The ability of cat superior colliculus (SC) neurons to integrate information from different senses is thought to depend on direct projections from regions along the anterior ectosylvian sulcus (AES). However, electrical stimulation of AES also activates SC output neurons polysynaptically. In the present study, we found that nitric oxide (NO)–containing (nitrergic) interneurons are a target of AES projections, forming a component of this cortico-SC circuitry. The dendritic and axonal processes of these corticorecipient nitrergic interneurons apposed the soma and dendrites of presumptive SC output neurons. Often, an individual cortical fiber targeted both an output neuron and a neighboring nitrergic interneuron that, in turn, contacted the output neuron. Many (46%) nitrergic neurons also colocalized with γ-aminobutyric acid (GABA), suggesting that a substantial subset have the potential for inhibiting output neurons. These observations suggest that nitrergic interneurons are positioned to convey cortical influences onto SC output neurons disynaptically via nitrergic mechanisms as well as conventional neurotransmitter systems utilizing GABA and other, possibly excitatory, neurotransmitters. In addition, because NO also acts as a retrograde messenger, cortically mediated NO release from the postsynaptic elements of nitrergic interneurons could influence presynaptic cortico-SC terminals that directly contact output neurons.
Output neurons in the deep layers of the cat superior colliculus (SC) project to brainstem regions involved in the control of head and eye movements (Sparks 1986; Stein and Meredith 1991; May 2005). These neurons also integrate information from 2 or more senses to facilitate overt behavior (Stein and Meredith 1993). This ability depends on influences from the anterior Ectosylvian Visual area (AEV) (Mucke et al. 1982; Olson and Graybiel 1987; Benedek et al. 1988; Jiang et al. 1994), auditory Field of the Anterior Ectosylvian Sulcus (FAES) (Clarey and Irvine 1986; Meredith and Clemo 1989), and somatosensory area IV (SIV) (Clemo and Stein 1982, 1983) subdivisions of the anterior ectosylvian sulcus (AES) (Jiang et al. 2002).
These unisensory cortico-SC projections (Wallace et al. 1993) are dense, highly organized, and target-specific SC neurons (Stein et al. 1982; McHaffie et al. 1988; Meredith and Clemo 1989; Meredith et al. 1992; Harting et al. 1997). A given multisensory SC neuron receives converging inputs from multiple AES subregions (Wallace et al. 1993; Jiang et al. 2001), and because many of the specifics of this cortico-SC circuit are not well understood, their powerful influences over SC multisensory integration (Wallace and Stein 1994; Jiang et al. 2001; Jiang and Stein 2003) are generally assumed to be monosynaptic (for a computational model of this circuit, see Rowland et al. 2006). However, electrical stimulation of AES also activates SC neurons with latencies indicative of polysynaptic inputs (Wallace et al. 1993), suggesting that AES modulates multisensory integration not only via direct contacts but also indirectly via interneurons, whose physiological properties are rarely sampled during electrophysiological experiments.
SC interneurons form a heterogeneous group whose axons have multiple and far-reaching ramifications, some of which may extend to the contralateral SC (Behan 1985; Ma et al. 1990; Behan and Appell 1992; Behan and Kime 1996; Olivier et al. 2000). Histological studies have shown that these small soma neurons have multipolar and fusiform morphologies with up to 5 primary dendrites that arborize onto the distal dendrites of neighboring neurons (Norita 1980; Mize 1988). Most of them express the inhibitory neurotransmitter γ-aminobutyric acid (GABA) (Mize et al. 1996; Behan et al. 2002) as well as the enzyme nitric oxide synthase (NOS) (Scheiner et al. 2000; Soares-Mota et al. 2001), the presence of which suggests that such interneurons produce nitric oxide (NO) (Bickford et al. 1999; McCauley et al. 2003), a diffusible gaseous molecule.
NO has a universal role in modulating the release of neurotransmitters throughout the brain and thus has profound effects on neuronal excitability and firing (for review, see Esplugues 2002). Although the specific functions of NO in the adult SC remain to be elucidated, many critical synaptic phenomena such as long-term potentiation and long-term depression appear to be initiated or regulated by NO (Garthwaite and Boulton 1995). Given that NO can also act in a retrograde manner with fast diffusion kinetics, it may be well-suited for modulating cortico-SC afferents that contribute to the enhancement or depression of SC neurons as they integrate information from different sensory modalities (Stein et al. 2004). Therefore, we sought to determine if nitrergic neurons were a component of the cortico-SC circuitry essential for SC multisensory integration. Of specific interest was determining the spatial relationships among cortical afferents, nitrergic neurons, and SC. It is also relevant to note that NO abnormalities have been implicated in a number of diverse neuropsychiatric disorders (Akyol et al. 2004; Sweeten et al. 2004). Thus, the present study may help us to understand how developmental challenges to the functional integrity of this cortical-midbrain circuit may contribute to the sensory integration anomalies that are manifested in autism (Iarocci and McDonald 2006), attention deficit/hyperactivity disorder, and sensory processing disorder (Miller et al. 2001).
