/. Embryo!, exp. Morph. Vol. 52, pp. 23-38, 1979
Printed in Great Britain © Company of Biologists Limited 1979
23
An experimental investigation into the possible
origin of pancreatic islet cells from
rhombencephalic neurectoderm
By ANN ANDREW 1 AND BEVERLEY KRAMER 1
From the Department of Anatomy, University of the Witwatersrand
Medical School, Johannesburg
SUMMARY
To determine whether or not any pancreatic islet cell type arises from rhombencephalic
levels of neurectoderm, lengths of presumptive rhombencephalon (containing potential
neural crest) of Black Australorp chick embryos at 6- to 9-somite stages were replaced isotopically and isochronically by neural tube of Japanese quail embryos. Some transplants
included mesencephalic regions. In some cases various levels of the rhombencephalon were
deleted and not replaced.
The quail nuclear marker was detected in cranial ganglia in operated embryos sacrificed at
3$ days of incubation and in enteric ganglia and cells accompanying some pancreatic nerves,
in embryos killed at 7 days of incubation. This provided evidence of normal migration of
crest cells from the grafts. Dopa was administered to the younger embryos, which were submitted to the formaldehyde-induced fluorescence procedure to demonstrate APUD (Amine
Precursor Uptake and Decarboxylation) cells. No pancreatic APUD cells exhibited the quail
nuclear marker. In 9- to 11-day embryos, A and B cells were identified by specific light and
electron microscopic features. None showed the quail marker. The marker was also absent
from those D cells seen and from cells of an as yet unidentified type, but not enough of these
were found to warrant a conclusion. All islet cell types were found in embryos from which
various levels of the rhombencephalon had been deleted.
It is concluded that at least A and B islet cells are not derived from the rhombencephalic
neurectoderm and probably not from mesencephalic levels. Their most likely origin remains
the endoderm, which was the accepted source until recently.
INTRODUCTION
Pancreatic islet cells are members of the APUD cell series (Pearse, 1966;
Andrew, 1976a). As such they are embraced by the proposal that all APUD
cells are derived from the neural crest or at least, the neurectoderm (Pearse,
1969; Pearse & Polak, 1971; Polak, Rost & Pearse, 1971; Pearse, Polak &
Bussolati, 1972). Many authors have accepted this view uncritically, whereas
1
Author's address: Department of Anatomy, University of the Witwatersrand Medical
School, Hospital Street, Johannesburg 2001, South Africa.
24
A. ANDREW AND B. KRAMER
Pearse (1975, 1977) has come to believe evidence to the contrary. It has been
shown that APUD cells present in the dorsal pancreatic bud of chick embryos
at 3-| days of incubation are not derived from trunk neurectoderm (Andrew,
19766); Pictet, Rail, Phelps & Rutter (1976) have found that the source of B
cells in the rat is not ectodermal; and Fontaine, Le Lievre & Le Douarin (1977)
have shown that APUD islet cells do not arise from vagal levels (levels of
somites 1-7) of the neural crest in chicks. In fact the latter investigation also
shows that the neural tube at this level - and thus the neurectoderm - is not the
source. Meanwhile similar work in our laboratory was testing hindbrain levels.
The derivation of enteric ganglia from the vagal neural crest (Andrew, 1970;
Andrew, 1971 (review); Le Douarin & Teillet, 1973) made such a source a
reasonable possibility. Furthermore, rhombencephalic levels had been mentioned
(Pearse, 1973) and mesencephalic levels indirectly suggested (Pearse & Polak,
1974), as the source of pancreatic APUD cells.
Apart from confirming the results of Fontaine et al. (1977), the present
investigation makes a twofold contribution. Firstly, it tests more rostral levels
of the neuroectoderm for a capacity to form islet cells. Secondly, it attempts to
specify the individual islet cell types covered by any conclusion regarding
possible origin from cephalic neurectoderm. In Fontaine's work, the APUD
reaction and a generalized islet cell stain were used to demonstrate islets: it
remained to ascertain which endocrine cell types were represented.
MATERIALS AND METHOD
(i) Experimental animals. These were chick embryos of the Black Australorp
breed and embryos of the Japanese quail {Coturnix coturnix japonica).
