Molecular genetic delineation of a deletion of chromosome 13q12→q13 in a patient with autism and auditory processing deficits

In a sporadic case of autism and language deficit due to auditory processing defects, molecular genetic studies revealed that a chromosomal deletion occurred in the 13q12→q13 region. No chromosome abnormalities were detected in the parents. We determined that the deletion occurred on the paternally derived chromosome 13. There are two previous reports of chromosome 13 abnormalities in patients with autism. The deletion in the subject described in this paper maps between the two chromosome 13 linkage peaks described by Bradford et al. (2001) in studies of subjects with autism and language deficits. The 9-Mb region deleted in the patient described here contains at least four genes that are expressed in brain and that play a role in brain development. They are NBEA, MAB21L1, DCAMKL1 and MADH9. These genes therefore represent candidate genes for autism and specific language deficits.

Through molecular genetic delineation of chromosome deletions in patients with autism spectrum disorders, candidate genes for these disorders may be identified. Here we report the fine structure mapping of a de novo chromosomal deletion in 13q in a patient with autism and language deficits due to auditory processing defects. The deletion includes the distal third of band 13q12 and the proximal two thirds of band 13q13. Results of our molecular cytogenetic studies and polymorphic marker analysis indicate that in this patient there is a deletion of approximately 9 Mb that extends from 28.52 Mb to 38.06 Mb. Results of polymorphic marker analysis revealed that the deletion arose on a paternally derived copy of chromosome 13. Given the advanced state of human genome sequencing it is now possible to determine which genes are deleted in our patient and to identify brain expressed genes that represent candidate genes for autism and language processing.
There are two other published reports of the co-occurrence of autism and deletions of the proximal region of chromosome 13. Ritvo et al. (1988) reported on the simultaneous occurrence of autism and retinoblastoma in a patient with a deletion that extended from 13q12 to 13q14. Steele et al. (2001) described a case of autism with a de novo deletion of 13q14 → q22.
Specific evidence for linkage of chromosome 13 markers and autism was reported by Bradford et al. (2001). These investigators analyzed the effect of incorporating language information and parental structural language phenotypes into the genome screening for autism. Their results revealed that two distinct peaks of linkage occurred on chromosome 13q. The deletion in our patient maps in a position between the map locations of the two linkage peaks described by Bradford et al. (2001).

Clinical evaluation
Our patient was born by normal vaginal delivery. Birthweight was 6 lb 1 oz, length was 21 inches. He had no feeding problems. Developmental milestones include smiling at one month, sitting at 9 months, walking alone at  GTG GAC TTT TGC T  D13S1229 For  GGT CAT TCA GGG AGC CAT TC  137-151  25,471 kb 0.74  60  Q sol'n  Rev CCA TTA TAT TTC ACC AAG AGG CTG C  D13S260 For  AGA TAT TGT CTC CGT TCC ATG A  158-173  26,424 kb 0.77  57  MgCl 2  Rev CCC AGA TAT AAG GAC CTG GCT A  D13S1493 For  ACC TGT TGT ATG GCA GCA GT  223-248  27,996 kb 0.80  57  MgCl 2  Rev GGT TGA CTC TTT CCC CAA CT  D13S1293 For  TGC AGG TGG GAG TCA A  119- 15 months. He had his first words at two years of age. When evaluated at 3 years and 8 months he had many single words. He could not easily put two words together. Clinical examination was normal and evaluation by a dysmorphologist revealed no abnormalities. MRI of the brain was carried out on a 1.5 Tesla Picker superconducting magnet. The sequences obtained included Sagittal T1-weighted, axial proton density and T2-weighted and coronal 3D gradient echo volume sequences. No mass lesions were seen; ventricles and sulci were within normal limits and no structural abnormalities were identified.

