Substance P NK1 receptor in the rat corpus callosum during postnatal development

Abstract Introduction The expression of substance P (SP) receptor (neurokinin 1, NK1) was studied in the rat corpus callosum (cc) from postnatal day 0 (the first 24 hr from birth, P0) to P30. Methods We used immunocytochemistry to study the presence of intracallosal NK1‐immunopositive neurons (NK1IP ‐n) during cc development. Results NK1IP ‐n first appeared on P5. Their number increased significantly between P5 and P10, it remained almost constant between P10 and P15, then declined slightly until P30. The size of intracallosal NK1IP ‐n increased constantly from P5 (102.3 μm2) to P30 (262.07 μm2). From P5 onward, their distribution pattern was adult‐like, that is, they were more numerous in the lateral and intermediate parts of the cc, and declined to few or none approaching the midline. At P5, intracallosal NK1IP ‐n had a predominantly round cell bodies with primary dendrites of different thickness from which originated thinner secondary branches. Between P10 and P15, dendrites were longer and more thickly branched, and displayed several varicosities as well as short, thin appendages. Between P20 and P30, NK1IP ‐n were qualitatively indistinguishable from those of adult animals and could be classified as bipolar (fusiform and rectangular), round–polygonal, and pyramidal (triangular–pyriform). Conclusions Number of NK1IP ‐n increase between P5 and P10, then declines, but unlike other intracallosal neurons, NK1IP ‐n make up a significant population in the adult cc. These findings suggest that NK1IP ‐n may be involved in the myelination of callosal axons, could play an important role in their pathfinding. Since they are also found in adult rat cc, it is likely that their role changes during lifetime.


| INTRODUCTION
The corpus callosum (cc), the largest fiber tract connecting the two cerebral hemispheres, is made up of axons whose cell bodies are located in layers II/III and V of the cerebral cortex (Innocenti, 1986).
The mature cc also contains both astrocytes and oligodedendrocytes (Innocenti, 1986), and many studies performed in different species, including humans, have described the presence of neurons.
SP also elicits a variety of effects by activating multiple subtypes of tachykinin receptors. Such effects appear to be involved not only in synaptic transmission, but also in synaptic plasticity during development of the mammalian central nervous system (CNS); in particular, several studies suggest that NK1 R may play a role in the synaptic plasticity associated with morphological and CNS functional development (Jonakait, Ni, Walker, & Hart, 1991;Ni & Jonakait, 1988;Quirion & Dam, 1986).
In many cases, these neurons, known as "intracallosal neurons" (Jovanov-Milošević, Petanjek, Petrović, Judaš, & Kostović, 2010), decrease during the postnatal period. In cats the number of MAP2positive intracallosal neurons drops from 570 at birth to about 200 in the adult cc. Moreover, their distribution changes with age. At first, they are found throughout the cc, whereas in the adult they are confined to the rostrum (Riederer et al., 2004). CR-positive neurons have been detected on the ventral border of mouse cc during the early stages of postnatal development (Revishchin et al., 2010). In the human cc, intracallosal neurons are particularly numerous in the second half of gestation and in the early postnatal years, but are sporadically found in the adult brain (Jovanov-Milošević et al., 2010). The above studies strongly suggest that the developing cc contains populations of transient neurons.
The present study was devised to gain insight into the possible involvement of SP in early postnatal cc development. To do this, an antibody against NK1 R (Shigemoto et al., 1993) was used to verify the presence of immunopositive intracallosal neurons in rats of different ages, from postnatal day 0 (P0) to P30 and to assess their size, morphology, and distribution during postnatal development.

| Animals
The study involved 43 Sprague Dawley albino rats of different ages whose care and handling were approved by the Animal Research Committee of Marche Polytechnic University in accordance with National Institutes of Health guidelines. All efforts were made to minimize animal suffering and to reduce the number of animals used.

