Leg bud mesoderm retains morphogenetic potential to express limb-like characteristics (“limbness”) in collagen gel culture

DEVELOPMENTAL DYNAMICS 193314-324 (1992)

Leg Bud Mesoderm Retains Morphogenetic Potential to Express Limb-Like Characteristics (“Limbness”)in Collagen Gel Culture KEITARO ISOKAWA, EDWARD L. KRUG, JOHN F. FALLON, AND ROGER R. MARKWALD Department of Cellular Biology and Anatomy, Medical College of Wisconsin, Milwaukee, Wisconsin 53226 (KJ., E.L.K., R.R.M.); Department of Anatomy, University of Wisconsin School of Medicine, Madison, Wisconsin 53706 (JPP.)

ABSTRACT Recent in situ hybridization studies have correlated expression of potential regulatory genes with pattern formation in limb bud mesoderm (Tabin: Cell 66:199-217,1991);however, the mechanism(s) controlling their expression in mesoderm and their relevance to the establishment of a limb morphogenetic pattern remain unknown. One likely candidate for regulating patterning events in limb mesoderm is the apical ectodermal ridge, as its removal in ovo results in a graded truncation of limb skeletal elements in the proximal-distal axis dependent upon the time of excision (Rowe and Fallon: J Embryo1 Exp Morph 68:l-7, 1982). In the present study, we investigate whether the hypothetical imprint of ridge ectoderm is retained in cultured mesoderm. Specifically, we sought to determine if a subpopulation of limb mesoderm that forms in collagen gel culture (Markwald et al: Anat Rec 226:91-107, 1990), retains any expression of “limbness” in the absence of limb ectoderm as characterized by the formation of a predictable number and distribution of limb-like chondrogenic elements in comparison to the temporal and spatial relationships of the in situ proximal, hindlimb skeletal structures. Accordingly, explants of undissociated mesoderm from stage 18-22 chicken leg buds were cultured without ectoderm on collagen gel lattices and the central subpopulation of mesoderm was examined morphologically. We show that this central subset of mesoderm will form chondrogenic cells which were not expressed uniformly throughout the subset, but rather distinct nodules or elements of cartilage were elaborated. Moreover, the number of elements expressed by the central subset increased with the age of the mesoderm at the time of explantation; spatially and temporally, the sequence of elements that formed always proceeded from the proximal, anterior margin of the subset to its distal, posterior border. The shapes of the initial elements (designated I and 11) resembled the forms of in situ proximal skeletal structures (girdle and femur-like),whereas more distal elements (111-V) were often fused and without structural similarity to in situ skeletal structures. When cultures were established from the posterior mesoderm of stage 19/20 or 21 mesoblasts, the 0 1992 WILEY-LISS, INC.

frequency of element I formation was reduced approximately one-half, whereas formation of more distal elements was unaffected. Conversely, element formation from the central subset established from isolated anterior mesoderm was virtually identical to intact mesoblasts, indicating a capacity to regulate for the loss of mesoderm as occurs in situ (Hampe: Archs Anat Microsc MorDh Exp 48:345-378, 1959). We interpret these findings to mean that limb mesoderm which forms the central subset is not merely competent to express in the absence of continuous ectodermal co-culture, a chondrogenic phenotype, but also retains intrinsic, Zimblike morphogenetic potential to form, regulate, and, to a variable extent, shape a reproducibly specific number of cartilage elements. Because the number of chondrogenic elements that formed in culture appeared stage-specific, a role for the apical ridge in specifying chondrogenic elements prior to its removal is supported. Indeed, our temporal findings correlate closely with the predictable truncation of leg skeletal structures following removal of the apical ridge in ovo, indicating that the endogenous ectodermal influence upon limb mesodermal morphogenesis can be assayed in vitro without the requirement of their continuous co-culture. 0 1992 Wiley-Liss, Inc. Key words: Limb bud, Differentiation, Pattern formation, Cartilage, Apical ectoderma1 ridge INTRODUCTION In the early phases of chicken leg bud development, a bulge of seemingly, homogeneous mesodermal cells are formed at specific axial levels. An ectodermal jacket surrounds the bulge and eventually differentiates apically into a specialized ridge or cap of cells

Received January 30, 1992; accepted March 6, 1992. Address reprint requestskorrespondence to Dr. Roger R. Markwald, Department of Cellular Biology and Anatomy, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226. Keitaro Isokawa’s current address is DeDartment of Anatomv. Nihon University School of Dentistry, 1-8-15Kanda-Surugadai, Zhiyoda-Ku, Tokyo 101 Japan.