Language testing
Extensive language assessment was carried out when the patient was 6 years old. At that time auditory processing skills were judged to be the subject's greatest weakness and were thought to affect his higher-level auditory comprehension and expressive language skills. Further detailed language evaluation was carried out at the age of 8 years and 9 months. The Token Test for children (TTTFC) (DiSimoni, 1978) assesses the ability to follow directions of increasing length and complexity. The subject's scores indicated continued difficulties with auditory processing skills of memory and sequencing. In addition as the complexity of the task increased his abilities to remember the direction and sequence correctly, decreased. The subject's overall score for the TTTFC was 476. The age mean for this test is 500 with a standard deviation of 5. The revised test of auditory perceptual skills (TAPS-R) (Gardner, 1996) assesses various areas of the individual's auditory-perceptual skills. The subject exhibited severe auditory difficulties in the areas of following directions and reasoning. Auditory word discrimination was an area of strength. The auditory perceptual quotient in this patient was 67. The age mean is 100 with a standard deviation of 15.
The CELF-3 test (Semel et al., 1995) assesses basic foundations of content and form that characterize mature language: word meanings (semantics) word and sentence structure (morphology and syntax) and recall and retrieval of spoken language (memory). Due to short attention span the entire CELF 3 test was not administered. The sub-tests administered revealed weakness in morphology and semantics. The Goldman-Fristoe-Woodcock test of auditory discrimination can be used as an index of an individual's ability to listen to and understand a message in the presence of competing sound (Brown et al., 1987). The subject was able to perform adequately when there was an absence of background noise, however when noise was introduced his ability to discriminate was significantly reduced.

Theory of mind assessment
This was examined using five tasks designed to assess a child's understanding of the appearance-reality distinction (Flavell et al., 1983;Sapp et al., 2000), of another's false beliefs and emotions (Baron-Cohen, 1989; adapted from Baron-Cohen et al., 1985;Harris et al., 1989;Wimmer and Hartl, 1991;Hughes et al., 2000), and of one's own false beliefs (Wimmer and Hartl, 1991). These tasks have been frequently used to assess theory of mind in typically developing children and in children with autism, and recent research into their psychometric properties indicate good test-retest reliability especially among aggregated scores of multiple measures (Hughes et al., 2000).
Theory of mind tests were administered twice over a two-week interval. The child's performance on the first order belief systems was inconsistent across tasks and over time. The child failed a second order false belief task at both testing sessions and exhibited some difficulty in responding verbally to questions tapping an understanding of the appearance of reality distinction. Although he showed some understanding of emotional states, he does not appear to appreciate the beliefs underlying these states. In sum the child appears to have a tentative understanding of the mental state of belief but only in response to very limited straightforward circumstances.

Evaluation for autism
There was no family history of autism. This child was administered the Autism Diagnostic Observation Schedule (ADOS) (Lord et al., 1989) (Module 3) at age 6 years, 2 months. His score measuring impairments in Communication quality was 2 (scores of 2 or higher indicate Pervasive Developmental Disorder Not Otherwise Specified (PDD-NOS); scores of 3 or higher indicate Autistic Disorder). His score measuring Qualitative Impairments in Reciprocal Social Interaction was 5 (scores of 4 or higher indicate PDD-NOS; scores of 6 or higher indicate Autistic Disorder). His combined Communication (2) and Social (5) scores was 7 (scores of 7 or higher indicate PDD-NOS; scores of 10 or higher indicate Autistic Disorder). The ADOS diagnosis for this child was PDD-NOS. This diagnosis was confirmed by clinical examination by pediatric neurologist, Pauline A. Filipek, M.D.
The Autism Diagnostic Interview (ADI) (Lord et al., 1994) was also administered to the mother of this boy when he was 6 years, 2 months of age. His score on the scale measuring qualitative impairments in reciprocal social interaction was 15 (scores of 10 or higher indicate Autistic Disorder). His score on the communication impairments scale was 12 (scores of 8 or higher indicate Autistic Disorder). On the scale measuring repetitive behaviors and stereotyped patterns, his score was 6 (scores of 3 or higher indicate Autistic Disorder). His profile also contained five indications of abnormal development before 36 months (scores of 1 or more indicate Autistic Disorder).

IQ testing
The Stanford-Binet test 4 th Edition (Thorndike et al., 1986) was administered to this child when he was 8 years, 7 months of age. His Test Composite (IQ) was 78 at the 8 th percentile and in the Slow Learner (IQ's of 68-78) Range relative to age-mates (mean = 100, standard deviation = 16). In a subscale pattern typical of autistic children, his Verbal Reasoning IQ was 70 (3 rd percentile) and comprised his lowest area score, while his Abstract/ Visual Reasoning IQ was significantly higher at 90 (27 th percentile). The Quantitative Reasoning IQ was 88 (23 rd percentile) and the Short-term Memory IQ was 80 (11 th percentile). The Memory for Beads visual memory subtest scaled score (mean = 50, SD = 8) was 45 at the 27 th percentile and the Memory for Sentences scaled subtest score was 38 at the 7 th percentile relative to age-mates.

Routine cytogenetic studies
Analysis of metaphases from peripheral blood revealed that the patient had an unbalanced chromosome complement with an interstitial deletion on chromosome 13. The breakpoints of the deletion were estimated to be 13q13.2 and 13q14.1. Chromosome analyses on peripheral blood from the parents revealed normal karyotypes.

