Ultrastructural and cytochemical study of neurones in the rat dorsal motor nucleus of the vagus after axon crush

Ultrastructural and Cytochemical Study of Neurones in the RatDorsal Motor Nucleus of the Vagus After Axon CrushV. NAVARATNAM, T.S. JACQUES, AND J.N. SKEPPER*Multi-Imaging Centre, Department of Anatomy, University of Cambridge, Tennis Court Road, Cambridge, UK

KEY WORDS vagus; crush injury; dorsal motor nucleus of the vagus; retrograde transport;horseradish peroxidase; cholinesterase; ultrastructure

ABSTRACT The cell populations in the dorsal motor nucleus of the vagus (DMNV) of the ratwere studied by light microscopy and transmission electron microscopy, including retrogradelabeling with horseradish peroxidase and histochemical demonstration of the distribution of theactivity of the enzymes acetylcholinesterase (AcChE) and butyrylcholinesterase (BuChE). Two typesof neurones were observed: 1) Larger Type A cells, which stain for both AcChE and BuChE and whichproject into the vagus nerve trunk, and 2) smaller Type B cells, which stain lightly for AcChE but notfor BuChE and which do not project into the vagus nerve.

Standardised vagal crush at the mid-cervical level causes loss of cholinesterase activity in Type Aneurones within a few days but has no effect on Type B neurones. Changes in nuclear morphology ofType A neurones are pronounced at 10 weeks postinjury, indicating that degeneration is irreversibleeven by this stage.

The number of Type A cells projecting to the vagus nerve reduces as a function of time, presumablyas these cells die. Only a small number of Type A neurones persist at 2 years postinjury. Microsc. Res.Tech. 42:334–344, 1998. r 1998 Wiley-Liss, Inc.

INTRODUCTIONStudies on the rat have demonstrated that neurones

in the dorsal motor nucleus of the vagus (DMNV)respond to axon injury in a more severe and prolongedmanner, including loss of cholinesterase activity, thando most other peripherally projecting neurones (Lewiset al., 1972; Lams et al., 1988), and it was suggestedthat, in the rat at least, these neurones have little or nocapacity for survival after axotomy, which is unusualfor peripheral neurones. However, care should be takenwith such an interpretation because in several speciesit has been shown that the DMNV contains not onlyneurones which project preganglionic parasympatheticnerve fibres into the vagal trunk, but also a significantpopulation which are not peripherally connected; pre-sumably, these are either intrinsic to the nucleus orthey project to other regions of the brain. (McLean andHopkins, 1981, 1982; Aldskogius et al., 1978; Ling et al.,1986, 1987). To resolve the issue of capacity for survival, itis important to distinguish the subsets of neuronesbefore characterising their responses to axon injury.

Since no single technique is able to identify allcomponents of each subpopulation, a combination of thefollowing strategies was used in the present study tocharacterise peripherally projecting DMNV neurones:1) retrograde labeling through the vagal trunk, usinghorseradish peroxidase (HRP); 2) cell size and morphol-ogy; the literature indicates that larger sized (Type A)cells are more likely to project into the vagus thansmaller (Type B) cells (see Ling et al., 1986, 1987); 3)cholinesterase cytochemistry for cholinergic neurones;and 4) immediate response to vagal trunk injury (inappropriate short-term experiments). The ultrastruc-tural changes induced by vagal crush in peripherally

projecting neurones, up to a period of 24 months, werestudied using transmission electron microscopy (TEM),including retrograde HRP labeling and histochemistryfor acetylcholinesterase (AcChE) and butyrylcholines-terase (BuChE). It is especially valuable to know whennuclear changes occur in these cells, since these wouldsuggest progression to an irreversible stage (for re-views, see Clarke, 1990; Trump and Berezesky, 1992;Martin, 1993). The patterns of change were comparedwith the reactions of nonprojecting cells. In addition,the number of surviving HRP-labeled neurones at eachstage was measured by stereological analysis.

MATERIALS AND METHODSOnset Controls

Twenty-one female Wistar rats 6 weeks of age wereused in this study. Five were used as onset controls; twowere anaesthetised with ketamine and rompun andperfused in preparation for the light and electronmicroscopical demonstration of cholinesterase activity;three others were used for retrograde HRP transportstudies. The latter were anaesthetised and the rightcervical vagus was exposed surgically. Using an ex-truded pipette connected to a microsyringe, 1.5 µl of20% horseradish peroxidase (Sigma type VI, diluted innormal saline) was injected into the vagal trunk, mid-way between the base of the skull and the recurrentlaryngeal branch, in a rostral direction. After 3 daysthey were terminally anaesthetised and processed forlight and electron microscopy (see below).

