Negative chemotaxis does not control quail neural crest cell dispersion

DEVELOPMENTAL

BIOLOGY

96, 542-551

(1983)

Negative Chemotaxis Does Not Control Quail Neural Crest Cell Dispersion

Negative chemotaxis has been proposed to direct dispersion of amphibian neural crest cells away from the neural tube (V. C. Twitty, 1949, Growth lS(Supp1. 9), 133-161). We have reexamined this hypothesis using quail neural crest and do not find evidence for it. When pigmented or freshly isolated neural crest cells are covered by glass shards to prevent diffusion of a “putative” chemotactic agent away from the cells and into the medium, we find a decrease in density of cells beneath the coverslip as did Twitty and Niu (1948, J Exp. Zool. 108,405-437). Unlike those investigators, however, we find the covered cells move slower than uncovered cells and that the decrease in density can be attributed to cessation of cell division and increased cell death in older cultures, rather than directed migration away from each other. In cell systems where negative chemotaxis has been demonstrated, a “no man’s land” forms between two confronted explants (Oldfield, 1963, E;cp. Cell Res. 30.125-138). No such cell-free space forms between confronted neural crest explants, even if the explants are closely covered to prevent diffusion of the negative chemotactic material. If crest cell aggregates are drawn into capillary tubes to allow accumulation of the putative material, the cells disperse farther, the wider the capillary tube bore. This is contrary to what would be expected if dispersion depended on accumulation of this material. Also, no difference in dispersion is noted between cells in the center of the tubes versus cells near the mouth of the tubes where the tube medium is freely exchanging with external fresh medium. Alternative hypotheses for directionality of crest migration in uiuo are discussed. INTRODUCTION

end were widely dispersed, while those near the open end did not migrate (Twitty, 1944). Further studies showed that single cells did not move extensively in these capillaries, but would migrate several cell lengths after two single cells contacted each other (Twitty and Niu, 1954). Cells also dispersed more in capillaries with narrow bores than in those with wide bores. Twitty and Niu proposed that a negative chemotactic agent was released by the neural crest cells and caused dispersion of the cells when it accumulated in sufficiently high concentrations, such as under a coverslip, at the sealed end of a capillary tube, or between two cells. We have repeated and extended the studies of Twitty and Niu using quail neural crest cells. While many of our observations are similar to those observed for amphibian neural crest, the control of quail crest cell dispersion in vitro does not appear to be due to negative chemotaxis.

A variety of mechanisms has been proposed, to account for neural crest cell dispersion away from the neural tube (for review see Abercrombie, 1965; Weston, 1982). Early studies by Twitty (1944, 1949) and Twitty and Niu (1948, 1954) suggested that amphibian neural crest cells produce a diffusible component which results in mutual repulsion of crest cells and subsequent migration away from each other. It was hypothesized that this “negative chemotaxis” drives neural crest cells away from their origin. The evidence presented by Twitty and Niu was based on two experimental designs. The first entailed placing a coverslip over part of an amphibian neural crest explant (Twitty and Niu, 1948). The cells underneath the coverslip appeared more dispersed (i.e., the cells were at lower density) after 24 hr, than those that were covered by medium only. These investigators also stated that the cells under the coverslips moved faster and migrated directionally away from each other although the data were never presented. In a second experiment, individual cells and aggregates of pigmented neural crest cells suspended in coelomic fluid were drawn into capillary tubes. One end was sealed, leaving the other end open to a drop of saline. Crest cells closest to the sealed I To whom all correspondence

METHODS

Neural

Neural crest cultures were generated as described previously (Loring et al., 1981). Briefly, neural tubes (NT) were dissected from 56-hr (22 somite) quail embryos (Coturnix coturnix japonica) with tungsten needles and digested with full strength Pancreatin (Gibco) at 37°C.

