Electron Hirschsprung's - Gut · Electron microscopyofmyentericnervesin Hirschsprung'sdisease andin normalbowel Results in Normal Bowel MYENTERIC GANGLIA The ganglia were composed

Gut, 1970, 11, 1007-1014

Electron microscopy of myenteric nerves inHirschsprung's disease and in normal bowel

E. R. HOWARD AND J. R. GARRETTFrom the Departments ofSurgery and Oral Pathology, King's College Hospital, London

SUMMARY The ultrastructure of the myenteric nerves of colon and rectum removed from10 children with Hirschsprung's disease has been studied and compared with normal infantbowel.

Distal aganglionic (Hirschsprung) bowel often showed a rich supply of nerves within themuscle layers and there was no obvious morphological abnormality of constituent axons. Thenumbers of nerves diminished as more proximal parts of the bowel were examined and thefewest nerves were found where ganglia first appeared. These ganglia were similar in structure tothe ganglia of normal bowel, and a striking feature of them all was the absence of collagenbetween constituent neuronal units. The larger nerve trunks of aganglionic bowel frequentlycontained myelinated axons and these have been observed within the myenteric plexus ofnormal rectum.

This study supports previous histochemical investigations of the nerves in bowel frompatients with Hirschsprung's disease and indicates that the condition is due to a complex andvariable abnormality of the arrangement of the nervous tissue of the bowel wall, involvingmyenteric nerves as well as ganglia.

A recent study of cholinesterase-positive andadrenergic nerves in the bowel, resected fromyoung children, suggested that the symptomsand signs of Hirschsprung's disease were due toabnormalities in the distribution of the myentericnerves as well as to an absence of ganglion cells(Garrett, Howard, and Nixon, 1969). It wasconsidered that, in addition to the absence ofcoordinated contraction and relaxation conse-quent upon aganglionosis, increased numbers ofcholinergic nerves in the distal aganglionic bowelcause a relative increase in its tone. Furthermorea deficiency of these nerves in the distal gang-lionic bowel suggested a zone capable only ofreduced motor activity. These conclusions werebased on the assumption that the nerves demon-strated by histochemical techniques were func-tional and capable of activating the muscle of thebowel wall but it is not possible to determine thepresence or absence of actual neuro-effector sitesat light microscopical level.Received for publication 10 July 1970.

In the present investigation tissues from a num-ber of subjects included in the previous workhave been examined by electron microscopy.The bowel from patients with Hirschsprung'sdisease has been compared with normal bowelin order to establish first the normality orabnormality of the individual axons, and secondwhether neuro-effector sites exist, so that thevalidity of the conclusions based on the histo-chemical observations could be assessed.

Materials and Methods

Gut resected because of Hirschsprung's diseasefrom 10 children aged from 3 months to 2 years10 months has been studied. Normal recto-sigmoid removed for an unrelated conditionfrom a child of 3 months was used as a control.All the above tissues were included in the histo-chemical study (Garrett et al, 1969). In addition,

on June 10, 2021 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.11.12.1007 on 1 Decem

ber 1970. Dow

nloaded from

http://gut.bmj.com/

1008 E. R. Howard andJ. R. Garrett

Fig. 1 Electron micrograph ofganglion in normal infant colon, stained with uranyl acetate to show collagen.Note the absence of collagen within the ganglion. A 'dark' cell is present in the lower part of the ganglion(x 1,600).

because of the difficulty in obtaining furtherjuvenile control tissue, normal rectum from twoadults was examined.

