A novel rearrangement of occludin causes brain ... - Springer Link

0 downloads 0 Views 839KB Size Report
Jun 21, 2013 - Christine Macgillivray • Ellen P. Wood • Philip D. Acott • M. Naeem Khan • Mark E. Samuels •. Jacek Majewski • Andrew Orr • Christopher R.
Hum Genet (2013) 132:1223–1234 DOI 10.1007/s00439-013-1327-y

ORIGINAL INVESTIGATION

A novel rearrangement of occludin causes brain calcification and renal dysfunction Marissa A. LeBlanc • Lynette S. Penney • Daniel Gaston • Yuhao Shi • Erika Aberg • Mathew Nightingale • Haiyan Jiang • Roxanne M. Gillett • Somayyeh Fahiminiya • Christine Macgillivray • Ellen P. Wood • Philip D. Acott • M. Naeem Khan • Mark E. Samuels Jacek Majewski • Andrew Orr • Christopher R. McMaster • Karen Bedard



Received: 9 April 2013 / Accepted: 9 June 2013 / Published online: 21 June 2013 Ó Springer-Verlag Berlin Heidelberg 2013

Abstract Pediatric intracranial calcification may be caused by inherited or acquired factors. We describe the identification of a novel rearrangement in which a downstream pseudogene translocates into exon 9 of OCLN, resulting in band-like brain calcification and advanced chronic kidney disease in early childhood. SNP genotyping and read-depth variation from whole exome sequencing initially pointed to a mutation in the OCLN gene. The high degree of identity between OCLN and two pseudogenes required a combination of multiplex ligation-dependent

M.A. LeBlanc and L.S. Penney contributed equally.

probe amplification, PCR, and Sanger sequencing to identify the genomic rearrangement that was the underlying genetic cause of the disease. Mutations in exon 3, or at the 5–6 intron splice site, of OCLN have been reported to cause brain calcification and polymicrogyria with no evidence of extra-cranial phenotypes. Of the OCLN splice variants described, all make use of exon 9, while OCLN variants that use exons 3, 5, and 6 are tissue specific. The genetic rearrangement we identified in exon 9 provides a plausible explanation for the expanded clinical phenotype observed in our individuals. Furthermore, the lack of polymicrogyria associated with the rearrangement of OCLN in our patients extends the range of cranial defects that can be observed due to OCLN mutations.

Electronic supplementary material The online version of this article (doi:10.1007/s00439-013-1327-y) contains supplementary material, which is available to authorized users. M. A. LeBlanc  D. Gaston  M. Nightingale  R. M. Gillett  A. Orr  K. Bedard (&) Department of Pathology, Dalhousie University, 5850 College St., Sir Charles Tupper Medical Building, Room 11-F, Halifax, NS, Canada e-mail: [email protected] L. S. Penney  E. P. Wood  P. D. Acott Department of Pediatrics, Dalhousie University, Halifax, NS, Canada Y. Shi  S. Fahiminiya  J. Majewski Department of Human Genetics, McGill University, Montreal, QC, Canada E. Aberg Maritime Medical Genetics, IWK Health Centre, Halifax, NS, Canada

C. Macgillivray Department of Ophthalmology, Capital Health, Halifax, NS, Canada M. N. Khan Department of Radiology, Dalhousie University, Halifax, NS, Canada M. E. Samuels Department of Medicine, University of Montreal, Montreal, QC, Canada A. Orr Department of Ophthalmology, Dalhousie University, Halifax, NS, Canada C. R. McMaster Department of Pharmacology, Dalhousie University, Halifax, NS, Canada

H. Jiang Department of Biostatistics, Princess Margaret Centre, Toronto, ON, Canada

123

1224

Introduction The finding of pediatric intracranial calcification has multiple causes, both acquired and genetic, with varying severities. An autosomal-recessive form of band-like calcification referred to as pseudo-TORCH syndrome (MIM [251290]) has been described (O’Driscoll et al. 2010; Reardon et al. 1994). This rare autosomal-recessive neurological disorder shows band-like calcification characterized by simplified gyration and polymicrogyria (BLCPMG) and has been attributed to genetic mutations in exon 3 and at the splice site between exon 5 and 6 of the OCLN gene (MIM [602876]) (O’Driscoll et al. 2010). The phenotypic overlap of pseudo-TORCH syndrome with Aicardi–Goutieres syndrome (MIM [225750]) has been recognized, although specific features remain associated with either syndrome (Crow et al. 2000, 2003). In this study, we ascertained two cases of brain calcification in a set of first cousins from a consanguineous family (Fig. 1). The disease incidence in the family is consistent with an autosomal-recessive pattern of inheritance. The individuals presented at birth with microcephaly and had seizures beginning at birth or at 2 months of age. Brain calcifications were noted on early imaging with the absence of obvious polymicrogyria. Both children demonstrated very limited growth, development, and cognitive abilities, along with advanced renal dysfunction. In these patients, we describe the identification of a novel deletion/ rearrangement of the OCLN gene, whereby an almost identical downstream OCLN pseudogene recombines into Fig. 1 Maritime brain calcification pedigree. A consanguineous Maritime Canadian family representing affected patients carrying a deletion in the OCLN gene. The proband, individual VI:3, is noted by an arrowhead. Clinically affected individuals are indicated with shaded symbols

123

Hum Genet (2013) 132:1223–1234

exon 9 of the OCLN gene. Studying and detecting genetic alterations in the later portion of the OCLN gene (beyond exon 5) was complicated by the fact that the region is duplicated and at some points triplicated.

