The role of Hox genes during vertebrate limb development

Jozsef Zakany1 and Denis Duboule1,2

The potential role of Hox genes during vertebrate limb
development was brought into focus by gene expression
analyses in mice (P Dolle, JC Izpisua-Belmonte, H Falkenstein,
A Renucci, D Duboule, Nature 1989, 342:767772), at a time
when limb growth and patterning were thought to depend upon
two distinct and rather independent systems of coordinates;
one for the anterior-to-posterior axis and the other for the
proximal-to-distal axis (see D Duboule, P Dolle, EMBO J 1989,
8:14971505). Over the past years, the function and regulation
of these genes have been addressed using both
gain-of-function and loss-of-function approaches in chick and
mice. The use of multiple mutations either in cis-configuration in
trans-configuration or in cis/trans configurations, has
confirmed that Hox genes are essential for proper limb
development, where they participate in both the growth and
organization of the structures. Even though their molecular
mechanisms of action remain somewhat elusive, the results
of these extensive genetic analyses confirm that,
during the development of the limbs, the various
axes cannot be considered in isolation from each other and that
a more holistic view of limb development should prevail
over a simple cartesian, chess grid-like approach of these
complex structures. With this in mind, the functional input of
Hox genes during limb growth and development can now
be re-assessed.
Addresses
1
National Research Centre Frontiers in Genetics, Department of
Zoology and Animal Biology, University of Geneva, Sciences III, Quai
Ernest Ansermet 30, 1211 Geneva 4, Switzerland
2
National Research Centre Frontiers in Genetics, School of Life
Sciences, Ecole Polytechnique Federale, Lausanne (EPFL),
Switzerland
Corresponding author: Zakany, Jozsef (jozsef.zakany@zoo.unige.ch)

Current Opinion in Genetics & Development 2007, 17:359366

This review comes from a themed issue on
Pattern formation and developmental mechanism
Edited by Ross Cagan and Christine Hartmann
Available online 20th July 2007
0959-437X/$ see front matter
# 2007 Elsevier Ltd. All rights reserved.
DOI 10.1016/j.gde.2007.05.011

Introduction
A functional account of this gene family during limb
development is difficult to discuss without a global view
of their expression dynamics, in particular, for the two
most important players: the HoxA and HoxD gene clusters. Deletions of both the HoxB and HoxC clusters
www.sciencedirect.com

(Figure 1) indeed did not elicit any abnormal phenotype

in limbs [1,2].
During limb development Hoxa and Hoxd genes belonging to identical groups of paralogy (located at the same
respective positions in the clusters) show similar, though
not identical, expression domains [5,6]. These complex
patterns are laid down in a temporal manner, with
anterior genes (e.g. groups 1 and 2) activated earlier
than posterior genes (e.g. groups 11 and 12), reminiscent
of their temporal activation during trunk formation and
extension. In early limb buds, the result of this sequential
activation, as far as Hoxd genes are concerned, is a
progressive restriction of expressing cells towards the
posterior margin of the bud. This Russian dolls strategy
bears many similarities with what is observed during
trunk development, hence it was hypothesized that the
underlying molecular mechanisms had been co-opted
from the trunk into newly emerging limbs, in the course
of evolution [7,8].
At this early stage, Hoxa genes are expressed similarly, yet
the anterior-to-posterior (A/P) asymmetry in transcript
distribution is no longer observed at a later developmental stage, such that genes expressed posteriorly early on
subsequently display an all-distal expression during handplate formation. In the case of the HoxD cluster, the late
phase of expression maintains an A/P bias in the expression of those genes, which were early on preferentially
expressed at the posterior margin. Altogether, both Hoxd
and Hoxa genes are expressed in two waves (Figure 2). In
the first phase, the general rules governing their expression seem to be similar for both clusters and involve a
collinear regulation in time and space, which resembles
the strategy implemented in the trunk. By contrast, the
second phases are quite distinct and may have evolved
separately, after cluster duplications occurred.

Loss-of-function and the P/D axis

Loss-of-function mutations of some Hoxa and Hoxd genes,
alone or in combination, strongly impact upon limb
morphology (Figure 3), in particular for genes belonging
to groups 913, with patterning defects generally corresponding to the expression domains of the inactivated
gene(s), except in those distal areas where the function
of more posterior genes is prevalent (see below). The
alignment of the phenotypes along the P/D limb axis thus
reflects gene order along the chromosome, as predicted by
the expression domains.
Unlike the situation in the trunk, and in contrast to
what was originally expected, loss-of-function phenotypes
Current Opinion in Genetics & Development 2007, 17:359366

360 Pattern formation and developmental mechanism

Figure 1

Predominant role of HoxA and HoxD clusters during limb development. On top, the full Hox genes complement is shown (left), along with the
associated wild-type morphology (right). The various schemes below illustrate full cluster deletions. Only the removal or either HoxA or HoxD
leads to a detectable phenotype, which is not drastic and mostly affects the digital plate [1,2,3,4]. However, the combined deletion of both
HoxA and HoxD leads to an early arrest of limb growth [3], pointing to a large functional redundancy between these two clusters (S: stylopod,
comprising the humerus and defining the upper arm; Z: zeugopod, comprising the radius and ulna and defining the lower arm; A: autopod,
comprising the ensemble of carpus and digits).