| Definition of stereotaxic levels
Animals came from three different litters and were examined at seven different ages. The day of birth (the first 24 h from birth) was considered as day 0 (P0).
Brains were removed and postfixed for 8-12 hr in the same fixative used for perfusion and then placed in increasing solutions of sucrose (10%, 20%, 30% in PB; at 4°C) for cryoprotection, until they sank. Each brain was cut in the sagittal plane into 60μm serial sections using a freezing microtome. Sections were collected in PB (0.1 mol/L; pH 7.4), mounted on subbed slides, stained with neutral red (Fluka Chemie GmbH, Buchs, Switzerland; 1% in aqueous solution), and covered by coverslip. They were analyzed by light microscopy to identify, in each age group, stereotaxic levels comparable with those of the adult. The stereotaxic levels selected were lateral (lat) 3.9, 2.9, 1.9, 0.9, 0.4. At these levels the following nuclei were easily recognizable even at P0: 3.9: fimbria (fi), hippocampus, parasubiculum (PaS), presubiculum (PrS), lamina dissecans entorhinal cortex (DsC).
A section in every four was placed in PB (0.1 mol/L; pH 7.4) and then mounted on subbed slides, stained with neutral red (1% in aqueous solution), and covered by coverslip. to all other stereotaxic levels was regularly reacted for NK1 immunocytochemistry. Sections from P0, P5, and P7 rats were reacted together with sections from P30 animals; the overlying cerebral cortex, caudate putamen (CPu), globus pallidus (GP), and mesencephalon were used as positive controls. The pattern of NK1 immunopositive neurons (NK1 IP-n ) in these CNS regions was consistent with previous studies (Barbaresi, 1998;Barbaresi et al., 2015;Horie et al., 2000;Mensà, 2013;Shigemoto et al., 1993).

| Characterization of the NK1 antibody
The NK1 R antibody was made in rabbit against a peptide corresponding to amino acid residues 349-407 of rat SP receptor; its specificity has been verified by preabsorption with trp E-SPR fusion protein, which abolished all staining (see Figure 2c of Shigemoto et al., 1993).

| Distribution
The distribution of NK1 IP-n in the cc was drawn using a camera lucida attached to a Leitz Orthoplan microscope equipped with a 10× objective (Leica, Wetzlar, Germany). Callosal boundaries were obtained by comparing the sections counterstained with neutral red with those reported in the atlas of Paxinos and Watson (1982) and Zilles (1985).
Three comparable lateral stereotaxic levels (lat 3.9, lat 1.9-2.00, lat 0.6-0.4) were selected to study the distribution of intracallosal NK1 IP-n in each of the following age: P5, P10, P15, and P30. All cc profiles were digitized with the Epson Perfection 3170 scanner (300 dpi resolution) connected to the Power Macintosh G5. Photographic montages were created in Adobe Photoshop CS4 Extended (Version 11.0; Adobe System, Inc., CA, USA).
Counts were performed by pooling together data from three adjacent sagittal sections at five stereotaxic levels (or stereotaxic levels comparable with those of the adult): lat 3.9, lat 2.9, lat 1.9, lat 0.9, and lat 0.4. Forty-five sections per age group were used for counting the number of NK1 IP-n , overall 270 sections. Student's t test was used for statistical comparisons. p ≤ .05 (*) was considered statistically significant.

| Soma size
Intracallosal NK1 IP-n were randomly selected for soma size analysis according to the following criteria: (1) neurons must be intensely labeled and must show a clearly distinguishable morphology; (2) cell bodies must be located centrally within the 60μm section depth in order to minimize the cutting of dendritic branches near the section surface; (3) dendrites must not be overly obscured by other heavily stained processes from nearby cells; (4) dendritic trees must not show discontinuity with their cell bodies. For each age group, soma size was obtained by pooling data from three different rats.
The outlines of all somata were drawn with a camera lucida attached to a Leitz Orthoplan microscope equipped with a 100× objective (Leica). Soma profiles were digitized with the Epson Perfection 3170 scanner (300 dpi resolution) connected to the Power Macintosh G5 (Apple Italia, Srl; Milano, Italy). The size of NK1 IP neurons, measured as square microns, was calculated using the NIH Image program (Rasband & Bright, 1995).

| Photomicrographs
Photomicrographs of NK1 IP-n were acquired using an Eclipse E 600 microscope (Nikon-Italia, Firenze, Italy) provided with a DS-Vi1 color camera (Nikon Instruments, Europe BV, Kingston, Surrey, UK).