Small, full-thickness strips of bowel were cutin a longitudinal direction from different levelsof each resected specimen adjacent to the tissuesexamined in the histochemical study. The tissueswere prepared as soon as the whole segment hadbeen removed in the operating theatre. Eachstrip, about 2 mm in width, was pinned with itscut surface downwards on cork in order to mini-mize distortion and then immersed in fixative.Initially phosphate-buffered glutaraldehyde fixa-tion (after the method of Sabatini, Bensch andBarmnett, 1963) was used for four hours. Sub-sequently it was found that better preservationwas achieved by fixing for two hours with theparaformaldehyde-glutaraldehyde mixture des-cribed by Karnovsky (1965). The tissues werewashed in 0.05M cacodylate buffer, pH 7.2,containing 7.5% sucrose. Each strip was thentransected into small blocks less than 2 mm wide,postfixed in 1 % osmium tetroxide, bufferedwith veronal acetate for two hours, dehydratedin alcohol, and embedded in Araldite. Ultra-thinsections were stained on grids with lead (Reynolds,1963) alone or after uranyl acetate and examinedin an E.M.6B electron microscope (AEI).

Fig. 2 Electron micrograph of normal adult rectumshowing a ganglion cell in close proximity to musclecells with a gap of about 1,sooA in places ( x 13,000).


http://gut.bmj.com

/G


ber 1970. Dow

nloaded from

http://gut.bmj.com/

Electron microscopy ofmyenteric nerves in Hirschsprung's disease and in normal bowel

Results in Normal Bowel

MYENTERIC GANGLIAThe ganglia were composed of extremely com-plicated arrangements of neurones, Schwanncells, and vast numbers of axons. The ganglionicmasses were surrounded by collagen and by thinfibroblastic-like cells in a single layer which wasoften incomplete. Although the majority ofganglia were confined to the intermyenteric zoneor submucosal regions, some were positionedpartly within the layers of the longitudinal orcircular muscle.The neurones were situated at the periphery of

the ganglia with much of their outer surfacenaked and unsupported (Figs. 1-3). This freeouter surface was often relatively close toadjacent muscle cell, sometimes with a gap ofless than 1 ,u (Fig. 2). The cytoplasm of theneurones was usually identifiable by its cyto-plasmic contents which included large numbersof mitochondria, endoplasmic reticulum, andfree and attached ribosomes. The preservation ofmitochondria was not always good, probablyas a consequence of the operative procedures.Occasional cells with a more dense cytoplasmwere seen, suggesting that there may be two typesof neurone but these cells usually had the leastwell preserved mitochondria of all.

Within the ganglia the majority of axons werebetween 0.4 and 1 g diameter but many wereoutside this range. The complexity of the arrange-ment of the axons and ganglion cells defiesaccurate description in the absence of serialreconstructions. It was impossible to defineaccurately the territory of each Schwann cell andusually a number of axons were present in eachinvagination of its plasma membrane. Sometimesthe axonal contents were not well preserved.Numerous vesicles wcre seen but without usingmore specific fixation (Richardson, 1966; Hdkfelt,1967) it is not possible to make any comment onwhether these were cholinergic, adrenergic, orneither. However, a fair number of large, dense-cored vesicles, about 1,o00A in diameter, wereobserved.A characteristic of the ganglia was the almost

complete absence of collagen within each gang-lionic mass (Fig. 1). Another striking feature wasthe paucity of definitive synaptic structures withthickening of the pre- and postsynaptic mem-branes, although such structures were occasionallyidentified (Fig. 4). However, numerous examplesof simple close contact between vesicle-containingaxons and the plasma membranes of neuronesand their processes, or with other axons, wereoften found and it seems possible that neuro-transmission can occur at these sites.

NERVES IN THE MUSCLE LAYERSLarge numbers of non-myelinated nerves werepresent in the muscle layers, particularly in the

circular muscle. The orientation of the axons,which ran in conjunction with Schwann cells,was similar to that of the circular and longitudinalmuscle fibres. The larger bundles were usuallysurrounded by a layer of perineural cells but thesmaller bundles were devoid of surrounding cells.Numerous individual Schwann-axon bundles,containing from one to 20 axons per Schwanncell, were found between muscle cells and occa-sional bare axons were seen but it was not possibleto be certain that an accompanying Schwann cellhad not been missed by the plane of the section.No distinct synaptic contacts were seen but axons,partially bared of Schwann cells or apparentlyunsupported and containing vesicles and mito-chondria, were frequently found with a freesurface in close proximity to a muscle cellseparated by a gap of less than s5oA (Figs. 5 and7). Such arrangements are considered to beneuro-effector sites. The adjacent muscle cellsoften showed concentrations of pinocytic vesicles.The axons were not always perfectly preservedbut numerous small agranular vesicles werefound and there were also many large granularvesicles of diameter about 1,OOOA.