Material and Methods
All animal husbandry and experimental procedures were performed in compliance with the National Institutes of Health “Guide for the Care and Use of Laboratory Animals” (NIH Publications No. 80-23, revised 1996) in facilities accredited by the American Association for the Accreditation of Laboratory Animal Care. Experimental protocols received prior approval by the Institutional Animal Care and Use Committee at Wake Forest University School of Medicine. Efforts were made to minimize animals’ discomfort and the number of animals used.
Animals (n = 4) were sedated with ketamine hydrochloride (30 mg/kg, intramurally [im]) and acepromazine maleate (0.05–0.1 mg/kg, im). Thirty minutes later, they were anesthetized with sodium pentobarbital (100 mg/kg, intaperitonially [ip]) and, following the loss of withdrawal reflexes to frankly noxious stimuli, were perfused transcardially with 0.9% saline wash followed by a fixative solution of 4% paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.3. The brains were blocked in situ, removed, and stored overnight at 4 °C in the same fixative solution. On the following day, the brainstem was sectioned at 50 μm on a vibratome in the coronal plane. Sections were rinsed several times in phosphate-buffered saline (PBS) containing 0.2% Triton X-100 (Tx) and, in order to reduce nonspecific binding, were blocked for 1 h in PBS–Tx (0.2%) containing 10% normal goat serum (NGS). Sections then were incubated overnight at 4 °C with neuronal nitric oxide synthase (nNOS) polyclonal antibody (1:2000, Immunostar Inc., Hudson, WI) in a solution containing PBS–Tx (0.2%), pH 7.4. After four 15-min rinses in PBS–Tx (0.2%), sections were incubated in biotinylated anti-rabbit secondary antibody for 2 h at room temperature (RT) (1:200; Vector Laboratories, Burlingame, CA). The Vector biotin–avidin procedure (Hsu et al. 1981) was used to link the antigen–antibody complex to horseradish peroxidase, which then was visualized with diaminobenzidine (DAB; 3,3 diaminobenzidine tetrahydrochloride). Finally, sections were mounted onto gelatin-coated slides, coverslipped using Cytoseal (Stephens Scientific, Riverdale, NJ), and dried overnight at 4 °C.
NOS–SMI-32 and NOS–GABA Double Labeling
Two series of adjacent sections from these same animals were used for double-labeling studies. Sections were rinsed 4 times in PBS–Tx (0.2%) and blocked for 1 h in the same buffer containing 10% NGS. They then were incubated overnight in a cocktail of either NOS (1:2000) and SMI-32 (1: 2500 mouse anti-SMI-32, Sternberger Monoclonals Inc., Jarrettsville, MD) or NOS (1:2000) and GABA (1:1000, mouse anti-GABA, Chemicon, Temecula, CA) primary antibodies. Following four 15-min rinses in PBS–Tx (0.2%), sections were incubated in a dilution of fluorescently labeled secondary antibodies (1:200; anti-rabbit conjugated to Alexa 594 for NOS and anti-mouse conjugated to Alexa 488 for SMI-32; anti-rabbit conjugated to Alexa 594 for NOS and anti-mouse conjugated to Alexa 488 for GABA; Molecular Probes, Eugene, OR) for 2 h at RT. Finally, the sections were rinsed in PBS, mounted, coverslipped, and maintained in the refrigerator at 4 °C.