(ii) Microsurgical techniques. Neural tube incorporating the neural crest was
transplanted from quail to chick embryos before any migration of neural crest
cells could have occurred in either. The general procedure was that adopted for
testing trunk levels (Andrew, 19766) and was that used in the similar work of
Fontaine et al. (1977). Credit for the method is due to Weston (1963) and
Le Douarin (e.g. Le Douarin & Teillet, 1973); it is predicated on the discovery of
the quail nuclear marker (large Feulgen-positive nucleoli) by Le Douarin
(1971).
Grafts were transferred from various levels between the mesencephalicmetencephalic junction and the boundary between somites 5 and 6 of quail
embryos to chick embryos of practically the same stage, from which identical
extents of the neural tube had been deleted. Some grafts also included the
mesencephalon. The lengths of the grafts varied from two subdivisions of the
brain (e.g. meten- and myelen-cephalon), through neural tube at the levels of
somites 1-5 only, to rhombencephalon as far as somite 5, and occasionally
mesencephalon and rhombencephalon to the level of somite 5.
Host embryos received grafts at stages between the 6- and the 9-somite stage;
Origin of pancreatic islet cells
25
transplantation sites in the 9-somite embryos were at the more caudal levels
only.
A few deletions of cephalic neural tube were performed on chick embryos
without replacement by quail neural tube.
(iii) Light microscopy. Operated embryos were fixed at various times and in
different ways. At 3-| days of incubation seven were submitted to dihydroxyphenylalanine (dopa) administration and freeze-dried; the formaldehydeinduced fluorescence (FIF) and wax-embedding procedures followed as
previously (Andrew, 1975, 19766). This was to demonstrate the APUD
reaction in islet cells. Fluorescent cells in serial sections were photographed
and the sections were then stained with the Feulgen method as before.
The pancreata of four embryos with grafts were fixed in Bouin's fluid at 9
days of incubation, and of two embryos at 11 days. The pancreas of one
operated embryo lacking a graft was fixed in the same way at 9 days. Ribbons
of four 5 jLim wax sections were mounted consecutively on series of six slides.
Alternate slides were stained with the Feulgen method, and the intervening
slides with Graumann's modification of Hale's colloidal iron method (Klessen,
1975) for B cells, Levene & Feng's (1964) modification of the phosphotungstic
acid-haematoxylin method for A cells, and the Bodian-protargol method
(Humason, 1962, p. 227) with a repeated exposure to protargol and subsequent
reduction as advocated by Dawson & Barnett (1944). The Bodian technique
reputedly demonstrates A and D cells (Soleia, Yassallo & Capella, 1968), but
here did not impregnate A cells. (It was used because Hellerstrom and Hellmans'
method for D cells was not giving satisfactory results.) The scheme allowed
examination of the Feulgen-stained nuclei for the nuclear marker, in sections
adjoining those stained for each islet cell type.
Several different forms of the aldehyde fuchsin method failed to stain B cells
in our material; the colloidal iron method was more successful. Counterstaining
with picro-ponceau (Humason, 1962, p. 165) and nuclear staining with Mayer's
haemalum were devised by Tadiello (unpublished). Nucleoli were better
stained than by the Feulgen method which gives a weak reaction on Bouin-fixed
tissue.
(iv) Electron microscopy. At 7 days of incubation, pancreata from five other
embryos with grafts were fixed in Karnovsky's fixative (4 % paraformaldehyde
and 0-5 % glutaraldehyde in Millonig's buffer at pH 7-4) at 4 °C for 2 h. They
were post-fixed in 1 % osmium tetroxide for 1 h at 4 °C and embedded in
araldite for electron microscopy. Thin sections were stained with uranyl acetate
and lead citrate.
The pancreata from three operated embryos lacking grafts were fixed for
electron microscopy at 7 days of incubation.
(v) Control tissue from normal embryos. Pancreata from normal chick and
quail embryos of 9 days' incubation were fixed as above for comparison with
tissue from operated embryos by light and electron microscopy. Proventriculus,
26
A. ANDREW AND B. KRAMER
gizzard and/or small intestine from 9- to 11-day operated embryos were fixed in
formalin containing 2 % calcium acetate, serially sectioned and stained by the
Feulgen method to enable identification of enteric ganglion cells derived from
the grafts.
RESULTS
It had been established previously that pancreatic APUD cells are present in
quail and chick embryos at 3 | days of incubation (Andrew, 1975, 19766). These
cells were shown to contain typical large Feulgen-positive nucleoli in the quail,
easily distinguishable from the smaller, more numerous masses of chromatin in
islet cells of the chick. The same applies to enteric and cranial sensory ganglia
fixed in formalin and to other nerve cells, neurilemmal cells and islets in 9-day
embryos fixed in Bouin's fluid. Haemalum used with colloidal iron (Figs. 7, 8),
and the Bodian protargol method (Figs. 12, 13) allowed distinction of quail and
chick islet cell nuclei, phosphotungstic acid-haematoxylin also acts quite as well.