should be addressed. 542

0012.1606/83

$3.00

Copgriyht

cc 1983

hy Academic

All

of reprodurtion

rights

in any

Press. form

Inc reserved

Crest Cultures

543

BKIEF NOTES

Tubes were removed to cold Hanks’ saline (Gibco) as soon as the somites loosened. Any remaining somitic, notochordal, or epithelial tissue was dissected away with sharp tungsten needles. The tubes were then transferred to 35-mm tissue culture dishes (Corning) containing F12 medium supplemented with 10% fetal calf serum, 3% IO-day chick embryo extract, and 1% penicillin-streptomycin solution (Gibco: 500 units penicillin and 5 mg steptomycin sulfate). Within a few hours, neural crest (NC) cells migrate from the neural tube onto the plastic dish. In addition, neural crest cells form spherical clusters on the neural tube (see Loring et al., 1981). After 48-72 hr these spherical clusters were removed from neural tubes with tungsten needles, washed in Hanks’ saline, and transferred to a culture dish containing fresh medium. The clusters can attach to the dish and the cells disperse to form a circular explant (see Fig. la). Experimental

Pmcedures

Cells beneath coverslips. Small squares of coverslips were positioned over or near cluster outgrowths or 18hr outgrowths from neural tube explants and fixed in place with small amounts of Vaseline or vacuum grease (Dow Corning) at each corner. Some of the coverslips rested on grease alone, while others sat on small feet made of shards of No. 1 coverglass. The height of the coverslips above the cells was measured by using the calibration on the fine focus knob of a Wild inverted microscope. We varied the coverslip height from 10 to 150 pm above the substratum. In a few cases, the coverslip actually touched the explant, since some neural crest cells migrated onto the top surface of the coverslip. Spreuding in capillary tubes. Neural crest cell dispersion in a confined space was tested by drawing freshly isolated clusters of identical size into capillary tubes of varying diameter (Microcaps, 250-1000 pm diameter). The tubes were filled with F12 medium and one end sealed with vacuum grease (Dow Corning). The tubes were placed in 60-mm petri dishes filled with medium, Time-Lapse Neural crest cell behavior and speed of movement were analyzed using time-lapse video recording. Cells were observed, using a Zeiss invertoscope D, and behavior was recorded, using an RCA No. TC 1005/VOl camera, a Panasonic time-lapse video tape recorder, and played back on an Electrohome monitor. Speed of movement was determined by tracing nuclear displacement at lo-min intervals. Displacement rates of crest cells beneath the coverslip and outside the coverslip were statistically compared using Stu-

dent’s t test and Wilcoxon and unequal variances.

ranking

assuming

unknown

RESULTS

Cells under Coverslips Neural crest cell behavior under coverslips was examined, using both pigmented crest (obtained 4 days after neural tube explanation) and 18-hr crest cells from freshly isolated neural tube explants. Using either cell type, cell density was reduced in the portion of the culture beneath the coverslip. The aggregates isolated from the neural tube are virtually 100% neural crest, since all these cells produce pigment (see Figs. la, b and Loring et al., (1981)). After the aggregates have been removed from the neural tube and replated, the cells are highly pigmented and have migrated from the cluster to form a circular explant (Figs. la, b). When coverslips are placed over half of such a circular explant and the cultures reexamined 48 hr later, several changes are observed. First, the cells have continued to disperse radially from the center of the explant (Figs. lc, d). Second, the density of cells beneath the coverslip is obviously decreased compared with the other half of the culture. These observations are similar to those of Twitty and Niu (1948). It should be noted, however, that the perimeter of the pigmented explants is still circular and that the cells beneath the coverslip have not dispersed any further from the center of the explant than from the uncovered side. Video tapes show that by 48 hr after the coverslip is applied, pigmented crest cells are virtually immobile and many cells beneath the coverslip are losing their pigment or rounding up and dying. Cell debris, pigment granules, and dead cells have collected just beneath the coverslip (Figs. le, f). When a coverslip is applied over portions of 18-hr outgrowths from neural tubes, or positioned at the edge of a 5-hr outgrowth so that the crest cells eventually migrate beneath it, a similar reduction in cell density is seen beneath the coverslip 24 hr later (Fig. 2). Again, the margin of the outgrowth is circular so that NC cells beneath the coverslip appear to be dispersing radially at the same rate as cells covered by the medium only. Higher magnification reveals few mitotic cells beneath the coverslip, whereas the rest of the outgrowth has a large number of mitotic cells (Fig. 2~). Video tapes also document the cessation of mitotis beneath the coverslip. Large numbers of cells have also died by this time (Fig. 2d). Eighteen-hour cell cultures, taped after the coverslips were positioned and analyzed for speed of movement, show that the speed of movement is statistically slightly faster for cells outside the coverslip than for cells beneath it (Table 1).

