Three to five sections were measured per antibody per hemisphere per animal generating around 90 –150 soma measurements and 30–50 neuropil measurements per hemisphere per antibody. Staining intensity data were generated by quantifying the ratio of each traced soma or neuropil luminance value over the average staining intensity of the white matter luminance values in that section. This transformed each data point into a percent darker than white matter score — also referred to as subunit staining intensity. For each animal average staining intensity was calculated for experimental and control area 3b layers II/III, IV, and V,VI soma and neuropil. This generated a total of six experimental and six experimental data points per animal.

To control for any differences in staining imposed by variability due to perfusion quality, nerve injury, or immunohistochemical reagents, averaged scores for each animal's experimental hemisphere data were compared to the control hemisphere data using paired comparisons. Despite possible differences in staining intensity, the effects reported in this study were consistent across animals. That is, a reported significant change in staining intensity is one that was consistent across the animals used in this study (See qualitative scatter-plots).

Quantification

To determine whether there was a general effect of nerve injury on receptor subunit expression, paired comparisons were made between data collected from experimental and control hemispheres. Receptor density varies between cortical layers (Geyer et. al. 1998; Rakic et.al. 1998; Shaw et. al. 1991; Young et al.,1990), as well as, between soma and neuropil (Karube, Kubota, and Kawaguchi 2004; Vickers et. al. 1993; Wang et. al. 2004). Because of this, two types of paired comparisons were used. One comparison determined whether significant increases or decreases were present between the layers of experimental and control area 3b cortex. In these analyses soma and neuropil data were collapsed across layer. A second comparison was used to determine whether soma or neuropil staining intensity significantly increased or decreased across layers of experimental and control area 3b cortex. In these analyses layer specific data was collapses across location. Both types of analysis lead to larger SEMs because of the natural variance in staining intensity between layer and location. Furthermore, the low number of animals used in these experiments decreased the power of all statistical analyses. While some analyses yielded results that approached significance (p < .1), only p values < .05 were considered significant.

Results

AMPAR and GABAR SUBUNIT EXPRESSION (1 MONTH)

GluR1 Subunit Staining Intensity (1 Month)

The average GluR1 staining intensities for cytoarchitectonic area 3b cortical layers were compared within animals (N=2) between experimental and control hemispheres one month after median nerve compression. GluR1 staining intensity scores were quantified in somata and neuropil of layers II/III, IV, and V/VI from a region of cortex that receives input from the median nerve (Figure 1A/B). Figure 1B presents a micrograph illustrating the region in which data was acquired. Samples from an area approximating digit two of the hand representation were taken immediately 1mm lateral to the central dimple (CD) of the central sulcus and extending no more than 2 mm laterally (demarcated by the bounding box; also see Sur et al., 1982). These conventions were the same for all data collected within this study.

Qualitative examples of somatic staining are presented in Figure 1C. The compared values for both animals are presented in Figure 1D. These data show no obvious trends for either animal. Figure 1E shows that there were no differences between experimental and control hemispheres for data collected within soma [mean ± SEM; experimental 105.52 ± 16.67 vs. control 104.97 ± 17.28; t (5) = .72 p = .50]. There were also no differences in neuropil staining intensity [mean ± SEM; experimental 70.05 ± 15.22 vs. control 68.80 ± 15.33; t (5) = .78 p = .47]. As shown in Figure 1F, comparing staining intensity data across experimental and control cortical layers revealed no significant differences in layers II/III [mean ± SEM; experimental 132.29 ± 9.48 vs. control 132.33 ± 10.91; t (3) = .021, p = .98], IV [mean ± SEM; experimental 46.78 ± 6.22 vs. control 45.40 ± 6.21 t (3) = 1.21, p = .31], or V/VI [mean ± SEM; experimental 84.29 ± 15.36 vs. control 82.93 ± 15.37 – t (3) = 1.06 p = .36].

