Hindlimb Representation in the Spiny Mouse Sensorimotor Cortex

ABSTRACT

The Cairo spiny mouse (Acomys cahirinus), a promising animal model in neuroscience, lacks a detailed neuroanatomical map of its cerebral cortex. This study aims to delineate the location and internal subdivision of the sensorimotor cortex, specifically the region controlling the hindlimb. For this purpose, retrograde tracing from the lumbar spinal cord using Fast Blue was combined with immunohistochemical characterization using neuronal markers (NeuN, SMI‐32, and calbindin 28 kDa) in adult animals. Fast Blue‐labeled neurons were identified in layer V within the cortical region defined as the sensorimotor cortex responsible for hindlimb control. Within this region, the general cytoarchitecture was defined using NeuN, which allowed visualization of all cortical layers. The border between the lateral and medial sub‐areas was clearly identified by SMI‐32 immunostaining, which showed denser labeling in the lateral zone (presumably the primary somatosensory cortex) compared to the medial zone (presumably the secondary motor cortex). Within the defined layers and sub‐areas, a comprehensive analysis of the calbindin‐expressing interneuron population was performed, revealing significant interareal differences in soma size and cellular density of calbindin‐positive neurons exclusively within layers III–IV. These findings provide the first detailed map of the hindlimb sensorimotor cortex in spiny mice, establishing a crucial neuroanatomical foundation for future studies using this novel model in sensorimotor research.

In the spiny mouse, the hindlimb sensorimotor cortex was identified using retrograde tracing by injecting Fast Blue into the lumbar spinal cord. In the identified area, NeuN labeling defined the general cytoarchitecture; SMI‑32 and calbindin 28 kDa labeling distinguished the lateral zone (primary somatosensory cortex) from the medial zone (primary motor cortex).

1. Introduction

The sensorimotor cortex is a large part of the neocortex crucial for multisensory processing and the capacity of mammals to learn and execute multiple movements (see Moreno‐López et al. ²⁰¹⁶; Stachowski and Dougherty ²⁰²¹). In mammals, it consists of the somatosensory cortex, responsible for the sensory processing related to the sense of touch, pressure, temperature, and proprioception (Krubitzer et al. 2011; Kaas et al. ²⁰¹⁸; Brooks and Cullen ²⁰¹⁹; O'Connor et al. ²⁰²¹), and the motor cortex, responsible for the motor control of the skeletal musculature (see Martin ²⁰⁰⁵; Strick et al. ²⁰²¹). Following injury to the sensorimotor cortex, an altered perception of sensory information or a complete loss of sensitivity, together with severe difficulties in motor control, has been reported (see Kearney and Lanius ²⁰²²; Asan et al. ²⁰²²).

Despite the more complex organization of the sensorimotor cortex in humans compared to that in rodents, the latter are one of the main animal models for the study of sensorimotor processing (see O'Connor et al. ²⁰²¹; Xu et al. ²⁰²²). As in other mammals, in rodents each subregion of the somatosensory cortex receives signals from a particular part of the body and is topographically mapped based on incoming signals; the total representation of the sensory body in the somatosensory cortex is known as “sensory homunculus” (Penfield and Boldrey 1937; Santiago et al. ²⁰⁰⁷). The following functional subdivisions of the somatosensory cortex have been identified in rodents: the primary (S1) and secondary (S2) areas (Krubitzer et al. 2011). The motor cortex is also topographically subdivided into functional regions according to the source of its motor output (Paxinos and Franklin 2019; Lee et al. ²⁰²²). However, the division of the motor cortex into subregions is under debate; in some studies, the primary motor (M1) and the secondary motor cortex (M2) have been reported (Allen Institute for Brain Science ²⁰¹¹; Paxinos and Franklin ²⁰¹⁹; Lee et al. ²⁰²²). In others, alternative names of motor cortical areas have also been used, including the agranular cortex (AGm), medial precentral cortex (PrCm), frontal orienting field (FOF), and so forth (Barthas and Kwan 2017; Lazari et al. ²⁰²⁴).

A combination of microstimulation and tracing techniques has revealed that in the sensorimotor cortex of rats and mice, the hindlimbs are represented caudomedially to the forelimb representation (Donoghue and Wise 1982; Sanderson et al. ¹⁹⁸⁴; C. X. Li and Waters ¹⁹⁹¹; Nandakumar et al. ²⁰²¹). A similar somatotopic organization has been observed in other rodents, such as squirrels (Cooke et al. 2012) and agouti (Rocha et al. 2007). As mentioned above, the multiple divisions/areas of motor cortex are debatable; for example, some authors have reported an absence of hindlimb representation within the AGm (secondary motor) area in rats (Wang and Kurata 1998).

