Development 99, 521-526 (1987)
Printed in Great Britain © The Company of Biologists Limited 1987
521
Intercalation and the cellular origin of supernumerary limbs in Xenopus
KEN MUNEOKA* and EDEN H. B. MURAD
Developmental Biology Center, University of California, Irvine, California 92717, USA
'Present address: Department of Biology, Tulane University, New Orleans, LA 70118, USA
Summary
The hypothesis that a specialized polarizing zone
controls the pattern of the anterior-posterior axis
during limb development in Xenopus has been tested
by analysing the cellular contribution to supernumerary limbs. Supernumerary limbs were generated by
grafting hindlimb buds contralaterally between
X. borealis and X. laevis to appose anterior and
posterior limb tissues. Cells derived from these two
species of Xenopus are readily identified by staining
with quinacrine. The analysis of cellular contri-
bution showed that supernumerary limbs consist of
approximately half anterior-derived (57 %) and half
posterior-derived (43 %) cells. These data are not
consistent with the polarizing zone theory but are
consistent with the hypothesis that both supernumerary limbs and normally developing limbs arise
from intercalary interactions between limb bud cells
with different positional values.
Introduction
to supernumerary limbs would be expected to be
primarily from responding, nonpolarizing tissue (anterior) origin. In contrast, the intercalation hypothesis predicts that cellular interactions are mutual and
reciprocal leading to cellular contribution from both
anterior and posterior limb tissues. However, regarding the intercalation hypothesis, although equal cellular contribution would be predicted, asymmetrical or
one-sided contribution to supernumerary limbs is not
difficult to understand given the evidence for directional or one-sided intercalation during pattern regulation (see Muneoka, Holler-Dinsmore & Bryant,
1986a). Thus, asymmetrical or one-sided cellular
contribution would not necessarily distinguish between the two views, whereas equal contribution
from graft and host would make the polarizing zone
model unlikely.
In amphibians, experiments in which cellular
contribution to supernumerary limbs have been
documented give varied results depending upon the
type of grafting operation utilized to stimulate supernumerary limbs. In the axolotl, supernumerary limbs
that form from contralateral grafts that appose anterior and posterior tissues consistently comprise
approximately equal numbers of cells of posterior and
anterior origin (Muneoka & Bryant, 1984a,b). However, a cellular analysis of supernumerary limbs
resulting from 180° blastema rotation shows a great
Currently there are two prevalent views about how
pattern is formed during the outgrowth of the vertebrate limb; the polar coordinate model (French,
Bryant & Bryant, 1976; Bryant, French & Bryant,
1981) and the polarizing zone model (Tickle, SummerbeU & Wolpert, 1975). The polar coordinate
model is based on the idea that limb-forming cells
have positional information and that intercalation
between cells with discontinuous information drives
patterning and outgrowth during development and
regeneration of the limb. The polarizing zone model
suggests that pattern is established during limb outgrowth with reference to a posterior polarizing zone
which is the source of a diffusible morphogen (for a
review of both models, see Bryant & Muneoka,
1986). These models were derived primarily from
tissue-grafting experiments which lead to pattern
regulation and supernumerary limb formation. One
way to discriminate between intercalation and the
existence of a specialized signalling region is to
examine the formation of supernumerary limbs following grafts to confront anterior and posterior tissue
utilizing cell marking techniques.
In theory, if a polarizing zone induces supernumerary limb formation and establishes pattern via
a diffusible morphogen, then cellular contribution
Key words: pattern formation, Xenopus, supernumerary
limbs, limb development, cell marker.
522
K. Muneoka and E. H. B. Murad
deal of variability, ranging from limbs derived solely
from anterior or posterior tissue to limbs which are
equally mixed (Maden & Mustafa, 1984). Such data
point to the equality of, as well as the potential of,
anterior and posterior tissues to participate in limb
formation and argue against the existence of a polarizing zone during limb formation" in urodeles. In
Xenopus the issue of cellular contribution to supernumerary limbs has been less well documented.