Current Opinion in Genetics & Development 2007, 17:359366

www.sciencedirect.com

The role of Hox genes during vertebrate limb development Zakany and Duboule 361

Figure 2

Two phases of expression of Hoxa and Hoxd genes. In both clusters, genes are activated in time, following the same general collinear strategy,
with posterior genes (e.g. group 13) transcribed in progressively more posterior cells of the limb buds. During the second (late) phase, Hoxa
and Hoxd patterns are still quite comparable, but with obvious differences, suggesting that different regulations are now implemented.

cannot be interpreted as classical homeotic transformations. Instead, they systematically involve loss or
reduction of skeletal elements. This effect, already visible
with single mutation [9], is best appreciated when compound mutants of the same group are examined (Figure 3).
In the combined absence of group 11 genes, lower arms or

zeugopods are very short [10], just as in combined absence

of group 13 genes where the autopods are lacking [11].
Accordingly, compound mutants of group 10 genes truncate regions more proximal than those affected by group 11
[12], and compound mutants of group 9 genes affect
regions more proximal than that of group 10 [13].

Figure 3

Forelimb phenotypes of compound mutant mice. In the absence of both Hoxa13 and Hoxd13, stylopods and zeugopods are normal whereas
the autopods are devoid of digits. In the absence of group 11 genes, stylopods are normal, unlike zeugopods, which are truncated. Only minor
defects are seen in the autopods. In the concomitant absence of both HoxA and HoxD clusters, autopods and zeugopods are absent,
whereas a proximal third of the stylopods is formed. Therefore, extension of the limb along its entire length critically depends upon Hox gene
function. (S: stylopod, comprising the humerus and defining the upper arm; Z: zeugopod, comprising the radius and ulna and defining the lower
arm; A: autopod, comprising the ensemble of carpus and digits.)
www.sciencedirect.com

Current Opinion in Genetics & Development 2007, 17:359366

362 Pattern formation and developmental mechanism

In these severe compound mutant combinations, some

parts of the limbs remain virtually untouched, suggesting
a relative genetic independence with respect to Hox
function. This is nevertheless not absolute, as mutation
in genes involved in the zeugopod (e.g. Hoxd11) can also
affect the morphology of the autopod, provided group 13
and 12 genes are absent. This functional hierarchy, added
to the redundant functions of paralogous genes, made it
difficult to anticipate the consequences of eliminating all
HoxA and HoxD functions during limb development. In
such mice, only the most proximal third of the forelimb
stylopod was still present [3]. This strong phenotype
demonstrates the growth promoting role of HOX products in limb development, as was proposed after inactivation of Hoxd13 alone [9]. As far as developmental
mechanisms are concerned, limb growth relies on both
the apical ectodermal ridge (AER) and the zone of
polarizing activity (ZPA) (see [14] for recent review);
below, we discuss the relationships between these signalling centers and Hox genes function.

Gain-of-function and the A/P axis

Gain-of-function approaches in chick suggested that various HOX proteins exert different functions. While
HOXD11 increases digit length and number, group 13
proteins have an opposite effect, in particular in the zeugopod, where they induce alterations similar to a group 11
loss-of-function (reduction of the bony elements) [15,16].
This is in agreement with posterior prevalence, the fact
that posterior HOX proteins antagonize the function of
more anterior ones, an effect that is also seen with loss-offunction alleles where altered phenotypes are usually not
scored when a more posterior protein is present.
Gain-of-function experiments carried out in transgenic
mice lead to similar conclusions [17,18]. While both
zeugopod and stylopod reductions are induced by group
13 products, ectopic expression of group 11 genes primarily induced stylopod reductions. Studies using missense
mutations in the Hoxd13 homeodomain indicated that
domains other than the homeodomain itself mediate
these effects [19]. For instance, a strong effect is associated with the expansion of a polyalanine tract in both
Hoxa13 and Hoxd13 [20], probably arguing for the
importance of proteinprotein interactions in the molecular mechanisms underlying Hox gene functions.
The functional relevance of mis-expression studies in vivo
or in vitro is nevertheless hard to evaluate, and chromosome
re-arrangements in vivo offer a powerful alternative. In
mice, targeted meiotic recombination (TAMERE) [21]
was used to produce systematic re-arrangements at the
HoxD locus, and comparisons between various alleles
helped assigning particular function to some gene products,
as illustrated by the following examples. First, after
deletion of both Hoxd12 and Hoxd13 [22], Hoxd11 is upregulated in the autopod, leading to massive polydactyly.
Current Opinion in Genetics & Development 2007, 17:359366