Over time, new rodent species with specific biological features or genetic modifications have become intensively studied in neuroscience. One of such new models, the Cairo spiny mouse, or Acomys cahirinus (referred to in this article as the spiny mouse), possesses several distinctive biological traits, such as precocial development (Brunjes 1990), a menstrual rather than estral cycle (Bellofiore et al. 2017), and a highly developed healing ability without forming a fibrotic scar (Seifert et al. 2012; Santos et al. ²⁰¹⁶; Maden et al. ²⁰¹⁸). The spiny mouse is being used as a new model organism for scientific research in different fields; however, its neuroanatomy and neuromorphology are largely unknown. Recent studies reported the ability of spiny mice to perform quadrupedal stepping after complete transection of the spinal cord at the thoracic level (Nogueira‐Rodrigues et al. ²⁰²²). These data suggested restoration of functional connections between the lumbar spinal cord and hindlimb‑representing sensorimotor cortex in these animals. However, even the exact locations of cortical areas in spiny mice remain undetermined. Therefore, mapping the sensorimotor cortex representing the hindlimbs became the aim of the present study.

To delineate the location and subdivision of the sensorimotor cortex in the spiny mouse, two approaches were used: (1) direct labeling of the neuronal population projecting to the spinal cord region responsible for hindlimb control, using the retrograde tracer Fast Blue (FB), and (2) neurochemical markers for (i) the general neuronal population, using the hexaribonucleotide binding protein‐3 (NeuN) (Mullen et al. 1992), (ii) large pyramidal neurons, using the SMI‐32 antibody, a marker for nonphosphorylated domains of high‐molecular‐weight N200kDa neurofilaments (Fuentes‐Santamaria et al. ²⁰⁰⁶), and (iii) the population of cortical interneurons, using the calcium‐binding protein calbindin 28 kDa (CB) (DeFelipe et al. ¹⁹⁹⁹). Previously, both markers were successfully used for the parcellation of cortical areas in rodents (Van Brederode et al. 1991; Sun et al. ²⁰⁰²; Boire et al. ²⁰⁰⁵; Van Der Gucht et al. ²⁰⁰⁷; Van De Werd et al. ²⁰¹⁰), carnivores (Van Der Gucht et al. 2001), and primates (Nimchinsky et al. 1997; Lewis and Van Essen ²⁰⁰⁰; Atapour et al. ²⁰²⁴).

2. Methods

2.1. Animals

All experiments and animal care conformed to European Parliament Directive 2010/63/EU on the protection of animals used in scientific research, and animal care and experiments were also conducted in accordance with the guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Furthermore, the Ethics Commission of the Pavlov Institute of Physiology (Protocol #02/17; 17.02.2022) approved all experiments that were conducted on animals during this research. Animals for this study were taken from the colony created and supported by the Laboratory of Neuromorphology of the Pavlov Institute of Physiology. The animals were raised and maintained in accordance with the conditions described previously in Shkorbatova et al. (2024). In this research, we used a total of 10 adult animals of both sexes aged 15.5–16.5 months. All of them were used for histochemical mapping, and three of them were used for the tracing experiment.

2.2. Surgery and Tracing

To map sensorimotor cortex neurons associated with the hindlimbs, the retrograde tracer FB (Sigma, F5756, USA) was injected into the lumbar spinal cord enlargement of three spiny mice (Figure 1A). For this purpose, animals were initially anesthetized in a chamber containing 5% isoflurane (IsoNic, Vetpharma Animal Health, Spain) delivered in air at a flow rate of 1.0 L/min. Anesthesia was then maintained via a mouse mask with 1.5%–2% isoflurane throughout the surgical procedure. The fur on the back was shaved, and the surgery area was disinfected with 70% ethanol. Then the animals were transferred to the surgical table. All further procedures were performed under aseptic conditions. A skin incision was made in the lower thoracic region of the back to expose the vertebrae. A full laminectomy of the T12 vertebra was made to access the spinal segments L2–L3 (Veshchitskii et al. 2025). Dura matter was not removed.

FIGURE 1.

Fast Blue tracing of cortical areas connected to the lumbar spinal cord enlargement. (A) General scheme of the Fast Blue tracing: regions of unilateral tracer injection in the spinal cord are displayed as dots in different colors for individual animals (green: #43, yellow: #298, and magenta: #301). The tracer moves retrogradely into the brainstem, where it switches to the contralateral side and ascends to the corresponding cortical areas. (B) Example of the location of Fast Blue‐labeled cells in the cortex at different anteroposterior (AP) levels: each cell is displayed as a blue point; scale bar is 1 mm. (C) Stereotaxic map with delineated areas of Fast Blue‐labeled cells in individual animals (green: #43, yellow: #298, and magenta: #301). AP: 0 is the bregma point. The left panel is a diagram showing the AP distribution of the number of labeled cells per animal. Dashed lines indicate the location of cells in layer V; solid lines represent the surface area projection (along microcolumns) of layer V Fast Blue‐labeled neurons; the punctuated line represents the border between the medial and lateral regions (averaged across animals) as defined by calbindin labeling. (D) Examples of Fast Blue injection sites displaying the electrode track and transverse distribution of tracer in the spinal cord for individual animals. In (B) and (D), arrows indicate direction: D, dorsal; V, ventral; R, rostral; C, caudal; M, medial; L, lateral.

