DlWfl14 I I I H Hhh

Parenting Children With Asperger's And High-functioning Autism

Teaching Autistic Children

Get Instant Access

11 tl3 14

12 T15


Polyglut amine repeat stretches

Zinc Leucine FOX finger zipper domain motif

Figure 31.2 Genomic structure and protein domains of FOXP2. Introns >50kb are indicated by //.

a single BAC clone that gave signals from both derivative chromosomes indicating it spanned the 7q31 breakpoint and may contain the disrupted SPCH1 gene. The gene disrupted by the translocation in CS, within the BAC clone, was characterized and designated FOXP2. FOXP2 was subsequently sequenced in KE family members and a G/A nucleotide change causing an arginine to histidine (R553H) substitution was found in exon 14 (Lai et al., 2001). The variant co-segregated perfectly with speech and language disorder and was not found in 364 control chromosomes.

FOXP2 genomic structure and function

FOXP2 is a member of the FOX family of transcription factors, which share a characteristic DNA-binding domain composed of a forked/ winged-helix (FOX). These transcription factors are involved in cellular differentiation and proliferation, signal transduction and pattern formation. FOXP2 also contains a long stretch of glutamine residues; expansion of polyglutamine domains have previously been implicated in other neurological diseases (Cummings and Zoghbi, 2000). However, no significant variation in the FOXP2 polyglutamine stretch has been detected in normal individuals or samples of individuals affected with neurodegenerative diseases, SLI and autism (Bruce and Margolis, 2002; Newbury et al., 2002). The polyglutamine stretch is probably stabilized by the mixture of CAG and CAA codons it contains.

The KE family mutation (R553H) occurs in the most conserved region of the forkhead domain and is adjacent to the residue that makes direct contact with target DNA promoter sequences. The CS translocation is localized in intron three of the gene. Initially 17 exons with alternatively spliced exon 3a and 3b were described from a 6.5 kb transcript. Three additional 5' exons, that may be alternatively spliced for exon 1, have subsequently been identified along with three internal exons 2a, 2b and 4a and a longer transcript of exon 10 (Bruce and Margolis, 2002). The principal protein isoform of FOXP2 is 715 amino acids long and includes two polyglutamine stretches, a zinc finger motif and a forkhead domain; alternative splicing results in isoforms with extra amino acids (Bruce and Margolis, 2002). FOXP2 is believed to span a genomic region of at least 603 kb and further exons and splice variants may yet be identified, Figure 31.2 shows the genomic structure of FOXP2.

Expression studies ofFOXP2

Previous studies of FOX family members have found that they are important regulators of embryogenesis. Northern blot analysis of FOXP2 shows widespread expression in adult and fetal tissues including strong expression in fetal brain and lung (Lai etal., 2001). The expression of FOXP2 and the mouse homolog Foxp2 has been studied in the developing brain of mouse and human.

The expression patterns were similar in mouse and human and suggest that FOXP2 has a conserved role in the development of corticostriatal and olivocerebellar circuits involved in motor control. Expression is found in the basal ganglia, which has been hypothesized to contribute to motor sequencing and procedural memory-like skills (Ferland et al., 2003; Lai et al., 2003; Lai et al., 2003; Takahashi etal., 2003). FOXP2 expression was also detected in the cerebral cortex, which is interesting given the importance of the insular cortex for speech co-ordination. The spatial expression of FOXP2 was found to correspond with structural abnormalities detected by magnetic resonance imaging (MRI) and functional differences detected by positron emission tomography (PET) in brains of affected members of the KE family. FOXP2 is expressed in the developing caudate nucleus within the basal ganglia where a bilateral reduction of gray matter is found in affected KE family members (Vargha-Khadem et al., 1998). The relative importance of the expression of FOXP2 in the developing brain has yet to be determined.

Proposed molecular mechanisms

Reports on other members of the FOX family, which cause autosomal dominant disease, suggest that mutations disrupting the function of the forkhead domain lead to haplo-insufficiency or unbalanced gene dosage during embryonic development. For example, mutations in FOXC1 cause hereditary glaucoma and can involve either deletion or duplication of the FOX gene (Nishimura et al., 2001). The molecular mechanisms of how disruption of the DNA binding site in FOXP2 affects neural development are yet to be described. Additional expression studies and the development of animal models will begin to uncover the role of FOXP2 in the pathways leading towards the uniquely human traits of speech and language (Marcus and Fisher, 2003). In addition, investigation of the intra-specific variation of the human FOXP2 gene showed that it contains amino-acid coding changes and a pattern of nucleotide polymorphism, which indicate that FOXP2 has been the target of selection during recent human evolution, consistent with the emergence of spoken language (Enard et al., 2002).

