INGRESSION OF DEEP CELLS OF THE MARGINAL REGION OF THE BLASTODERM DURING EARLY GASTRULATION OF FUNDULUS HETEROCLITUS
By J.P. Trinkaus, Department of Biology, Osborn Memorial Laboratories, P.O. Box 6666, New Haven, CT 06511-8155
It became apparent at the zebrafish meeting in Cold Spring Harbor last May that there is a certain amount of interest in my recent discovery that the gastrula of Fundulus heteroclitus does not undergo involution. Because this research has not yet been published or even presented formally, many of the questions raised received incomplete answers. With summer and the annual Fundulus season upon us, I will not be able to compose a formal paper (with proper illustrations) on this work for some time yet. In the meantime, it occurs to me that, because of the importance of the matter, it might be useful for a number of you to know about my results so far. Hence, this informal progress report.
The subject of involution during the gastrulation of teleosts has been the subject of controversy for many years. The modern attack was launched some sixty years ago by Oppenheimer (1936) and Pasteels (1936), using the best marking technique available at the time. Now, with the papers of Ballard (1966, 1973) on the trout and those of Thorogood and Wood (1987) and Wood and Timmermans (1988) on the rosy barb and Warga and Kimmel (1990) on the zebrafish, using better methods for following cells, it has seemed important to me to reinvestigate the problem in Fundulus. The combined clarity and larger size of the Fundulus egg (1.8 mm diam) and the availability of DIC optics make it particularly favorable material for direct observation of the movements of individual cells in vivo (see Trinkaus, 1973; Trinkaus and Erickson, 1983; Trinkaus, Trinkaus and Fink, 1992; Trinkaus, 1992; Trinkaus, 1993).
Methods
Direct minute by minute observation of the motile activities of a large number of individual deep cells with an inverted microscope, equipped with DIC optics, and recorded by time-lapse video, with imaging at appropriate intervals. My observations thus far have concentrated on the motile behavior of deep cells of the upper cell layer of the dorsal and lateral-ventral marginal regions of the blastoderm from late "blastula" (Armstrong and Child stage 12 1/2) through early epiboly (stages 13-14 3/4) to full formation of the germ ring at one third epiboly (stage 15 1/2). This is a period of over 4 hours at 22C. In each sequence, the motile fate of cells in the upper visible layer of deep cells was traced and recorded. The trajectories of all traceable cells were followed to avoid subjectivity. All of this tracking was of cells observed en face, as the blastoderm expands vegetally in epiboly. Unfortunately, individual cell movements are not discernable in profile views of the marginal region of the blastoderm of Fundulus (as they are in the rosy barb). Because of this, in part, I have no knowledge of the fate of cells once they sink beneath the upper deep cell layer. This will require cell marking àla Warga and Kimmel (1990). This is, of course, an important question. Until we have this information, we will not know for Fundulus in which directions cells move after ingression nor how much of the hypoblast is formed from upper level cells that have undergone ingression or to delamination of this rather thick region of the blastoderm (a question that is likewise still largely unanswered in the trout, rosy barb, and zebrafish).
The questions before me in Fundulus were: what is the pattern of movements of deep cells and where and when do they undergo ingression or internalization during early gastrulation, in relation to epiboly, formation of the germ ring and convergence? Quite frankly, what I expected to find was classical involution, as reported for certain teleosts, and thought that I would be able to observe certain aspects of it better and in more detail in the Fundulus blastoderm because of the outstanding optical properties of the egg. If so, this would lay a better basis for a later analysis of the mechanism of involution. What I found now follows.
Results
1) Just after the earliest detection of epiboly (stages 13-13 1/4) up to the first appearance of the germ ring (stage 14 3/4) and beyond until about one third epiboly (stages 15-15 1/2) deep cells at or near the blastoderm margin undergo ingression. Most of this occurs behind the margin of the blastoderm, away from the margin. Typically, cells at the margin recede from the margin and then undergo ingression, one to three cells away. Some marginal cells ingress directly at the margin, but only a minority. Of 34 cells initially at the margin of the dorsal region of the blastoderm only 7 underwent ingression there. Of 28 cells initially at the margin of the lateral-ventral region, only 5 underwent ingression at the margin. In another series at higher magnification, of 8 cells initially at the margin, only one underwent ingression there.
2) Cells that are near, but not at the margin, almost always ingress from their submarginal position. In a dorsal series, where the cells were initially 1-3 cells away from the margin, 3 cells moved to the margin to ingress there, whereas 24 cells either underwent ingression in place, moved away from the margin or toward but not to the margin before undergoing ingression. I have rarely observed movement of submarginal cells to the margin to ingress solely there. All submarginal cells move about a bit but do not move consistently toward or to the margin prior to their ingression. The ingression of cells farther away from the margin is highly variable. For example, in one series of 26 cells, anywhere from 3-6 cells distant from the margin only 1 cell moved to the margin to undergo ingression there. The rest either remained stationary (12 cells), moved toward the margin (8 cells) or actually moved farther away from the margin, as far as 8-9 cells away (5 cells). The only generalization that can be made about the ingressive behavior of cells of the marginal region of the Fundulus blastoderm from stages 13 1/4-15 1/2 is that relatively few cells remain at the margin or move to the margin of the blastoderm to undergo ingression there. Thus, one can conclude that in Fundulus there is no involution. Cells undergo ingression in the marginal region of the blastoderm, but mainly away from the margin. It is, of course, of interest that about 1/3 of the submarginal cells observed in these sequences moved toward the margin (but ingressed before they reached the margin). This may provide an explanation for observations based on random labeling of cells that have led many of us to believe that involution is a normal feature of teleost gastrulation.