The tracer was injected into the spinal cord using a 5‐µL Hamilton syringe (Hamilton Company, Romania) fitted with a glass micropipette (inner diameter: ∼130 µm and outer diameter: ∼450 µm). Three unilateral injections of 0.2 µL each (1.0% in distilled water) were administered on the right side of the spinal cord equidistantly at rostral, middle, and caudal regions. Injection coordinates were as follows: 0.8 mm from the spinal cord surface (dorsoventral coordinates) and 0.5 mm from the midline of the blood vessel of the spinal cord (mediolateral coordinates), targeting the central region of the intermediate gray matter according to our stereotaxic scheme (Veshchitskii et al. 2025). After injection, the micropipette was kept in the spinal cord for 1 min.

Following the injection, the paravertebral muscles were sutured with PGA #5 (Lintex, Russia), and the skin was sutured with Ethilon Polyamide #6 (Ethicon, USA). After surgery, animals were allowed to recover from anesthesia in a warm chamber before being returned to their home cages. Postoperative analgesia was maintained via subcutaneous administration of ketorolac (Ketanov, S.K. Therapy S.A.) twice a day for 3 days in a dosage of 1 mg/kg, and an antibiotic (Vetbicin‐3, Zoopharm; 360,000 ED/kg) was administered subcutaneously once after surgery. The animals were allowed to survive 6 days after the intraspinal injections.

2.3. Transcardial Perfusion and Tissue Preparation

Prior to perfusion, all animals were deeply anesthetized with a mixture of Xyla (2 mg/kg) (Interchemie werken “De Adelaar” BV, Netherlands) and Zoletil (20 mg/kg) (Virbac, France), and a sufficient degree of anesthesia was determined by the absence of any reaction and even by the absence of increasing respiratory rate to strong compression of the finger or the pinna with tweezers. Thereafter, transcardial perfusion was conducted with 0.9% sodium chloride (Ecoservice, Russian Federation), followed by 4% buffered paraformaldehyde (Sigma–Aldrich, USA). Subsequently, the brains and spinal cord were removed from the skull and vertebral column, respectively, postfixed overnight in 4% buffered paraformaldehyde, and then sequentially immersed in 20% and 30% sucrose solutions for cryoprotection until they sank. A spinal cord segment was defined as the region between the caudal‑most parts of the dorsal rootlet attachment zones connected to the neighboring dorsal root ganglia. Using a freezing microtome (Reichert, Austria), a continuous series of 40‑µm coronal brain slices and 50‑µm transverse spinal cord slices was prepared. All slices were collected in 24‐well plates in PBS with 0.1% NaN₃, washed three times for 10 min each on the rotary shaker under gentle agitation in the same solution, and stored at 4°C until application.

2.4. Immunohistochemistry

For immunohistochemical (IHC) staining, free‐floating tissue slices were first washed in phosphate‐buffered saline (PBS) (Rosmedbio, Russian Federation) (pH 7.4) three times: for 0.5, 5, and 10 min, respectively, at room temperature (RT). The same washing regimen was applied to all subsequent steps of the IHC procedure. Endogenous peroxidase activity was then inhibited by 0.3% hydrogen peroxide in distilled H₂O (dH₂O) for 30 min at RT. After washing, nonspecific antigens were blocked by incubating in 3% bovine serum albumin (Biolot, Russian Federation) for 60 min at RT. Without washing, the brain slices were then treated with individual primary monoclonal antibodies against NeuN, SMI‐32, and CB (Table 1) for 72 h under gentle agitation on the shaker at 4°C. After washing, the slices were incubated with biotinylated secondary antibodies (Table 1) for 24 h while agitating at 4°C. Following washing, the slices were treated with horseradish conjugated avidin–biotin–peroxidase complex (ABC Elite system, Vector Laboratories, USA) for 60 min at RT and washed again. The final revealing was performed in 0.03% hydrogen peroxide solution using diaminobenzidine (Sigma Aldrich, USA) as a chromogen with nickel ammonium sulfate intensification for 2–5 min at RT under visual microscopic control of reaction quality. The slices were then washed in dH₂O, mounted on gelatinized glass slides, dehydrated in ethanol of ascending concentrations, cleared in xylene, and coverslipped using Bio Mount mounting media (Bio Optica, Italy).

