It is established that glutamate is a neurotransmitter in most excitatory neurons in the central nervous system. Glutamate is loaded into synaptic vesicles by vesicular glutamate transporters (VGluTs) prior to exocytotic release. At least three VGluT isoforms have been cloned, namely VGluT1, VGluT2, and VGluT3. VGluT1 was initially characterized as a brain-specific Na+-dependent inorganic phosphate transporter (Ni et al.,1994; Bellocchio et al.,2000). VGluT2 was cloned as a differentiation-associated Na+-dependent inorganic phosphate transporter (Aihara et al.,2000). A third member was cloned and termed VGluT3 (Fremeau et al.,2002; Gras et al.,2002; Schäfer et al.,2002), which is structurally and functionally related to VGluT1 and VGluT2. In situ hybridization and immunohistochemical studies demonstrated that the mRNA and protein of VGluT1 and VGluT2 are present in the glutamatergic neurons of the large brain area (Ni et al.,1994; Kaneko et al.,2002; Li et al.,2003a,b). Whereas, VGluT3 mRNA and protein are found in specific cholinergic and serotonergic neurons (Fremeau et al.,2002; Gras et al.,2002; Schäfer et al.,2002).
In the peripheral sensory ganglia, such as dorsal root and trigeminal ganglion, the presence of VGluT1 mRNA and protein has been demonstrated in all sizes of sensory neurons, although predominately in medium–large neurons. VGluT2 is present in sensory ganglion, however the proportion and size of VGluT2 neurons is somewhat controversial (Li et al.,2003a,b; Oliveira et al.,2003; Todd et al.,2003; Landry et al.,2004; Brumovsky et al.,2007). Intriguingly, little is known about the occurrence of VGluT3 in sensory ganglia. In addition, the presence of VGluT has been documented in the peripheral endings of neurons in the dorsal root and trigeminal ganglion. In the trigeminal nervous system, muscle spindles of jaw closer display VGluT1 immunoreactivity (Pang et al.,2006; Lund et al.,2010). This originates from neurons in the trigeminal mesencephalic nucleus showing immunoreactivity for VGluT1 (Pang et al.,2006). Neurons in the trigeminal mesencephalic nucleus innervate the muscle spindle of the jaw closer and periodontal mechanoreceptors.
The periodontal ligament receives dense sensory innervation by nociceptive-free nerve endings and mechanoreceptive specialized endings. Although various types of mechanoreceptors have been reported in the periodontal ligament, the Ruffini ending is an essential mechanoreceptor (for review, Maeda et al.,1999; Wakisaka et al.,2000). Morphologically, rat periodontal Ruffini endings are characterized as having expanded axon terminals and an association with specialized Schwann cells called lamellar or terminal Schwann cells. In addition, the axon terminals of periodontal Ruffini endings contain various sized vesicles (Byers1985; Maeda et al.,1989). Recently, we reported the presence of synapse-related proteins such as the 25 kDa synaptosomal-associated protein (SNAP-25) around the vesicle membranes at the axon terminal of periodontal Ruffini endings (Honma et al.,2010). Taken together with other evidence, we speculated that VGluT is present in the periodontal Ruffini endings. Thus, we examined the presence and localization of VGluT isoforms in the periodontal ligament of the rat incisor at both light and electron microscopic levels.
MATERIALS AND METHODS
Eight male Sprague-Dawley rats, weighing 200–250 g, were purchased from Nihon Doubutsu (Osaka, Japan). The protocol for animal experiments was reviewed and approved by the Intramural Animal Use and Care Committee of the Osaka University Graduate School of Dentistry, prior to the commencement of experiments.
Three animals were euthanized with an overdose of chloral hydrate (600 mg/kg, i.p.), and the trigeminal ganglia was carefully removed. Total RNA was isolated using the SV Total RNA isolation system (Progema, Madison, WI) according to the manufacturer instructions. First-strand cDNA synthesis was performed by reverse transcription with 2 μg of total RNA. The sequences of PCR primers used are shown in Table 1.