*Correspondence to: Dr. Jeremy N Skepper, Multi-Imaging Centre, Universityof Cambridge, Department of Anatomy, Tennis Court Road, Cambridge, CB2 3DY,UK. E-mail: [email protected]

Received 4 November 1997; accepted in revised form 18 December 1997

MICROSCOPY RESEARCH AND TECHNIQUE 42:334–344 (1998)

r 1998 WILEY-LISS, INC.

Vagal CrushThe remaining 16 animals were similarly anaesthe-

tised and the exposed right vagus was subjected to a

standardised crush injury at a point midway betweenthe base of the skull and the origin of the recurrentlaryngeal nerve. A pair of smooth-jawed haemostats

Fig. 1. Vibratome section (100 µm) of hindbrain of a rat taken atthe level of the DMNV 3 days after administration of HRP to the rightcervical vagus. Cell bodies in the right DMNV stain strongly for HRP.Scale bar 5 500 µm.

Fig. 2. Adjacent vibratome section to Figure 1 after incubation todemonstrate AcChE activity. The intense staining of the neuropilmasks staining of individual cell bodies in the DMNV. Scale bar 5 500µm.

Fig. 3. Adjacent vibratome section to Figure 2 after; incubationBuChE activity. Cell bodies and processes of both the right and leftDMNV are strongly stained. Scale bar 5 500 µm.

Fig. 4. Vibratome section (100 µm) of hindbrain of a rat taken atthe level of the DMNV 3 days after a crush injury to the cervical vagus.The section has been incubated to demonstrate BuChE activity.Enzyme activity is strong in the cell bodies and processes of the leftDMNV, but noticeably attenuated in those of the right DMNV. Scalebar 5 500 µm.

Fig. 5. Vibratome section (100 µm) of hind brain of a rat taken atthe level of the DMNV 10 weeks after a crush injury to the cervicalvagus. The section has been incubated to demonstrate BuChE activity.Enzyme activity is strong in the cell bodies and processes of the leftDMNV but absent from those of the right DMNV. Scale bar 5 500 µm.

Fig. 6. Vibratome section (100 µm) at the level of the DMNV 10weeks after a crush injury to the right cervical vagus. HRP wasinjected into the right cervical vagus 3 days before termination. Cellbodies and processes of the right DMNV are stained. Scale bar 5 500µm.

335VAGAL CRUSH AND NEURONE SURVIVAL

were closed tightly about the vagal trunk for 10 secondsand released. A loose, black silk ligature was tiedaround the vagal trunk distal to the crush injury andwas sutured to the adjacent sterno-mastoid muscle.This served to identify the site of injury in the long-term experiments.

Two rats were terminated at 3 days after the injuryand two others were terminated 1 week after the injury;their brainstems were removed and the DMNV exam-ined by light and electron microscopy after cholinester-ase cytochemistry. Six of the remaining rats werechosen by lottery at 10 weeks postinjury, injected withHRP just proximal to the site of the crush injury andleft for 3 days, terminally anaesthetised and processedfor light and electron microscopy. Six rats were left for24 months before processing; two died under anaesthe-sia and the remaining four of the long-term lesionedanimals were injected with HRP and processed asabove.

Fixation and SamplingThe animals were terminally anaesthetised with

ketamine and rompun, a midline thoracotomy wasperformed, and a 19-gauge needle was clamped into theleft ventricle. An incision was made in the right auricleand the animals were exsanguinated by perfusion withbuffered saline containing 10 mmol/l PIPES buffer, pH7.4, 139 mmol/l sodium chloride, 2.7 mmol/l potassiumchloride, 2 mmol/l calcium chloride, 5 mmol/l sodiumnitrite, 19.4 mmol/l glucose, and 2.5% polyvinylpyrrol-idone (FW 40,000) at room temperature. This wasfollowed by perfusion with cold (4°C) fixative for 15minutes containing 3% glutaraldehyde, 0.5% formalde-hyde, 2 mmol/l calcium chloride, and 2.5% polyvinylpyr-rolidone (FW 40,000) in 100 mmol/l PIPES buffer, pH6.0. Fixation was terminated by perfusion with 500 mlof cold (4°C) buffer containing 100 mmol/l PIPES buffer,2 mmol/l calcium chloride, and 2.5% polyvinylpyrrol-idone (FW 40,000).