GluR2/3 Subunit Staining Intensity (1 Month)

The average GluR2/3 staining intensities for cytoarchitectonic area 3b cortical layers were compared within the soma and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 2) one month after median nerve compression. Figure 2A presents qualitative examples of GluR2/3 expression in control and experimental sections of II/III, IV, and V/VI layers of cortex. The compared data show a trend towards increased GluR2/3 staining intensity in both animals across layers and within soma and neuropil (Figure 2B). As seen in Figure 2C, statistical comparison revealed a significant increase in neuronal soma staining intensity between control and experimental hemispheres [mean ± SEM; experimental 91.65 ± 16.48 vs. control 83.61 ± 15.95 – t (5) = 3.09 p = .027]. This significant increase was also found within the neuropil [mean ± SEM; experimental 69.84 ± 13.00 vs. control 57.28 ± 13.03 – t (5) = 4.89, p = .004]. In figure 2D, paired comparison between the cortical layers of experimental and control hemispheres revealed a significant increase in layer II/III [mean ± SEM; experimental 122.32 ± 8.64 vs. control 112.66 ± 10.20 – t (3) = 3.65, p = .035], IV [mean ± SEM; experimental 41.99 ± 3.00 vs. control 35.95 ± 4.66 – t (3) = 3.43, p = .041], and V/VI subunit staining intensity [mean ± SEM; experimental 77.93 ± 7.60 vs. control 62.71 ± 9.89 – t (3) = 4.03, p = .027].

Figure 2.

Changes in GluR2/3 receptor subunit staining 1 month after nerve injury. Figure 2A: Photomicrographs showing qualitative examples of GluR2/3 soma staining between control and experimental II/III (II/III), IV (VI), and V/VI (V/VI) cortical layers 1 month after nerve injury: scale bar 5 μm. Figure 2B: Qualitative scatter-plot showing compared GluR2/3 staining intensity data for all animals across layer and location one month after nerve injury. Figure 2C: Bar histogram showing the quantified difference in GluR2/3 receptor subunit staining between soma and neuropil one month after nerve injury. Figure 2D: Bar histogram showing the quantified difference in GluR2/3 receptor subunit staining between cortical layers II/III, IV, and V/VI one month after nerve injury. * = p < .05; ** = p < .01; *** = p <.001.

GABA_A α 1 Subunit Staining Intensity (1 Month)

The average GABA_A 1 staining intensities for cytoarchitectonic area 3b cortical layers were compared within the soma and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 2) one month after median nerve compression. Figure 3A presents qualitative examples of GABA_A 1 expression in control and experimental sections of II/III, IV, and V/VI layers of cortex. The compared values from both animals show trends indicative of reduced staining intensity for neuropil and layer IV and V/VI in the experimental hemisphere, but no change in layer II/III soma staining intensity (Figure 3B). As seen in figure 3C, paired comparison revealed a non-significant effect of nerve-injury on GABA_A 1 subunit staining intensity within experimental soma (mean ± SEM; experimental 76.26 ± 16.63 vs. control 79.87 ± 14.61 – t (5) = 1.48 p = .197). Staining intensity within the neuropil; however, was significantly decreased [mean ± SEM; experimental 49.79 ± 12.54 vs. control 57.05 ± 11.12 – t (5) = 4.69, p = .005]. Figure 3D shows the paired comparisons for subunit staining intensity between experimental and control cortical layers. As the trends across both animals indicated, there was not a significant difference in staining intensity within layer II/III [mean ± SEM; experimental 106.51 ± 11.19 vs. control 106.50 ± 9.49 – t (3) = .004, p = .996]. There was a significant decrease in staining intensity for measurements taken in the experimental hemispheres of cortical layers IV [mean ± SEM; experimental 29.34 ± 4.11 vs. control 38.04 ± 3.42 – t (3) = 6.84, p = .006], and V/VI [mean ± SEM; experimental 53.23 ± 8.80 vs. control 60.84 ± 7.84 – t (3) = 3.58, p = .037].

Figure 3.

