Reticulo- and trigemino-hypoglossal connections: A quantitative comparison of ultrastructural substrates

Brain Research

Bullerin,

0361-9230/88 $3.00 + .OO

Vol. 21,pp. 795-803.Pergamon Press plc, 1988.printed in the U.S.A.

Reticula- and Trigemino-Hypoglossal Connections: A Quantitative Comparison of Ultrastructural Substrates ROSEMARY

C. BORKE,

MARTIN

E. NAU AND DONALD

B. NEWMAN

USUHS, F. Hebert School of Medicine, Department of Anatomy 4301 Jones Bridge Road, Bethesda, MD 20814-4799 Received

19 November

1987

BORKE, R. C., M. E. NAU AND D. B. NEWMAN. Reticula- and trigemino-hypoglossal connections: A quantitative comparison oj’ultrastructural substrates. BRAIN RES BULL 21(5) 795-803, 1988.-Axon terminals were identified and characterized by electron microscopy after injections of horseradish peroxidase (HRP) into the spinal V nucleus (SPVN) or the medullary reticular formation adjacent to the XIIth nucleus. The synaptic organization and topology of these two different populations of hypoglossal afferents (T-XII and R-XII respectively) were determined by quantitative comparisons. Significant differences were obtained in the ratios of morphological types of terminals, sizes of axonal endings and their location on postsynaptic structures. Axon terminals containing spherical vesicles (S-terminals) and those with flattened/pleomorphic vesicles (F-terminals) were anterogradely labeled with HRP from both injection sites. However, the S/F ratio for R-XII terminals was 1.2:1compared to 2.6: 1 for T-XII afferents. Asymmetrical membrane densities (Gray Type I) were the predominant form ofjunctional specialization for S-terminal synapses. Asymmetrical densities with subjunctional dense bodies/bars (S-Taxi) were associated with a higher proportion of T-XII synapses than R-XII synapses. Almost all of the F-terminals from both sources had symmetrical densities (Gray Type II). The average diameter of R-XII terminals was greater than that ofT-XII terminals. R-XII-F terminals were the largest terminals. The majority of axon terminals from both sources formed axodendritic synapses. However, R-XII terminals had a higher incidence (10% vs. 3%) of axosomatic contacts. The proportion of R-XII-F-terminals decreased from the central toward the distal dendrites, whereas the opposite was found for T-XII-F and T-XII-S-terminals. In contrast to these findings, R-XII-S-terminals were more uniformly distributed on dendrites of all sizes. The variations in synaptology and spatial distribution may be morphological expressions of different functional properties associated with primary and secondary sources of hypoglossal afferents. Horseradish Anterograde tracing Axon terminals in the XIIth nucleus

peroxidase Medullary reticular formation Synaptology and spatial distribution

Spinal V nucleus

ALTHOUGH motoneurons in the hypoglossal nucleus provide the single source of efferent innervation to tongue mus-

havior involved in mastication, swallowing and oral grooming (50). The second source of hypoglossal input, the dor-

culature, the structural and functional aspects of neural pathways mediating tongue movements have not been completely elucidated. Physiological evidence suggests that in subprimates, spatially-separated sources, namely the periphery (23, 35, 39) and cortical (24, 34, 40) areas, influence the activity of hypoglossal motoneurons by multisynaptic connections. Brain stem intemeurons have been implicated by physiological data to be interposed between the corticofugal and peripheral afferents and hypoglossal efferents (25, 40, 45). Recent retrograde tracer experiments in rats have identified the anatomical loci of intramedullary neurons that project to the hypoglossal nucleus (1250). The medullary reticular formation adjacent to the XIIth nucleus constitutes the primary source of hypoglossal afferents. This region receives input from diverse levels of the neuraxis (30) and by its monosynaptic connections with the hypoglossal nucleus as well as all oral motor nuclei, functions to organize the synergies of complex oral-lingual be-