LIMB MORPHOGENESIS IN CULTURE

termed the apical ectodermal ridge. How the bulge initially forms and how it is subsequently transformed morphogenetically into a specific pattern of chondrogenic skeletal structures remain two of the most challenging problems of limb development. While the initial formation of the bud appears independent of ridge ectoderm (Carrington and Fallon, 19881, formation of the ridge appears to be required for the mesoderm to express chondrogenic elements in a specific proximal to distal sequence. Removing the ridge ectoderm truncated the normal progression of chondrogenic elements along the proximal-distal axis (Rowe and Fallon, 1982). Similarly, Kosher et al. (1979) found that if intact limb buds were placed into organ culture, the organ rudiments formed polarized, “patterned outgrowths of chondrogenic structures, whereas, if the apical ridge ectoderm were removed, only rounded, unorganized cartilaginous masses developed. Recently, the expression of potential regulatory genes in mesodermal cells has been described and correlated with the presence or absence of a functional ridge ectoderm (Tabin, 1991; Coelho et al., 1991; Robert et al., 1991). Thus, the ectoderm would appear to be a logical source of morphogenetic signals for regulating pattern formation in the mesodermal bulge. However, to test any “ectodermal hypothesis” related to mesodermal pattern formation, it would appear necessary to determine if a morphogenetic imprint of previous ectoderma1 ridge interaction in situ was retained and expressible by mesodermal cells once placed into culture. In particular, the identification and function of specific ectodermal signals or the significance of regulatory gene expression in mesoderm would seemingly require knowing how much inherent memory of limb-like characteristics was retained by the mesoderm in the absence of continuous ectodermal influence. One complexity in studying ectodermal-mesodermal interactions in limb development is that in ovo, presumptive chondrogenic areas are not recognized until stages 23-24, when some, but not all, of the mesodermal cells form cellular condensations which presage their expression of type I1 collagen a t stage 25 (Linsenmayer et al., 1973; Thorogood and Hinchliffe, 1975; Dessau et al., 1980). This suggests that only subpopulations of limb mesoderm are competent, to form prechondrogenic structures. The existence of prechondrogenic subpopulations of cells is supported by marking experiments using peanut agglutinin (Aulthouse and Solursh, 1987), type I1 collagen cDNA (Swalla et al., 1988; Stirpe and Goetinck, 19891, and an antibody against a mesoderm proteoglycan termed PG-M (Shinomura et al., 1990) and by in vitro experiments, including micromass culture (Ahrens et al., 1977, 1979; Solursh et al., 19811, separation of progenitor cells (Sasse et al., 1984; Kosher and Rodgers, 1987) and cell cloning (Rutz et al., 1982). Thus, the isolation or segregation of such presumptive subpopulations should seemingly be considered in attempting to specifically

315

characterize or assess any morphogenetic imprints on the mesoderm arising from the ectoderm or elsewhere. The purpose of the present study was to evaluate the morphogenetic performance of a prechondrogenic subpopulation of early limb mesoderm (stages 18-22) to express characteristics of “limbness” in 3-d collagen gel culture. We define “limbness” according to Zwilling (1968) as not merely the potential of limb mesoderm to express a chondrogenic phenotype but by its predisposition to form a proximal to distal sequence of differentiated chondrogenic elements whose number changes with age in a manner characteristic of in situ limb mesoderm. Because homotypic and heterotypic cell interactions seem to be essential for chondrogenesis in the limb (Solursh and Reiter, 1980; Archer et al., 19851, we considered the maximal retention of naturally occurring cell-to-cell associations as a critical component of the culture approach. Indeed, as Frederick and Fallon (1982) have shown, recombinant limbs may lose their limb-like characteristics when dissociated anterior and posterior mesodermal cells are intermixed prior to grafting. Therefore, we have utilized undissociated limb mesoderm, as opposed to micromass-type cultures, to maintain the relative positions of mesodermal cells and to allow for interactions among the cells. We have shown previously that when the spatial organization of limb mesoderm is retained, the cells will segregate in a collagen gel culture system into 3 subpopulations of mesenchyme (Markwald et al., 1990). Here, we now show that the central subpopulation that forms from the intact mesodermal bulge of hindlimbs explanted between stages 18 and 22 will give rise to cells which form cartilaginous elements whose spatial and temporal sequence of expression show apparent similarity to the in situ development of proximal leg skeletal structures. These data indicate that intact mesoderm retains its inherent memory of limbness without the requirement for ectodermal co-culture or factors extrinsic to the limb. Rather its ability to express a characteristic limb pattern depends upon the length of time in association with the ectoderm in vivo (i.e., before its removal and explantation into culture).

RESULTS Undissociated limb mesodermal explants, or “mesoblasts”, were prepared (for detail, see Fig. 1and Experimental Procedures). These intact mesoblasts cultured on a lattice of hydrated collagen sort out into three distinct populations (i.e., surface, central, and seeded subsets) (Markwald et al., 1990). In the present report, results pertinent t o the chondrogenic central subset will be described. Temporal Changes in the Potential of Central Subsets to Form Cartilage Elements in Culture Central subsets established from “whole mesoblasts” (see Fig. 1)of stage 18-22 leg buds laid down a series of cartilage elements sequentially along the proximo-

316

ISOKAWA ET AL.

(apnial)

4

P

Fig. 1. Dissection criteria. b: baseline of limb. d: cutting line for mesoblasts dissected out deeply. P: the most laterally projected point in the distal tip of limb. M: the point dividing the baseline into two equal segments, anterior and posterior halves, 3.5-somites in length. This point is located at the 30th somite or just between 30th and 31st somite, depending on variations among individual embryos. W: the point located 1.5 somites anterior to M when estimated on the baseline. Shaded area indicates wedge-shaped tissue discarded in preparation of mesoblast explants without antero-proximal mesoderm. Hatched area shows distal tip mesoblast explants.

distal axis according to the age of the mesoblast at the time of explantation. We designated these elements as I, 11,111,IV, and V in successive order from proximal to distal. Using this designation, the temporal sequence of elements formed from stage 18-22 mesoblasts after 6 days in culture is given in Figure 2 and Table 1. It is apparent that as the stage of the leg bud from which the mesoblast was removed increased, the potential to form additional more distally positioned, cartilage elements also increased. The most proximal cartilage element (element I), was the first to form. The frequency of its formation increased between stages 18/19- to 22 (Table 1). This does not necessarily mean that the central subsets established by younger mesoblasts have less potential t o form element I, as it is possible that in the younger limb buds, the mesodermal cells responsible for element I formation are located proximal to the baseline cut (line “b” in Fig. 1)performed to remove these limb buds. This contention was supported by results of deeply dissected stage 18/19- mesoblast cultures (line “d” in Fig. 11,in which 92% (11cases out of 12) showed element I formation (vs. 47% for baseline cut mesoblasts; Table 1).Based on this observation, the variation in size and shape of element I demonstrated in Figure 2 could be interpreted as a result of minute differences in the location of the baseline cut. It is interesting to note that the largest pieces of element I formed an element with morphological characteristics suggestive of a girdle-like structure (Fig. 2ej). Indeed,

