morphogenesis of intestinal villi -...

/ . Embryol exp. Morph. Vol. 34, 3, pp. 723-740, 1975 7 2 3

Printed in Great Britain

Morphogenesis of intestinal villi

II. Mechanism of formation of previllous ridges

By DAVID R. BURGESS1

From the Department of Zoology, University of California, Davis

SUMMARY

Villi lining the avian intestine originate from longitudinal folds (previllous ridges) runningthe length of the embryonic intestine. The morphogenetic events that occur in the epitheliumduring initial ridge formation in the chick embryo duodenum were examined by light andelectron microscopy. The epithelium, in cross-section, undergoes three stages prior to theformation of ridges; termed the circle (4^-6 days), the ellipse (6-8£ days), and the triangle(8̂ —9 days). At about 9 days of development three ridges form with three more forming oneday later. The mechanisms responsible for folding of the epithelium were examined. Micro-dissection followed by organ culture demonstrated that constriction by the surroundingcircular smooth muscle cannot account for folding of the epithelium. Mitotic pressurewithin the epithelium also cannot account for folding since there is no difference in thenumber of epithelial cells per cross-section between the ellipse and the triangle stages andthe epithelial tube is not restricted from expanding. Active constrictions in groups ofepithelial cells, mediated by bands of microfilaments, are thought to cause folding. Bundlesof microfilaments are localized in the apical region of all epithelial cells at all stages studiedand are localized in the basal region of those cells occupying the crests of the formingridges. Cytochalasin B-treatment prevented ridge formation and disrupted the bundles ofmicrofilaments.

INTRODUCTION

The development of shape or form, usually termed morphogenesis, is ofprime interest in the study of development. An example of an epithelial sheetthat undergoes dramatic change in form during development is the intestinalepithelium, which becomes shaped into finger-like villi that protrude into thelumen of the gut. The intestinal epithelium in the chicken does not form villidirectly, but first forms longitudinal folds, termed previllous ridges, runningthe length of the intestine (Hilton, 1902).

The processes by which sheets of epithelial cells form curved or lobulatedstructures have been under intensive investigation, and several possiblemechanisms have been proposed to account for such morphogenetic events.One proposed mechanism involves mitotic pressure within an epithelium.According to this concept the addition, by mitosis, of cells to the epitheliumwould force it to fold or buckle z/the epithelium were confined by some external

1 Author's address: Friday Harbor Laboratories, Friday Harbor, Washington 98250,U.S.A.

46 E M B 34

724 D. R. BURGESS

force such as surrounding tissues or mesenchyme (Zwann & Hendrix, 1973).Another proposed mechanism to account for folding of epithelial cell sheetsis based on 'active' contractions by individual cells in the tissue. Such activecell contractions by a group of cells within a sheet can account for folding if:(1) contraction is restricted either to the apical or basal portion of the cells; and(2) the cells retain their adhesions within the epithelium.

Localized contractions within individual cells are thought to be mediated bycontractile intracellular microfilaments, 4-6 nm in diameter (see Wessells et al.1971; Schroeder, 1973; Spooner, 1973). Contractile microfilaments in epithelialcells have been proposed to supply the force required for cell-shape changesresponsible for amphibian neurulation (Baker & Schroeder, 1967; Schroeder,1970; Burnside, 1971; Karfunkel, 1971), pancreas morphogenesis (Wessells &Evans, 1968), lens invagination (Wrenn & Wessells, 1969), morphogenesis ofsalivary epithelium (Spooner & Wessells, 1970, 1972) and morphogenesis ofoviducal epithelia (Wrenn & Wessells, 1970; Wrenn, 1971).

Several lines of evidence support the contention that intracellular microfila-ments are involved in active cell constrictions which are thought to causefolding of epithelial cell sheets. There is a spatial and temporal correlationbetween the appearance of microfilament bundles and cell shape changes infolding epithelial sheets (Schroeder, 1970; Burnside, 1971). Also, microfilamentsdemonstrate a structural and biochemical similarity to actin, a contractileprotein of muscle (Ishikawa, Bischoff & Holtzer, 1969; Pollard & Weihing,1974).

Cytochalasin B (CB) has been widely used in the investigation of cell motilitybecause it inhibits many kinds of cell movements, as first described by Carter(1967). The list of folding in epithelia inhibited by CB includes morphogenesisof oviduct epithelium (Wrenn & Wessells, 1970; Wrenn, 1971), morphogenesisof salivary gland epithelium (Spooner & Wessells, 1970, 1972), and neurulation(Karfunkel, 1971). CB not only prevents folding in these systems, but alsodisrupts the structure of the bands of microfilaments within those epithelialcells that change shape during the folding process (Wrenn & Wessells, 1970;Wrenn, 1971; Cloney, 1972; Spooner & Wessells, 1972).