Materials and methods Clinical ascertainment and consent Approval for the research study was obtained from the IWK Health Centre research ethics board (Project #1002678; REB file 4812). Informed consent was obtained from individuals or their guardians for all samples used in this study. DNA was obtained from blood or saliva samples using standard methods. SNP genotyping Whole-genome high-density single nucleotide polymorphism (SNP) genotype scanning was performed at the McGill University and Genome Quebec Innovation Centre, using the Illumina HumanOmni (Illumina, Inc., San Diego, CA) panel with one million markers. Data were scanned using the Bead Array Reader (Illumina, Inc.), plate Crane Ex, and Illumina BeadLab software (Illumina, Inc.). Initial quality control and export of data were done using Illumina’s GenomeStudio software. Six individuals were genotyped.

Hum Genet (2013) 132:1223–1234

Homozygosity mapping Homozygosity mapping (HM) analysis was performed on 1M resolution SNP genotyping data using the quantitative homozygosity mapping program of Huqun et al. (2010) using the two affected patients (patients #1991 and #1994) and four unaffected relatives (individuals #1992, #1996, #2000 and #2010). The HM software was run using an estimated genotyping error rate of 0.003 (default parameter) and a minimum region of homozygosity (RHS) cut-off of 2 cM. Whole exome sequencing Whole exome sequencing was performed on two affected individuals: patient #1991 and patient #1994. A total of 3 lg of DNA was used for exome capture with the Agilent SureSelect all exon 38 Mb kit. Massively parallel sequencing was performed with 100-bp paired-end reads using the Illumina HiSeq 2000 sequencer, at the McGill University and Genome Quebec Innovation Centre as previously described (Alfares et al. 2011). Reads were assembled against the human genomic reference sequence (hg19) using the Burrows-Wheeler Aligner (BWA) (Li and Durbin 2009). Genomic variants were called using the genome analysis toolkit (GATK) pipeline (McKenna et al. 2010), and annotated with Annovar (Wang et al. 2010). All variants were compared against dbSNP, 1,000 Genomes Project and a pool of over 300 in-house exomes from unrelated projects (to help filter artifactual false-positive variant calls). Potentially damaging variants included nonsynonymous mutations (missense and nonsense), splicesite variants, and frameshift changes due to insertions and/ or deletions (indels). Exome variants were further filtered by their location in regions of the genome not excluded by HM. This filtering strategy led to the identification of 20 candidate genes. Mutations in candidate genes found in the exomes of two affected subjects were tested by Sanger sequencing, and in the case of OCLN by multiplex ligationdependent probe amplification (MLPA), and were then screened for in other related patients by Sanger sequencing. Copy number variation analysis We analyzed the exomes of the two individuals for copy number variations with a novel algorithm based on exome read depth (Shi and Majewski 2013). Read depth of the two individuals were compared against the distribution of reads in 70 unrelated controls using the same exome sequencing protocol. For each sample, we normalized the read depth at each exon to reads per million mapped reads (RPKM) (Mortazavi et al. 2008). We then segmented the affected exomes into regions of similar sample depth to control

1225

read-depth ratio using DNAcopy (Olshen et al. 2004). Finally, for each region in the affected individuals we calculated the statistical deviation (as a p-value) from the distribution of the controls. Microarray was performed to look for clinically significant deletions or duplications. Microarray-based comparative genomic hybridization (aCGH) was performed using an oligonucleotide-based 125-K feature whole-genome microarray (SignatureChip Oligo SolutionTM version 3.0, custom designed by Signature Genomics, manufactured by Roche NimbleGen, Madison, WI), according to previously described methods (Duker et al. 2010). Clinical oligonucleotide-based CGH results were normal. Multiplex ligation-dependent probe amplification (MLPA) Genomic DNA was examined for insertions or deletions by MLPA analysis using the MLPA P200-A1 kit (MRCHolland, Amsterdam, Holland) along with custom probes. Custom probes were designed to either target a discriminating nucleotide between the OCLN gene and the pseudogene(s) or to amplify both the OCLN gene and the pseudogene(s) (Table S1). Primers designed to target discriminating nucleotides have the discriminating nucleotides located at the 30 end of the 50 primer (highlighted in green). Primers designed to target both the OCLN gene and the pseudogene(s) have no discriminating nucleotides within their sequence. The MLPA was validated using probes directed to the SAG gene (NG_009116.1) in a family with patients affected by Oguchi syndrome (MIM [258100]), previously identified as having two, one or no copies of the gene in unaffected or affected family members, respectively (Guernsey 2008, unpublished data). Data were analyzed using GeneMarker V.1.6 (Soft Genetics, Inc.). PCR and direct Sanger sequencing Selected regions were amplified from genomic DNA by PCR. Amplified fragments were purified from an agarose gel, and sequenced using Sanger fluorescent sequencing and capillary electrophoresis. Sequencing was performed at the DNA Diagnostic Laboratory at the IWK Health Centre in Halifax, Nova Scotia, Canada. Sequence traces were analyzed using MutationSurveyor V.3.97 (Soft Genetics, Inc.). To allow sequencing of individual chromosomal copies of DNA sequences, TA-cloning of PCR product was performed when needed using the pCR2.1-TOPO cloning kit from Invitrogen according to manufacturer’s instructions. To rule out Aicardi–Goutieres syndrome (AGS) as the cause of the genetic disease in the proband (patient #1991),

