1. The water-soluble carcinogenic hydrocarbon, 1 : 2 : 5 : 6-dibenzanthracene-α-β-endo-succinate, has been implanted in carefully graded doses into gastrulae of Triton alpestris.

  2. An optimal dose for neural tube formation was found at 0·0125γ per embryo, and for palisade inductions at 0·000125γ per embryo.

  3. It is suggested that the action of this hydrocarbon at least is direct, imitating the action of the naturally occurring organizer substance, and not indirect, by injuring the ventral ectoderm so as to liberate the masked organizer substance which it contains. The effective range of dosage resembles that of many biologically active substances.

  4. A parallelism exists between the frequency and the size of the neural tube inductions.

From the beginning of gastrulation in the Amphibia, a stimulus is exerted by the invaginating archenteron roof upon the overlying dorsal ectoderm, which causes the latter to form the neural tube, the main axis of the future organism. The dorsal lip of the blastopore, where the invagination mainly proceeds, is therefore called the “organization centre” or “organizer”, and the process of effective stimulus and response is called “induction”. Further details will be found in the well-known monographs of Huxley & de Beer (1934), Dalcq (1935) and the recent Silliman Lectures of Spemann himself (1938).

When the induction process was subjected to further analysis, the important fact was discovered by Holtfreter (1933) that the substance or substances responsible for the stimulus (i.e. the “evocator” in the terminology of Needham et al. (1934)) was heat-stable. It was not destroyed by boiling the tissue. But Holtfreter also found that parts of the gastrula which never normally possess inductive power, such as the ventral ectoderm and yolk endoderm, acquire it after being boiled. Although this was at first interpreted as implying the existence of an inhibitory substance in these parts, the view which has become more generally accepted supposes there to be an inactive complex of the organizer substance, with protein and possibly with polysaccharide, in all the regions of the blastula. In normal development this inactive complex is broken down, with the consequent liberation of the active substance, only in the dorsal lip of the blastopore. By artificial treatment, however, such as boiling or treating with organic solvents, the complex can be broken down in ventral ectoderm also, so that it then acquires inducing power. For a discussion of these facts from the biochemical point of view, the review of Needham (1936) may be consulted.

Since the first discovery that dead material can perform neural inductions, a considerable number of experiments has been made in which chemically prepared fractions or pure chemical substances have been tested for organizer action. Thus Fischer et al. (1935) have obtained the induction of neural tubes with preparations of nucleoprotein and various fatty acids, while Waddington & Needham (1935) found similar activity on the part of a number of synthetic polycyclic hydrocarbons. But in general the significance of Holtfreter’s discovery of the “masked” organizer substance in the ventral ectoderm has not been sufficiently appreciated. The difficulty may be understood by reference to the diagram (Fig. 1). The only test for organizer activity is the implantation of the substance to be investigated into the cavity of the blastocoele, where it will come into contact with the presumptive epidermis, and, if it is effective, divert it from its normal fate to form a neural tube. But unfortunately the presumptive epidermis and the ventral ectoderm are the same thing, and it is therefore impossible in any given case to be sure that the substance implanted has not acted injuriously on the tissue, liberating the organizer substance contained in it. This objection is all the more plausible since it has been shown by Waddington et al. (1936) that it is not necessary to kill the ventral ectoderm in order to liberate the organizer ; it suffices to act upon it with certain dyes in fairly dilute concentration. In Fig. 1, X is the implanted substance ; it may act to give the response of induction, R, directly, but it may also act by liberating the masked organizer, (E), which then performs the induction.

Fig. 1.

Diagram to illustrate direct and indirect induction. X, chemical substance; (E), masked organizer substance; E, free, active, organizer substance; R, response of neuralization; blc, blastocoele cavity ; dl bp, dorsal lip of the blastopore ; pnp, presumptive neural plate ; ve, ventral ectoderm ; ye, yolk endoderm; yp, yolk-plug. The ectoderm is artificially thickened for diagrammatic purposes.

Fig. 1.

Diagram to illustrate direct and indirect induction. X, chemical substance; (E), masked organizer substance; E, free, active, organizer substance; R, response of neuralization; blc, blastocoele cavity ; dl bp, dorsal lip of the blastopore ; pnp, presumptive neural plate ; ve, ventral ectoderm ; ye, yolk endoderm; yp, yolk-plug. The ectoderm is artificially thickened for diagrammatic purposes.

It is difficult to see what other method can be adopted in order to distinguish between these direct and indirect forms of induction, save a study of dosage. If the required dose of a substance X is very small then there is less probability that it is acting by “injuring” the ventral ectoderm and so liberating the active substance of the ectoderm itself. If on the other hand a large amount has to be used, the probability of such an effect is greater. As a corollary it would follow that those chemical substances which are active in the most minute concentrations, are most likely to be chemically related to the organizer substance itself.

In the present investigation it was decided to make use of a water-soluble hydrocarbon introduced by Cook (1931)—1 : 2: 5 : 6-dibenzanthracene-α-β-endo-succinate. Its strongly carcinogenic properties are described by Barry et al. (1935) and Parsons (1936). It was tested as an organizer substance or evocator by Waddington (1938), who found it to be very active, though not among the most powerfully acting compounds. He obtained 39 % of neural tube inductions. The value of this substance lies in the fact that it is very soluble and can be made up in accurate dilutions.