Results in Hirschsprung's Disease

AGANGLIONIC BOWEL

Intermuscular zoneLarge nerve trunks characteristic of this conditionwere readily found in the intermuscular zone ofall cases (Fig. 9). Theywere composed ofhundredsof axons enclosed within a sheath of perineuralcells and collagen. Smaller trunks were alsopresent. The majority of axons were not myel-inated but in six out of the 10 cases the presenceof some myelinated axons was immediatelyapparent. The myelinated axons were betweenabout 1 0and26,uin diameter,and.werefoundonlyin the ensheathed trunks. Their ultimate des-tination is not known. The non-myelinated axonswere usually less than 1.5 ,u in diameter andgenerally ran in groups of two to four axons perSchwann cell. Each Schwann-axon bundle wasseparated from its neighbour by collagen. Theaxons contained tubules, vesicles, and mito-chondria, and showed no obvious pathologicalfeatures.

Because of the somewhat unexpected findingof myelinated axons in the intermuscular zone,two necropsy specimens of normal adult rectumwere specially stained for myelinated nerves. Thetissues were fixed in formalin, postfixed in theosmium-tetroxide mixture, and paraffin sectionswere examined by light microscopy. In bothspecimens scattered myelinated axons were found,including occasional ones in ganglia (Fig. 10).

1009 on June 10, 2021 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.11.12.1007 on 1 D

ecember 1970. D

ownloaded from

http://gut.bmj.com/

E. R. HowardandJ. R. Garrett

Fig. 3 Electron micrograph ofganglion in descendingcolon from a patient with Hirschsprung's diseaseshowing peripheral position ofa neurone ( x 4,400).

Fig. 4 Electron micrograph ofganglion in thetransverse colon from a patient with Hirschsprung'sdisease showing two axo-somatic synapses withthickening of the synaptic membranes (arrow)( x 20,000).



j.com/


ecember 1970. D

ownloaded from

http://gut.bmj.com/

Electron microscopy ofmyenteric nerves in Hirschsprung's disease and in normal bowel

Nerves in the muscle layersThe most distal segment of aganglionic bowel inseven out of the 10 cases contained numerousnon-myelinated nerves in the muscle layers,especially in the circular muscle. The nervesappeared similar to those in normal bowel and

showed no obvious pathological features. Some-times there were large numbers of axons perSchwann cell but this was not markedly differentfrom the normal. In any one section a number ofnerves appeared to be in transit, for the Schwann-axon bundle was separated from the muscle

Fig. 5 Electron micrograph ofnormal bowel showing Fig. 6 Electron micrograph ofaganglionic bowela vesiculated axon in very close proximity to a showing a vesiculated axon in very close proximitycircular muscle cell ( x 26,000). to a circular muscle cell ( x 15,000).

Fig. 7 Electron micrograph ofnormal bowel showing Fig. 8 Electron micrograph of aganglionic bowela vesiculated axon in very close proximity to a showing a vesiculated axon in very close proximitylongitudinal muscle cell ( x 24,000). to a longitudinal muscle cell (x 15,000).



j.com/


ecember 1970. D

ownloaded from

http://gut.bmj.com/


layers by a moderate amount of collagen.Numerous sites similar to those considered to beneuro-effector sites in the normal, in which anaxon was in close proximity to a muscle cell witha gap of 500A or less, were readily found (Figs.6 and 8). In these positions the axons tended tocontain increased numbers of vesicles and themuscle cell often showed concentrations ofpinocytic vesicles.As one ascended the aganglionic bowel the

numbers of nerves in the circular muscle alwaysdecreased.