in some instances, element I cartilage had an opening (arrows in Fig. 2d,e) that might correspond to the obturator foramen in vivo. Such openings did not occur in the other more distal elements (11-V). The second and only other element to be expressed in the central subsets of stage 18 mesoblasts was rounder and more nodule-like than element I (Fig. 2a). Element I1 showed a marked tendency to elongate with time in culture in a distal-posterior direction within central subsets established from stages 19- to 20- mesoblasts (Fig. 2 M ) . Histological sections of the elongated element I1 indicated that histodifferentiation was not homogeneous, but rather different zones were recognizable which were mindful of quiescent, proliferative, and hypertrophic regions characteristic of in situ chondrogenic models of long bones (Fig. 3). As shown in Figure 2, the elongation of element I1 was a time dependent phenomenon; after stage 20-, as additional chondrogenic elements formed within central subsets, element I1 remained more cuboidal or block-shaped and variably fused with one or more distal elements. Element I11 first appeared in the central subsets of stage 20 and 21 mesoblasts, apparently adjoined OT connected to an element I1 to form an apparent transitional element designated element II/III that varied in shape from somewhat cylindrical to triangular or Yshaped (cf. Fig. 2c,f-i). In some cases more than two cartilage masses were delineated faintly (indicated with dots in Fig. 2i). Serial sections revealed more than two or three whorls within this cartilage (not shown). Eventually, element IUIII completely segregated into three pieces in stage 22 mesoblast cultures: element 11, anterior element 111, and posterior element I11 (Fig. 2j,k). Element IV appeared at the distal end of central subsets in 24% of stage 20 and in 48% of stage 21 mesoblast cultures (Fig. 2g-1, Table 1). Element V, the most distal element to form, was expressed in only about one-half (56%) of stage 22 cultures (Fig. 21, Table 1). In cultures of this stage, cell death was observed frequently in the area corresponding to the center of the original explant, which resulted in posterior element I11 being replaced with cell debris (asterisk in Fig. 21). By superimposing outlines of individual cartilage elements from several cultures, we observed that the spatial expression of a particular cartilage element within a central subset was highly reproducible within given staged groups of mesoblast cultures. Moreover, a relationship was revealed between the anatomical coordinates of the explanted mesoblast and the specific element of cartilage that subsequently formed in culture. Element I consistently developed in an anteroproximal direction from the center of the original explant (Fig. 4a,c), whereas element II/III (Fig. 4b,d) also formed anterior to the center, but more distally than for element I. In contrast, element IV formed in a postero-distal direction from the center of the original explant (Fig. 4e).

LIMB MORPHOGENESIS IN CULTURE

a

b

Q

d

f

e

9

h

i

k

Q

a I

317

Fig. 2. Cartilage formed by whole leg mesoblasts after 6 days in culture: (a) stage 18, (b) stage 19-, (c) stage 19, (d) stage 20-, (e-g) stage 20, (h,i) stage 21, (j-I) stage 22. I through V represent individual cartilage elements as described in the text. 11/111 is a fusion of elements II and Ill. The line drawings represent the outlines of individual cartilage elements shown in the photographs. The proximal, distal, anterior, and posterior directions of original explants are oriented in left, right, top, and bottom, respectively. Arrows in (d) and (e) indicate presumptive obturator foramen. Dots in (i) indicate ill-defined cartilage mass within element 11/111. Asterisk in (I)indicates debris from cell death. Alcian blue staining. Bar = 500 pm.

318

ISOKAWA ET AL.

Fig. 3. Histological section of stage 19 whole mesoblast after 6 days in culture. Cartilage elements I and II are apposed to each other (arrow) without a distinct intervening perichondrium. Elongated element II shows three zones of chondrocytes (i.e., quiescent, proliferative, and hypertrophic) from proximal (left, next to element I) to distal (right). Proximal quiescent chondrocytes are in a whorl arrangement. Alcian blue and Kernechtrot double staining. Bar = 500 pm.

anterior

a

1 Ccenter 01oriainal exDiant

Fig. 4. Locations of cartilage elements relative to the original explants: (a) element I, stage 19/20-, (b) element II, stage 19/20-, (c) element I, stage 21, (d) elements 11/111, stage 21, (e) element IV, stage 21. Outlines of cartilage elements after 6 days in culture were superimposed. Each consists of 5 randomly-selected cases. Bar = 500 pm.

Does the Number of Cartilage Elements Formed by a Mesoblast Change With Respect to Time in Culture? To address this question, stage 19/20- mesoblasts were analyzed in culture over a 10-day time course (Fig. 5). As was characteristic for culture day 6 of stages 19120-, two chondrogenic elements formed, a small element I and an elongated larger element 11. Both of these elements appeared after 2 days in culture (Fig. 5a) and were clearly demonstrated by day 3 (Fig. 5b). Histological sections of the central subsets showed that often after 2-3 days of culture, the alcian blue positive elements (condensations) were not demarcated by a perichondrium (data not shown); however, by day 6, both elements were ensheathed by a non-alcianophilic perichondrium (see Fig. 3). When cultures were permitted to grow for longer than 6 days, no additional cartilage elements developed, although the absolute size of existing cartilage elements increased (cf. Figs.

