The effect of nerve section on the incidence and distribution of gap junctions in the odontoblast layer of the cat.код для вставкиСкачать
THE ANATOMICAL RECORD 218:458-465 (1987) The Effect of Nerve Section on the Incidence and Distribution of Gap Junctions in the Odontoblast Layer of the Cat G.R. HOLLAND Division of Endodontics, Faculty of Dentistry, University of Alberta, Edmonton, Alberta, Canada T6G 2N8 ABSTRACT Gap junctions are numerous in the odontoblast layer of the dental pulp and may link sensory axons to odontoblasts. If these junctions do link axons and odontoblasts, they, together with the axons, should disappear after cutting the pulpal nerves centrally. Under general anesthesia the inferior alveolar nerve on one side of two young adult cats was sectioned. Under general anesthesia the animals were perfused with fixative 56 hours later and the corona1 dental pulp prepared for electron microscopy. Ultrathin sections were examined from the level of the pulpal cornu and levels approximately one, two, and three mm below this. The incidence of cell processes and gap junctions was measured at different distances from the pulp predentin junction, and operated and control sides compared. The odontoblast layer at the level of the cornu differed from elsewhere in having, on the control side, a greater density of cell processes and gap junctions and in having clearly recognizable axons approaching to within 5 to 10 pm of the predentin. The only statistically significant changes after nerve section occurred in this layer and consisted of a decline in the incidence of cell processes and of gap junctions that link one cell process to another. There was no significant difference between the operated and control sides in the number of gap junctions linking cell processes to recognizable cell bodies. The odontoblast layer in the pulpal cornu contained substantial numbers of unsheathed axons, many presumably en route to the dentin. These axons may participate in gap junctions that link them to other cell processes, possibly even other axons. No clear evidence was found of junctions involving axons and identifiable odontoblast cell bodies. A synapselike contact between odontoblasts and nerves would add structural support to two ideas; that the odontoblast may act as a sensory receptor and that sensory nerves have a trophic effect on odontoblastic activity. Other evidence for both these ideas is inconclusive in part due to the technical difficulties in examining such a remote and well-concealed cell, and in part due to conceptual difficulties in defining what constitutes a receptor and what is a direct neurotrophic effect (for review, see Holland, 1986a). Electrophysiological methods of recording the activity of odontoblasts, although promising direct evidence of excitability and synaptic transmission, are difficult. The most successful attempt thus far (Magloire et al., 1979) did not positively establish the receptive role or the neural link. Neural resection experiments (Avery et al., 1974)designed to establish a trophic function are fraught by secondary effects such as the loss of propioceptive input. Other structural studies have been equivocal. Conventional anatomical methods are cursed with the difficulty of identifying unsheathed axons in the processabundant odontoblast layer. The potentially most informative approaches using anterograde transport of ra@ 1987 ALAN R. LISS, INC. dioactive labels in nerve fibers (Byers and Kish, 1976; Byers, 1977, 1979) and tracing of nerve fibers in serial sections (Holland, 80a) could not demonstrate any convincing specialized junctions involving axons. Although earlier studies (Frank, 1968; Arwill, 1968) had reported synapselike structures in dentinal tubules, subsequent observations of labeled axons in dentin (Byers, 1977) or as a result of comparing intracellular filament and tubule populations (Holland, 1986)could not confirm these earlier findings. Certainly anything like a classical chemical synapse is absent from the odontoblast layer, and in our view the dentin a s well (Holland, 1986). The only potentially