In the remaining three cases few nerves werefound in the muscle layers at all levels of theaganglionic bowel including the most distal part.These three cases had been shown previously topossess very few cholinesterase positive oradrenergic nerves in the muscle layers (Garrett

>1'

Fig. 9 Electron micrograph ofpart of a large nervetrunk in aganglionic bowel showing two myelinatedand many unmyelinated axons ( x 4,800).

Fig. 10 Small myelinated axon within ganglion ofnormal adult rectum (osmium tetroxide, x 560).

et al, 1969), had presented with the mildestsymptoms, and did not show contraction of thedistal bowel on radiology.

GANGLIONIC BOWELIn five of the 10 specimens ganglionic tissue frombowel proximal to the aganglionic segment wasincluded.

GangliaWherever ganglia occurred they appeared essen-tially similar to those in normal bowel (Figs. 3and 4), and no obvious abnormalities weredetected. The sampling error and the smallnessof each sample in any electron-microscopicalsection makes it impossible to comment upontheir relative size or frequency.

Nerves in the muscle layersInvarably in the most distal ganglionic tissuevery few muscular nerves and correspondinglyfew 'neuro-effector sites' were found. In twocases with more proximal ganglionic tissues thenumbers of myenteric nerves and the 'neuro-effectot sites' approached normal. Thus it isconcluded that a variable length of ganglionicbowel, contiguous with the aganglionic segment,has a deficient myenteric innervation.

Discussion

It should be mentioned that imperfections in thepreservation of the tissues were not infrequent,the nerve tissue always being more affected thanthe muscle which was generally very well pre-served. These defects were probably the con-sequence of extensive handling of the tissue andthe time lag between devascularization andfixation, both of which are inevitable in operationsof this nature. Sometimes there was irregularpreservation within individual blocks, never-theless sections satisfactory enough for reasonableassessment were always achieved.The ganglia in the control tissues and those

present in some proximal parts of the Hirsch-sprung's specimens showed great complexity.They were essentially similar to those describedby Richardson (1958) in the small intestine of therabbit. The absence of collagen within the gangliawas totally dissimilar to the appearances withinconventional autonomic ganglia, such as sym-pathetic ganglia in the rat (Elfvin, 1963) orparasympathetic ganglia in the chorda of thecat (Garrett, 1966). In conventional autonomicganglia each neural unit, composed of a neuroneand processes, supporting cells and axons, isclearly separated from its neighbours by collagen.Human myenteric ganglia have a structure morelike primitive brain tissue and it is suggested thatthey should be considered as a system separate


http://gut.bmj.com

/G


ber 1970. Dow

nloaded from

http://gut.bmj.com/

1013 Electron microscopy ofmyenteric nerves in Hirschsprung's disease and in normal bowel

from, but influenced by, the parasympatheticand sympathetic divisions of the autonomicsystem.The peripheral distribution of the neurones with

part of their outer surface unsupported is alsoquite distinctive. We wonder whether a similararrangement with the free surface of neurones inclose proximity to muscle cells helps to explainthe proposition of Ambache and Freeman (1968),from work on the guinea-pig, that a non-cholin-ergic motor substance released by the gangliareaches the muscle layers by simple diffusion.Smith (1970) describes two morphological typesof myenteric neurone, and it is possible that thedarker cells occasionally seen in the present studyrepresent the argyrophilic cells described by her.However, their mitochondria were always poorlypreserved and it is possible that the morphologicalappearances were the consequence of a metaboliccatastrophe. The significance of the large granularvesicles frequently seen in axons is not under-stood. Their presence has been noted in Auer-bach's plexus of a human embryo a mere 12 mmlong by Kubozoe, Daikoku, and Takiti (1969),and similar vesicles have been found in bothadrenergic and cholinergic axons (Hokfelt, 1966).Although no obvious abnormalities were de-