In the chick the lack of large nucleoli in islet cells contrasted with the more
prominent ones present in exocrine cells. The very large nucleoli of quail cells
were seen to be strikingly different from the small chick nucleoli in electron
micrographs of 9-day pancreatic islets (Figs. 16-20), though of course not every
section through a quail nucleus passes through the centre of a nucleolus.
Of 63 operated embryos, 21 (33 %) survived. As a rule the grafts had healed
in position well, but the brain was often somewhat malformed. The development of the pancreas usually appeared normal; sometimes it was a little small.
The criterion for successful transplantation was migration of neural crest cells
from the graft. Therefore in the embryos sacrificed at 3f days of incubation,
quail nuclei were sought in cranial sensory ganglia as well as in the operated
region of the brain (Fig. 1). In embryos fixed at 9-11 days, cells from the vagal
neural crest should have reached the foregut and at least the upper reaches of
the small intestine (see Le Douarin & Teillet, 1973). Since relevant levels (levels
of somites 1-7, Le Douarin & Teillet, 1973) had been almost entirely included
in all these grafts, quail nuclei in enteric ganglia (Fig. 5, cf. Fig. 4) and pancreatic
nerves (Figs. 6 and 15, cf. Fig. 14) indicated the success of the operation.
Thirteen out of eighteen embryos which had received transplants were successful
on these bases (Table 1).
(1) Formaldehyde-inducedfluorescence.In one of the embryos treated to show
FIF and scored as successfully operated, fluorescence was lacking in the
pancreas (probably due to some technical mishap). The others all exhibited
fluorescent APUD cells in apparently normal numbers in the pancreas. In none
of these cells were there quail nuclei (Table 2; Figs. 2 and 3).
(2) Light microscopy (LM). The staining of A and B islets for light microscopy
was very pale, as has been found for the embryonic pancreas by others (Schweisthal, Frost & Brinn, 1975). Nevertheless, in combination with the characteristic
Origin of pancreatic islet cells
27
Table 1. The occurrence of quail nuclei in derivatives of the rhombencephalic
neural crest in chick embryos with quail rhombencephalic grafts
No. of
embryos with
grafts
3f-day embryos
Cranial ganglia
7- to .11 -day embryos
Enteric ganglia
Pancreatic nerves
Totals
Quail nuclei
K
,
>
Absent
Present
No. of
successfully
operated embryos
7
2
5
5
3
3
8
8
8
8
13
11
18
Table 2. The occurrence of quail nuclei in pancreatic APUD cells in 3%-day chick
embryos with quail rhombencephalic grafts
Successfully
operated
5*
FIF in
islets
4
Quail nuclei
in APUD cells
0
* Number of embryos.
Table 3. The occurrence of quail nuclei in A and B pancreatic islet cells of7- to
11 -day chick embryos with quail rhombencephalic grafts
A cells
Light microscopy
Electron microscopy
Totals
B cells
Light microscopy
Electron microscopy
Totals
Successfully
operated
Quail nuclei
present
5*
3
8
0
0
0
5
3
8
0
0
0
* Number of embryos.
arrangement of small rounded groups of cells as B islets and large masses of
folded cords as A islets, the staining was adequate for the identification of the
two types of islets.
In all successfully grafted embryos prepared for light microscopy, colloidal
iron staining demonstrated B islets, in which the nuclei were typical of chick
cells (Table 3; Fig. 8, cf. Fig. 7). Adjacent Feulgen-stained sections confirmed
the lack of quail nuclei. Likewise, phosphotungstic acid-haematoxylin-stained
A islets were seen, by comparison with adjoining Feulgen- and nearby colloidal
iron-stained sections also, to consist entirely of chick cells (Table 3; Figs. 10
28
A. ANDREW AND B. KRAMER
Origin of pancreatic islet cells
29
and 11, cf. Fig. 9). As in the pancreas of normal 9-day chick and quail embryos,
Bodian-positive cells were sparsely scattered in the exocrine parenchyma. They
were found in most, but not all, the pancreata of embryos with grafts. All
nucleoli visible were small (Fig. 13, cf. Fig. 12); adjoining Feulgen sections
revealed no quail nuclei except in nerves.