synaptic junction available in profusion is the gap junction in the odontoblast layer. The gap junction is known, in the central nervous system, to be capable of acting synaptically (Pappas and Bennet, 1966). Whereas these junctions are widely distributed in the peripheral pulp (Holland, 19771, no study has thus far established with Received December, 1986; accepted March 24, 1987. NERVE SECTION AND ODONTOBLAST GAP JUNCTIONS any degree of certainty that they link axons and odontoblasts. The most positive finding in this direction is the cytological similarity between the processes linked to odontoblasts and axons (Holland, 1980b).One obvious approach, which may establish whether these junctions involve axons or not, is to examine the effects of nerve section. Recent studies have examined the effect of denervation on the apical (Holland and Robinson, 1984) and dentinal (Holland et al., 1986) innervation. This study examines the changes brought about by shortterm denervation on the distribution of gap junctions in the odontoblast layer from some of the animals prepared for these earlier studies. MATERIALS AND METHODS Two young adult cats were used in these experiments. The animals were anesthetized with a mixture of alphaxalone and alphadolone acetate (Saffan, Glaxo Laboratories, U.K.). Body temperature was maintained with a n electric blanket controlled from a rectal thermistor. The left inferior alveolar nerve was exposed in the mandibular canal in the region of the anterior border of the masseter muscle. The nerve was sectioned and the cut ends replaced in apposition. Each animal was given penicillin postoperatively. Electrophysiological observations were made on these animals and have been reported elsewhere (Holland et al., 1986). Fifty-six hours after nerve section, the animals were, under general anesthesia, perfused with a fixative mixture via the common carotid arteries. The perfusion was initiated 459 with a prewash of 300 mosmol. pH 7.4 phosphate buffer (Hayat, 1970) in which 2.7% dextran had been added (Dextran T40, Pharmacia Fine Chemicals, Sweden: Aldskogius and Arvidsson, 1978). The perfusion was continued and completed with a similar mixture to which 5% glutaraldehyde had been added. After overnight immersion of the head in buffered glutaraldehyde without dextran, the lower canine teeth were removed and decalcified in a mixture of 4% EDTA and l%glutaraldehyde. The crowns of the teeth were slit transversely at 1 mm intervals and the resulting discs processed by washing in phosphate buffer, postfixation in 2% phosphate-buffered osmium tetroxide, en bloc staining with uranyl acetate, alcoholic dehydration, and embedment in Araldite resin. Ultrathin sections of each block from the corona1 half of the crown were cut. To ensure that sections close to the pulpal cornu were examined in each tooth, the block containing the cornu was step sectioned until the pulp was reached. Ultrathin sections were taken from the block containing the cornu and the three, 1-mm blocks below it. A single section containing a complete cross section of the pulp was cut a t each level from both control and operated teeth, a total of 16 sections. This enabled the position of each section to be related to the pulpal cornu. The sections were stained with uranyl acetate and lead citrate and examined in the electron microscope. Montage electron micrographs were taken of segments of the odontoblast layer from its junction with the predentin to a level 40 pm toward the central Fig. 1 , Low power micrograph of a typical area examined. This section was taken 1-2 mm below the pulpal cornua from a cat whose inferior alveolar nerve on the same side had been transected 2 days earlier. PD = predentin, OD = odontoblast. G.R. HOLLAND 460 TABLE 1. Total no. of Drocesses Distance from PDJ (pm) Level 1 Control operated 2 Control operated 3 Control operated 4 Control operated 5-10 0-5 10-15 68.3 (19.1) 114.0 (21.0)* 114.0 (8.5)* 38.5 (18.8) 35.5 (30.9) 48.0 (11.9) 60.0 (26.2) 46.5 (17.7)" 66.8 (5.7) 50.6 (8.9) 40.6 (20.7) 7.2 (8.5) 56.3 (15.6) 56.3 (10.1) 52.3 (28.5) 51.0 (22.1) 47.0 (6.3) 