Changes in GABA_A α1 receptor subunit staining 1 month after nerve injury. Figure 3A: Photomicrographs showing qualitative examples of GABA_A α1 somal staining between control and experimental cortical layers II/III, IV, and V/VI 1 month after nerve injury: scale bar 5 μm. Figure 3B: Qualitative scatter-plot showing compared GABA_A α1 staining intensity data for all animals across layer and location one month after nerve injury. Figure 3C: Bar histogram showing the quantified difference in GABA_A α1 receptor subunit staining between soma and neuropil one month after nerve injury. Figure 3D: Bar histogram showing the quantified difference in GABA_A α1 receptor subunit staining between cortical layers II/III, IV, and V/VI one month after nerve injury. * = p < .05; ** = p < .01; *** = p <.001.

GABA_BR1a Subunit Staining Intensity (1 Month)

The average GABA_BR1a staining intensities for cytoarchitectonic area 3b cortical layers were compared within the somata and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 2) one month after median nerve compression. Figure 4A presents qualitative examples of GABA_BR1a staining intensity in the soma of control and experimental sections of II/III, IV, and V/VI layers of cortex. The compared data for both animals in Figure 4B show no clear trends for changes in subunit staining intensity. Statistical comparison (Figure 4C) indicates that there were no differences in staining intensity for soma [mean ± SEM; experimental 68.07 ± 14.89 vs. control 69.40 ± 14.22 – t (5) = .418 p = .693] or neuropil measurements [mean ± SEM; experimental 46.65 ± 11.88 vs. control 46.13 ± 11.75 – t (5) = .418 p = .693]. Furthermore Figure 4D indicates that there were also no significant difference for subunit staining intensity within experimental cortical layers II/III [mean ± SEM; experimental 97.51 ± 9.05 vs. control 96.00 ± 9.41 – t (3) = .659 p = .556], IV [mean ± SEM; experimental 27.66 ± 3.88 vs. control 29.15 ± 4.70 – t (3) = 1.78, p = .17], or V/VI [mean ± SEM; experimental 48.15 ± 7.58 vs. control 46.91 ± 9.08 – t (3) = .678 p = .546].

Figure 4.

Changes in GABA_BR1a receptor subunit staining 1 month after nerve injury. Figure 4A: Photomicrographs showing qualitative examples of GABA_BR1a soma staining between control and experimental cortical layers II/III , IV, and V/VI 1 month after nerve injury: scale bar 5 μm. Figure 4B: Qualitative scatter-plot showing compared GABA_BR1a staining intensity data for all animals across layer and location one month after nerve injury. Figure 4C: Bar histogram showing the quantified difference in GABA_BR1a receptor subunit staining between soma and neuropil one month after nerve injury. Figure 4D: Bar histogram showing the quantified difference in GABA_BR1a receptor subunit staining between cortical layers II/III, IV, and V/VI one month after nerve injury. * = p < .05; ** = p < .01; *** = p <.001.

GABA_BR1b Subunit Staining Intensity (1 Month)

The average GABA_BR1b staining intensities for cytoarchitectonic area 3b cortical layers were compared within the somata and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 2) one month after median nerve compression. Figure 5A presents qualitative examples of GABA_BR1b expression in control and experimental sections of II/III, IV, and V/VI layers of experimental and control cortex. Compared data for both animals shows a clear trend towards a decrease in subunit staining intensity across all layers within soma and neuropil (Figure 5B). Statistical comparison in Figure 5C shows significant decreases for staining intensity within the soma [mean ± SEM; experimental 78.47 ± 15.17 vs. control 92.72 ± 15.85 – t (5) = 2.76, p = .0398] and neuropil of neurons within the experimental hemisphere [mean ± SEM; experimental 51.78 ± 14.37 vs. control 64.11 ± 13.29 – t (5) = 3.35, p = .020]. Figure 5D shows that the significant decrease was also seen in experimental cortical layers II/III [mean ± SEM; experimental 106.70 ± 10.10 vs. control 119.70 ± 10.48 – t (3) = 3.21, p = .048], IV [mean ± SEM; experimental = 32.57 ± 6.33 vs. control 42.53 ± 5.58 – t (3) = 4.02, p = .027], and V/VI [mean ± SEM; experimental 56.11 ± 13.08 vs. control 73.00 ± 11.24 – t (3) = 3.78, p = .034].

Figure 5.