sal portion of the spinal V nucleus (SPVN), in part receives peripheral afferents from intraoral structures (21). Some of these neurons are associated with the transfer of information from primary trigeminal axons to cranial motoneurons including the XIIth nucleus for sensori-motor reflexes (2, 12, 50). The sensory cells of spinal V nucleus also project to the primary source of hypoglossal afferents, the medullary reticular formation (30,35). Recent work of the principal investigator delineated the ultrastructural features of reticula- (R-XII) and trigemino-hypoglossal (T-XII) axon terminals (911). The aim of the current study was to compare quantitatively the synaptic organization and spatial distribution of primary and secondary hypoglossal afferents to see whether presumed differences in functional properties have an anatomical substrate in synaptic arrangements. Horseradish peroxidase (HRP) served as an anterograde tracer, permitting ultrastructural comparison of the synaptology of reticula- and trigemino- hypoglossal connections.

795

796

BORKE. NAU AND NEWMAK METHOD

G&i

Preparations

Hypoglossal neurons were studied in sections of the rat medulla prepared by modification (37) of the Adams-Golgi technique (1). Brains of four anesthetized rats were fixed by transcardiac perfusion with physiological saline, followed by lf% phosphate-buffered formalin and a mordant solution of chloral hydrate (5%) and potassium dichromate (5%) in I@% buffered formalin. Blocks of tissue through the medulla were isolated, impregnated with silver nitrate (I%), dehydrated, embedded in Epon and cut in the transverse plane. Golgiimpregnated neurons in the hypoglossal nucleus were projected through a prism over the third ocular tube of a Biophot microscope. Silhouettes of the shafts of the first and second order dendrites from 46 neurons were traced. Their diameters were measured on a Microplan II digitizing tablet from drawings enlarged to a final magnification of x720. The range of measurements obtained from the Go&-impregnated material was used to group the dendritic profiles measured from electron micrographs into two size ranges: largeintermediate and small. Anterograde

Lubeling

Experinwnts

Anterograde labeling experiments were carried out in 16 male rats of the Sprague-Dawley strain. The present study utilized specimens obtained from previous studies (9-l 11 describing the qualitative aspects of ultrastructural labeling of reticula- and trigemino-hypoglossal connections. A summary of the labeling experiments is included in this communication. The rats were anesthetized with sodium pentobarbital (30 mg/kg) or % chloral hydrate (5 ml/kg). Single injections of OS-2@% HRP (40-100 nl) were made through glass micropipettes, stereotaxically guided into SPVN, pars interpolaris (n=8) or into the medullary reticular formation (n=8) adjacent to the XII nucleus [i.e., nucleus reticularis parvocellularis (RPc) and the medial aspect of nucleus reticularis gigantocellularis (RGc) as detailed previously (9)]. The possibility that some labeled axon terminals in the hypoglossal nucleus after reticular formation injections resulted from uptake by damaged SPVN axons coursing through this region was considered by monitoring the incidence of retrograde labeling of SPVN neurons. In addition, adjacent vibratome sections were stained with silver degeneration techniques (16) to confirm that axons had not been damaged by the pressure of the injection. Control

Experiments

The control series was comprised of 2 rats per injection site. All procedures were identical to those followed for the experimental series except that a small volume (40 nl) of tracer vehicle (0.1 M Tris buffer) without HRP was injected into one of the two aforementioned sites. Electron

Microscop)

After a survival period of 18-50 hr, control and experimental animals were reanesthetized and killed by transcardisc perfusion of Ringer’s solution, aldehydes (41) and a final solution of 0.1 M phosphate buffer. Transverse sections of the medulla were cut at 50 pm on a Vibratome and reacted histochemically for HRP according to the cobalt-glucose oxidase method (20). Alternate sections were mounted and counterstained with neutral red to evaluate the position and

extent of the injection site. Remaining reacted sections were processed for electron microscopy according to previously detailed procedures (9 I 1). Ultrathin sections of the hypoglossal nucleus were left unstained for ultrastructural evaluation. @rutltitrrtii’e