SLI and autism

Although large extended pedigrees affected by autism and SLI do exist, no clear Mendelian pattern of inheritance has been consistently observed. There are also very few families affected by autism over several generations because autistic individuals very rarely have children. The strategy for locating genomic regions likely to harbor susceptibility genes has relied upon whole genome screens for linkage. The majority of these screens used non-parametric or model free statistics to analyse linkage data, because the genetic model for SLI or autism remains unknown, but various parametric models have also been used for analysis. A larger number of studies have been carried out for autism compared to SLI but for both disorders the identification of susceptibility genes remains elusive and narrowing critical regions has also proved challenging.

Genome screens for SLI

Three whole genome screens have been performed for SLI and several susceptibility loci have been identified (Bartlett et al., 2002; SLIC, 2002; SLIC, 2004). The study by Bartlett et al. included branches of five large families with a history of language and reading impairments. Following assessment with a wide battery of tests, individuals were assigned an affection status based on their spoken language, reading ability or history of language impairment. Multiple parametric analyses were performed using different penetrance levels and disease allele frequencies, which corresponded to a population prevalence of ~7%. Significant evidence for linkage was found at 13q21 for reading impairment with a two-point LOD of

3.62 (under a recessive model). There was also evidence for linkage at 2p22 for language impairment and 17q23 for reading impairment with two-point LODs of 2.28 and 1.92 (under recessive and dominant models respectively). Bartlett et al. have extended their initial study to examine regions of linkage previously found for both SLI and autism in new samples ascertained for SLI (Bartlett et al., 2004). They suggest that studies of individuals with SLI may help to identify a genetic component which is common to both disorders, given that some children with autism have language abilities similar to children with SLI. They find evidence for linkage at 13q21 and more modest linkage signals on 7q and 2q which overlap with autism genome screen results in their extended sample.

In contrast to Bartlett et al., the SLI Consortium (SLIC) followed a non-parametric quantitative trait locus (QTL) linkage approach where no prior assumptions were made about the genetic basis of the disease. It is important that phenotypes with detectable heritabilities are selected for QTL analysis. The SLIC sample included 98 families containing a proband with standard language scores >1.5 SD below the mean for their age. Three quantitative language measures were analysed: the CELF-R receptive and expressive language scales and a non-word repetition test. Two regions with maximum multipoint LOD scores of 3.55 were detected on chromosome 16q and 19q. The first locus on 16q was linked to the non-word repetition measure while the second locus on 19q was linked to the CELF-R expressive language score. A second genome screen using an independent sample of 86 multiplex families by the SLI Consortium supports these findings with a MLS of 2.84 generated on 16q and a MLS of 2.31 on 19q (SLIC, 2004). Combining the data from both SLI Consortium genome screens provides strong evidence for a susceptibility gene on 16q with an MLS of 7.46. The positions of the loci implicated in SLI by these genome screens are shown in Figure 31.3.

Genome screens for autism

Nine whole genome screens have been performed for autism to date and loci have been implicated on all chromosomes except 11, 12, 14, 20, and 21(MLS >1.0). The lack of suitable quantitative measures for autism has resulted in the use of large numbers of affected sib pairs for qualitative linkage analysis and the majority of studies have employed non-parametric linkage analysis. Figure 31.3 shows linkage reported by whole genome screens for autism. The top four linkages reported to date are on 3q25-27 with a maximum multipoint LOD score (MMLS) of 4.21, 2q21-33 with a MMLS score of 3.74, 7q22.2-31 with a MMLS of 3.20 and 13q22 with a MMLS of 3.0 (Barrett et al., 1999; IMGSAC, 2001a; Auranen et al., 2002). The most consistently reported linkage from genome screens is to chromosome 7q near the SPCH1 locus, as shown in Figure 31.4. Convincing replication of linkage results for complex genetic disorders has been equivocal and attempts have been made to perform a meta-analysis of published linkage results. These meta-analyses have used methods such as: ranking linkage results across genome screens to analyse between scan heterogeneity, analysis allowing for the posterior probability of linkage in different subsets, and using multiple scan probabilities. (Badner and Gershon, 2002; Bartlett et al., 2005; Cantor et al., 2005; Trikalinos et al., 2006). These meta-analyses of data find support for linkage reported to chromosome 1, 7q and 17q11 but suffer from a number of statistical limitations and none include all published genome screens. Additional analysis of autism linkage data has focused on sex-limited and parent-of-origin effects by looking at linkage signals driven by same-sex sibling pairs and parental sharing (Lamb et al., 2005).