Photographic montages of intracallosal neurons were created in
Adobe Photoshop CS4 Extended (Version 11.0; Adobe System, Inc.); all images were cropped to appropriate size and adjusted only for brightness and contrast.

| Postnatal day 0 (P0)
The immunocytochemical procedure yielded excellent Golgi-like staining of neurons and their processes at all ages studied.
Intracallosal NK1 IP-n were not detected at P0, but a dense plexus of intensely labeled fibers (probably glial processes; Horie et al., 2000) extending from the base of the medulla to the floor of the fourth ventricle was found in the same sections, at the most medial levels of the medulla oblongata (Figure 1a (Figures 6j and   9a). The number of NK1 IP-n peaked at P10, then slightly declined from P10 to P15 (Figure 3). Their distribution was similar to that of the adult. NK1 IP-n were seen along the rostrocaudal extension of the cc, but showed differences along its lateromedial dimension, being more numerous at the lateral and intermediate levels (Figures 2 and 7b,c).  (Figures 2 and 7d) and their number showed a further slight reduction (Figure 3).

| Postnatal day 20 (P20) to postnatal day 30 (P30)
As reported in a recent study (Barbaresi et al., 2015), intracallosal  (Figures 3 and 7e). The dendrites could be followed up for hundreds of microns and bore several swellings and some spines along their course (Figures 8 and   9). Although dendritic spines were not counted, they seemed to be less numerous than those of P 15 rats (Figures 8 and 9). NK1 IP-n dendrites formed a dense network along the rostrocaudal extension of the cc (Figure 9e,h). The dendrites could often be followed as far as the overlying white matter (Figures 8, P20-A, P30-B, and 9h).

| DISCUSSION
This study examined the distribution of rat intracallosal neurons expressing SP receptor NK1 on the cellular membrane at different postnatal life and decreases with aging, although regional variability in NK1 ontogeny has been described.
In the rat trigeminal motor nucleus, NK1 receptor expression peaks at P7 and then declines (Tanaka-Gomi et al., 2007).
Immunocytochemical and western blot analyses indicate that its expression in the hypoglossal nucleus also decrease postnatally (Adachi, Huxtable, Fang, & Funk, 2010). In a RT-PCR study, Taoka et al. (1996) documented a transiently high level of NK1-IR mRNA between days 0 and 3, followed by a gradual reduction, in the rat cerebral cortex, hippocampus, and cerebellum. In the striatum, SP receptor binding sites have been seen to form dense patches between P1 and P7 and to decrease thereafter, whereas high densities of binding sites has been reported in most brain stem nuclei of neonatal but not adult rat (Quirion & Dam, 1986). (1986)  These data also suggest that, in the cc the reduction in intracallosal NK1 IP-n is not due to a dilution related to the cc volumetric expansion, but due to a reduced ability of intracallosal neurons to express the SP receptor.

Charlton and Helke
A key finding was the lack of intracallosal NK1 IP-n at P0 and their progressive increase in size and number during postnatal development. The absence of intracallosal NK1 IP-n at P0 seems to be specific for two reasons: (1) in the same sections, dense labeling was observed in the medulla oblongata, caudate putamen, hippocampus, and cerebral cortex, in accordance with previous immunocytochemical, HPLCradioimmunoassay, and autoradiographic studies (Ardelt, Karpitskiy, Krause, & Roth, 1996;Diez-Guerra, Veira, Augood, & Emson, 1989;Horie et al., 2000;Mensà, 2013;Quirion & Dam, 1986); (2) P0 sections were processed with those from P30 animals, where labeling was similar to that reported in a previous study (Barbaresi et al., 2015). However, these findings do not rule out the possibility that NK1 expression in intracallosal neurons at P0 was below the detection threshold of the immunocytochemical techniques used in the study.