544

DEVELOPMENTALBIOLOGY

VOLUME96.1983

b

la

d Confrontation

of Aggregates

If neural crest cells are producing a negative chemotactic factor, then it would be expected that a “no man’s land” would form between two confronted explants where the factor might build to a high concentration (see Carrel1 and Ebeling, 1922; Oldfield, 1963).

fresh plastic culture dish, the cells migrate radially away from the center of the cell clusters and invariably merge with each other in the space between the explants (Figs. 3a, b). Intermingling of the cells between the explants still occurs even if a coverslip is used to trap the pu-

RKIW

NOTES

f FIG;. 1. (a) A 3-day-old aggregate which was removed from a neural tube and replated in fresh medium. Cells were allowed to adhere and migrate for 48 hr. (h) The same aggregate from a after 3 more days in culture. The cells continue to divide and migrate radially from the center of the explant. (c) A 4-day-old aggregate which was removed from the neural tube, plated in fresh medium and photographed 3 days later. Appearance of the same aggregate shown in Id just prior to addition of a coverglass. (d) The cluster outgrowth in c, 48 hr after the coverslip was applied. Crest cells have migrated equally in all directions, hut the cells are at lower density underneath the coverglass (CGJ. The edges of the coverglass are marked by two arrows. (e) This 4-day-old aggregate was replated for 48 hr, a coverslip was applied, and photographed after an additional 48 hr. The plane of focus is above the substratum and shows an accumulation of small pigment granules collected underneath the coverglass (CG). The cells have dispersed equally in all directions hut are at a lower density heneath the coverglass. (f) A -I-day-old aggregate treated as in e but photographed 24 hr after positioning of the coverglass. The density of cells is lower beneath the coverslil,, and some dead cells and pigment granules are beginning to collect. Scale bar = 100 pm.

tative chemotactic agent in the confrontation zone (Fig. 3~). Under these latter conditions the cells are of lower density beneath the coverslips than in the uncovered portions of the explants, although they are evenly dispersed. The explants still retain their circular profile even when they meet underneath the coverslip. Neutral Crest Cells in Capillary

Tubes

Two-day unpigmented aggregates or 4- to 6-day pigmented aggregates of identical size were drawn by capillarity into fine glass tubing 250 to 1000 pm in diameter. In some experiments, one end of the tube was plugged with silicone vacuum grease. Visual observation reveals no consistent difference in dispersion from aggregates that are close to the plugged end versus aggregates near the open end of the same capillary tube (Fig. 4a). Contrary to observations by Twitty (1944), the cells are more dispersed in the wider capillary tubing than in the narrower one (Figs. 4b-e).

DISCUSSION

Several important pieces of evidence from our experiments suggest that quail NC cells do not disperse under these conditions as a result of negative chemotaxis. (1) The perimeters of the circular explants or cluster outgrowths remain circular after moving beneath a coverslip. If the covered cells dispersed faster or were more persistent in their direction of movement, as Twitty and Niu (1948) suggested, they should have migrated farther from the center of the explant than cells on the uncovered side of the explant. This is especially true when two confronted explants are covered, but there too, the profile of both the confronted aggregates remains circular (2). The cells underneath the coverslips actually move slower, not faster, than those that are uncovered (see Table 1). (3) Cells disperse equally well whether they are in the plugged end of a capillary, where chemotactic material should accumulate, or at the open end, where it should disperse. Furthermore, the cells disperse more in wider capillaries than in nar-