Therefore, Hoxd11 stimulates digital condensations, a

function that is normally counteracted by the Hoxd12
and Hoxd13 products. Second, in the presence of a posterior
HoxD mini-cluster (Hoxd11Hoxd13), these three genes are
expressed ectopically, throughout the early developing
limb bud, leading to a double posterior limb [23]. In both
cases, gain-of-function effects involve the loss of the
anterior-to-posterior (A/P) polarity in the distal limb and
thus illustrate the role of Hox patterning along the A/P axis.
Therefore, while combined loss-of-functions generate
severe truncations along the P/D axis, gain-of-functions
point to their importance for patterning along the A/P axis.
A tentative explanation of this genetic conflation of developmental axes arises from the identification of some of the
target molecules and mechanisms.

Hox genes and the ZPA

ZPA function is mediated by the SHH signalling molecule,
produced by posterior limb bud cells [24]. Initial evidences
that Shh was regulated by Hox genes came from the overexpression of Hoxb8 [25] and Hoxd12 [17]. Recently, gainof-function and loss-of-function alleles have confirmed this
epistatic relationship. On the one hand, expression of the
posterior Hoxd11Hoxd13 genes in the early anterior limb
bud induced double posterior limbs because of mirrorimage Shh patterns [23]. On the other hand, deletions of
both HoxA and HoxD clusters resulted in the absence of Shh
expression in its normal domain [3]. Furthermore, genetic
complementation analysis using various HoxD deficiencies
indicated that genes belonging to paralogy groups 10, 12
and 13 (and probably 11) are capable of mediating Shh
activation [26]. From the ozd chick mutant, it is also known
that A/P asymmetric Hoxd11Hoxd13 expression is established in the absence of Shh [27]. This genetic interaction is
probably direct, since biochemical evidence [28] suggests
direct binding of HOX products to a limb regulatory
sequence [29] upstream Shh.
We conclude that Hox genes control Shh expression in
posterior limb bud cells, following the early phase of
collinear activation that restricts expression of the
relevant Hox genes posteriorly. Subsequently, Shh signalling will impact upon the late phase of Hox gene expression, in particular by skewing the expression domains
toward the posterior aspect, thus modulating the various
digital identities, an effect that is abrogated in the
absence of the Gli3 gene product. Consistently, inactivation of Shh results in a truncation [30] less severe than
the full HoxA-/D-truncation due to the persistence of the
early phase of Hox gene expression. The dual role of Hox
genes in both the A/P and the P/D axial systems are
largely explained by this single genetic interaction, which
makes these two axes interdependent on each other.

Hox genes and the AER

The early phase of Hox gene activation is not solely
dedicated to the activation of Shh. In the absence of both
www.sciencedirect.com

The role of Hox genes during vertebrate limb development Zakany and Duboule 363

HoxA and HoxD clusters, forelimb development is indeed

arrested much before it does when Shh function is abrogated on its own. In both cases, however, the effect is
mediated by an insufficiency of the AER, suggesting that
AER formation is an important outcome of Hox function
preceding induction of Shh transcription.
Also, internal HoxD cluster deletions assayed in the
absence of the HoxA cluster induced variable extent of
stylopod truncations [26], showing that Hoxd9 and Hoxd8
contribute significantly to proximal limb development,
even though these gene products cannot induce Shh
expression [26]. Interestingly, the additional proximal
defects observed in their absence are reminiscent of a
combined absence of Fgf8 and Fgf4 in the AER [31,32].
The genetic dependence of AER formation on both Fgf10
[33] and HoxA/D functions suggests that mesenchymal
Fgf synthesis may mediate Hox gene function [34].
From these data, the hypothesis is proposed (Figure 4)
whereby induction of both the AER and the ZPA involves
Hox genes function. While induction of the ZPA (Shh)
involves groups 1013, AER formation and maintenance
primarily involve groups 811 and originally require a
relatively low activity of groups 12 and 13, which tend to
counteract AER function [34]. Consequently, initial ZPA
induction also depends on the absence of group 12 and 13
products, since ZPA induction is promoted by the AER in
the form of a positive feedback loop [35]. Maximal group
12 and 13 expression inhibits AER activity, thereby
terminating distal patterning eventually. This inhibition
of AER function is less pronounced around the ZPA (Shh
stimulates ZPA maintenance), which may account for the
over-representation of ZPA-derived cells in the autopod
[36].
This view is supported by Fgf4/Fgf8 double mutants, where
both Fgf10 and Shh are detected, even if dramatically
reduced [31]. Also, in the absence of Shh, Fgf10 expression
is scored. Finally, the establishment of the double posterior
polarity observed in some HoxD mutant limb buds involves
symmetrical Fgf10 expression [34]. Furthermore, in the
very small Fgfr2 mutant limb buds, Fgf10 is present [37]
suggesting that its activation is largely Fgf signalling independent and intrinsic to mesenchymal cells. However, the
fact that the humerus is more developed in HoxA/D double
mutants than in Fgf10 mutants indicates that Fgf10 is not
under the exclusive control of these two gene clusters
[25,38]. It is equally clear that Hox function is not sufficient
for initial Fgf10 induction. Rather, it may contribute to a set
of cues including Tbx5 [39] as well as other components
such as RA [40,41] and Wnt [42] signalling pathways.