(3) Electron microscopy (EM). In all three pancreata which were from successfully operated embryos and which were studied by electron microscopy, A, B
and D islet cells were identified by reference to the typical features described in
chick embryos (Dieterlen-Lievre, 1963, 1965; Machino, 1966; Machino &
Sakuma, 1968; Przbylski, 1967; Macerollo, 1977) and adult fowls (Kobayshi
& Fujita, 1969; Mukami & Mutoh, 1971). Large A islets (Fig. 17) contained A
cells (Fig. 18) having highly electron-dense, round profiles and more or less
loosely-fitting membranes. The size of the granules was variable, often smaller
than at older stages, as has been noted by Macerollo 1977). As in normal
embryos, B islets (Fig. 19) were smaller. In some cells the granules resembled
those of older chicks and adults in having electron-dense bars enclosed in
spherical membranes (Fig. 20). In others, some mature granules were present,
but most profiles were round and filled with electron-dense, finely granular
matrix; some of these granules had a less dense peripheral zone. D cells occurred
in very small numbers. Their granules were round, varied in electron density
and had fairly tightly-fitting membranes.
Some of those islet cells seen by electron microscopy to contain very few
secretory granules, had fairly large nucleoli. However, no others and no cells
recognizable as A, B or D cells showed the large nucleolus of quail cells (Table
3; Figs. 17-20). D cells were too infrequent for certainty that the nuclear profiles
FIGURES 1-6
Fig. 1. Section through the rhombencephalon of a 3^-day chick embryo which
received a graft of quail myelencephalon at the 74-somite stage. The junction
between chick (below) and quail neural tube (above) shows the difference between
chick and quail nuclei. On the right is a cranial sensory ganglion containing quail
cells. Formol acetate; Feulgen.
Fig. 2. Dopa-provoked formaldehyde-induced fluorescence in the dorsal pancreatic bud of the same operated embryo referred to in the legend for Fig. 1.
Freeze-dried; formaldehyde vapour-fixed.
Fig. 3. The same section illustrated in Fig. 2 subsequently stained with the Feulgen
method which shows that the nuclei of the fluorescent cells in Fig. 2 are chick
nuclei like the rest in the field. Freeze-dried; formaldehyde vapour-fixed; Feulgen.
Fig. 4. Nuclei in an intramural ganglion of the gizzard of a normal 11-day chick
embryo, for comparison with Fig. 5. Formol acetate; Feulgen.
Fig. 5. Quail nuclei with prominent nucleoli (cf. chick nuclei in Fig. 4) in an enteric
ganglion of an 11-day chick embryo which received a graft of quail mesen- and
rhomben-cephalon at the 7-somite stage. Formol acetate; Feulgen.
Fig. 6. Quail nuclei associated with a pancreatic nerve in an 11-day chick embryo
which had received a graft of quail mesen- and rhomben-cephalon at the 8-somite
stage. Bouin; colloidal iron, haemalum and picro-ponceau.
3
liMB 52
30
A. ANDREW AND B. KRAMER
Origin of pancreatic islet cells
31
Table 4. Levels of successful transplants and deletions of neural tube and neural
crest
Level
Mesencephalon
Metencephalon
Myelencephalon
Somite 1
Somite 2
Somite 3
Somite 4
Somite 5
No. of times
transplanted
No. of times
deleted
4
11
13
12
11
11
11
9
0
3
3
2
2
2
2
2
seen were not in sections that had missed the middle of a quail nucleolus, whereas A and B cells were sufficiently numerous. Quail nucleolar features were also
lacking in islet cells with polymorphic granule profiles. The matrix of the
granules was moderately electron dense and the bounding membranes were
tightly fitting. The profiles varied from round through elliptical to pear-shaped.
The granules in some cells were small. These cells were sparse in the islets, more
so in one specimen than in another.
The number of times each level of the neural tube had been transplanted in
successfully operated embryos is tabulated (Table 4). Levels between the metencephalon and somite 5 are represented 9-13 times, the mesencephalon 4 times.
In four embryos which received no grafts (1 LM; 3 EM), the meten- and
myelen-cephalon were deleted three times each, the somite levels twice. In all
FIGURES 7—13
Fig. 7. B islet in the pancreas of a normal 1.1-day quail embryo to show the quail
nuclear features with the B cell-staining procedure. Bouin; colloidal iron, haemalum
and picro-ponceau.