546

DEVELOPMENTAL BIOLOGY

VOLUME 96, 1983

FIG. 2. (a, b) NT shown in this figure were cultured for 18 hr, and then a coverglass (edges marked by arrowheads) was applied over a part of the neural crest explant. The cells were videotaped for another 18 hr and the cultures were fixed and photographed. Cells underneath the coverslip are at lower density but have not dispersed faster than the uncovered cells, since the perimeter of the whole explant remains circular. Note that many cells beneath the coverslip at the perimeter of explant in a have begun to round up and retract their processes. These rounded cells can be distinguished from flattened ones since they are more refractile. The density of cells is higher near the edges of the coverglass than in the center. Time lapse tapes showed these cells had pushed under the coverglass from the surrounding high density culture, contrary to the observation of Twitty and Niu (1948) that crest cells backed out from under the coverslip. (c) High magnification of an area from b. A piece of substratum debris is marked by a curved arrow in both micrographs. The edge of the coverglass is marked by arrowheads. There are many more rounded cells outside than under the coverslip, which time-lapse shows to be mitotic cells, (d) High magnification of NC cells beneath a coverslip in cultures similar to those in a and b. After only 18 h many dead cells appear (arrows). These can be identified since they are swollen and no longer attached to the substratum. Scale bar = 100 pm.

rower ones. These results are just the opposite from what one would expect, if the cells produced a negative chemotactic material.

The reduced cell density underneath the coverslip appears to be related in large part to cell health. In partitular, the newly migrating NC cells stopped dividing

as evidenced by time-lapse observations, and both 24hr NC and older pigmented cells appeared to be dying beneath the glass coverslips. In older cultures (Figs. le, f), cells outside the coverslip were well spread and not losing pigment while those beneath the coverslip were rounded, far less pigmented with pigment granules budding off their upper surfaces. These results could be explained by the following: (1) There may have been a pH change beneath the coverslip due to accumulating cell metabolites which could not diffuse away. (2) There might have been a change in oxygen tension underneath

the coverslip (see Moscona et ul., 1965). Quail NC cells seem particularly sensitive to narrow spaces in tissue culture, compared with other fibroblasts (Newgreen et al., 1979; Erickson and Nuccitelli, unpublished data). This is particularly puzzling in light of the original report by Perry et al. (195’7) that small numbers of cells need to condition their medium in order to spread and divide, and can do this best in narrow capillaries. Regardless, the decrease in cell density of quail NC aells is not due to rapid cell dispersal, but probably some unidentified perturbation of cell division or cell health.

548

DEVELOPMENTALBIOLOGY TABLE1 SPEEDOFMOVEMENTOFNEURALCRESTCELLS' Rate of translocation* (pm/min)

Cells outside coverslip Cells beneath cover-slip

0.82 -+ 0.32 0.64 f 0.21

N 36 30

Note. The difference between the means was significant using both Student’s t test (t = 2.651; P < 0.0102) and Wilcoxon rank test (P < 0.0033). Both tests assumed unknown and unequal variances. “Neural tubes were allowed to attach to a plastic substratum for 5 hr. A coverslip was positioned near the explant so that neural crest cells could migrate beneath it. Filming of cultures was begun 18 hr after neural tube explantation. Both covered and uncovered neural crest cells were filmed simultaneously. *Cell movement was traced on the video screen by recording the position of the cell nucleus every 10 min. Observation time for each cell was at least 180 min.

It seems unlikely that poor cell health would mask the effects of a negative chemotactic agent, since migratory cells beneath and outside the coverslips displayed the same mode and similar rates of locomotion. Cell aggregates of identical size disperse to the same extent no matter their position in a capillary tube. Furthermore, the cells disperse farther in wide bore capillaries than in narrow ones. One possible explanation is that the narrow bore represents an angle of curvature that is too steep to allow migration around the circumference, so that each cell can only migrate along the length of the tube (Dunn and Heath, 1976; Fisher and Tickle, 1981; Dunn, 1982). The wide-bore pipets with their wider angle of curvature, might allow migration in all directions rather than only along the long axis; the cells could therefore disperse more rapidly. How, then, can we account for the results of Twitty and Niu? It could be that amphibian neural crest cells respond to different cues than do fowl or mammalian neural crest. Unpublished works, using Xenopus and Triturus, suggest that these species do not show negative chemotaxis either (R. Keller, personal communication). Alternatively, the observations of Twitty and Niu may be explained by cell death, as we have observed for quail crest, although they do claim to see an increase of speed of movement beneath the coverslip. In any case, it appears that we must seek another explanation for fowl neural crest dispersion away from the neural tube in viva.

Several mechanisms have been suggested for directional migration of the neural crest cells away from the neural tube in the embryo (for a recent review, see Weston (1982)). While these are still subject to speculation, the following seem most probable, or at least have not been ruled out, based on some experimental evidence.