123

1226

PCR and Sanger sequencing was performed on the entire coding regions and at the highly conserved exon–intron splice junctions of known causative genes. These included TREX1 (AGS1/AGS5), RNASEH2B (AGS2), RNASEH2C (AGS3) and RNASEH2A (AGS4). This sequencing was performed at the clinical lab, Centogene, in Germany.

Results Clinical assessment and phenotype We ascertained a total of eight DNA samples from a family with an apparent autosomal-recessive form of brain calcification. These samples comprised the proband (patient #1991), her unaffected mother and father (#1992 and #2000), the paternal sister (#2009), her partner (#1996) and his father (#1995), as well as their affected daughter (patient #1994) and unaffected son (#2010) (Fig. 1). A more complete clinical description of both patients is included in the supplemental data. The proband (patient #1991) is a 4.5-year-old female. She weighed 2.87 kg at birth and was noted to be microcephalic. A cranial ultrasound performed at the delivering hospital identified bilateral calcifications. At 5 weeks of age she displayed severe microcephaly [occipital–frontal circumference: 32.5 cm (\3rd percentile)] and failure to thrive [weight: 2.92 kg (\3rd percentile)]. A computed tomography (CT) scan of the brain was performed at 5 weeks and compared to normal control CTs (Fig. 2a, c). Our patient’s CT (#1991) revealed bilateral, symmetrical calcifications involving the ventrolateral thalami and asymmetrical calcifications in the basal ganglia. There was bilateral subcortical calcification with possibly some involvement of the cortex. The cerebellum was spared. The lateral ventricles were slightly enlarged. A follow-up head magnetic resonance imaging (MRI) identified the same pattern of calcification compared to normal controls, and we also noted that there was no evidence of a migrational defect in the patient (Fig. 2b, d). A follow-up CT scan performed at 3 months identified progression of the calcifications of the thalami, basal ganglia and subcortical white matter. At 5 months, patient #1991 was found to have hypercalcemia (peak calcium = 3.01 mmol/l) and hypernatremia (peak sodium = 161 mmol/l). Renal imaging identified small echogenic kidneys. Follow-up renal ultrasounds continued to show asymmetry of kidney size and evolution of bilateral echogenic changes consistent with cortical calcifications. A diethylene-triamine-penta-acetic acid (DTPA) renal scan confirmed asymmetry of function (right = 36 %, left = 64 %) and significant reduction of glomerular filtration rate (GFR) at 18.0 ml/min/1.73 m2, confirming chronic kidney disease (CKD) stage 4 (CKD

123

Hum Genet (2013) 132:1223–1234

stage 4 = GFR of 15–30 ml/min/1.73 m2). Three annual follow-up DTPA renal scans continue to confirm CKD stage 4. Tube feedings were implemented to ensure adequate fluid intake of 150 % of maintenance water requirements. The patient (#1991) developed seizures at 2 months of age, with lip smacking and twitching. This progressed to status epilepticus at 3 months. She has profound developmental impairment, cerebral palsy and failure to thrive. She is nonverbal and largely non-communicative. She has no independent mobility and no independent use of her hands. She is followed by ophthalmology for cortical visual impairment with nystagmus and has only slight response to light. There is no evidence of liver or cardiac dysfunction and no evidence of thrombocytopenia or involvement of other organs. Investigations were not suggestive of a TORCH infection, alpha-interferon levels in CSF were normal, and no mutations were discovered on sequencing of the TREX1, RNASEH2B, RNASEH2C and RNASEH2A genes associated with Aicardi–Goutieres syndrome. A microarray analysis on the DNA of this patient did not identify any clinically significant deletions or duplications. Review of the proband’s family history revealed a female paternal first cousin (patient #1994) with a seizure disorder, cerebral palsy, global developmental delay and brain calcifications. This child was born at term [weight: 3.3 kg (25th–50th percentile)] with microcephaly [head circumference: 32 cm (3rd percentile)]. The child developed neonatal seizures consisting of bicycling movements at 10 h of age. A head CT scan at 1 day of age revealed underdevelopment of the frontal lobes and a small region possibly suspicious for polymicrogyria in the perisylvian region of the left frontal lobe (Fig. 2e). There was dense calcification in the subcortical white matter of the bilateral frontal lobes, the lentiform nuclei and the anterior thalami. A head MRI at 2.5 years revealed a markedly dilated ventricular system with progressive cortical and subcortical brain atrophy (Fig. 2f). There was the appearance of linear subcortical, thalami and lentiform nuclei calcifications in a band-like pattern. TORCH testing was normal. This patient (#1994) also has significant renal dysfunction. She displayed acute renal failure that was responsive to intravenous fluid replacement and a baseline hypernatremia (peak sodium = 168 mmol/l). Her renal ultrasound confirmed small echo-dense kidneys in the typical location without hydronephrosis or structural anomalies. A DTPA renal scan at 6 year 5 months confirmed asymmetry of function (right = 38 %; left = 62 %) and significant reduction of GFR at 34.3 ml/min/1.73 m2 CKD stage 3 (CKD stage 3 = GFR of 30–60 ml/min/1.73 m2). Normalization of her sodium was achieved with provision of adequate fluid intake of 175 % of maintenance water requirements.