Changes in GABA_BR1b receptor subunit staining 1 month after nerve injury. Figure 5A: Photomicrographs showing qualitative examples of GABA_BR1b soma staining between control and experimental cortical layers II/III, IV, and V/VI, 1 month after nerve injury: scale bar 5 μm. Figure 5B: Qualitative scatter-plot showing compared GABA_BR1b staining intensity data for all animals across layer and location one month after nerve injury. Figure 5C: Bar histogram showing the quantified difference in GABA_BR1b receptor subunit staining between soma and neuropil one month after nerve injury. Figure 5D: Bar histogram showing the quantified difference in GABA_BR1b receptor subunit staining between cortical layers II/III, IV, and V/VI one month after nerve injury. * = p < .05; ** = p < .01

AMPAR and GABAR Subunit Expression (5 Months)

GluR1 Subunit Staining Intensity (5 Months)

The average GluR1 staining intensities for cytoarchitectonic area 3b cortical layers were compared within the soma and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 3) five months after median nerve compression. Figure 6A presents qualitative examples of GluR1 staining intensity in the soma of control and experimental sections of II/III, IV, and V/VI layers of cortex. The compared data from all three animals only show a trend of decreased staining intensity in layers II/III of experimental cortex (Figure 6B). Comparison of staining intensity data collected from the soma [mean ± SEM; experimental 83.42 ± 9.24 vs. control 83.67 ± 9.61 – t (8) = .493 p = .63] and neuropil [mean ± SEM; experimental 49.22 ± 6.45 vs. control 47.16 ± 7.19 – t (8) = 2.11 p = .067] of experimental and control hemispheres show no significant difference (Figure 6C). Figure 6D shows a significant decrease in receptor subunit expression in layer II/III [mean ± SEM; experimental 90.80 ± 8.21 vs. control 92.35 ± 8.25 – t (5) = 2.83, p = .036], but no changes for layers IV [mean ± SEM; experimental 38.39 ± 4.54 vs. control 36.29 ± 5.00 – t (5) = 3.39, p = .019], or V/VI [mean ± SEM; experimental 69.77 ± 10.68 vs. control 67.61 ± 11.72 – t (3) = 1.06 p = .36].

Figure 6.

Changes in GluR1 receptor subunit staining 5 months after nerve injury. Figure 6A: Photomicrographs showing qualitative examples of GluR1 soma staining between control and experimental II/III, IV, and V/VI cortical layers five months after nerve injury: scale bar 5 μm. Figure 6B: Qualitative scatter-plot showing compared GluR1 staining intensity data for all animals across layer and location five months after nerve injury. Figure 6C: Bar histogram showing the quantified difference in GluR1 receptor subunit staining between soma and neuropil five months after nerve injury. Figure 6D: Bar histogram showing the quantified difference in GluR1 receptor subunit staining between cortical layers II/III, IV, and V/VI five months after nerve injury. * = p < .05

GluR2/3 Subunit Staining Intensity (5 Months)

The average GluR2/3 staining intensities for cytoarchitectonic area 3b cortical layers were compared within the soma and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 3) five months after median nerve compression. Figure 7A presents qualitative examples of GluR2/3 expression in control and experimental cortical layers of II/III, IV, and V/VI. A trend towards increased staining intensity is seen for all three animals across experimental cortical layers, soma, and neuropil (figure 7B). Paired comparisons shown in Figure 7C indicate a significant nerve-injury induced increase in staining intensity for neuronal soma [mean ± SEM; experimental 62.61 ± 8.08 vs. control 59.82 ± 7.81 – t (8) = 4.05 p = .0037] and neuropil within the experimental hemisphere [mean ± SEM; experimental 41.10 ± 5.77 vs. control 38.63 ± 5.7 – t (8) = 4.08, p = .003]. As seen in Figure 7D, There was also a significant increase in staining intensity for layers II/III [mean ± SEM; experimental 73.31 ± 7.25 vs. control 70.05 ± 7.42 – t (5) = 2.92, p = .032], IV [mean ± SEM; experimental 29.07 ± 3.45 vs. control 27.19 ± 3.57 – t (5) = 7.19, p = .0008], and V/VI of the experimental hemisphere [mean ± SEM; experimental 53.18 ± 6.73 vs. control 50.18 ± 6.46 – t (5) = 3.77, p = .013].