Mcthotl.~

Randomly selected electron micrographs containing labeled axon terminals in the hypoglossal nucleus were utilized for quantitative evaluation. Electron micrographs (x22,500-47,500) were analyzed on a Microplan II digitizing tablet interfaced to an IBM-PC. The evaluated population of labeled terminals consisted of I) 563 labeled R-XII terminals from 8 rats and 2) 576 labeled T-XII terminals from 8 rats. Previous ultrastructural studies (6, 8, 46) classified axon terminals in the XlIth nucleus according to synaptic vesicle shape: those containing spherical vesicles (S-terminals) and those containing flattened and/or pleomorphic vesicles (Fterminals). Based on these findings, counts were tabulated of the I) number of labeled S-terminals and 2) number of labeled F-terminals. Labeled S- and F-terminals forming synapses (indicated by the presence of an active zone and a junctional membrane specialization in the plane of section) were tabulated and characterized according to the type of junctional specialization and the type of synapse (i.e., axosomatic, axodendritic or axoaxonic). Finally, the diameters of the labeled presynaptic terminals and postsynaptic dendrites of synaptic pairs were measured. Values reported in the text include the means and standard errors of the means. The relations between a number of different parameters were investigated using SYSTAT software. The statistical significance observed between the mean values for 1) diameters of S- and F-terminals from the two sources and 2) diameters of dendrites associated with labeled R-XII and T-XII terminals were analyzed by Student’s t-tests. Comparisons between the frequency distribution of S- and F-terminals for the two sources on dendrites of different sizes were made using chi-square (4x2) tests. RESULTS

Somata of hypoglossal neurons emit 3-6 primary dendrites, (meanz4.27 pm 20.77) which range in diameter from 2.52 pm to 9.24 /*rn (mean=4.42 wrn +-1.23). These first order dendrites divide into secondary dendrites, ranging in diameter from I .45 pm to 4. I 1 pm (mean=2.32 Frn ~0.61). The second order dendrites taper into small, distal dendrites with diameters of 90%). A greater proportion of axosomatic synapses were formed by R-XII terminals (1%) than T-XII endings (3%). Unequivocal axoaxonic synapses were not identified.

FIG. 4. Postsynaptic dendrite diameters. Frequency distribution of reticula-hypoglossal (R-XII) and trigemino-hypoglossal (T-XII) axon terminals with spherical (S) and flattened (F) vesicles synapsing on dendrites of increasing diameters (2.50 pm). Significant differences C~~0.05) in the distribution of R-XII(S) vs. T-XII(S) and R-XII(F) vs. T-XII(F) were obtained using chi-square 4x2 con-

tingency tests.

Spmtirrl

distribution

of’

Dixon

terminuls

on

dendrites.

Cross-sectional diameters of dendrites measured in electron micrographs formed the basis for determining the spatial distribution of axon terminals from the two sources of hypoglossal input. Only dendrites that formed synapses with labeled axon terminals were measured. The criteria used for identifying synapses included the presence of an active site and a junctional membrane specialization. Measurements were compared to the diameters of the first and second order dendrites measured in Golgi impregnations. On this basis, dendrites were arranged in two groups: large-intermediate (>1.45 pm) and small (cl.45 pm). The overlap in the size range of primary and secondary dendrites in the Golgi material and the infrequent sampling in electron micrographs of the most proximal part of the dendritic tree in the transverse plane prevented a more precise correlation. Thus, the present data given only an approximate distribution of the two sources of afferents on the dendritic tree, but certain differences in spatial arrangement are revealed. The diameters of dendrites ranged from 0.33-6.02 pm. Paralleling the findings for presynaptic structures, the mean diameter for dendrites (mean=2.10 pm kO.82) contacted by R-XII terminals was greater than the average diameter for dendrites (mean=1.79 pm kO.64) postsynaptic to T-XII terminals. Statistically significant differences in the average dendritic diameters of R-XII-F synapses (mean=2.17 wrn t0.78) and T-XII-F synapses (mean= 1.75 pm kO.60) were established by Student’s t-test (p