| Comparison with other studies
The progressive increase in the number of intracallosal NK1 IP-n seen at different postnatal ages, especially from P0 to P5 and from P5 to P10, and the large number of neurons found in the adult (Barbaresi et al., 2015), contrast with most previous studies, since a prominent feature of cc development is the presence of transient neurons and fibers (Innocenti, 1986).  (Ding & Elberger, 2000) and then gradually decreases to become very rare in the adult cc (Ding & Elberger, 2000;Woodhams et al., 1985). Reelin-expressing cells are seen in the rat and mouse cc at P7, P14, and P21 (Misaki, Kikkawa, & Terashima, 2004); however, since the authors provide no information either on the number of intracallosal neurons found at the various postnatal ages or on the number found in the adult, these data are difficult to evaluate and to compare with those found in the present study.
A recent immunocytochemical study has shown that chain migrating interneurons positive for Sp8 (a transcription factor) transiently cross the cc during the second postnatal week (Cai, She, & Wang, 2015). Moreover, GABA-like immunoreactive (ir) axons have been reported in the rat cc until postnatal day 6; they were grouped in dense bundles and most of them disappeared in the older rats (Cobas, Alvarez-Bolado, & Fairén, 1988). Very few and sparse GABA-containing fibers persist in the adult rat cc (Ottersen & Storm-Mathisen, 1984).
These fibers could be axons of cortical cells projecting transiently through the cc. In an in vivo study performed in rat pups (from P0 to P1), combining retrograde labeling with electrophysiology and immunocytochemistry, Kimura and Baughman (1997) found a population of GABAergic callosal neurons accounting for between 21% and 57% of the whole callosal population. In the adult, GABAergic callosal F I G U R E 6 Photomicrographs of intracallosal NK1 IP-n at three postnatal ages (P5-P7-P10). a-e: P5; f: P7; g-j: P10. a-d: Four intracallosal NK1 IP-n showing different morphologies. (a) An ovoid NK1 IP-n with a thick principal dendrite directed toward the ependymal cc region. (b and d) Two round NK1 IP-n close to the ependymal cc region, whose dendrites are directed toward the dorsal, posterior and anterior cc regions. (c) Two adjacent NK1 IP-n in the middle of the cc. (e) Medial cc region, probably between lateral at 0.9 and 0.4 mm (comparable with those of the adult): no neurons are found at this cc level. Several NK1 IP-n are visible over the cc, in the IG. (f) An ovoid NK1 IP-n with dendrites directed in all directions, including the ependymal cc region. Arrowhead: a growth bud. (g) Polygonal NK1 IP-n . (i) Several intracallosal NK1 IP-n . A neuron (asterisk) sends its dendrites into the CPu. (j) A bipolar NK1 IP-n close to the ependymal region of the cc. Arrowheads in d and f indicate growth buds at branching points. Calibration bars: 25 μm in a-d, f, g, i, j; 100 μm in e; 10 μm in h neurons are reduced to 0.7-1% of the whole callosal population (Fabri & Manzoni, 2004;Gonchar et al., 1995).
The developing rat cc also contains a transient population of NPY ir and SOM ir fibers. Both fiber populations initially increase, they peak at P10, then decrease to mature levels. Only few NPY ir and SOM ir axons are found in the adult cc (Ding & Elberger, 2000). These fibers could be axons sent through the cc to the contralateral hemisphere by transitory NPY ir and SOM ir neuronal populations found in the rat cerebral cortex (Ding & Elberger, 1994. The second important finding of our study regards the distribution of intracallosal NK1 IP-n during postnatal development. At variance with earlier reports, we found that from P5 onward their distribution was similar to that described in the adult cc (Barbaresi et al., 2015), in that they were more numerous in the lateral cc and gradually decreased approaching the midline, where they were few or absent; in contrast, in cat and human cc, the distribution of intracallosal neurons in early postnatal development is different to that of the adult cc.