Surgery and Tracer Injections into Different AES Subregions
Three sets of injections into the AES were performed to detail cortical convergence patterns onto SC NOS-immunostained neurons: 1) injections of biotin dextran amine (BDA) conjugated to Alexa 594 into FAES and injections of BDA conjugated to Alexa 488 into AEV (n = 4), 2) injections of BDA conjugated to Alexa 488 into AEV and injections of BDA conjugated to Alexa 594 into SIV (n = 4), and 3) injections of BDA conjugated to Alexa 488 into FAES and injections of BDA conjugated to Alexa 594 into SIV (n = 4). Cats were pretreated with ketamine hydrochloride (30 mg/kg, im) and acepromazine maleate (0.05–0.1 mg/kg, im), intubated through the mouth, and then anesthetized for surgery with isoflurane (0.5–3%). Each animal was positioned in a stereotaxic head holder, and a craniotomy was performed to provide access to the cortical area of interest. The coordinates of the injection sites were according to locations of interest as determined by previous studies (Stein et al. 1983; McHaffie et al. 1988; Meredith and Clemo 1989). The extent and boundaries of the AEV (visual), FAES (auditory), and SIV (somatosensory) were identified electrophysiologically by the predominant sensory-specific responses of individual neurons using a tungsten microelectrode (1–3 MΩ). Visual search stimuli consisted of moving and flashed lights; auditory search stimuli included broadband (20–20 000 Hz) noise bursts, clicks, claps, and whistles; and somatosensory search stimuli consisted of mechanical taps, manual compression of the skin, and rotation of joints. Once an AES subdivision was identified and mapped, the needle tip of a 10-μL Hamilton syringe was positioned near its center and lowered 1–1.5 mms.
A 10-kDa dextran amine conjugated to BDA or Alexa 594 or Alexa 488 (Molecular Probes) was used as anterograde tracer. Tracers were dissolved in PB to yield a concentration of 10% and were pressure-injected into different AES subregions over a 20-min period. Following the injections, the syringe was left in place for 15 min before being withdrawn. The cortex was covered with Gelfoam, and the muscle and skin were sutured closed. The animal was kept on a heating pad until mobile and then returned to its home cage. Analgesics (burtorphanol tartrate, 0.1–0.4 mg/kg/6 h) and antibiotics (cefazolin, 20 mg/kg/bid) were administered as needed for 7–10 days.
Triple-Labeling Fluorescent Studies: Dual Tracer Injections into AES Subregions Combined with NOS Immunocytochemistry
Sections from the animals that received dual tracer injections were subsequently processed for NOS immunocytochemistry. After a 15-day survival period to facilitate anterograde transport of the tracer, animals were sedated with ketamine hydrochloride (30 mg/kg, im) and acepromazine maleate (0.05–0.1 mg/kg, im), anesthetized with sodium pentobarbital (100 mg/kg, ip), and perfused transcardially as above. Brains were stereotaxically blocked in situ, removed from the skull, and sectioned on a vibratome at 50 μm in the coronal plane. Cortical sections containing the injection site were examined microscopically in order to assure that the tracer did not spread to adjacent cortical areas. Similarly, ipsilateral SC sections were analyzed for the presence of stained fibers using a fluorescent microscope with the appropriate filters for either green or red epifluorescent illumination. After several rinses with PBS, sections were blocked with 10% NGS in PBS containing 0.2% Tx for 1 h. Sections then were incubated overnight at 4 °C in anti-NOS primary antibody. The next day, they were rinsed several times in PBS–Tx (0.2%) and incubated in a 1:200 dilution of anti-rabbit biotinylated secondary antibody (Vector Laboratories) for 2 h at RT Finally, sections were washed and incubated in a streptavidin solution (1:200; Molecular Probes) for 1 h at RT, washed, mounted on gelatin-coated slides, and coverslipped using Gel Mount (Biomeda, Foster City, CA).
Triple-Labeling Nonfluorescent Studies: Single Tracer Injections into AES Combined with B-nicotinamide Adenine Dinucleotide Phosphate-diaphorase (NADPH-d) Histochemistry and SMI-32 Immunocytochemistry
After BDA injections into the AES, animals (n = 3) were anesthetized, perfused, and treated as above. For β-nicotinamide adenine dinucleotide phosphate-diaphorase (NADPH-d) histochemistry, sections were rinsed several times in Tris buffer (pH 7.1) and then incubated in a solution containing 50 mL Tris buffer, 12.5 mg nitroblue tetrazolium, 50 mg NADPH, and 0.3% Triton-X at 37 °C for 2 h. To identify BDA-stained fibers, sections were rinsed 4 times for 15 min in Tris buffer and then incubated in the biotin–avidin peroxidase complex for 2 h at RT. The peroxidase was visualized with a nickel-intensified DAB reaction to produce a black reaction product. For SMI-32 immunocytochemistry, sections were preincubated for 1 h in 10% NGS and then incubated in SMI-32 monoclonal antibody diluted in PBS–Tx (0.2%) overnight at 4 °C. The following day, after several rinses in PBS–Tx (0.2%), sections were incubated in a dilution of biotinylated anti-mouse secondary antibody (1:200; Vector Laboratories) for 2 h at RT. After washing in PBS–Tx (0.2%), the sections were incubated in ABC for 1 h and then reacted with DAB and hydrogen peroxide to produce a brown reaction product.