Table 1. PCR primers for amplification of VGluT gene isoforms
PCR amplification was carried out under the following conditions: 94°C for 30 sec, 57°C for 30 sec, 72°C for 30 sec; these steps were repeated for 35 cycles followed by a final elongation step at 72°C for 7 min. Amplification products were analyzed by 1.5% agarose gel electrophoresis and visualized with ethidium bromide under UV illumination.
Five animals were deeply anesthetized with chloral hydrate (600 mg/kg, i.p.) and transcardially perfused with 0.02 M phosphate-buffered saline (PBS; pH 7.4) and then with a mixture of 4% paraformaldehyde and 0.05% glutaraldehyde in 0.1-M phosphate buffer (PB, pH 7.2). Trigeminal ganglion and the lower jaw was removed and further fixed in 4% paraformaldehyde in 0.1 M PB for 2–3 days. Lower jaws were decalcified with 7.5% ethylene diamine tetraacetic acid for 4–6 weeks at 4°C under gentle agitation. Decalcifying solution was changed at least twice a week. All specimens were soaked in 20% sucrose/PBS for cryoprotection. Lower jaws were sectioned with a cryostat at a thickness of 30 μm (for light microscopic analysis) or 50 μm (for electron microscopic analysis), collected in PBS and treated as floating sections. Trigeminal ganglion was sectioned at a thickness of 30 μm and also treated as floating sections.
Sections were treated with PBS containing 0.03% H2O2 to inactivate endogenous peroxidase activity for 30 min at room temperature. Following preincubation in PBS containing 3% normal goat serum (Vector Laboratories, Burlingame, CA) and 1% bovine serum albumin (BSA; Sigma, St. Louis, MO) for 30 min, sections were incubated with one of the following polyclonal primary antibodies raised in guinea pig for 16–18 hr at room temperature; anti-VGluT1 (1:5000; Chemicon International, Temecula, CA), anti-VGluT2 (1:5000; Chemicon) and anti-VGluT3 (1:5000; Chemicon). Following several rinses with PBS, sections were incubated with goat biotinylated anti-guinea pig IgG (1:100; Vector), and subsequently incubated with Avidin-Biotin Complex (Vector). The antigen-antibody binding sites were visualized by incubation with 0.05-M Tris-HCl buffer containing 0.08% diamino benzidine (DAB) and 0.003% H2O2. Reactions were enhanced by 0.1% nickel ammonium sulfate (NAS). Following the DAB reaction, immunostained sections were mounted on gelatin-subbed glass slides and lightly counter-stained with methyl green. Sections were dehydrated through a concentration ascending series of ethanol washes, cleared with xylene and cover slipped with Permount (Fisher Scientific, NJ).
For electron microscopic analysis, sections were immunostained according the protocol mentioned above, although NAS intensification was omitted. The immunostained sections were further fixed with 1% OsO4 reduced with 1.5% potassium ferrocyanide for 60 min at 4°C. Sections were dehydrated through a concentration ascending series of ethanol washes, transferred to propylene oxide and embedded with Epok 812 resin. Ultrathin sections were prepared with a diamond knife and examined with a transmission electron microscope (HT-7650; Hitachi, Tokyo, Japan) following brief staining with uranyl acetate and lead citrate.