5c and 2d). Thus, no new chondrogenic elements could be initiated within the central subset of a stage 19120mesoblasts beyond the two that were detected initially after 2-3 days in culture. Only when cultures were initiated with the next older stage mesoblast could additional, or structurally more complex, chondrogenic elements be formed (Table 2).

Formation of Cartilage Elements From Either Anterior or Posterior Halves of Limb Mesoblasts A characteristic of limb development in vivo is the ability to compensate for, or “regulate,” the loss of mesoderm in forming chondrogenic elements. Hampi5 (1959) demonstrated specifically for hindlimbs of stages 19 and 20 that a normal limb would be formed if some or all the mesoderm distal to line “b” in Figure 1 were removed (provided an ectodermal jacket remained). His results showed that the potential for the remaining limb tissue to regulate diminished progressively as the age of the embryo increased. Accordingly, we sought to determine if the apparent “pattern” of cartilage element formation observed in our culture system could be modified by removing specific regions of the mesoblast prior t o placing it into culture. To do so, some portion or all of the anterior or posterior regions of stages 19120- and 21 mesoblasts were cultured separately and the resulting cartilage patterns compared after 6 days. In initial experiments, we removed just the anteroproximal area of stage 19120- hindlimbs (shaded area in Fig. 1). The outcome of such extirpations was that the first element was reduced in 47% of the cases (n = 15). In a more extensive series of experiments, we then removed the entire anterior half of the mesoblast, culturing only the posterior half. The outcome was virtually identical to removing just the anterior proximal portion of the mesoblast (i.e., approximately a 50%reduction in element I formation but no loss of more distal elements) (Table 2; Fig. 6a,c). Moreover, these findings did not change if older mesoblasts (stage 21) were used. Basi-

LIMB MORPHOGENESIS IN CULTURE

319

Fig. 5. Time course of stage 20- whole mesoblast cultures: (a) day 2, (b) day 3,(c) day 10. The proximal, distal, anterior,and posterior directions of original explants are oriented in left,right, top, and bottom, respectively. Polarization of alcian blue-positivecartilage elements to the anterior region is apparent in (a),because outer boundary of the original explant remains well-delineated.Light dark area indicated by arrows in (a)and (b) is actually brownish and does not represent specificalcian blue staining. Bars = 500 pm.

TABLE 1. Pattern of Cartilage Elements Formed by Whole Leg Mesoblasts After 6 Days in Culture" Stage 18/1919/2020 21 22

Total cultures 19 46 45 31 16

I 47% (9) 89% (41) 93% (42) 97% (30) 100%(16)

Incidence of cartilage element I1 I11 IV 100% (19) 0% (0) 0% (0) 100%(46) 0% ( 0 ) 0% (0) 100%(45P 24% (11) 100%(3lIb 48% (15) 100% (16) 100% (16) 100% (16)

V 0% (0) 0% (0) 0% (0) 6% (2) 56% (9)

T h e number in parenthesis refers t o cases observed. I through V represent individual cartilage elements as explained in Results. bAt these stages, cartilage corresponding to element I1 and I11 occurred as a fusion (element II/IIII.

TABLE 2. Pattern of Cartilage Elements Formed by the Anterior or Posterior Half of Leg Mesoblasts After 6 Days in Culture" Stage 19/2021

Limb region anterior posterior anterior posterior

Total cultures 29 27 25 25

Incidence of cartilage element I I1 I11 IV 86% (25) 100%(29) 0% (0) 0% (0) 41% (11) 100%(27) 0% (0) 0% (0) 100% (25) 100% (25Ib 48% (12) 40% (10) 100% (25Ib 52% (13)

T h e number in parenthesis refers to cases observed. I through V represent individual cartilage elements as explained in Results. bAt this stage, cartilage corresponding to element 11 and I11 occurred as a fusion (element IUIII).

cally, in our culture system, posterior halves of stage 19/20-through stage 21 mesoblasts were unable to form the first element of cartilage about half the time. In contrast, the corresponding anterior halves of the mesoblasts formed the characteristic number of elements for each stage 100% of the time (Table 2; Fig. 6b,d). These data suggest that both anterior and posterior halves of the original mesoblast are capable of compensating for the loss of each other in terms of expressing a specific number of chondrogenic elements.

While such compensatiodregulation did not change as the complexity of the chondrogenic pattern increased with age of the hindlimb, regulation for element I specifically was reduced for posterior mesoderm. In addition, the anatomical sites a t which the chondrogenic elements formed in cultures of posterior mesoderm appeared to be shifted anteriorly, relative to the original location of the explant (Fig. 7). This kind of shift was never observed in cultures of anterior mesoderm.

320

ISOKAWA ET AL.

A unique feature of tip mesoblast cultures was the occurrence of free chondrocytes outside the organized cartilage mass, within the lattice of the collagen gel. These cells were rounded and had a marked alcian blue-positive, pericellular matrix, and invariably were polarized to the hypertrophic side of the cartilage mass (Fig. 8). These cells were observed in 85%(23 cases out of 27) of stage 19/20-distal tip mesoblast cultures. Conversely, cultures of stage 19/20-mesoblasts without the distal tip region produced chondrogenic elements characteristic of those formed by whole mesoblasts (data not shown).

Fig. 6. Comparison of cartilage developed by anterior and posterior halves of mesoblasts after 6 days in culture: (a) anterior, stage 20-, (b) posterior, stage 20-, (c) anterior, stage 21, (d) posterior, stage 21. The proximal, distal, anterior, and posterior directions of original explants are oriented in left, right, top, and bottom. Control culture for stages 20- and 21 showed the same pattern with Figure 2d,h. Cartilage was visualized by alcian blue staining. Bar = 500 pm.