The brains were removed and the hindbrain sepa-rated. Starting from a random position caudal to theDMNV, the hindbrain was sectioned coronally using avibratome set at a nominal thickness of 100 µm. Precisecalibration revealed this to produce a mean sectionthickness of 74 µm. Sections were collected sequentiallyin five regularly recurring sites for ultrastructure andlight and electron microscopical demonstration of retro-gradely transported HRP and endogenous AcChE andBuChE.

The sections selected for ultrastructural study byTEM were trimmed to an area containing both rightand left DMNV and returned to primary fixative at 4°Cfor 4 hours. They were rinsed in PIPES buffer, treatedwith 2% osmium ferricyanide containing 2 mmol/lcalcium chloride for 2 hours at 4°C, rinsed in water, andbulk stained with uranyl acetate. They were subse-quently dehydrated in an ascending series of ethanolsolutions and embedded in Araldite. To ensure that thesections remained flat for en face sectioning, they wereplaced in a drop of Araldite onto Araldite castings thathad been trimmed in an ultramicrotome to produce aflat surface. A carbon-coated coverslip (10 mm diam-eter) was placed carbon side down on top of the sectionsto flatten them. The sections were cured at 60°C for 48hours; if the surrounding drop of Araldite was minimal,the carbon-coated coverslip separates easily with alittle mechanical pressure to allow en face sectioningwith minimal loss of tissue. Otherwise, a brief immer-sion of the coverslip in liquid nitrogen caused it toseparate from the casting. Thin sections (40–50 nm) inthe original coronal plane were cut using a diamondknife on a Reichert Ultracut E. They were double-stained with uranyl acetate and lead citrate and viewedat 80 kv in a Philips EM 300.

Demonstration of Horseradish PeroxidaseActivity

Sections were rinsed twice in 0.1 M PIPES buffer, pH6.0, then incubated in 1 ml PIPES buffer (pH 6.0), 50 µl1% ammonium paratungstate, and 50 µl 0.1% ethanolictetramethyl benzidine for 30 minutes. 10 µl of 0.3%hydrogen peroxide was added and the sections wereagitated for 25 minutes. They were rinsed four timeswith 0.1 M PIPES buffer (pH 6.0), mounted on slides,stained with cresyl violet, and coverslipped for lightmicroscopy. Sections to be examined by TEM wereincubated in 1 ml 0.1 M PIPES buffer, pH 6.0, 20 µl 1%cobaltous chloride, 20 µl 2.5% aqueous diaminobenzi-dine tetrahydrochloride, and 20 µl 0.3% hydrogen perox-ide for 10–15 minutes. They were rinsed four timeswith 0.1 M PIPES buffer (pH 6.0) and treated as per theultrastructural studies above. This method is essen-tially that of Weinberg and Van Eyck (1991) with thesubstitution of phosphate buffer by PIPES buffer andthe use of a brief period of fixation at pH 6.0. Controlsfor endogenous horseradish peroxidase activity werecarried out using the same incubation media on slices oftissue from rats that had not been injected with tracerenzyme.

Fig. 7. Vibratome section (100 µm) at the level of the DMNV 24months after a crush injury to the right cervical vagus. HRP wasinjected into the right cervical vagus 3 days before termination. A fewcell bodies and processes of the right DMNV are stained. Scale bar 5500 µm.

336 V. NAVARATNAM ET AL.

StereologyThe mean volume of the DMNV was estimated using

the Cavalieri principle (Michel and Cruz-Orive, 1988).The area of the DMNV was estimated by planimetry oneach of the slices mounted for light microscopy using aKontron Videoplan. The areas were summed and multi-plied by the mean distance between each section to givethe mean volume of each DMNV. The numerical densityof cells demonstrating HRP activity (estimated usingthe optical dissector) multiplied by the mean volume ofthe DMNV gave an estimate of the total number ofneurones demonstrating HRP activity.