Cameron & Fallon (1977) used grafts involving the
Oxford single-nucleolate mutant to provide evidence
that supernumerary outgrowths that resulted from
180° limb bud rotation appeared to be solely derived
from anterior tissue. Based on these data, it was
suggested that a posterior polarizing zone is present
in the developing Xenopus limb bud. However, since
cell marking experiments in the axolotl give variable
results when supernumerary limbs are generated by
180° rotation (Maden & Mustafa, 1984), it is possible
that the results from Xenopus represent a subset of
the possible types of contribution patterns. Because
of the importance of resolving the issue of whether a
morphogen or intercalation is the most-likely candidate for the control of limb patterning, we have
reinvestigated this issue utilizing a high-resolution
cellular marker for studies on Xenopus (Thiebaud,
1983). This cellular marker takes advantage of the
cytological difference between cells from two closely
related species of Xenopus (X. laevis and X. borealis)
and virtually every cell in histological preparations
can be identified. We have utilized this marker to
investigate the cellular contribution to supernumerary limbs that form following contralateral
grafts to appose anterior and posterior tissue since
this grafting strategy gives consistent results in axolotl
studies. Our results demonstrate equal cellular contribution to supernumerary limbs in Xenopus which
are analogous to those found from similar experiments in the axolotl (Muneoka & Bryant, 1984a,b).
These data are consistent with the idea that intercalation drives the formation of supernumerary limbs in
amphibians and provide evidence against the existence of a polarizing zone during limb development in
Xenopus.
Materials and methods
All experiments were performed on Xenopus laevis and X.
borealis spawned at the University of California, Irvine.
Larvae were maintained at 20 ± 1 °C in 20 % Steinberg's
solution buffered with humic acid and changed three times
a week, and were fed on nettle powder. Paired X. laevis and
X. borealis larva at stage 52 (Nieuwkoop & Faber, 1975)
were anaesthetized in 1:4000 MS222 prior to operations.
Limb buds were amputated at the presumptive ankle with
iridectomy scissors and exchanged between X. laevis and X.
V
a
P
u
d
p
a
Fig. 1. Diagram showing the grafting procedure used in
this experiment. Hindlimb buds from stage-52 larvae
from X. borealis (stippled) and X. laevis were grafted
contralaterally (left to right) and oriented to confront
anterior and posterior bud tissues while aligning dorsal
and ventral tissues. Inset shows the final graft orientation
of a X. borealisrightlimb bud grafted onto a X. laevis left
hindlimb (a, anterior; p, posterior; d, dorsal).
borealis. Anterior-posterior contralateral grafts were made
such that anterior tissue apposed posterior tissue while the
ventral/dorsal tissues remained aligned (Fig. 1). Grafts
were held in place with strips of lens paper placed longitudinally over the graft junction. Following the operation,
animals were individually housed and checked daily to
record the survival of the gTaft. Camera-lucida drawings
were made weekly to monitor the formation of supernumerary structures. The animals were sacrificed just prior
to metamorphosis and the experimental limbs were fixed in
Carnoys for 6-8 h. Fixed limbs were either photographed
or drawn with a camera lucida to aid in the subsequent
reconstruction of the skeletal pattern from serial sections.
Following fixation, the limbs were decalcified for two days
in 10% versene (pH = 6-0), embedded in wax and cut at
7 fan into serial longitudinal sections. The sections were
stained with quinacrine (Thiebaud, 1983) and analysed by
fluorescence microscopy.
A quantitative analysis of the cellular contribution to
each supernumerary limb was made in the region of the
articulation between the metatarsal and first phalangeal
element of each digit. Counts were made of the number of
cells of X. laevis and X. borealis origin in the cartilage, the
soft tissue surrounding the cartilage (i.e. primarily loose
connective tissue and muscle, although positive identification of muscle tissue was not possible in all preparations)
and the epidermis, in dorsal, central and ventral longitudinal sections of each digit. Thus, for each digit data were
collected in 15 longitudinal fields, diagrammed in a representation of a transverse section in Fig. 2A. Since the cell
counts showed that the epidermis behaves independently of
Cellular origin of supernumerary limbs in Xenopus
B
Dermis
523
Cartilage Dermis
Dorsal
Central
Ventral
Fig. 2. Diagrams showing the regions of each digit scored for cellular contribution. (A) Regions analysed for cell
contribution in longitudinal sections projected onto a transverse section of a single digit. Cell counts were made of the
epidermis (e), the cartilage (c) and the soft tissue (d) between the epidermis and cartilage. A total of 15 fields was
counted for each digit of the supernumerary limbs analysed. (B) Diagram showing how the data are summarized in
Table 1. Since the cell counts of the epidermis showed that its behaviour was independent of the internal tissues (see
text), these counts were omitted from the data summary. The data are displayed for the cartilage and surrounding soft
tissue (dermis) in dorsal, central and ventral regions of each digit.
the mesodermal cells (see below), only the data from
mesodermal cell counts are displayed in summarizing the
results (Table 1) as shown in Fig. 2B. The data are
expressed in Table 1 as the frequency of X. borealis cells
(X. borealis cells/total cells); a frequency of 1-0 indicates
100 % contribution from X. borealis tissue, a frequency of
0-0 indicates 100 % contribution from X. laevis tissue.