Fig. 7. Locations of cartilage elements relative to the original anterior or posterior half mesoblasts: (a) anterior, stage 20-, (b) posterior, stage 20-, (c) anterior, stage 21, (d) posterior, stage 21. Each outline consists of 5 randomly-selected cases of cartilage formed after 6 days in cultures. Correlate these drawings with photographs in Figure 6. Bar = 500 pm.

Distal Tip Mesoblast Cultures In these experiments, expression of chondrogenic elements was compared in cultures established from just the distal tip of stage 19t20- mesoblasts (subridge mesoderm, see Fig. 1)with those explanted from mesoblasts which were missing this portion. In cultures of distal tip mesoderm, one small piece of cartilage was observed which was ensheathed by a perichondrium (Fig. 8). This single element of cartilage appeared to correlate with the distal part of element I1 formed in cultures of whole mesoblasts, because this cartilage mass showed a similar tendency to elongate. Three distinct zones, quiescent, proliferative, and hypertrophic, were observed in cases where the elongation of this element was marked, but only the hypertrophic zone was observed in less elongated cartilage.

DISCUSSION While other culture systems provide for expression of a chondrogenic phenotype (Solursh and Reiter, 1980; Archer et al., 1985; Kosher and Rodgers, 1987; Cottrill et al., 1987),our data suggest that the approach used in this study is unique in that it provides for the isolation of a subpopulation of mesoderm that not only retains its potential for overt chondrogenic cytodifferentiation, but also its potential for the formation of a predictable distribution (“pattern”) of chondrogenic elements. Moreover, for the collective reasons discussed below, we have interpreted the in vitro formation of chondrogenic elements to suggest that the cells of the central subset retain reproducible characteristics of “limbness.” To determine limbness, we sought to compare our in vitro findings to those obtained in vivo by Hampe (1956,1957,1959),who in an extensive series of experiments established fate maps for the mesoderm of the hindlimb bud through stage 23. Two of his major conclusions were (1) that as the age of the hindlimb increases, the mesoderm becomes increasingly competent to produce a greater number of skeletal elements, and (2) that each of the elements is laid down progressively in a proximal to distal sequence. In this regard, for our culture system, we have shown the temporal expression of chondrogenic elements by cells of the central subset was not immutable, but rather changed in a highly reproducible and characteristic manner, according to the age of the mesoblast used to establish the culture. For a specific stage, the pattern of element formation could not be modified, regardless of the length of time in culture. Spatially, the formation of elements within the central subset was not random but occurred in a proximal to distal sequence. Consistent with the previously established fate maps and the concept of limbness, most of the cells responsible for forming element I were in the proximo-anterior region of the central subset, while element IV forming cells appeared in the distal-posterior region of the central subset with the other elements in between. The highly reproducible manner in which the elements were expressed temporally and spatially in culture is also consistent with the concept of “positional information,” defined by Summerbell et al. (1973) to explain the

LIMB MORPHOGENESIS IN CULTURE

321

Fig. 8. Stage 20- distal tip mesoblast cultures after 6 days: (a) whole mount preparation, (b) histological section. Free chondrocytes (arrows) show alcian blue-positive pericellular matrix and are polarized to the hypertrophic side of the main cartilage mass. Bars = 500 pm.

characteristic pattern by which skeletal elements form within limb mesoderm. The capacity of both anterior and posterior mesoblasts to compensate for the surgical removal of mesoderm is also consistent with Hampe’s (1959) concept of “regulation” and further suggests retention of limbness in the cultured cells. The decrease in ability of posterior tissue to form element I, while it retains the ability to form more distal structures, may indicate that mesoderm from the posterior region of the limb either loses the potential to regulate expression of the most proximal elements earlier than anterior mesoderm does or has less initial potential for chondrogenesis. Possibly, the more anteriorly directed formation of chondrogenic elements seen uniquely in cultures of posterior mesoderm (Fig. 7) supports the second alternative, a finding consistent with the lack of alcian blue staining in the posterior region of whole mesoblast cultures a t day 2 or 3 (Fig. 5a,b). In either event, the observation that the central subset derived from the anterior half of a hindlimb bud can fully express the normal sequence of limb-like, chondrogenic elements seems to rule out the requirement of a polarizing signal after stage 18 from posterior mesoderm (compare with Fallon and Crosby, 1975; Crosby and Fallon, 1975; also see Honig, 1983a,b) to permit the process, exactly the outcome predicted by Hampe’s in vivo ablation experiments. Interestingly, we also found that the shapes of the chondrogenic elements which formed in culture resembled to varying degrees the morphology of in situ skeletal structures. The structural similarities were greatest for the initial chondrogenic elements (I and 11)that formed within central subsets, but diminished considerably for the last two elements (IV and V). Thus, depending upon their sequence of formation in culture, in vitro expression of chondrogenic elements appears to produce heterogeneous but reproducibly similar structures. Given the extraordinary differences between the culture and in vivo environments, including the ab-