Demonstration of Cholinesterase ActivitySubstrate solutions for the demonstration of AcChE

and BuChE activity were prepared as follows: 200 mgacetylthiocholine iodide or 220 mg butyrylthiocholineiodide was placed into 30 ml centrifuge tubes anddissolved in 8 ml of water. They were precipitated with14 ml of 0.1 M cupric sulphate (5H2O), added drop bydrop over a whirlimixer. This was allowed to stand for10 minutes and spun down hard for 10–20 minutes. Thesupernatant was aspirated and 124 mg of glycine wasadded to every 20 ml to stabilise it.

The sections were washed in succinate buffer, pH 5.5,for 10 minutes with four changes. They were thenincubated for 3–4 hours in complete media made up asfollows: 8 ml of substrate solution, 8 ml of isotonicsodium sulphate, 2 ml of water, and 1 ml of succinic

acid, pH adjusted to 5.5 with 0.1 M sodium hydroxide.They were rinsed for 1 hour in 2–4 changes of bufferwash. The reaction product was fastened by treatmentfor 1 hour in two changes of sulphide buffer. They wererinsed in buffer, stained with cresyl violet, and mountedfor light microscopy or embedded for TEM as describedpreviously (Lewis and Knight, 1992). Attempts weremade to double-label for both cholinesterase activityand horseradish peroxidase. These were unsuccessful.If reaction for cholinesterase followed that for horserad-ish peroxidase, the HRP final reaction product dis-solved. Similarly, if reaction for HRP followed that forcholinesterase, the cholinesterase final reaction prod-uct was dissolved.

RESULTSLight Microscopy

HRP administration at the time of vagal injuryproduces numerous labeled neurones throughout theDMNV (Fig. 1) and this staining is confined to theneuronal somata and processes. AcChE activity inonset controls is present not only in the soma anddendrites of DMNV neurones, but also extensivelythroughout the surrounding neuropil (Fig. 2). On theother hand, BuChE activity, like HRP labeling, islargely confined to the soma and processes of theDMNV neurones (Fig. 3), a property that has made itan attractive marker for these cells in past studies.

Fig. 8. Thin section (50 nm) through the right DMNV of a control rat showing a typical Type Aneurone. The nucleus contains a prominent nucleolus and the endoplasmic reticulum is present in thecytoplasm as lamellae. Scale bar 5 5 µm.


Fig. 9. Thin section (50 nm) through the right DMNV of a control rat showing a typical Type B neurone. Thenucleus contains a prominent nucleolus and is lobate. Endoplasmic reticulum is sparse. Scale bar 5 2 µm.

Fig. 10. Thin section (50 nm) through a Type A neurone in the rightDMNV of a rat, 1 week after a crush injury to the right cervical vagus.The endoplasmic reticulum is dispersed and lipofuscin bodies areprolific. Scale bar 5 1 µm.

Fig. 11. Thin section (50 nm) through a Type A neurone in the rightDMNV of a control rat 3 days after injection of HRP into the rightcervical vagus. HRP reaction product is in lysosomes and present inthe cytoplasm as long crystals. Scale bar 5 2 µm.

Within three days of a crush injury to the right vagaltrunk, all cholinesterase activity is reduced in thecorresponding DMNV (Fig. 4) and disappears within 7days of such an injury. At 10 weeks postinjury, BuChEactivity is still absent from the right DMNV (Fig. 5),whereas HRP-labeled neurones are still recognisable,although much reduced in number (Fig. 6). Twenty-fourmonths after injury, AcChE and BuChE are still largelyabsent from the right DMNV, though occasional neu-rones at various levels of the nucleus show someactivity (see below). HRP labeling was present in two ofthese animals but absent in the other two (Fig. 7).

Electron MicroscopyUltrastructural examination confirms the presence

of two contrasting neuronal populations in the DMNV:Type A of modest size that appear mostly bipolar inprofile within thin plastic sections. They possess anovoid nucleus with a prominent nucleolus and theircytoplasm contains characteristic stacks of rough endo-plasmic reticulum or nissl substance (Fig. 8). Thesecond type (Type B) are smaller in profile with a smallratio of cytoplasm to nucleus, little rough endoplasmicreticulum, and a highly lobate nucleus (Fig. 9). Within 3days of injury to the vagus, the rough endoplasmicreticulum of the Type A neurones begins to lose itstypical lamellae. At 1 week after injury, the roughendoplasmic reticulum is considerably disorganisedand dispersed and lipofucsin bodies become prominent

in many cell profiles (Fig. 10). Type B neurones appearunchanged. This pattern persisted in neurones of theDMNV examined at 10 weeks and 24 months postin-jury.