TABLE 1.

The list of antibodies used.

Antibody	Dilution	Source animal	Target molecule	Manufacturer and cat. #
Anti‐NeuN	1:3000	Mouse	Hexaribonucleotide binding protein‐3	Millipore, USA; cat. # MAB377; RRID: AB_2298772
SMI‐32	1:15,000	Mouse	Nonphosphorylated domains of high‐weight neurofilaments	BioLegend, USA; cat. # 801701; RRID: AB_2564642
Anti‐calbindin‐D‐28K	1:10,000	Mouse	Calcium‐binding protein 1 (28 kDa)	Sigma–Aldrich, USA; cat. # C9848; RRID: AB_476894
Anti‐mouse IgG	1:600	Horse	Primary antibodies	Vector Labs, Inc., USA; cat. # BA‐2000; RRID: AB_2313581

2.5. Image Acquisition and Analysis

Images of wet‐mounted brain slices with Fast Blue‐labeled (FB+) neurons were acquired using a fluorescence microscope EVOS M5000 with 20× objective (Thermo Fisher Scientific, USA). Images of brain slices after the IHC reaction and of unstained wet slices of the spinal cord for reconstruction of tracer injection sites were acquired using the following setup: Olympus CX31 light microscope with 4×, 10×, and 20× objectives (Olympus Corporation, Japan), VideoZavr Standart VZ‐18C23‐B camera (VideoZavr, Russian Federation), and a personal computer with VideoZavr Scan 2.4 software package (VideoZavr, Russian Federation).

The FB+ region was determined by examining every third unstained slice using a fluorescent microscope, and at least 10 slices containing FB+ cells were included in the analysis. For immunohistochemistry, three to five slices through the FB+ region were used for NeuN, SMI‐32, or CB staining. Slices for all stainings were adjacent to those selected for FB analysis. The slices for different antibodies were manually aligned using Fiji ImageJ software (ver. 1.54f) (Schneider et al. 2012); the curvature of the cortex and underlying brain structures were used for the overlap (see details below). Morphometric analysis for immunohistochemically labeled cells (neuronal soma area and cellular density), as well as quantification of FB+ cells, was manually performed using the open‐source Fiji ImageJ software. In slices with FB, the absence of nonspecific autofluorescence was checked by photographing the same slices using different excitation (Ex) / emission (Em) wavelengths: Texas Red (Ex: 585/29; Em: 628/32), GFP (Ex: 482/25; Em: 524/24), and DAPI (Ex: 357/44; Em: 447/60). FB+ cells were observed only with the DAPI cube.

To align the cortical region containing FB+ neurons projecting to the lumbar spinal cord on a dorsal‐view stereotaxic map, the reconstruction was performed according to the following algorithm. Individually for each animal, coronal fluorescent images of the brain were stacked and aligned using the midline and dorsal surface of the brain as reference points. Then, on each image, the distances from the midline to the nearest and farthest FB+ cells were measured. As a result, the distances between the obtained values reflected the mediolateral distribution of FB+ cells at their location depth (layer V) in each brain slice. Additionally, the projection of the obtained coordinates onto the cortical surface was performed along the cortical minicolumns. Then, all points were reproduced and connected in Fiji ImageJ, and dorsal‐view cortical maps were obtained for both the actual location of FB+ cells in layer V (shown as dashed lines in Figure 1C) and their projection onto the cortical surface (shown as solid lines in Figure 1C). For the latter, the border between the medial and lateral regions was delineated using the aligned CB‐labeled FB+ slices in the middle and extremums of the FB+ region; the schematic averaged border line is shown as a dotted line in Figure 1C. The direct colocalization of FB and IHC labeling was performed for three animals; these superpositions were used as reference patterns. For the remaining animals, histochemical mapping was conducted based on the above‐mentioned reference patterns.

To calculate the AP coordinates of the FB+ region, five additional animals were used. In these animals, two tracks were made 2 mm lateral to the midline at the AP levels of bregma and lambda using a steel needle (0.35 mm in diameter) and a stereotaxic apparatus (RWD, China). After the correspondence of bregma and lambda levels to the basic brain structures on the coronal brain slices was repeatedly established, one of these animals was used as a reference animal for the present work. In this animal, periodic unstained photographs of coronal slices were taken, and their precise AP coordinates were calculated. The AP coordinates of the FB+ area in the experimental animals were established by direct comparison of the unstained slices from the experimental animals with those from the reference animal. The additional reference group was created to avoid the potentially significant impact of needle tracks (which are required for brain AP coordinate definition) in the experimental group.