sence of ectoderm in the culture system, one might expect only random chance for any structural similarities sufficient to even suggest potential relationships to in vivo skeletal structures. Yet, for cartilage element I, data were consistently observed that suggested more than a coincidental relationship between this element and the most proximal in vivo skeletal structure, the girdle. For example, element I included an opening reminiscent of an obturator foramen. The in vitro anlagen for cartilage element I developed invariably from the alcian blue-positive condensation located in the most proximo-anterior location of the central subset; similarly, in preliminary, in vitro experiments we found that the girdle developed from the most anteroproximal condensation to stain with alcian blue within the intact limb. Also, element I had features similar to a large cartilage mass produced in organ cultures of stage 17-21 leg buds that was identified by Suzuki and Ide (1987) as girdle-like. If this putative anatomical assignment for element I is correct, then it follows that element 11, anterior element 111, and posterior element I11 may be in vitro correlates of the femur, tibia, and fibula, respectively. In this context, it may be significant that the second cartilage element (putative femur), which forms in mesoblasts obtained from stage 20 or 21, frequently appeared Y-shaped; similarly, the in vivo cartilage blastema for the femur, which initially appears a t stage 25, is also Y-shaped as shown by anti-type I1 collagen (Dessau et al., 1980) or alcian blue whole mount staining (unpublished data). Without specific markers, we obviously cannot unequivocally equate any of the chondrogenic elements formed in vitro with in vivo correlates; however, the similarities in morphology were an unexpected outcome of the study which further supports our conclusion that limb-like characteristics are retained in this culture system. At the very least, we feel that these elements may represent chondrogenic foci that correlate with in vivo skeletal structures. What apparently is required to allow formation of a

322

ISOKAWA ET AL.

limb-like sequence of chondrogenic elements in culture is the preservation of naturally occurring cellular and/ or cell-matrix associations within either anterior or posterior mesoderm. If limb cells are dissociated before being placed in culture, a micromass of chondrogenic tissue forms without any specific indication of limbness (Ahrens and Solursh, 1979).Thus, a critical component for expressing limbness in culture appears to be the retention of endogenous cellular and/or cell-matrix associations. Our present data would further suggest that the association requirement for expression of limbness is greater than either the continued presence of an ectodermal influence or retaining both anterior and posterior regions of the mesoblast. Finally, our findings also point to the potential source(s) of an endogenous signal that permits the temporal specification of skeletal structures. A striking correspondencewas observed between the temporal potential for central subsets to form chondrogenic elements in culture and the results of excising the ridge ectoderm in ovo at different stages of development (Rowe and Fallon, 1982). Central subset mesoderm that has become determined to form a particular part of the leg under apical ectodermal ridge influence retains that developmental program in the absence of the ridge, both in vivo and in our culture system. In this context, it is noteworthy that only the cells located in the distal-most region of the limb mesoderm (i.e., directly beneath the apical ridge) behaved as if they were not yet fixed in terms of positional memory. Explants from this region produced free cells that migrated into the gel lattice (seeded subset) that were competent to express a chondrogenic phenotype but lacked potential to be organized into a distinct element enclosed by perichondrium. It is widely recognized that the distal mesoderm is part of a “progress zone” which is a region of undetermined cells under ridge inf hence (Summerbell, 1974a,b; Tickle et al., 1975; Carrington and Fallon, 1988). One predictable outcome of “truncating” that interaction by prematurely separating the interacting tissues is that the mesoderm might show reduced potential to form chondrogenic elements, exactly as observed in the distal tip cultures. Similarly, our findings with the distal tip cultures exactly paralleled those observed by Kosher et al. (19791, who found that morphogenesis of the subridge mesoderm in organ culture depended upon the continued presence of ridge ectoderm. Thus, we interpret these findings to indicate that while the inability of distal cells to specify a chondrogenic element reflects an interrupted interaction with the apical ridge, the ability of more proximal cells to reproducibly express, in the absence of a continued ectodermal influence, a specific pattern of elements reflects the outcome of an already completed inductive interaction. These data suggest to us that a proximal prechondrogenic subpopulation of mesoderm retains its memory of limbness when placed into collagen gel culture and, as such, the present culture approach could

provide the opportunity to investigate the homotypic or heterotypic cellular interactions involved in the temporal and spatial formation of chondrogenic elements in developing limb buds and, accordingly, could serve as a potential assay for ectodermal ridge function.

EXPERIMENTAL PROCEDURES Fertilized eggs from White Leghorn chickens were incubated to appropriate stages at 375°C in a humidified atmosphere. Embryos were removed, rinsed, and placed in Earle’s balanced salt solution (EBSS), and staged carefully according to the morphology of their leg buds, especially the lengthlwidth ( L N ) ratios (Hamburger and Hamilton, 1951). As this ratio for leg buds younger than stage 20 is not defined in the article, we used the following criteria: stage 18, LMr = 8.06.0; stage 19-, LIW = 6.0-5.0; stage 19, LMr = 5.03.5; stage 20-, LMr = 3.5-3.0. Mesoblast Preparation Undissociated mesodermal explants, “mesoblasts,” were prepared by dissecting leg buds from the body and removing the ectodermal jacket. Whole mesoblast explants were obtained from stage 18 to 22 leg buds by dissecting along the baseline of limb buds (line “b” in Fig. 1).Some stage 18119- leg buds were dissected to include body mesoderm (line “d” in Fig. 1). For stage 19120- and 21 anterior mesoblast and posterior mesoblast explants, limb buds were first cut along the line “ P M (Fig. 11, followed by a baseline cut. For stage 19120- mesoblasts without the antero-proximal portion, a wedge-shaped region of the limb bud (shaded in Fig. 1)was removed and discarded, and the remaining limb tissue was processed and cultured. Distal tip, or subridge, mesoblasts at stage 19120- were defined as the distal-most 0.1 mm of mesodermal tissue (hatched in Fig. 1):The remaining portion of the mesoblast without the tip was also cultured. After the dissection described above, limb bud pieces were digested briefly in 0.5%trypsin (GIBCO) in EBSS a t 37°C. When the ectodermal jacket became loosened (ranging from 4 min for stage 18/29 tip mesoblast preparations to 12 min for stage 22 whole mesoblast preparations), tissues were washed and transferred into icecold EBSS containing 2% chicken serum, and the ectoderm was removed carefully with forceps. Mesoblasts from which the ectoderm had been detached during digestion were not used. Culture Methods Hydrated collagen gels (rat tail tendon type I collagen; Collaborative Research) were prepared as described previously (Bernanke and Markwald, 1982) in 4-well tissue culture plates (Nunc; 0.3 mllwell, 1 mg collagenlml final concentration). After polymerization, the three-dimensional lattices of collagen were rinsed twice with minimal essential medium, alpha modification (MEMa; GIBCO), and then once with MEMa containing 1%chicken serum (SPAFAS), 5 Fg insulinlml,

