Brain Research 864 (2000) 152–156www.elsevier.com/ locate /bres

Short communication

Neurocalcin-immunoreactive primary sensory neurons in the trigeminalganglion provide myelinated innervation to the tooth pulp and

periodontal ligamenta,b , c a,b*Hiroyuki Ichikawa , Hiroyoshi Hidaka , Tomosada Sugimoto

aSecond Department of Oral Anatomy, Okayama University Dental School, 2-5-1 Shikata-cho, Okayama 700, JapanbBiodental Research Center, Okayama University Dental School, 2-5-1 Shikata-cho, Okayama 700, Japan

cDepartment of Pharmacology, Nagoya University School of Medicine, 65 Tsurumai, Showa-ku, Nagoya 466, Japan

Accepted 15 February 2000

Abstract

The distribution of neurocalcin-immunoreactive (NC-ir) primary sensory neurons was examined in the trigeminal ganglion (TG),mesencephalic trigeminal tract nucleus (Mes5) and intraoral structures. NC-ir primary sensory neurons were located in the TG but not theMes5. The coexpression study demonstrated that virtually all NC-ir TG neurons exhibited S100-immunoreactivity (-ir). In the tooth pulp,NC-ir nerve fibers were observed in the subodontoblastic and odontoblastic layers. Immunoelectron microscopic and retrograde tracingmethods revealed that myelinated pulpal axons derived from the TG mostly exhibited the ir. In the periodontal ligament, bush-like endingsshowed NC-ir. These endings were morphologically identical to Ruffini-like endings. The present study suggests that NC-ir trigeminalprimary sensory neurons have their cell bodies in the TG. Their peripheral axons are probably myelinated. Such neurons include pulpalnociceptors and low-threshold mechanoreceptors. 2000 Elsevier Science B.V. All rights reserved.

Themes: Sensory systems

Topics: Somatic and visceral afferents

Keywords: Immunohistochemistry; Neurocalcin; Periodontal ligament; S100; Tooth pulp; Trigeminal ganglion

Neurocalcin (NC) is a calcium-binding protein (CaBP) number is small [13]. Therefore, NC-ir TG cells arewhich was isolated from bovine brain [15]. Immuno- unlikely to be proprioceptors.histochemical studies have revealed that this CaBP is In this study, we examined the distribution of NC-irlocated in the dorsal root (DRG) and trigeminal ganglia neurons in the TG, Mes5 and oro-facial tissues. The(TG) [9,12]. In these ganglia, NC-immunoreactive (-ir) coexpression of this protein with S100, a marker forneurons are mostly medium-sized or large [9]. The small- medium-sized and large primary sensory neurons [3,4,7],est class of DRG and TG neurons was rarely NC-ir. In the was also investigated in the TG by a double immuno-spinal nervous system, NC has been detected in peripheral fluorescence method. In addition, retrograde tracing andaxons innervating muscle spindles [10,12]. Thus, NC-ir immunoelectron microscopic methods were performed toDRG neurons appear to include muscular proprioceptors. know the origin and ultrastructure of NC-ir nerve fibers.In the trigeminal system, however, proprioceptors innervat- Adult male Sprague–Dawley rats (200–300 g) wereing the masticatory muscles and periodontal ligaments used. Four animals were anesthetized with ether to thehave their cell bodies in the mesencephalic trigeminal tract level at which respiration was markedly suppressed, andnucleus (Mes5). Although primary neuronal cell bodies transvascularly perfused with 50 ml of saline followed byinnervating stretch receptors in the extrinsic eye muscles 500 ml of 4% formaldehyde in 0.1 M phosphate bufferare located in the ophthalmic division of the TG, their (pH 7.4). The TGs, brainstems containing the Mes5,

maxillae and mandibles were dissected and post-fixed withthe same fixative for 30 min. Maxillae and mandibles were

*Corresponding author. Fax: 181-86-222-4572. decalcified with 4.13% ethylene diaminetetraacetic acid