For double immunofluorescence, sections were incubated with a mixture of polyclonal rabbit anti-VGluT1 (1:500; Synapse System, Göttingen, Germany) and polyclonal guinea pig anti-VGluT2 (1:500) or polyclonal guinea pig anti-VGluT3 (1:500). Some sections were incubated with a mixture of polyclonal rabbit anti-protein gene product 9.5 (PGP 9.5; 1:1000; Ultraclone, UK) and one of the polyclonal antibodies against the VGluT isoform raised in guinea pig at a dilution of 1:1000 for 16–18 hr at room temperature. The sections were then incubated with a mixture of Alexa Fluor 488-conjugated anti-rabbit IgG (1:1000; Molecular Probes) and Cy3-conjugated anti-guinea pig IgG (1:500; Molecular Probes) for 90 min at room temperature. Following cover-slip with Vectashield (Vector), the immunostained sections were viewed with a fluorescence microscope (AxioSkop 2; Carl Zeiss) and images captured with CCD camera software (AxioCom Ver. 4.0; Carl Zeiss), or conforcal laser scanning microscopy (LSM700, Carl Zeiss). The images were transferred to Adobe Photoshop (ver. 5.0).
Immunohistochemical controls were performed by omission of the primary antibody, secondary antibody or ABC complex. There were no observable immunoreactions in controls.
The cross sectional areas of trigeminal ganglion cells showing each isoform of VGluT immunoreactivity was measured using Win ROOF software (ver. 5.0, Mitani Co. Osaka, Japan). Comparisons of the measured cell size means were assessed by the Mann-Whitney U-test. P-values of < 0.05 were considered statistically significant.
VGluT mRNA and Protein in the Trigeminal Ganglion
RT-PCR for mRNA of each VGluT subtype revealed the presence of mRNA for all subtypes (Fig. 1).
Immunoreactions for all VGluT subtypes were localized in the cytoplasm of trigeminal ganglion neurons. Double-labeling of VGluT1 and VGluT2, and VGluT1 and VGluT3 revealed co-localization of individual VGluT subtypes (Figs. 2 and 3). The size range (and mean ± SD) of VGluT1- and VGluT2-immunoreactive trigeminal neurons from 20 randomly selected double-labeled images was 205.6–2564.0 μm2 (784.8 ± 342.6 μm2; n = 499) for VGluT1, and 74.1–2564.0 μm2 (786.0 ± 340.8 μm2; n = 401) for VGluT2. There was no statistical difference in the cross sectional area distribution between each group. Approximately 79.8% (398/499) of VGluT1-immunoreactive neurons displayed VGluT2 immunoreactivity, and approximately 99.3% (398/401) of VGluT2-immunoreactive neurons exhibited immunoreactivity for VGluT1. The size range (and mean ± SD) of neurons double-labeled with VGluT1 and VGluT2 was 205.6–2564.0 μm2 (791.2 ± 336.6 μm2; n = 398) (Fig. 2a–d).
Double-labeling of VGluT1 and VGluT3 immunoreactivity showed co-localization of these two proteins in trigeminal ganglion neurons (Fig. 3a–c). The size range (and mean ± SD) of VGluT1- and VGluT3-immunoreactive neurons was 271.3–2433.4 μm2 (809.0 ± 354.1 μm2; n = 469) for VGluT1, and 271.3–2205.9 μm2 (815.2 ± 358.8 μm2; n = 406) for VGluT3. There was no significant difference between the two groups. Approximately 84.6 % (397/469) of VGluT1-immunoreactive neurons displayed VGluT3 immunoreactivity, and approximately 97.8 % (397/406) of VGluT3-immunoreactive neurons showed VGluT1-immunoreactivity. The size range (and mean ± SD) of neurons expressing both VGluT1 and VGluT3 was 271.3–2205.9 μm2 (825.9 ± 355.6 μm2; n = 397) (Fig. 3d).
VGluT in Periodontal Ligament
In the lingual periodontal ligament of the rat incisor, two types of expanded nerve endings, namely Ruffini endings were recognized as revealed by PGP 9.5 immunohistochemistry (Fig. 4a). Ruffini endings composed of relatively thick nerve fibers and those composed of relatively thin nerve fibers. Double immunolabeling of PGP 9.5 and VGluT1 revealed almost all PGP 9.5 immunopositive elements showed immunoreactions for VGluT1 (Fig. 4a–c). Under electron microscopy, reaction products were localized in the axon terminal of Ruffini endings, which is filled with large numbers of mitochondria (Fig. 4d). These immunoreactions were densely detected around the vesicle membranes (Fig. 4e). The Schwann sheath surrounding immunopositive axon terminals was devoid of immunoreactions (Fig. 4d).