Figure 7.

Changes in GluR2/3 receptor subunit staining 5 months after nerve injury. Figure 7A: Photomicrographs showing qualitative examples of GluR2/3 soma staining between control and experimental II/III, IV, and V/VI cortical layers five months after nerve injury: scale bar 5 μm. Figure 7B: Qualitative scatter-plot showing compared GluR2/3 staining intensity data for all animals across layer and location five months after nerve injury. Figure 7C: Bar histogram showing the quantified difference in GluR2/3 receptor subunit staining between soma and neuropil five months after nerve injury. Figure 7D: Bar histogram showing the quantified difference in GluR2/3 receptor subunit staining between cortical layers II/III, IV, and V/VI five months after nerve injury. * = p < .05; ** = p < .01; *** = p <.001.

GABA_A α 1 Subunit Staining Intensity (5 Months)

The average GABA_A 1 staining intensities for cytoarchitectonic area 3b cortical layers were compared within the somata and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 3) five months after median nerve compression. Figure 8A presents qualitative examples of GABA_A 1 expression in control and experimental cortical layers II/III, IV, and V/VI. The compared data show a trend towards a significant increase in receptor subunit staining intensity for layer II/III of the experimental hemisphere (Figure 8B). In figure 8C, paired comparison revealed no significant change in GABA_A 1 subunit staining intensity between experimental and control hemisphere soma [mean ± SEM; experimental 84.53 ± 9.73 vs. control 83.26 ± 9.63 – t (8) = 2.00 p = .080] and neuropil [mean ± SEM; experimental 52.63 ± 8.44 vs. control 51.92 ± 7.27 – t (8) = .499, p = .63]. Figure 8D shows that staining intensity data from experimental II/III layers did show a significant increase in staining intensity when compared to control data [mean ± SEM; experimental 99.30 ± 7.11 vs. control 95.60 ± 7.88 – t (8) = .3.98, p = .0105]. At the same time, staining intensity data for IV [mean ± SEM; experimental 37.86 ± 4.71 vs. control 38.45 ± 3.94 – t (8) = .656, p = .540], and V/VI cortical layers [mean ± SEM; experimental 68.57 ± 10.03 vs. control 68.72 ± 9.61 – t (8) = .108, p = .917] showed no significant difference.

Figure 8.

Changes in GABA_A α1 receptor subunit staining 5 months after nerve injury. Figure 8A: Photomicrographs showing qualitative examples of GABA_A α1 soma staining between control and experimental II/III, IV, and V/VI cortical layers five months after nerve injury: scale bar 5 μm. Figure 8B: Qualitative scatter-plot showing compared GABA_A α1 staining intensity data for all animals across layer and location five months after nerve injury. Figure 8C: Bar histogram showing the quantified difference in GABA_A α1 receptor subunit staining between soma and neuropil five months after nerve injury. Figure 8D: Bar histogram showing the quantified difference in GABA_A α1 receptor subunit staining between cortical layers II/III, IV, and V/VI five months after nerve injury. * = p < .05

GABA_BR1a Subunit Staining Intensity (5 Months)

The average GABA_BR1a staining intensities for cytoarchitectonic area 3b cortical layers were compared within the somata and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 3) five months after median nerve compression. Figure 9A presents qualitative examples of GABA_BR1a staining intensity in the soma of control and experimental cortical layers II/III, IV, and V/VI. Figure 9B shows that there was a trend towards increased subunit expression in layers IV, and V/VI of experimental hemispheres for all three animals. This trend was also seen for the soma and neuropil measurements across layers. Paired comparisons shown in figure 9C indicates a significant increase in GABA_BR1a subunit staining intensity for both neuronal soma [mean ± SEM; experimental 64.45 ± 7.76 vs. control 61.14 ± 8.21 – t (8) = 3.08 p = .015] and neuropil of experimental hemispheres [mean ± SEM; experimental 42.11 ± 5.33 vs. control 39.09 ± 5.85 – t (8) = 2.78 p = .023]. Figure 9D shows that this increase was not present in cortical layers II/III [mean ± SEM; experimental 74.05 ± 7.16 vs. control 73.36 ± 7.27 – t (5) = .659 p = .905]; however, there were significant increases in layers IV [mean ± SEM; experimental 31.40 ± 3.08 vs. control 28.25 ± 3.32 – t (5) = 2.64, p = .045] and V/VI of experimental hemispheres [mean ± SEM; experimental 54.39 ± 6.69 vs. control 48.73 ± 6.98 – t (5) = .5.35 p = .003].

