You are on page 1of 11

Clin Genet 2001: 59: 111

Printed in Ireland. All rights reser6ed

Mini Review

Human HOX gene mutations


Goodman FR, Scambler PJ. Human HOX gene mutations.
Clin Genet 2001: 59: 1 11. Munksgaard, 2001
HOX genes play a fundamental role in the development of the vertebrate central nervous system, axial skeleton, limbs, gut, urogenital tract
and external genitalia, but it is only in the last 4 years that mutations
in two of the 39 human HOX genes have been shown to cause congenital malformations: HOXD13, which is mutated in synpolydactyly, and
HOXA13, which is mutated in Hand-Foot-Genital syndrome. Here we
review the mutations already identified in these two genes, consider
how these mutations may act, and discuss the possibility that further
mutations remain to be discovered both in developmental disorders and
in cancer.

FR Goodman and PJ Scambler


Molecular Medicine Unit, Institute of Child
Health, London, UK
Key words: Hand-Foot-Genital syndrome
HOX genes HOXA13 HOXD13 limb
malformations polyalanine tract
expansions synpolydactyly
Corresponding author: Dr FR Goodman,
Room 213, Molecular Medicine Unit, Institute of Child Health, 30 Guilford Street,
London WC1N 1EH, UK. Tel: +44 20
7242 9789 ext: 2432; fax: +44 20 7404
6191; e-mail: fgoodman@hgmp.mrc.ac.uk
Received 13 July 2000, revised and accepted for publication 8 August 2000

The HOX genes encode a highly conserved family


of transcription factors, which play a fundamental
role in embryonic morphogenesis. First discovered
in the fruit fly, Drosophila, they were later found to
be structurally and functionally conserved
throughout the animal kingdom (1). In humans, as
in most vertebrates, there are 39 HOX genes organised into four clusters named HOXA through
HOXD, which are believed to have arisen from a
single ancestral cluster by duplication and divergence. Each cluster is located on a different chromosome, and contains 9 11 genes. In general, the
order of the genes within each cluster corresponds
closely to their temporal and spatial expression
patterns during development: genes at the 3% end of
each cluster are expressed early, in anterior and
proximal regions, whereas genes at the 5% end of
each cluster are expressed later, in posterior and
distal regions. HOX proteins help pattern the developing embryo along both the primary (rostrocaudal) and secondary (limb and genital bud) body
axes. They therefore play an important role in the
development of the central nervous system, axial
skeleton, gut and urogenital tract, as well as the
limbs and external genitalia (2). At the molecular
level, they are believed to act by controlling the
transcription of specific sets of target genes, although few such genes have yet been identified. All
HOX proteins bind specific DNA sequences near
their target genes via a highly conserved 60 amino

acid DNA-binding domain, the homeodomain,


which is encoded by a 180-bp sequence element,
the homeobox. Most, if not all, are also thought to
interact with specific DNA-binding partners, forming multimeric protein complexes, which may activate or repress transcription of their targets (3).
Ten years ago, many researchers believed that
HOX gene mutations would not be found in humans. Some held that mutations in genes of such
fundamental importance would cause severe developmental defects (as they do in Drosophila), resulting in early intra-uterine death. Others pointed to
the probable functional redundancy among vertebrate HOX genes, and suggested that mutations in
any one HOX gene would cause, at most, minor
abnormalities, which would escape medical attention. In fact, mice carrying targeted mutations in
individual HOX genes were subsequently found to
have clear but generally non-lethal phenotypes, but
it was not until 1996 that a human malformation
syndrome was first shown to be caused by mutations in a HOX gene. This syndrome was synpolydactyly (SPD), and the gene in question was
HOXD13 (4). In the following year, Hand-FootGenital syndrome (HFGS) was shown to be caused
by a mutation in the closely related HOXA13 gene
(5). These two genes are still the only HOX genes
known to be mutated in human disorders. Both lie
at the 5% end of their respective clusters, and play
important roles in the development of the distal
1

Goodman and Scambler

limbs and the lower urogenital tract. In this review,


we first describe the mutations already discovered

in these genes and then consider how these mutations may act, and whether further such mutations
remain to be found.

Synpolydactyly and HOXD13

Fig. 1. Limb abnormalities caused by HOXD13 mutations. A


and B: Photographs of a child heterozygous for a 9-residue
polyalanine tract expansion showing 3/4 finger and 4/5 toe
syndactyly, both with digit duplication in the syndactylous
webs. C: Radiograph of a child heterozygous for a 7-residue
polyalanine tract expansion showing 4th finger duplication
(broad proximal phalanx, bifid middle phalanx) and soft tissue
syndactyly between the 3rd and duplicated 4th fingers together
with 5th finger clinodactyly (hypoplastic middle phalanx). D:
Radiograph of a child heterozygous for an 8-residue polyalanine tract expansion showing broad 5th metatarsal and duplication of the phalanges of the 5th toes. E: Radiograph of a
child heterozygous for a 14 residue polyalanine tract expansion
showing 3rd finger duplication (all phalanges, bifid 3rd
metacarpal) with soft tissue syndactyly between the 4th and
duplicated 3rd fingers, 5th finger camptodactyly (hypoplastic
middle phalanx) and an enlarged capitate bone. F: Radiograph
of an adult heterozygous for a 1-bp deletion in the homeobox
(Fig. 2C) showing broad hallux, attempted duplication of the
2nd metatarsal in the 1st web space and middle phalanx
hypoplasia of toes 2 to 5.

