|  The
ability to speak is uniquely human. Although other species are capable
of communicating with each other, the use of speech is a distinguishing
feature of Homo sapiens. Is it possible that this trait is determined
by specific genes? This question has been debated by linguists for many
decades, and both sides of the argument have vocal supporters. The recent
isolation of a gene that appears to be critical for the normal acquisition
of language is a remarkable achievement and important to clinicians and
researchers alike. The researchers’ overall strategy is instructive
because a similar strategy is likely to prove successful with other childhood
psychiatric disorders including autism and Tourette syndrome. Each of
the component steps in this process have been reviewed in prior columns:
identifying an extended family for study, using linkage or karyotyping
techniques to pinpoint the chromosomal location of the gene, pulling out
the mutated gene, and determining the normal function of the encoded protein.
Approximately
5% of children are significantly impaired in acquiring expressive and/or
receptive language despite the absence of any detectable deficits in hearing,
intelligence, or socioeconomic status. Twin studies have established that
genetic factors play a significant role in certain developmental language
disorders such as expressive language disorder without articulation difficulties,
which is the most frequent expressive language disability, and expressive
language disorder with articulation difficulties. These disorders have
significantly higher concordance rates in monozygotic twins than in dizygotic
twins.
In
1990, researchers described a three-generation family containing members
with a severe language disorder (Fig.
1). Initial characterization of the family showed that approximately
half the family members were affected. There was an equal distribution
among males and females, an inheritance pattern consistent with an autosomal
dominant form of transmission.
The
family was initially described as suffering from a very specific defect
in the use of grammatical rules. However, continued characterization led
to a broader phenotype not limited to difficulties in grammatical skills,
but pervading virtually all facets of grammar and expressive language
acquisition. Affected individuals suffered from deficits in several areas
of language processing, including the ability to break up words into their
constituent phonemes. There were also indications of cognitive impairments,
with verbal components more profoundly depressed than performance scores.
Moreover, all affected members had severe impairments in fine orofacial
movements. The speech dyspraxia was so severe and disabling that the speech
of a number of the subjects was unintelligible; many affected children
were taught sign language to improve their ability to communicate. This
profile indicates that the disorder disrupted intellectual, linguistic,
and orofacial functions.
Functional
and structural brain-imaging studies were then performed on some of the
affected probands. The results suggest a key site of pathology in the
basal ganglia. Bilateral reduction in the volume of the caudate nucleus
was present in all affected family members studied. The authors speculate
that these neural abnormalities arise early in development and persist
into maturity, leading to the observed intellectual impairments and the
orofacial dyspraxia. Functional abnormalities were also found in several
motor-related areas of the frontal cortex that project to the affected
neostriatal regions. It was suggested, however, that the abnormalities
in the frontal regions were secondary to the primary bilateral lesions
within the basal ganglia.
The
availability of approximately 30 affected and unaffected family members
allowed investigators to localize the gene responsible for the observed
phenotype. Linkage analysis was performed with fluorescence-based typing
techniques. Such techniques rely on the availability of microsatellite
markers, which are DNA sequences evenly spaced throughout the human genome.
The closer a marker is to a gene of interest, the less likely a recombination
event will occur between that marker and the mutated gene. That is, there
is a very high probability that DNA sequences on either side of the affected
gene will be inherited together in affected family members. The markers
need to be “informative,” that is, one needs to be able to
distinguish different variants between the various subjects. Only then
can one determine whether one of these markers is tightly linked with
the gene that causes the disorder.
A
region of chromosome 7 (7q31) was implicated as the area of interest.
Two closely linked markers gave maximal pairwise lod scores of approximately
6. A lod score of 3 means that the odds are 10 to the third power for
the likelihood of linkage compared with odds of no linkage. Because of
the logarithmic nature of lod scores, a lod score of 6 indicates that
the odds for the likelihood of a gene being found at that site were now
a million times greater than the odds for no linkage. The linkage analysis
localized the gene to a region of chromosome 7 containing several million
nucleotides.
At
this point, a fortuitous event occurred that made the task of identifying
and cloning the gene considerably easier. An unrelated patient was discovered
with a very similar expressive language disorder. Moreover, this individual
had a specific chromosomal rearrangement at 7q31, the same region implicated
in the previous linkage analysis.
The
chromosomal abnormality in this new individual was a translocation. Translocations
are exchanges of genetic material between two chromosomes and, along with
other chromosomal abnormalities such as microdeletions or inversions,
are of considerable use to researchers as they often act as beacons to
chromosomal regions in which mutated genes will be found. A well-known
example of this is the Philadelphia chromosome, a translocation that occurs
between chromosomes 9 and 22. The translocation is found in more than
90% of people with chronic myelogenous leukemia (CML). The breakpoints
that disrupt each chromosome were finely mapped, and the genes that lie
at each breakpoint were identified. As a consequence of the translocation,
the gene for ABL, a tyrosine kinase, is translocated from its
normal position on chromosome 9 to chromosome 22, where it is fused into
the BCR gene. The functional consequence of this breakage and
reconnection is the creation of a new gene containing a portion of both
the BCR and ABL genes. The protein translated by this
new gene is a novel, chimeric protein that contains the enzymatic portion
of the protein tyrosine kinase but lacks its regulatory domain. Thus the
protein is constitutively active and unable to be regulated. As the normal
function of the abl protein is to regulate cell differentiation, cell
division, and cell adhesion properties, the newly formed oncogene leads
to the uncontrolled proliferation of white cells observed in CML.