To reconstruct the spinal segments containing the tracer injection sites, images of unstained wet transverse spinal cord slices of segments L1–L3 were aligned using the central canal and the borders of the white and gray matter as reference points, then stacked and saved as a TIFF file. Using the “Orthogonal Views” function (Ferreira and Rasband 2012) of Fiji ImageJ, the injection sites were reconstructed in horizontal and sagittal planes. The precise rostrocaudal location of these sites is displayed as dots in the lumbar segments in the spinal cord scheme in Figure 1A. The injection sites were defined as vertical tracks filled with a yellow solution of Fast Blue, clearly visible on the slices. Afterward, the presence of tracer in these tracks on these slices was verified under a fluorescence microscope as maximally bright zones.

2.6. Statistics

Data are presented as mean ± SD and were processed in the GraphPadPrism ver.8 (GraphPad, USA) software. Normality was assessed using the Kolmogorov–Smirnov test. The Wilcoxon test (for paired intraareal comparisons) and the Mann‐Whitney U test (for unpaired interareal comparisons) were used.

3. Results

3.1. Retrograde Labeling of Cortical Neurons Projecting to the Lumbar Spinal Cord

Since only the approximate location of the sensorimotor cortex (with no indications of exact borders) has been established for the spiny mouse (Vitorino et al. 2022), the approximate locations of the somatosensory and motor cortices were estimated using the Allen Mouse Brain Atlas (Allen Institute for Brain Science ²⁰¹¹). To map the region of the sensorimotor cortex responsible for the hindlimbs, we first assessed the location of neurons that project directly to the lumbar spinal cord using retrograde labeling.

Consistent with previous data (Akintunde and Buxton ^{1992a, 1992b}), unilateral spinal cord injection of a retrograde tracer (Figure 1D) resulted in predominant neuronal labeling within the contralateral cortex (Figure 1B), a pattern attributed primarily to the crossing dorsal corticospinal tract. In contrast, only weak labeling was observed in the ipsilateral cortex, corresponding either to projections from the noncrossing ventral corticospinal tract or to tracer leakage into the second half‐cord. In the present study, exclusively the contralateral labeling was analyzed.

Within the contralateral sensorimotor cortex, FB+ neurons were revealed at the rostrocaudal levels AP(−0.6)–AP(+2.1). As shown in Figure 1C, the number of labeled cells was dependent on the rostrocaudal location of the slice, being maximal around the AP(0.5)–AP(+1.5) level. Along the mediolateral axis, these cells occupied a 1.5–3.0 mm region from the midline. Note that because the brightness of FB+ neurons was dependent on the microscope settings applied individually for each slice, we omitted an analysis of cell area.

As a result of the first task, the region of the sensorimotor cortex responsible for hindlimb control was delineated. The next task was to distinguish between the somatosensory and motor subregions using a combination of two neurochemical markers.

3.2. NeuN Staining

Despite the availability of well‐developed techniques for revealing both areal and layered subdivision of the cerebral cortex, general visualization of basic cortical cytoarchitecture using Nissl staining or its analogs remains a useful tool in neuroscience (Palomero‐Gallagher and Zilles ²⁰¹⁹; Meystre et al. ²⁰²⁴). In the present study, we used NeuN staining, which allows visualization of most neuronal cells.

Using NeuN labeling, most cortical layers are distinguishable: (1) layer I, containing only solitary labeled neuronal somas; (2) a band including layers II–III, containing neuronal somas of multiple shapes; (3) layer IV, containing distinctly smaller neuronal somas, including a population of very small, round‐shaped cells; (4) layer V, containing neuronal somas of multiple shapes but predominantly large pyramidal neurons; and (5) layer VI, containing neuronal somas of multiple shapes, generally smaller than those in layer V (Figure 2A). NeuN staining did not allow for a clear subdivision of the area containing FB+ cells into lateral and medial parts, but numerous very small, round‐shaped neurons within the lateral part of layer IV were clearly visible, making it possible to distinguish between the granular and dysgranular areas (Figure 2A).

FIGURE 2.

Immunohistochemical labeling of the Fast Blue‐traced cortical zone. NeuN (A), calbindin (B), SMI‐32 (C), and Fast Blue (D) labeling on adjacent slices are shown. In the upper panels, labeled coronal slices are subdivided into cortical areas 29/30, lateral (Lat) and medial (Med) sub‐areas, and barrel cortex. In the magnified lower panels, cortical layers I–VI and the borders between Lat/Med and Med/30 are shown. The scale bars are 1000 µm for upper panels and 400 µm for lower panels.