| Hypothesis on the functional role of the intracallosal NK1 IP-n
The presence and numerical growth of intracallosal NK1 IP-n between P5 and P30 could be related to the myelination process of cc axons.
A recent double-labeling immunofluorescent study has demonstrated that in the adult rat cc nearly all intracallosal NK1 IP-n colocalize with neuronal nitric oxide synthase (nNOS), the enzyme that synthesizes NO (Barbaresi et al., 2015). It may thus be hypothesized that NK1 IP-n containing nNOS could be present even in the cc of younger animals.
This hypothesis is supported by the increase in NO-producing neurons in the deep white matter of the rat cerebral cortex during postnatal development (Clancy, Silva-Filho, & Friedlander, 2001). The excitation of these neurons by SP (via synaptic contact or volume transmission) F I G U R E 7 Number of NK1 IP-n detected on different stereotaxic planes from lateral (lat 3.9) to medial (lat 0.4  Paxinos and Watson (1982). Three rats per age (CC-Nk1/1; CC-NK1/2; CC-NK1/3) are shown. Data for each rat come from pooling three adjacent sections could lead to NO production through different second messenger systems (Bredt & Snyder, 1990;Khawaja & Rogers, 1996;Quartara & Maggi, 1997;Saria, 1999;Vincent, 1994) subsequently to its release.
In turn, released NO would stimulate the growth and differentiation of oligodendrocytes, which are responsible for the myelination of callosal axons (Garthwaite, Hampden-Smith, Wilson, Goodwin, & Garthwaite, 2015;Tanaka, Markerink-Van Ittersum, Steinbush, & De Vente, 1997) that occurs during the first month of life (Seggie & Berry, 1972;Valentino & Jones, 1982). The importance of NO in cc myelination processes is also demonstrated by other studies. nNOS-deficient mice show a delay in remyelination following chemical demyelination (Liñares et al., 2006), and Sprague Dawley pups inhaling NO-enriched air during first postnatal week show increased myelination of cc axons (Olivier et al., 2010).
As in the adult (Barbaresi et al., 2015), NO released from intracallosal NK1-expressing neurons may also be involved in cerebrovascular control mechanisms (Iadecola, 2004) or in modulating arterial blood flow during cerebral ischemia in rat pups (Bonnin et al., 2012).
In parallel with their number, the size of intracallosal NK1 IP-n also underwent a considerable increase. Since measurements were not made in the plane of the nucleolus (Anamizu, Seichi, Tsuzuki, & Nakamura, 2006;Offord, Ota, Oenning, & Dyck, 1974), their actual size could not be measured; however, our experimental approach allowed document- the growth of projecting (pyramidal) and local circuit neurons accelerates in the first 3 weeks of postnatal development to achieve adult size at the end of the fourth week (Miller, 1984). Between P5 and P10, the dendritic tree of NK1 IP-n increased gradually, as described in cerebral cortex neurons (Miller, 1984). Since the dendrites NK1 IP-n extend along the anteroposterior axis of the cc and toward the overlying cerebral cortex, it is possible for intracallosal NK1 IP-n to be activated by SPergic sibility is that SP released from these cortical neurons could act on intracallosal NK1 IP-n in a paracrine-like manner, since SP can diffuse across a significant distance from its site of release to bind to a receptor (Liu et al., 1994;Nakaya et al., 1994;Vruwink, Schmidt, Weinberg, & Burette, 2001;Wolansky, Pagliardini, Greer, & Dickson, 2007). In addition, the present data showed that from P10 onward neurons were more often grouped in clusters, with their dendrites forming a dense network close to the ependymal layer or to the middle and dorsal cc. Even more frequently, NK1 IP-n were found close to the ependymal layer; this suggests that they may be in contact with cerebrospinal fluid (CSF) through their dendrites, axons, or perikarya, and that they may belong to the CSFcontacting neuronal system found in many vertebrate periventricular brain regions (Vigh et al., 2004). Although CSF contains a relatively high amount of SP in the adult brain (Muñoz & Coveñas, 2014), little is known about its SP concentration in early stages of the rat cerebral development. High levels of SP have been reported in CSF of fetuses and children (Tam, Dockray, & Lister, 1985). If this also applies to the rat, then ependymal NK1 IP-n could be activated, more intensely than in adults, via volume transmission by diffusion of SP from CSF through the intercellular space (Abbadie, Skinner, Mitrovic, & Basbaum, 1999;Barbaresi et al., 2015;Ramer, 2008), thus playing an important role in neurodevelopment processes as suggested for humans (Tam et al., 1985). Such neurons could also be involved in CSF composition and in the regulation of its pH and osmolality (Vigh et al., 2004).
An important and intriguing feature of intracallosal NK1 IP-n , found in the present study, is the presence of dendritic filopodia mixed with dendritic spines, which seemed to be particularly numerous between P10 and P15 and then declined during the postnatal development.
What is the function of these dendritic protrusions? Although there is evidence that both dendritic filopodia and spines are involved in synaptogenesis during development of several of CNS regions (Ziv & Smith, 1996), their functional role in the cc is less clear. Moreover, an exuberant and transient projection, formed by an excess of callosal projecting neurons and callosal axon branching, has been described in the cc during the first stage of development (Innocenti, 1986;Kadhim, Bhide, & Frost, 1993;O'Leary, Stanfield, & Cowan, 1981). Are these spines a target for transient axonal branching of permanent callosal axons? The findings described above suggest that, during development, dendritic protrusions grow and search for nearby axons to synapse with; thereafter they disappear due, for example, to retraction of transient axonal branching of permanent callosal axons (Kadhim et al., 1993). In support for this hypothesis, a similar sequence of events seems to take place in the development of the afferent innervation of other CNS regions (Ramoa, Campbell, & Shatz, 1988;Saito et al., 1992).
According to our findings, dendritic filopodia decreased as age increased, and only spines were found on NK1 dendrites at P20 and P30. Dendritic spines are sites of excitatory synaptic transmission, and their structure and density are important measures of synaptic function (Pannese, 1994;Peters, Palay, & Webster, 1991). It is therefore likely that intracallosal NK1 IP-n receive some synaptic contact.
The hypothesis is supported by an early electron microscopic study describing synapses on intracallosal neurons (Ling & Ahmed, 1974).
Moreover, NK1 IP-n receiving neurochemically diverse synaptic inputs have been described in other fibrous tracts such as the dorsal columns of several mammals (Abbadie et al., 1999;Ramer, 2008).