323

LIMB MORPHOGENESIS IN CULTURE

5 WE transferridml, and 5 ng seleniudml (Collaborative-Research), and‘antibiotics (100 units penicillidml and 100 p,g streptomycidml, GIBCO). Gels were incubated with supplemented MEMa at 37°C in a humidified incubator with 5% C02/95% air atmosphere at least overnight before use. Medium was exchanged once more 2 hr prior to explanting mesoblasts. Mesoblasts were placed on a collagen gel (1 mesoblast/well) from which the medium had been removed, and were oriented to make sure their ventral or their dorsal surface was towards the gel surface. The anteroposterior and proximodistal axes were recorded on the plate lid, and the location of the mesoblast marked precisely by a minute dot a t the bottom of plate. After allowing mesoblasts to attach to the gels for 6 hr, 0.5 ml of supplemented MEMa was added to each culture. Mesoblasts were incubated 6 days in most cases. Medium was changed every two days.

Detection of Cartilage and Analysis Cultures were fixed in 2.5%glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 2 hr, rinsed several times in this buffer, and then stained overnight in 0.1% alcian blue 8GS (CHROMA) adjusted to pH 1.0. The cultures were washed repeatedly with cacodylate buffer until the gels became clear, then stored in glycerol containing 0.05% sodium azide. Observations were performed using a Nikon inverted microscope. The outline of alcian blue-positive cartilage was traced in a consistent and reproducible manner utilizing a 1 x 1 mm grid in an eyepiece of the microscope and recording paper with 0.25 x 0.25 inch grids a t a fixed magnification. The information about orientation and location of the original explant was transferred also to the recording paper. Some of the drawings thus obtained were input into a computer and processed (Adobe Illustrator; ADOBE) (i.e., the images were oriented and superimposed according to the dot that indicated the original location of the mesoblast). There was no detectable difference in the results between explants cultured ventral or dorsal side down, thus, we oriented most photographs and drawings in a dorsal view of the right leg bud. After this analysis, cultures were trimmed, dehydrated in graded ethanol and xylene, and embedded in paraffin. Serial sections perpendicular to the gel surface were prepared and restained with 0.1% alcian blue (pH 1.0) for 30 min (the initial whole mount staining did not penetrate into the center of the cartilage mass). Counterstaining was done with 0.1% Kernechtrot (CHROMA) in 5% aluminum sulfate for 10 min. ACKNOWLEDGMENTS The authors are grateful to Terri Murtha-Wardon for her expert technical assistance and t o Sue TjepkemaBurrows and Carolyn Snyder for preparing the figures. This work was supported by an NIH grant HD20743.

REFERENCES Ahrens, P.B., Solursh, M., and Reiter, R.S. (1977) Stage-related capacity for limb chondrogenesis in cell culture. Dev. Biol. 60:6982. Ahrens, P.B., Solursh, M., Reiter, R.S., and Singley, C.T. (1979) Position-related capacity for differentiation of limb mesenchyme in cell culture. Dev. Biol. 69:436-450. Archer, C.W., Rooney, P., and Cottrill, C. (1985) Cartilage morphogenesis in vitro. J . Embryol. Exp. Morphol. 90:33-48. Aulthouse, A.L. and Solursh, M. (1987) The detection of a precartilage, blastema-specific marker. Dev. Biol. 120:377-384. Bernanke, D.H. and Markwald, R.R. (1982) Migratory behavior of cardiac cushion tissue cells in a collagen-lattice culture system. Dev. Biol. 91:235-245. Carringtan, J.L. and Fallon, J.F. (1988) Initial limb budding is independent of apical ectodermal ridge activity: Evidence from a limoless mutant. Development 104361-367. Coelho, C.N.D., Krabbenhoft, K.M., Upholt, W.B., Fallon, J.F., and Kosher, R.A. (1991) Altered expression of the chicken homeoboxcontaining genes GHox-7 and GHox-8 in the limb buds of limbless mutant chick embryos. Development 113:1487-1493. Cottrill, C.P., Archer, C.W., Hornbruch, A,, and Wolpert, L. (1987) The differentiation of normal and muscle-free distal chick limb bud mesenchyme in micromass culture. Dev. Biol. 119:143-151. Crosby, G.M. and Fallon, J.F. (1975) Inhibitory effect on limb morphogenesis by cells of the polarizing zone coaggregated with pre- or postaxial wing bud mesoderm. Dev. Biol. 46:28-39. Dessau, W., Von der Mark, K., Von der Mark, K., and Fisher, S. (1980) Changes in the patterns of collagens and fibronectin during limb bud chondrogenesis. J. Embryol. Exp. Morphol. 5751-60. Fallon, J.F. and Crosby, G.M. (1975) Normal development of the chick wing following removal of the polarizing zone. J. Exp. Zool. 193: 449-455. Frederick, J.M. and Fallon, J.F. (1982) The proportion and distribution of polarizing zone cells causing morphogenetic inhibition when coaggregated with anterior half wing mesoderm in recombinant limbs. J . Embryol. Exp. Morph. 67:13-25. Hamburger, V. and Hamilton, H.L. (1951) A series of normal stages in the development of the chick embryo. J. Morphol. 38:49-92. Hampe, A. (1956) Sur la topographie des ebauches presomptives dans le bourgeon de membre posterieure du poulet. C. R. Hebd. SBanc. Acad. Sci. Paris 243:970-973. Hampe, A. (1957) Sur le rBle du mesoderme et de l’ectoderme du bourgeon de patte dans les 6changes entre le poulet et le canard. C. R. Hebd. Seanc. Acad. Sci. Paris 24k3179-3181. Hampe, A. (1959) Contribution B l’etude du developpment et de la regulation des deficiences et des excedents dans la patte de l’embryon de poulet. Archs. Anat. Microsc. Morph. Exp. 48:345-378. Hauschka, S. and Rutz, R. (1983) Regional distribution of myogenic and chondrogenic precursor cells in vertebrate limb development. In: “Limb Development and Regeneration, Part B,” Kelley, R.O., Goetinck, P.F., and MacCabe, J.A. (eds). New York: Alan R. Liss, Inc., pp 303-312. Honig, L.S. (1983a) Polarizing activity of the avian limb examined on a cellular basis. In: “Limb Development and Regeneration, Part A,” Fallon, J.F. and Caplan, A.I. (eds). New York Alan R. Liss, Inc., pp 99-108. Honig, L.S. (1983b) Does anterior (non-polarizing region) tissue signal in the developing chick limb? Dev. Biol. 97:424-432. Kosher, R.A., and Rodgers, B.J. (1987) Separation of the myogenic and chondrogenic progenitor cells of undifferentiated limb mesenchyme. Dev. Biol. 121:376-388. Kosher, R.A., Savage, M.P., and Chan, S.C. (1979) In vitro studies on the morphogenesis and differentiation of the mesoderm subjacent to the apical ectodermal ridge of the embryonic chick limb-bud. J . Embryol. Exp. Morph. 50:75-97. Linsenmayer, T.F., Toole, B.P., and Trelstad, R.L. (1973) Temporal and spatial transitions in collagen types during embryonic chick limb development. Dev. Biol. 35:232-239. Markwald, R.R., Bolender, D.L., Krug, E.L., and Lepera, R. (1990) Morphogenesis of precursor subpopulations of chicken limb mesen-

