American Journal of Medical Genetics (Neuropsychiatric Genetics) 96:765–770 (2000)

Analysis of a 1-Megabase Deletion in 15q22-q23 in an Autistic Patient: Identification of Candidate Genes for Autism and of Homologous DNA Segments in 15q22-q23 and 15q11-q13

Moyra Smith,* Pauline A. Filipek, Charles Wu, Maureen Bocian, Simin Hakim, Charlotte Modahl, and M. Anne Spence Department of Pediatrics, University of California, Irvine, California

We have identified a one megabase deletion sequences or to non-coding homologous in the 15q22-15q23 region in a patient with DNA sequences. The PTPN9 gene encodes a autism, developmental delay, and mild dys- non-receptor protein tyrosine phosphatase. morphism. Genes that map within the dele- The Slp-1 [hUNC24] gene is expressed tion region and genes that are interrupted mainly in the brain. Am. J. Med. Genet. or rearranged at the deletion breakpoints (Neuropsychiatr. Genet.) 96:765–770, 2000. are candidate genes for autism. Fluroes- © 2000 Wiley-Liss, Inc. cence in situ hybridization studies in this patient revealed that part or all of the PML KEY WORDS: autism; 15q22-q23; 15q11-q13; gene is absent from one chromosome 15 and PTPN9; SLP-1[hUNC24] a BAC clone containing the D15S124 gene locus hybridizes to only one chromosome 15. BAC clones containing the PTPN9, and INTRODUCTION SLP-1[hUNC24] genes showed markedly re- duced hybridization in the 15q22-q23 region Autism is characterized by impairments in reciprocal on one chromosome 15 in the patient. These social interaction and communication, restricted and BACs also hybridize to the 15q11-q13 region stereotyped patterns of interests and activities, and the in close proximity to SNRPN and HERC2, presence of developmental abnormalities by 3 years of and in this region there is equal intensity of age [Bailey et al., 1996]. Autism most likely represents signal on the normal and on the deleted a genetically heterogeneous condition. There is evi- chromosome. There are previous reports of dence from linkage studies [Bass et al., 1998] and from deletions and duplications of the 15q11-q13 cytogenetic studies [Steffenburg et al., 1996; Schroer et region in patients with autism. Our patient al., 1998] that a locus on chromosome15q11-15q13 represents the first report of a 15q22-q23 de- plays a role in the development of autism. Schroer et al. letion. Hybridization of the PTPN9 and [1998] proposed that the GABA receptor A genes and Slp-1 Bac clones to the 15q11-q13 and the the UBE3A gene represent candidates for autism. We 15q22-q23 regions of chromosome 15 may be undertook an analysis of chromosome 15 gene loci by due to the presence of PTPN9 or SLP-1 gene fluorescence in situ hybridization (FISH) in a patient sequences or to the presence of other gene with autism, developmental delay, and mild dysmor- phology who exhibited a number of features suggestive of Angelman syndrome. The results of high-resolution karyotype analysis were normal. The Angelman region fluorescence in situ hybridization; was examined by FISH using SNRPN and GABRB3 ס Abbreviations: FISH probes, each in combination with the PML gene as a ס National Center for Biotechnical Information; PCR ס NCBI polymerase chain reaction. control probe for chromosome 15. SNRPN and Contract grant sponsor: National Institute of Child Health and GABRB3 genes mapped to the 15q11-q13 region and Human Development (NICHD); Contract grant number: HD yielded normal signals. The PML gene probe hybrid- 35458 (M.A. Spence, principal investigator); Contract grant spon- sor: Funds from the National Alliance for Autism; Contract grant ized to only one member of the chromosome 15 pair. number: 26165 (Moyra Smith, principal investigator). This finding indicated that there was a deletion in the *Correspondence to: Moyra Smith, Department of Pediatrics, chromosome 15q22-q23 region. Further studies were Medical Sciences 1, Room C214, University of California, Irvine, undertaken to define the extent of the deletion, to rule CA 92697. E-mail: [email protected] out