Molecular cytogenetic studies
Cultured white blood cells (lymphoblastoid cell lines) from the patient and his parents were used to produce slides with spreads of metaphase chromosomes and interphase nuclei. We utilized information from the Human Genome projects as archived on the NCBI website (http://www.ncbi.nlm. nih.gov/), to identify a series of linearly ordered BAC clones on chromosome 13q. BAC clones were ordered from Research Genetics/In Vitrogen. BAC clone preparations were plated out on agar and single colonies of each specific BAC were isolated and grown overnight in liquid culture medium. BAC clone DNA was extracted by alkaline lysis according to the procedure recommended by Research Genetics. DNA from BAC clones was labeled using Spectrum Green dUTP (Vysis) or Spectrum Red dUTP (Vysis) and nick translation. Labeled BAC clone DNA was ethanol precipitated along with Cot1 human DNA to block repetitive sequences. These precipitates were then dried, and dissolved and used for FISH studies. Hybridization of probes to chromosome preparations and pre-and posthybridization washes were carried out according to procedures recommended by Vysis.

Analysis of polymorphic markers and genotyping
The purpose of these studies was to further define extent of the deletion and to determine the parental origin of the deletion chromosome. The polymorphic markers used, their map position, PCR amplification primers, and conditions are summarized in Table 1.
The markers were amplified using PCR. The reaction mix of 25 Ìl contained 0.50 Ìg of DNA, 2.5 Ìl of Qiagen 10× buffer (100 mM Tris, 500 mM KCl), 1.3 ÌM forward primer, 1.3 ÌM reverse primer, 0.30 mM dNTPs, and 1.2 U Taq. We also added 5.0 Ìl of Qiagen Q solution and/or 1.5 mM of MgCl 2 . The thermocycling was carried out in an MJ Research Inc PTC-100 and MJ Research Inc Minicyclers. The amplification program consisted of an initial incubation at 94°C for 1 min, followed by 30 cycles of denaturing at 92°C for 40 s, annealing at either 57°C or 60°C for 40 s, and extension at 75°C for 90 s, with a final extension at 75°C for 5 min. For individual marker conditions, see Table 1.
Once the fragment was amplified, the PCR product was run on a polyacrylamide gel to confirm that the correct fragment was amplified. The 25-ml gel consisted of 6 % acrylamide:bis (19:1), 10 % glycerol, 1× TBE, 2.1 mM ammonium persulfate, and 0.24 % N,N,N),N)-Tetramethylethlyenediamine. +/+ Indicates that the BAC clone mapped to both members of the chromosome 13 pair; +/-indicates that the BAC clones hybridized to only one member of the chromosome 13 pair (bold italies). In the case of polymorphic markers +/+ indicates that the child inherited a different allele from each parent; +/-indicates that the child inherited an allele from only one parent (bold italics).
After the amplification was verified, the PCR product was diluted. The dilutions ranged from 1:1 to 1:30, depending on the intensity of the fragment as determined by examining the gel. Approximately 0.1-0.2 ng of diluted PCR product was then added to 2 fM of Applied Biosystems Genescan-500 Tamra internal land standard and 2.0 Ìl deionized formamide. The mixture was then denatured for 5 min at 94°C in the thermocyclers mentioned above. Following the denaturing, the samples were immediately placed on ice to prevent renaturation. The samples were then run on the Applied Biosystems 377 sequencer. The gel used for the Sequencer was the Long Ranger Singel from Biowittaker Molecular Applications. The gel was analyzed using Applied Biosystems Genescan and Genotyper programs. We compared the marker fragment sizes of the proband with those of the parents for discrepancies.