The whole of the present work was done on gastrulae of Triton alpestris by the standard “Einsteckung” method of Mangold and Spemann. The developing embryos were selected very strictly into groups of the same morphological age as judged by the form of the blastopore.

The stock solution of the hydrocarbon was made up by dissolving 16·6 mg. in 3 c.c. (5·53 mg./c.c.). Crystalline egg albumen was added when necessary, forming the implantation mass, in the proportion of 100 mg. to a c.c. 1 c.mm. of the implantation mass therefore contained exactly 5 γ of the hydrocarbon, so that, since the average amount implanted into one blastocoele cavity was 0·25 c.mm., the dose per embryo was 1·25γ. This was the strongest dose given. From the original stock solution successive dilutions were prepared from 1/10 to 1/100,000, so that the smallest dose per embryo was 0·0000125γ, with all intermediate stages at dilutions of 10.

The results will be seen in Table I and Fig. 2. The system on which the classification of inductions was made was that used by Waddington et al. (1935). A signifies a well-formed neural tube, B followed by a variable number of plus signs signifies a “palisade induction”, i.e. the appearance of a more or less large number of neural cells in a row, but not joined up to form a tube, B followed by a minus sign indicates a slight degree of neuralization such as is sometimes produced by the implantation of egg albumen alone. C indicates a rod-like mass of neural tissue, with no lumen; this is rather a rare reaction. Apart from B–, the other negative classes are O, in which there is no reaction at all on the part of the ectoderm, and D, in which the ectoderm has formed a thickening, usually lenticular, composed of cuboidal cells without any morphological differentiation.

Table I.

Statistics of inductions with 1 : 2 : 5 : 6-dibenzanthracene-α-β-endo-succinate

Statistics of inductions with 1 : 2 : 5 : 6-dibenzanthracene-α-β-endo-succinate
Statistics of inductions with 1 : 2 : 5 : 6-dibenzanthracene-α-β-endo-succinate
Fig. 2.

× all inductions; • neural tubes; ∘ palisades.

Fig. 2.

× all inductions; • neural tubes; ∘ palisades.

The first noteworthy point in the results is that the inductive action of this hydrocarbon shows an optimal dose at 0·0125γ per embryo. At this dilution the number of neural tubes induced is a little over 40 % of the number of embryos in which the implant was retained, but on each side of it the percentage of neural tubes falls off. It may be mentioned that the criterion taken for a neural tube can be seen by reference to the illustrations in Waddington (1938). While it is not easy to picture the mechanism of the optimal dose effect in precise chemical terms, the fact that there is an optimal concentration makes it rather unlikely that the substance injures the tissue on which it is acting. If it did so, and liberated the masked organizer substance, its effect would be expected to increase with concentration.1

Another interesting point is that the palisade inductions rise to a maximum at concentrations lower than the optimum for neural tube inductions. This supports the impression hitherto generally current, that these are weaker manifestations of the same process as that involved in neural tube inductions. It may even be that they are early stages of tube formation, due to a slower action of the weaker concentrations.

In the work of Waddington (1938) on the inductive activity of various carcinogenic and oestrogenic hydrocarbons, it was found that a fairly far-reaching parallelism existed between the frequency of neural tube inductions and the actual volume of neural tissue induced. The more neural tubes were induced in a series, the larger their average volume was, with the exception of 3-4-benzpyrene, which induced 50 % of neural tubes, but all very small. This parallelism is found to hold good in the present work also; it will be seen from Fig. 2 that the average volume rises on each side of the optimal concentration to a peak at 85 c.mm. × 10−4. This level compares with the largest volumes of Waddington’s series.

It will be noticed in Table I that in the weakest concentration of all one neural tube was found, and it is interesting that in this case its size was extremely small, only 5 c.mm. × 10−4. A further examination of the effects to be found in these extremely small concentrations would be worth while in future laying seasons. This tube has been omitted from the graph on account of its exceptional position.

Perhaps the most important point about the whole investigation, however, is the low level of dose required to produce neural inductions. It is instructive to calculate the doses required in mg./kg. wet weight, and to compare the resulting figures with others in the literature for various biological responses. Assuming that a newt gastrula weighs approximately 4 mg. we find from Fig. 2 that the maximum neuralizing activity (including both neural tubes and palisade inductions) occurs when the gastrula receives 0·001 γ. The volume of the blastocoele cavity is taken as one-quarter of the egg volume, namely 1 c.mm., and the implantation mass one-quarter of the blastocoele cavity. Such a dose corresponds to 0·25 mg./kg. wet weight. Effects are still obtained, however, at a dilution one hundred times lower, i.e. 0·0025 mg./kg. wet weight.