VOLUME96. 1983

(1) Contact inhibition (Abercrombie and Heaysman, 1954; Abercrombie, 1965) is the mechanism by which cell movement is inhibited in the direction of contact with another cell. As a result, cells will move radially away from regions of high cell density to low cell density, and they will form parallel arrays (Abercrombie and Heaysman, 1954; Erickson, 1978). Neural crest cells display contact inhibition (Newgreen et al., 1979) and contact paralysis in vitro on a plane substratum (Gooday and Thorogood, personal communication; Erickson, unpublished results) although one report comparing speed of movement with number of cell contacts suggests in a three-dimensional collagen gel, that they do not (Davis and Trinkaus, 1981). In addition, SEM shows that neural crest cells are aligned parallel with their direction of movement (Bancroft and Bellairs, 1976; Tosney, 1978; Lofberg et al., 1980; Erickson and Weston, submitted for publication). However, contact inhibition has not been directly demonstrated in the embryo. (2) Persistence of directional migration could be accounted for by the phenomenon of “nudging” first noted in Fund&us deep cells (Tickle and Trinkaus, 1976; Trinkaus, 1978), in which touching the side of a cell induces bleb formation on the opposite side. Crest migration away from the neural tube could thus be due to the cells closest to the tube contacting the cells in front of them thereby inducing lamellipodium formation away from the neural tube. Interestingly, Twitty and Niu (1954) noted that single cells in a capillary do not migrate extensively, but, when one cell contacts another, they will disperse in opposite directions over several cell lengths. The “cell nudging” phenomenon could explain this observation. (3) Alternatively, the cells could be structurally polarized (Albrect-Buehler, 1977; Holmes and Trelstad, 1977; Gotlieb et al., 1981) and have an endogenous intuition of direction. This seems unlikely, since, during development, the cells are initially randomly arranged on top of the tube and then become aligned as they continue their lateral migration (Tosney, 1978; Bancroft and Bellairs, 1976). (4) Neural crest cells could follow a chemotactic or adhesive gradient (haptotaxis) (Carter, 1967; Harris, 1973). While there is substantial evidence suggesting that fibronectin is the migratory substratum preferred by the neural crest in the embyro (Newgreen and Thiery 1980; Newgreen et al., 1982; Erickson and Turley, 1983), it has not been demonstrated to exist as a gradient, although our techniques, at present, may not be sufficient to resolve such a gradient. Furthermore, when neural crest cells are grafted into the ventral NC pathway, the cells are capable of migrating dorsally (i.e., opposite from their usual direction), suggesting that

BRIEF NOTES

FIG. 3. (a) Three-day-old aggregates were removed from the neural tube, positioned opposite each other, and grown for an additional 4 days in culture. The cells disperse equally in all directions. (b) The same confronted aggregates as in Fig. Za, 48 hr later. The margins of outgrowth have met, and cells intermingle in the confrontation zone between the two explants. (c) Two 4-day-old aggregates were removed from the neural tube, positioned opposite each other, and grown for 3 days. A coverslip was then placed over the confrontation zone, well before the outgrowths met (edges of the coverslip are marked by arrowheads) and photographed 2 days later. The density of cells is decreased beneath the coverslip, as seen in Figs. le-g; but even under these conditions, the outgrowths meet and cells mingle at a uniform density. Notice also that the circular margins of the outgrowths are retained heneath the coverglass. Scale bar = 100 pm.

FIG. 4. (a) Four neural crest aggregates in a 550-pm-diameter capillary tube where the left end (marked 0) is plugged with vacuum grease and the right end (marked 3.03 mm) is open and freely exchanging with fresh medium. The degree of cell dispersion is variable, but there does not appear to be any correlation with position in the tube. (b, c) Two neural crest aggregates of equal size which have spread for 48 hr in a 250-*m-bore capillary tube, The cells are densely packed. Note that the cells are quite flat and have short cell processes. (d, e) Two neural crest aggregates originally equal in size to those in b and c which have attached and spread for 48 hr in a 550.pm-bore capillary tube. The cells have dispersed much further, are less dense, and many of the cells have long, thin cell processes. The number of cells in aggregates b-e, respectively, when photographed were 70, 86, 102, 98. The number of cells in b and c was probably underestimated, since the density was so high. The original number of cells in each was nearly identical as calculated from aggregate diameter (see Loring et al. 1981). (a) Scale bar is in mm. (b-e) Scale bar = 100 pm. 550