Several groups have analyzed genome screen data based on phenotypic subsets with the aim of delineating more homogeneous groups and strengthening linkage signals. Measures relating to language ability in particular have been used to subgroup families. Two studies have grouped sib

Figure 31.3 Loci detected by SLI and autism genome screens.

pairs based on phrase speech delay (>36 months) (Buxbaum et al., 2001; Shao et al., 2002). Both studies found increased support for linkage on chromosome 2q with MLSs of 2.86 and 3.32. An alternative approach was taken by Bradford et al., who incorporated information on parental language phenotypes (Bradford et al., 2001). When parent's history of language difficulties was taken into consideration they found that a subset of sib pairs where both probands had severe language delay were responsible for linkage signals on chromosome 7q and 13q. QTL analysis has also been carried out by Alarcon et al. using measures from the ADI for ''age at first word,'' ''age at first phrase'' and ''repetitive stereotyped behavior.'' Evidence for a QTL using a measure of ''age at first word'' overlapped with a QTL for repetitive behavior found on chromosome 7q (Alarcon et al., 2002). The addition of 139 multiplex families to their study provided evidence for QTLs for ''age at first word'' at 3q13 and 17q and evidence for a language QTL on 16q which overlaps linkage found for SLI. The evidence for a language QTL on 7q diminished in the enlarged sample but ranking families according to language covariates gave support for a language-related locus. (Alarcon et al., 2005).

Although these studies provide intriguing evidence that a susceptibility locus specifically relating to language exists for autism, their results are somewhat disparate and implicate already established susceptibility loci. Whether grouping sib pairs by language defines a particularly homogeneous group is debatable: language may be simply acting as a marker for autism severity. Attempting to utilise parental language data and measures from the ADI for analysis is also problematic given the unknown heritability of these traits. A rigorous QTL approach, as used by SLIC in their SLI genome screen, helps guard against the problems of unknown trait heritability and diminishing sample sizes when subsets are analysed. Suitable robust measures are not yet available for autism but exploratory analyses based on language endophenotypes ascertained from the ADI-R have been performed as described above.

Cytogenetic studies for autism

The identification of FOXP2 was aided by the characterization of the translocation of patient CS. Similarly, chromosomal abnormalities could give clues to the location of genes in multifactorial disease such as SLI and autism. Of particular interest in autism is the 15q11-q13 chromosomal region, which is most frequently reported as containing chromosomal abnormalities in autistic individuals and overlaps the Prader-Willi Angelman critical region (PWACR). There also appears to be a cluster of cytogenetic abnormalities in individuals with autism on chromosome 7, as shown in Figure 31.4 (visit www.chr7.org/ for more information on chromosome 7). Vorstman et al. have reviewed cytogenetic regions of interest (CROI) associated with an autistic phenotype. They identify regions of overlap between CROIs and linkage and association studies on chromosome 7, 15q11-q13 and 5p14 as well as novel regions with multiple case reports of cytogenetic abnormalities such as 22q11 and 22q13 (Vorstman et al., 2006). However, only ~3-6% of autistic individuals possess chromosomal abnormalities and the cytogenetic evidence must be viewed with caution. The genetic pathways disrupted by these comparatively large-scale genomic abnormalities could provide information about unknown susceptibility gene variants, but may also be unrelated.

Association studies for autism

Numerous association studies have been carried out for autism both focused on candidate genes and using genome screen data (Vorstman et al., 2006). In general, these studies suffer from a number of limitations: they have only relatively modest sample sizes of between 100-200 cases or trios, the p-values reported have only moderate significance (P < 0.05 but >0.001) and insufficient

SNPs have been genotyped to completely capture linkage disequilibrium across the candidate gene. There is however some evidence for association on chromosome 7q for autism and SLI, which support the linkage results, but these associations do not cluster closely together and have only a moderate significance (IMGSAC, 2001b; O'Brien etal., 2003). Similarly, investigations into other candidate loci such as the serotonin transporter (5-HTT) gene on chromosome 17 and the gamma-aminobutyric acid (GABA) receptor on chromosome 15, have not been consistently replicated, casting doubt on the meaning of these associations (Lamb et al., 2002). A systematic survey of a candidate region by association mapping with dense single nucleotide polymorphisms (SNP) maps, rather than narrowly focusing on a single gene, is now becoming more feasible and may present a useful way to localize susceptibility genes (Botstein and Risch, 2003).