SPD (MIM 186000) is a rare, dominantly-inherited


limb malformation in which there is a distinctive
combination of syndactyly and polydactyly (Fig.
1A,B,C,D). Typically, patients have 3/4 finger and
4/5 toe syndactyly, with partial or complete digit
duplication in the syndactylous web. There may
also be clinodactyly, camptodactyly or brachydactyly of the 5th fingers, and variable syndactyly
of the 2nd to 5th toes, often with middle phalanx
hypoplasia. Incomplete penetrance and variable
expressivity are common. One to all four limbs can
be involved, and the severity of involvement ranges
from partial skin syndactyly to complete reduplication of a digit, extending as far proximally as the
metacarpals/tarsals.
The clue to the molecular basis of SPD came
from studies of a remarkable family from an isolated Turkish village, with 182 affected living individuals, and a history going back over at least 140
years (6). The SPD locus was mapped to chromosome 2q31, where the HOXD gene cluster is located (7), and mutations in HOXD13 were
subsequently identified in three affected American
families (4). Moreover, these mutations were of a
kind never seen before. In normal individuals,
exon 1 of HOXD13 contains an imperfect trinucleotide repeat sequence encoding a 15-residue
polyalanine tract (Fig. 2A,B). Each affected family
harboured an expansion of this repeat, resulting in
an additional 7, 8 or 10 alanine residues, respectively (Fig. 2E). Similar expansions have now been
reported in 22 further SPD families (810), and
appear to have remained stable in size over at least
seven generations, consistent with the observed
absence of genetic anticipation.
A study of families with different-sized expansions has since found that the larger the expansion,
the more complete the penetrance, and the more
severe the phenotype (9). Affected individuals from
a family with a 14-alanine expansion, the largest
discovered so far, have the most severe limb abnormalities, including involvement of the first digits
and distal carpals (Fig. 1E). Affected males from
this family also have hypospadias, a genital malformation not usually associated with SPD, but
consistent with HOXD13 expression in the developing genital tubercle. The few reported SPD homozygotes show the same genotypephenotype
correlation. Thus, a patient homozygous for a 7-

Human HOX gene mutations

Fig. 2. Wild-type and mutant HOXD13. A: Genomic structure of HOXD13, showing sites of reported mutations. D/F= deletion
and frameshift. Exp = expansion. B: Wild-type HOXD13 protein. C and D: Predicted proteins resulting from a 1-bp deletion in the
homeobox (D/F1) and a 14 bp deletion in exon 1 (D/F2), respectively. The former would contain the first 278 wild-type amino acids
followed by 33 novel amino acids, and would lack the last 49 amino acids of the homeodomain. The latter would contain the first
107 wild-type amino acids followed by 115 novel amino acids, and would lack the entire homeodomain. E: Predicted protein
resulting from 7- to 14-residue expansions of the polyalanine tract.

alanine expansion has very short fingers, 3/4/5


syndactyly, short round metacarpals and absent
trapezium (4), but 7 patients homozygous for a
9-alanine expansion have much more severe abnormalities, with cats paw hands containing six to
eight rudimentary digits, complete soft tissue syndactyly and malformation/shortening of all
metacarpals and carpals (11).
A subtly different form of SPD found in two
unrelated families is caused by two intragenic deletions in HOXD13, affecting the homeobox in the
one family and exon 1 in the other (12) (Fig.
2A,C,D). All deletion carriers from both families
share a novel foot malformation, consisting of
partial duplication of the 2nd metatarsals in the 1st
web spaces, and often also of the 4th metatarsals in
the 4th web spaces, together with broad halluces
and short 1st metatarsals (Fig. 1F). These abnormalities are not found in patients with SPD caused
by HOXD13 polyalanine tract expansions, and
extend as far anteriorly as the 1st digits, which are
not normally affected in SPD. Some deletion carriers from these two families also have 3/4 synpoly-

dactyly in the hands and 4/5 synpolydactyly in the


feet, but only at low penetrance.
Hand-Foot-Genital syndrome and HOXA13

Hand-Foot-Genital syndrome (MIM 140000) is


another rare, dominantly-inherited condition, in
which distal limb abnormalities (Fig. 3AD) are
accompanied by malformations of the lower genito-urinary tract. In the limbs, the most striking
feature is first digit hypoplasia, producing short,
proximally placed thumbs with hypoplastic thenar
eminences and short, medially deviated halluces.
There is also ulnar deviation of the 2nd fingers,
clinodactyly/brachydactyly of the 5th fingers,
brachydactyly of the 2nd to 5th toes, and delayed
ossification, fusion and shortening of the carpals
and tarsals. These abnormalities are fully penetrant, bilateral and symmetrical with little variation in severity. By contrast, the genito-urinary
tract malformations are incompletely penetrant
and variably severe. Genital abnormalities include
hypospadias in males and Mullerian duct fusion
3

Goodman and Scambler

defects in females. The latter range from isolated


longitudinal vaginal septum to double uterus with
double cervix, and have led to foetal loss or neonatal death in at least three families (13 15). Urinary

Fig. 3.