Results
Results of molecular cytogenetic studies and analysis of polymorphic DNA markers undertaken in blood samples and cell lines from a child with autism and a deficit in higher level auditory comprehension and expressive language are presented in Tables 2-4 and in Figs. 1, 2. FISH studies revealed that six BAC clones assigned to chromosome 13q12 → q13, in the region between 28,521 kb and 38,058 kb, hybridized to only one member of the chromosome 13 pair in the patient's cells. Results of these studies are summarized in Table 2. An example of FISH studies on metaphase chromosomes is shown in Fig. 1. This figure illustrates that clone RP11-90M5 (labeled red) hybridizes to both members of the chromosome 13 pair. Clone RP11-307O13 (labeled green) hybridizes to only one member of the chromosome 13 pair.
The presence of a deletion on one member of the chromosome 13 pair was confirmed by results of polymorphic marker analysis. Results of analysis of polymorphic markers are summarized in Table 3. ABI tracings of a subset of the markers are illustrated in Fig. 2. The paternal DNA sample is designated 50.101, the maternal sample is 50.102 and their child's sample is 50.201. The child is heterozygous for markers D13S1493 and D13S291. He is hemizygous for the marker D13S218 which maps at approximately 33,019 kb and he was also hemizygous for D13S325 which maps at approximately 37,160 kb. For both markers the child failed to inherit a paternal allele. As may be seen from Table 3 the child inherited both maternal and paternal alleles for markers at other locations on chromosome 13.
We utilized information in NCBI databases to determine a list of genes that map within the region we showed to be deleted. These genes are listed in Table 4. We considered four of the genes in this region to be of particular interest since they are expressed in brain and play a role in brain development. These genes are: NBEA, MAB21L1 (which maps within NBEA), DCAMKL1 (which contains a doublecortin domain) and MADH9.  Fig. 2. Genotype profiles of polymorphic markers. Note that for markers D13S1493 (top left) and D13S291 (bottom right) the proband (50.201) inherited an allele from both the father (50.101) and the mother (50.102). For markers D13S218 and D13S325 the proband (50.201) failed to inherit a marker from father (50.101).