3) Prior to the beginning of epiboly, Fundulus deep cells do not undergo ingression. I have not observed any cells of the marginal region of the blastoderm leave the surface and sink in during the blastula period (stage 12), whether at the margin or as far away as 6 cells distant from the margin. A few cells were observed to undergo ingression at stage 13-13 1/4, but most ingression occurs between stages 13 1/2 and 14 3/4 (144 of a total of 167 cells thus far traced). After stage 14 3/4, when a faint germ ring is first apparent, up to stage 15 1/2, only small numbers of cells undergo ingression. Ingression definitely tapers off as the germ ring forms and convergence toward the embryonic shield begins. In sum, ingression of deep cells of Fundulus takes place mainly during the period between the onset of the epiboly and the appearance of the germ ring. Perhaps the commencement of epiboly provides a stimulus. Incidentally, more ingression was observed at the very beginning of epiboly (stages 13-13 1/3) by submarginal cells than by those at the margin (12 cells versus 1 cell). This is consistent with the hypothesis that ingression of these deep cells is somehow favored by a submarginal location.
4) Ingression is most active in the dorsal marginal region, just lateral to the nubbin that represents the presumptive embryonic shield, and quite sporadic in the ventral marginal region. For example, in the dorsal region, all cells (74) at or within a few cells of the margin were found to undergo ingression. In contrast, in the lateral-ventral region, of a total of 51 cells traced from a position at or near the margin, 36 underwent ingression but 15 others did not. In addition, in the dorsal half of this sequence more cells underwent ingression (21 ingression; 5 no ingression) than into the ventral half (16 ingression; 9 no ingression). It appears, therefore, that there is a dorso-ventral gradient of ingression, suggesting that ingression in the marginal region of the blastoderm depends on environment factors, possibly stemming from the presumptive embryonic shield. Interestingly, this apparent gradient correlates with two well-established phenomena. First, in Fundulus, the germ ring is evident first dorsally and only later ventrally. Second, the directional movements of ventral germ ring cells during convergence are much less efficient than those of dorsal germ ring cells (Trinkaus and Trinkaus, unpub.).
5) When deep cells engage in ingression they do so by means of so-called blebbing movement (Trinkaus, 1973; Trinkaus and Erickson, 1983). This is the characteristic mode of deep cell movement during the late blastula stage and at the very beginning of gastrulation. Movement by means of filolamellipodia becomes dominant only when cells start converging toward the embryonic shield. This makes sense, because cells involved in blebbing movement are non-contact inhibiting and thus readily invade dense masses of cells, in contrast to filolamellipodial cells, which are contact inhibiting and hence non-invasive (Trinkaus, Trinkaus, and Fink, 1992). If deep cells were not non-contact inhibiting during this beginning phase of gastrulation, they would not undergo ingression, i.e., invade masses of other cells. Indeed, the diminution of ingression at stages 15-15 1/2 is correlated with a change to a dominant filolamellipodial, contact-inhibiting mode of cell movement in the now definitive germ ring and with convergence toward the embryonic shield.
Some Thoughts
There is apparent evidence for ingression, but not for involution, of cells of the upper deep cell layer in the marginal region of the Fundulus blastoderm during early gastrulation. Moreover, this ingression is associated with the subsequent appearance of the germ ring. If we define involution (sometimes called "wheeling in") as a) movement in the upper deep cell layer toward and to the blastoderm margin; b) sinking in at the margin; and then c) moving backward in the lower deep cell layer, we lack evidence for involution in Fundulus. In Webster's Unabridged Dictionary, involution is defined as "in botany, rolled inward at the edges as involute leaves"; "in anatomy, a part formed by rolling or curling inward". See also The Concise Oxford Dictionary. This is the way I (and I think most embryologists) have always thought of it. It has been likened to a caterpillar crawling to the edge of a table, crawling down over the edge and then crawling back on the underside of the table. Ingression is less precisely defined. It means "to go into", "to enter in", "the act of entering". My evidence is consistent with generalized cellular ingression in the marginal region of the Fundulus early gastrula blastoderm. Although I have not studied the fate of these cells after ingression, it seems certain that they join and contribute to the lower cell layer(s) of the marginal region of the blastoderm. Possibly they contribute to the hypoblast.
Now that the evidence indicates that involution at the blastoderm margin does not occur in Fundulus gastrulation, what name can we give to the fascinating process I have described? I do not yet have a name. Suffice it to say that the Fundulus embryo undergoes topographically and temporally localized ingression during the beginning phase of gastrulation. Also, how does this phenomenon in Fundulus relate to what has been reported in other teleosts? Possibly Fundulus is different. I personally doubt this, particularly since the recent work of Shih and Fraser on the zebrafish. But, on the other hand, there are important differences in gastrulation among the Amphibia (Dettlaff, 1993), a phylogenetic category whose gastrulation has been investigated far more than that of the teleosts. Clearly what is required is a similar continued examination of the details of gastrulation in the superb teleost material. It is only after this that meaningful comparisons can be made. And, most importantly, it is also only after this that we will be in a sound position to begin meaningful investigation of the mechanisms of the process in any given species.
Finally, hovering over all these varying results is the evolutionary question, and an accompanying general preference by many embryologists for simplicity. Do vertebrates gastrulate in homologous ways? It is becoming increasingly evident that in some ways they do not. This gives rise to another riveting evolutionary question. No matter how they gastrulate, all vertebrate gastrulae give rise directly to a remarkably similar larva, the so-called "pharyngula" (Ballard, 1981). Perhaps in the end it will be more important to understand how this has come to pass than to search for homologies in gastrulation that may not be there.
Return to Contents