A
similar strategy was used to isolate and characterize the gene responsible
for the language disorder described above. A first step was to identify
any genes located at the translocation breakpoint. A team led by Anthony
Monaco at the Wellcome Trust Center for Human Genetics at Oxford used
fluorescence in situ hybridization (FISH) to map the 7q31 region.
To
understand how this is done, one must first understand certain elements
of the molecular technique used. Human genomic libraries contain all the
sequences of DNA found on our chromosomes. The chromosomes are first cut
into stretches of DNA much smaller than the original full complement of
3 billion nucleotides. The much smaller sections are easier to work with
and manipulate. To be useful to the investigators, the DNA segments are
inserted into other organisms that are capable of replicating not only
their own DNA but also the inserted human fragments. When needed, it is
relatively easy to separate the human DNA inserts from the other DNA present.
Considerable
effort has gone into generating these genomic libraries. Human cells are
often grown in the laboratory in cultures. Experimental tricks have been
devised to isolate specific cell lines that contain only one or another
human chromosome. Over time, each of the human chromosomes was isolated
and cut into smaller portions. These smaller portions were placed into
vectors to allow their replication by the host organism. In fact, these
initial inserts were enormous (up to a million bases long) and could be
replicated only after placing them in yeast. The resulting libraries were
thus called yeast artificial chromosomes (YACs) and, although very useful,
they still proved too large to easily isolate individual genes.
Libraries
were then constructed in which the average genomic insert size was on
the order of tens of thousands of bases. These smaller segments could
now be inserted into bacteria instead of yeast. These libraries are called
bacterial artificial chromosomes (BACs). Once again, the entire human
genome was digested and the resulting component parts were placed into
a suitable vector. Individual BAC clones are currently available, with
each containing a relatively small piece of human genomic sequence. The
position of the insert on the human genome has also been determined. Thus
several BAC clones that were adjacent to the 7q31 region were labeled
with fluorescent markers and allowed to hybridize with chromosomes isolated
from the affected individual. The fluorescent label on the BAC clones
allows microscopic visualization of the chromosomes to which the BAC clone
binds. The BAC clone that spanned the breakpoint on chromosome 7 was identified
in this way.
The
researchers had thus reduced the search from billions of nucleotides on
the full complement of human DNA to a more manageable tens of thousands
of nucleotides present on the BAC clone. Moreover, because the sequence
of each BAC clone is known, the researchers were now able to look at that
sequence to determine whether any genes were present. This is now done
through computer-assisted searches that identify particular patterns in
the nucleotide sequence. Specifically, they look for long stretches of
nucleotides that encode potential open-reading frames, that is, the sequence
of nucleotides that can theoretically be translated into a protein. The
researchers found one such open-reading frame on the isolated BAC clone
and also discovered that the full sequence of the gene was not present.
Additional sequences from adjacent BAC clones were obtained, and the longest
open-reading frame that was present before encountering a STOP codon was
generated.
This
candidate gene was then translated into its predicted amino acid sequence.
Comparisons were made between this sequence and all known proteins, and
highly homologous regions were sought. Such similarities between a known
and an unknown protein provide hints about the potential function of the
unknown protein. The carboxyl terminal of the protein sequence contained
an 84-amino acid sequence that resembled a domain present in a group of
transcription factors known as the forkhead family of proteins. The novel
gene was designated FOXP2 in accordance with the nomenclature
standards for this rapidly growing family of transcription factors. The
researchers found that the gene is expressed in fetal tissues, particularly
the brain. Moreover, preliminary characterization of its expression pattern
in mouse brain suggested that it could be found within the neopallial
cortex and the developing cerebral hemispheres.
The
forkhead gene family comprises transcription factors with a conserved
100-amino acid DNA-binding motif. It was originally identified in Drosophila,
where mutations cause ectopic expression of head structures within the
gut. Members of this family have been found in many species and are known
to be key regulators of embryogenesis. Similar to all transcription factors,
they function as master control genes that regulate the transcription
of other genes required for normal development of the tissues in which
they are found. A number of human disorders have already been found to
be caused by mutations in specific FOX genes including congenital
glaucoma and thyroid agenesis. Many of the mutations that have been found
in these disorders are missense changes that result in the substitution
of critical amino acids in functional domains of the transcription factor.
With
this knowledge, the investigators went back to the original extended family
to search for any mutations in the FOXP2 gene. A single missense
mutation that changed a guanine to an adenine was detected in exon 14
of the gene in all affected family members, and this mutation cosegregated
perfectly with the language disorder. As it was possible that this change
in nucleotide might be a normal variant within the population, 364 unrelated
individuals were also screened. None of them had the observed change in
sequence, indicating that it does not represent a naturally occurring
polymorphism. Moreover, the functional consequence of the mutation was
to change an arginine to a histidine within the critical DNA-binding motif.
The amino acid arginine is found at that position in all forkhead proteins
isolated to date and is believed to be required for either the binding
of the transcription factor to DNA sequences at specific target genes
or regulating transcription of those genes.
In
conclusion, the isolation of a gene that affects the acquisition of expressive
language is a remarkable achievement. It suggests that the transcription
factor FOXP2 is necessary for normal embryonic development of
brain areas related to expressive language and that the genes that it
regulates are required for the establishment of neural networks involved
in this complex developmental milestone. It is likely that future work
will reveal many other genes, such as those regulated by FOXP2,
that affect specific components of expressive language development. It
should also be clear that none of this work was possible without the initial
work and interest of astute clinicians.
top of page
|