Discussion
We report the presence of a chromosomal deletion in 13q that includes the distal third of band 13q12 and the proximal two thirds of band 13q13, a region corresponding to the region extending from 28.5 Mb to 38.06 Mb. The question arises whether hemizygosity for specific genes on chromosome 13 contributes to or is sufficient to lead to autism and language deficits. In reviewing consequences of haplo-insufficiency Fisher and Scambler (1994) proposed that some developmental pathways are particularly susceptible to dosage effects because of exquisite sensitivity to levels of critical proteins. They also postulated that proteins involved in the assembly of intermolecular multi-subunit complexes in which subunit stoichiometry is important, would be deleteriously affected by hemizygosity. They proposed further that the level of expression of the nondeleted gene may influence the phenotype.
Hemizygous deletions are particularly deleterious when they occur in chromosomal regions that are imprinted, e.g. the Angelman region on chromosome 15q (Clayton-Smith and Laan, 2003). If, for example, maternal genes in the 13q12 → q13 region are silenced through imprinting, the deletion of the paternally derived 13q12 → q13 region would be particularly harmful. Evidence for imprinting in the proximal region of 13q was reported by Demenais et al. (2001) and by Velagaleti et al. (2001).
The 9-Mb region deleted in the patient described here contains at least four genes that are expressed in brain and that play a role in brain development. They are NBEA, MAB21L1 and DCAMKL1 and MADH9. These genes therefore represent candidate genes for autism and specific language deficits. Neurobeachin (NBEA) is a neuron-specific multi-domain protein with a high-affinity binding site for a regulatory subunit of protein kinase A, (Wang et al., 2000). It contains a BEACH WD 40 sequence domain in its C terminal region. At least ten distinct mammalian proteins with BEACH WD40 domains have been identified. Neurobeachin is found throughout the cytoplasm of the neuronal cell body and neuron processes. It also occurs in a sub-population of synapses specifically at the post-synaptic plasma membrane. Neurobeachin functions as a protein kinase anchor protein AKAP. The anchoring of protein kinase to specific sites is associated with many transfer processes including hormone secretion and modulation of neurotransmitter receptors and ion channels. Wang et al. (2000) postulate that neurobeachin plays a role in post-Golgi neuronal membrane trafficking. Recent studies on the domain structure of neurobeachin crystals have led to the identification of plekstrin type domains. Neurobeachin EST's in Unigene HS3821 are from neuronal and endocrine tissues. At the protein level Wang et al. (2000) detected neurobeachin only in brain lysates, including forebrain, cerebellum and brainstem. They postulated that the low level neurobeachin mRNA in endocrine tissue is not translated into protein. Developmental changes in neurobeachin expression have been documented in mouse brain. Expression is highest in neonatal brain and in adult brain levels are 50 % of those in neonatal brain.
A gene designated MAB21L1 is contained within the neurobeachin gene. This gene is homologous to a C. elegans gene Mab21 that plays a role in neural development. The MAB21L1 gene was reported with the designation CAGR by Margolis et al. (1999). Within MAB21L1 there is a CAG repeat that is unstable and undergoes repeat expansion. There are no definitive reports of neurological diseases associated with expansion of this gene. Wong et al. (1999) analyzed Mab21 gene expression in mouse embryos. They reported that both Mab21 and Mab21L1 genes are expressed in overlapping domains at all stages of embryogenesis. Mab21 gene expression was first detected in the neuroepithelium of the cephalic neural folds in E8 embryos. In E10-E14.4 embryos both Mab21 genes were expressed in the encephalic vesicle and in the neural tube.
Expression also occurred in the optic tissue, mid-brain branchial arches and limb buds. Mab21L1 was also found in the otic pits.
DCAMKL1 encodes a brain-specific transmembrane kinase. The N-terminal 345-amino-acid region shows 78 % homology to doublecortin. The C-terminal 427-amino-acid region contains two transmembrane domains and is 98 % homologous to a calmodulin-dependent kinase-like domain. DCAMKL1 is thought to play a role in cortical development. Several different splice variants of DCAMKL1 occur. Burgess and Reiner (2002) reported that DCAMKL1 splice alternatives are differentially expressed in embryonic and adult brain. One splice variant of this gene is designated CPG16 (candidate plasticity gene 16). Burgess and Reiner (2000) demonstrated that DCAMKL1 (designated DCKLl1 by them) is microtubule associated and that it is expressed in the growth cones of post-mitotic neurons. Vreugdenhil et al. (2001) reported that in hippocampus of adult rats alternate splice products of the Dcamkl1 (Dckl1) gene are expressed. They postulated that these different transcripts play a role in controlling neuronal plasticity.
The doublecortin-encoding gene, DCX, maps to the X chromosome Xq22.3 → q23 and encodes doublecortin, a brain-specific putative signaling protein that is mutated in several neuronal migration defects including X-linked lissencephaly and subcortical band heterotopia, (Sossey-Alaoui and . Developmental cortical migration defects have been reported in autism on the basis of neuroradiologic and neuropathologic studies (Piven et al., 1990). Corbo et al. (2002) induced targeted mutations in the DCX gene in mice. They reported that in females heterozygous for DCX mutations and in hemizygous males there was disruption of lamination in the hippocampus. This was most severe in the CA3 region. Behavioral studies indicated that the abnormal cytoarchitecture correlated with defects in context and cued conditioned fear tests.
MADH9 is a member of the SMAD family of proteins that mediate the TGF beta signaling pathway. TGF beta signaling plays a role in proliferation and differentiation of many different cell types. TGF beta signaling pathways have been shown to play a role in differentiation at synaptic junctions (Packard et al., 2003). MADH9 is expressed in brain and in many other tissues. Ritvo et al. (1988) reported the simultaneous occurrence of autism, sporadic retinoblastoma and reduced esterase D activity and a deletion on chromosome 13. In their patient who met DSMIII and National Society for Autistic children criteria for autism, the red cell esterase D levels were 50 % of normal. Chromosomes were analyzed using trypsin-Giemsa banding and a deletion that included band 13q13 and portions of bands 13q12 and 13q14. Esterase D and the retinoblastoma gene (RB) map telomeric to the deletion in our patient. It is of interest to note that the patient with autism that Ritvo described and our patient have deletions of 13q13 and a portion of 13q12. Steele et al. (2001) described a case of autism with a de novo deletion of 13q13 → 13q22. They noted that the RB gene was not deleted. It seems likely that the deletion in our patient is more centromeric than the deletion described in the autism case reported by Steele et al. (2001). Bradford et al. (2001) reported the effect of incorporating linkage information and parental structural language phenotypes into analyses in two chromosomal regions where the highest MMLS/het LOD scores were found in their genome screening for autism, namely chromosome 13q and 7q. The results of their updated linkage analyses revealed that two distinct peaks occurred on chromosome 13q. One peak was obtained with marker D13S217, which maps at 23.351 Mb and another peak was obtained with marker D13S800 which maps at 71.843 Mb. D13S217 and D13S800 are the only chromosome 13 markers for which Bradford et al. (2001) reported linkage results. The deletion in the subject we describe here maps between the two chromosome 13-linkage peaks described by Bradford et al. (2001).

Chromosome 13 and autism
In summary, results of cytogenetic and linkage studies reported in the literature and results of studies in our patient, indicate that there may be two autism and language deficit determining loci on chromosome 13q, one at 13q12 → 13q13 and another at 13q22. Molecular genetic studies of the patient that we report revealed a deletion in the 13q12 → q13 region. Four of the loci that map in this region play a role in brain development: these are NBEA, MAB21L1 (which maps within NBEA), DCAMKL1 and MADH9. Further studies in other autistic subjects are required to confirm if one or more of these genes play a role in the etiology of autism.