Introduction

Similar to other primates1,2,3, in the marmoset monkey, the primary motor cortex (M1, Brodmann Area 4) is formed by distinct cytoarchitectural subdivisions. According to the contemporary view, the representations of the limb and body axial musculatures are encompassed within cytoarchitectural area 4ab, located medially, whereas the representations of the head musculature (including facial and oral) are located laterally, in cytoarchitectural area 4c4,5. In particular, A4ab contains the unique “gigantopyramidal” neurons, known as Betz cells in primates6,7,8. In contrast, in A4c, layer 5 is populated by smaller pyramidal neurons4. Betz cells are characterised by their pyramidal shape, large size, and prominent apical dendrites that are oriented along a vertical axis. Although more sparsely distributed, they can be up to 20 times larger in volume when compared to other pyramidal cells in humans9.

Betz cells have specialised molecular and physiological characteristics6. Given their distinctive action-potential properties10,11 and direct connectivity onto the ventral horn of the spinal cord12, they are well placed for initiation and modulation of fine movement. Betz cells can be identified using markers of pyramidal neurons, such as antibodies against non-phosphorylated neurofilament (e.g. SMI-32)13,14, as well as general markers of neuronal nuclei (e.g. NeuN)14. In addition, some of these cells specifically express nitric oxide synthase (NOS)15 and/ or the calcium-binding protein parvalbumin (PV)14. However, no single molecular criterion has been shown yet that distinguishes adult Betz cells from other layer 5 extratelencephalic projection neurons16.

There is a large body of evidence to indicate that the local vasoactive intestinal polypeptide-positive (VIP+) interneurons exert inhibition onto excitatory pyramidal cells, either directly or indirectly, to shape their outputs17,18,19,20,21,22. Here, we present evidence that staining M1 with antibodies against VIP reveals not only a subset of GABAergic interneurons, as expected from previous studies23,24,25,26,27, but also Betz cells. The spatial and laminar distributions of VIP+ interneurons and Betz cells are compared.

Materials and methods

Animals

The study is reported in accordance with ARRIVE guidelines. Materials were sourced from nine healthy adult marmosets (Callithrix jacchus) as detailed in Table 1. The experiments were conducted according to the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. All experimental protocols were approved by the Monash University Animal Ethics Committee. The animals had no veterinary record of serious acute or chronic health conditions. Some of the animals were also used in an anatomical tracing experiment unrelated to the present study; for details of the tracer experiments, see28,29.

Table 1 Details of sex, age, and staining.
Full size table

Tissue perfusion and processing

The marmosets were euthanized with an overdose of sodium pentobarbitone (100 mg/kg, i.v.) and transcardially perfused with 0.1 M heparinized phosphate buffer saline (PBS; pH 7.2), followed by 4% paraformaldehyde (PFA) in 0.1 M PBS. The brains were dissected and post-fixed in 4% PFA for one hour prior to immersion in PBS, with increasing concentrations of sucrose (10%, 20%, and 30%), over several days at 4 °C. The one-hour post-fixation was a critical step for VIP visualisation. Longer post-fixation times were not successful unless heat-induced antigen retrieval steps were used which produced variable degrees of success. For consistent staining across cases, we exclusively used one-hour post-fixation for data presented in this study. Using a cryostat, frozen 40 μm coronal sections were obtained. Five sequential series were collected, with the first two being stained for myelin30 (modified version31), and for cell bodies (Nissl stain). These series were used to delineate boundaries of the motor cortex4. The other series were used for immunostaining, as detailed in Table 1. In cases also used for anatomical tracing, the sections used for the present analyses were sourced from the hemisphere contralateral to the tracer injections.

Immunostainings were conducted by incubating sections in blocking solution (10% normal horse serum and 0.3% Triton X-100 in 0.1 M PBS) for 1 h at room temperature before undergoing incubation in primary antibody at 4 °C for 42–46 h (Table 2). For immunohistochemistry using DAB (IHC), a biotinylated horse anti-mouse IgG secondary antibody incubation was applied for 30 min, followed by treatment with Avidin–Biotin Complex (ABC) reagent and DAB (3,3'-Diaminobenzidine) substrate working solution. Alexa Fluor secondary antibodies were used for immunofluorescence staining (IF) (Table 2).