Double immunolabeling of VGluT1 and VGluT2 exhibited that these two subtypes were co-localized in the Ruffini endings (Fig. 5). In contrast to VGluT1 and VGluT2, VGluT3 immunoreactivity was not detected in the periodontal ligament of the rat incisor even when apparent Ruffini endings were recognized with PGP 9.5 (Fig. 6).
The present immunohistochemical study reveals the presence of all VGluT subtypes in trigeminal ganglion neurons. Moreover, both VGluT1 and VGluT2 were expressed in the periodontal Ruffini endings of the rat incisor, while VGluT3 was not detected.
The presence of VGluT1 and VGluT2 mRNA and protein has been studied in dorsal root and trigeminal ganglion (Li et al.,2003a,b; Oliveira et al.,2003; Todd et al.,2003; Landry et al.,2004; Brumovsky et al.,2007). Based on these studies, VGluT1 is present in all sizes of sensory neurons, predominately in many medium–large sensory neurons. Our present results are in agreement with these studies; VGluT1 is present predominately in medium–large trigeminal neurons of the rat. However, the proportion of sensory neurons expressing VGluT2 is somewhat controversial. Li et al. (2003b) found that, similar to VGluT1, VGluT2 is present in all sizes of trigeminal neurons, predominately in medium–large neurons, and VGluT2-immunoreactive neurons were more frequently observed than VGluT1-immunoreactive neurons in the rat trigeminal ganglion. They also revealed that approximately 90% of VGluT1 neurons co-localized with VGluT2 and 83% of VGluT2 neurons were VGluT1-immunoreactive. However, in the mouse dorsal root ganglion, Brumovsky et al. (2007) reported that VGluT2 is expressed in small–medium dorsal root ganglion with moderate co-expression of VGluT1 and VGluT2 usually in medium–large-neurons (less than half of all VGluT1 neurons were also VGluT2-positive). Our present results were similar to those by Li et al. (2003b), although we found that VGluT1 is expressed more frequently than VGluT2. Also, approximately 80% of VGluT1-neurons expressed VGluT2 and almost all VGluT2 neurons showed VGluT1. The differences in results may be accounted for by the different antibodies used. Even though it is safe to say that many VGluT1-immunoreactive trigeminal ganglion neurons co-expressed VGluT2 and vice versa. In contrast to VGluT1 and VGluT2, little is known about VGluT3 in the sensory ganglion. In the dorsal root ganglion, both VGluT3 mRNA and protein was rarely detected (Oliveira et al.,2003; Landry et al.,2004). However, we could clearly detect VGluT3 mRNA by RT-PCR analysis and many VGluT3-immunoreactive neurons. These neurons also expressed VGluT1 in the trigeminal ganglion.