Targets from gene expression profiling

Hox target genes in limbs have been isolated mostly from
either Hoxa13 or Hoxd13 loss-of-function or gain-of-function, or from genetic alterations including these two
www.sciencedirect.com

Figure 4

A genetic model of Hox function in determining both AER and ZPA

activities during mammalian forelimb development. Owing to a
collinear sequence of activation, transcription of specific sets of Hox
genes (depicted in red) leads to temporal control of the target genes
Fgf10 and Shh, thus modulating the activity of key signalling centers,
AER and ZPA, respectively. In each of the three limb regions, the
stylopod, zeugopod and autopod, the genetically predominant
group is indicated by larger typeface. Any member of groups 811
is capable of inducing Fgf10 (encircled in green oval), whereas any
member of the overlapping subset of groups 1013 is capable of
inducing Shh (encircled in blue oval). During early budding,
preceding AER formation, groups 19 are sequentially activated all
across the emerging limb bud mesenchyme (see Figure 2). They
contribute to Fgf10 expression in all mesenchyme cells, which in
turn promotes AER induction. The phenotype of isolated group 9
loss-of-function illustrates this interaction, where the skeletal defects
are localized to the stylopod probably involving an early transitory
AER deficit. With the appearance of group 10 transcripts in posterior
limb bud cells, Shh is induced, marking the onset of the ZPA. Fgf10
transcript accumulation also becomes restricted to the posterior
limb bud mesenchyme, where it predominantly depends on group 10
and 11 genes. With continued distal outgrowth, a third phase of Fgf10
pattern appears as simultaneous posterior/proximal and posterior/
distal domains. The establishment of the posterior/distal Fgf10
expression domain is coincident with the appearance of the more
distal expression domains of group 10 and 11 genes (see Figure 2)
and a distal shift in Shh expression extending into the presumptive
autopod. This phase is equivalent to the AER/ZPA positive feedback
mechanism, and is eventually responsible for the formation of distal
stylopod and the zeugopod, which together give rise to the largest
proportion of distal outgrowth. The phenotype of isolated group 11
loss-of-function illustrates this interaction, where the skeletal defects
are localized to the zeugopod (see Figure 3), probably involving a
combined transitory AER and ZPA deficit. Upon activation of
groups 12 and 13, both maximal in the posterior/distal region, Shh
and ZPA maintenance continues. By contrast, Fgf10 expression
becomes attenuated because of inhibition of group 10 and 11
functions by groups 12 and 13 via posterior prevalence. Full levels
of group 12 and 13 proteins inactivate group 10 and 11 functions,
compatible with the decays of both AER and ZPA. The phenotype
of isolated group 13 loss-of-function illustrates this interaction,
where the skeletal defects are localized to the autopod (see Figure 3).
The truncations and gene expression changes seen in simultaneous
absence of the HoxA and HoxD clusters highlighted the signalling
defects in both Fgf10-mediated AER function and Shh-mediated
ZPA function.

genes. Amongst many potential candidates, an interesting

set of target genes seems to emerge from various
approaches, which are involved in endochondral bone
formation, that is, a process where Hox genes expectedly
Current Opinion in Genetics & Development 2007, 17:359366

364 Pattern formation and developmental mechanism

play a role because of both their loss-of-function phenotypes and their expression in proliferating chondrocytes
[43]. Overexpression of Hoxa13 thus affects the expression of Enpp2, an enzyme produced in precartilaginous
condensations that modulates cell motility [44].
Loss-of-function of Hoxa13 was also reported to modulate
expression of bone morphogenetic proteins like BMP2
and BMP7 [45], and exogenous administration of these
proteins could partially restore the characteristic Hoxa13
mutant digit defect. However, the loss-of-function of
these potential target genes failed to reveal a clear and
direct association with Hox gene function [46]. Another
gene product possibly modulated by Hoxd gene function
in developing limb buds is SHOX2, the mouse cognate of
the human short stature gene [47], which may be
repressed by distal Hoxd genes, even though genetic
redundancy makes a clear validation of this point difficult.
Interestingly, another confirmed target gene is Sox9, a key
transcription factor for the commitment of mesenchymal
cells into the chondrogenic fate [48].