tected in the ganglia from proximal tissue inHirschsprung's disease this does not necessarilyimply that they can mediate normal reflexactivity. It is possible that their input is impairedand certainly the innervations of the circularmuscle in the most distal ganglionic bowel wasalways deficient.The large nerve trunks in the intermuscular

zone of the aganglionic bowel were found to becholinesterase positive, and thus probably cholin-ergic, some years ago by Kamijo, Hiatt, andKoelle (1953). It is considered that they mayrepresent nerves of the sacral parasympatheticout-flow which would normally be going tomyenteric ganglia were ganglia present. Myel-inated axons have not been previously observedin these nerve trunks (Bodian, 1960; Smith, 1970),although the present investigation indicates thatthey are probably present in a large proportionof cases. Their function is not known but sincethey have now been found in normal gut (Fig. 10)they may be part of an afferent system.The numbers of nerves seen by electron micro-

scopy in the muscle layers of bowel resectedbecause of Hirschsprung's disease showed aclose correlation with the distribution of nervesdemonstrated histochemically on adjacent tissues(Garrett et al, 1969). Our conclusions based onthis distribution pattern, which were summarizedat the beginning of this paper, are thereforesupported by the present study. On the otherhand the normal bowel appeared to contain moremuscular nerves than had been anticipated by thelight microscopical study, and contrary toexpectations there were often many axons perSchwann cell. It has already been suggested thatthe

thicker, darker cholinesterase nerve staining oftenseen in the most distal aganglionic bowel was dueto the presence of more axons per Schwann cellthan in normal bowel, but electron microscopydoes not support this idea. We would nowsuggest that the apparent discrepancy in nervesize may be due either to the more stronglycholinesterase-positive axons in Hirschsprung'sdisease, or to a proportion of the axons in normalbowel being cholinesterase negative. If the latterbe true, and considering that few adrenergicnerves are present in the muscle layers (Bennett,Garrett, and Howard, 1968), is it possible that aproportion of axons in normal bowel representa non-cholinergic, non-adrenergic system whichis absent in aganglionic bowel?

It is now generally agreed that smooth muscleinnervation is by an en-passant type of closeapproximation between axons containing vesiclesand the muscle cells. Bennett and Rogers (1967)considered that a gap of 1,000A in the taenia coliof guinea pigs would allow neuro-activation. Inthe present study numerous sites of close associa-tion were found in normal bowel and in distalaganglionic bowel and not infrequently the gapwas 500A or less. Thus it would seem that in thewell innervated parts of Hirschsprung tissueneuro-effector sites do exist and there is somefunctional evidence in support of this. Forexample, uncoordinated contractions are oftenpresent in the aganglionic bowel and these can bestopped by low spinal anaesthesia (Ehrenpreis,1946; Bodian, Stephens, and Ward, 1949). Itwould appear that even if nerves are pregan-glionic they sometimes achieve a functionalrelationship with the muscle cells.

Thus, as well as the deficiency of ganglia inHirschsprung's disease, the pathology is in-fluenced by the distribution of myenteric nerveswhich exhibit no obvious morphological abnor-malities. Furthermore the deficiency of neuro-effector sites constantly found in the ganglionicbowel adjacent to the aganglionic segmentsuggests that this zone is capable of only reducedmotor activity and radiological studies supportthis idea (Ehrenpreis, 1946; State, 1965).

We are indebted to Mr H. H. Nixon for theclinical material used in this study. We are alsovery grateful to Mr K. J. Davies, Miss M. Egan,and Miss M. Davidson for technical assistance.During part of this study E.R.H. was assisted

by an MRC grant.

References

Ambache, N., and Freeman, M. A. (1968). Atropine-resistantlongitudinal muscle spasms due to excitation of non-cholinergic neurones in Auerbach's plexus. J. Physiol.,199, 705-727.


http://gut.bmj.com

/G


ber 1970. Dow

nloaded from

http://gut.bmj.com/


Bennett, A., Garrett, J. R., and Howard, E. R. (1968). Adrenergicmyenteric nerves in Hirschsprung's disease. Brit. med. J.,1,487-489.