the occurrence of uniparental disomy, and to exam- Received 26 January 2000; Accepted 13 March 2000 ine the SNRPN methylation status to determine © 2000 Wiley-Liss, Inc. 766 Smith et al. whether imprinting defects were present. This is the The following tests revealed no abnormalities: brain- first report of a deletion in the 15q22-q23 region in a stem auditory evoked potentials, computed tomogra- patient with autism. It is likely that the autistic behav- phy scan, creatine phosphokinase levels, routine high- iors in our patient are due to the absence of one copy of resolution karyotype, DNA for fragile-X syndrome, a specific gene or genes (haploinsufficiency) or to the sleep-deprived electroencephalography, free/total/acyl interruption or rearrangement of specific genes within carnitine profile, quantitative serum amino acids, lac- the breakpoint regions. Genes that map within the de- tate, pyruvate, and urine organic acids. Physical and letion region on 15q22-q23 and genes that map at the dysmorphology examinations at 7 years of age revealed deletion breakpoint region therefore represent candi- a height in the 10th percentile, weight in the 5th per- date genes for autism. We identified a series of over- centile, and head circumference at 50th percentile. The lapping BAC clones that map in the distal breakpoint patient (Fig. 1a) has blond hair, light blue lacy irises, region of the deletion in our patient. These overlapping and skin, hair, and eye color much lighter than that of BACs, which contain the sequence of two different the parents. She has a long, narrow face, slightly flat genes, PTPN9, and SLP-1, the human homolog of the occiput with no horizontal groove, minimal epicanthal Caenorhabditis elegans UNC24 gene, hybridize to the folds, slightly concave nasal base and bridge, thin up- 15q22-q23 region and to the 15q11-q13 region. This per lip, normal mouth width, wide dental spacing in finding indicates that homologous DNA sequences ex- the upper jaw, narrow anterior palate with broad lat- ist in the 15q22-q23 region, which is deleted in our eral palatine ridges, relatively prominent pointed chin, patient with autism, and in the 15q11-q13 region, mild pectus excavatum, poor muscular contours of which has been shown to be duplicated deleted or in- thighs and calves, flat nails, lax fingers, and somewhat verted in other cases of autism that have been reported transparent skin on chest and shoulders. The patient is in the literature [Bundey et al, 1994; Steffenburg et al., hypotonic, and deep tendon reflexes are brisk. She 1996; Cook et al., 1997]. walks with an immature gait, often with her arms in an uplifted flexed position. She has occasional episodes of hand flapping and intermittent episodes of excessive CLINICAL REPORT chewing and mouthing behaviors and teeth grinding. The mother was a 37-year-old gravida 3 woman, and There is no tongue protrusion, no excessive smiling, pregnancy was uneventful with no teratogenic expo- and no unprovoked outbursts of laughter. The child has sures. This infant was delivered at term by cesarean frequent infections. The results of a complete develop- section for possible fetal distress, with a birth weight of mental and neuropsychological evaluation, and a quan- 6 pounds 13 ounces. Hypotonia and severe global de- titative brain magnetic resonance imaging scan will be velopmental delay were evident by 6 to 12 months: the presented elsewhere (Filipek et al., unpublished manu- child sat at 10 months, crawled at 20 months, and script). walked at 26 to 29 months. She developed at least seven spontaneous words after 18 months, had good METHODS AND RESULTS eye contact, pointed, and waved bye-bye, before losing all of these skills by 29 months of age. Evaluation for suspected autism was carried out at 4.2 years of age, Cultured lymphoblastoid cell lines were used to pre- when she was evaluated as meeting the criteria for pare slides with metaphase spreads and interphase nu- autistic disorder by the DSM-IV, the Autism Diagnos- clei. FISH