Hum Genet (2013) 132:1223–1234

Fig. 2 CT and MRI from normal control patient, proband (sample #1991) and proband’s cousin (sample #1994). a Normal control CT and b normal control MRI. c Patient #1991 CT at 5 weeks of age demonstrates thalamic and basal ganglia calcification along with mild band-like calcification in subcortical location of both frontal lobes. d Patient #1991 MRI at 5 weeks of age. Note the band-like gyriform calcification at the grey–white junction and mild degree of atrophy present in the frontal lobes with no migrational abnormality,

1227

specifically polymicrogyria. e Patient #1994 CT scan at the age of 1 day shows atrophy and lack of gyral pattern in both frontal lobes, and calcification in bilateral subcortical locations of frontal lobes, basal ganglia and thalami. Note the enlargement of ventricles. f Patient #1994 MRI at the age of 2 years and 6 months shows marked progression of atrophy of both cerebral hemispheres and brainstem

123

1228

Hum Genet (2013) 132:1223–1234

Fig. 3 Identification of a deletion found within the shared genomic region of affected individuals. a Ideogram of results from quantitative homozygosity mapping, where each horizontal bar represents a chromosome. Two neighboring runs of homozygous SNPs (RHS) for both affected individuals, spanning 2 cM or more, are identified on chromosome 5 and are shaded in black. The separation between the two RHS was less than 1 Mb and they were merged together for all future analyses. b Exome data analysis identified a deletion within exon 9 of OCLN. In control patients, the read-depth appears normal while only four reads were observed in the affected individuals. Data

visualized using integrated genomics viewer (Thorvaldsdottir et al. 2012; Robinson et al. 2011). c The OCLN gene is located on chromosome 5 (q.13.2). In addition to the OCLN gene (blue), q13.2 also contains pseudogene I (brown) and pseudogene II (green), with their orientation indicated by arrows. Each gene and pseudogene is divided in numbered *1 kb fragments identified by the name of the primers used to amplify them. Discriminating nucleotides (DN) used in the amplicons are available and presented in Table 1. Black regions between the OCLN gene and pseudogenes are not to scale and represent the genomic region found in that area

At 6.5 years of age, this patient (#1994) is wheelchair dependent with no head or trunk control, and is nonverbal. Her growth is significantly below the third percentile but following her own curve. Her head circumference is significantly below the second percentile. She is followed by ophthalmology for exotropia and significant cortical visual impairment with a normal retinal exam. There is no evidence of liver or cardiac dysfunction and no evidence of thrombocytopenia or involvement of other organs. A more complete clinical description of both patients is included in the supplemental data.

Defining the region by homozygosity mapping

123

The two affected individuals in this family are first cousins. The absence of the clinical phenotype in the parents of both affected individuals, and the presence of consanguinity within the family, suggests the likelihood of a homozygous mutation as the cause of this disorder. We performed SNP genotyping using a high-density SNP panel. Homozygosity mapping using the quantitative homozygosity mapping (qHM) software (Huqun et al. 2010) identified a large region of approximately 15.6 Mbp between 58,353,809 and

Hum Genet (2013) 132:1223–1234

1229

Fig. 4 Identification of the deleted region within exon E09 of OCLN by multiplex ligation-dependent probe amplification (MLPA). MLPA probes for the coding region of E09 and the 30 UTR were designed to amplify only the OCLN gene and not the pseudogene(s). The discriminating nucleotide specific to the OCLN gene is located at the 30 end of the 50 primer, therefore ligation between the two probes will not occur within the pseudogenes. Probes for exons E02, E07, E08

amplify both the OCLN gene and pseudogene(s). The size (bp) of the amplicon is indicated across the top horizontal axis, and the vertical axis indicates signal intensity (arbitrary units). a MLPA probes for the coding region of exon E09 and the 30 UTR do not amplify in affected samples. All other probes for OCLN, the size standard (red), and the probe for a control gene, SAG, amplify in patients. b All MLPA probes amplify in the control samples

Table 1 Discriminating nucleotides identified within the OCLN gene and pseudogene(s) Section

Position in OCLN gene

Discriminating nucleotide ID

OCLN (BC029886.1)

Pseudogene(s) (LOC100170939, LOC647859)