The three-dimensional changes that occur in the duodenum of the chickembryo during previllous ridge formation have been described by a variety oftechniques (Hilton, 1902; Coulombre & Coulombre, 1958; Grey, 1972). Thefirst paper in this series described the very precise and predictable morpho-genetic folding of the embryonic chick duodenum that leads to the establishmentof definitive villi (Grey, 1972). The mechanisms responsible for the conversionof the tubular epithelium into an intricate pattern of previllous ridges have not,however, been explored. This paper describes the central events involved in theestablishment of the first previllous ridges and reports on experiments designedto illuminate the mechanisms of ridge formation.

Morphogenesis of intestinal villi. II 725

MATERIALS AND METHODS

Embryos

Eggs from a commercial line of White Leghorn chickens were obtained froma local hatchery and incubated in a forced-draft incubator. All embryos weresacrificed by decapitation and staged according to Hamburger & Hamilton(1951). The proximal end of the duodenum from the pylorus to the apex of theloop was used for all parts of the study.

Organ culture

Embryos were dissected in warm Hank's balanced salt solution (HBSS) andfragments 1-2 mm in length were cut from the proximal end of the duodenum.Tissues were cultured in Eagle's Minimal Essential Medium containing 10%fetal calf serum and 100 i.u./ml penicillin and 100 /*g/ml streptomycin (MEM)(Grand Island Biological Co. or Pacific Biological Co.). Fragments werecultured floating in MEM in 35 mm culture dishes (Falcon Plastics, Inc.) in anatmosphere of 5% CO2 in air. For some experiments varying amounts ofmesenchyme were removed prior to culture. This was done either by dissectionwith tungsten needles or with sharpened fine forceps. Alternatively, treatmentof stage-34 intestines with 1% trypsin in calcium- and magnesium-free Hank'sbalanced salt solution for 30 min at 4 °C allowed the successful separation ofepithelium from mesenchyme.

Fragments of intact duodena from stage-34 embryos were also cultured for24 to 36 h in glucose-free MEM (GFMEM). GFMEM was prepared fromEagle's Minimal Essential Medium without glucose (Pacific Biological Co.) towhich fetal calf serum was added to achieve a final concentration of 10%. Thefetal calf serum had been dialyzed for 2 days against 0-85 % NaCl and for1 day against glucose-free HBSS (GFHBSS).

Effect of Cytochalasin B on epithelial morphogenesis

In some experiments 1- to 2-mm segments of intact duodena from stage-34and stage-36 embryos were cultured for 24-36 h in MEM. To some of thesecultures Cytochalasin B (CB) (Imperial Chemical Industries, Ltd) was addedto achieve a final concentration in the medium of 1-0/tg/ml. For controlcultures a volume of dimethylsulfoxide (DMSO), the solvent for CB, equal tothe volume of CB solution added to the experimental cultures was added to thecontrol dish. In order to test for reversibility of the drug's effect, CB-treatedfragments from stage-34 embryos that had been cultured for 24-36 h were washedthree to four times with fresh MEM for 30 mins to 1 h and then reincubatedfor 24-36 h.

MicroscopyTissues were fixed at room temperature in 2% glutaraldehyde buffered in

0-1 M cacodylate buffer, pH 7-2, and postfixed in 1% osmium tetroxide.

726 D. R. BURGESS

Tissues were embedded in Epon and sectioned with glass knives on a Porter-Blum MT-1 ultramicrotome or with a diamond knife on a Porter-Blum MT-2ultramicrotome. One-micrometer sections, stained with a solution of 1 %toluidine blue and 1% boric acid, were routinely taken for light microscopy.Thin sections were stained in 2 % aqueous uranyl acetate and lead citrate andexamined with a Hitachi HU-11E electron microscope.

Some tissues and cultured fragments were prepared for scanning electronmicroscopy. The tissues were fixed, postfixed, and dehydrated as usual, thendehydrated by passage through a graded series of amyl acetate. Final prepara-tion was by the critical point method as utilized by Grey (1972). The fragmentswere coated with gold and silver, using a vacuum evaporator equipped with arotating specimen holder, and examined in a Cambridge Stereoscan electronmicroscope.

RESULTS

Major developmental changes in the configuration of the intestinal epithelium

Between 8 and 12 days of incubation the embryonic intestine constructs theprevillous ridges upon which the definitive villi later form. These early eventsof ridge formation can be easily followed in cross-sections of the proximalduodenal loop. At about A\ days of incubation (stage 24) the intestinal epi-thelium appears in cross-section as a thick-walled circular tube with a smalllumen (Fig. 1 A). From stage 26 to approximately stage 30 (5-7 days of incuba-tion) the cross-sectional profile of the epithelium becomes elliptical and thelumen begins to expand (Fig. 1C-F). The circumference of the ellipse increasesbetween stage 30 and stage 34 (7-8 days of incubation). Most of the increaseoccurs in the long axis; the width remains fairly constant (Table 1).

