CELL JUNCTIONS AND NEURAL TUBE FORMATION IN CYNOPS ORIENTALIS

In the early embryogenesis of amphibia following gastrulation the mor-phogenetic movements of neurulationbecome actively in progress.Underthe action of the invaginating meso-derm the dorsal ectoderm is inducedto differentiate into neural plate whichwill be delimited by the neural foldsanteriorly and laterally.Later theneural plate becomes thickened withthe neural folds gradually grow medially till they fuse in the mid-lineto form the neural tube. It has already been shown bythe vital staining experiments thatthe morphogenetic movements arehighly coordinated between differentgerm layers and within the same germlayer and epiboly takes place in theectoderm as a whole,not separately by individual cells.Concerning thecoordination of cells during epibolywe have suggested that the tight junctions may play an important role.But how are the cells coordinatedafter their transformation into neuraltissues? How does the neural platemove as a whole in the formation ofthe neural tube? Disassembly of tight junctions andassembly of gap junctions were repor-ted during the neurulation of Ranapipiens and there is evidence ofdisruption of tight junctions in thesame process in the chick embryos.According to these authors,it seemsthat tight junctions are dispensableduring neural tube morphogenesis.Contradictory to these works,we havefound the continuous presence of bothtight and gap junctions during the developmental period from late gastrulato the closure of the neural tube. MATERIALS AND METHODS Embryos of Cynops orientalis serve as theexperimental material.Six stages as shownin table 1 were chosen.In order to insureonly the presumptive neuroepithelium andneuroepithelium cells were fractured and observed,the specimens were taken microsurgi-cally so that the neural folds,the presumptive caudal mesoderm and the mesodermunderlying the neural plate were not included.The extirpated specimens were folded laterallyalong the mid-line of the plate so that afterfixation and treatment with glycerol they wereinserted into the hole of the specimen holderin such a way that the fracture occurred atthe level about mid-trunk region. RESULTSLate gastrula (stages 14 and 15) (Figs.1—2) The apical borders of the cellsare provided with beltlike tight junctions whose degree of tightness varies.At some portion the junction is con-sisted of only 4—5 loosely connectedsealing strands (Fig.1),while at otherportion the sealing strands are morethan 10 and are tightly interwoven(Fig.2).Gap junctions are situated ??beneath the tight junctions andaround the junctions scatteredparticles are frequently found.Early neurula (stage 16) (Figs.3—4) Figures 3 and 4 are taken fromadjacent portions of a single cell.Thevariation in the degree of tightnessof the sealing strands of the tightjunction is apparent,from more than10 closely interwoven sealing strands(Fig.3) to 5-6 loosely connected ones(Fig.4).Gap junctions appear nearthe tight junction,sometimes connec-ted to the free ending of a sealingstrand with scattered particles nearthe site of connection (Fig.3),othersare in intimate contact with thestrand (Fig.9) and there may be someat a distance from the tight junction.Mid-neurula (stage 18) (Fig.5) Just as in late gastrula and earlyneurula,belt-like tight junctions arepresent at the apical border of theneuroepithelium cells.The junctionsvary at different portions of a singlecell,not only in the number but alsoin the degree of tightness of the sealingstrands.There may be only 2—3strands parallel to the cell surfaceor there may appear more than 10closely interwoven ones as shown infigure 5.Five gap junctions have beenobserved in this portion of the cell,situated beneath the tight junction,connected to the free ending of a strandor surrounded at all sides by the strands.The last mentioned condition hasfrequently been observed in cells ofall the other stages.(Figs.10—12).Late neurula (stages 20—22) (Fig.6) Shortly before the closure of theneural tube as the neural folds meetat the mid-dorsal line,the neuroepithelium cells are likewise provided withwell differentiated tight junctions,with sealing strands of free endings andin form of a loop isolated from allothers.Gap junctions are frequentlyobserved,often with 3—4 close toeach other (Fig.6).The sizes ofgap junctions vary.The largest at-tains 1.76μm~2 consisted of thousandsof particles with scattered ones nearby(Fig.13). DISCUSSIONSThe persistent occurence of cell junctions in all stages of neurulation Cell junctions,both tight junctions and gap junctions,are presentin the presumptive neuroepitheliumand neuroepithelium cells of all thestages,from late gastrula to the stageshortly before the closure of theneural tube (Figs.1—6).Althoughcell junctions occur persistently duringthe entire process of neurulation,yetwhat observed do not represent astable state.As for the tight junctions,the degrees of tightness of thesealing strands vary,not only in thenumber,from only 2—3 (Fig.5,lowerleft