A variety of heterologous tissues have been tested for the ability to promote cartilage differentiation in isolated chick-embryo somites, using a modified chorioallantoic grafting technique. Of the 12 tissues tested only 3- and 4-day embryonic ectoderm promoted somite chondrogenesis in somites that fail to chondrify when grafted in isolation. This activity of ectoderm was evident in grafts of somites isolated with adjacent ectoderm, and in grafts of somites recombined with ectoderm derived from several sources. Four-day embryonic limbbud ectoderm, including the apical ridge, was capable of promoting somite chondrogenesis, but to no greater extent than dorsal trunk ectoderm of the same age. It is suggested that the ability of embryonic ectoderm to promote cartilage differentiation in isolated somites is associated with its ability to synthesize basement membrane material (sulphated glycosaminoglycans and collagen), in association with adjacent somite mesoderm.

A problem of importance in any proposed inductive interaction is the specificity of that interaction. The existence of other tissues capable of promoting the same differentiation casts doubt on a unique tissue-specific activity of the inducer, but may also reveal more general mechanisms whereby the phenotypic differentiation of a particular tissue type is favoured.

Recent experiments on the organ culture of isolated chick-embryo somites have revealed that in the in vitro environment factors unrelated to the presence of inducing tissues may promote the differentiation of cartilage (Ellison, Ambrose & Easty, 1969). It may be that nutritional supplementation of the organ culture medium, as employed by Ellison et al. (1969) is merely making good deficiencies that artificially depress the intrinsic chondrogenic potential of the somites. Alternatively, it is possible that the ‘stabilization’ of phenotypic differentiation observed in these cultures does reflect a response of the somites that may have a definite morphogenetic role in vivo.

Previous experiments on the specificity of the response of somites to spinal cord and/or notochord induction are inconclusive. On the one hand, no activity of tissues other than spinal cord and notochord has been demonstrated in organ culture (Avery, Chow & Holtzer, 1956; Stockdale, Holtzer & Lash, 1961). On the other hand, chorioallantoic cultures of somites indicated that adjacent tissues other than spinal cord and notochord might promote somite chondrogenesis (Seno & Büyüközer, 1958). Against the organ culture experiments may be adduced the fact that under the conditions employed in these experiments, somites that are now known to possess considerable intrinsic chondrogenic potential failed to differentiate cartilage. These conditions were thus manifestly suboptimal for demonstrating somite differentiation, and may have failed to reveal low levels of cartilage promoting activity in other tissues. The chorioallantoic culture experiments have, however, been subjected to alternative explanations (Stockdale et al. 1961), and being isolation rather than recombination experiments do not afford a critical analysis of the effect.

This controversial aspect of somite chondrogenesis has been examined using the modified chorioallantoic grafting technique reported previously (O’Hare, 1972). In this study both isolation and recombination experiments have been carried out on a more extensive scale than previously attempted. The modified chorioallantoic grafting technique, using a Millipore filter as graft vehicle, has been shown to be at least as sensitive in demonstrating somite chondrogenesis as the best organ culture techniques and superior to them in some respects, and thus affords a good method for testing the specificity of the somite response.

Preparation of somites and other tissues for grafting were as detailed in the previous paper (O’Hare, 1972), solutions of 3 % trypsin being used to dissociate the tissue.

Grafts were assembled on pieces of HA-grade Millipore filter prior to grafting. The slightly adhesive nature of the Millipore surface permitted the tissues to remain in close proximity during and after the grafting procedure. Grafts were transferred to the chorioallantoic membrane of 9- to 10-day host embryos, with the grafted tissues under the Millipore filter in direct contact with the chorioallantoic epithelium.

After 9 days’ culture, graft sites were identified by the presence of the Millipore filter, excised, fixed, embedded and serially sectioned. Sections were stained in alcian blue/haematoxylin/eosin and scored for differentiated derivatives.

Isolation of somites with ectoderm and endoderm

Grafts were made of groups of four stage 9 –12 posterior somites together with the adjacent ectoderm and endoderm. These somites have been previously shown (O’Hare, 1972) to consistently fail to differentiate cartilage when grafted in isolation.

In one set of experiments grafts were prepared by manual dissection with ectoderm and endoderm remaining undisturbed in association with the somites. A second set of grafts was made in which somites plus ectoderm/endoderm were treated with 3 % trypsin for 1 min before grafting. In the latter grafts the ectoderm and endoderm were still loosely adherent to the surface of the somites, but the close connexion between the tissues seen in vivo was largely disrupted. Results of these experiments are presented in Table 1.

Table 1.

Differentiation of somites isolated with ectoderm and endoderm

Differentiation of somites isolated with ectoderm and endoderm
Differentiation of somites isolated with ectoderm and endoderm

In the presence of undisturbed ectoderm and endoderm, stage 9 –12 posterior somites will differentiate cartilage in 21 % of grafts. Isolated somites consistently fail to differentiate cartilage in spite of graft viability being demonstrated by occasional nephric tubules found in the grafts.