0006-8993/00/$ – see front matter 2000 Elsevier Science B.V. All rights reserved.PI I : S0006-8993( 00 )02175-2

H. Ichikawa et al. / Brain Research 864 (2000) 152 –156 153

2disodium salt (EDTA) in 0.1 M phosphate buffer (pH 7.4) NC-ir neurons measured 228–3400 mm (mean1S.D.52for 1 week at room temperature. All materials were soaked 14961627 mm , n5668). These neurons were mostly

2in a phosphate-buffered 20% sucrose solution overnight, medium-sized (400–800 mm ; 10.5% or 70/668) or large2frozen sectioned at 15 mm and thaw-mounted on gelatin- (.800 mm ; 88.6% or 592/668). Small NC-ir neurons

2coated glass slides. For the demonstration of NC, an ABC were very rare in the TG (,400 mm ; 0.9% or 6/668). In(avidin–biotin–horseradish peroxidase complex) method the Mes5, primary sensory neurons were devoid of the irwas performed. Sections were incubated with rabbit anti- (Fig. 1B). The double immunofluorescence method re-NC serum (1:6000) [12] followed by the incubation with vealed the coexpression of NC and S100 throughout thebiotinylated goat anti-rabbit IgG and ABC complex (Vec- ganglion (Fig. 1C, D). Virtually all (1994/2002 or 99.6%)tor Laboratories). The immunoreactivity (ir) was visualized NC-ir neurons expressed S100-ir and 94% (1994/2121) ofby nickel ammonium sulfate-intensified diaminobenzidine S100-ir neurons exhibited NC-ir.reaction. For cell size analysis, the microscopic image The inferior alveolar nerve contained numerous NC-ir(3215) of the cell body was projected over a digitizer nerve fibers (Fig. 1E). These nerve fibers mostly showed atablet using a drawing tube. The cross-sectional area of thick, smooth appearance. NC-ir smooth nerve fibers werethose cell bodies that contained the nucleolus was re- also seen in small bundles within the tooth pulp, periodon-corded. tal ligament, and the deep layer of gingiva and oral

For the coexpression of NC and S100 in the TG, mucosa. However, distinct NC-ir nerve endings could besections were incubated with a mixture of the rabbit observed only in the tooth pulp and periodontal ligamentanti-NC serum (1:200) and mouse anti-S100 monoclonal of incisor and molar teeth (Fig. 1F, G). In the root pulp ofantibody (1:1000, Sigma) overnight at room temperature. molar teeth, NC-ir nerve fibers projected straight until theySubsequently, sections were incubated in a mixture of reached the coronal pulp. No NC-ir fiber was detected inlissamine rhodamine B chloride-conjugated donkey anti- the vicinity of the odontoblast in the root pulp. Accom-rabbit IgG (1:500, Jackson ImmunoResearch Labs) and panied by blood vessels, NC-ir nerve fibers ascendedfluorescein isothiocyanate-conjugated donkey anti-mouse toward the pulp horn. In the subodontoblastic layer, theseIgG (1:100, Jackson ImmunoResearch Labs). nerve fibers ramified repeatedly, formed nerve plexuses