GABA_BR1b Subunit Staining Intensity (5 Months)

The average GABA_BR1b staining intensities for cytoarchitectonic area 3b cortical layers were compared within the soma and neuropil of layers II/III, IV, and V/VI for the experimental and control hemispheres of animals (N = 3) five months after median nerve compression. Figure 10A presents qualitative examples of GABA_BR1b expression in control and experimental sections of II/III, IV, and V/VI layers of cortex. The compared data shows a clear trend in all three animals of decreases in subunit staining intensity across all layers for both soma and neuropil (Figure 10B). As seen in figure Figure 10C, paired comparisons revealed a significant nerve-injury induced decrease in GABA_BR1b subunit staining intensity between experimental and control hemispheres for soma [mean ± SEM; experimental 63.61 ± 8.09 vs. control 79.24 ± 7.97 – t (8) = 4.57, p = .0018] and neuropil measurements [mean ± SEM; experimental 41.05 ± 6.51 vs. control 53.89 ± 7.97 – t (8) = 3.43, p = .008]. Figure 10D shows that there was also a significant decrease in experimental cortical layers II/III [mean ± SEM; experimental 76.29 ± 6.85 vs. control 97.09 ± 6.93 – t (5) = 5.53, p = .002], IV [mean ± SEM; experimental = 28.77 ± 3.92 vs. control 36.44 ± 3.90 – t (5) = 2.60, p = .047], and V/VI [mean ± SEM; experimental 51.60 ± 7.07 vs. control 66.30 ± 8.22 – t (5) = 3.28, p = .021].

Figure 10.

Changes in GABA_BR1b receptor subunit staining 5 months after nerve injury. Figure 10A: Photomicrographs showing qualitative examples of GABA_BR1b soma staining between control and experimental II/III, IV, and V/VI cortical layers five months after nerve injury: scale bar 5 μm. Figure 10B: Qualitative scatter-plot showing compared GABA_BR1b staining intensity data for all animals across layer and location five months after nerve injury. Figure 10C: Bar histogram showing the quantified difference in GABA_BR1b receptor subunit staining between soma and neuropil five months after nerve injury. Figure 10D: Bar histogram showing the quantified difference in GABA_BR1b receptor subunit staining between cortical layers II/III, IV, and V/VI five months after nerve injury. * = p < .05; ** = p < .01

Discussion

In this study we continue a series of experiments that investigated the immunostaining of AMPA and GABA receptor subunits in cytoarchitectonic area 3b of monkey cortex following median nerve compression. This injury model was selected because it offers the opportunity examine the changes in several targeted neurotransmitter subunits as the cortex undergoes the process of topographic reorganization and peripheral nerve reinnervation (Merzenich et al., 1983a, b; Wall et al., 1983; Schroeder et al., 1997; Wall et al., 1983). To that end, we find that the pattern of subunit expression provides insight into the receptor phenotypes that are altered during cortical reorganization (Garraghty et. al. 2006), and also suggests that changes in cortical receptor expression persist even following regenerating nerve injuries.

Unmasking & Developmental Recapitulation

The unmasking phase of somatosensory reorganization occurs immediately after the loss of peripheral inputs (Merzenich et. al. 1983b; Calford and Tweedale 1991; Myers et al. 2000). During this phase, masked latent inputs are disinhibited and immediately expressed as islands of novel somatotopic representations. The shift in receptor expression during the unmasking phase is characterized by a significant decline in GABA_A receptors in the deprived cortex (Wellman et. al. 2002; Garraghty et. al. 2006). This neuronal response to pathophysiological conditions likely serves as an essential priming step for the induction of developmental-like plasticity in the adult system (Rivera et al., 1999, 2005).