Patient

Control

39

Chr5: 68,847,081

DN39A

A

G

A/G

A/G

40

Chr5: 68,848,415

DN40A

C

G

G

C/G

42

Chr5: 68,849,631 Chr5: 68,850,198

DN42A DN42B

C C

T G

T G

C/T C/G

43

Chr5: 68,850,644

DN43A

G

Ca

C

G/C

a

Chr5: 68,851,387

DN43B

T

C

C

T/C

44

Chr5: 68,851,563

DN44A

T

Ca

C

T/C

46

Chr5: 68,853,486

DN46A

A

Ga

A/G

A/G

Chr5: 68,853,907

DN46B

A

Ga

A/G

A/G

Chr5: 68,854,696

DN47A

T

Ga

T/G

T/G

Chr5: 68,855,071

DN47B

C

Ta

C/T

C/T

a

47 48 a

Chr5: 68,855,412

DN48A

C

T

C/T

C/T

Chr5: 68,855,766

DN48B

A

Ca

A/C

A/C

Represents discriminating nucleotide (DN) for two pseudogenes locations

123

1230

74,002,657 (hg19) on chromosome 5, which was shared between the two affected individuals (Fig. 3a). A small apparent break of \1 Mb was observed within the homozygous shared region; however, the outer bounds of the entire region were used for all analyses. Sequencing of genes within the homozygous region The largest candidate region on chromosome 5 contained 83 annotated genes. We started by Sanger sequencing the coding regions of a subset of candidate genes within the region that included SGTB, OCLN, TRIM23, RGS7BP, HTR1A, ANKRA2, FOXD1, FCH02, MAP1B, MARVELD2, SLC30A5, C5orf43, FLJ37543, KIF2A, DIMT1L, RNF180, SREK1IP1, NLN, TMEM171, and NAIP. No point mutations were observed in any genes sequenced. It should be noted that although the coding exons of OCLN had been sequenced and no mutations were observed, the lack of discriminating nucleotides between the OCLN gene and the pseudogene in the coding portion of exon 09 would explain why a deletion or rearrangement was not identified using this method. We then moved to whole exome sequencing on the two affected individuals. Again, no obvious point mutations were identified in any of the genes within the candidate autozygous region, or elsewhere in the genome. However, within this region there were 28 genes that contained exon(s) that were not covered by whole exome sequencing. These exons were further analyzed by PCRbased Sanger sequencing. Four of these genes fell in a duplicated region and were not re-sequenced. One exon could not be amplified in either control samples or in the affected samples. The exons for the other 24 genes were sequenced and no point mutations were observed. Data from the whole exome sequences were examined for evidence of structural or copy number variations. The algorithm (Shi and Majewski 2013) prioritizes rare variants and it identified a deletion in the last exon of OCLN at a significance of p \ 0.005. This exon is well covered in all the control samples, with an average of 400 reads mapping to the exon (Fig. 3b), while only four reads were detected in the affected samples. Amplification of the homologous pseudogene (Fig. 3c) with a sequencing error making the product similar to the OCLN gene, or a small amount of contamination, could account for this observation even in the presence of a deletion. The data implied that the affected individuals may carry a homozygous deletion in exon 9 of the OCLN gene. Verification of the deletion within OCLN by MLPA Exon 9 encompasses the last portion of the coding region of OCLN as well as the 30 UTR. Although the genomic region including exons 5–9 of OCLN is duplicated, and an

123

Hum Genet (2013) 132:1223–1234

area starting in the 30 UTR is triplicated in the consensus human genome sequence, there are several locations where individual nucleotides differ between the true gene and the pseudogenes. These differentiating nucleotides allow for the design of amplicons for detecting copy number that are specific to the true gene. Using MLPA with probes designed to specifically hybridize with the OCLN gene, we observed a loss of signal within exon 9. Probes designed to target either the coding region of exon 9 or the 30 UTR did not amplify; however, MLPA probes designed to amplify exons 2, 7 and 8 confirmed each was present (Fig. 4a, b). Verification of the deletion within OCLN by PCR and DNA sequencing To further define the deleted region of the OCLN, PCR amplification was performed across specific OCLN segments. PCR primers were designed to amplify ten fragments, each of approximately 1,000 base pairs (bp), within the region where the deletion was suspected (Table S2). From these, segments were selected based on the presence of a nucleotide that could discriminate between the OCLN gene and the pseudogenes (Table 1). This allowed us to determine by Sanger sequencing if we had amplified the OCLN gene, the pseudogene(s), or both (Figure S1A and S1B). Four amplicons, labeled F40/R40, F42/R42, F43/ R43 and F44/R44 detected only the pseudogene(s) in patients, and detected both the OCLN gene and the pseudogene(s) in unaffected controls (Fig. 5a). If the absence of the OCLN gene in patients was due to a straightforward deletion occurring between F40 and R44, we would expect F39/R46 to amplify an 8,000-bp product representing the OCLN gene and the pseudogene I in the control samples, and to amplify an 8,000-bp product representing only pseudogene I in patients. Under this model, we would also predict a smaller product with a minimum size of 2,000 bp (since F39/R39 and F46/R46 amplify 1,000-bp fragments in both the OCLN gene and the pseudogene). However, no smaller product was detected. Furthermore, we confirmed based on discriminating nucleotides DN39A, DN42A, DN42B, and DN44A, that the 8,000-bp product represented both the OCLN gene and the pseudogene in controls, but only the pseudogene in the patients (data not shown). To determine where the breakpoint for the deletion was located, a series of PCR products using the forward primer F39 and reverse primers at increasing distances were amplified. Primer R39 was used to sequence and detect the discriminating nucleotide DN39A. In patients, only the pseudogene was detected in all amplicons from F39/R40 to F39/R46, while in controls all segments representing both the OCLN gene and the pseudogene were detected (Fig. 5b).