abnormalities include ectopic ureteric orifices,


vesico-ureteric reflux and pelvi-ureteric junction
obstruction,
and
can
lead
to
chronic
pyelonephritis, renal insufficiency and renal transplant (1416).
The clue to the molecular basis of HFGS came
from studies of the spontaneous mouse mutation
Hypodactyly (Hd). Heterozygous Hypodactyly
mice (Hd/ +) have distal limb abnormalities virtually identical to those of HFGS patients, including
hypoplastic 1st digits, short 2nd and 5th middle
phalanges, and fusions and/or delayed ossification
of many of the carpals and tarsals (17). Initial
mapping of the Hd locus to the vicinity of the
mouse HOXA gene cluster (18) was followed by
the identification of a mutation in HOXA13 (17)
(Fig. 4G). The striking similarity of the limb abnormalities in Hd/ + mice and HFGS patients
then prompted an examination of the HOXA13
gene in the family in which HFGS had first been
described (13), leading to the discovery that affected individuals were indeed heterozygous for a
nonsense mutation in the homeobox (5) (Fig.
4A,C).
Five further mutations in HOXA13 have recently been identified in families with HFGS (19).
Three are also nonsense mutations, two in exon 1
and the third in the homeobox (Fig. 4A,DF). The
fourth new mutation, however, is a polyalanine
tract expansion, very similar those in HOXD13
that cause SPD. Exon 1 of HOXA13 contains three
imperfect trinucleotide repeat sequences encoding
polyalanine tracts, which in normal individuals are
14, 12 and 18 residues long respectively (Fig.
4A,B). The new mutation expands the last of these
tracts by an additional 8 alanine residues (Fig.
4H), and has remained stable in size over at least
three generations. All the HOXA13 mutations
mentioned so far produce a phenotype typical of
HFGS. The fifth new mutation, however, produces
Fig. 3. Limb abnormalities caused by HOXA13 mutations. A
and B: Photographs of a child heterozygous for a nonsense
mutation in exon 1 (Fig. 4F) showing short proximally-placed
thumb, ulnar deviation of the index finger, 5th finger
clinodactyly and short medially-deviated hallux. C and D:
Radiographs of the same child at age 8 years, showing short
1st metacarpal and metatarsal, small pointed 1st distal phalanges, hypoplastic 2nd and 5th middle phalanges in the hand,
absent ossification of the 3rd to 5th middle phalanges and
hypoplasia/absence of the 2nd to 5th distal phalanges in the
foot, and delayed ossification of the carpal centres. E and F:
Photographs of a child heterozygous for a missense mutation
in the homeobox (Fig. 4I) showing extremely small thumb and
absent hallux. G and H: Radiographs of the same child at age
5 years, showing extremely short 1st metacarpal, absent 1st
metatarsal apart from rudimentary base, hypoplasia/absence
of all middle phalanges, delayed ossification of the carpal
centres and abnormal tarsals.

Human HOX gene mutations

a more severe limb phenotype, including extremely


short thumbs, absent halluces and marked hypoplasia of all middle phalanges (Fig. 3E H). This is

a missense mutation in the homeobox, predicted to


replace asparagine 51 of the homeodomain by a
histidine (Fig. 4A,I).

Fig. 4. Wild-type and mutant HOXA13. A: Genomic structure of HOXA13, showing sites of reported mutations. X =nonsense
mutation. Exp = expansion. M =missense mutation. B: Wild-type HOXA13 protein. C F: Predicted proteins resulting from the
four nonsense mutations. X1 and X2, both in the homeobox, would remove the last 20 and 24 amino acids respectively, including
three of the four key residues responsible for contacting target DNA in the case of X1, and all four key residues in the case of X2.
X3 and X4, both in exon 1, would remove the last 193 and 253 amino acids, respectively including the entire homeodomain. G:
Mutant HOXA13 protein in the Hypodactyly mouse, caused by a 50-bp deletion in exon 1. The first 25 wild-type amino acids are
followed by 275 novel amino acids. H: Predicted protein resulting from 8-residue expansion of the third polyalanine tract. I:
Predicted protein resulting from missense mutation in the homeobox, which replaces asparagine 51 of the homeodomain by
histidine.

Goodman and Scambler


Hemizygosity for the HOXA and HOXD gene
clusters

A recent report describes a patient with a de no6o


interstitial deletion on chromosome 7p14 encompassing the entire HOXA cluster (20). This patient
has distal limb and genital abnormalities typical of
HFGS, as well as facial dysmorphism, mild mental
retardation, a persistent patent ductus arteriosus
and velopharyngeal insufficiency. His HFGS phenotype probably reflects haploinsufficiency for
HOXA13, while his cardiac and velopharyngeal
abnormalities resemble those found in mice homozygous for a targeted deletion in HOXA3 (21),
and may well be due to haploinsufficiency for
HOXA3.
The consequences of hemizygosity for the
HOXD cluster are less clear. At least 30 patients
have been reported with cytogenetically visible
deletions involving chromosome 2q31 (2226).
Most have multiple malformations, including
limb abnormalities (camptodactyly, a wide sandal gap with syndactyly/brachydactyly of the 2nd
to 5th toes, split hand and/or split foot,
monodactyly) and male genital abnormalities
(cryptorchidism, micropenis, feminisation of the
external genitalia). Six of these deletions (2326)
have now been shown to remove the entire
HOXD cluster, but a seventh, found in a patient
with bilateral split foot, appears to stop short
centromeric to the cluster (22), suggesting that
a gene underlying the ectrodactyly spectrum of
limb abnormalities is located not within the
cluster, but in the region centromeric to it. A
submicroscopic deletion, which removes just the
five most 5% HOXD genes (HOXD9 HOXD13 )
and the adjacent EVX2 gene, causes a limb phenotype similar to SPD (27). This may reflect disturbed regulation of the remaining 3% HOXD
genes, but more likely results from haploinsufficiency for the six deleted genes, uncomplicated by
the effects of haploinsufficiency for any neighbouring genes.

Functional effects of the HOXD13 and HOXA13


mutations

The mutations in HOXD13 and HOXA13 discovered so far fall into three distinct classes: truncation mutations, polyalanine tract expansions and
an amino acid substitution in the homeodomain,
each of which probably acts by a different
mechanism.
6

Truncation mutations

The two intragenic deletions in HOXD13 (Fig.