Bennett, M. R., and Rogers, D. C. (1967). A study of the inner-vation of the taenia coli. J. cell. Biol., 33, 573-596.

Bodian, M. (1960). Pathological aids in the diagnosis and manage-ment of Hirschsprung's disease. In Recent Advances inClinical Pathology, edited by S. C. Dyke. Series 3, pp. 385-392. Churchill, London.

Bodian, M., Stephens, F. D., and Ward, B. C. H. (1949). Hirsch-sprung's disease and idiopathic megacolon. Lancet, 1, 6-11.

Ehrenpreis, T. (1946). Megacolon in the newborn. A clinical androentgenological study with special regard to the patho-genesis. Acta chir. scand., 94, Suppl., 112.

Elfvin, L. G. (1963). The ultrastructure of the superior cervicalsympathetic ganglion of the cat. 2: The structure of thepreganglionic end fibers and the synapses as studied byserial sections. J. Ultrastruct. Res., 8, 441-476.

Garrett, J. R. (1966). The innervation of salivary glands. II. Theultrastructure of nerves in normal glands of the cat. J. roy.micr. Soc., 85, 149-162.

Garrett, J. R., Howard, E. R., and Nixon, H. H. (1969). Auto-nomicnerves in rectum andcolon inHirschsprung's disease.Arch. Dis. Childh., 44, 406-417.

Hokfelt, T. (1966). Electron microscopic observations on nerveterminals in the intrinsic muscles of the albino rat iris.Acta physiol. scand., 67, 255-256.

Hokfelt, T. (1967). Electron microscopic studies on brain slicesfrom regions rich in catecholamine nerve terminals. Actaphysiol. scand., 69, 119-120.

Kamijo, K., Hiatt, R. B., and Koelle, G. B. (1953). Congenitalmegacolon. A comparison of the spastic and hypertrophiedsegments with respect to cholinesterase activities, andsensitivities to acetylcholine, D.F.P. and the barium ion.Gastroenterology, 24, 173-185.

Karnovsky, M. J. (1965). A formaldehyde-glutaraldehyde fixativeof high osmolality for use in electron microscopy. (Abstr.)J. cell. Biol., 27, 137-138A.

Kubozoe, T., Daikoku, S., and Takiti, S. (1969). Electronmicrok.scopic observations on Auerbach's plexus in a 12 mm.human embryo. J. neurovisc. Relat., 31, 291-307.

Reynolds, E. S. (1963). The use of lead citrate at high pH as anelectron-opaque stain in electron microscopy. J. cell. Biol.,17, 208-212.

Richardson, K. C. (1958). Electronmicroscopic observations onAuerbach's plexus in the rabbit, with special reference tothe problem of smooth muscle innervation. Amer. J. Anat.,103,99-135.

Richardson, K. C. (1966). Electronmicroscopic identification ofautonomic nerve endings. Nature (Lond.), 210, 756.

Sabatini, D. D., Bensch, K., and Barrnnett, R. J. (1963). Cyto-chemistry and electronmicroscopy. The preservation ofcellular ultrastructure and enzymatic activity by aldehydefixation. J. cell. Biol., 17, 19-58.

Smith, B. (1970). Disorders of the myenteric plexus. Gut, 11,271-274.

State, D. (1965). Rationale for segmental colon resection in thetreatment of congenital megacolon (Hirschsprung'sdisease). In Current Surgical Management, III, edited byE. H. Ellison, S. R. Friesen, and J. H. Mulholland,pp. 427-433. Saunders, Philadelphia.


http://gut.bmj.com

/G


ber 1970. Dow

nloaded from
http://gut.bmj.com/

Documents

Electron Hirschsprung's - Gut · Electron microscopyofmyentericnervesin Hirschsprung'sdisease andin normalbowel Results in Normal Bowel MYENTERIC GANGLIA The ganglia were composed