Table 2 List of primary and secondary antibodies used for immunostaining.
Full size table

To test for the specificity of anti-VIP primary antibodies, we ran control immunohistochemistry experiments in which either the primary antibodies were removed (Supplementary Fig. S1) or pre-absorbed with the VIP peptide (absorption test, Supplementary Fig. S2). Pre-absorbed VIP primary antibody was made through incubation with the VIP peptide (Cat# ABIN2958145) in the blocking solution for 24 h at 4 °C (molar ratio of peptide to antibody; 100: 1 for Cat# L108/92 and 400:1 for Cat# HPA017324). Sections were incubated in the pre-absorbed antibody solution for 42–46 h, followed by the application of secondary antibodies as described above (Supplementary Fig. S2).

Subsequently, Aperio Scanscope AT Turbo (Leica Biosystems) was used to scan myelin, Nissl, and IHC stained sections. Fluorescent images from the motor cortex were captured using a Nikon C1 laser scanning confocal microscope.

Quantification of VIP+ neurons and statistical analysis

The three distinct regions of the primary motor cortex, namely areas 4a, 4b (encompassing representations of the axial and limb musculatures), and 4c (the representation of the head musculature), along with the caudal subdivisions of the dorsal premotor area (A6DC) were delineated using previously defined criteria4,32.

Four VIP-stained sections were selected from each subdivision of the primary motor cortex from three cases that underwent IHC (CJ226, CJ227, and CJ240). Using standard features of the Aperio Image Scope software, square counting frames (150 × 150 μm) were aligned across the cortical thickness. The uppermost frame was aligned at the border of layers 1 and 2. The frames were equidistant (10 μm apart) and covered all cortical layers up to the white matter33. The number of counting frames varied depending on the thickness of the cortex in each section. Each counting frame was bounded by 2 exclusion lines (left and bottom) and 2 inclusion lines (right and top), as previously described33. A cell was counted if it was completely within the counting frame or if the cell body contacted the inclusion lines.

The number of counted cells per frame was averaged in each section and used to calculate the neuronal density. The section thickness (40 μm) and a shrinkage factor of 0.801 obtained in a previous study34 were considered when calculating the number of cells per volume unit. For the analysis presented in Fig. 5e, we measured the VIP fluorescence intensity of 8 neurons from each type (Betz cell or interneuron) in one section as a relative quantity, using built-in functions of ImageJ software. We further calculated the corrected total fluorescence [CTF = integrated density − (area of selected cell \(\times\) mean fluorescence background)] (Fig. 5f), to best assess the amount of VIP expression in these two cell types. Integrated density is the sum of the values of the pixels in the selected part of the image, i.e. cell soma. Mean fluorescent background was obtained by averaging the value of fluorescence intensity across random points in the area between cells.

Statistical analysis was conducted using GraphPad Prism v9.2.0 (Graph-Pad Software, La Jolla, CA, USA) and involved the use of unpaired t-test or one-way ANOVA followed by post hoc Tukey’s tests. Statistically significant differences were those with a p-value < 0.05. Figures were created using Adobe Illustrator.

Results

Immunostaining against VIP in the motor cortex led to the visualisation of giant Betz cells of layer 5, as well as a subset of putative inhibitory neurons (Fig. 1). While the expression of VIP in interneurons has been demonstrated in other species22,23,24,25,35, our data also indicates weak VIP expression in Betz cells, which was confirmed using three different antibodies and validated by control experiments involving the removal of the primary antibody (Table 2 and Supplementary Fig. S1) and absorption test (Supplementary Fig. S2). The VIP antibodies used in different analyses are given in the corresponding figure legends.

Fig. 1
figure 1

Vasoactive intestinal polypeptide (VIP) is expressed in giant pyramidal (Betz) cells of primary motor cortex. Representative images of primary motor cortex (M1, Area 4a and Area 4b) are shown using myelin (a), Nissl (b) and VIP staining (c) in CJ226. Arrowheads point to the areal boundaries5. Red lines identify layers 2 to 6, while layer 5 is indicated by small black lines in (a–d). Schematic of stained layer 5 Betz cells is indicated in (d). Dashed rectangles in (b,c) are shown at higher magnification in (e,f) in which Betz cells are clearly visible within layer 5 (between the two dashed lines). Solid rectangles in (b,c) are magnified in (g,i) representing A4a and in (h,j) representing A4b. Red arrows point to Betz neurons. Scale 1 mm for (a–d), 200 µm for (e,f) and 100 µm for (g–j). VIP antibody used in production of images is from Abcam (Cat# ab30680).