2C,D) and the four nonsense mutations in
HOXA13 (Fig. 4CF) are predicted to result in
proteins lacking part or all of the homeodomain,
and might therefore be expected to cause a loss-offunction. Similar mutations in other homeodomain
proteins, such as PAX6 (28), EMX2 (29), PITX2
(formerly RIEG) (30) and HLXB9 (31, 32), appear
to act as null alleles, whether because the mutant
protein is unstable or cannot bind target DNA.
Moreover, as mentioned above, haploinsufficiency
for HOXA13 appears to cause typical HFGS (20).
The abnormalities in patients with the HOXD13
and HOXA13 truncation mutations differ, however, from those in the corresponding knock-out
mice.
Mice carrying targeted disruptions of the
HOXD13 homeobox have been made by two
groups, with virtually identical results (33, 34).
Less than half the heterozygotes have any abnormalities, and these consist only of a rudimentary
post-axial digit, usually in the forelimbs, sometimes accompanied by minor metacarpal and
carpal defects. Homozygotes have more marked
limb abnormalities, including delayed chondrification and ossification of most distal limb bones,
shortening of the digits due to middle phalanx
hypoplasia/aplasia and, in about half, an extra
post-axial digit. The lowest sacral vertebra and
internal anal sphincter are also abnormal, and
males are infertile, with malformation of the penian bone and accessory sex glands (33, 35, 36).
Neither heterozygotes nor homozygotes, however,
have the distinctive foot phenotype or the central
synpolydactyly seen in humans with HOXD13
truncation mutations.
Mice with a targeted disruption of the HOXA13
homeobox, and a targeted deletion of the entire
HOXA13 gene also share a common phenotype
(37, 38). Heterozygotes have far milder limb abnormalities than humans with HOXA13 truncation
mutations. In the forelimbs there is merely occasional fusion of the phalanges of digit 1, while in
the hind limbs there is often malformation of the
claw and distal phalanx of digit 1, with occasional
2/3 soft tissue syndactyly. Homozygotes die in
utero at 11.515.5 dpc with severe urinary and
genital tract abnormalities, including displaced
ureteric orifices, absent bladder, absent distal Mullerian ducts in females and premature stenosis of
the umbilical arteries, which may explain the mutations early lethality. They also have absent 1st
digits, with hypoplasia and webbing of digits 2 to
5, and delayed chondrification of the digits, carpals

Human HOX gene mutations

and tarsals. Interestingly, these limb defects are


also far milder than those in Hypodactyly mice.
Hd/+ mice have limb abnormalities virtually
identical to human HFGS patients, as described
above, while Hd/Hd mice have only a single incompletely formed digit on each paw (17).
The phenotypic differences between humans
with HOXD13 and HOXA13 truncation mutations
and the corresponding knock-out mice can be explained in several ways. First, the truncated human
proteins, although unable to bind DNA specifically, might nevertheless exert a deleterious functional effect through their remaining N-terminal
portions. The Drosophila homeodomain protein
fushi tarazu can regulate target genes even when its
homeodomain has been partially deleted, if expressed at sufficiently high levels, because it can
still interact with DNA-binding partners through
its N-terminal region (39). If this is also true of
HOXD13 and HOXA13, however, the truncated
proteins must be stably expressed and translocated
to the nucleus, and their effects must be mediated
by their first 107 and 135 amino acids respectively,
as these are the only regions shared by the different
truncated forms of the two proteins.
A second possibility is that the consequences of
haploinsufficiency for HOXD13 or HOXA13 are
rather different in humans and in mice, perhaps
because of human/mouse differences in sensitivity
to reduced gene dosage or in the role normally
played by the two genes. The latter is especially
likely in the developing female reproductive tract,
as normal female mice have two separate uterine
horns, which only join together at the level of the
cervix, and Mullerian duct fusion defects in mice
manifest only as a double cervix and/or a septate
vagina. Differences of this kind now seem to be the
explanation for the different phenotypes seen in
humans and mice carrying loss-of-function mutations in another homeobox gene, HLXB9 (32). The
limb abnormalities in Hypodactyly mice are very
similar to those in humans with HFGS, however,
and are markedly more severe than those in mice
with targeted HOXA13 mutations, even when each
mouse mutation is crossed onto the genetic background of the other (40). The Hypodactyly mutant
protein retains only the first 25 amino acids of
HOXA13, followed by 275 amino acids with no
wild-type counterpart (Fig. 4G). This protein is
stable, localises to the nucleus, and has been found
to cause limb reduction defects when strongly
overexpressed, but only in three out of 15 transgenic mice (41). A gain-of-function effect exerted
by this protein may underlie the Hypodactyly phenotype but, if so, the consequences in mice appear
to be very similar to those of haploinsufficiency for
HOXA13 in humans.

The final, and perhaps the most likely, alternative is that the human truncation mutations and
the Hypodactyly mutation all act as null alleles,
and that the targeted mouse mutations have not
produced a straightforward loss-of-function. The
insertion of a selectable marker cassette during the
construction of targeted mouse mutations is
known to be able to disrupt the expression of
neighbouring genes, especially if the genes in question are closely linked or clustered and share regulatory elements, as the HOX genes do (42). This
possibility could be explored by generating mice
carrying targeted HOXD13 and HOXA13 mutations unaccompanied by additional regulatory sequences, using methods that eliminate the selection
cassette, such as the Cre/loxP site specific recombinase system.
Polyalanine tract expansions