Trigeminal and dorsal root ganglion neurons are morphologically pseudo-unipolar neurons, i.e., their axons extend both peripherally and central bilaterally. As axon terminals lack the biosynthesis machinery necessary to produce proteins, all VGluT isoforms were found to be synthesized in the cell bodies of trigeminal ganglion neurons as shown by RT-PCR analysis. Those proteins are transported peripherally and centrally bilaterally. Therefore, it is reasonable that VGluTs are expressed in the medullary dorsal horn (Li et al.,2003a) where trigeminal neuron central processes terminate. Although occurrence and distribution of VGluT in the sensory ganglia has been well documented by many groups, little is known about the expression of VGluT in the peripheral nerve terminals. It is generally accepted that there is a linear correlation between cell diameter and conduction velocity in the dorsal root ganglion. This indicates that the medium–large neurons innervate the proprioceptors and mechanoreceptors, while small neurons innervate nociceptors. Wu et al. (2004) reported the presence of VGluT1 in the muscle spindles of rat triceps surae muscle. In the trigeminal system, the muscle spindles of the rat masseteric nerve displayed VGluT1 immunoreactivity that originated from the trigeminal mesencephalic nucleus neurons (Pang et al.,2006, Lund et al.,2010). Thus, it is safe to say that VGluT1 present in the nerve endings originates from medium–large neurons, because muscle spindles are innervated by medium–large sensory neurons. As shown in this study, the presence of VGluT1 in the periodontal Ruffini endings is in agreement with this speculation as periodontal Ruffini endings of the rat incisor are mechanoreceptors originating from medium–large trigeminal ganglion neurons. In addition to proprioceptors and mechanoreceptors, VGluT1-immunoreactive nerve fibers have been documented in the epidermis of mouse hind paw glabrous skin. These terminate as free nerve endings, which may originate from small dorsal root ganglion neurons (Brumovsky et al.,2007). Although the rat incisor periodontal ligament also contains nociceptive free nerve endings, we could not detect VGluT1 in the periodontal free nerve endings due to sparse distribution of free nerve endings in the ligament. Our preliminary experiments demonstrated VGluT1 in the free nerve endings in gingivae and palatal epithelium, and intrapulpal nerve fibers of the molar teeth (unpublished observations).
This study showed that many VGluT1-immunoreactive trigeminal ganglion neurons co-expressed VGluT2 immunoreactivity, suggesting that sensory nerve endings co-express VGluT1 and VGluT2. This was confirmed with VGluT1 and VGluT2 co-localization in the periodontal Ruffini endings.
In contrast to VGluT1 and VGluT2, there is only one report on the presence of VGluT3 in peripheral tissues. Nunzi et al. (2004) reported that all three VGluT subtypes are present in the free nerve endings of mouse palatine mucosa and in the Merkel-neurite complex. In this study, we found that the periodontal Ruffini endings of rat incisors showed both VGluT1 and VGluT2, and lacked VGluT3, although many VGluT1-immunoreactive trigeminal ganglion neurons also expressed VGluT3. There are two possible explanations; periodontal Ruffini endings of rat the incisor originated from trigeminal neurons having both VGluT1 and VGluT2, although lacked VGluT3. Another is that the levels of VGluT3 in the periodontal Ruffini endings are too low to detect with current methods. Although the periodontal Ruffini endings in the trigeminal neurons have all three VGluT subtypes.
Functional significance of VGluT in primary sensory endings remains unknown. In the central nervous system, glutamate is loaded into synaptic vesicles by VGluT. Thus, VGluT is a good marker for glutamatergic neurons. However, in the peripheral axon terminals there is no apparent synapse with few exceptions such as the Merkel-neurite complex. Axon terminals of the periodontal Ruffini endings contain many types of vesicles (Byers, 1986; Maeda et al.,1989). Including the presence of SNAP-25 and syntaxin, members of synapse-related proteins, around the vesicles of periodontal Ruffini ending axon terminals (Honma et al.,2010). In this study, we demonstrated that VGluT1 is present around the vesicles of the axon terminals of periodontal Ruffini endings. However, it is uncertain whether glutamate may be released from the periodontal Ruffini endings and function as a neurotransmitter because glutamate is the most abundant excitatory neurotransmitter in the central nervous system. In the muscle spindles, Bewick et al. (2005) proposed that glutamate release has an autogenic effect on excitability of the sensory endings. Thus, similar to the muscle spindle, we speculated that periodontal Ruffini endings released glutamate from axon terminals and this might contribute to the excitability of the endings. Besides the function as a neurotransmitter, glutamate is a general metabolic substrate. The presence of VGluT in non-neuronal tissues such as calcified tissues has been reported (see review Hinoi et al,2004). Therefore, glutamate has metabotropic effects in the Ruffini endings.