Specificity of the Antibodies
To ensure the specificity of antibodies, SC sections were incubated in the absence of the corresponding primary antibody, which was replaced with either buffer or nonimmune serum from the same species (Burry 2000). No immunostaining was detected under these conditions. In addition, the specificity of all the antibodies used has been extensively characterized by western blot as well as by preabsorption with the appropriate competitive control in different brain regions (Sherry and Ulshafer 1992; Eliasson et al. 1997; Van der Gucht et al. 2001; Gotti et al. 2005).
The terminologies used to define SC boundaries and subdivisions of the AES are in accordance with the previous studies. Briefly, although the SC is generally divided into 2 functional compartments (superficial and deep), its individual layers have been identified (Kanaseki and Sprague 1974). Thus, the superficial laminae consist of stratum zonale (SZ), stratum griseum superficiale (SGS), and stratum opticum (SO), whereas the deeper layers are an amalgam of intermediate and deep layers and include stratum griseum intermediale (SGI), stratum album intermedium, stratum griseum profundum (SGP), and stratum album profundum. As noted in the Introduction, the AES cortex has been subdivided into 3 modality-specific zones: auditory (FAES), visual (AEV), and somatosensory (SIV). In addition, SC neurons were described according to both soma shape and number of dendrites emerging from the cell body. In agreement with a previous study in the cat (Norita 1980), 3 neuronal categories were used: small (5–20 μm), medium (21–35 μm), and large (36–65 μm).
In the present study, nitrergic neurons, labeled with either the nNOS or the enzyme NADPH (Scheiner et al. 2000), were defined as those neurons that produce NO, a diffusible gaseous molecule that can have excitatory effects not only in surrounding neurons but also in the neurons in which it is formed. Additionally, an apposition was considered only when the stained terminal and the immunostained neuron were in focus in the same focal plane and there was no detectable gap between them (Pilowsky et al. 1992; El Manira et al. 1997; Makeham et al. 2001). Although an apposition does not necessarily mean synaptic contact or synapse, it has been demonstrated with electron microscopy that in more than half of the cases appositions are indeed real synapses (Pilowsky et al. 1992).
Measurements of the Soma Diameter
The soma diameters of NOS-, GABA-, and SMI-32-immunostained neurons were measured using the public domain image analysis software Scion Image for Windows (Scion, Frederick, MD; v beta 4.0.2). Three fields (2.97 × 104 μm2; dorsal, middle, and ventral) were sampled using a 60× objective in every fifth section of the ipsilateral deep SC throughout the rostrocaudal extent of the nucleus. Only neurons with a well-defined soma, nucleus, and nucleolus were measured.
For quantification of immunostained neuronal profiles, a densitometric analysis similar to that described previously was used (Alvarado et al. 2004; Fuentes-Santamaria et al. 2005). Sections were examined under blue (Alexa 350), red (Alexa 594), and green (Alexa 488) epifluorescent illumination within the first 48 h of cutting. Analysis was performed in every fifth section, 250 μm apart, extending through the rostrocaudal extent of the nucleus. In each section, 3 fields (6.39 × 104 μm2; dorsal, middle, and ventral) were sampled. All the images were captured, using a 40× objective, with a Spot RT Slider digital camera (Diagnostic Instruments, Sterling Heights, MI) attached to an Olympus BX50 microscope. Scion Image, which can be customized with a built-in Pascal-like macro language (Xu and Pitot 1999; Xu et al. 2000), was used to process and analyze captured images. A macro based on the signal-to-noise ratio was written for that purpose (Alvarado et al. 2004). A monochrome (gray) image containing a grayscale of pixel intensities from 0 (white) to 255 (black) was digitized, and no contrast or brightness corrections were performed on the images.
Double-labeled neurons were identified by the presence of either NOS and SMI-32 or NOS and GABA within the soma of the same neuron. In order to quantify immunostained neurons, a consistent threshold was set as 2 standard deviations above the mean gray value of the field, and profiles exceeding this threshold were identified as immunostained (Alvarado et al. 2004; Fuentes-Santamaria et al. 2005). Following this determination, the images were converted into binary and the number of immunofluorescence profiles (neurons) per field was determined. To verify neurons colocalizing 2 different antibodies, the binary image corresponding to a given antibody was pseudocolored (red) and then merged with the other pseudocolored image (green). Then, images were merged, and the resultant image was used to quantify the colocalizing neurons.