During the 6-8 h required for the embryo to advance from stage 34 to stage35 the elliptical tube of epithelium is transformed into a triangular-shaped tube(Fig. 1G). Immediately following its formation, the sides of the triangle appearto fold toward the lumen so that the first three previllous ridges are formed(Fig. 1H-J).

During the 24-36 h of development after formation of the first three ridges,one previllous ridge usually forms in the location occupied by the valley

FIGURE 1

Cross-sections (1 ftm Epon sections) through duodenal fragments from embryosranging from 4£ to 12 days of development. The cross-sectional shape of theepithelium at different stages of development is represented in this series of micro-graphs. All x 145. (A) Circle, stage-24 embryo; (B) small ellipse, stage-26 embryo;(C) small ellipse, stage-29 embryo; (D) elongated ellipse, stage-30 embryo; (E)elongated ellipse, stage-33 embryo; (F) elongated ellipse, stage-34 embryo; (G)forming triangle, stage-34 + embryo; (H) triangle, stage-35 embryo; (I) three-ridge,stage-36 embryo; (J) three-ridge, stage-36 embryo; (K) six-ridge, stage-37 embryo;(L) formed previllous ridges, stage-38-39 embryo.


46-2

728 D. R. BURGESS

between two established ridges so that a total of six previllous ridges arepresent (Fig. IK). After this period ridge formation becomes more irregular.There are generally about eight previllous ridges formed by eleven days andsixteen by thirteen days (Fig. 1L and Grey, 1972).

Table 1. Cross-sectional dimensions of epithelial tubes at differentstages of development

Embryonicstage*

Shape of Average Number oflumen Measurements (/*m) Range (/tm) samples

4* (24)5-7 (26-30)

7_8 (30-34)

8-8* (34-35)

OvalSmall ellipse

Elongated ellipse

Triangle

DiameterLengthWidth

LengthWidthBaseHeight

127151106

181114177145

120-130142-16293-115

160-19698-135

172-184140-152

510

9

5

* The first numbers refer to days of incubation. The figures in parentheses are theHamburger-Hamilton stage.

Formation of circular smooth muscle

Coulombre & Coulombre (1958) suggested that contraction by the developingcircular smooth muscle of the epithelial tube could cause the formation of theprevillous ridges by forcing the epithelium to buckle. It was necessary, therefore,to determine if the forming circular smooth muscle has, at the time immediatelyprior to folding, the contractile machinery necessary for contraction.

Up to the six-ridge stage there are about 15-20 layers of concentricallyarranged, loosely packed mesenchymal cells around the epithelium. Beyondthese loosely packed cell layers are 10-20 cell layers of tightly packed mesen-chymal cells which will form circular smooth muscle (Fig. 1). By stage 34(ellipse stage) organized contractile apparatuses are evident in many of thesecells.

The hypothesis that the band of smooth muscle produces folding of theepithelium was tested by removing the mesenchyme from one side of theelliptically-shaped epithelium in intestinal segments of stage-34 embryos (Fig.2A). The mesenchyme was removed by dissection using sharpened fine forceps.This operation removed the forming circular smooth muscle but did notremove the two to six loosely packed mesenchymal cell layers adjacent to theepithelium. The circular band of smooth muscle was therefore prevented frombeing continuous: consequently it could exert little or no force on the intactepithelial tube. When such experimentally altered fragments were cultured for24-36 h, all epithelial tubes folded to form three ridges (Fig. 2B). Looselypacked mesenchymal cells repopulated the region previously occupied by the

Morphogenesis of intestinal villi. II 729smooth muscle; these invading mesenchyme cells did not, however, formsmooth muscle. In all cultured fragments the epithelium overgrew the ends ofthe tube and covered the surface of the whole fragment, but this overgrowthappeared to have no effect on folding of the epithelium.

The conclusion that smooth muscle is not required for folding of the epi-thelium is strengthened by evidence from a more radical experiment. Twelveellipse-stage intestines from eight embryos were cut open lengthwise to formflat fragments with the epithelium on the surface. This surgery was done byslitting duodenal fragments open lengthwise with a sharpened tungsten needlecutting through both mesenchyme and one side of the epithelial tube. Imme-diately after such surgery the epithelial sheet was at first slightly curved (Fig. 3),but flattened out within five or six hours of culture. Ridge formation in thesefragments was delayed by about 24 h as compared to cultured intact duodenalfragments. After culturing in MEM for 48 h the epithelia formed ridges in nineof the twelve fragments. Four of the fragments had formed six or more ridges(Fig. 4); the others formed two or three ridges. In most cases the ridges wereshorter than normal and highly irregular in contour (Fig. 4). Whether theirregular ridges that form in these experimental cases do so by the samemechanism as the very regular ridges that normally form is not clear. It doesseem possible, however, to rule out the notion that the growing epitheliumsimply buckled because it was prevented from spreading. Extensive spreading ofthe epithelium occurred not only in those fragments that formed ridges, butalso in those that did not.