Functional evolution
It is likely that collinear Hox genes regulation was originally deployed to specify positions along the main body
axis of an ancestral bilaterian animal. In mammals, this
genetic system is at work in a variety of organs or
structures such as the limbs, the central nervous system,
the vertebral column, the gut, the genitals, the excretory
apparatus and the hematopoietic system. These successive co-options may have been tolerated because of a
higher functional flexibility following large-scale cluster
duplications [49]. Here again, the evolution of tetrapod
appendages from an ancestral fin-like structure (the finto-limb transition) may have involved important regulatory modifications affecting both the HoxA and HoxD
clusters.
In early gnathostome, such as catshark, the Hox system
was probably implemented in median fin to mark-off the
A/P polarity. These Hox gene expression patterns in fins
are reminiscent of the early limb bud pattern seen in
mouse or chick embryo (the early phase), perhaps reflecting a stage, in limb evolution, at which Hox-dependent A/
P patterning emerged in the absence of a P/D patterning
role [50]. In the skate, another gnathostome, Shh expression, is not detectable in early postbudding fins [51]. In
aggregate, these two sets of data suggest that the emergence of A/P specification in an ancestral limb bud did not
require Shh function. Studies of HoxA and HoxD clusters
in bony fish early pectoral fins [52,53] are also consistent
with the priority of Hox collinear function over Shh for the
establishment of the A/P polarity. The AER is present in
both shark and skate (as well as in cyclostomes) and, at
least in skate, retinoic acid signalling, a likely regulator of
coordinated Hox cluster transcription, is operative [51].
Therefore, an important step in the elaboration of the
Current Opinion in Genetics & Development 2007, 17:359366

tetrapod limb may have been the secondary acquisition of

Shh control by posterior Hox genes and the emergence of
the second phase of Hox gene expression. Whether this
was instrumental in the distalization of Hox function in
basal tetrapods and when did an Fgf10 orthologue enter
into the target set remain open for the moment.
It is thus probable that the proximo-distal extension and
patterning role of Hox gene clusters involved a second
phase of Hox function, downstream Shh signalling, to
control distal outgrowth of the limb and specification of
the terminal elements. Although this could have occurred
at the origin of tetrapods, recent analyses of developing
chondrichtyan pectoral fins have shown both Shh functionality and the most distal expression of posterior Hoxd
genes [51]. In this latter case, although some more functional analyses are required, it was concluded that both
phases of Hox gene expression, with the expected Shh
function in-between, already existed in fish pectoral fins
and was subsequently abrogated from teleostean fishes,
which evolved a large exoskeleton at the expense of a
distal endoskeleton. In this scenario, the entire molecular
machinery necessary for a tetrapod limb to develop was
already present in fishes, ready to accompany a progressive shift towards the emergence of paired limbs [51].

Conclusions
The function and regulation of Hox genes are admittedly
best understood in both developing limbs and rhombencephalic crest cells [54]. How much of what we have
learned by studying the limbs can be transposed to the
patterning of the major body axis. As far as collinear
regulation is concerned, the early phase of Hox expression
in limb buds presents many similarities with the mechanism that dictates progressive Hox activation during
trunk extension, at least in its most caudal parts [7,26].
By contrast, the late phase of Hox expression in limbs is
probably an ad hoc regulatory innovation, which accompanied the evolution of the most distal limb pieces.
Regarding Hox gene function, the parallel with the main
body axis is more difficult to establish. As discussed above,
although both Fgf10 and Shh are plausible mediators of Hox
function during the early and late growth of limb buds,
respectively, none of these latter two genes appears to be of
key importance for the A/P patterning of the main body
axis. In contrast to the limb bud, Shh expression in the main
axis is not restricted to posterior Hox genes expression
domains, even though the maximal redundancy of the
system, in the trunk, makes a genetic demonstration of
this point virtually impossible. Likewise, Fgf10 is preeminently involved with budding morphogenesis and
seems to be rather inert in axial patterning.
Altogether the similarities (in regulation) and differences
(in functions) observed for Hox genes between the main
body axis and the limb patterning illustrate the rather
www.sciencedirect.com

The role of Hox genes during vertebrate limb development Zakany and Duboule 365

nonspecific functional outcome of this genetic system and

its flexibility to be co-opted, at various times, in different
contexts. Although the central molecular mechanism (i.e.
to dispatch a set of molecular cues in time and space) is
used essentially unmodified, both the upstream regulators and the downstream effectors may be different and
adapted to the local context.