Fig. 8. Two B islets in an 11-day chick embryo which had received a graft of quail
mesen- and rhombencephalon at the 8-somite stage. The nuclei are chick nuclei
(cf. Fig. 7). Bouin; colloidal iron, haemalum and picro-ponceau.
Fig. 9. A islet tissue in the pancreas of a normal 11-day quail embryo to show the
nuclear features with the A cell-staining procedure. Bouin; phosphotungstic acidhaematoxylin.
Fig. 10. A islet tissue in a 9-day chick embryo which had received a graft of quail
mesen- and rhomben-cephalon at the 7-somite stage. Bouin; phosphotungstic acidhaematoxylin.
Fig. 11. The same site in an adjacent section to that shown in Fig. 10. It has been
stained with Feulgen; the nuclei are chick nuclei. Bouin; Feulgen.
Fig. 12. A Bodian-positive cell in the pancreas of a normal 11-day quail embryo,
showing the large nucleolus (N). Bouin; Bodian.
Fig. 13. A Bodian-positive cell in the pancreas of a 9-day chick embryo which had
received a graft of quail mesen- and rhomben-cephalon at the 7-somite stage. The
nucleus is a chick nucleus (cf. Fig. 12). Bouin; Bodian.
3-2
32
A. ANDREW AND B. KRAMER
Fig. 14. An electron micrograph showing a neurilemmal (Schwann) cell associated
with a pancreatic nerve in a normal 9-day quail embryo. The nucleolus is very
large as are the nucleoli of nearby mesenchymal cells.
Fig. 15. Two neurilemmal cells with quail nuclei associated with a pancreatic
nerve in a 7-day chick embryo which had received a graft of quail metencephalon
and rostral myelencephalon at the 7-somite stage. The mesenchymal nucleus at the
top is a chick nucleus.
Fig. 16. Part of an A islet from the pancreas of a normal 9-day quail embryo. Note
the large nucleoli seen in some of the quail cells.
Origin of pancreatic islet cells
33
these embryos, A, B and D cells and cells with polymorphic granule profiles
contained nuclei similar to those in specimens with successful transplants - they
were, of course, chick nuclei. The islet cell types were present in comparable
arrangement. Cells accompanying pancreatic nerves and others constituting
enteric ganglia seemed as numerous as in grafted embryos - they, too, naturally
contained chick nuclei.
DISCUSSION
It is clearly of paramount importance in this experiment that transplantation
should have been performed before any migration of neural crest cells had
started at the levels transplanted. The crest first appears at mesencephalic levels
and at the 6-somite stage (Holmdahl, 1928; Romanoff, 1960, p. 222). Lateral
migration of the cranial crest begins at the 8-somite stage (Romanoff, 1960,
p. 223). Thus this is the latest stage for transplantation of the mesencephalon.
On a similar basis, transplantation of the rhombencephalon should not be done
after the 9-somite stage. These limits were adhered to.
It is also vital to know whether the migration of quail neural crest cells occurs
normally in chick hosts. This is confirmed in the present experiment for the
migration of cells from the vagal crest which gives rise to enteric ganglion cells
in the small intestine (Andrew, 1970; Le Douarin & Teillet, 1973) and which
supplies parasympathetic pancreatic nerves with cells which include neuroblasts
and neurilemmal (Schwann) cells (Fontaine et al, 1977). In younger embryos,
in which not many such cells have yet arrived at these sites, the presence of
quail cells in cranial sensory ganglia was an indication of normal migration
from the grafts.
The complete absence of quail nuclear features from the pancreatic APUD
cells exhibiting dopa-provoked FIF at 3f days of incubation showed that none
of the APUD cells had arisen from mesen- or rhomben-cephalic levels of the
neurectoderm. The APUD cells of normal older embryos (9 and more days of
incubation) include A, B and D cells (Andrew, 1976a), and all three types are
present in the dorsal pancreatic bud of normal 3f-day embryos (DieterlenLievre, 1965; Przbylski, 1967; Andrew, 1977; Macerollo, 1977). It was therefore
likely, but not proven, that the APUD cells in 3|-day operated embryos included
the three cell types. To clarify this point a number of embryos was allowed to
develop to 7—11 days of incubation.
At these times, A, B and D cells could be recognized by their ultrastructure
and/or staining reactions. It was intended that the Bodian reaction would distinguish D cells. The small number of impregnated cells provoked us to test this
assumption. A combination of impregnation and immunocytochemical staining
for the D cell hormone, somatostatin, showed that only some D cells were
revealed by the Bodian method.