Although smooth muscle was not required for epithelial folding, it was foundthat mesenchyme was required for folding morphogenesis. Intact epithelialtubes from early ellipse-stage intestines, as isolated by brief trypsin treatment,cultured for up to 48 h in MEM failed to undergo folding morphogenesis.

Cellular dynamics in the epithelium associated with previllous ridge formation

If the duodenal epithelium was restricted from expanding by the surroundingmesenchyme and smooth muscle concomitant with cell proliferation in theepithelium, the epithelial tube would be forced to buckle. The possibility thatcell proliferation might play a role in the folding of the intestinal epitheliumwas therefore examined.

In order to correlate the growth pattern (i.e. addition of cells) with themorphogenetic folding of the intestine, the number of cells in cross-sections ofthe epithelium was counted at all stages from the circle (4^ days) to the three-ridge (10 days) stage. These data are summarized in Fig. 7. No significant in-crease in cell number occurs between 4y and 6\ days, i.e. during the periodwhen the epithelium changes from the circle to the ellipse stage. Thus thereappears to be no correlation between growth and shape of the epitheliumduring this period. Between 6\ and 8 days the cell number increases 38 %. Thisincrease is accompanied by a lengthening of the ellipse. A major shape change

46-3

730 D. R. BURGESS


occurs between 8 and 8^ days when the epithelium changes from elliptical totriangular in shape. There is, however, no addition of cells to the cross-sectionalarea during this period. A second prominent increase in the number of cellsoccurs between the ninth and tenth day of development; during this period thenumber of cells increases over 50%. This second saltation, unlike the first,occurs concomitantly with a shape change in the epithelium: the appearanceof ridges in the epithelium.

The possibility that mitotic pressure (during the second saltation) played acausal role in the formation of ridges was explored further. Mitotic pressurecan effect the folding of an epithelium only if the epithelium is somehow con-fined to a restricted volume. An attempt was made, therefore, to estimate thevolume available to the intestinal epithelium as it acquires the increased numberof cells shown in Fig. 7. Such an estimate can be made by measuring thediameter of a circle drawn around the outer limits of the epithelium. Comparisonof the triangle stage and the three-ridge stage shows, however, that a largeincrease in the diameter occurs between these two stages (an average of 162 jmnfor six triangles and 216 /im for three three-ridges). Such an increase indicatesthat during the transformation of the triangle to the three-ridge stage, theepithelium is not completely restricted by the surrounding mesenchyme.

FIGURES 2-6

Fig. 2A. Drawing representing the fragment of tissue produced by removing halfthe mesenchyme from an ellipse-stage piece of intestine.

Fig. 2B. Epithelial tube from ellipse-stage intestine cultured 24-36 h. The mesen-chyme layers had been removed as shown in Fig. 2, and during the culture periodloose-packing mesenchyme cells filled in the area that had been extirpated (blackarrows). The circular smooth muscle did not invade the region that had beenremoved, but remained in its previous location (white arrows). A total of eightfragments from six embryos were altered and cultured for this experiment. Allformed ridges, x 190.

Fig. 3. Scanning electron micrograph of an ellipse-stage intestine that had beenslit open lengthwise and fixed immediately. The epithelium (outlined by brackets)remains slightly curved as seen from the luminal surface in this view. The mesenchyme(mes) lies on either side of the epithelium, bordering it. x 260.

Fig. 4. Scanning electron micrograph of a slit-open ellipse-stage intestine that hadbeen cultured for 48 h. Numerous uneven ridges formed that were wavy andbranching, x 250.

Fig. 5. Thin section of the luminal area of an epithelial cell from an ellipse-stageintestine showing a fibrillar region (arrow) extending away from the intermediatejunction region and terminating a short distance away in the cytoplasm. Thisregion appears fuzzy in many cells indicating that the microfilament band had beencut in cross-section, x 42 500.

Fig. 6. A glancing section of the luminal region of epithelial cells showing a tangen-tially cut intermediate junction region. A large microfilament band (arrows) withmany organized filaments is seen running parallel with the junction in this region,x 25000.

732 D. R. BURGESS

Coulombre & Coulombre (1958) reported that the diameter of the wholeduodenum does not increase while the overall length of the proximal loop doesincrease from 6 to 12 days of development.

6 7

Days of incubation

10

Fig. 7. Graph plotting the number of cells per cross-sectional view of the intestinalepithelium versus age of the embryo. This graph demonstrates two phases ofgrowth of the epithelium during the morphogenetic period studied. The line isdrawn through the mean values with the ranges represented by the bars. The follow-ing number of epithelia were used in determining the values plotted: circle, 7;small ellipse, 10; elongated ellipse, 11; triangle, 12; three-ridge, 3.