Full size image

Betz cells

VIP staining intensity was strongest in the largest layer 5 pyramidal cells of A4ab, which are located more medially in this area, as defined by Nissl and myelin staining (Fig. 1a–c)5. This region corresponds to the representations of the leg and trunk musculatures4. The locations of stained Betz cells are shown by black circles in Fig. 1d. VIP staining intensity in Betz cells diminishes toward the more lateral parts of A4b (Fig. 1e–j). The cell body and apical dendrites of Betz cells were clearly visible in the VIP staining, similar to the observations with Nissl staining, with a more pronounced delineation of the apical dendrites (Fig. 1e–j). There was no VIP staining in layer 5 pyramidal cells of cytoarchitectural areas A4c and 6DC (Fig. 2), which lack the gigantopyramidal characteristics4,32.

Fig. 2
figure 2

Expression of vasoactive intestinal polypeptide (VIP) is limited to the largest layer 5 pyramidal cells. Representative images of Area 4c (a–e) and A6DC (caudal subdivisions of the dorsal premotor area, f–j) are shown using myelin (a,f), Nissl (b and g) and VIP staining (c and h) in CJ226. Arrowheads point to the areal boundaries5. Red lines in (a–c and f–h) identify layers 2–6, while layer 5 is indicated by small black lines. Dashed rectangles in (b,c,g,h) are shown at higher magnification in (d,e,i,j), respectively. Large pyramidal cells are visible within layer 5 (between the two dashed lines) only in Nissl staining (d,i). Scale 1 mm for (a–c) and (f–h), 200 µm for (d–e) and (i–j). VIP antibody used in production of images is from Abcam (Cat# ab30680).

Full size image

We confirmed the weak presence of VIP staining for Betz cells by colocalization with the neuronal marker (NeuN), and the calcium-binding protein PV (Fig. 3). PV-positive (PV+) axonal endings, which are known to heavily innervate Betz cells14, were very evident in our material, forming perisomal networks that delineated the Betz cell bodies (Figs. 3 and 4). We also found that PV is present in the soma of a subset of Betz cells in marmoset (Fig. 4), as reported previously in the human brain14. The PV signal was far less intense in Betz cell bodies, compared to interneurons or PV+ terminal nets delineating soma boundaries, but was clearly distinguishable from the background, and particularly evident by comparison with PV negative (PV-) Betz cells. Examples of PV+ and PV- Betz cells alongside interneurons are shown in Fig. 4. The two other calcium-binding proteins tested, calbindin D28-K (CB) and calretinin (CR), were not expressed within Betz cells (Fig. 4). This is despite the known expression of CB in some populations of pyramidal cells in primates, including in the marmoset36,37. The largest of Betz cells reached up to 25 µm in soma diameter in the marmoset brain (Figs. 3 and 4).

Fig. 3
figure 3

Co-localisation of vasoactive intestinal polypeptide (VIP) with a neuronal marker (NeuN) and parvalbumin (PV) in the Betz cells. (a) Confocal images of cortical layer 5 showing immunoreactivity to VIP either with neuronal marker NeuN (top) or PV (bottom). The two cyan dashed lines identify layer 5. (b) Insets at higher magnification. White arrows point to the Betz cells in all images. Scale 100 µm (a) and 50 µm (b). VIP antibody used in production of images is from Atlas Antibodies (Cat# HPA017324).

Full size image
Fig. 4
figure 4

Giant pyramidal cells of the primary motor cortex are identified by parvalbumin (PV) but no other calcium binding proteins. Examples of NeuN (a), calbindin-D28k (b), calretinin (c) and parvalbumin (d) staining of area 4A. The dashed squares in (a–d) are shown in higher magnification in (e–h), respectively. White arrows in (e) point to Betz cells. Betz cells in (i–k) are taken from sections not shown here and those in (l,n) are magnified from panel h. Asterisks (in panels k and n) point to Betz cells that do not contain PV in their soma. PV content is also compared to an interneuron visible in (m), highlighted by a red arrow in panel (h). Scale 200 µm for (a–d), 100 µm for (e–h) and 25 µm for (i–n).

Full size image

VIP+ inhibitory neurons

VIP staining revealed a subset of inhibitory neurons in the marmoset motor cortex (Fig. 5a–c). VIP+ interneurons had different morphologies likely including bipolar cells, double bouquet cells, and bitufted cells38 (Fig. 5a). Co-staining for VIP and GABA confirmed the inhibitory nature of VIP+ interneurons (Fig. 5b,c). The arbitrary fluorescence intensity was much higher in the VIP+ interneurons than in VIP+ Betz cells (Fig. 5d,e, unpaired t-test; 3165 ± 281 vs. 1333 ± 125, p < 0.0001). However, considering the somatic area, the corrected total fluorescence within the cell bodies was similar (Fig. 5f, unpaired t-test, 18,347 ± 2473 vs. 14,404 ± 258, p = 0.16).