Hum Genet (2013) 132:1223–1234

Fig. 5 The use of discriminating nucleotides between the OCLN gene and the pseudogene(s) by PCR and Sanger sequencing identifies a structural rearrangement within the OCLN gene. a The OCLN gene (blue), the pseudogene I (brown) and the pseudogene II (green) are divided into numbered *1 kb fragments identified by the name of the primers used to amplify them. Amplicons of these *1 kb fragments using primers F39/R39, F46/R46, F47/R47 and F48/R48 are from both the OCLN gene and the pseudogene(s), in both affected and control samples. Using primers F40/R40, F42/R42, F43/R43 and F44/ R44, only the pseudogene(s) was amplified in affected samples. b PCR products of increasing size were amplified using primer F39 with different reverse primers, and sequenced at position DN39A. Only the F39/R39 amplicon yielded both the OCLN gene and the pseudogene in affected samples. All other amplicons in the affected samples were only identified as the pseudogene according to DN39A. Control samples amplified both the OCLN gene and the pseudogene in all amplicons. c PCR was used to amplify DNA from patients using primers F42/R48 and the product was TA-cloned and sequenced at

1231

positions DN48A and DN48B, where the real gene was detected and at position DN42A, where the pseudogene was detected. The controls had either both locations as the real gene or both locations as the pseudogene. d A heterozygous call at DN48A/B was identified in fragment F43/R48 in both affected and control samples due to the amplification of both the pseudogene II and the fusion of OCLN gene to pseudogene I. e Proposed model for the rearrangement on chromosome 5. The pseudogene I (brown) translocates into the OCLN gene disrupting the normal genomic sequence. Some parts of the translocated region from F39 to F46 of pseudogene I may remain within its original region (beige). However, the remaining region (beige) cannot be entirely intact, as no pseudogene fragment was amplified with F42/R48. Pseudogene II (green) remains in its original form. The OCLN gene (blue) is now disrupted with genomic DNA from the pseudogene I along with extra genomic material represented by the black box. The contents and size of the whole genomic DNA incorporation is unknown

123

1232

Working from the 30 end, we amplified a segment with F42/R48 in patients and controls. As expected, controls were heterozygous for both the OCLN gene and the pseudogene according to the sequencing results at DN48A and DN48B. However, patients were homozygous for the OCLN gene at DN48A and DN48B. No pseudogene(s) was detected at these discriminating nucleotides. In this same segment, the patients were homozygous for the pseudogene at DN42A (Fig. 5c). We considered the possibility that because of the complexity of this region, an annotation error might exist in the database for hg19. Cloning and sequencing of the segment F42/R48 from controls found the database to be correct. In controls, clones either had both DN42A and DN48B corresponding to the OCLN gene or had both corresponding to the pseudogene. Cloning and sequencing of F42/R48 from affected individuals, on the other hand, showed evidence of a rearrangement. The segment amplified corresponded to the pseudogene at DN42A, DN43A and DN44A but the OCLN gene at position DN46A, DN47A, DN48A and DN48B (Fig. 5c). Similarly, the fragment F39/R46 amplified from an affected individual had the pseudogene at DN39A and the OCLN gene at DN46A (data not shown). This is suggestive of a rearrangement of pseudogene I into the OCLN gene. As we were not able to amplify a product with F39 from the OCLN gene paired with any reverse primer from the translocated pseudogene I, we predict that a large amount of DNA linked to pseudogene I was also translocated. Furthermore, the fragment F43/R48 was heterozygous for the OCLN gene and pseudogene at positions DN48A and DN48B in patients and controls. This is consistent with the fact that these primers would be expected to amplify the triplicated section, pseudogene II, as well as the disrupted OCLN gene fused with the pseudogene I (Fig. 5d). Based on the observation that segments between F40/ R40 and F44/R44 only amplify the pseudogene and do not detect any OCLN gene in patient samples (Fig. 5a), and based on the fact that the section amplified by F42/R48 (Fig. 5c) is not present in pseudogene II and therefore must have come from pseudogene I, we propose a model where a section of pseudogene I, along with a larger section of genomic DNA, is translocated into the OCLN gene (Fig. 5e).

Discussion Herein, we describe the identification of a novel deletion/ rearrangement of OCLN that results in band-like brain calcification, limited cognitive abilities, failure to thrive and renal dysfunction in early childhood. Identification of this genomic rearrangement is particularly interesting, as it was complicated by almost identical downstream