was carried out by standard techniques. tic Interview- Revised [Lord et al., 1994], and the Au- Probes for FISH included commercially available tism Diagnostic Observation Schedule-Generic [Lord et probes from Oncor, Gaithersburg, MD [GABRB3/PML al., 1989]. The Autism Diagnostic Interview algorithm and SNRPN/PML] and Vysis, Downer’s Grove, IL [SN- scores are presented in Table I. The child’s medical RPN and PML/RARA]. We isolated a series of BAC history is significant for a lack of feeding problems, clones for loci in the 15q22-q23 region [National Center excessive drooling, sleep disturbance, hypermotoric be- for Biotechnology Information (NCBI), 1999]. BAC havior, gait ataxia, or seizures. The infant was noted to clones corresponding to D15S124 and D15S160 genes have strabismus, which was treated by eye patching. were isolated by screening BAC libraries with oligo- nucleotide primers corresponding to these loci. BAC clones containing sequences corresponding to the genes TABLE I. Algorithm Scores for the Autism TLE3, PTPN9, PML, SLP1(hUNC24), CHRNA5, and Diagnostic Interview ETFA were identified by searching for genomic DNA Patient Autism clones with matching sequences in the NCBI genome Categories score cutoffa survey sequence database. BAC clones thus identified were ordered from Research Genetics, Huntsville, AL. Qualitative impairment of reciprocal social interaction 20 10 Individual BAC clones were cultured, and DNA was Impairment of communication 14 7 isolated and used for polymerase chain reaction (PCR) Repetitive behaviors and to confirm that each BAC clone was positive for the stereotyped patterns 7 3 corresponding gene sequences. Purified BAC DNA was Abnormal development labeled with spectrum green or spectrum orange (Vy- evident before 36 months 5 1 sis), using the Vysis labeling system. Labeled BAC aNote that scores at or above the cutoff indicate autism DNA was hybridized to human Cot 1 DNA to block Analysis of 1Mb Deletion in 15q22-q23 in an Autistic Patient 767

Fig. 1. a: Patient at 7 years of age. Note the posturing, prominent chin, and somewhat flattened nasal bridge. b: FISH using D15S124 BAC. Note the single hybridization signal in metaphase, and in interphase the nuclei from patient. c: FISH with SNRPN probe (red) and PTPNN BAC (green). Note the normal hybridization of SNRPN to the 15q11-q13 region. d: FISH with SLP-1 BAC. This BAC hybridizes to the 15q11-q13 and 15q22-q23 regions. The signal in the latter region is markedly reduced on one chromosome in the patient. repetitive sequences. Thereafter, it was hybridized to Results of FISH Studies metaphase and interphase nuclei on prepared slides. Following the washing of slides according to standard FISH on the patient’s chromosomes revealed normal protocols, slides were examined using direct fluores- hybridization with SNRPN, GABRB3, and HERC2 cence microscopy at appropriate wavelengths. BAC probes in the 15q11-q13 region. PML probes [Vysis, clones used and their map positions are listed in Table II. Oncor] hybridized to only one member of the chromo- 768 Smith et al.

TABLE II. Results of FISH Studies on Patient Chromosomes Signal on #15 in each metaphase Number of Locus Centirays Map position signals Level of signal SNRPN 24.20 15q11-q12 2 Equal on both HERC2 25.16, 38.5 15q11-q13 2 Equal on both GABRA 38.85 15q12-q13 2 Equal on both TLE3 247.46 15q22-q23 2 Equal on both D15S124 251.04 15q22-q23 1 PTPNN 254.56 15q22-q23 2 Trace on one PTPN9 254.56 15q22-q23 2 Reduced on one PTPN9 15q11-q13 2 Equal on both PML (Oncor) 254.58 15q22-q23 1 PML (Vysis) 15q22-q23 1 PML Bac 15q22-q23 2 Reduced on one D15S160 258.61 15q23 2 Reduced on one D15S160 258.61 15q11-q13 2 Equal on both SLP-1 258.61 15q23 2 Reduced on one 15q11-q13 2 Equal on both Snurportin 15q23 2 Equal on both ETFA 268.81 15q24 2 Equal on both CHRNA5 273.04 15q24 2 Equal on both some 15 pair, indicating the presence of a deletion in mosomes from the mother and father of the patient. the 15q22 region. Results obtained with