References and recommended reading

Papers of particular interest, published within the annual period of
review, have been highlighted as:
 of special interest
 of outstanding interest
1.

Suemori H, Noguchi S: Hox C cluster genes are dispensable for

overall body plan of mouse embryonic development. Dev Biol
2000, 220:333-342.

2.

Medina-Martinez O, Bradley A, Ramirez-Solis R: A large targeted

deletion of Hoxb1Hoxb9 produces a series of single-segment
anterior homeotic transformations. Dev Biol 2000, 222:71-83.

3.


Kmita M, Tarchini B, Zakany J, Logan M, Tabin CJ, Duboule D:

Early developmental arrest of mammalian limbs lacking HoxA/
HoxD gene function. Nature 2005, 435:1113-1116.
The early lethality induced by deleting the HoxA cluster was overcome by
using Prx1Cre deleter mice in a conditional approach allowing for multiple
clusters deletions.
4.

Zakany J, Kmita M, Alarcon P, de la Pompa JL, Duboule D:

Localized and transient transcription of Hox genes suggests a
link between patterning and the segmentation clock. Cell 2001,
106:207-217.

5.

Haack H, Gruss P: The establishment of murine Hox-1

expression domains during patterning of the limb.
Dev Biol 1993, 157:410-422.

6.

Nelson CE, Morgan BA, Burke AC, Laufer E, DiMambro E,

Murtaugh LC, Gonzales E, Tessarollo L, Parada LF, Tabin C:
Analysis of Hox gene expression in the chick limb bud.
Development 1996, 122:1449-1466.

7.

Tarchini B, Duboule D: Control of Hoxd genes collinearity

during early limb development. Dev Cell 2006, 10:93-103.

8.

Shubin N, Tabin C, Carroll S: Fossils, genes and the evolution of

animal limbs. Nature 1997, 388:639-648.

9.

Dolle P, Dierich A, LeMeur M, Schimmang T, Schuhbaur B,

Chambon P, Duboule D: Disruption of the Hoxd-13 gene
induces localized heterochrony leading to mice with neotenic
limbs. Cell 1993, 75:431-441.

10. Davis AP, Witte DP, Hsieh-Li HM, Potter SS, Capecchi MR:
Absence of radius and ulna in mice lacking hoxa-11 and
hoxd-11. Nature 1995, 375:791-795.
11. Fromental-Ramain C, Warot X, Messadecq N, LeMeur M, Dolle P,
Chambon P: Hoxa-13 and Hoxd-13 play a crucial role in the
patterning of the limb autopod. Development 1996,
122:2997-3011.
12. Wellik DM, Capecchi MR: Hox10 and Hox11 genes are required
to globally pattern the mammalian skeleton. Science 2003,
301:363-367.
13. Fromental-Ramain C, Warot X, Lakkaraju S, Favier B, Haack H,
Birling C, Dierich A, Dolle P, Chambon P: Specific and redundant
functions of the paralogous Hoxa-9 and Hoxd-9 genes in
forelimb and axial skeleton patterning. Development 1996,
122:461-472.
14. Tickle C: Making digit patterns in the vertebrate limb.
Nat Rev Mol Cell Biol 2006, 7:45-53.
15. Morgan BA, Izpisua-Belmonte JC, Duboule D, Tabin CJ: Targeted
misexpression of Hox-4, 6 in the avian limb bud causes
apparent homeotic transformations. Nature 1992,
358:236-239.
www.sciencedirect.com

16. Goff DJ, Tabin CJ: Analysis of Hoxd-13 and Hoxd-11

misexpression in chick limb buds reveals that Hox genes
affect both bone condensation and growth. Development 1997,
124:627-636.
17. Knezevic V, De Santo R, Schughart K, Huffstadt U, Chiang C,
Mahon KA, Mackem S: Hoxd-12 differentially affects preaxial
and postaxial chondrogenic branches in the limb and
regulates Sonic hedgehog in a positive feedback loop.
Development 1997, 124:4523-4536.
18. Williams ME, Lehoczky JA, Innis JW: A group 13 homeodomain is

neither necessary nor sufficient for posterior prevalence in the
mouse limb. Dev Biol 2006, 297:493-507.
An important observation from such mis-expression studies is the dispensability of the homeodomain for posterior HOX proteins to inhibit
more anterior functions. Because this capacity is independent of DNA
binding, it may be mediated by direct proteinprotein interactions and
does not importantly involve transcriptional regulation.
19. Caronia G, Goodman FR, McKeown CM, Scambler PJ,
Zappavigna V: An I47L substitution in the HOXD13
homeodomain causes a novel human limb malformation by
producing a selective loss of function. Development 2003,
130:1701-1712.
20. Utsch B, Becker K, Brock D, Lentze MJ, Bidlingmaier F, Ludwig M:

A novel stable polyalanine [poly(A)] expansion in the HOXA13
gene associated with hand-foot-genital syndrome: proper
function of poly(A)-harbouring transcription factors
depends on a critical repeat length? Hum Genet 2002,
110:488-494.
The expansion of a reiterated alanine sequence, located outside the
homeodomains, causes a range of semidominant inherited malformations in human patients. From a genetic perspective, these triplet expansions seem to produce proteins that not only loose their function but also
impact upon the functions of other Hox gene products, somewhat related
to a dominant negative effect of certain transcription factors.
21. Herault Y, Rassoulzadegan M, Cuzin F, Duboule D: Engineering
chromosomes in mice through targeted meiotic
recombination (TAMERE). Nat Genet 1998, 20:381-384.
22. Kmita M, Fraudeau N, Herault Y, Duboule D: Serial deletions and
duplications suggest a mechanism for the collinearity of Hoxd
genes in limbs. Nature 2002, 420:145-150.
23. Zakany J, Kmita M, Duboule D: A dual role for Hox genes in limb
anteriorposterior asymmetry. Science 2004, 304:1669-1672.
24. Riddle RD, Johnson RL, Laufer E, Tabin C: Sonic hedgehog
mediates the polarizing activity of the ZPA. Cell 1993,
75:1401-1416.
25. Charite J, de Graaff W, Shen S, Deschamps J: Ectopic
expression of Hoxb-8 causes duplication of the ZPA in the
forelimb and homeotic transformation of axial structures.
Cell 1994, 78:589-601.
26. Tarchini B, Duboule D, Kmita M: Regulatory constraints in the
evolution of the tetrapod limb anteriorposterior polarity.
Nature 2006, 443:985-988.
27. Ros MA, Dahn RD, Fernandez-Teran M, Rashka K, Caruccio NC,
Hasso SM, Bitgood JJ, Lancman JJ, Fallon JF: The chick
oligozeugodactyly (ozd) mutant lacks sonic hedgehog
function in the limb. Development 2003, 130:527-537.
28. Capellini TD, Di Giacomo G, Salsi V, Brendolan A, Ferretti E,

Srivastava D, Zappavigna V, Selleri L: Pbx1/Pbx2 requirement for
distal limb patterning is mediated by the hierarchical control of
Hox gene spatial distribution and Shh expression.
Development 2006, 133:2263-2273.
The binding to the Shh limb enhancer is observed even for those HOXD
products that are unable to elicit a transcriptional response from Shh in
vivo (for example Hoxd9), suggesting that various HOX products may
have either activating or repressing properties when bound to this target
site.
29. Sagai T, Hosoya M, Mizushina Y, Tamura M, Shiroishi T:
Elimination of a long-range cis-regulatory module causes
complete loss of limb-specific Shh expression and truncation
of the mouse limb. Development 2005, 132:797-803.
30. Chiang C, Litingtung Y, Harris MP, Simandl BK, Li Y, Beachy PA,
Fallon JF: Manifestation of the limb prepattern: limb
Current Opinion in Genetics & Development 2007, 17:359366

366 Pattern formation and developmental mechanism

development in the absence of sonic hedgehog function.

Dev Biol 2001, 236:421-435.
31. Boulet AM, Moon AM, Arenkiel BR, Capecchi MR: The roles of
Fgf4 and Fgf8 in limb bud initiation and outgrowth.
Dev Biol 2004, 273:361-372.
32. Sun X, Mariani FV, Martin GR: Functions of FGF signalling from
the apical ectodermal ridge in limb development. Nature 2002,
418:501-508.
33. Sekine K, Ohuchi H, Fujiwara M, Yamasaki M, Yoshizawa T,
Sato T, Yagishita N, Matsui D, Koga Y, Itoh N et al.: Fgf10 is
essential for limb and lung formation. Nat Genet 1999,
21:138-141.
34. Zakany J, Zacchetti G, Duboule D: Interactions between Hoxd
and Gli3 genes control the limb apical ectodermal ridge via
Fgf10. Dev Biol 2007, 306:883-893.
35. Niswander L, Jeffrey S, Martin GR, Tickle C: A positive feedback
loop coordinates growth and patterning in the vertebrate limb.
Nature 1994, 371:609-612.
36. Scherz PJ, Harfe BD, McMahon AP, Tabin CJ: The limb bud Shh
Fgf feedback loop is terminated by expansion of former ZPA
cells. Science 2004, 305:396-399.
37. Xu X, Weinstein M, Li C, Naski M, Cohen RI, Ornitz DM, Leder P,
Deng C: Fibroblast growth factor receptor 2 (FGFR2)-mediated
reciprocal regulation loop between FGF8 and FGF10 is
essential for limb induction. Development 1998,
125:753-765.
38. Stratford TH, Kostakopoulou K, Maden M: Hoxb-8 has a role in
establishing early anteriorposterior polarity in chick forelimb
but not hindlimb. Development 1997, 124:4225-4234.