None of the A or B cells identified by light or electron microscopy and none
34
A. ANDREW AND B. KRAMER
Origin of pancreatic islet cells
35
of the Bodian-positive cells was seen to contain a quail nucleus. Too few D cells
were distinguished by either light or electron microscopy to warrant a conclusion
that none were quail cells. The only quail nuclei seen in the pancreas, however,
were those associated with nerves, as found too by Fontaine et ai (1977) in the
similar experiment which concerned vagal levels and to which reference has
already been made. A cell type differing in ultrastructural appearance from A,
B and D cells was encountered and showed too, no sign of quail nuclear features.
Although polymorphic, the granule profiles were very different from enterochromaffin granules (which are also characterized by 'polymorphic' profiles) in
size, electron density and even shape. Cells answering to a similar description
were seen by Pearse, Polak & Heath (1973) in early embryonic pancreas of the
mouse and were thought to be primitive islet cells. We have, however, seen the
same cell type scattered in the exocrine parenchyma in 18-day chick embryos.
It might be a cell type responsible for the production of one of the recently discovered pancreatic hormones.
The number of times each level of the hindbrain was transplanted, and the
number of successfully operated embryos justify a conclusion as regards the
hindbrain: the mesencephalon was transplanted less often but the results are in
line with those for the rhombencephalon.
The results of the few operations in which deletions, but no transplantations,
were performed, are consistent with an origin of islet cells from a source other
than the hindbrain and its neural crest. Alone, the latter results are not significant as the niche left by a lack of hindbrain-derived neurectodermal cells
could have been filled by crest cells from trunk levels. Such a phenomenon must
have occurred in the case of enteric ganglia and pancreatic nerves which were
populated by apparently normal numbers of cells.
The present experiments extend the findings of Fontaine et al. (1977). We can
now state that not only the vagal neural crest but also the entire rhombencephalon and its neural crest - and probably the mesencephalic neurectoderm - make
no contribution to the major pancreatic islet cell types, A and B cells. No contrary indications were found for the D cells seen or for the as yet unidentified
FIGURES
17-20
Fig. 17. Part of an A islet in a 7-day chick embryo which had received a quail
graft of caudal mesencephalon and entire rhombencephalon at the 7-^-somite stage
The nuclei are chick nuclei (cf. Fig. 16).
Fig. 18. An A cell from the pancreas of a 7-day chick embryo which had received a
graft of quail metencephalon and rostral myelencephalon at the 7-somite stage.
(The magnification of the nucleolus is twice that in Figs. 14-16.)
Fig. 19. Part of a B islet from the same operated chick as Fig. 17. The nuclei are
chick nuclei (cf. Fig. 16).
Fig. 20. A B cell from the same operated embryo as Figs. 17 and 19. The nucleus is a
chick nucleus.
36
A. ANDREW AND B. KRAMER
islet cell type, but enough observations were not forthcoming to justify a positive conclusion regarding either of them. Trunk levels of the neurectoderm have
already been shown not to be the source of pancreatic APUD islet cells of
3|-day embryos (Andrew, 19766). It remains only to specify the cell types subject to this conclusion for trunk levels.
A little experimental evidence in line with an endodermal origin of islet cells
has been forthcoming. In an experiment in which mesoderm from a quail
embryo was combined with chick endoderm, the donors were young enough to
preclude the presence of neural crest cells. In the pancreatic islet tissue which
differentiated, the A and B cells contained chick nuclei and were thus endodermal
in origin (Andrew, 1978).
Pearse (1977) now believes that gastro-intestinal and pancreatic endocrine
cells are endodermal but favours the idea that they may nevertheless arise from
'neurendocrine-programmed ectoblast'. Father to this thought is the remarkable
correspondence between the products of a number of APUD cells in the gastrointestinal tract and the products of some cells of accepted neural or neural crest
origin (Pearse, 1976, 1977). There is reason to believe that islet cells are not
derived from ectoderm (or ectoblast), but they certainly stem from the epiblast
as this is the source of most of the endoderm (Fontaine & Le Douarin, 1977).
However, that the same inductive influence is exerted on potential islet cells and
on classically defined neurectoderm (so programming the former in a neurendocrine manner) seems to us a remote possibility because of their different
location - the potential islet cells in the presumptive endoderm, the potential
neurectoderm in the presumptive ectoderm.