Medial to the medial sub‐area, two regions are visible: granular and agranular regions; at the rostrocaudal levels caudally from AP1.5, these correspond to areas 30 and 29 of the retrosplenial cortex (Merkulyeva et al. 2026) (Figure 2A). Lateral to the lateral sub‐area, barrel‐like structures were revealed (Figure 2A).

3.3. SMI‐32 Staining

Using SMI‐32 labeling, the area containing FB+ cells could be subdivided into five horizontal bands: (1) layer I, containing no labeled somas and no neuropil staining; (2) a band including layers II–III, containing densely stained neuropil and large, intensively stained neuronal somas; (3) a band including layer IV, containing no labeled neuronal somas but clearly visible neuropil staining; and (4) a band including layer V and the upper part of layer VI, containing large, intensively stained neuronal somas and neuropil (Figure 2B). Note that most neuronal somas within all layers were obscured by the intensely stained, dense neuropil. As a result, in the present study, SMI‐32 staining was used not for quantitative counting but for the qualitative subdivision of the area containing FB+ cells. According to it, a well‐visible border between the granular (lateral) and dysgranular (medial) sub‐areas can be seen. The lateral sub‐area had more densely packed neuronal somas and darker‐stained neuropil within layers II–III and V compared to the medial sub‐area (Figure 2B).

Medial to the medial sub‐area, two regions are visible: both exhibiting strong SMI‐32 labeling in layers II–III and V–VI. Furthermore, in the more medial region, at rostrocaudal levels caudal to AP1.5 and corresponding to the retrosplenial cortex, well‐defined bundles formed by labeled neuronal processes can be seen (Merkulyeva et al. 2026) (Figure 2B). Lateral to the lateral sub‐area, SMI‐32 staining was generally similar but contained a narrow zone with weaker SMI‐32 staining in layer II (Figure 2B).

3.4. CB Staining

3.4.1. General Pattern of Staining

CB labeling allowed to divide the area containing FB+ cells into four horizontal bands: (1) layer I, containing no labeled cells but showing neuropil staining; (2) a band including layer II and the upper part of layer III, containing small, densely packed neurons; (3) a band including the lower part of layer III and all of layer IV, containing or not containing very small, densely packed cells (note that this band was clearly evident only within the lateral sub‐area); and (4) a band including layers V and VI, containing large, rare neurons (Figure 2C).

Medial to the medial sub‐area, two regions are visible: a more medial region with extremely weak CB staining, and a more lateral region with staining similar to that in the medial sub‐area but with a lower number of stained cells in layers V–VI (at rostrocaudal levels caudally from AP1.5, these regions correspond to areas 30 and 29 of the retrosplenial cortex [Merkulyeva et al. ²⁰²⁶]) (Figures 2C and 3A,B). Lateral to the lateral sub‐area, a sharp increase in the number of small CB‐immunopositive (CB+) neurons in layer IV was observed, and barrel‐like structures were revealed (Figures 2C and 3A,B). The border between these sub‐areas closely corresponded to that revealed by SMI‐32 staining (Figure 2B).

FIGURE 3.

Calbindin‐positive immunohistochemical labeling of the lateral and medial cortical sub‐areas. (A) Border region between lateral (on the right) and medial (on the left) cortical sub‐areas with adjacent NeuN labeling. Plates 1–5 in (A) are magnified in (B); I–VI, cortical layers; WM, white matter. The graphs of the cellular density (С) and soma area (D) of calbindin‐positive cells. *p < 0.05. The scale bars are 400 µm for (A) and 40 µm for (B).

According to all these comparisons, most of the FB+ cells occupied the lateral part of the cortex (Figure 2D).

3.4.2. Soma Area of the CB‐Stained Neurons

The soma area of CB+ cells was analyzed in all layers, excluding layer I, due to the very low number of cells. In the medial sub‐area, the average soma area of CB+ cells was 147–187 µm² in layers II–III and 102–138 µm² in layers V–VI. In the lateral sub‐area, an average soma area was 151–180 µm² in layers II–III, 79–123 µm² in layers III–IV, and 114–183 µm² in layers V–VI. Within the lateral sub‐area, the soma area in layers II–III and V–VI was significantly larger than that in layer IV (p = 0.0156, p = 0.0156). Within the medial sub‐area, significant differences were observed between layers II–III and V–VI (p = 0.0156). At the same time, no interareal differences were observed for either layers II–III (p = 0.7104) or V–VI (p = 0.4557).