Role of cell-shape changes in the formation of the first set of previllous ridges

If cells in precise regions of the epithelium actively contracted to changetheir shape, folding of the epithelial sheet could occur. An ultrastructuralanalysis of the epithelial cells was therefore undertaken with particular attentionfocused upon the location of cytoplasmic microfilaments. The first question tobe answered concerned the region within individual epithelial cells in whichmicrofilaments were most abundant.

Microfilaments are most conspicuous in the luminal neck where they appearto insert as large bundles into the intermediate junction region (Fig. 5), distalto the desmosomes with their associated tonofilaments. They are present inthis region throughout the stages studied. In this apical region of flask-shapedcells, the bands of microfilaments primarily course around the apical perimeterof the cell at the level of the intermediate junction and usually do not connectthe cell at directly opposite points on the surface. This arrangement is apparentfrom most micrographs, which demonstrate a fibrillar or fuzzy zone emanatingfrom the region of the intermediate junction region and apparently ending inthe cytoplasm a short distance away. However, when the tight junctionai


region is cut tangentially over an extended area, a band of microfilaments canbe observed running parallel with the membrane (Fig. 6). Bands of micro-filaments are also localized along the bases of the epithelial cells, parallel withand immediately adjacent to the basal plasma membrane (Fig. 8).

The second major question to be answered was whether the apical or basalgroups of filaments were prominent only in cells in certain regions of theepithelium. No preferential localization could be observed in the apical bandof microfilaments. When the tube is triangular in cross-section, apical micro-filamentous bands are as frequent in the cells making up the sides of the triangleas in those cells in the corners. This general observation is true for the ellipsestage as well.

Basal microfilaments are, however, most easily resolved as organized bandsin the cells comprising the sides of the triangular-shaped tube as they buckleinward to form the first previllous ridges (Fig. 8). The basal surfaces of thesecells are narrower in diameter than those of cells in corners of the triangle andare usually folded into extended pseudopods.

Cytochalasin B was used to study the possible role of microfilaments in con-trolling active cell-shape changes, and thus the formation of previllous ridges.Control explants of whole duodenal fragments from stage-34 embryos culturedeither in the presence or absence of DMSO formed three previllous ridgeswithin 24-36 h. A common characteristic of all control cultures was the rapidoutgrowth of the epithelium from the cut ends of the tubes. Aside from thetighter packing of peripheral mesenchyme in some fragments, the appearanceof the epithelium and mesenchyme in control tissues cultured 24-36 h wasindistinguishable from normal three-ridge-stage intestines at both the light andelectron microscope level (Fig. 9B).

Isolated ellipse-stage intestines grown in the presence of 1 ^g/ral CBdid not fold (Table 2 and Fig. 9C). The general ultrastructural appearanceof CB-treated intestinal fragments was generally comparable to that of nor-mal intestines. The structure, location, and orientation of microtubules anddesmosomal tonofilaments of CB-treated intestines did not differ from controlcultures or normal tissue. The mesenchymal layers were generally less denselypacked than normal.

Some significant effects of CB were, however, noted. In contrast to controltissues, the epithelium in CB-treated fragments did not spread out from theends of the tubes. Also, in CB-treated intestines the apical surfaces of the cellsbulged into the lumen, reminiscent of normal cells in the circle stage of develop-ment. This bulging effect ranged from epithelium which could not be dis-tinguished from control cultures to epithelium in which many of the cellsappeared to have extruded portions of apical cytoplasm into the lumen.

Particular attention was paid to the structure of the microfilament bands inthe epithelium of CB-treated tissues. There was a subtle difference between thestructure of these bands in the luminal region of CB-treated cells and those of

734 D. R. BURGESS

11 A

Morphogenesis of intestinal villi. II 735control cultures. The response of microfilaments in the apical region of the

cells was variable. The apical cytoplasm of many epithelial cells exposed to CB

exhibited an increased granularity or denseness in place of organized bands of

microfilaments (Fig. 10). In the apical region of some epithelial cells from CB-

Table 2. Effect of CB on the formation and maintenance of ridges infragments of embryonic intestine in organ culture

Experiment Results

Ellipse-stage intestines cultured 24-36 h in: Number of fragments folded*(1) MEM+DMSO (control) 25/30(2)MEM + CB 5/50(3) MEM + CB, wash, MEM + DMSO 7/151

(recovery from CB)Intestines with established ridges cultured 24 h in: Number maintaining ridges

(1) MEM + DMSO (control) 9/9(2)MEM + CB 12/13

* Folding is defined as the development either to the triangle stage or to the distinctthree-ridge stage.

t Those that recovered formed abnormal-looking triangle stages.

treated fragments a web of microfilaments with many properly oriented micro-filaments was present. Basal bands of microfilaments were not observed in CB-treated epithelia and the region normally occupied by such bands appeared asa fuzzy zone.