The authors wish to thank sincerely the Medical Research Council of South Africa, the
Senate Research Committee and the Council's Faculty of Medicine Research Grant of the
University of the Witwatersrand, Johannesburg, for grants supporting this project. For their
dedicated and skilled technical assistance we thank Mrs Grietje Holmes, Mrs Janet Layzell
and Mr Walter Tadiello. We appreciate the continued interest and encouragement of the
Head of the Department of Anatomy, Professor P. V. Tobias. We wish to pay tribute to
Professor Nicole Le Douarin, Director of the Laboratoire d'Embryologie of the C.N.R.S.
and the College de France for the method we have followed, and acknowledge the assistance
so generously given one of us with technical procedures by her colleague, Dr Christiane Le
Lievre. Our gratitude is due too, to Dr Francoise Dieterlen of the same institute for suggestions concerning the staining of islet cells, and to Dr Brian Follett of the Department of
Zoology, University of North Wales, for advice on the breeding of quails.
REFERENCES
ANDREW, A. (1970). The origin of intramural ganglia. III. The 'vagal' source of enteric
ganglion cells. /. Anat. 107, 327-336.
ANDREW, A. (1971). The origin of intramural ganglia. IV. The origin of enteric ganglia: a
critical review and discussion of the present state of the problem. /. Anat. 108, 169-184.
ANDREW, A. (1975). APUD cells in the endocrine pancreas and the intestine of chick embryos. Gen. comp. Endocrinol. 26, 485-495.
A. (1976a). Pancreatic APUD cells in older chick embryos, with special reference
to their identity. Histochemistry 49, 213-219.
ANDREW,
Origin of pancreatic islet cells
37
A. (19766). An experimental investigation into the possible neural crest origin of
pancreatic APUD (islet) cells. J. Embryol. exp. Morph. 35, 577-593.
ANDREW, A. (1977). Pancreatic D cells in very young chick embryos. Gen. comp. Endocrinol.
31, 463-465.
ANDREW, A. (1978). The embryonic origin of pancreatic islet (APUD) cells in the chick. In
Molecular Endocrinology (ed. I. Maclntyre & M. Szelke), Amsterdam, New York:
Elsevier/North Holland, pp. 355-359.
DAWSON, A. B. & BARNETT, J. (1944). Bodian's protargol method applied to other than
neurological preparations. Stain Tech. 19, 115-118.
DIETERLEN-LIEVRE, F. (1963). Demonstration de l'activite precoce des cellules A du pancreas
chez l'embryon de poulet. C.r.hebd. Seanc. Acad. Sci., Paris 256, 1597-1599.
DIETERLEN-LIEVRE, F. (1965). Etude morphologique et experimentale deladifferenciationdu
pancreas chez l'embryon de poulet. Bull. biol. Fr. Belg. 99, 3-116.
FONTAINE, J., LE LIEVRE, C. & LE DOUARIN, N. M. (1977). What is the developmental fate of
the neural crest cells which migrate into the pancreas in the avian embryo? Gen. comp.
Endocrinol. 33, 394-404.
FONTAINE, J. & LE DOUARIN, N. M. (1977). Analysis of endoderm formation in the avian
blastoderm by the use of quail-chick chimaeras. /. Embryol. exp. Morph. 41, 209-222.
HOLMDAHL, D. E. (1928). Die Entstehung und weitere Entwicklung der Neuralleiste. Z.
mikrosk.-anat. Forsch. 13, 99-298.
HUMASON, G. L. (1962). Animal Tissue Techniques. San Francisco and London: W. H.
Freeman.
KLESSEN, C. (1975). Zur Darstellung der B Zellen des Tnselsorgans mit der Eisenbindungs
Reaktion. Histochemie 45, 203-212.
KOBAYASHI, S. & FUJITA, T. (1969). Fine structure of mammalian and avian pancreatic islets
with special reference to D cells and nervous elements. Z. Zellforsch. mikrosk. Anat. 100,
340-363.
LE DOUARIN, N. (1971). Caracteristiques ultrastructurales du noyau interphasique chez la
caille et chez le poulet et utilisation de cellules de caille comme 'marqueurs biologiques' en
embryologie experimentale. Annls Embryol. Morph. 4, 125-135.
LE DOUARIN, N. M. & TEILLET, M.-A. (1973). The migration of neural crest cells to the wall
of the digestive tract in avian embryos. /. Embryol. exp. Morph. 30, 31-48.