Preparation of Figures and Statistical Analysis
Photoshop (Adobe v5.5) and Canvas (Deneba v6.0) were used to adjust size, brightness, and contrast of publication images. All data are expressed as mean ± standard deviation and analyzed statistically using a Student's t-test. Statistical significance was set at the level of P < 0.05.
Numerous neurons throughout the rostral–caudal extent of the deep (multisensory) layers of the SC were labeled with either the nNOS or the enzyme NADPH-d (Scheiner et al. 2000) that is known to colocalize with nNOS (Hope et al. 1991). Labeling was particularly evident at midcollicular levels and consisted mainly of small (5–20 μm) darkly stained neurons (Fig. 1A, also see white arrows in 1B). Because NADPH histochemistry reveals proximal and distal dendrites as well as axons, features that usually are not seen with NOS immunohistochemistry, it was chosen as the marker for nitrergic neurons in the present analysis. Accordingly, NADPH-stained bipolar neurons with dendritic arbors that were oriented either vertically (Fig. 1C,D) or horizontally (Fig. 1E), as well as multipolar neurons with radially directed arbors (Fig. 1F,G), were labeled. Although far fewer in number, a second population of neurons with medium (21–35 μm; black arrow in Fig. 1B) or large (>35 μm; black arrowhead in Fig. 1B) soma also was observed (Scheiner et al. 2000). Based on their size, laminar distribution, cell morphology and location, this latter population is likely to be the descending component of the tectoreticular spinal tract (Grantyn A and Grantyn R 1982; Moschovakis and Karabelas 1985; Munoz and Guitton 1991; Meredith et al. 1992).
Small or Medium-to-Large Nitrergic Neurons Represent Distinct Populations of Neurons: NOS and SMI-32 Double Labeling
Because the neurofilament protein SMI-32 is preferentially expressed in output neurons of the cat SC (Fuentes-Santamaria et al. 2006), we wanted to determine whether the medium-to-large soma (Fig. 2A–C) nitrergic neurons also fell into this category. The percentage of NOS-immunostained neurons (Fig. 2A–C) colocalizing with SMI-32 (Fig. 2D–F) revealed that the majority (84.36 ± 4.87%) of neurons with medium-to-large soma were double labeled, whereas only a minority (5.23 ± 0.49%) of NOS-immunostained neurons were SMI-32 immunostained (inset in Fig. 2I; also see Fig. 2G,I). This indicates that small and medium-to-large soma nitrergic neurons represent 2 distinct neuronal populations of presumptive interneurons and output neurons, respectively.
SMI-32 Immunostained Neurons Receive Appositions from Small Nitrergic Neurons
The distribution and relationship between these 2 populations of neurons then were sought. Specifically, we wanted to examine whether SC output neurons, whose descending axons form the predorsal bundle (brown neurons immunostained with SMI-32 in Fig. 3), also receive close appositions from NADPH neurons (blue neurons in Fig. 3). NADPH histochemistry in the SGI and SGP, in addition to staining neurons with different morphologies, also stained fibers (Fig. 3A). Particularly in the SGI, these fibers were organized as a series of periodic patches that ran rostrocaudally and spread across the mediolateral extent of the SC (arrows in Fig. 3A). These patchy areas consisted of fibers that had both “en passant” boutons and terminal boutons. Some of these patches in the SGI were seen interconnecting with the SGP and periaqueductal gray through vertically oriented streams of fibers (Fig. 3A). Clusters of NADPH-stained neurons were contained both within and outside these patchy areas (Fig. 3A,B). Conversely, medium-to-large SMI-32 immunostained neurons were located outside these patches (arrows in Fig. 3B), where they were frequently in close proximity to blood vessels (asterisk in Fig. 3B). The axons from NADPH-stained neurons (red arrowheads in Fig. 3) often could be seen apposing the soma (white arrowheads in Fig. 3C) or dendritic processes (white arrowheads in Fig. 3D) of SMI-32-immunostained neurons, indicating a likely interaction between these 2 neuronal populations. Further morphological evidence for such interactions was the presence of NADPH-stained dendritic processes apposing the dendrites of either medium (white arrowhead in Fig. 3E) or large (white arrowheads in Fig. 3F) SMI-32-immunostained neurons.