Results
Anterior-posterior apposed (A/P) contralateral limb
bud grafts in X. laevis are known to result in the
formation of supernumerary limbs (Maden, 1981). In
our experiment, interspecific limb bud exchanges
between X. laevis and X. borealis resulted in the
formation of supernumerary limbs in 17 of 23 (74 %)
cases. All except one of these were single (sometimes
incomplete) supernumerary limbs forming either anterior or posterior to the grafted limb bud. From
these, nine of the most complete supernumerary
limbs (six posterior and three anterior) were selected
for detailed analysis of limb pattern and cellular
contribution; four of these were complete 5-digit
limbs, three were 4-digit limbs and two possessed 3
digits. The criteria used to reconstruct the skeletal
pattern of the limb were those of Muneoka et al.
(1986b). All nine of the supernumerary limbs selected
for analysis resulted from X. borealis limb bud grafts
onto X. laevis hosts.
Cellular contribution was analysed using the differential quinacrine staining of X. laevis and X. borealis
tissue (Fig. 3A,B) described by Thiebaud (1983). The
nuclei of X. laevis and X. borealis cells are easily
distinguishable when viewed with a fluorescence
microscope; X. laevis nuclei fluoresce uniformly
whereas X. borealis nuclei appear mottled with bright
fluorescent spots. This staining pattern is true for all
cell types in the limb and virtually every cell in
histological sections is scorable, thus making this cell
marker ideal for detailed analyses of cellular contribution (Thiebald, 1983). In addition, we found that
the differential staining of quinacrine is unaffected by
our decalcification procedure.
The analysis of cellular contribution to the epidermis of the supernumerary limbs indicated that in all
cases and for all regions counted, the epidermis was
derived entirely from the host stump. In fact in eight
of the nine limbs analysed the grafted Limb itself was
also covered entirely with host-derived epidermis.
This observation suggests that limb epidermis migrates distally over time and behaves independently
of the subepidermal mesoderm.
The results from the analysis of cellular contribution to the mesodermal derivatives in the supernumerary limbs are shown in Table 1. These data
clearly show that supernumerary limbs resulting from
A/P contralateral limb bud grafts in Xenopus are
composed of cells derived from both anterior and
posterior tissues at the graft junction. The frequencies shown in Table 1 were used to quantify the
overall cellular contribution from anterior versus
posterior tissue. Such an analysis indicates that the
cellular contribution from anterior tissue to supernumerary limbs ranges from 46% to 66% with a
mean of 57 % with posterior tissue contributing to the
remaining limb regions. Thus, the overall contributions from anterior and posterior tissue are close to
equal although there is variability from limb to limb.
The position of the graft junction was variable from
limb to limb, but generally fell within digits 2, 3 or 4
(Table 1). In addition, the nature of the boundary
524
K. Muneoka and E. H. B. Murad
between anterior-derived and posterior-derived cells
was highly variable; in some cases the boundary
appeared fairly abrupt yet in other cases it was
rather diffuse. An example of a section through an
anterior supernumerary limb with an abrupt graft/
host boundary is shown in Fig. 3C. One striking
aspect of the data is that there are peninsulas of hostor graft-derived cells which intrude into regions of
opposite origin. These appear as 'islands' in Table 1.
These peninsulas are predominately but not exclusively confined to the soft tissue between cartilage and
epidermis and are not regions of lymphocyte accumulation which would be indicative of graft rejection.
Furthermore, they are in most instances host-derived
cells which appear to have invaded graft-derived
regions of the supernumerary limb.
Fig. 3. (A) Quinacrine-stained cells from X. borealis showing the characteristic mottled staining of the nucleus used to
identify these cells in chimaeric tissues. (B) Quinacrine-stained cells from X. laevis showing uniformly bright nuclear
staining. (C) Low-magnification photograph of a longitudinal section through a 5-digit anterior supernumerary limb
(limb no. 7) which possessed an abrupt boundary (white line) between X. laevis and X. borealis cells. The boundary
parallels the posterior edge of the third metatarsal and proximally runs between the tibiale and fibulare. Distally the
boundary is shifted anteriorly to bisect the articulation between the metatarsal and the first phalangeal element.