of HOX group 13 transcription factors with and without

monomeric DNA binding capability. Dev Biol 2005,
279:462-480.
45. Knosp WM, Scott V, Bachinger HP, Stadler HS: HOXA13
regulates the expression of bone morphogenetic proteins 2
and 7 to control distal limb morphogenesis. Development 2004,
131:4581-4592.
46. Bandyopadhyay A, Tsuji K, Cox K, Harfe BD, Rosen V, Tabin CJ:
Genetic analysis of the roles of BMP2, BMP4, and BMP7 in limb
patterning and skeletogenesis. PLoS Genet 2006, 2:e216.
47. Cobb J, Dierich A, Huss-Garcia Y, Duboule D: A mouse model for

human short-stature syndromes identifies Shox2 as an
upstream regulator of Runx2 during long-bone development.
Proc Natl Acad Sci USA 2006, 103:4511-4515.
Shox2 is expressed early on, proximally, surrounding prechondrogenic
condensation as well as inside, yet weakly. Its inactivation leads to the
disappearance of Runx2 positive cells, thus abrogating chondrocyte
differentiation and, consequently, long bone development.
48. Akiyama H, Stadler HS, Martin JF, Ishii TM, Beachy PA,

Nakamura T, de Crombrugghe B: Misexpression of
Sox9 in mouse limb bud mesenchyme induces
polydactyly and rescues hypodactyly mice. Matrix Biol 2007,
26:224-233.
Sox9 expression is lost in Hoxa13Hd mutant mice in digit one, which is the
most sensitive skeletal element in that mutant. Forced Sox9 expression in
wild-type mice stimulates cell proliferation leading to polydactyly and
partially rescues the digit deficit in Hoxa13Hd mutants.
49. Wagner GP, Amemiya C, Ruddle F: Hox cluster duplications and
the opportunity for evolutionary novelties. Proc Natl Acad Sci
USA 2003, 100:14603-14606.

39. Hasson P, Del Buono J, Logan MP: Tbx5 is dispensable for

forelimb outgrowth. Development 2007, 134:85-92.

50. Freitas R, Zhang G, Cohn MJ: Evidence that mechanisms of fin

development evolved in the midline of early vertebrates.
Nature 2006, 442:1033-1037.

40. Niederreither K, Vermot J, Schuhbaur B, Chambon P, Dolle P:

Embryonic retinoic acid synthesis is required for forelimb
growth and anteroposterior patterning in the mouse.
Development 2002, 129:3563-3574.

51. Dahn RD, Davis MC, Pappano WN, Shubin NH: Sonic hedgehog
function in chondrichthyan fins and the evolution of
appendage patterning. Nature 2007, 445:311-314.

41. Yashiro K, Zhao X, Uehara M, Yamashita K, Nishijima M, Nishino J,

Saijoh Y, Sakai Y, Hamada H: Regulation of retinoic acid
distribution is required for proximodistal patterning and
outgrowth of the developing mouse limb. Dev Cell 2004,
6:411-422.

52. Metscher BD, Takahashi K, Crow K, Amemiya C, Nonaka DF,

Wagner GP: Expression of Hoxa-11 and Hoxa-13 in the
pectoral fin of a basal ray-finned fish, Polyodon spathula:
implications for the origin of tetrapod limbs. Evol Dev 2005,
7:186-195.

42. Hill TP, Taketo MM, Birchmeier W, Hartmann C: Multiple roles of

mesenchymal beta-catenin during murine limb patterning.
Development 2006, 133:1219-1229.

53. Sordino P, van der Hoeven F, Duboule D: Hox gene expression in

teleost fins and the origin of vertebrate digits. Nature 1995,
375:678-681.

43. Zakany J, Duboule D: Synpolydactyly in mice with a targeted

deficiency in the HoxD complex. Nature 1996, 384:69-71.

54. Serpente P, Tumpel S, Ghyselinck NB, Niederreither K,

Wiedemann LM, Dolle P, Chambon P, Krumlauf R, Gould AP:
Direct crossregulation between retinoic acid receptor {beta}
and Hox genes during hindbrain segmentation. Development
2005, 132:503-513.

44. Williams TM, Williams ME, Kuick R, Misek D, McDonagh K,

Hanash S, Innis JW: Candidate downstream regulated genes

Current Opinion in Genetics & Development 2007, 17:359366

www.sciencedirect.com