This range stands in some contrast with the dosages which have been employed by some of the workers who have implanted chemical substances. Thus Fischer et al. (1935) used 5 % free fatty acids emulsified in agar-agar, so that 1 c.mm. would have contained 50γ fatty acid, i.e. a dose of 12·5 mg./kg. wet weight. Their-nucleo-protein preparations, moreover, had to be implanted as such, so that each gastrula must have received about 250γ, or 62·5 mg./kg. wet weight. More recently Barth (1937) has reported effects (neural tube inductions?) using 0·05 % digitonin in egg albumen, i.e. a dose of 31·5 mg./kg. wet weight. It will be seen that these amounts are all of the order of a hundred or a thousand times the dosages used in the present work. The latter, moreover, fall within the range of activity of hormones, vitamins, and other stimulating substances, for which a few data are assembled in Table II. The comparison between these is, of course, a little difficult, since hormones and drugs are injected in single doses, the vitamin figures are for daily requirements, and the carcinogens, which usually work very slowly, are administered by successive injections. However, the main point, that the dose of water-soluble hydrocarbon producing the maximum inductive effect on the gastrula ectoderm lies in the same range as that of many biological stimulating substances, is clearly established.

Table II.

Activities of biological stimulating substances

Activities of biological stimulating substances
Activities of biological stimulating substances

Since this is the case, it would seem more likely that the action of this substance (and hence, by implication, of the other hydrocarbons active in neural induction) is direct, that is to say, in imitation of the normal organizer, rather than indirect, by liberating the organizer substance masked in the ventral ectoderm.

The present set of experiments did not include any controls on pure egg albumen alone, but details of such experiments will be found in the earlier papers from these laboratories, especially Waddington (1938). No greater differentiation than the B– type has ever been observed on implantation of crystalline egg albumen alone.

This work was done during the tenure of a Rockefeller travelling Fellowship from Peiping Union Medical College and Chung-Cheng Medical College. It forms part of a plan of investigations proposed by Dr J. Needham and Dr C. H. Waddington, to both of whom I am indebted for help and advice. The hydrocarbon derivative was kindly provided by the Research Department of the Chelsea Royal Cancer Hospital.

Barry
,
G.
,
Cook
,
J. W.
,
Haslewood
,
G. A. D.
,
Hewett
,
C. L.
,
Hieger
,
I.
&
Kennaway
,
E. L.
(
1935
).
Proc. roy. Soc. B
,
117
,
318
.
Barth
,
L. G.
(
1937
).
Biol. Bull. Wood’s Hole
,
73
,
346
.
Chen
,
K. K.
,
Chen
,
A. L.
&
Anderson
,
R. C.
(
1936
).
J. Pharmacol
.
25
,
579
.
Cook
,
J. W.
(
1931
).
J. chem. Soc.
3273
.
Cook
,
J. W.
,
Dodds
,
E. C.
&
Lawson
,
W.
(
1936
).
Proc. roy. Soc. B
,
121
,
133
.
Dalcq
,
A.
(
1935
).
L’Organisation de l’œuf chez les Chordés. Paris
.
Deelman
,
H. T.
(
1935
).
Proc. Coherence Leeuwenhoek-Vereeniging
,
4
,
37
.
Dodds
,
E. C.
,
Goldberg
,
L.
,
Lawson
,
W.
&
Robinson
,
R.
(
1938
).
Nature, Land
.,
142
,
34
.
Fischer
,
F. G.
,
Wehmeier
,
E.
,
Lehmann
,
H.
,
Jühling
,
L.
&
Hultsch
,
K.
(
1935
).
Ber. dtsch. chem. Ges
.
68
,
1196
.
Harris
,
L. J.
(
1938
).
Private communication to the author
.
Holtfreter
,
J.
(
1933
).
Arch. EntwMech. Org
.
128
,
584
,
Huxley
,
J. S.
&
De Beer
,
G. R.
(
1934
).
Elements of Experimental Embryology. Cambridge
.
Needham
,
J.
(
1936
).
Proc. R. Soc. Med
.
29
,
31
.
Needham
,
J.
,
Waddington
,
C. H.
&
Needham
,
D. M.
(
1934
).
Proc. roy. Soc. B
,
114
,
393
.
Parsons
,
L. D.
(
1936
).
J. Path. Bact
.
43
,
1
.
Shear
,
M. J.
(
1936
).
Amer. J. Cancer
,
26
,
322
.
Spemann
,
H.
(
1938
).
Embryonic Development and Induction. New Haven
.
Waddington
,
C. H.
(
1938
).
Proc. roy. Soc. B
,
125
,
365
.
Waddington
,
C. H.
&
Needham
,
D. M.
(
1935
).
Proc. roy. Soc. B
,
117
,
310
.
Waddington
,
C. H.
,
Needham
,
J.
&
Bracket
,
J.
(
1936
).
Proc. roy. Soc. B
,
120
,
173
.
Waddington
,
C. H.
,
Needham
,
J.
,
Nowiński
,
W. W.
&
Lemberg
,
R.
(
1935
).
Proc. roy. Soc. B
,
117
,
289
.
Went
,
F. W.
&
Thimann
,
K. V.
(
1937
).
Phytohormones. New York
.
1

It may be of interest that when the slides were kindly examined by Dr Glücksmann, pycnotic nuclei were not found to any greater extent at the higher dosages than at the lower ones. This again suggests that degeneration phenomena did not play a large part.