13.2 (9.4) 31.9 (15.1) 36.8 (16.4)* 54.5 (16.8) 42.5 (14.5) 45.0 (9.3) 9.0 (6.7) 15-20 20-25 25-30 30-35 35-40 89.3 (23.7)* 32.5 (5.0) 54.8 (15.8) 41.6 (11.7) 37.8 (13.1) 27.7 (10.7) 28.8 (13.1) 24.3 (11.4) 82.0 (18.7)" 18.5 (1.0) 44.5 (9.2) 42.2 (18.7) 22.5 (18.1) 29.0 (11.0) 16.5 (3.9)* 28.0 (4.8) 88.0 (9.0)* 15.0 (10.2) 21.5 (4.9) 36.8 (11.5) 19.3 (1.5) 23.8 (9.1) 20.0 (8.5) 30.8 (6.1) 53.7 (47.0)* 22.8 (7.8) 22.0 (8.5) 30.2 (8.4) 14.0 (4.5) 16.3 (2.5) 11.5 (12.0) 15.7 (2.3) 42.0 (21.5) 15.8 (10.3) 23.5 (2.1) 19.2 (4.8) 16.7 (0.6) 12.3 (5.0) 5.0 (0.0) 16.5 (0.71) *Statistically significantly different (P < 0.05). 140 r 120 t 140 ,- T I \ E 2 40 i -1 0 0 1 2 3 4 Level Below Pulpal Cornu. ( m m ) 0 0 I I I I 10 20 30 40 Distance From Pulp-Dentin Junction ( p m ) Fig. 2. The incidence of cell processes 5-10 pm from the pulp-dentin junction at different levels in the corona1 pulp. Filled circles represent control values, open circle counts after nerve section. Bar lines show standard deviations. Only a t the level 0 to 1 mm below the cornu are the differences statistically significantly different (P < 0.05). Fig. 3. The incidence of cell processes at different distances from the pulp-predentin border at the level of the pulpal cornu. Solid circlescontrol, open-operated. All but the first and last pairs of points are statistically significantly different (P < 0.05). pulp. The original magnification was 1 , 2 5 0 ~and with enlargement a t printing of 4.5 x gave a final total magnification of 5,625 X. Each montage was subdivided into four strips 20 pm wide and 40 pm long with the long axis a t right angles to the pulp-predentin border. Each strip was then subdivided every 5 pm along its length. This resulted in a series of eight zones at different distances from the predentin each 100 pm2 in area. This resulted in 268 zones, each localized in terms of distance from the predentin and position in the tooth. This ensured that in comparing observations on operated and control sides, comparisons were being made from similar locations. Previous studies have shown that both gap junctions and axons show regional variations in distribution (Holland, 1978, 1980a). The following observations and counts were made from each zone: The number and length of gap junctions linking one cell process to another. The number and length of gap junctions linking cell processes to cell bodies. The number and length of gap junctions linking one cell body to another. The number of cell processes present. The number of axons present. Structures that may have been degenerating nerve terminals and the relationship of these structures to other cells and processes. Cell bodies were defined a s profiles that contained nuclear profiles. Axons were identified by their relationship with Schwann cells. It was realized that a t the magnifications used, gap junctions could not successfully be differentiated from tight junctions, but our earlier studies have shown that the vast majority of these junctions are restricted to the cell to cell contacts adjacent to the predentin. Junctional lengths were measured using a digitizing pad linked to a small computer (Bioquant, R & M Biometrics, Nashville, TN). The observer was unaware at the time of measurement of the origin, NERVE SECTION AND ODONTOBLAST GAP JUNCTIONS 1407 tween operated and control counts were not significant in 21 of 24 cases. - Cell to Cell Junctions (Fig. 5) Junctions with nuclear profiles on each side were rare; only 16 were observed in total. The majority were in the region immediately adjacent to the predentin. Whereas more were present in the control tissue than the operated, differences were not statistically significant. 