FIGURES 8-11

Fig. 8. Basal region of epithelial cell occupying the crest of a forming ridge of athree-ridge-stage intestine. A band of organized microfilaments (arrows) is presentrunning parallel with the basal plasma membrane, x 52000.Fig. 9. (A) An ellipse-stage intestine prior to culture. (B) A triangle-stage intestineformed while cultured in MEM + DMSO for 24 h. The epithelium folded normallywhile the mesenchyme became more tightly packed. (C) An ellipse-stage fragmentcultured for 24 h in 1 /tg/ml CB. Folding of the epithelium did not proceed duringthis period. (A), (B), and (C) x 180.Fig. 10. Masses of finely granular material in the region normally occupied bymicrofilament bands in an epithelial cell from ellipse-stage intestine cultured for 24 hin the presence of 1 /tg/ml CB. x 32000.Fig. 11. (A) Thick section of a triangle-stage intestine dissected free of envelopingcircular smooth muscle and most of the mesenchyme and cultured in MEM for6 h. The triangle form was maintained in preparations like this one and in triangle-and three-ridge-stage intestines isolated free of mesenchyme with the use oftrypsin. x 240. (B) Six-ridge-stage intestine cultured for 24 h in the presence of1 /tg/ml CB. Although large intercellular spaces formed in the mesenchyme thestructure of formed ridges was maintained in 12 of 13 intestinal fragments culturedin CB. x 185.

736 D. R. BURGESS

To determine if the CB effects were due to a general toxic effect not observableultrastructurally, tissues were allowed to recover from CB treatment. Of 15intestinal fragments (from 10 embryos) allowed to recover from CB treatment,only seven fragments recovered to form ridge-like structures. In those that didrecover, the diameter of the epithelial tube and the height of the epithelial cellswas decreased. The epithelium did, however, recover to the extent that it grewout the cut ends of the tubes to cover the mesenchyme.

CB has also been shown to inhibit sugar transport across the plasma mem-brane (Kletzien & Perdue, 1973). In order to test whether CB prevented ridgeformation by preventing uptake of glucose by the epithelial cells, fragments ofintestines without ridges were cultured in glucose-free MEM (GFMEM). After24-48 h, eight of nine fragments from six embryos had formed ridges inGFMEM. The epithelial cells appeared normal in all respects, but the mesen-chyme appeared abnormal in that large numbers of intercellular spaces ap-peared. That folding of the epithelium occurred in the absence of glucosesuggests that the effect of CB in preventing folding of the intestinal epitheliumis independent of an effect on transport of glucose.

Stability of previllous ridges

Although the data presented in the preceding sections indicate that themechanism of formation of ridges is an intrinsic property of the epithelium,there remained the possibility that the circular muscle layers act to maintainridges, once they have formed. In order to test this possibility the mesenchymelayers were removed either manually, a procedure that leaves some mesenchymeadherent to the epithelium, or enzymically, a procedure that produces epithelialtubes that are completely free of mesenchyme. Three- and six-ridge-stageepithelial tubes that were freed from mesenchyme by either of these techniquesmaintained their original shape for up to 6 h in culture (Fig. 11 A). Longertimes were not examined.

Established ridges were also stable in the presence of CB at a concentrationof 1 -0 /£g/ml. Thirteen fragments of whole three- to six-ridge-stage intestineswere cultured for 24 h in the presence of CB. The structure of the ridges wasmaintained in 12 of 13 cultured fragments (Fig. 11B).

DISCUSSION

The morphogenetic changes in the epithelium and mesenchyme of the earlychick embryo intestine are quite precise and predictable. The epitheliumexhibits three distinct stages during the establishment of the first three previllousridges; these stages have been termed the circle, ellipse, and triangle. Theobservation that villar morphogenesis begins with the formation of threeprevillous ridges is consistent with the finding of Hilton (1902). These findingsconflict with those of Coulombre & Coulombre (1958) who reported that the


embryonic chick duodenum initially forms only two previllous ridges. Sinceridge formation begins with a starting number of three, the Coulombre's con-tention that ridges are added in a geometric progression (2, 4, 8, 16, 32) needsto be re-examined. The apparent discrepancy in the number of initial previllousridges as reported in this study versus that reported by the Coulombres may bedue to the region of the developing duodenum examined by the Coulombres.Nevertheless, there appears to be an average of eight previllous ridges in the11-day duodenum (Grey, 1972). The transition to this stage from the three-ridge stage has not been studied.

This investigation sought to determine the location of the force(s) thatgenerates the previllous ridges. The first question, therefore, was whether thefolding mechanism or force was extrinsic to the epithelium (i.e. located insurrounding mesenchymal or muscular layers) or, alternatively, whether thefolding mechanism resides within the intestinal epithelium itself.