As well as being the first human HOX gene mutations to be discovered, the polyalanine tract expansions in HOXD13 that underlie SPD (Fig. 2E)
were also the first instance of a novel class of
mutation, and helped draw attention to the functional importance of HOX proteins N-terminal
regions, hitherto comparatively neglected in favour
of the homeodomain. Similar pathological expansions have subsequently been identified in
HOXA13, as described above (Fig. 4H), as well as
in two non-homeodomain transcription factors,
RUNX2 (formerly CBFA1) (43) and ZIC2 (44),
which result in cleidocranial dysplasia (MIM
119600) and holoprosencephaly (MIM 603073) respectively. (Smaller expansions in PABP2 cause
oculopharyngeal muscular dystrophy, a late-onset
neuromuscular disorder (45), but this protein is not
a transcription factor and the pathological mechanism involved may be different.) In all four transcription factors, the normal polyalanine tracts are
short (1518 residues), and show little or no polymorphism in length, while the expansions are also
short (715 extra residues) and meiotically stable.
They therefore differ sharply from other diseasecausing expansions of trinucleotide repeats, and
are probably caused not by strand slippage but by
unequal crossing-over between two normal alleles
that have become misaligned during replication
(46).
Polyalanine tracts are a common motif in both
homeodomain and non-homeodomain transcription factors, but their normal function is not understood. One possibility is that they act as flexible
spacer elements between other functional domains
(47); another is that they are involved in binding
other proteins with which the transcription factors
7

Goodman and Scambler

interact (48, 49). It is also unclear what the effect


of expanding such tracts would be. In the case of
HOXD13, at least, expansions do not perturb the
mutant proteins stability or ability to enter the
nucleus (FRG, unpublished data). Moreover, the
penetrance and severity of the associated phenotype increase with increasing expansion size, as
described above, consistent with a progressive
gain-of-function effect. This interpretation is supported by the observation that the phenotypes
caused by expansions in HOXD13 and RUNX2
are subtly different from those caused by probable
loss-of-function mutations in the same genes (4, 9,
12, 43), although this does not appear to be the
case for HOXA13 (19) or ZIC2 (44). The expansions may, therefore, prevent protein protein interactions that normally occur or allow novel
interactions to take place without affecting the
mutant proteins ability to bind their normal DNA
targets, so allowing them to exert a dominant
negative effect over the remaining wild-type
protein.
An SPD-related phenotype occurs in mice with a
targeted deletion at the 5% end of the HOXD cluster
created using the Cre/loxP system, which removes
HOXD11, HOXD12 and HOXD13 without leaving
behind a neomycin resistance cassette (50). The
only abnormality in heterozygotes is mild shortening of digits 2 and 5, due to middle phalanx
hypoplasia. Homozygotes, however, have several
limb abnormalities in common with SPD heterozygotes and homozygotes, including shortening,
fusion, webbing and polydactyly of the digits, and
in particular central synpolydactyly, as well as
small malformed carpals. These mice lack both
copies of HOXD11, HOXD12 and HOXD13, so
their phenotype supports the hypothesis that mutant HOXD13 protein carrying an expanded
polyalanine tract exerts a dominant negative effect
over HOXD11 and HOXD12 as well as over wildtype HOXD13 protein (50), acting as a super
dominant negative.
Substitution in the homeodomain

The only missense mutation so far identified in a


HOX gene is the amino acid substitution in
HOXA13 described above (Fig. 4I), which affects
an invariant asparagine residue in the homeodomains recognition helix. Replacing this residue,
which directly contacts target DNA (51), with histidine would probably not affect the synthesis or
stability of the resultant protein, but would almost
certainly perturb its interactions with DNA. The
side chain of histidine is bulky but partially positive at neutral pH, and so may permit DNA bind8

ing. The mutant protein may, therefore, recognise


new targets or fail to recognise some of its normal
ones, while still retaining its capacity to regulate
transcription, thus exerting a gain-of-function effect, which may explain the exceptional severity of
the associated limb phenotype (Fig. 3EH).
Are there more HOX gene mutations to be found?

The discovery of mutations in HOXD13 and


HOXA13 suggests that mutations in other HOX
genes may also cause human malformation syndromes. Most such mutations are unlikely to be
simple null alleles, however, as a patient hemizygous for the entire HOXA cluster has a phenotype attributable to haploinsufficiency for just
HOXA3 and HOXA13 (20), and patients hemizygous for the five most 5% HOXD genes and
EVX2 have a phenotype similar to that caused by
mutations in HOXD13 alone (27). But gain-offunction mutations, such as the polyalanine tract
expansions and homeodomain substitution described above, may produce different phenotypes.
Other such mutations may include polyalanine
tract contractions, amino acid substitutions perturbing interactions with DNA-binding partners or
co-factors, mutations involving the regulatory regions of single or multiple genes, and chromosomal re-arrangements affecting the regulation of an
entire cluster.
A search for HOXA1 mutations in sporadic
cases of hindbrain malformation has so far proved
negative (52), as has a search for HOXA10 mutations in sporadic cryptorchidism (53). In a large
family with dominantly-inherited Duane syndrome
(sixth cranial nerve palsy-MIM 604356) accompanied by other cranial nerve palsies and deafness,
the underlying gene has recently been mapped to
chromosome 2q31. Although mutations in the coding regions of HOXD1, HOXD3 and HOXD4,
which are important in hindbrain development,
have been excluded, a regulatory mutation remains
possible (54). The gene underlying dominantlyinherited mesomelic dysplasia of the Kantaputra
type (MIM 156232) has also recently been mapped
to the vicinity of the HOXD cluster (55). A similar
mesomelic dysplasia accompanied by cervical and
lumbosacral vertebral abnormalities is associated
with a balanced reciprocal translocation that involves chromosome 2q31 (56). Interestingly, the
limb phenotype in these two conditions closely
resembles that in the radiation-induced semi-dominant mouse mutant Ulnaless. The Ulnaless locus is
tightly linked to the mouse HOXD cluster, and the
expression patterns of HOXD11 -HOXD13 are perturbed in Ulnaless limb buds, but an underlying