Candidate gene investigations for SLI and autism

Following the identification of FOXP2, a number of DNA samples from language impaired patients have been screened for mutations (Bruce and Margolis, 2002; Meaburn et al., 2002; Newbury et al., 2002; Wassink et al., 2002; Gauthier et al., 2003; O'Brien et al., 2003; Gong et al., 2004; MacDermot et al., 2005). In general, findings have been negative except for a single study investigating FOXP2 in probands affected by verbal dyspraxia. Three novel coding changes were found including one missense stop codon in exon 7 that is likely to have functional significance (MacDermot et al., 2005). Screening samples of individuals with autism, language impairment and idiopathic neurodegenerative movement disorders has detected some polymorphisms in the poly-glutamine stretch of the gene and silent mutations (Newbury and Monaco, 2002). It is worth noting that a study of Chinese Han trios, affected by autism found some evidence for association with FOXP2 but this has not be replicated in other samples and may represent a population specific effect (Gong et al., 2004). It is possible that variants



Region of cytogenetic abnormality/breakpoint


46, XX, t(5;7) (q14;q35)



46, XX t(7;13) (q32;q12)


46, XY, t(7;21) (q32;q22.3)


46, XY, dup(7) (q31.32;q34)

Autism and developmental delay

46, XY, t(5;7) (q15;q31.32)


Patient CS (FOXP2 disrupted)

46, XY, t(5;7) (q22;q31.2)


Language impaired/ behavioral problems

46, XY, t(2;7) (p23;q31.3)


46. XY, inv(7) (p12.2;q31.3)


46, XY, t(7;13) (q31.3;q21)


46, XX, t(7;11) (q31.2;q25)



46. XY, inv(7) q22;q31.2)


46, XY, t(6;7) (p11.2;q22)

Autism, deafness and hyperactivity

46,XY, t(7;12) (q21.4;q15)



46, XY, del(7)(q21.13)


Autism and mental retardation

46, X?,t(7;20) (q11.2:pp11,2)



46, XY, t(1;7;21) (p22;q;q)

Region of linkage

cm fjc^^1^ 'tco «"«"iri (¿^ cm cmcmcoiof- co co co coco co eo

11 1111 111


.— t— i\J CNl t- cm cm t- cmo-rvl-t- o" co co lf> cmco cooj —

.— t— i\J CNl t- cm cm t- cmo-rvl-t- o" co co lf> cmco cm fjc^^1^ 'tco «"«"iri (¿^ cm cmcmcoiof- co co co coco co eo

Figure 31.4 Cytogenetic cases associated with autism and region of linkage on chromosome 7. PDDr, pervasive developmental disorder.

exist in regulatory regions of FOXP2 that have not yet been detected and affect speech and language, but it is unlikely that coding variants in the gene play a major role in the genetic susceptibility of autism or SLI.

A large number of genes have been screened for mutations in autistic individuals and a handful of missense variants detected. Few of these studies are powerful enough either to convincingly exclude a candidate gene or demonstrate a substantial contribution to the genetic susceptibility to autism. Table 31.2 shows a selection of candidate gene studies for autism and information about replication studies. One of the few studies to detect numerous missense mutations in a single gene was for reelin (RELN) whose protein product is important for neuronal migration during brain development (Bonora et al., 2003). Mice naturally null for the reelin protein have a characteristic reeling gait and possess neuroanatomical abnormalities in neuron positioning in the cerebral cortex, cerebellum and hippocampus. RELN maps beneath the peak of linkage reported by IMGSAC (International Molecular Genetic Study of Autism Consortium) at 7q22 and preferential transmission of a triplet repeat 5' of the gene has been reported (Persico et al., 2001). Thirteen missense mutations were identified in total by Bonora et al., five of which were not present in controls. It remains difficult to assess the significance of these results since no behavioral phenotype could identify individuals with mutations. Furthermore, the numbers of families with missense mutations are insufficient to explain the strength of linkage detected on chromosome 7. Additional studies of reelin in independent autism samples have been carried out but the evidence they present is contradictory, as shown in Table 31.2.

Was this article helpful?

0 0
Joy Of Modern Parenting Collection

Joy Of Modern Parenting Collection

This is a collection of parenting guides. Within this collection you will find the following titles: Issues, rule and discipline, self esteem and tips plus more.

Get My Free Ebook

Post a comment