Several lines of evidence militate against the notion that the force is extrinsicto the epithelium. Folding occurs when the continuity of the circular layer ofsmooth muscle is disrupted (Fig. 2). Folding occurs even when the epithelialtube is slit open lengthwise to permit culture of the intestinal fragment as a flatsheet (Fig. 4). The diameter of the epithelial tube also increases as folding occurs(Fig. 1); although this observation does not, of itself, rule out the possibility thatthe folding force is extrinsic, it does establish that the space available to theepithelium does not remain fixed throughout the period of folding. Theseexperiments and observations, taken together, provide clear evidence that amechanism for folding based on mechanical constriction, as proposed byCoulombre & Coulombre (1958), is not applicable to the initial formation ofprevillous ridges.

The question was then addressed as to what factors intrinsic to the epitheliumcould be responsible for folding. Mitotic pressure within the epithelium seemsto be an unlikely mechanism, since certain periods of folding occur without aconcomitant increase in cell number (Fig. 7). For cell proliferation to be re-sponsible for folding, the epithelium would also have to be restricted fromexpanding. Since: (1) ridges form when the epithelial tube is slit open length-wise, thereby permitting the epithelium to migrate laterally beyond its normallimits, and (2) the circular sheath of mesenchyme surrounding the epitheliumexpands in diameter during folding, there is no evidence to suggest that theepithelium is restricted to a fixed diameter during folding.

Several lines of evidence are consistent with the hypothesis that cytoplasmicmicrofilaments are responsible for epithelial folding. Microfilaments arepresent in the basal region of only those cells on the apices of the folding ridges.The presence of microfilaments in only those cells with narrow basal ends isconsistent with the notion that microfilaments are responsible for the narrowingand buckling of these cells, and that the cellular constrictions are responsiblefor folding of the epithelium. Similar basal buckling was observed in salivary

738 D. R. BURGESS

gland epithelial cells during cleft formation (Spooner & Wessells 1970, 1972).By contrast, apical microfilaments are present in all cells of the epitheliumduring all stages studied.

Intracellular microfilaments have been localized in a number of other foldingepithelial sheets, including the neural plate (Schroeder, 1970; Burnside, 1971;Karfunkel, 1971), forming glands in the oviduct (Wrenn, 1971), developinglens (Wrenn & Wessells, 1969) and the developing salivary gland (Spooner &Wessells, 1970, 1972). In all these cases except for the salivary gland and theintestinal epithelium, microfilaments are preferentially localized in thoseregions of the cells which change shape and only in those cells whose shapechange could bring about the final conformation of the epithelium. An interest-ing point is that in the neural plate, and also in the lens, cell-shape changesneed occur only once in order to establish the correct final shape of the organ.In the cases of the epithelium from the salivary gland and the intestine, bothof which undergo numerous foldings, apical microfilaments are present in allepithelial cells and are not preferentially localized in cells whose changes inshape could account for folding morphogenesis. One explanation of thepresence of microfilaments in all cells of an epithelium during folding is thatall cells in such continually folding epithelia should possess microfilaments,since all cells ultimately change shape. Folding in such epithelia cannot becontrolled by the mere presence or absence of microfilaments. Instead, thecontrol mechanism must involve the selective stimulation of contraction ofmicrofilaments in only certain cells.

Further evidence that microfilaments are responsible for active cell constric-tions leading to folding of the intestinal epithelium comes from experimentsusing CB on cultured ellipse-stage intestines. CB inhibited folding of theepithelial cells. The locations previously occupied by microfilament bands werereplaced by highly granular dense material. The apparent dissolution of bandsof microfilaments and the concomitant prevention of folding in the intestineimplicates microfilaments as being responsible for folding of previllous ridges.Similar correlations have been noted in other developing epithelia (Spooner &Wessells, 1972). The failure of CB to abolish microfilaments in all cells of theintestinal epithelium is inconsistent with the reports of the effects of this agent inother developing epithelia. The reason for this difference is unclear. CB ap-parently does not act by disrupting sugar transport since ridge formation pro-ceeds normally in medium lacking glucose, indicating that the effects of CBon sugar transport and on folding are independent.

The forces that act to maintain established ridges once they are formed arenot well understood. It has been proposed that collagen fibers surroundingthe developing salivary epithelium act to stabilize established clefts (Bernfield &Banerjee, 1972). Isolated three-ridge-stage intestinal epithelia, however, main-tain their folded configuration even with little or no mesenchyme or otherextracellular materials present. This result indicates a degree of intrinsic stability


to the folded intestinal epithelium. Such stability is unaffected by CBtreatment.

There are a number of problems that remain unresolved. At the tissue levelthe central remaining questions concern the factors that regulate the numberof ridges and their time of appearance. At the cellular level it is not known whatfactors control the placement and activation of contraction of microfilamentsin the epithelial cells. Further work on the developing intestine may provideanswers to some of these questions.