Retrograde Transport of HRPAt all time points studied, only Type A neurones

transported HRP from the vagal trunk. HRP stainingwas strong in Type A neurones in onset controls with amean of 2,032 6 117 (n 5 3) cells labeling (Fig. 11).Labeling was also strong at 10 weeks postinjury butwith fewer cells demonstrating transport, 763 6 93(n 5 6), (Fig. 12). At 24 months after vagal crush HRP,labeling (administered 3 days before sampling) was lessintense. The staining pattern was weaker and thecrystals typical of the substrate tetramethylbenzidinewere much more dispersed (Fig. 13). Transport of HRPwas not demonstrable in two of the surviving rats andonly 250 and 157 cells were labeled in the remainingtwo rats. However, at all time points, less than 0.05% ofcell profiles with the morphological characteristics ofType A neurones remained unlabeled by HRP. Thisindicates that at each time point surviving Type Aneurones projected axons into the vagus nerve trunk.

Ultrastructural Distribution of CholinesteraseActivity

Reaction product for BuChE was prominent in thecytoplasm of the Type A neurones and largely absent in

Fig. 12. Thin section (50 nm) through a Type A neurone in the rightDMNV of a rat 10 weeks after a crush injury to the right cervicalvagus. HRP was injected into the right cervical vagus 3 days beforetermination. HRP reaction product is in lysosomes and is also presentin the cytoplasm as long crystals. Scale bar 5 2 µm.

Fig. 13. Thin section (50 nm) through a Type A neurone in the rightDMNV of a rat 24 months after a crush injury to the right cervicalvagus. HRP was injected into the right cervical vagus 3 days beforetermination. Moderate staining is present in lysosomes and diffuselythrough the cytoplasm. Scale bar 5 2 µm.


the Type B neurones (Fig. 14) of the onset control rats.The reaction product tended to appear as lumps orcrystals, which obscured the underlying structures(Fig. 14). AcChE reaction product was much finer intexture and gave much better structural resolution. Itfilled the cisterns of rough endoplasmic reticulum of theType A neurones, the nuclear membrane, the Golgiapparatus, and appeared at the surface of the neuro-lemma (Fig. 15). Type B neurones also stain for AcChE(Fig. 16), but much less intensely, with reaction productappearing in the nuclear membrane, in short cisterns ofendoplasmic reticulum, and occasionally on the neuro-lemma. Type A neurones at time zero had many nerveterminals abutting the perikarya. These frequentlydemonstrated reaction product for AcChE in the synap-tic gutter (Fig. 17). As early as day 3 after injury, nerveterminals on the perikarya were less evident and theendoplasmic reticulum was becoming disorganised.AcChE activity fell and was less intense at the nuclearmembrane and neurolemma. Within the cytoplasm itwas restricted to those cisterns of endoplasmic reticu-lum that were still intact. (Fig. 18).

At 10 weeks postinjury, Type B neurones were unaf-fected. Type A neurones exhibited residual cholinester-

ase activity only within short lengths of persistentendoplasmic reticulum (Fig. 19); many lipofuscin drop-lets were present in these cells. We also observedclumping or condensation of nuclear chromatin andcytoplasm in several Type A neurones with reducedAcChE. At 24 months postinjury, no return of BuChEactivity was seen. However, some Type A neurones inthe right DMNV had regained a reasonably normaldistribution of AcChE activity within their endoplasmicreticulum (Fig. 20), which was interspersed with lipofus-cin droplets. There were a small number of nerveterminals abutting the soma of such cells, but it mustbe emphasised that they were a minority of the persist-ing Type A neurones. The majority of Type A neuroneswere similar to those found at 10 weeks, demonstratingeither absent or dispersed endoplasmic reticulum.

DISCUSSIONOur observations confirm that the DMNV in the rat

contains two subpopulations of neurones, similar toprevious reports on the rat by Aldskogius et al. (1978),in the cat by McClean and Hopkins (1981, 1982), and inthe monkey and hamster by Ling et al. (1986, 1987): 1)Type A neurones, which are larger, stain for AcChE and

Fig. 14. Thin section (50 nm) through the right DMNV of a control rat after incubation to demonstrateBuChe activity. Reaction product is present in the form of large granules in the cytoplasm of Type Aneurones (A) but absent from Type B neurones (B). Scale bar 5 2 µm.