123

Hum Genet (2013) 132:1223–1234

pseudogenes of the latter part of OCLN. Mutations in other parts of the OCLN gene not duplicated within the genome, namely in exon 3 and at the exon 5/6 intron junction, have been reported to cause brain calcification with patients presenting with microcephaly, polymicrogyria, seizures, and severely impaired development but with no reported extra-cranial phenotype (O’Driscoll et al. 2010). In contrast, our proband had no evidence of polymicrogyria upon routine cranial MRI and CT and the proband’s cousin had only subtle suggestions of polymicrogyria on a cranial CT. Additionally, our patients demonstrated primarily subcortical calcification at the grey–white junction while the calcification in the previous report was cortical (O’Driscoll et al. 2010). Also differing from previous cases, both our patients exhibited significant renal impairment with inadequate ability to concentrate urine with resultant hypernatremic dehydration at baseline with aggravation during intercurrent illness. Renal dysfunction had not been included in the phenotype of the patients previously described with OCLN mutations (O’Driscoll et al. 2010). However, on further review, a patient from family F275 (O’Driscoll et al. 2010) was reported to have died from renal failure in the original description of the patient (Abdel-Salam and Zaki 2009). Additionally, a patient from family F386 was reported to have mild renal dysplasia (O’Driscoll et al. 2010), and a patient from family F375 (O’Driscoll et al. 2010) in the original description was reported to have several episodes of unexplained hypernatremia (Briggs et al. 2008). The renal phenotype was not further expanded upon in these families and there did not appear to be a consistent pattern of renal disease among the patients. The renal dysfunction experienced by our patients appears to be more severe and consistent than the previously reported cases. Functionally redundant isoforms of OCLN may explain the previous human phenotypes observed due to OCLN mutations in exon 3 or at the splice site between exon 5 and 6, which were confined to the brain and characterized by early-onset seizures, severe progressive microcephaly, and developmental arrest (O’Driscoll et al. 2010). There are at least seven alternatively spliced isoforms of OCLN although not all make use of exons 3–6. While it appears that there are no tissues where the dominant isoform is one that lacks exon 3 (OCLN-ex3del, OCLN-ex3del-4del, OCLN-ex3partdel), many peripheral tissues express one or all of these isoforms to some extent. Therefore, it is possible that in patients where the mutation is limited to exon 3, the other isoforms can compensate in the periphery. The brain has very limited expression of OCLN-ex3del, and no expression of the other two isoforms lacking exon 3, suggesting that the brain may be dependent on isoforms containing exon 3. Exon 9, however, is found in all known OCLN isoforms and its deletion likely explains the

Hum Genet (2013) 132:1223–1234

phenotypes in our cohort that extend beyond the brain. Therefore, we speculate that the significant change in renal development and function observed in our patients relate to the loss of peripheral OCLN. In the previous report, a single patient with a mutation affecting splicing at exon 5–6 was also identified (O’Driscoll et al. 2010). The patient was described at 3 months of age with the phenotype confined to the brain. There are OCLN isoforms that lack exons 5 and 6, although to what extent these isoforms are expressed in other tissues are not known. Alternatively, it is conceivable that this individual could develop other clinical features that have not yet been detected. Understanding the consequences of isoform-specific mutations versus subtle differences that may exist between individuals that have alternative isoforms present in their periphery, and individuals where the expression of OCLN is expected to be completely lost, could potentially shed light on the function of this protein. In addition, the lack of polymicrogyria and the extra-cranial phenotype in our patients extends the range of phenotypes that can be observed due to OCLN mutations, and expands our understanding of the protein function. Consistent with a total loss of OCLN manifesting peripheral phenotypes, the Ocln-/- mouse also displays developmental impairment and brain calcification, and shows signs of peripheral abnormalities (Saitou et al. 2000). However, outside of the previously described brain phenotype and newly described renal phenotype, our patients did not have obvious involvement of other organs. This study also draws attention to overcoming the challenges associated with detecting mutations in genes where there are pseudogenes present. The University of Chicago Genetic Services offers testing for mutations in OCLN associated with pseudo-TORCH, either alone or as part of its polymicrogyria panel, which covers exons 2–5 along with their intron–exon boundaries. This screen would not have detected the OCLN mutation in our patients. A test that includes screening for copy number variation in OCLN would be a useful addition for the molecular diagnosis of individuals with this now expanded phenotype, which appears to not necessarily include polymicrogyria. Furthermore, this test would identify mutation carriers who are at risk of imparting this recessive disorder to their children. Acknowledgments We are grateful to the family members who generously contributed their time and materials for this research. The authors would like to acknowledge the significant contributions to this work that were made by Dr. Duane Guernsey. Dr. Guernsey was a valued member of our research team who passed during the completion of this project. The following agencies provided funding for this project: Genome Canada, Genome Atlantic, Nova Scotia Health Research Foundation, Nova Scotia Research and Innovation Trust, Dalhousie Faculty of Medicine, Capital District Health Authority, IWK Health Centre Foundation, and Capital Health Research Fund.

1233 The authors would like to acknowledge the contribution of the Genome Quebec High Throughput Sequencing Platform. M.A.L is supported by a trainee award from the Beatrice Hunter Cancer Research Institute with funds provided by The Terry Fox Strategic Health Research Training Program in Cancer Research from CIHR. M.E.S is supported by the Centre de Recherche du CHU Ste-Justine. Conflict of interest of interest.