locus-specific These studies revealed normal hybridization of the BAC clones that were used to define the limits of the PML probes (PML Vysis, PML Oncor) and of the deletion are described in Table II and are illustrated in SNRPN and GABRB3 probes. Figures 1b–d. The proximal breakpoint of the deletion lies distal to TLE3, which is mapped at 247.46 centi- Analysis of Polymorphic Microsatellite rays. The D15S124 BAC clone hybridized to one chro- Repeat Markers mosome 15 only, (Fig. 1b) indicating that DNA con- DNA was prepared from peripheral white blood cells tained within the D15S124 BAC was absent from the and/or lymphoblastoid cell lines, from the patient and deletion chromosome. The PTPNN BAC gene (Fig. 1c) from her mother and father. Fluorescence-tagged prim- showed only a trace of hybridization to the 15q22-q23 ers specific for each microsatellite repeat polymor- on the deleted chromosome 15. Signal at 15q22-q23 phism were used to generate PCR products that then with PTPN9, D15S160, and SLP-1 BACs (Fig. 1d) was were analyzed by electrophoresis (Alf system, Pharma- markedly reduced on the deletion chromosome. These cia, Uppsala, Sweden). GABRA5 and D15S125 studies results indicate that the distal breakpoint of the dele- indicate that the child does not represent a case of uni- tion lies within the chromosomal region that hybridizes parental disomy, since she has inherited one allele to the overlapping BACs, PTPN9, D15S160, and SLP1 from each parent (Table III). It is not possible to be (258.61 centrirays). The deletion is therefore approxi- certain which parental chromosome underwent dele- mately 10 centirays [cR3000] in size, which corre- tion since the parents each share one allele for many of sponds to a genetic map distance of 1 centimorgan and the markers tested. a physical map distance of approximately 1 megabase. The Snurportin BAC maps distal to the deletion. FISH Analysis of SNRPN Imprinting Status revealed that the BAC clones PTPN9, D15S160, and This analysis was carried out using the methylation- SLP-1 hybridized to both the 15q22-15q23 and the specific PCR analysis developed by Kubota et al. 15q11-q13 region (Fig. 1d) on the deleted chromosome [1997]. Bisulphite-modified DNA from our patient and and the normal chromosome 15 in the patient and on her parents was used in PCR experiments with the control chromosomes. The 15q11-q13 signal obtained with these probes maps adjacent to, and is partly su- perimposed on, a signal obtained with the SNRPN TABLE III. Analysis of Chromosome 15 probe (Vysis) and with the HERC2 BAC clone. DNA Microsatellite Polymorphisms sequencing revealed that the PTPNN and PTPN9 Map position BACs partly overlap. The PTPN9 BAC hybridizes to Polymorphic marker cR cM Father Mother Child the 15q11-13 chromosome region and to the 15q22-q23 GABRB3 38.85 14 BC AC BC region, while the PTPNN BAC hybridizes only to the GABRA5 38.85 14 A BC AB 15q22-q23 region. Evidence for the overlap of the D15S970 56.8 AB A AB D15S160 and SLP-1 BACs was obtained through PCR D15S993 59.3 BC AC C experiments, which showed that a product could be D15S125 231.16 63.8 B A AB generated from both of these BAC clones when D15S124 251.04 71 B AB B D15S160 locus primers were used for PCR. FISH stud- D15S197 253.26 71.3 B AB B ies were caried out on metaphase and interphase chro- D15S160 258.61 71.8 A A A Analysis of 1Mb Deletion in 15q22-q23 in an Autistic Patient 769

SNRPN-M and the SNRPN-P primer sets. The ex- q13 region is due to the presence of PTPN9 or SLP-1 pected 174-base pair maternal PCR fragment and the gene sequences. It is possible that hybridization to both 100-base pair paternal fragment were generated from the 15q22-23 and the 15q11-q13 regions is due to the parental DNA and from patient DNA. presence of other gene sequences or noncoding DNA homologous DNA sequences in these BAC clones. The DISCUSSION PTPN9 gene encodes a nonreceptor protein tyrosine phosphatase. Phosphorylation of the tyrosine residues The results of our study revealed the presence of an