3.4.3. Cellular Density of the CB‐Stained Neurons

In the medial sub‐area, the average cellular density of CB+ cells was 0.37–0.62 cells/µm² in layers II–III and 0.02–0.03 cells/µm² in layers V–VI. In the lateral sub‐area, an average cellular density was 0.44–0.73 cells/µm² in layers II–III, 0.47–1.1 cells/µm² in layers III–IV, and 0.02–0.04 cells/µm² in layers V–VI (Figure 3C). Within the lateral sub‐area, the cellular density in layers V–VI was significantly lower than that in layers II–III and IV (p = 0.0156, p = 0.0156). Within the medial sub‐area, cellular density in layers V–VI was also significantly lower than that in layers II–III (p = 0.0156). At the same time, no inter‐areal differences were observed for either layers II–III (p = 0.2593) or V–VI (p = 0.3829).

4. Discussion

In the present study, injection of the retrograde tracer FB into the lumbar enlargement region (segments L2–L3) of the spinal cord in the spiny mouse resulted in FB+ cells in layer V of the cortex at the rostrocaudal levels AP(‐0.6)–AP(+2.1) and within the mediolateral extent between the cingulate and barrel cortex. The location of this region corresponds well to the sensorimotor region controlling the hindlimbs in rats (Barth et al. 1990; Akintunde and Buxton ^{1992a, 1992b}; Khazipov et al. ²⁰¹⁵; Olivares‐Moreno et al. ²⁰¹⁷) and in mice (Paxinos and Watson 1998; Allen Institute for Brain Science ²⁰¹¹; Kamiyama et al. ²⁰¹⁵). According to Olivares‐Moreno et al. (2017), neurons projecting to the lumbar spinal cord were restricted to areas corresponding to the medial part of M1 and hindlimb area S1. Miller (1987) reported that labeled cells were located mainly along the border of somatosensory areas 3 and 4, with full coverage of the motor area. Both Olivares‐Moreno et al. (2017) and X. G. Li et al. (1990) reported an absence of lumbar‐projecting neurons in area S2. In the present study, no labeled cells were found in the lateral regions where S2 can be situated.

Taking into account that in rodents the somatosensory areas occupy a more lateral position relative to the motor areas (see brain atlases for the house mouse, Mus musculus [Paxinos and Watson ¹⁹⁹⁸; Allen Institute for Brain Science ²⁰¹¹], the brown rat, Rattus norvegicus [Khazipov et al. ²⁰¹⁵], and the Mongolian gerbil, Meriones unguiculatus [Radtke‐Schuller et al. ²⁰¹⁶]), the medial sub‐area in the spiny mouse can be considered a motor region and the lateral sub‐area a somatosensory region. However, the question remains: What specific cortical areas are labeled by the FB injection? Are they S1, M1, or M2 ones? To address this, we superimposed the pattern of the FB+ cells with patterns of the CB and SMI‐32 staining.

A dense SMI‐32 labeling was revealed in the motor and sensory areas of rats, with clearly visible bands of pyramidal cells in layers III, V, and VI (Hiscock et al. 1998; Kirkcaldie et al. ²⁰⁰²; Boire et al. ²⁰⁰⁵); an abrupt change in this pattern at the transition with the cingulate/retrosplenial regions was also observed (Kirkcaldie et al. 2002; Boire et al. ²⁰⁰⁵). A similar transition was observed in spiny mice in our previous study (Merkulyeva et al. 2026). In the present study, we also identified a clear transition between two sub‐areas containing FB+: the lateral sub‐area exhibited a dense band of SMI‐32 staining in layers II–III, which was nearly absent in the medial region. In a study of the hamster neocortex, Boire et al. (2005) reported intense SMI‐32 staining in layers III, V, and VI of area S1 and weak staining in area M2. Moreover, authors reported a presence of the clear visible border between areas M1 and M2 (Boire et al. 2005). Regarding the S1–M1 transition, they noted denser SMI‐32 staining in area S1. In rats, a significant reduction of SMI‐32 staining was revealed in area M2, but similarly dense staining was reported for areas S1 and M1 (Kirkcaldie et al. 2002). Weak SMI‐32 staining in area M2 was also reported in mice (Van Der Gucht et al. 2007). Therefore, we propose that the weakly stained medial area in our study corresponds to area M2. However, which area is located laterally: S1 or M1?

The most striking feature distinguishing neighboring somatosensory and motor areas is the presence of layer IV. Area S1 (corresponding to Brodmann area 3) is granular, with a well‐defined layer IV, whereas the motor cortex (corresponding to Brodmann area 4) is agranular, lacking a distinct layer IV (Krubitzer et al. 2011). However, in rats, the hindlimb motor representation overlaps with the sensory representation within the granular area S1 (Hall and Lindholm 1974; Donoghue and Wise ¹⁹⁸²); this region contains both densely packed granule cells in layer IV and large pyramidal cells in layer V (Tennant et al. 2011). In NeuN‐stained slices, a well‐defined granular layer is visible in the lateral sub‐area, which also contains well‐defined pyramidal neurons, whereas no granular cells are seen in the medial sub‐area. These data support the identification of the medial and lateral sub‐areas as motor and somatosensory areas, respectively.