Treatment of the somites plus ectoderm and endoderm with trypsin before grafting results in a fall in the incidence of both cartilage and distinguishable ectodermal derivatives to 10%. There is no correlation, however, between the presence in the final 9-day graft of ectodermal (or endodermal) derivatives and of cartilage.

Striated muscle and bone were not observed in any of these grafts. Nephric tubules were found in about 10% of grafts of isolated somites and of somites plus ectoderm and endoderm, but were unrelated to the presence of cartilage.

Recombination of somites with ectoderm

In these grafts, isolated stage 9 –12 posterior somites in groups of four were recombined with a piece of isolated ectoderm prepared from another part of the embryo. The following sources of ectoderm were tested: 2-day trunk ectoderm, 3-day trunk ectoderm, 4-day trunk ectoderm, and 4-day limb-bud ectoderm. Results are presented in Table 2.

Table 2.

Differentiation of somites recombined with ectoderm

Differentiation of somites recombined with ectoderm
Differentiation of somites recombined with ectoderm

The recombination of isolated somites with 2-day trunk ectoderm failed to result in any cartilage differentiating from the somites. Recombination of somites with 3- and 4-day trunk ectoderm did, however, result in a considerable incidence of cartilage. The incidence of detectable differentiated ectodermal derivatives rose from 42% with 3-day ectoderm to 60% with 4-day ectoderm plus somites, while the incidence of cartilage fell from 30% to 23%. There was no association of cartilage with differentiated ectodermal derivatives in the final graft, cartilage being found in grafts both with and without ectodermal derivatives.

The activity of 4-day limb-bud ectoderm, including apical ectoderm ridge, was almost exactly the same as that of dorsal trunk ectoderm of the same age.

A total of 70 grafts of 3- and 4-day isolated ectoderm was made without encountering any instances of cartilage differentiating due to accidental inclusion of mesenchyme. The incidence of distinguishable ectodermal derivatives in these isolated ectoderm grafts was 30 –40 %, being higher with older material. Ectoderm usually differentiated as keratinized epithelial vesicles. In a few instances, combined ectodermal/mesodermal structures such as feathers were found in grafts of 2-day somites with 3- or 4-day ectoderm, but even in these grafts ectoderm normally differentiated as keratinized vesicles.

The incidence of ectodermal differentiation reported here is probably an underestimate, as only large ectodermal formations could be distinguished from the keratinized epithelial ‘pearls’ that arise as non-specific manifestations of chorioallantoic epithelial metaplasia (Moscona, 1959).

Somites plus heterologous tissues

A total of 164 grafts of stage 9 –12 posterior somites in groups of four were made in association with a variety of heterologous tissues. The tissues tested included 3-day optic vesicle, 3-day otic vesicle, 4-day myocardium, 4-day mesonephros, 9-day heart ventricle, 9-day liver, 9-day intestine, 9-day choroid plexus, 9-day forebrain, 9-day spinal cord and polyoma transformed fibroblasts.

In all cases considerable growth and differentiation of these tissues took place in the grafts, but in no case did cartilage differentiate from the somites. The viability of the grafted 2-day tissues was demonstrated by the presence of nephric tubules in 10% of the grafts.

The results presented here show that the adjacent ectoderm and endoderm is capable of promoting somite chondrogenesis. This is in agreement with the results of Seno & Büyüközer (1958), and contrasts with the failure of somites to undergo chondrogenesis when cultured in vitro with ectoderm (Stockdale et al. 1961 ; Lash, 1963). The possibility that the cartilage arising in grafts of somites plus ectoderm and endoderm is of extra-somitic origin, as suggested by Stockdale et al. (1961), would seem most unlikely in the present study. There is no association whatever between the incidence of nephric tubules and of cartilage in the same grafts. Such an association would be expected if the cartilage arose from accidentally included lateral mesoderm adjacent to the nephrotome.

Cartilage-promoting activity is even more clearly revealed in the grafts of isolated somites recombined with 3- and 4-day ectoderm. A stage specificity in cartilage-promoting activity by ectoderm is suggested by the fact that 2-day ectoderm does not possess this activity, whereas 3- and 4-day are both active, with 4-day ectoderm less active than 3-day ectoderm in spite of a higher incidence of differentiated ectodermal derivatives.

Although the cartilage-promoting activity of ectoderm would appear to be stage specific it is not apparently spatially specific. Thus limb-bud ectoderm results in almost exactly the same incidence of cartilage differentiating in somite/ectoderm grafts as dorsal trunk ectoderm of the same age. The specific role of the limb-bud apical ectodermal ridge in the development of the limb mesenchyme (see Amprino, 1965) does not appear to influence its activity with respect to somite chondrogenesis (apical ectodermal ridge was deliberately included in all limb-bud ectoderm-containing grafts).