BRIEF

such a gradient is not present during early stages of migration (Erickson et al., 1980). (5) Neural crest, as well as other embryonic migratory cells, may be guided by electrical fields (Trinkaus, 1982; Jaffe, 1982). Indeed, neural crest cells will migrate to the cathode end of an imposed electrical field in vitro (Nuccitelli and Erickson, submitted; Keller, personal communication). Direction of neuron outgrowth (Jaffe and Poo, 1979; Pate1 and Poo, 1982) and neutrophil migration (Orida, 1982) can also be controlled by imposing an electrical field. While currents are known to leave the primitive streak in the early chick embryo (Jaffe and Stern, 19’79), and large currents are present around the neural tube at the onset of crest migration (Erickson and Nuccitelli; unpublished results), it is not known if these currents can direct neural crest cell migration. We would like to thank J. P. Trinkaus and R. D. Grey for critical reading of the manuscript. This research was supported by NSF Grant PCM-8004524 and NIH Grant PHS DE0 5630-01 awarded to Carol A. Erickson.

REFERENCES ABERCROMBIE, M. (1965). The locomotory behavior of cells. 11~“Cells and Tissues in Culture” (E. Willmer, ed.), pp. 177-202. Academic Press, New York. AUERCROMBIE, M., and HEAYSMAN, J. E. M. (1954). Observations on the social behavior of cells in tissue culture. II. “Monolayering” of fibroblasts. Exp. Cdl Res. 6, 293-306. AI.BRE~HT-BUEHI,ER, G. (19’77). Phagokinetic tracks of 3T3 cells: parallels between the orientation of track segments and of cellular structures which contain actin or tubulin. Cell 12, 333-339. BAN(‘RO~T, M., and BFXLAIRS, R. (1976). The neural crest cells of the trunk region of the chick embryo studied by SEM and TEM. Zoom 4, 73-85. CARRF:LL, A., and ERE:LING, A. H. (1922). Pure cultures of large mononuclear leucocytes. J. ILrp Afed. 36, 365-386. CARTER, S. (1967). Haptotaxis and the mechanisms of cell motility. Nature lLnndon) 213, 256-260. DAVIS, E. M., and TRINKAIJS, J. P. (1981). Significance

of cell-to-cell contacts for the directional movement of neural crest cells within a hydrated collagen lattice. .I Ewhr~ol. Exp. Mwrphol. 63, 29-51. Dt:NN, G. A. (1982). Contact guidance of cultured tissue cells: A survey of potentially relevant properties of the substratum. In “Cell Behavior” (R. Bellairs, A. Curtis, and G. Dunn, eds.), pp. 247-280. Cambridge liniv. Press, Cambridge. DIIXX, G. A., and HEATII, J. P. (1976). A new hypothesis of contact guidance in tissue cells. Exp. Cell Rrs. 101, 1-14. ERICKSON, C. A. (19%). Analysis of the formation of parallel arrays by BHK cells i7~ r)itro. &p (b/l Res. 115, 303-315. ERIVKSON, C. A., TOSNEY, K. W., and WESTON, J. A. (1980). Analysis of migratory hehavior of neural crest and fibroblastic cells in embryonic tissues. I1c7,. Hiol. 77, 142-156. EKI(‘KSO~‘, C. A., and TIJRLEY, E. A. (1983). Substrata formed by combinations of extracellular matrix components alter neural crest cell motility in ktro. J. Cdl Sci. In press. FISHER, P. E., and TICKLE, C. (1981). Differences in alignment of normal and transformed cells on glass fibers. Exp. Cell Rex 131, 407410.