For demonstration of the origin of NC-ir nerve fibers and reached the base of the odontoblastic layer (Fig. 1F).innervating the tooth pulp, two male rats (300–350 g) Some of them entered the odontoblastic layer. However,were used. Under deep anesthesia by i.p. injection with NC-ir nerve fibers could not be observed in the dentine. Inethyl carbamate (650 mg/kg) and pentobarbital sodium the incisor tooth pulp, a few smooth nerve fibers were(20 mg/kg), 1% fluorogold (FG, Fluorochrom) in distilled observed accompanying blood vessels. However, no NC-irwater was applied to the right first and second maxillary nerve fiber was detected in the vicinity of odontoblasts. Inmolar pulps. After 3 days, the animals were reanesthetized that part of the lingual periodontal ligament of the incisorwith ether, and transvascularly perfused with 4% formalde- which was adjacent to the alveolar bone, NC-ir nerve fibershyde. The right TG and the part of the brainstem including left the nerve bundles and ran towards the tooth root. In thethe Mes5 were frozen-sectioned and mounted on gelatin- ligament, halfway between the bone and tooth surfaces,coated glass slides. The brainstem of the rats, that had each fiber formed bush-like endings consisting of a few toreceived tracer application, was cryosectioned at 50 mm several terminal twigs at the distal end (Fig. 1G). Theand examined with fluorescent illumination for possible portion of the lingual periodontal ligament apposed to theFG labeling. Then, 12-mm thick sections of the ganglia tooth root and the tissue covering the labial portion of thewere processed for the coexpression of NC and S100 as incisor tooth were devoid of NC-ir nerve endings. In thedescribed above. molar periodontal ligament, a few bush-like endings were

For electron microscopic analysis, two additional ani- observed around the apical foramen. Terminal twigs withinmals were similarly fixed except that 0.05% glutaraldehyde such endings in the molar tooth appeared to be fewer thanwas added to the fixative. Unfrozen 50-mm thick sagittal in the incisor tooth (not shown).sections of the EDTA-decalcified mandibles were cut with At 3 days after FG application to the upper molar tootha Microslicer (Dosaka EM) and processed for an ABC pulps, many cell bodies were labeled in the TG (Fig. 1H).method as described above. Following diaminobenzidine They were located in the maxillary division of the gang-reaction, the sections were postfixed in 1% osmium lion. All these FG-labeled cells are considered to betetroxide in 0.1 M phosphate buffer (pH 7.4), dehydrated primary neurons innervating the tooth pulp and the possi-through a graded series of alcohols, and embedded in bility of periodontal spread of FG through the apicalPolybed 812. Ultrathin sections were examined after foramen is negligible, because the labeled neuron was notstaining with lead citrate for 1 min. observed in the Mes5. The retrograde labeling and im-

The specificities of the antibodies used have been munohistochemical methods revealed that FG-labeled TGalready described elsewhere [4,12]. neurons contained NC- and S100-irs (Fig. 1H–J). A total

As described previously [9], numerous NC-ir neuronal of 53.9% (83/154) and 59.1% (91/154) of the labeledcell bodies were scattered throughout the TG (Fig. 1A). cells were immunoreactive for NC and S100, respectively.

154 H. Ichikawa et al. / Brain Research 864 (2000) 152 –156

Fig. 1. Micrographs of NC-ir (A–C, E–G, I, K–M), S100-ir (D, J) and FG (H) in the TG (A, C, D, H–J), Mes5 (B), inferior alveolar nerve (E), tooth pulp(F, K, L) and periodontal ligament (G, M). Panels C, D and panels H–J are from the same sections, respectively. Numerous TG neurons exhibit NC-ir (A),while Mes5 primary sensory neurons are devoid of it (arrows in B). In the TG, NC-ir neurons (arrows in C) mostly coexpress S100 (arrows in D). Theinferior alveolar nerve contains NC-ir thick nerve fibers. In the molar tooth pulp, NC-ir nerve fibers make subodontoblastic (sub) nerve plexuses and someof them enter the odontoblastic layer (od). Bush-like endings consisting of some terminal twigs contain NC-ir in the lingual periodontal ligament of theincisor tooth (arrows in G). TG neurons retrogradely labeled from the molar tooth pulp with FG (arrows in H) show both NC-ir (arrows in I) and S100-ir(arrows in J). In the root pulp, myelinated and large unmyelinated axons exhibit NC-ir (arrows in K). Arrowheads in K point to NC-immunonegativeunmyelinated axons. In the tooth pulp, NC-ir nerve endings make close contact with odontoblasts (arrows in L). Ruffini-like endings exhibit NC-ir in thelabial periodontal ligament of the incisor. Arrowheads in M point to finger-like projections. Bars5100 mm (A–F, H–J), 50 mm (G), and 1 mm (K–M).