The authors declare that there are no conflicts

References Abdel-Salam GM, Zaki MS (2009) Band-like intracranial calcification (BIC), microcephaly and malformation of brain development: a distinctive form of congenital infection like syndromes. Am J Med Genet A 149A(7):1565–1568. doi:10.1002/ajmg. a.32894 Alfares A, Nunez LD, Al-Thihli K, Mitchell J, Melancon S, Anastasio N, Ha KC, Majewski J, Rosenblatt DS, Braverman N (2011) Combined malonic and methylmalonic aciduria: exome sequencing reveals mutations in the ACSF3 gene in patients with a nonclassic phenotype. J Med Genet 48(9):602–605. doi: 10.1136/jmedgenet-2011-100230 Briggs TA, Wolf NI, D’Arrigo S, Ebinger F, Harting I, Dobyns WB, Livingston JH, Rice GI, Crooks D, Rowland-Hill CA, Squier W, Stoodley N, Pilz DT, Crow YJ (2008) Band-like intracranial calcification with simplified gyration and polymicrogyria: a distinct ‘‘pseudo-TORCH’’ phenotype. Am J Med Genet A 146A(24):3173–3180. doi:10.1002/ajmg.a.32614 Crow YJ, Jackson AP, Roberts E, van Beusekom E, Barth P, Corry P, Ferrie CD, Hamel BC, Jayatunga R, Karbani G, Kalmanchey R, Kelemen A, King M, Kumar R, Livingstone J, Massey R, McWilliam R, Meager A, Rittey C, Stephenson JB, Tolmie JL, Verrips A, Voit T, van Bokhoven H, Brunner HG, Woods CG (2000) Aicardi-Goutieres syndrome displays genetic heterogeneity with one locus (AGS1) on chromosome 3p21. Am J Hum Genet 67(1):213–221. doi:10.1086/302955 Crow YJ, Black DN, Ali M, Bond J, Jackson AP, Lefson M, Michaud J, Roberts E, Stephenson JB, Woods CG, Lebon P (2003) Cree encephalitis is allelic with Aicardi-Goutieres syndrome: implications for the pathogenesis of disorders of interferon alpha metabolism. J Med Genet 40(3):183–187 Duker AL, Ballif BC, Bawle EV, Person RE, Mahadevan S, Alliman S, Thompson R, Traylor R, Bejjani BA, Shaffer LG, Rosenfeld JA, Lamb AN, Sahoo T (2010) Paternally inherited microdeletion at 15q11.2 confirms a significant role for the SNORD116 C/D box snoRNA cluster in Prader–Willi syndrome. Eur J Hum Genet (EJHG) 18(11):1196–1201. doi:10.1038/ejhg.2010.102 Huqun, Fukuyama S, Morino H, Miyazawa H, Tanaka T, Suzuki T, Kohda M, Kawakami H, Okazaki Y, Seyama K, Hagiwara K (2010) A quantitatively-modeled homozygosity mapping algorithm, qHomozygosityMapping, utilizing whole genome single nucleotide polymorphism genotyping data. BMC Bioinform 11 (Suppl 7):S5. doi:10.1186/1471-2105-11-S7-S5 Li H, Durbin R (2009) Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25(14):1754–1760. doi:10.1093/bioinformatics/btp324 McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, Garimella K, Altshuler D, Gabriel S, Daly M, DePristo MA (2010) The genome analysis toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res 20(9):1297–1303. doi:10.1101/gr.107524.110

123

1234 Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B (2008) Mapping and quantifying mammalian transcriptomes by RNASeq. Nat Methods 5(7):621–628. doi:10.1038/nmeth.1226 O’Driscoll MC, Daly SB, Urquhart JE, Black GC, Pilz DT, Brockmann K, McEntagart M, Abdel-Salam G, Zaki M, Wolf NI, Ladda RL, Sell S, D’Arrigo S, Squier W, Dobyns WB, Livingston JH, Crow YJ (2010) Recessive mutations in the gene encoding the tight junction protein occludin cause band-like calcification with simplified gyration and polymicrogyria. Am J Hum Genet 87(3):354–364. doi:10.1016/j.ajhg.2010.07.012 Olshen AB, Venkatraman ES, Lucito R, Wigler M (2004) Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics 5(4):557–572. doi:10.1093/biosta tistics/kxh008 Reardon W, Hockey A, Silberstein P, Kendall B, Farag TI, Swash M, Stevenson R, Baraitser M (1994) Autosomal recessive congenital intrauterine infection-like syndrome of microcephaly, intracranial calcification, and CNS disease. Am J Med Genet 52(1): 58–65. doi:10.1002/ajmg.1320520112

123

Hum Genet (2013) 132:1223–1234 Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz G, Mesirov JP (2011) Integrative genomics viewer. Nat Biotechnol 29(1):24–26. doi:10.1038/nbt.1754 Saitou M, Furuse M, Sasaki H, Schulzke JD, Fromm M, Takano H, Noda T, Tsukita S (2000) Complex phenotype of mice lacking occludin, a component of tight junction strands. Mol Biol Cell 11(12):4131–4142 Shi Y, Majewski J (2013) FishingCNV: a graphical software package for detecting rare copy number variations in exome-sequencing data. Bioinformatics. doi:10.1093/bioinformatics/btt151 Thorvaldsdottir H, Robinson JT, Mesirov JP (2012) Integrative genomics viewer (IGV): high-performance genomics data visualization and exploration. Brief Bioinform. doi:10.1093/bib/ bbs017 Wang K, Li M, Hakonarson H (2010) ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 38(16):e164. doi:10.1093/nar/gkq603