LEVENE, C. & FENG, P. (1964). Critical staining of pancreatic alpha granules with phosphotungstic acid haematoxylin. Stain Tech. 39, 39-41.
MACEROLLO, P. (1977). Pancreatic islet cells in very young chick embryos with special
reference to D cells. Gen. comp. Endocrinol. 33, 139-146.
MACHINO, M. (1966). Electron microscopic observations of pancreatic islet cells of the early
chick embryo. Nature, Lond. 210, 853-854.
MACHINO, M. & SAKUMA, H. (1968). The secretory activity of the pancreatic delta cells of the
young chick and the chick embryo. /. Endocrinol. 40, 129-130.
MUKAMI, S. T. & MUTOH, K. (1971). Light and electron microscope studies of the pancreatic
islet cells in the chicken under normal and experimental conditions. Z. Zellforsch. mikrosk.
Anat. 116, 205-237.
PEARSE, A. G. E. (1966). Common cytochemical properties of cells producing polypeptide
hormones with particular reference to calcitonin and the thyroid C cells. Vet. Rec. 79, 587590.
PEARSE, A. G. E. (1969). The cytochemistry and ultrastructure of polypeptide-hormone
producing cells of the APUD series and the embryologie, physiologic and pathologic implications of the concept. /. Histochem. Cytochem. 17, 303-313.
PEARSE, A. G. E. (1973). Cell migration and the alimentary system: endocrine contributions
of the neural crest to the gut and its derivatives. Digestion, Basel 8, 372-385.
PEARSE, A. G. E. (1975). The endocrine division of the nervous system: a logical extension of
the APUD concept. Folia anat. iugoslavica 4, 5-20.
PEARSE, A. G. E. (1976). Peptides in brain and intestine. Nature, Lond. 262, 92-94.
PEARSE, A. G. E. (1977). The APUD concept and its implications: related endocrine peptides
in brain, intestine, pituitary, placenta and anuran cutaneous glands. Med. Biol. 55,115-125.
ANDREW,
38
A. ANDREW AND B. KRAMER
A. G. E. & POLAK, J. M. (1971). Neural crest origin of the endocrine polypeptide
(APUD) cells of the gastrointestinal tract and pancreas. Gut 12, 783-788.
PEARSE, A. G. E. & POLAK, J. M. (1974). Endocrine tumours of neural crest origin: neurolophomas, apudomas and the APUD concept. Med. Biol. 52, 3-18.
PEARSE, A. G. E., POLAK, J. M. & BUSSOLATI, G. (1972). The neural crest origin of gastrointestinal and pancreatic endocrine polypeptide cells and their distinction by sequential
immunofluorescence. Folia Histochem. Cytochem, Krakow 10, 115-120.
PEARSE, A. G. E., POLAK, J. M. & HEATH, C. M. (1973). Development, differentiation and
derivation of the endocrine polypeptide cells of the mouse pancreas. Immunofluorescence,
cytochemical and ultrastructural studies. Diabetologia 9, 120-129.
PICTET, R. L., RALL, L. B., PHELPS, P. & RUTTER, W. J. (1976). The neural crest and the
origin of the insulin-producing and other gastrointestinal hormone-producing cells.
Science, N.Y. 191, 191-193.
POLAK, J. M., ROST, F. W. D. & PEARSE, A. G. E. (1971). Fluorogenic amine tracing of
neural crest derivatives forming the adrenal medulla. Gen. comp. Endocrinol. 16, 132-136.
PRZBYLSKI, R. J. (1967). Cytodifferentiation of the chick pancreas. T. Ultrastructure of the
islet cells and the initiation of granule formation. Gen. comp. Endocrinol. 8, 115-128.
ROMANOFF, A. L. (1960). The Avian Embryo. Structural and Functional Development. New
York: Macmillan.
SCHWEISTHAL, M. R., FROST, C. C. & BRINN, J. E. (1975). Stains for A, B and D cells in fetal
rat islets. Stain Tech. 50, 161-170.
SOLCIA, E., VASSALLO, G. & CAPELLA, C. (1968). Selective staining of endocrine cells by basic
dyes after acid hydrolysis. Stain Tech. 43, 257-263.
WESTON, J. A. (1963). A radioautographic analysis of the migration and localisation of trunk
neural crest cells in the chick. Devi Biol. 6, 279-310.
PEARSE,
{Received 26 August 1978, revised 26 January 1979)