Injection Sites and Projection from AES
To examine whether nitrergic neurons in multisensory layers also receive converging inputs from multiple AES subregions, small paired injections of different fluorescent dextran amines were made into the AES (Fig. 4A–C). Only those cases in which the tracer did not extend to adjacent areas of the cortex were included in the present study. Inspection of the injection sites revealed that the BDA deposits in each case were restricted to the cortical area of interest. Representative BDA injection sites are illustrated in Figure 4A–C. In agreement with previous studies in the cat (Stein et al. 1983; Norita et al. 1986; McHaffie et al. 1988; Meredith and Clemo 1989; Harting et al. 1997), anterogradely labeled AES fibers and boutons were distributed mainly in the ipsilateral multisensory layers (although a very weak projection to the contralateral SC also was noted). No labeling was observed in SZ, SGS, or SO. Cortical afferents from the AES were observed throughout the rostrocaudal and mediolateral extent of SGI and SGP and were distributed into 2 distinctive tiers organized as a series of periodic patches. The first and more dorsal tier of patches was located in SGI, whereas the second one was distributed in the ventral part of SGP (not shown).
Presumptive Nitrergic Interneurons Receive Inputs Primarily from a Single AES Subregion
After dual injections of fluorescent tracers were made into different subregions of the AES (Fig. 4A–C), the distribution of cortical afferents from FAES, AEV, and SIV was investigated (Fig. 4D–F). Parent cortical fibers from AES frequently were seen branching several times, rising from 1–2 thinner collaterals at each branching point. Each of these collaterals had a large number of varicosities, where either rounded terminal boutons (white arrowheads in Fig. 4D–F) or en passant boutons (white arrows in Fig. 4D–F) were observed. Because NOS failed to stain all but the most proximal portion of dendrites in most instances, this precluded examining many potential apposition sites. Nevertheless, varicosities were seen apposing the soma and occasionally the proximal dendrites of small NOS-immunostained neurons. Commonly, each one of these cortical branches was observed bearing a large number of swellings and contacting either the same nitrergic neuron multiple times or contacting multiple neurons sequentially (Fig. 4D–F). Although most of the small NOS-immunostained neurons usually received inputs from one AES subregion (Fig. 4D–F), a few medium-sized neurons received converging inputs from multiple subregions (yellow asterisk in Fig. 4F). High-magnification images illustrating these cortical appositions onto NOS-immunostained neurons are shown in Figure 4G–J.
Projections from a Single AES Axon Can Contact both Presumptive Nitrergic Interneurons and SMI-32-Immunostained Output Neurons
Injections of BDA restricted to individual subregions of AES were made to determine whether individual corticofugal axons contact both nitrergic interneurons and output neurons. Anterogradely labeled fibers (asterisk in Fig. 5A points to one of these stained fibers) were seen apposing the dendritic and axonal processes of both small NADPH-stained neurons and SMI-32-immunostained neurons (Fig. 5A). A single fiber from the AES was observed presynaptic to the soma and dendrites of multiple NADPH-stained neurons (black arrowheads in Fig. 5B). In addition, when the axonal processes of SMI-32-immunostained neurons were labeled, cortical fibers were observed apposing their stained axons (red arrowheads in Fig. 5B). In some cases, a cortical fiber contacted an SMI-32-immunostained neuron that also received appositions from an NADPH-stained neuron that was not itself observed being contacted by the cortical fiber (white arrowheads point to these appositions in Fig. 5C). Interestingly, a single cortical fiber, bearing numerous en passant boutons (Fig. 5D–F), often was observed apposing both small nitrergic neurons (black arrowheads) and output neurons (yellow arrowheads) multiple times. Moreover, individual cortical afferents could be seen apposing SMI-32-immunostained output neurons (yellow arrowheads in Fig. 5E,F) and nitrergic interneurons (black arrowheads in Fig. 5E,F) that, in turn, contacted SMI-32-immunostained output neurons (white arrowheads in Fig. 5E,F). Less frequently, an individual cortical fiber targeted a nitrergic interneuron that contacted an immunostained SMI-32 output neuron that was not apposed by the same cortical fiber (data not shown).
GABAergic Neurons Are a Subset of Nitrergic Neurons and Receive Projections from the AES
In order to further characterize the nature of nitrergic neurons that are interposed between cortical afferents and SC output neurons, we investigated whether NOS-immunostained neurons (Fig. 6A–C) colocalized with the inhibitory neurotransmitter GABA (Fig. 6D–F). In agreement with a previous study (see Mize, 1988), the majority of GABAergic immunostained neurons had small soma (5–20 μm Fig. 6D,F), although a few medium-soma (21–35 μm, Fig. 6E,F) neurons were occasionally observed immunostained. Quantification and colocalization of NOS- and GABA-immunostained neurons revealed that almost half (46.07 ± 2.30%) of the NOS-immunostained neurons also were GABAergic, whereas the vast majority (88.94 ± 1.66%) of GABAergic neurons were NOS immunostained (inset in Fig. 6I; also see Fig. 6G,I). Arrows in Figure 6G,H are pointing to some NOS-immunostained neurons that did not colocalize with GABA. These data suggest that a substantial subset of nitrergic interneurons in the deep layers of the SC is GABAergic.