H. Ichikawa et al. / Brain Research 864 (2000) 152 –156 155

While 95.2% (79/83) of NC-ir FG-labeled neurons tal Ruffini-like endings derived from the TG are low-showed S100-ir and 86.8% (79/91) of S100-ir FG-labeled threshold mechanoreceptors which are thought to beneurons exhibited NC-ir (Fig. 1H–J). activated by touch, pressure and movement of teeth during

An immunoelectron microscopic method revealed that chewing, swallowing and speech [1]. Cell bodies ofNC-ir was localized to the axoplasm but not Schwann cells primary sensory neurons innervating periodontal Ruffini-in the tooth pulp and periodontal ligament. At a higher like endings are also located in the Mes5 [2]. However,power magnification, the immunoreaction products were NC-ir Ruffini-like endings are unlikely to originate fromextensively distributed over the neurofilaments, microtu- the Mes5 because Mes5 primary sensory neurons werebles and the membrane. In the root pulp, 61.4% (35/57) of devoid of the ir.axons were immunoreactive for NC (Fig. 1K). In all, 92% In conclusion, we have examined the distribution of(23/25) of myelinated axons (the short diameter of ir NC-ir primary sensory neurons in the trigeminal system.axons51.0–3.8 mm) exhbited NC-ir; 37.5% (12/32) of Such neurons were located in the TG but not the Mes5. Inunmyelinated axons showed the ir and such unmyelinated the TG, virtually all NC-ir neurons coexpressed S100-ir.axons were mostly large (the short diameter of ir unmyeli- These neurons supplied the tooth pulp with myelinatednated axons .0.4 mm; 91.7% or 11/12) (Fig. 1K). In the axons. In the periodontal ligament, Ruffini-like endingsodontoblastic layer, NC-ir neurites made close contact with also showed NC-ir. NC-ir TG neurons are probablyodontoblasts (Fig. 1L). No specialized junction was de- myelinated, and include pulpal nociceptors and low-thres-tected between the neurites and odontoblasts. NC-ir nerve hold mechanoreceptors.fibers could not be observed in the dentine. In the lingualperiodontal ligament, the morphology of NC-ir bush-likeendings was identical to that of Ruffini-like endings which Referenceshave been described previously [6,8,11,14]. Their axop-lasm was enriched with mitochondria and formed finger-

[1] M.R. Byers, W.K. Dong, Comparison of trigeminal receptor locationlike projections that extended into the surrounding inter- and structure in the periodontal ligament of different types of teethcellular space (Fig. 1M). Schwann cells in these endings from the rat, cat, and monkey, J. Comp. Neurol. 279 (1989)

117–127.formed lamellar sheets around the axoplasm.[2] M.R. Byers, A. O’Connor, R.F. Martin, W.K. Dong, MesencephalicThe present study demonstrated the presence of NC-ir

trigeminal sensory neurons of cat: Axon pathway and structure ofprimary sensory neurons in the TG but not the Mes5. Mes5mechanoreceptive endings in periodontal ligament, J. Comp. Neurol.

primary sensory neurons are exclusively muscular and 250 (1986) 181–191.periodontal proprioceptors. Unlike in the spinal system, [3] H. Ichikawa, Y.-F. He, T. Sugimoto, S100 protein-immunoreactive

trigeminal neurons innervating the rat molar tooth pulp, Somatosens.therefore, trigeminal proprioceptors appear to be devoid ofMotor Res. 15 (1998) 128–133.NC-ir. This study also showed the coexpression of NC and

[4] H. Ichikawa, C.J. Helke, Coexistence of S100b and putativeS100. Virtually all NC-ir TG neurons exhibited S100-ir. Intransmitter agents in vagal and glossopharyngeal sensory neurons of

the TG, S100-ir neurons lack ir for tachykinin, a marker the rat, Brain Res. 800 (1998) 312–318.for unmyelinated primary sensory neurons [5,7]. In addi- [5] H. Ichikawa, D.M. Jacobowitz, T. Sugimoto, Calretinin-immuno-tion, capsaicin which is a neurotoxin for unmyelinated reactive neurons in the trigeminal and dorsal root ganglia of the rat,