corner) to as many as more than10 (Figs.5,middle part;Fig.3),but ??also in the arrangement,in the form.of loosely connected loops (Figs.1and 6) or tightly interwoven networks(Fig.5,middle part;Fig.3).Suchvariations usually occur at differentportions of a single cell (Figs.3,4 and5). The observations of the ultrathinsections of lathanum treated speci-mens showed that for all these stagesthe tight junctions may exist in twodifferent conditions:La may be impeded at the surface of the cell (Fig.7) or it may flow across the junctions(Fig.8).These observations cor-respond very well with what have beenfound with the freeze-etching replicas.The former represents the portion withnumerous tightly interwoven strands,while the latter is the site wherestrands are few and loosely arranged.It is very probable that all what wehave observed represent the dynamicprocess of the continous morphological changes of the tight junctions. The persistent changes of gapjunctions are the assembly and disassembly of the particles with whichthe junctions are consisted.If theobserved and recorded gap junctionsare divided into groups according totheir sizes (Table 2),it can be seenthat in all the different stages thearea varies from below 0.02μm~2(Fig.14) to above 0.50μm~2 (Fig.13) with most between 0.02—0.50μm~2.The differences in sizes are veryevident among different cells of acertain stage and even within asingle cell.Scattered particles are fre-quently observed around the junctions,they may appear at one side (Fig.1and 13) or near the site of connectionwith the sealing strand of the tightjunctions (Figs.3 and 6).Whetherthe scattered particles are motivatingto or away from the junctions cannot yet be determined,but the varia-tions in junction sizes and the pre-sence of scattered particles around thejunctions are strong evidences that thedynamic changes of the junctions are ??persistently in progress. The observations mentioned aboveare in contradiction with what re-ported in Rana pipiens.Accord-ing to their observations,the tightjunctions present in the yolk plugstage became gradually loosened andfragmented and entirely disassembled at the neural tube stage,and si-multaDeously gap junctions displayeda process of assembly with a few par-ticles gathered together at first andthe number of particles increased andthey subsequently blended togetherinto larger aggregates.In chick em-bryos disruption of tight junctionshave been reported and large gapjunctions also seemed to disappearduring the process of neurulation.The contradictions in the results,es-pecially those with Rana pipiens,aredifficult to be explained,since it canhardly be attributed to the differencein experimental materials.The modeof fixation was suspected that theextirpation of the embryonic mater-ials before fixation might cause someeffect on the cell junctions.In orderto clarify this suspicion we carried outanother set of experiments with theintact embryos fixed overnight andthe neuroepithelium then extirpated.The replicas from such specimens didnot show any difference.They veri-fied the presence of tight and gapjunctions during all the stages exa-mined.Tight junctions and morphogeneticmovements The persistent occurrence andthe continuous dynamic changes ofcell junctions during the process ofneurulation should be considered withtheir pertaining functions.The func-tion of the tight junctions is general-ly accepted to be a permeabilitybarrier.They are present at the api-cal borders of the cells of tubular orvesicular structures,acting as thebarrier between the intercellular spacesand the external environment toimpede the outflow of the fluid con-tent.The correlations of the tightjunctions and the morphogenesis havealready been noted by some authors.In the formations of bile canaliculi,follicles of thyroid gland and kid-ney tubules,during or just beforethe appearance of the lumen the tightjunctions were always present in thecells destined to join in the formationof the luminal wall.Madara et al.,on the freeze-etching observations ofthe junctional complexes in fetal ratsmall intestine,pointed out the formation and the reorganization of thejunctional complexes during the trans-formation from the primitive strati-fied to simple columnar epitheliumand the development of the intestinallumen.This function of the tightjunctions is no doubt performed inthe early embryogenesis of amphibia.The tight junctions of the ectoderm ??cells of early gastrula impede the out-flow of the blastocoel fluid to main-tain the hydrostatic pressure of theblastocoel;those of the neural tubecells will likewise keep the hydrostaticpressure of the central canal. Besides the function to be a per-meability barrier,however,the tightjunctions may play very importantrole in the morphogenetic movementof early embryonic development.This has not been received enoughattention in the previous literature.During gastrulation as mesoderm andendoderm move to the interior theectoderm cells have to spread in orderto overgrowth the entire embryo.As