Using CB as a marker for interneurons (see DeFelipe ¹⁹⁹⁷), we observed several bands of CB staining similar to those previously described in the rat cortex (Van Brederode et al. 1991; Staiger et al. ²⁰⁰⁴). However, no distinct differences in CB staining between motor and somatosensory areas have been reported in rats (Van Brederode et al. 1991). Similarly, in the cerebral cortex of mice, no differences in the population of the CB+ cells between motor and somatosensory cortex were found (Carretta et al. 2003, Carretta et al. ²⁰⁰⁴). Notably, Carretta et al. (2003) and Staiger et al. (2004) did not document a clear band of small labeled cells in layer IV. In contrast, we identified a clear border between two sub‐areas based on the presence of small CB+ cells within the granular layer IV of SS1, which was also shown in rats but probably manifested itself to a lesser extent (Sánchez et al. 1992; Sun et al. ²⁰⁰²). The presence of layer IV is a classic morphological delimiter of the SS1/M1 border, which is used in anatomical atlases of rats (Swanson 2018) and mice (Allen Institute for Brain Science ²⁰¹¹). Several interspecies differences in somatosensory cortex morphology have been reported. For example, in rats, unlike mice, a specific population of barrels (the posteromedial barrels) is filled with small neurons throughout layer IV (Welker and Woolsey 1974). Therefore, we propose that the small CB+ cells observed in layer IV of the spiny mouse sensorimotor cortex may be specific to this species.

Recently, Lazari et al. (2024), using a combination of techniques (axonal tracing, functional MRI, myelin mapping, gene expression, and optogenetics), proposed a new division of the motor cortex in mice. The authors identified the location and structure of three new areas: anterior–lateral motor cortex (ALM), anterior–lateral M2 (aM2), and posterior–medial M2 (pM2). Together with M1, these areas only partially overlap the well‐known areas M1 and M2 presented in multiple references, including Paxinos Mouse Brain Atlas (Paxinos and Franklin 2019) and Allen Mouse Brain Atlas (Allen Institute for Brain Science ²⁰¹¹). According to this new scheme, all four motor areas are visible at the level of striatum (Lazari et al. 2024). If a similar organization exists in the cerebral cortex of the spiny mouse, the border between the lateral and medial sub‐areas could correspond to a border between different motor regions. However, because the lateral sub‐area clearly contains a granular layer, we are more confident that the border identified here separates somatosensory and motor areas. Recent data on the rat sensorimotor cortex also indicate a more complex organization—in particular, selective loci responsible for hip flexion, retraction, and extension were revealed within area M1, which is responsible for the hindlimbs (Halley et al. 2020). It has also been shown that selective loci of the motor cortex receiving projections from particular hindlimb muscles (Maurer et al. 2023) and selective loci of the somatosensory cortex receiving projections from particular trunk dermatomes (Nandakumar et al. 2021) exist within the rodent somatosensory cortex. Based on these data, we suppose that the sensorimotor cortex of the spiny mice may have a similarly complex organization and that this may be reflected in the cytoarchitectonic structure revealed in the present study.

4.1. Limitations

Due to the limited sample size, potential sex differences in the structure of the somatosensory cortex of the spiny mouse were not assessed.

Author Contributions

Conceptualization: Natalia Merkulyeva. Data curation: Natalia Merkulyeva. Formal analysis: Anton Beljajev, Aleksandr Veshchitskii, and Aleksandr Mikhalkin. Investigation: Anton Beljajev, Polina Shkorbatova, Aleksandr Veshchitskii, and Aleksandr Mikhalkin. Methodology: Anton Beljajev, Polina Shkorbatova, Aleksandr Veshchitskii, and Aleksandr Mikhalkin. Project administration: Natalia Merkulyeva. Resources: Natalia Merkulyeva. Software: Anton Beljajev, Aleksandr Veshchitskii, and Aleksandr Mikhalkin. Validation: Natalia Merkulyeva. Visualization: Anton Beljajev, Aleksandr Veshchitskii, Aleksandr Mikhalkin, and Natalia Merkulyeva. Writing – original draft: Anton Beljajev, Aleksandr Veshchitskii, Aleksandr Mikhalkin, and Natalia Merkulyeva. Writing – review and editing: Anton Beljajev, Aleksandr Veshchitskii, Aleksandr Mikhalkin, Natalia Merkulyeva, and Polina Shkorbatova.

Funding

The study was supported by the State funding allocated to the Pavlov Institute of Physiology, Russian Academy of Sciences (No. 1021062411653‐4‐3.1.8).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The original contributions presented in this study are included in the article and its Supporting Information; further inquiries can be directed to the corresponding author.