Fig. 16. Type B neurones from the same section as Figure 15. These contain much smaller butsignificant amounts of AcChE reaction product, which is largely confined to the endoplasmic reticulum,neurolemma, and nuclear membrane. Scale bar 5 2 µm.

Fig. 15. Adjacent section to Figure 14 incubated to demonstrate AcChE activity. Type A neuronescontain much reaction product, particularly in their endoplasmic reticulum. Scale bar 5 2 µm.

Fig. 18. Thin section of a vibratome section from a rat that has received a crush injury to the cervicalvagus 3 days previously. The section has been incubated to demonstrate AcChE activity. Enzyme activityis reduced both at the neurolemma and in the synaptic gutters. Scale bar 5 2 µm.

Fig. 17. Type A neurone from the same section as Figure 15. Nerve terminals (t) frequently abut thesoma of the cells. Intense AcChE reaction product is found at the synaptic gutter of these terminals. Scalebar 5 2 µm.


BuChE and project into the vagal trunk; 2) Type Bneurones, which are smaller, stain lightly for AcChEbut not for BuChE and do not project into the vagusnerve. Type A neurones respond rapidly to cervicalcrush by losing cholinesterase activity within a fewdays. Only a very small proportion of these cells regainenzyme activity in the long term. In the medium andlong term after axon crush, the majority of Type Aneurones gradually die out, so that after 2 years veryfew cells survive. The results of the stereological analy-sis of Type A neurones labeled with HRP showed some2,000 cells labeled in the control group. At 10 weeks thisfigure was reduced to less than 40% of control levels. At24 months, cells in only two out of four survivinganimals could be labeled and these were approximately10% of control values. On the other hand, Type B cellsare apparently unaffected by vagal crush.

The vast majority of Type A neurones were retro-gradely labeled with HRP and they displayed a reactionto axon crush within 3 days. It follows that they are asubstantial source of efferent fibres to the vagus nerve.The small minority of Type A profiles not labeled withHRP may indicate less than 100% uptake of the in-jected tracer enzyme.

None of the Type B neurones were retrogradelylabeled with HRP, nor did they show any response tovagal crush. It seems certain that these neurones do notproject down the vagus to the crush or labeling site atthe mid-cervical level. They are either local interneu-rones or project to other brain areas; McLean and

Hopkins (1982) were able to establish projections toother parts of the brain for about 10% of these smallerneurones in the cat.

All neurones in the uninjured rat DMNV expressAcChE activity within the endoplasmic reticulum oftheir somata, but only a proportion of cells, restricted tothe Type A variety, contain BuChE as well. Thus,BuChE seems to be a more reliable marker of pregangli-onic vagal neurones in the rat than AcChE, although itsfunctional significance is unclear. AcChE is also foundon synaptic membranes abutting DMNV neurones,thus accounting for the neuropil staining under thelight microscope. Vagal crush causes profound reduc-tion of AcChE and BuChE activity in Type A neurones ofthe DMNV, it is rapid in onset, and long-lasting. Within3 days there is depletion of cholinesterase activity, bothAcChE and BuChE, throughout Type A neurones, andby 10 weeks the reduction is almost complete. Type Bcells remain normal both in terms of enzyme activityand morphology.

By 10 weeks, cholinesterase activity is virtuallyabsent in Type A neurones and unaffected in Type Bneurones. However, a fair proportion of Type A cells canstill be demonstrated to be alive by retrograde labelingwith HRP, although there are only a third of thenumber of cells labeled in comparison with onset con-trols (i.e., 3 days after labeling in the absence of a crushinjury); however, even those Type A cells present ex-hibit an abnormal appearance under the electron micro-scope, suggesting they might die in the long term. Onthe other hand, Type B cells remain unchanged. Therecovery of AcChE activity in a small number of Type Aneurones is in keeping with the observations of Lams etal. (1988). It is, however, impossible to discount thepossibility that this may simply be due to the crushinjury affecting less than 100% of the vagal axons insome subjects.