Additionally, in order to verify that this particular component of nitrergic interneurons is involved in the cortico-SC circuitry essential for SC multisensory integration, cortical BDA injections were made and the tissue coprocessed for GABA immunocytochemistry. Labeled cortical fibers bearing numerous boutons along its length were seen apposing the soma of small GABAergic immunostained neurons (arrows in Fig. 6J–L). Frequently, these cortical fibers contacted either the same neuron multiple times or multiple neurons sequentially (Fig. 6J–L).
The present observations reveal that nitrergic interneurons in the deep SC receive descending inputs from individual AES subregions known to be essential for multisensory integration. The dendritic and axonal processes of these corticorecipient interneurons apposed the soma and dendrites of presumptive multisensory output neurons. Often, an individual cortical fiber targeted both an immunostained output neuron and a neighboring nitrergic interneuron that, in turn, contacted the same output neuron. Double-labeling observations indicate that a large proportion of nitrergic neurons also colocalize with GABA, suggesting that a large subset of these corticorecipient interneurons is inhibitory. The intimate relationship of nitrergic interneurons with cortical afferents and SC output neurons, coupled with the nonsynaptic nature of NO neuromodulation, may allow such interneurons to have considerable influence in crafting the final integrated multisensory signal in SC output neurons and modulating the premotor signals by which these output neurons control eye movements.
Neuroanatomical studies using NADPH or NOS reveal that the SC contains numerous NO-stained elements (Scheiner et al. 2000, 2001; Soares-Mota et al. 2001; Behan et al. 2002). Histochemical staining for NADPH is especially prevalent in the more caudal aspects of the multisensory layers, where it is largely contained within afferent fibers arising predominantly from the mesencephalic pedunculopontine and lateral dorsal tegmental nuclei (Beninato and Spencer 1986; Hall et al. 1989; McHaffie et al. 1991; Scheiner et al. 2000). In these regions, 90% of nitrergic neurons colocalize with the neurotransmitter acetylcholine (Beninato and Spencer 1986; Scheiner et al. 2000), indicating that nitrergic fibers in the SC from these extrinsic sources are cholinergic. By contrast, the intrinsic nitrergic neurons described herein, while they have a similar areal distribution as the extrinsic nitrergic fibers, are considerably smaller in number relative to the total SC neuropil (Scheiner et al. 2000). Moreover, unlike their extrinsic counterparts, SC nitrergic interneurons do not colocalize acetylcholine (Scheiner et al. 2000) but, as the present data reveal, almost half (46%) colocalize GABA, an incidence considerably higher than the 5% colocalization previously reported in ferret (Behan et al. 2002). Whether this discrepancy is indicative of a species difference or reflective of methodological factors (including the specific GABA and/or NOS antibodies used and different processing techniques) is not known. At present, what other neurotransmitter(s) is/are coexpressed by the remaining non-GABAergic nitrergic interneurons remains to be determined. However, given that there is no evidence that SC neurons contain glycine (Fujiwara et al. 1998), the other major inhibitory neurotransmitter, it is likely that the remaining corticorecipient nitrergic interneurons colocalize an excitatory neurotransmitter(s). This postulate is consistent with polysynaptic activation of SC output neurons following electrical activation of AES (Wallace et al. 1993).
Surprisingly little consideration has been devoted to the potential contribution of interneurons to multisensory processing in the SC. A previous study indicated that after electrical stimulation of AES neurons, a small percentage (14%) of SC output neurons were activated orthodromically with latencies presumed to be polysynaptic (Wallace et al. 1993). These data, however, likely underestimate the contribution of interneurons because subthreshold currents cannot be determined with extracellular recordings. Particularly germane in this context is our observation that almost half of the nitrergic interneurons coexpress GABA, whose hyperpolarizing synaptic influences could not have been detected in such experiments. Nevertheless, given the neuroanatomical distinctiveness of intrinsic SC nitrergic neurons and their relative paucity, it is remarkable that corticotectal afferents contact such interneurons with such frequency.