The effect of disrupting the ectoderm/endoderm-somite association by brief trypsinization, together with the effect of recombination of somites with older ectoderm, strongly suggests that the presence or absence of basement membrane material (BMM) in association with the somite influences its subsequent differentiation. The origin and characteristics of BMM found in association with somites and composed largely of sulphated glycosaminoglycans and collagen will be discussed in detail in a subsequent communication, but it has been found that the deposition of basement membrane material between somites and ectoderm is most marked in the 3-day embryo. This suggests that the activity of 3-day ectoderm in promoting somite chondrogenesis is related to its ability to reform BMM destroyed during isolation of the somites.

The importance of interface materials in other morphogenetic interactions has been well documented (see Grobstein, 1967), and they appear to play an important part in regulating the morphogenesis of epithelio-mesenchymally derived organ systems. The effect of ectoderm described in the present paper, like indeed all experimental interference in morphogenetic systems, may be artifactual in the sense that the somite never suffers in vivo deprivation or disruption of associated interface materials. Nevertheless, it is possible that the gradual deposition of such materials around the somite may constitute an important factor in temporal coordination of somite differentiation.

The absence of cartilage-promoting activity in any of the other tissues tested, including 9-day spinal cord, demonstrates that the mere presence of actively growing and metabolizing tissues (a situation not always attained in organ cultures) is not of itself an adequate stimulus for somite chondrogenesis. This may be contrasted with recent results concerning the specificity of the classic interaction between the spinal cord and metanephrogenic mesenchyme in the mouse which leads to the formation of kidney tubules. It would now appear that a variety of non-specific neural tissues (Lombard & Grobstein, 1969) and unrelated mesenchymes (Unsworth & Grobstein, 1970) can act to ‘induce’ tubules. The induction or promotion of somite chondrogenesis does, however, seem to be restricted to tissues that have some relationship to the somites in vivo. This, in turn, implies that these interactions do have genuine morphogenetic significance in vivo.

I am grateful to Professor E. J. Ambrose for encouragement, and Drs G. C. Easty and M. L. Ellison for helpful discussion. This investigation has been supported by grants to the Chester Beatty Research Institute (Institute of Cancer Research: Royal Cancer Hospital) from the Medical Research Council and the Cancer Campaign for Research.

Amprino
,
R.
(
1965
).
Aspects of limb morphogenesis in the chicken
. In
Organogenesis
(Ed.
R. L.
Dehaan
&
H.
Ursprung
), pp.
255
281
.
New York
:
Holt, Rinehart and Winston
.
Avery
,
G.
,
Chow
,
M.
&
Holtzer
,
H.
(
1956
).
An experimental analysis of the development of the spinal column. V. Reactivity of chick somites
.
J. exp. Zool.
132
,
409
425
.
Ellison
,
M. L.
,
Ambrose
,
E. J.
&
Easty
,
G. C.
(
1969
).
Chondrogenesis in chick embryo somites in vitro.
J. Embryol. exp. Morph.
21
,
331
340
.
Grobstein
,
C.
(
1967
).
Mechanisms of organogenetic tissue interaction
.
Natn. Cancer Inst. Monogr.
26
,
279
299
.
Lash
,
J. W.
(
1963
).
Studies on the ability of embryonic mesonephros expiants to form cartilage
.
Devl Biol.
6
,
219
232
.
Lombard
,
M.
&
Grobstein
,
C.
(
1969
).
Activity of various embryonic and postembryonic sources for induction of kidney tubules
.
Devl Biol.
19
,
41
51
.
Moscona
,
A.
(
1959
).
Squamous metaplasia and keratinization of chorionic epithelium of the chick embryo in egg and in culture
.
Devl Biol.
1
,
1
23
.
O’hare
,
M. J.
(
1972
).
Differentiation of chick embryo somites in chorioallantoic culture
.
J. Embryol. exp. Morph.
27
,
215
228
.
Seno
,
T.
&
Büyükôzer
,
I.
(
1958
).
Cartilage formation in somite grafts of chick blastoderm
.
Proc. natn. Acad. Sci. U.S.A.
44
,
1274
1284
.
Stockdale
,
F.
,
Holtzer
,
H.
&
Lash
,
J.
(
1961
).
An experimental analysis of the development of the spinal column. VII. Response of the dissociated somite cells
.
Acta Embryol. Morph, exp.
4
,
40
46
.
Unsworth
,
B.
&
Grobstein
,
C.
(
1970
).
Induction of kidney tubules in mouse metanephrogenic mesenchyme by various embryonic mesenchymal tissues
.
Devl Biol.
21
,
574
556
.