Numbers indicate the 5 digits observed in this section of this supernumerary limb.
Cellular origin of supernumerary limbs in Xenopus
525
Table 1. Cellular contribution to supernumerary limbs
Digit number
Limb no.
1
2
3
4
5
B/B/B
B/B/B
B/B/B
B/B/B
B/B/B
B/B/B
37/L/L
52/L/L
5/L/L
L/L/L
L/L/L
L/L/L
B/B/B
B/B/B
B/B/B
52/B/B
63/B/67
64/B/70
70/45/L
90/77/L
32/B/40
L/L/L
L/L/L
L/L/L
B/B/B
B/B/B
B/B/B
B/62/B
B/26/L
B/35/23
B/L/L
B/18/L
B/L/L
Posterior supernumerary limbs
1
B/B/B
B/B/B
B/B/B
2
3
4
B/B/B
B/B/B
B/B/B
B/B/75
93/B/81
79/B/81
L/L/L
L/47/70
50/92/44
L/90/L
14/34/L
36/66/B
L/L/L
L/L/L
L/L/L
5
B/B/B
B/B/B
B/B/B
B/B/B
B/B/B
B/B/B
B/B/B
B/B/B
B/B/B
17/45/L
39/35/L
57/S7/L
L/L/L
L/L/L
L/L/L
6
96/B/69
89/B/77
98/B/82
79/B/6
66/B/52
85/B/77
23/L/L
L/L/L
3/L/L
B/B/B
B/B/B
B/B/B
Anterior supernumerary limbs
7
L/L/L
L/L/L
L/L/L
L/L/L
L/L/L
L/L/L
29/33/B
L/27/B
32/31/B
B/B/B
B/B/B
64/B/48
8
L/L/3
L/3/L
L/L/L
16/L/28
L/L/19
L/L/L
96/B/95
97/B/85
92/65/3
90/B/99
69/92/95
92/85/77
9
L/L/L
L/L/L
L/L/L
L/L/24
L/8/B
L/37/91
B/B/40
B/B/50
B/B/67
B/B/B
B/B/B
B/B/B
B, 100 % X borealis cells.
L, 100 % X foevis cells.
Numbers represent the percent of X. borealis cells.
Discussion
Supernumerary limbs are known to form following
contralateral grafts to appose anterior and posterior
hindlimb bud tissue in Xenopus (Maden, 1981). Using
the X. laevis/X. borealis cell marker we have found
that such supernumerary limbs consist of about equal
numbers of cells from both anterior and posterior
origin. Because of the excellent resolution of the X.
borealis/X. laevis cell marker used in this study we
were able to make detailed observations of the
behaviour of some limb tissues during supernumerary
limb formation. We have found that the epidermis
behaves independently of the mesoderm during
supernumerary limb formation: the host epidermis
apparently migrates distally and replaces the grafted
epidermis. Such observations are consistent with
those described in urodeles (Lheureux, 1983;
Muneoka & Bryant, 1984a; Tank, Connelly & Bookstein, 1985). In addition, the graft/host boundary of
mesodermal tissues appears convoluted to varying
degrees, resulting from.peninsulas of primarily hostderived cells intruding into regions which are otherwise of graft origin. These peninsulas are found in
areas between cartilage and epidermis suggesting that
they may represent populations of invading muscle
cells from more proximal limb regions as has been
shown to occur in the developing chick limb bud
(Newman, Pautou & Kieny, 1981).
The data from this study suggest that the cellular
interactions that result in supernumerary limbs are
mutual and reciprocal, and thereby argue against the
526
K. Muneoka and E. H. B. Murad
existence of a polarizing zone in the developing limb
bud of Xenopus. Our results are strikingly different
from a previous report by Cameron & Fallon (1977)
where it was argued that supernumerary limbs which
result from 180° limb bud rotation in Xenopus arose
exclusively from anterior limb tissues. This conclusion was based on two pieces of evidence. First, an
analysis of many supernumerary limbs suggested that
the dorsal-ventral orientation (based on curvature of
the digits) followed that expected from the anterior
component of the interaction. However, Maden
(1980) subsequently has demonstrated that one cannot rely on the curvature of digits but must analyse
muscle patterns histologically to accurately determine
the dorsal-ventral orientation of supernumerary
limbs in amphibians. In addition, the correlation
between dorsal-ventral pattern and cellular contribution to supernumerary limbs is not absolute
(Maden & Mustafa, 1984). Second, an assessment of
cellular contribution to supernumerary limbs using
the Oxford single-nucleolate mutant showed cellular
contribution from only anterior-derived tissue. Since
analogous studies in the axolotl (Maden & Mustafa,
1984) showed a wide range in contribution patterns to
supernumerary limbs (from solely anterior to solely
posterior), it is possible that Cameron & Fallon
(1977) may have observed only a subset of this
variability.