551

NOTES

GOTLIEB, A. I., MC MAY, L., SUBRAHMANYAN, L., and KALNINS. V. 1. (1981). Distribution of microtubule organizing centers in migrating sheets of endothelial cells. J. Cell Biol 91, 589-594. HARRIS, A. K. (1973). Behaviour of cultured cells on substrata of variable adhesiveness. Exp. Cell Res. 77, 285-297. HOLMES, L. B., and TRELSTAD, R. L. (1977). Patterns of cell polarity in the developing mouse limb. Dev. Biol. 59, 164-173. JAFFE, L. F. (1982). Developmental currents, voltages, and gradients. In “Developmental Order: Its Origin and Regulation” (S. Subtelny and P. B. Green, eds.), pp. 183-215. Alan R. Liss, New York. JAFFE, L. F., and Poo, M.-M. (1979). Neurites grow faster towards the cathode than the anode in a steady field. J. Exp. Zool. 209.115-128. JAFFE, L. F., and STERN, C. D. (1979). Strong electrical currents leave the primitive streak of chick embryos. Science 206, 569-571. LOFBERG, J., AHLORS, K., and FALLSTROM, C. (1980). Neural crest cell migration in relation to extracellular matrix organization in the embryonic axolotyl trunk. Dev. Biol. 75, 148-167. LORI%, J., GLIMELIIIS, B., ERICKSON, C., and WESTOP~,J. A. (1981). Analysis of developmentally homogenous neural crest cell populations i,l uif~o. Dec. Bid. 82, 86-94. MOXONA, A., TROWELL, 0. A., and WILLMER, E. N. (1965). In “Cells and Tissues in Culture” (E. Willmer, ed.), Chap. 2. Academic Press, New York. NEWGREEN, D. F., GIBBINS, I. L., SAUTER, J., WALLENFELS, B., and WUTZ, R. (1982). Illtrastructural and tissue culture studies on the rate of fibronectin, collagen and glycosaminoglycans in the migration of neural crest cells in the fowl embryo. Cell Tissue Res. 221, 521-549. NEWGREEN, D. F., RITTERMAN, M., and PETERS, E. A. (1979). Morphology and behaviour of neural crest cells of chick embryo in vitro. Cell Tissue Res. 203, 125-138. OLDFIELD, F. E. (1963). Orientation behavior of chick leucocytes in tissue culture and their interactions with fibroblasts. Exp. Cell Res. 30, 125-138.

ORIDA, N. (1982). Directional protrusive pseudopodial activity and motility in macrophages induced by extracellular electric fields. Cell Motil.

2, 243.

PATEL. N. and Poe, M.-M. (1982). Orientation of neurite growth by extracellular electric fields. J. Neurosci. 2, 483-496. PERRY, V. P., SANFORD, K. K., EVANS, V. J., HYATT, G. W., and EARLE, W. R. (1957). Establishment of clones of epithelial cells from human skin. J. N&. Cnllcer Iwst. 18, 709. TICKLE, C., and TRINKAUS, J. P. (1976). Observations on nudging cells in culture. Nuture (Lwdm~ 261, 413-414. TOSNEY, K. T. (1978). The early migration of neural crest cells in the trunk region of the avian embryo: an SEM-TEM study. De?: Biol. 62, 317-333.

TRINKAUS, J. P. (1978). Mediation of cell surface behavior by intercellular contact. Zoor/ 6, 57-63. TRINKAIJS, J. P. (1982). Some thoughts on directional cell movement during morphogenesis. A “Cell Behaviour” (R. Bellairs, A. Curtis, and G. Dunn, eds.), pp. 471-498. Cambridge Univ. Press, Cambridge. TWITTY, V. C. (1944). Chromatophore migration as a response to mutual influences of the developing pigment cells. J. Exp. 2001. 95,259290.

T~ITTY, V. C. (1949). Developmental analysis of amphibian pigmenlf(Supp1. 9), 133-161. tation. Grodh TU’ITTY, V. C., and Nrrr, M. C. (1948). Causal analysis of chromatophore migration. J. Exp. Zool. 108, 405-437. TWITTY, V. C., and NI~J, M. C. (1954). The motivation of cell migration, studied by isolation of embryonic pigment cells singly and in small groups in vitro. J Exp Zool. 125, 541-573. WESTON, J. A. (1982). Motile and social behavior of neural crest cells. 111“Cell Behaviour” (R. Bellairs, A. Curtis, and G. Dunn, eds.), pp. 429-470. Cambridge Univ. Press, Cambridge.