120 07 %ul 100- a, 0 0 & .c 0 L 461 80 - T a, n E 3 z Distance From Pulp-Dentin Junction ( p m ) Fig. 4. The incidence of cell processes at different distances from the pulp-predentin border at a level 3-4 mm below the pulpal carnu. There is no statistically significant difference between control (solid circle) and operated (open circle) values at any point. operated or control, or position of the sample under examination. In all, 2,900 junctions were counted and measured. In comparing incidence and lengths between operated and control sides, a simple nonparametric ranking test, the median test, was used. RESULTS No gross differences in appearance were obvious between tissue taken from various levels or between operated and control sides. The conditions of fixation resulted in a sizeable extracellular space and clear delineation of cell processes (Fig. 1).No plausible degenerating nerve terminals were found. Distribution of Processes (Table 1) At different levels in the crown (Fig. 2) In control teeth the largest concentration of processes was at the level of the pulpal cornu. The incidence declined in the more apical levels. In operated teeth there was no consistent pattern of change between different levels; cornual counts were not significantly higher than those at lower levels. At different distances from the pulp-predentin border (Figs. 3,4) In control teeth there was a decline in the number of processes from the pulp-predentin junction toward the central pulp. This was particularly marked a t the highest level near the cornu (Fig. 3). At lower levels (Fig. 41, the decline was much less marked. The pattern in operated teeth was similar, but at the cornual level the number of processes was always lower than in the control teeth and the difference was statistically significant (Table 1).At the more apical levels, the differences be- Process to Cell Junctions (Fig. 6) The number of junctions linking processes to clearly recognizable cell bodies in control tissue declined in the coronal to apical direction, although most of the decline occurred between the level of the cornu and the next level apically. There was no similar regional pattern in teeth from the operated side. At the level of the cornu, peripheral regions adjacent to the predentin showed higher incidences than did more central areas (Fig. 7), although this pattern was not evident a t lower levels (Fig. 8). The incidence of these junctions in the cornual level was always lower in operated than in control sides; however, a t only one point were these differences statistically significant (Fig. 7). There was no consistent pattern of difference between operated and control sides at lower levels (Fig. 8). When total lengths of cell process junctions were compared, a similar pattern was found. Process to Process Junctions (Table 2; Fig. 9) In unoperated teeth the incidence was highest in the level immediately adjacent to the cornu and declined apically, with most of the decline occurring within the first millimeter. The incidence also declined from peripheral to central pulp (Figs. 101, but again the fall was a step rather than a slope. After denervation the incidence of process to process junctions fell, but only in the most coronal level where in the region 10 to 35 pm from the pulp-predentin junction differences between counts on operated and control sides were statistically significant (Fig. 10). When the distribution of process to process junctions was expressed as their total length per 100 p area, the pattern was similar. Significant differences between operated and control sides were again limited to the juxtacoronal level. Axons No axons were recognizable in the first 5-pm strip adjacent to the predentin a t all levels. They first appeared 5-10 pm from the predentin a t the most coronal level. A general increase in number occurred toward the central pulp, with the highest counts in the most central and apical levels. Axons were present in operated teeth, but none were closer to the predentin than 20-25 pm. Counts in operated teeth were lower than in control, but low numbers precluded valid statistical comparison. DISCUSSION AND CONCLUSIONS The data from unoperated teeth described quantitatively some of the regional patterns within the odontoblast layer and peripheral pulp at different levels in the tooth. It is clear that the layer in the cornu was substantially different from the layer elsewhere. The incidence of cell processes is higher, as are the numbers of junc- 462 G.R. HOLLAND