Human HOX gene mutations

mutation in any one HOXD gene has been excluded, as has a large deletion or chromosomal
re-arrangement in the region (57, 58). In all three
cases, the causative mutation probably affects a
hitherto unidentified cis-acting regulatory element
controlling the entire cluster.
In addition to their role in development, HOX
genes, especially members of the HOXA and
HOXB clusters, are important in normal haematopoiesis. Germline mutations in HOX gene may
therefore cause inherited haematological disorders,
while somatic mutations may play a part in haematological malignancies (59, 60). Recurrent chromosomal translocations producing NUP98/HOXA9
and NUP98/HOXD13 fusion proteins have been
identified in patients with acute myeloid
leukaemia, chronic myeloid leukaemia and
myelodysplastic syndrome (61 65). These fusion
proteins are thought to contribute to leukaemic
transformation by misregulating normal HOX
gene targets in myeloid cells (66). Activation of
HOXA7 or HOXA9 by proviral integration can
lead to myeloid leukaemia in mice (67), as can
overexpression of HOXA9 and HOXA10 in mouse
haematopoietic cells (68, 69). Several DNA-binding partners of the HOX proteins, including members of the PBX and MEIS families, are themselves
implicated in leukaemogenesis (59, 60), and co-activation of MEIS1 and HOXA9 may be common
in myeloid leukaemia (70).
Finally, HOX genes are expressed in many tissues during adult life, including the gut, kidneys,
genital tract and skin, and there is increasing evidence that they are often misexpressed in solid
tumours (71). Somatic HOX gene mutations, particularly regulatory mutations, may therefore have
a wider role in oncogenesis. Interestingly, HOXA5
has recently been shown to activate transcription
of the key tumour suppressor gene p53, and loss of
p53 expression occurs secondary to loss of HOXA5
expression in many human breast cancer cell lines
and tumours (72). This may reflect a somatic lossof-function mutation in HOXA5, but is usually
caused by transcriptional silencing associated with
methylation of the genes promoter region.
HOXA5 may thus itself be an important tumour
suppressor gene, and other HOX genes may prove
to play a similar role in different tumour types. In
cancer as in developmental disorders, the search
for mutations in the human HOX genes may have
only just begun.
References
1. Krumlauf R. HOX genes in vertebrate development. Cell
1994: 78: 191201.

2. Mark M, Rijli FM, Chambon P. Homeobox genes in


embryogenesis and pathogenesis. Pediatr Res 1997: 42:
421 429.
3. Mann RS, Affolter M. HOX proteins meet more partners.
Curr Opin Genet Dev 1998: 8: 423 429.
4. Muragaki Y, Mundlos S, Upton J, Olsen BR. Altered
growth and branching patterns in synpolydactyly caused
by mutations in HOXD13. Science 1996: 272: 548 551.
5. Mortlock DP, Innis JW. Mutation of HOXA13 in handfoot-genital syndrome. Nat Genet 1997: 15: 179 180.
6. Sayli BS, Akarsu AN, Sayli U, Akhan O, Ceylaner S,
Sarfarazi M. A large Turkish kindred with syndactyly type
II (synpolydactyly). 1. Field investigation, clinical and
pedigree data. J Med Genet 1995: 32: 421 434.
7. Sarfarazi M, Akarsu AN, Sayli BS. Localization of the
syndactyly type II (synpolydactyly) locus to 2q31 region
and identification of tight linkage to HOXD8 intragenic
marker. Hum Mol Genet 1995: 4: 1453 1458.
8. Akarsu AN, Stoilov I, Yilmaz E, Sayli BS, Sarfarazi M.
Genomic structure of HOXD13 gene: a nine-polyalanine
duplication causes synpolydactyly in two unrelated
families. Hum Mol Genet 1996: 5: 945 952.
9. Goodman FR, Mundlos S, Muragaki Y et al. Synpolydactyly phenotypes correlate with size of expansions in
HOXD13 polyalanine tract. Proc Natl Acad Sci USA
1997: 94: 7458 7463.
10. Baffico M, Baldi M, Cassan PD et al. Synpolydactyly:
clinical and molecular studies on four Italian families. Eur
J Hum Genet 1997: 5 (suppl. 1): A142.
11. Akarsu AN, Akhan O, Sayli BS, Sayli U, Baskaya G,
Sarfarazi M. A large Turkish kindred with syndactyly type
II (synpolydactyly). 2. Homozygous phenotype? J Med
Genet 1995: 32: 435 441.
12. Goodman FR, Giovannucci-Uzielli ML, Hall C, Reardon
W, Winter R, Scambler P. Deletions in HOXD13 segregate
with an identical, novel foot malformation in two unrelated families. Am J Hum Genet 1998: 63: 992 1000.
13. Stern AM, Gall JCJ, Perry BL, Stimson CW, Weitkamp
LR, Poznanski AK. The hand-foot-uterus syndrome. J
Pediatr 1970: 77: 109 116.
14. Halal F. The hand-foot-genital (hand-foot-uterus) syndrome: family report and update. Am J Med Genet 1988:
30: 793 803.
15. Donnenfeld AE, Schrager DS, Corson SL. Update on a
family with hand-foot-genital syndrome: hypospadias and
urinary tract abnormalities in two boys from the fourth
generation. Am J Med Genet 1992: 44: 482 484.
16. Poznanski AK, Kuhns LR, Lapides J, Stern AM. A new
family with the hand-foot-genital syndrome-a wider spectrum of the hand-foot-uterus syndrome. Birth Defects Orig
Artic Ser 1975: 11: 127 135.
17. Mortlock DP, Post LC, Innis JW. The molecular basis of
hypodactyly (Hd): a deletion in HOXA13 leads to arrest of
digital arch formation. Nat Genet 1996: 13: 284 289.
18. Innis JW, Darling SM, Kazen-Gillespie K, Post LC, Mortlock DP, Yang T. Orientation of the HOXA complex and
placement of the Hd locus distal to HOXA2 on mouse
chromosome 6. Mamm Genome 1996: 7: 216 217.
19. Goodman FR, Bacchelli C, Brady AF et al. Novel
HOXA13 mutations and the phenotypic spectrum of
Hand-Foot-Genital Syndrome. Am J Hum Genet 2000: 67:
197 202.
20. Devriendt K, Jaeken J, Matthijs G et al. Haploinsufficiency of the HOXA gene cluster, in a patient with handfoot-genital syndrome, velopharyngeal insufficiency and
persistent patent ductus Botalli. Am J Hum Genet 1999:
65: 249 251.