on proteins plays an important role in the regulation of approximately 1-megabase deletion in the 15q22-15q23 neural growth and differentiation [Van Vactor et al., region in a patient with autism, developmental delay, 1998]. This process is tightly controlled by the opposing and mild dysmorphism. Genes that map within the de- activities of protein tyrosine kinases, which add active letion region and genes that are interrupted or rear- phosphate residues and protein tyrosine phosphatases ranged at the proximal and distal breakpoint regions and remove phosphate residues. The SLP-1 gene en- are candidate genes for autism. A BAC gene containing codes a bipartite protein, a stomatin-like N-terminal the D15S124 locus hybridizes to only one chromo- domain and a nonspecific lipid transfer protein C- some15 in this patient; it is likely that genes contained terminal domain. This bipartite SLP-1 protein is the within this BAC are deleted. Through DNA sequenc- human homolog of the C. elegans UNC24 protein ing, we identified an EST clone in this BAC Image [Seidel and Prohaska, 1998]. SLP-1 is expressed clone 1621746 [gb AI003855]. The gene that is repre- mainly in the brain, with the highest levels of expres- sented by this EST clone has not yet been character- sion in frontal lobe, cerebral cortex, caudate nucleus, ized. Part or all of the PML gene is absent from one amygdala, temporal lobe, putamen, substantia nigra, chromosome in our patient. PML is a ubiquitously ex- and hippocampus. The UNC24 gene in C. elegans is pressed nuclear phosphoprotein that is localized within designated as a behavioral gene, since it is required for specific bodies within the nucleus [Ruthardt et al., normal locomotion and interacts with genes that affect 1998]. PML is involved in growth control, and overex- the response to volatile anesthetics. This protein is ap- pression of PML induces growth suppression [Le et al., parently involved in the function of the lipid bilayer 1996]. In patients with promyelocytic leukemia, the [Barnes et al., 1996]. Further studies are required to PML gene forms a fusion product with the retinoic acid determine whether genes mapping within the deletion receptor alpha [deThe et al., 1990]. In our patient, or at the deletion breakpoints, as defined by our pa- there was no evidence of formation of a PML/RARA tient, are altered or missing in other patients with au- fusion gene product. Goy et al. [1995] described the tism. presence of a megabase insertion deletion restriction Ј fragment length polymorphism 3 to the PML gene. It ACKNOWLEDGMENTS is unlikely that a polymorphism accounts for the find- ings in our patient, since neither of the parents show We thank the family of our patient for their coopera- evidence of a deletion of PML. BACs containing PTPN9 tion and interest. We are indebted to graduate stu- and SLP1 genes showed markedly reduced hybridiza- dents, especially Tracy Laulhere, for their assistance in tion in the 15q22-q23 region on one chromosome 15. It the assessment core of the autism program project. is likely that the distal breakpoint of the deletion in our Note that a letter of permission to publish the patient’s patient maps within the chromosome 15 DNA se- photographs accompanies this manuscript. quences present in these BACs. The question then arises whether the deletion, interruption, or rearrange- REFERENCES ment of PTPN9 or SPL-1 played a role in the causation of autism in this patient. Of particular interest is the Bailey A, Phillips W, Rutter M.1996. Autism: towards an integration of fact that the PTPN9 BAC and the SLP-1 BAC hybrid- clinical, genetic, neuropsychological, and neurobiological perspectives. ize to both the 15q22-q23 region and the 15q11-q13 J Child Psychol Psychiatry 37:89–126. region in the patient and in controls. In our patient Bass MP, Wolpert CM, Menold MM, Donnelly SL, Ravan SA, Hauser ER, Church TD, Slotterbeck B, Curtis AM, Zimmerman A, Gilbert JR, with autism, there is a deletion in the chromosome Wright HH, Abramson RK, DeLong GR, Vance JM, Curraro ML, Peri- 15q22-q23 region. In the literature, there are a number cak-Vance MA. 1998. Genomic screen for autistic disorder. Am J Hum of cases reported with abnormalities of chromosome Genet 63:A281. 