Fig. 5. Two odontoblast cell bodies (OD) linked by a gap junction (G) and also by a desmosomelike junction (D) (operated side). Fig. 6. A cell process (CP contacting a n odontoblast cell body (OD) by a gap junction (arrow) (control side). NERVE SECTION AND ODONTOBLAST GAP JUNCTIONS 463 14 12 10 8 4 6 T 4 T 2 0 101 20’ - - 30 0 0 40 J. Distance From Pulp-Dentin Junction (pm) Distance From Pulp-Dentin Junction ( p m ) Fig. 7.The incidence of gap junctions linking processes to nuclei containing cell bodies at the level of pulpal cornu, only one 15-20 ~m from the predentin are control (open circles) and operated (closed circles) statistically significantly different (P < 0.05). Fig. 8. The incidence of gap junctions linking cell processes to cell bodies containing nuclei 3 to 4 mm below the level of the pulpal cornu. There are no statistically significant differences between control (open circle) and operated (closed circles) sides. TABLE 2. No. of process to process junctions Distance from PDJ (pm) Level 1 Control operated 2 Control operated 3 Control operated 4 Control operated 0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40 11.0 (10.1) 5.5 (7.7) 15.0 (7.3)* 1.8 (3.0) 13.8 (14.8) 0.7 (1.6) 6.7 (8.1) 1.8 (1.7) 28.3 (13.4) 14.3 (11.4) 14.8 (7.0) 11.4 (10.1) 11.3 (8.8) 9.8 (5.5) 10.0 (11.0) 4.8 (1.5) 25.3 (2.5)* 4.5 (3.1) 17.0 (7.8) 12.0 (7.3) 9.5 (7.0) 12.8 (6.0) 6.8 (3.7) 5.8 (3.1) 27.0 (2.6)” 6.8 (2.2) 7.0 (5.0) 10.6 (8.2) 7.8 (4.9) 12.8 (3.9) 8.7 (7.0) 2.8 (2.5) 14.0 (O)* 5.3 (3.0) 9.3 (5.3) 12.2 (6.6) 2.0 (O)* 11.2 (8.7) 3.0 (1.2) 4.5 (3.4) 15.5 (3.2)* 6.8 (5.7) 4.0 (2.9) 11.8 (6.3) 2.5 (2.5) 8.0 (6.3) 6.3 (4.0) 9.0 (5.4) 15.3 (12.9)* 7.0 (5.5) 3.0 (1.4) 6.4 (5.2) 2.5 (1.9) 6.3 (5.2) 3.0 (4.2) 1.7 (1.5) 17.3 (21.6) 4.0 (3.2) 8.0 (2.9) 5.4 (3.4) 3.3 (4.2) 2.0 (2.3) 2.0 (2.8) 1.0 (1.4) *Statistically significantly different (P < 0.05). tions linking these processes to cell bodies or other processes. Although there were no morphologically recognizable axons in the first 5-pm strip adjacent to the predentin a t any level, they extend most nearly to the dentin at this level. We have previously demonstrated that the juxtacornual dentin is the most densely innervated (Holland, 1981; Holland et al, 1986) and it seems reasonable to suggest that a t least some of the additional processes counted are unsheathed axons. The effect of nerve section is also only measurable in the cornu. The reduction in the number of cell processes after denervation is statistically significant, indicating a loss of axons. Whereas not all apical axons have degenerated by 56 hours postsection (Holland and Robinson, 1984), a great many do, and as there is a virtually total loss of intradentinal axons at this time (Holland et al, 1986), it would appear that degeneration occurs not only centrifugally from the site of section but centripetally from the terminals. The reason for selecting the short survival period was twofold. First, to try and catch degenerating axonal figures. This failed. The second reason was to minimize the secondary effects of denervation that may result from changes in blood flow and loss of sensory input. We have no assurance, however, that any of these effects (e.g., increased synthetic activity of the odontoblasts) did not occur and that the morphological changes measured were not related. Whereas nerve section causes a clear effect on the incidence of cell processes, the effect on the distribution of junctions is less obvious. No statistically significant reduction in the number of process to cell junctions occurred, suggesting that the processes involved in these junctions do not belong to the proportion of the population that were lost (i.e., were not axons). The reduction in cell to cell junctions was significant a t the cornual level. One interpretation of this would be that these are G.R. HOLLAND 464 Fig. 9. Cell processes linked by gap junctions (arrows)(operated side). axon to axon junctions. Matthews has presented neurophysiological evidence of coupling (Matthews and Holland, 1976) of axons supplying the dental pulp. This current data would seem