Fig. 5
figure 5

Antibody against vasoactive intestinal polypeptide (VIP) stains inhibitory interneurons and Betz cells. Representative images of different morphologies of VIP-positive (VIP+) inhibitory neurons of the primary motor cortex (a). White arrowheads point to the dendritic bifurcations. Inhibitory nature of VIP+ interneurons is shown by co-localisation with gamma aminobutyric acid (GABA) that is missing in Betz cells (b,c). Fluorescence intensity comparison for a VIP+ inhibitory neuron and Betz cell (d). Cyan and yellow arrows point to VIP+ inhibitory neurons and Betz cells, respectively in b–d. Mean ± SEM density of arbitrary fluorescence intensity obtained from variable somatic locations (e) and corrected total fluorescence for somatic area (f) for VIP+ neurons. Scale bar in (d, 20 µm) applies to all fluorescence images. VIP antibodies used in production of images are from Atlas Antibodies (Cat# HPA017324, a,d) and Neuro Mab (Cat# L108/92, b,c).

Full size image

In all subfields of the primary motor cortex, VIP+ inhibitory neurons were mostly concentrated in the supragranular layers and their density gradually decreased toward layer 6 (Fig. 6a–c). The mean density of VIP+ interneurons was significantly higher in A4c compared to the other subdivisions of primary motor cortex [Fig. 6d, one-way ANOVA: F (2, 33) = 21.26, p < 0.0001; Tukey's test; A4c (4745 ± 146/mm3) vs. A4a (3065 ± 224/mm3) or A4b (3410 ± 199/mm3), p < 0.0001)]. VIP+ interneurons corresponded to 4.5, 5.2, and 6.6% of the total neuronal population in cytoarchitectural subdivisions A4a, A4b, and A4c, respectively (Fig. 6e).

Fig. 6
figure 6

Layer profile of vasoactive intestinal polypeptide-positive (VIP+) interneurons of the primary motor cortex. Representative VIP-stained strips from different parts of the primary motor cortex including areas A4a, A4b and A4c (a). Individual cells are indicated by yellow arrows. Small red lines identify the border with layer 2 and white matter. Limits of layer 5 are shown by small black lines. Quantification of VIP+ interneuron density across cortical depth (b). Some VIP+ interneurons are shown at higher magnifications in (c). Scale 100 µm in (a) and 20 µm in (c). Mean ± SEM density of VIP+ interneurons (d) and its percentage, as ratio of total neurons stained by NeuN in our previous study33 (e). Black circles in (d) represent values obtained from individual sections. VIP antibody used in production of images is from Abcam (Cat# ab30680).

Full size image

Discussion

The primary motor cortex is not only involved in motor control, but also in sensory integration, behavioural strategising, working memory, and decision-making39. Its susceptibility in several diseases, including neurodegenerative conditions and ageing40,41,42,43,44, warrants further understanding of its neurochemistry. Here, we demonstrate that VIP, a peptide with widespread distribution in many organs, is expressed in the largest excitatory pyramidal neurons of the primate neocortex, adding to the intricacy of distinctive gene expression and physiological features of the giant layer 5 Betz cells6. This finding extends VIP expression to a new class of cortical neurons, raising new questions about the physiological role of this peptide in excitatory neurons of the motor cortex.

The specificity of VIP staining was confirmed using three different primary antibodies. Control experiments, in which these antibodies were omitted, along with an absorption test supported the specificity (see Supplementary Figs. S1 and S2), ruling out alternative explanations such as non-specific background staining. Our findings are supported by transcriptomic data in humans, suggesting the weak presence of VIP mRNA in cortical pyramidal cells45, including layer 5 extratelencephalic projecting glutamatergic cortical neurons of motor cortex (https://cellxgene.cziscience.com/gene-expression).

VIP expression in Betz cells

VIP is widely distributed in the nervous system as well as in many other peripheral organs, acting both as a neuromodulator and neuroendocrine releasing factor46,47,48. In the brain, VIP regulates the secretion and expression of neurotrophic factors49,50, has neuroprotective and anti-inflammatory effects46,48, and is up-regulated in neurons and immune cells after injury and/or inflammation48,51.

Owing to VIP’s versatile and widespread effects, it is currently difficult to pinpoint its possible role in Betz cells without additional functional assessments. However, the particularly large size of Betz cells may imply unique metabolic/homeostatic demands, for which VIP may contribute by having neuroprotective effects.