In long-term experiments (24 months), HRP labelingdemonstrates very few surviving neurones. Type A cellsare difficult to find under the electron microscope buthealthy-looking Type B cells, all of which exhibit charac-teristic light AcChE activity, are found without diffi-culty. BuChE activity has disappeared completely.

As pointed out earlier, such a failure to recover fromaxon injury by rat vagal preganglionic neurones isunusual among peripheral neurones. However, sincemost peripheral ganglia associated with the vagusreceive preganglionic inputs from both sides, it ispossible that functional loss for organs such as theheart and gut may not be extensive after unilateralvagal lesions.

Indeed, the capacity for sprouting by the pregangli-onic fibres in the contralateral vagus may compensatefor cell and axon loss on the lesioned side. We hope topursue this aspect in future studies by examiningchanges in synaptic density in ganglia after vagalcrush.

ACKNOWLEDGMENTSThe authors are grateful to Ms. J.M. Powell for

technical assistance.

Fig. 19. Thin section of a vibratome section from a rat that hasreceived a crush injury to the cervical vagus 10 weeks previously. Thesection has been incubated to demonstrate AcChE activity. Enzymeactivity is reduced in this Type A Neurone (A). Scale bar 5 2 µm.


REFERENCESAldskogius, H., Barron, K.D., and Regal, R. (1978) Axon reaction in

dorsal motor vagal and hypoglossal neurons of the adult rat. Lightmicroscopy and RNA-cytochemistry. J. Comp. Neurol., 193:165–167.

Clarke, P.G.H. (1990) Developmental cell death: Morphological diver-sity and multiple mechanisms. Anat. Embryol., 181:195–213.

Lams, B.E., Isacson, O., and Sofroniew, M.V. (1988) Loss of transmitter-associated enzyme staining following axotomy does not indicatedeath of brainstem cholinergic neurons. Brain Res., 475:401–406.

Lewis, P.R., and Knight, D.P. (1992) Cytochemical staining methodsfor electron microscopy. In: Practical Methods in Electron Micros-copy, Vol. 14. A.M. Glauert, ed. Elsevier, New York, pp 217–228.

Lewis, P.R., Blundell Jones, P., Breathnack, S.M., and Navaratnam, V.(1972) Regenerative capacity of visceral preganglionic neurones.Nat. New Biol., 236:181–182.

Ling, E.A., Wong, W.C., Yick, T.Y., and Leong, S.K. (1986) Ultrastruc-tural changes in the dorsal motor nucleus of the monkey followingbilateral cervical vagotomy. J. Neurocytol., 15:1–15.

Ling, E.A., Shieh, J.Y., Wen, C.Y., Yick, T.Y., and Wong, W.C. (1987)The dorsal motor nucleus of the vagus nerve of the hamster:

Ultrastructure of vagal neurons and their responses to vagotomy. J.Anat., 152:161–172.

Martin, S.J. (1993) Apoptosis: Suicide, execution or murder? TICS,3:141–144.

McLean, J.H., and Hopkins, D.A. (1981) A light and electron micro-scopic study of the dorsal motor nucleus of the vagus nerve in thecat. J. Comp. Neurol., 195:157–175.

McLean, J.H., and Hopkins, D.A. (1982) Ultrstructural identificationof labelled neurons in the dorsal motor nucleus of the vagal nervefollowing injection of horseradish peroxidase into the vagus nerveand brainstem. J. Comp. Neurol., 206:243–252.

Michel, R.P., and Cruz-Orive, L.M. (1988) Application of the Cavalieriprinciple and vertical sections method to lung: Estimation of volumeand pleural surface area. J. Microsc., 150:117–136.

Trump, B.F., and Berezesky, I.K. (1992) The role of cytosolic Ca1in cell injury, necrosis and apoptosis. Curr. Opin. Cell Biol., 4:227–232.

Weinberg, R.J., and Van Eyck, S.L. (1991) A tetramethylbenzidine/tungstate reaction for horseradish peroxidase histochemistry. J.Histochem. Cytochem., 39:1143–1148.

Fig. 20. Thin section of a vibratome section from a rat that has received a crush injury to the cervicalvagus 24 months previously. The section has been incubated to demonstrate AcChE activity. Enzymeactivity is reduced in the neuropil but strong in this particular Type A neurone. Scale bar 5 2 µm.


Documents

Ultrastructural and cytochemical study of neurones in the rat dorsal motor nucleus of the vagus after axon crush