A few days after nerve injury, GABAR withdrawal (Casasola et al., 2004 and Silva-Barrat et al., 1989) induces an intermediate state of plasticity characterized by a pattern of neuronal activity and receptor correlates that resemble a recapitulation of developing neocortex (physiology - Lenz et al., 1989, 1998; Dykes et al., 1995; Bergenheim et al., 2004; Sun et al., 2005; receptor expression - Mowery and Garraghty 2009; Mowery et. al. 2011; Sarin et al., 2012). For example, spontaneous neuronal activity takes on the bursting characteristics of developing neurons (Feller, 1999), and AMPA and GABA_A/B subunits are up and down regulated into a pattern of expression described in developing systems (AMPA - Eybalin et al., 2004, Ho et al., 2007, Kumar et al., 2002; Whitney et al., 2008; GABA - Fritschy et al., 1999; McLean et al., 1996; Fukuda et al., 1993). This intermediate state has been suggested to be an important priming step for continued reorganization (Cusick et. al. 1990).

Reorganization and Onset of Reinnervation

The second stage of the cortical response to nerve injury involves the progressive reorganization of somatosensory receptive fields in the weeks following nerve injury (Merzenich et. al. 1983a; Schroeder et al., 1997). One month after median nerve transection, the cortex has undergone reorganization (Garraghty and Muja, 1996); however, with median nerve compression, the reinnervation of the hand by the regenerating median nerve is just beginning (see Figure 2; Wall et al. 1983). While the physiological state of the cortex was not investigated in this study, the pattern of results we report here is similar to those described for AMPA and GABA_A/B receptors in experiments that use a non-regenerating form of nerve injury. For example, AMPA receptor numbers were elevated across all cortical layers one to two months after nerve transection (Garraghty et al., 2006). For the nerve transected monkeys, we argued that the increase in AMPA receptors paralleled reports of increases in AMPA receptor expression in the hippocampus found after NMDA receptor-dependent LTP (e.g., Tocco et al., 1992; Maren et al., 1993). Presumably, NMDA receptor-dependent plasticity mechanisms are operating in the deprived cortex after nerve compression as well. A significant elevation in the GluR1/2 containing AMPAR would presumably be accompanied by a significant increase in GluR1 receptor expression which we find no evidence of. Therefore, we would posit that the previously unrevealed AMPAR species would be those containing the GluR2/3 subunits. Which were significantly elevated across all layers and locations.

The pattern of expression for GABA_A α1 subunit staining intensity is very similar to that found 1-2 months after median nerve transection (Garraghty et al., 2006; Fig. 2). In that study we reported a significant decrease in the number of GABA_A receptors in layers VI and V/VI. In the current study GABA_A α1 subunit expression is also significantly lowered in IV and V/VI layers while indicating no change in layers II/III. Given that the 1 subunit is predominantly found in the mature GABA_A receptor phenotype (1 α1, 2 α2, and 2 y2 subunits – Golshani et al., 1997) we maintain that the result from the current study corroborates our previous finding.

In our previous study GABA_B expression was significantly reduced in IV and V/VI layers during reorganization; however, the isoform of the GABA_B receptor remained unknown (Garraghty et. al. 2006). The presynaptic GABA_BR1a subunit staining intensity is not significantly different from controls at this time suggesting this receptor isoform did not contribute to the significant decrease in GABA_B receptor expression. Instead our results suggest that the GABA_BR1b containing postsynaptic isoform (Benke et. al. 1999) is significantly down-regulated after nerve injury. A parsimonious explanation for this highly significant decrease in subunit density is that prolonged activation of NMDA receptors induces a large scale endocytocis of the postsynaptic GABA_B receptor (Terunuma et. al. 2010). Furthermore, the postsynaptic GABA_B receptor has been shown to inhibit the function of NMDA receptors (e.g., Otmakhova and Lisman, 2004). Thus, a significant reduction in postsynaptic GABA_B would in turn serve to facilitate the ongoing process of NMDA dependent reorganization (Garraghty and Muja 1996).