Betz cells, with their large body size and long myelinated axons, are amongst the most vulnerable neurons52, which are affected in ageing and motor neuron disease40,42,44. High energy requirement and reliance on long distance axonal transport, along with a large cell surface area, increases exposure to undesirable factors53, likely rendering Betz cells in need of extra regulatory protection. In line with this hypothesis, it has been shown that VIP mRNA is induced in facial motor neurons of mice following the application of an inflammatory stimulus directly to the nerve51. Moreover, the expression of PV in Betz cells14,54 may add another layer of protection through calcium regulation. Consistently, the proportion of PV-containing Betz cells decreases with age in humans55, making them more susceptible to degeneration42.

The presence of VIP in the largest pyramidal cells may be related to the high demands of these cells for energy substrate. VIP induces concentration-dependent glycogenolysis in mouse cortical slices, which may indicate a regulatory mechanism for the local control of energy metabolism56. Such a hypothesis is consistent with the maximal VIP staining in the largest pyramidal cells, which are located more medially in A4ab. Overall, the expression of VIP in Betz cells was weak and whether they actually produce VIP and use it as a neurotransmitter or neuromodulator needs to be clarified by further studies.

VIP+ interneurons in the motor cortex

VIP+ interneurons were primarily located in the supragranular layers of the cortex in the areas examined in this study, similar to observations in other species23,24,25,26,35. VIP+ interneuron density across cortical layers showed a sharp decline toward the white matter, with far fewer VIP+ interneurons present in the infragranular layers. Supra- and infragranular VIP+ interneurons may have different functional roles as the two populations are transcriptionally distinct in mice57. Although the drop in the density of VIP+ interneurons in the infragranular layers was consistent in the three examined areas, the mean density of VIP+ interneurons and their relative proportion to the total neuronal density were higher in A4c compared to the areas 4a and 4b. This may reflect regional differences in network organisation for facial versus limb and body movements. Together with the differences in expression of VIP in Betz cells, this result further highlights the fact that functionally defined cortical areas are not necessarily homogeneous, a principle which has been demonstrated in terms of cellular structure58 (including present results), connections59,60 and physiological properties61,62.

Overall, we report a 5–7% contribution of VIP+ interneurons to the neuronal population in the motor cortex in the marmoset monkeys. Our density estimates for VIP+ interneurons in the motor cortex (3065–4745/mm3) are 50–100% higher than those in the mouse motor cortex26. Based on the 23% prevalence of GABAergic interneurons in marmoset’s motor cortex6 and total neuronal density33, we estimate that VIP+ interneurons of the motor cortex constitute around 20–29% of the total interneuron population. These estimates are higher than those reported in several cortical areas of mice, which are around 11–15% of total GABAergic neurons22,25,63,64. Analysis of single-nucleus RNA expression in marmosets suggests that although the relative percentage of VIP+ neurons is much smaller in A4ab compared to prefrontal or somatosensory cortices, they constitute the second highest population of interneurons in A4ab27. The observed morphologies of VIP+ interneurons in the marmoset motor cortex were similar to those reported in other species23,38,65.

In analyses that take into consideration the soma size of neurons, the arbitrary fluorescence intensities in the VIP+ interneurons were similar to the Betz cells, indicating a comparatively similar expression of VIP. While the modes of connectivity of VIP+ interneurons with Betz cells remain to be elucidated, current literature suggests that VIP+ interneurons provide inhibitory inputs to pyramidal neurons mainly through forming disinhibitory circuits17,19,64. However, direct inhibition of pyramidal neurons by this population has also been reported18,20,21.

The abundance of studies on VIP+ interneurons in rodents has shed light on the specific roles of this neuronal population, which include motor integration17,66,67 and gain control19. However, we are far from attaining a similar understanding of their function in primates, where only limited studies have been made available23,27,68,69, possibly due to the technical difficulty of successful staining for VIP, including visualisation of interneurons. Here, we have determined that reliable VIP staining in the marmoset can only be achieved with short periods of post-fixation, an observation that may facilitate further studies.

Conclusion

The distribution of VIP+ neurons in the marmoset motor cortex demonstrates that this neuropeptide is expressed both in interneurons and in large pyramidal cells with putative projections to the spinal cord (Betz cells). The specific presence of VIP in Betz cells hints at a different set of functions for this neuropeptide, as well as differences in neuronal circuitry involved in generating and controlling different types of body movements.