The data presented here do not support the polarizing zone hypothesis as an explanation for anteriorposterior limb patterning in Xenopus, and are consistent with the idea that intercalation between anterior
and posterior cells in the limb bud drives the formation of supernumerary limbs (French et al. 1976;
Bryant et al. 1981). These results are similar in many
respect to analogous cellular contribution studies
performed on the developing and regenerating axolotl limb where supernumerary limbs appear to result
from intercalary interactions between anterior and
posterior cells (Muneoka & Bryant, 1984a,b).
The authors thank Drs Susan Bryant, Rosemary Burton,
David Gardiner, Stanley Sessions and Nancy Wanek for
numerous discussions and critical comments on the manuscript. We thank Dr Danuta Krotoski for providing the
Xenopus borealis larvae. Research supported by PHS grant
HD 06082 and a gift from the Monsanto Company.
References
BRYANT, S. V., FRENCH, V. & BRYANT, P. J. (1981).
Distal regeneration and symmetry. Science 212,
993-1002.
S. V. & MUNEOKA, K. (1986). Views of limb
development and regeneration. Trends in Genetics 2,
153-159.
CAMERON, J. & FALLON, J. F. (1977). Evidence for
polarizing zone in the limb buds of Xenopus laevis. Devi
Biol. 55, 320-330.
BRYANT,
FRENCH, V., BRYANT, P. J. & BRYANT, S. V. (1976).
Pattern regulation in epimorphic fields. Science 193,
969-981.
LHEUREUX, E. (1983). The origin of tissues in the Xirradiated regenerating limb of the newt, Pleurodeles
waltlii. In Limb Development and Regeneration, Part A
(ed. J. F. Fallon & A. I. Caplan), pp. 455-465. New
York: Alan R. Liss, Inc.
MADEN, M. (1980). Structure of supernumerary limbs.
Nature, Lond. 287, 803-805.
MADEN, M. (1981). Experiments on anuran limb buds
and their significance for principles of vertebrate limb
development. /. Embryol. exp. Morph. 63, 243-265.
MADEN, M. & MUSTAFA, K. (1984). The cellular
contributions of blastema and stump to 180°
supernumerary limbs in the axolotl. J. Embryol. exp.
Morph. 84, 233-253.
MUNEOKA, K. & BRYANT, S. V. (1984a). Cellular
contribution to supernumerary limbs in the axolotl,
Ambystoma mexicanum. Devi Biol. 105, 166-178.
MUNEOKA, K. & BRYANT, S. V. (19846). Cellular
contribution to supernumerary limbs resulting from the
interaction between developing and regenerating
tissues in the axolotl. Devi Biol. 105, 179-187.
MUNEOKA, K., HOLLER-DINSMORE, G. V. & BRYANT, S.
V. (1986a). Pattern discontinuity, polarity and
directional intercalation in axolotl limbs. /. Embryol.
exp. Morph. 93, 51-72.
MUNEOKA, K., HOLLER-DINSMORE, G. & BRYANT, S. V.
(19866). Intrinsic control of regenerative loss during
limb development in Xenopus laevis. J. exp. Zool. 240,
47-54.
NEWMAN, S. A., PAUTOU, M.-P. & KIENY, M. (1981). The
distal boundary of myogenic primordia in chimeric
avian limb buds and its relation to an accessible
population of cartilage progenitor cells. Devi Biol. 84,
440-448.
NIEUWKOOP, P. D. & FABER, J. (1975). Normal Table of
Xenopus laevis (Daudin). Amsterdam: North-Holland
Publishing Co.
TANK, P. W., CONNELLY, T. G. & BOOKSTEIN, F. L.
(1985). Cellular behavior in the anteroposterior axis in
the regenerating forelimb of the axolotl, Ambystoma
mexicanum. Devi Biol. 109, 215-223.
THIEBAUD, CH. H. (1983). A reliable new cell marker in
Xenopus. Devi Biol. 98, 245-249.
TICKLE, C , SUMMERBELL, D. & WOLPERT, L. (1975).
Positional signalling and specification of digits in chick
limb morphogenesis. Nature, Lond. 254, 199-202.
(Accepted 4 December 1986)