to support the direct gap junction as the anatomical substrate of this phenomenon. The cornual odontoblast layer is a special region possessing a n abundance of unsheathed axons, some of which may participate in low resistance gap junctions with other axons. No statistically defensible evidence for the innervation of the odontoblast was obtained. ACKNOWLEDGMENTS I would like to thank Dr. P.P. Robinson for preparing the animals, Mrs. Enid Pehowich for her skillful technical assistance, and Mrs. Holly Ridyard for expertly typing the manuscript. This study was supported by the Alberta Heritage Foundation for Medical Research (grant #EG-3098) and the Medical Research Council of Canada (grant #MA-9681). LITERATURE CITED I 0 Distance From Pulp-Dentin Junction ( p m ) Fig. 10. The distribution of junctions linking one cell process to another at the level of the pulpal cornu. Control values (open circles) and operated values (solid circles) are statistically significantly different between 10 and 35 Fm from the predentin. Aldskogius, H., and J. Arvidsson (1978) Nerve cell degeneration and death in the trigeminal ganglion of the adult rat following peripheral nerve transection. J. Neurocvlom. 7t229-250. Arwill, T. (1968) The ultrastructure of the'pulpo-dentinal border zone. In: Dentine and Pulp: Their structure and relations. N.B.B. Symons, ed. London, Livingstone, pp. 147-167. Avery, J.K., C.F. Cox, and R.E. Corpron (1974)The effects of combined nerve resection and cavity preparation and restoration on response dentine formation in rabbit incisors. Archs. Oral Biol., 19539-548. Byers, M.R. (1977)Fine structure of trigeminal receptors in rat molars. In: Pain in the Trigeminal Region. D.J. Anderson and B. Matthews, eds. Amsterdam, Elsevier, pp, 13-24. NERVE SECTION AND ODONTOBLAST GAP JUNCTIONS Byers, M.R. (1979)Large and small trigeminal receptors nerve endings and their associations with odontoblasts in rat molar dentin and pulp. In: Advances in Pain Research and Therapy, vol. 3. J.J. Bonica, J.C. Liebeskind, and D.G. Albe-Fessard, eds. New York, Raven Press pp. 265-270. Byers, M.R., and S.J. Kish (1976)Delineation of somatic nerve endings in rat teeth by radioautography of axon-transported protein. J. Dent. Res., 55:419-425. Frank, R.M. (1968a) Attachment sites between the odontoblast process and the intradental nerve fibre. Archs. Oral Biol., 13:833-834. Hayat, M.A. (1970) Principles and Techniques of Electron Microscopy, vol. I. New York, Van Nostrand Reinhold, pp. 242-243. Holland, G.R. (1977) Structural Relationships in the Odontoblast Layer. In: Pain in the Trigeminal Region. D.J. Anderson and B. Matthews, eds. Amsterdam, Elsevier, pp. 25-33. Holland, G.R. (1980a) Non-myelinated nerve fibres and their terminals in the sub-odontoblastic plexus of the feline dental pulp. J. Anat., 130:457-467. Holland, G.R. (1980b) Microtubule and microfilament populations of cell processes in the dental pulp. Anat. Rec., 198:421-426. Holland, G.R. (1981)The incidence of dentinal tubules containing more 465 than one process in the cuspal dentin of cat canine teeth. Anat. Rec., 200:437-442. Holland, G.R. (1986a) Odontoblasts and nerves: Just friends. Proc. Finn. Dent. Soc. 82179-189. Holland, G.R. (198613) Nerves in Dentine. In: Recent Progress in Oral Biology. S.W.J. Lisney and B. Matthews, eds. Bristol, Univ. Bristol Press, pp. 168-185. Holland, G.R., and P.P. Robinson (1984)Evidence for the persistence of axons at the apex of the cat’s lower canine tooth after section of the inferior alveolar nerve. Anat. Rec., 208:175-183. Holland, G.R., B. Matthews, and P.P. Robinson (1986)An electrophysiological and morphological study of the innervation and reinnervation of cat dentine. J. Physiol. 386:31-43. Magloire, H., H. Vinard, and A. Joffre (1979) Electrophysiological properties of human dental pulp cells. J. Biol Buccale., 7:251-262. Matthews, B., and G.R. Holland (1975) Coupling between nerves in teeth. Brain Res., 93:354-358. Pappas, G.D., and M.V.L. Bennet (1966)Specialized junctions involved in electrical transmission between neurons. Ann. N.Y. Acad. Sci., 137:495-508.