Goodman and Scambler


21. Chisaka O, Capecchi MR. Regionally restricted developmental defects resulting from targeted disruption of the
mouse homeobox gene HOX-1.5. Nature 1991: 350: 473
479.
22. Boles RG, Pober BR, Gibson LH. Deletion of chromosome 2q24-q31 causes characteristic digital anomalies: case
report and review. Am J Med Genet 1995: 55: 155 160.
23. Collins AL. 2q31.3 is important in distal limb morphogenesis. Presented at Proceedings of the Eighth Manchester
Birth Defects Conference 1996.
24. Nixon J, Oldridge M, Wilkie AO, Smith K. Interstitial
deletion of 2q associated with craniosynostosis, ocular
coloboma, and limb abnormalities: cytogenetic and molecular investigation. Am J Med Genet 1997: 70: 324 327.
25. Del Campo M, Jones MC, Veraksa AN et al. Monodactylous limbs and abnormal genitalia are associated with
hemizygosity for the human 2q31 region that includes the
HOXD cluster. Am J Hum Genet 1999: 65: 104110.
26. Slavotinek A, Schwarz C, Getty JF, Stecko O, Goodman
F, Kingston H. Two cases with interstitial deletions of
chromosome 2 and sex reversal in one. Am J Med Genet
1999: 86: 7581.
27. Goodman F, Majewski F, Winter R, Scambler P. Haploinsufficiency for HOXD8 HOXD13 and EVX2 causes atypical synpolydactyly. Am J Hum Genet (Supplement) 1999:
65: A298.
28. Prosser J, van Heyningen V. PAX6 mutations reviewed.
Hum Mutat 1998: 11: 93108.
29. Brunelli S, Faiella A, Capra V et al. Germline mutations in
the homeobox gene EMX2 in patients with severe schizencephaly. Nat Genet 1996: 12: 9496.
30. Semina EV, Reiter R, Leysens NJ et al. Cloning and
characterisation of a novel bicoid-related homeobox transcription factor gene, RIEG, involved in Rieger syndrome.
Nat Genet 1996: 14: 392399.
31. Ross AJ, Ruiz-Perez V, Wang Y et al. A homeobox gene,
HLXB9, is the major locus for dominantly inherited sacral
agenesis. Nat Genet 1998: 20: 358361.
32. Hagan DM, Ross AJ, Strachan T et al. Mutation analysis
and embryonic expression of the HLXB9 Currarino syndrome gene. Am J Hum Genet 2000: 66: 15041515.
33. Dolle P, Dierich A, LeMeur M et al. Disruption of the
HOXD-13 gene induces localised heterochrony leading to
mice with neotenic limbs. Cell 1993: 75: 431441.
34. Davis AP, Capecchi MR. A mutational analysis of the 5%
HOXD genes: dissection of genetic interactions during
limb development in the mouse. Dev 1996: 122: 1175
1185.
35. Kondo T, Dolle P, Zakany J, Duboule D. Function of
posterior HOXD genes in the morphogenesis of the anal
sphincter. Dev 1996: 122: 26512659.
36. Podlasek CA, Duboule D, Bushman W. Male accessory
sex organ morphogenesis is altered by loss of function of
HOXD-13. Dev Dyn 1997: 208: 454465.
37. 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. Dev
1996: 122: 29973011.
38. Warot X, Fromental-Ramain C, Fraulob V, Chambon P,
Dolle P. Gene dosage-dependent effects of the HOXA-13
and HOXD-13 mutations on morphogenesis of the terminal parts of the digestive and urogenital tracts. Dev 1997:
124: 47814791.
39. Copeland JWR, Nasiadka A, Dietrich BH, Krause HM.
Patterning of the Drosophila embryo by a homeodomaindeleted Ftz polypeptide. Nature 1996: 379: 162165.