Reinnervation and Recovery

In this study we used a form of nerve compression injury that has been reported to induce peripheral reinnervation of the hand and distal digit tips ~ 3 months after nerve compression (Wall et al., 1983). That is, the centrally-projecting afferents once again activated neurons topographically across the appropriate sector of cortical area 3b. For this study a survival duration of five months was selected to promote full regeneration of the median nerve; however, it is important to note that the complete recovery from peripheral nerve injury in humans and animal models has been reported to vary on an individual basis (eg Gordon et. al. 2007; Ronchi et. al 2009). No electrophysiological data was acquired from these animals and the state of functional reactivation of cortex is unknown. Therefore the results from this study should be interpreted as coming from a region of cortex that has undergone acute deprivation and reorganization with possible reactivation by original peripheral receptors following partial or full reinnervation of the compressed nerve.

Regardless, the pattern of subunit expression is different from that reported at 1 month post injury; suggesting a different state of plasticity. Furthermore it varies between layers II/III, and layers VI, V, and VI. Specifically the upper layers have significantly elevated levels of GluR1 and GluR2/3. This could suggest that both the GluR1/2 and GluR2/3 receptor isoforms of AMPA receptor have been up-regulated, providing evidence of increased NMDA dependent LTP within this system (Tocco et al., 1992; Maren et al., 1993). At the same time, we report a significant increase in GABA_A α1 subunit, and this could indicate a concomitant increase in GABA_A receptor subunits. Finally the presynaptic GABA_B receptor is not significantly different from controls while the postsynaptic GABA_B receptor is significantly decreased. From a plasticity perspective this pattern suggests that inhibitory mechanisms have been up-regulated in layer II/III to balance the heightened excitatory tone inherent to an increase in AMPA receptors and a decrease in the NMDAR inhibiting post synaptic GABA_B receptor (Otmakhova and Lisman 2004).

The second pattern of subunit expression is seen in layer IV, V, and VI. In these layers, GluR1 is not significantly different from controls, while GluR2/3 remains significantly elevated. Like the results reported at 1 month, this suggests that the GluR2/3 containing AMPARs remain up-regulated. Unlike layer II/III, GABA_A α1 is not significantly different from controls while auto-regulatory presynaptic GABA_BR is significantly up-regulated. This suggests decreased GABA neurotransmitter release and subsequently lower inhibitory tone. At the same time the postsynaptic GABABR is significantly decreased, again suggesting disinhibition of the NMDAR (Otmakhova and Lisman 2004). This is not surprising considering that prolonged NMDAR activation promotes a significant endocytosis of these receptors (Terunuma et. al. 2010) and reorganization is an NMDA dependent process (Myers et. al. 2000). Taken together this pattern of subunit expression suggests that these layers are more plastic. That is, a decrease in inhibitory tone, specifically for NMDAR, and an increase in the excitatory mediating GluR2/3 AMPAR suggest that there is an increased probability of long-term potentiation within these neural networks (Davies et al., 1991; Davies and Collingridge, 1996; Mott and Lewis 1992; Stäubli, Scafidi, and Chun 1999).

Conclusion

In many sensory systems, traces of transient reorganization persist long after recovery (Froemke, Merzenich, and Schreiner 2007; Woolf et. al. 1995). At the same time, original somatotopic representations are retained in the reorganized cortex of humans with long standing nerve injuries (Shady 1994; Moore and Schady 2000 Halligan et. al. 1993). These original somatotopic representations are perceived during electrical activation of nerve stumps suggesting that the cortex retains traces of past plasticity. Here we provide evidence that the primary somatosensory cortex has layer specific patterns of plasticity related receptor expression after a regenerating nerve injury. It is possible that these persistent up and down regulations of receptor subunits facilitate or represent the preservation of neuroanatomical changes that occur during reorganization.

• Nerve compression injuries induce reorganization prior to reinnervation.

• GluR2/3 containing AMPAR are significantly elevated following reinnervation.

• Inhibitory tone returns following reinnervation.

• Presynaptic GABA autoregulation is significantly elevated in layers VI, V, and VI.

• Postsynaptic NMDAR inhibition is significantly reduced following reinnervation.