324

ISOKAWA ET AL.

chyme in three-dimensional collagen gel culture. Anat. Rec. 226: 91-107. Robert, B., Lyons, G., Simandl, B.K., Kroiwa, A., and Buckingham M. (1991) The apical ectodermal ridge regulates Hox-7 and Hox-8 gene expression in developing chick limb buds. Genes and Development 52363-2374. Rowe, D.A. and Fallon, J.F. (1982) The proximodistal determination of skeletal parts in the developing chick leg. J . Embryol. Exp. Morph. 68:l-7. Rutz, R., Haney, C., and Hauschka, S. (1982) Spatial analysis of limb bud myogenesis: A proximodistal gradient of muscle colony-forming cells in chick embryo leg buds. Dev. Biol. 90:399-411. Sasse, J., Horwitz, A,, Pacifici, M., and Holtzer, H. (1984) Separation of precursor myogenic and chondrogenic cells in early limb bud mesenchyme by a monoclonal antibody. J. Cell Biol. 99:18561866. Shinomura, T., Jensen, K.L., Yamagata, M., Kimata, K., and Solursh, M. (1990) The distribution of mesenchyme proteoglycan (PG-M) during wing bud outgrowth. Anat. Embryol. 181:227-233. Solursh, M. and Reiter, R.S. (1980) Evidence for histogenic interactions during in vitro limb chondrogenesis. Dev. Biol. 78:141-150. Solursh, M., Reiter, R.S., Ahrens, P.B., andvertel, B.M. (1981) Stageand position-related changes in chondrogenic response of chick embryonic wing mesenchyme to treatment with dibutyryl cyclic AMP. Dev. Biol. 83:9-19. Stirpe, N.S. and Goetinck, P.F. (1989) Gene regulation during cartilage differentiation: Temporal and spatial expression of link protein

and cartilage matrix protein in the developing limb. Development 107:23-33. Summerbell, D. (1974a) Interaction between the proximo-distal and antero-posterior co-ordinates of positional value during the specification of positional information in the early development of the chick limb-bud. J. Embryol. Exp. Morphol. 32:227-237. Summerbell, D. (1974b) A quantitative analysis of the effect of excision of the AER from the chick limb-bud. J . Embryol. Exp. Morph. 32:651-660. Summerbell, D., Lewis, J.H., and Wolpert, L. (1973) Positional information in chick limb morphogenesis. Nature 244:492-496. Suzuki, H.R. and Ide, H. (1987) Positional heterogeneity of interaction between fragments of avian limb bud in organ culture. Dev. Biol. 124:41-49. Swalla, B.J., Upholt, W.B., and Solursh, M. (1988) Analysis of type I1 collagen RNA localization in chick wing buds by in situ hybridization. L Dev. Biol. 12551-58. Tabin, C.J. (1991) Retinoids, homeoboxes, and growth factors: Toward molecular models for limb development. Cell 66:199-217. Tickle, C., Summerbell, D., and Wolpert, L. (1975) Positional signalling and specification of digits in chick limb morphogenesis. Nature 254199-202. Thorogood, P.V. and Hinchliffe, J.R. (1975) An analysis of the condensation process during chondrogenesis in the embryonic chick hindlimb. J . Embryol. Exp. Morphol. 33581-606. Zwilling, E. (1968) Morphogenetic phases in development. Dev. Biol. 2(~~pp1):184-207.