The author would like to thank Drs Ursula K. Abbott, John H. Crowe and Robert D.Grey for helpful suggestions during the course of this work and Dr Grey for critical readingof the manuscript. The author also thanks Mr John Mais and Ms Carolyn Ishikawa forexcellent technical assistance.

REFERENCES

BAKER, P. C. & SCHROEDER, T. E. (1967). Cytoplasmic filaments and morphogenetic move-ment in the amphibian neural tube. Devi Biol. 15, 432-450.

BERNFIELD, M. R. & BANERJEE, S. D. (1972). Acid mucopoly-saccharide (glycosaminoglycan)at the epithelial-mesenchymal interface of mouse embryo salivary glands. /. Cell Biol.52, 664-673.

BURNSIDE, B. (1971). Microtubules and microfilaments in newt neurulation. Devi Biol. 26,416-441.

CARTER, S. B. (1967). Effects of cytochalasins on mammalian cells. Nature, Lond. 213,261-264.

CLONEY, R. A. (1972). Cytoplasmic filaments and morphogenesis: effects of cytochalasin Bon contractile epidermal cells. Z. Zellforsch. mikrosk. Anat. 132, 167-192.

COULOMBRE, A. J. & COULOMBRE, J. L. (1958). Intestinal development. I. Morphogenesis ofthe villi and musculature. J. Embryol. exp. Morph. 6, 403-411.

GREY, R. D. (1972). Morphogenesis of intestinal villi. I. Scanning electron microscopy ofthe duodenal epithelium of the developing chick embryo. / . Morph. 137, 193-214.

HAMBURGER, V. & HAMILTON, H. L. (1951). A series of normal stages in the developmentof the chick embryo. /. Morph. 88, 49-92.

HILTON, W. A. (1902). The morphology and development of intestinal folds and villi invertebrates. Am. J. Anat. 1, 459-504.

ISHIKAWA, H., BISCHOFF, R. & HOLTZER, H. (1969). Formation of arrowhead complexeswith heavy meromyosin in a variety of cell types. / . Cell Biol. 43, 312-328.

KARFUNKEL, F. (1971). The role of microtubules and microfilaments in neurulation inXenopus. Devi Biol. 25, 30-56.

KLETZIEN, R. F. & PERDUE, J. F. (1973). The inhibition of sugar transport in chick embryofibroblasts by cytochalasin B. /. biol. Chem. 248, 711-719.

POLLARD, T. D. & WEIHING, R. R. (1974). Actin and myosin and cell movement. CriticalRev. Biochem. 2, 1-65.

SCHROEDER, T. E. (1970). Neurulation in Xenopus laevis. An analysis and model based uponlight and electron microscopy. / . Embryol. exp. Morph. 23, 427-462.

SCHROEDER, T. E. (1973). Cell constriction: contractile role of microfilaments in divisionand development. Am. Zool. 13, 949-960.

SPOONER, B. S. (1973). Microfilaments, cell shape changes, and morphogenesis of salivaryepithelium. Am. Zool. 13, 1007-1022.

SPOONER, B. S. & WESSELLS, N. K. (1970). Effects of cytochalasin B upon microfilamentsinvolved in morphogenesis of salivary epithelium. Proc. natn. Acad. Sci., U.S.A. 66,360-364.

SPOONER, B. S. & WESSELLS, N. K. (1972). An analysis of salivary gland morphogenesis:role of cytoplasmic microfilaments and microtubules. Devi Biol. 27, 38-54.

740 D. R. BURGESS

WESSELLS, N. K. & EVANS, J. (1968). Ultrastructural studies of early morphogenesis andcytodifferentiation in the embryonic mammalian pancreas. Devi Biol. 17, 413-446.

WESSELLS, N. K., SPOONER, B. S., ASH, J. F., BRADLEY, M. O., LUDUENA, M. A., WRENN,J. T. & YAMADA, K. M. (1971). Microfilaments in cellular and developmental processes.Science, N.Y. 171, 135-143.

WRENN, J. T. (1971). An analysis of tubular gland morphogenesis in chick oviduct. DeviBiol. 26, 400-415.

WRENN, J. T. & WESSELLS, N. K. (1969). An ultrastructural study of lens invagination in themouse. / . exp. Zool. 171, 359-368.

WRENN, J. T. & WESSELLS, B. K. (1970). Cytochalasin B: effects upon microfilamentsinvolved in morphogenesis of estrogen-induced glands of oviduct. Proc. natn. Acad.Sci., U.S.A. 66, 904-908.

ZWANN, J. & HENDRIX, R. W. (1973). Changes in cell and organ shape during early develop-ment of the-ocular lens. Am. Zool. 13, 1039-1050.

(Received 29 April 1975, revised 4 July 1975)

morphogenesis of intestinal villi -...

Documents