15q11-q13, including deletions, duplication, and inver- Barnes TM, Jin Y, Horvitz HR, Ruvkun G, Hekimi S.1996. The Caenorhab- sions, [Bundey et al., 1994; Steffenburg et al., 1996; ditis elegans behavioral gene unc-24 encodes a novel bipartite protein similar to both erythrocyte band 7.2 (stomatin) and nonspecific lipid Cook et al ,1997; Schroer et al. 1998]. In the 15q11-q13 transfer protein. J Neurochem 67:46–57. region, the PTPN9 and SLP-1 BAC signal maps are Bundey S, Hardy C, Vickers S, Kilpatrick MW, Corbett JA. 1994. Dupli- adjacent to and partly overlap the SNRPN and HERC2 cation of the 15q11-13 region in a patient with autism, epilepsy and signal. The HERC2 gene and its duplicated copies are ataxia. Dev Med Child Neurol 36:736–742. transcribed elements that flank the Prader-Willi- Cook EH Jr, Lindgren V, Leventhal BL, Courchesne R, Lincoln A, Shulman Angelman gene region. Domains within this gene en- C, Lord C, Courchesne E. 1997. Autism or atypical autism in mater- nally but not paternally derived proximal 15q duplication. Am J Hum code a guanine nucleotide exchange element and an E3 Genet 60:928–934 ubiquitin ligase [Ji et al., 1999]. Further analysis of de The H, Chomienne C, Lanotte M, Degos L, Dejean A. 1990. The t(15;17) fragments of these BAC clones will be required to dem- translocation of acute promyelocytic leukaemia fuses the retinoic acid onstrate whether or not the hybridization to the 15q11- receptor alpha gene to a novel transcribed locus. Nature 347:558–561. 770 Smith et al.

Goy A, Passalaris T, Xiao YH, Miller WH Jr, Siegel DS, Zelenetz AD. 1995. National Center for Biotechnology Information. 1999. Bethesda, MD. The PML gene is linked to a megabase-scale insertion/deletion restric- Genemap 99 database http://www.ncbi.nlm.nih.gov/. tion fragment length polymorphism. Genomics 26:327–333. Ruthardt M, Orleth A, Tomassoni L, Puccetti E, Riganelli D, Alcalay M, Ji Y, Walkowicz MJ, Buiting K, Johnson DK, Tarvin RE, Rinchik EM, Horsthemke B, Stubbs L, Nichols R.D. 1999. The ancestral gene for Mannucci R, Nicoletti I, Grignani F, Fagioli M, Pelicci PG.1998. The transcribed, low-copy repeats in the Prader-Willi/Angelman region en- acute promyelocytic leukaemia specific PML and PLZF proteins local- codes a large protein implicated in protein trafficking, which is defi- ize to adjacent and functionally distinct nuclear bodies. Oncogene 6: cient in mice with neuromuscular and spermiogenic abnormalities. 1945–1953. Hum Mol Genet 8:533–542. Schroer RJ, Phelan MC, Michaelis RC, Crawford EC, Skinner SA, Cuccaro Kubota T, Das S, Christian SL, Baylin SB, Herman JG, Ledbetter DH. M, Simensen RJ, Bishop J, Skinner C, Fender D, Stevenson RE. 1998. 1997. Methylation-specific PCR simplifies imprinting analysis. Nat Autism and maternally derived aberrations of chromosome 15q. Am J Genet 16:16–17. Med Genet76:327–336. Le XF, Yang P, Chang KS. 1996. Analysis of the growth and transforma- tion suppressor domains of promyelocytic leukemia gene, PML. J Biol Seidel G, Prohaska R.1998. Molecular cloning of hSLP-1, a novel human Chem 271:130–135. brain-specific member of the band 7/MEC-2 family similar to Cae- Lord C, Rutter M, Goodes S.1989. Autism diagnostic observation schedule: norhabditis elegans UNC-24. Gene 225:23–29. a standardized observation of communicative and social behavior. J Steffenburg S, Gillberg CL, Steffenburg U, Kyllerman M. 1996. Autism in Autism Dev Disord 19:185–212. Angelman syndrome: a population-based study. Pediatr Neurol14:131– Lord C, Rutter M, Le Couteur A. 1994. Autism Diagnostic Interview- 136. Revised: a revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. J Autism Van Vactor D, O’Reilly AM, Neel BG.1998. Genetic analysis of protein Dev Disord 24:659–685. tyrosine phosphatases. Curr Opin Genet Dev 8:112–126.