Brain Res. 617 (1993) 96–102.primary sensory neurons has no effect on S100-ir TG[6] H. Ichikawa, D.M. Jacobowitz, T. Sugimoto, Coexpression ofneurons. Thus, S100-ir neurons are considered to be

calretinin and parvalbumin in Ruffini-like endings in the rat incisormyelinated in the TG. The present coexpression finding periodontal ligament, Brain Res. 770 (1997) 294–297.suggests that NC-ir TG neurons have myelinated axons. It [7] H. Ichikawa, D.M. Jacobowitz, T. Sugimoto, S100 protein-immuno-is likely that NC is localized to myelinated non-prop- reactive primary sensory neurons in the trigeminal and dorsal root

ganglia of the rat, Brain Res. 748 (1997) 253–257.rioceptors in the trigeminal system.[8] H. Ichikawa, C. Xiao, Y.-F. He, T. Sugimoto, Parvalbumin-immuno-In this study, intraoral structures contained NC-ir nerve

reactive nerve endings in the periodontal ligaments of rat teeth,fibers and endings. NC-ir nerve endings were observed in Arch. Oral Biol. 41 (1996) 1087–1090.the subodontoblastic and odontoblastic layers of the molar [9] S. Iino, M. Kato, H. Hidaka, S. Kobayashi, Neurocalcin-immuno-tooth pulp. Our retrograde tracing method revealed that positive neurons in the rat sensory ganglia, Brain Res. 781 (1998)

236–243.NC-ir pulpal nerve fibers originated from the TG. These[10] S. Iino, S. Kobayashi, H. Hidaka, Neurocalcin-immunopositivefindings suggest that NC-ir TG neurons innervate pulpal

nerve terminals in the muscle spindle, Golgi tendon organ and motornociceptors. The present electron microscopic method also endplate, Brain Res. 808 (1998) 294–299.demonstrated that myelinated axons in the root pulp mostly [11] K. Kannari, O. Sato, T. Maeda, T. Iwanaga, T. Fujita, A possiblecontained NC-ir. NC-ir unmyelinated radicular axons mechanism of mechanoreception in Ruffini endings in the periodon-

tal ligament of hamster incisors, J. Comp. Neurol. 313 (1991)observed in this study may be preterminal segments of368–376.myelinated stem axons.

[12] K. Okazaki, S. Iino, S. Inoue, S. Kobayashi, H. Hidaka, DifferentialIn the periodontal ligament, bush-like endings showed distribution of neurocalcin isoforms in rat spinal cord, dorsal root

NC-ir. The fine structure of these periodontal endings was ganglia and muscle spindle, Biochem. Biophys. Acta 1223 (1994)identical to that of Ruffini-like endings [1,6,11]. Periodon- 311–317.

156 H. Ichikawa et al. / Brain Research 864 (2000) 152 –156

[13] J.D. Poter, B.L. Guthrie, D.L. Sparks, Innervation of monkey incisors, An immunohistochemical investigation of neurofilamentextraocular muscles: localization of sensory and motor neurons by protein and glia specific S-100 protein, Cell Tissue Res. 251 (1988)retrograde transport of horseradish peroxidase, J. Comp. Neurol. 218 13–21.(1983) 208–219. [15] M. Terasawa, A. Nakano, R. Kobayashi, H. Hidaka, Neurocalcin: a

[14] O. Sato, T. Maeda, S. Kobayashi, T. Iwanaga, T. Fujita, Y. novel calcium-binding protein from bovine brain, J. Biol. Chem. 267Takahashi, Innervation of periodontal ligament and dental pulp in rat (1992) 19596–19599.