| CONCLUSION
The main findings of this study may be summarized as follows: (1) intracallosal neurons expressing NK1, the principal SP receptor, are visible since P5; (2) at P5, their distribution is already similar to that seen in the adult; (3) their number increase between P5 and P10 then declines, but unlike other intracallosal neurons, NK1 IP-n make up a significant population in the adult cc; (4) intracallosal NK1 IP-n size increase with age; (5) starting at P20, intracallosal NK1 IP-n form a heterogeneous population. These neurons may act as temporary targets for ingrowing callosal axons, take part in the mechanism of axonal myelination of callosal fibers, and play an important role in callosal pathfinding. Since they are also found in adult rat cc, it is likely that their role changes during lifetime (Friedlander & Torres-Reveron, 2009;Rockland & Nayyar, 2012).

ACKNOWLEDGMENTS
The authors are grateful to Professor Ryuichi Shigemoto (Division of Cerebral Structure, National Institute for Physiological Sciences, Okazaki, Japan) for generously supplying the NK1 antibody and to Dott.ssa Silvia Modena for the language review (Word Designs; F I G U R E 9 Photomicrographs of intracallosal NK1 IP-n at different postnatal ages (P15, P20, P30). (a) A round intracallosal NK1 IP-n whose dendrites are oriented in all directions (P15). (b) A triangular neuron; the apical dendrite crosses the white matter and reaches the cerebral cortex (P15). The framed area, enlarged in c, shows both dendritic appendages and dendritic spines (arrowheads). (d) A bipolar neuron bearing several dendritic appendages (arrowheads) and dendritic spines (P15). Coronal section. (e) Cluster of NK1 IP-n in the splenium (P20). (f) Two NK1 IP-n in the medial cc (P20). Some NK1 IP-n are in the IG, over the cc. (g) A bipolar NK1 IP-n close to the ependymal region of the cc (P30). (h) Several NK1 IP-n in the middle cc region (P30). (i) A probable bipolar NK1 IP-n in the splenium (P30). Calibration bars: 500 μm in a, b, and h; 250 μm in e, g, and i; 100 μm in f; 10 μm in c and d www.silviamodena.com). Grant sponsor: Università Politecnica delle Marche-Ricerca Scientifica d'Ateneo 2013-2014.

CONFLICTS OF INTEREST
None declared.