Given the anatomical and neurochemical complexity existing among cortical afferents, SC output neurons, and nitrergic interneurons, there are 2 possible routes by which AES influences can be exerted. One is a monosynaptic circuit that involves direct convergence of AES afferents on SC output neurons (Fig. 7A), the other, suggested by the present data, is a polysynaptic circuit that involves output neurons and nitrergic interneurons (Fig. 7B,C). Taken together, the data indicate that corticorecipient nitrergic interneurons are positioned to convey cortical influences onto SC multisensory output neurons disynaptically via conventional neurotransmitter systems utilizing GABA and other, presumably excitatory, neurotransmitters. In addition, and perhaps more importantly, because NO also acts as a retrograde messenger, cortically mediated NO release from the postsynaptic elements of these nitrergic interneurons could have a widespread influence on the adjacent neuropil and, particularly, on cortical afferents that establish direct contact with SC output neurons (Fig. 8). This is consistent with previous electron microscopic observations showing that interneurons participate in serial synapses and/or “triad-like” arrangements with cortical afferents and dendritic processes of output neurons (Norita 1980), including AES afferents (Harting et al. 1997). Whether this intimate relationship between cortical afferents and nitrergic interneurons is unique to AES afferents, or whether it is a feature of cortical afferents in general, is not known.
Although the SC has one of the highest concentrations of NO in the brain (Bredt et al. 1991; Vincent and Kimura 1992), and the distribution and morphology of its nitrergic neurons are well documented (Behan et al. 2002; Scheiner et al. 2000; Soares-Mota et al. 2001), nothing is known about the role NO plays in the normal physiology of SC output neurons. During development, however, NO has been implicated in the refinement of visual pathways in the superficial SC layers (Mize et al. 1996; Chalupa and Snider 1998). Given the intimate association of nitrergic interneurons with AES terminals, it is useful to consider this relationship from a developmental perspective. Cortical control over multisensory integration in the deep SC is absent at birth and develops gradually over a protracted postnatal period (Wallace and Stein 1997). Interestingly, NADPH staining in the deeper layers does not reach its peak intensity until day 35 (Scheiner et al. 2001), a time when multisensory neurons begin displaying adult-like integrative properties (Wallace and Stein 1997).
Physiological evidence suggests that only unisensory neurons within the different AES subregions converge on individual SC multisensory neurons (Wallace et al. 1993). As a consequence, the synthesis of cross-modal inputs occurs within the SC output neuron itself; it is not simply relayed from cortical multisensory neurons. The present anatomical results allude to a similar segregation of cortical afferents, as the present observations indicate that the majority of the nitrergic interneurons received inputs from a single AES subregion and not converging inputs from multiple subregions. Although these data are suggestive, a very strong caveat mitigates drawing firm conclusions. Because only the soma and proximal dendrites of these interneurons were immunostained, one cannot exclude the possibility that converging inputs from AES subdivisions included unlabeled, more distal dendrites. Moreover, demonstrating convergence would likely require an optimal topographical registration among AES injection sites; this becomes even more critical as the surface area of the target neuron becomes smaller. Also favoring such a possibility is the observation that, because direct cortico-SC projections to multisensory SC neurons target distal dendrites, proximal dendrites, and the soma, convergence can involve a variety of different patterns (Fuentes-Santamaria et al. 2004). Although a few medium-sized NOS-immunostained neurons were observed receiving converging inputs from more than one subregion of the AES, it is likely that these medium neurons, which colocalized SMI-32, are indeed projection neurons and not interneurons.
Because NO is a highly diffusible gas that can spread over several hundred microns, its effects can be exerted at a considerable distance from the release site, thereby influencing not only nitrergic interneurons but also nearby large multisensory output neurons. These effects can be both synaptic and nonsynaptic (Kiss and Vizi 2001; Vizi et al. 2004). Thus, the net effect of activating nitrergic interneurons is likely to reflect a number of complex interactive processes that can be engaged simultaneously and/or with some temporal disparities. For example, in those synaptic configurations involving interneurons with both NO and GABA, a GABAergic inhibitory effect can be exerted directly on the multisensory output neuron (similar to the effects observed in the paraventricular nucleus, see Bains and Ferguson 1997, and prepositus hypoglossi, see Moreno-Lopez et al. 2002). In addition, NO release also is known to modify the effects of several excitatory neurotransmitters (for review, see Esplugues 2002), including those that are widely distributed in the multisensory layers of the SC, such as acetylcholine and glutamate (for review, see Binns 1999). Understanding the timing and interplay of the various synaptic and nonsynaptic processes induced by corticorecipient nitrergic interneurons, although challenging, could provide critical insights into how AES facilitates the integration of cross-modal inputs that is so important in the control of overt behavior.
The National Institutes of Health (NS36916, NS22543, NS35008, EY016716); Wallace Research Foundation; Siebert Neuroscience Endowment.
We thank N. London for editorial assistance. Conflict of Interest: None declared.