10

40. Post LC, Innis JW. Altered HOX expression and increased
cell death distinguish Hypodactyly from HOXA13 null
mice. Int J Dev Biol 1999: 43: 287 294.
41. Post LC, Margulies EH, Kuo A, Innis JW. Severe limb
defects in Hypodactyly mice result from the expression of a
novel, mutant HOXA13 protein. Dev Biol 2000: 217: 290
300.
42. Olson EN, Arnold HH, Rigby PW, Wold BJ. Know your
neighbors: three phenotypes in null mutants of the myogenic bHLH gene MRF4. Cell 1996: 85: 1 4.
43. Mundlos S, Otto F, Mundlos C et al. Mutations involving
the transcription factor CBFA1 cause cleidocranial dysplasia. Cell 1997: 89: 773 779.
44. Brown SA, Warburton D, Brown LY et al. Holoprosencephaly due to mutations in ZIC2, a homologue of
Drosophila odd-paired. Nat Genet 1998: 20: 180 183.
45. Brais B, Bouchard J-P, Xie Y-G et al. Short GCG expansions in the PABP2 gene cause oculopharyngeal muscular
dystrophy. Nat Genet 1998: 18: 164 167.
46. Warren ST. Polyalanine expansion in synpolydactyly
might result from unequal crossing-over of HOXD13. Science 1997: 275: 408 409.
47. Karlin S, Burge C. Trinucleotide repeats and long homopeptides in genes and proteins associated with nervous
system disease and development. Proc Natl Acad Sci USA
1996: 93: 1560 1565.
48. Han K, Manley JL. Transcriptional repression by the
Drosophila even-skipped protein: definition of a minimal
repression domain. Genes Dev 1993: 7: 491 503.
49. Licht JD, Hanna-Rose W, Reddy JC et al. Mapping and
mutagenesis of the amino-terminal transcriptional repression domain of the Drosophila Kruppel protein. Mol Cell
Biol 1994: 14: 4057 4066.
50. Zakany J, Duboule D. Synpolydactyly in mice with a
targeted deficiency in the HOXD complex. Nature 1996:
384: 69 71.
51. Gehring WJ, Qian YQ, Billeter M et al. HomeodomainDNA recognition. Cell 1994: 78: 211 223.
52. Capra V, Moroni A, DeMarco P, Faiella A, Boncinelli E,
Cama A. Human HOXA1 gene and hindbrain malformations. Am J Hum Genet (Supplement) 1999: 65: A144.
53. Kolon TF, Wiener JS, Lewitton M, Roth DR, GonzalesJr
ET, Lamb DJ. Analysis of homeobox gene HOXA10 mutations in cryptorchidism. J Urol 1999: 161: 275 280.
54. Appukuttan B, Gillanders E, Juo SH et al. Localization of
a gene for Duane retraction syndrome to chromosome
2q31. Am J Hum Genet 1999: 65: 1639 1646.
55. Fujimoto M, Kantaputra PN, Ikegawa S et al. The gene
for mesomelic dysplasia Kantaputra type is mapped to
chromosome 2q24-q32. J Hum Genet 1998: 43: 32 36.
56. Ventruto V, Pisciotta R, Renda S et al. Multiple skeletal
familial abnormalities associated with balanced reciprocal
translocation 2;8(q32;p13). Am J Med Genet 1983: 16:
589 594.
57. Peichel CL, Prabhakaran B, Vogt TF. The mouse Ulnaless
mutation deregulates posterior HOXD gene expression and
alters appendicular patterning. Dev 1997: 124: 3481 3492.
58. Herault Y, Fraudeau N, Zakany J, Duboule D. Ulnaless
(Ul), a regulatory mutation inducing both loss-of-function
and gain-of-function of posterior HOXD genes. Dev 1997:
124: 3493 3500.
59. Look AT. Oncogenic transcription factors in the human
acute leukemias. Science 1997: 278: 1059 1064.
60. van Oostveen J, Bijl J, Raaphorst F, Walboomers J, Meijer
C. The role of homeobox genes in normal hematopoiesis
and hematological malignancies. Leukemia 1999: 13:
1675 1690.

Human HOX gene mutations


61. Nakamura T, Largaespada DA, Lee MP et al. Fusion of
the nucleoporin gene NUP98 to HOXA9 by the chromosome translocation t(7;11)(p15;p15) in human myeloid
leukaemia. Nat Genet 1996: 12: 154158.
62. Borrow J, Shearman AM, Stanton VP et al. The
t(7;11)(p15;p15) translocation in acute myeloid leukaemia
fuses the genes for nucleoporin NUP98 and class I homeoprotein HOXA9. Nat Genet 1996: 12: 159167.
63. Raza-Egilmez SZ, Jani-Sait SN, Grossi M, Higgins MJ,
Shows TB, Aplan PD. NUP98 -HOXD13 gene fusion in
therapy-related acute myelogenous leukemia. Cancer Res
1998: 58: 42694273.
64. Wong KF, So CC, Kwong YL. Chronic myelomonocytic
leukemia with t(7;11)(p15;p15) and NUP98 /HOXA9 fusion. Cancer Genet Cytogenet 1999: 115: 7072.
65. Hatano Y, Miura I, Nakamura T, Yamazaki Y, Takahashi
N, Miura AB. Molecular heterogeneity of the NUP98 /
HOXA9 fusion transcript in myelodysplastic syndromes
associated with t(7;11)(p15;p15). Br J Haematol 1999: 107:
600604.
66. Kasper LH, Brindle PK, Schnabel CA, Pritchard CE,
Cleary ML, van Deursen JM. CREB binding protein interacts with nucleoporin-specific FG repeats that activate

67.

68.

69.

70.

71.
72.

transcription and mediate NUP98 -HOXA9 oncogenicity.


Mol Cell Biol 1999: 19: 764 776.
Nakamura T, Largaespada DA, Shaughnessy JD Jr,
Jenkins NA, Copeland NG. Cooperative activation of
HOXA and Pbx1 -related genes in murine myeloid
leukaemias. Nature Genet 1996: 12: 149 153.
Thorsteinsdottir U, Sauvageau G, Hough MR et al. Overexpression of HOXA10 in murine hematopoietic cells perturbs both myeloid and lymphoid differentiation and leads
to acute myeloid leukemia. Mol Cell Biol 1997: 17: 495
505.
Kroon E, Krosl J, Thorsteinsdottir U, Baban S, Buchberg
AM, Sauvageau G. HOXA9 transforms primary bone
marrow cells through specific collaboration with Meis1a
but not Pbx1b. EMBO J 1998: 17: 3714 3725.
Lawrence HJ, Rozenfeld S, Cruz C et al. Frequent co-expression of the HOXA9 and MEIS1 homeobox genes in
human myeloid leukemias. Leukemia 1999: 13: 1993 1999.
Cillo C, Faiella A, Cantile M, Boncinelli E. Homeobox
genes and cancer. Exp Cell Res 1999: 248: 1 978.
Raman V, Martensen SA, Reisman D et al. Compromised
HOXA5 function can limit p53 expression in human breast
tumours. Nature 2000: 405: 974 978.

11

You might also like