Clinical Significance of Copy Number Variants Involving KANK1 In

European Journal of Medical Genetics 62 (2019) 15–20

Contents lists available at ScienceDirect

European Journal of Medical Genetics

journal homepage: www.elsevier.com/locate/ejmg

Clinical signiﬁcance of copy number variants involving KANK1 in patients with neurodevelopmental disorders T

∗ Rena J. Vanzoa, , Hope Twedea, Karen S. Hoa,b, Aparna Prasada, Megan M. Martina, Sarah T. Southa, E. Robert Wassmana a Lineagen, Inc., Salt Lake City, UT, United States b University of Utah, Department of Pediatrics, United States

ABSTRACT

Copy number variants (CNV)s involving KANK1 are generally classified as variants of unknown significance. Several clinical case reports suggest that the loss of KANK1 on chromosome 9p24.3 has potential impact on neurodevelopment. These case studies are inconsistent in terms of patient phenotype and suspected pattern of inheritance. Further complexities arise because these published reports utilize a variety of genetic testing platforms with varying resolution of the 9p region; this ultimately causes uncertainty about the impacted genomic coordinates and gene transcripts. Beyond these case reports, large case-control studies and publicly available databases statistically cast doubt as to whether variants of KANK1 are clinically significant. However, these large data sources are neither easily extracted nor uniformly applied to clinical interpretation. In this report we provide an updated analysis of the data on this locus and its potential clinical relevance. This is based on a review of the literature as well as 28 patients who harbor a single copy number variant involving KANK1 with or without DOCK8 (27 of whom are not published previously) identified by our clinical laboratory using an ultra-high resolution chromosomal microarray analysis. We note that 13 of 16 patients have a documented diagnosis of autism spectrum disorder (ASD) while only two, with documented perinatal complications, have a documented diagnosis of cerebral palsy (CP). A careful review of the CNVs suggests a transcript-specificeffect. After evaluation of our case series and reconsideration of the literature, we propose that KANK1 aberrations do not frequently cause CP but cannot exclude that they represent a risk factor for ASD, especially when the coding region of the shorter, alternate KANK1 transcript (termed “transcript 4” in the UCSC Genome Browser) is impacted.

1. Introduction variant's clinical relevance. However, when they include differing phenotypes leading to conflicting conclusions, case reports fail to When interpreting genomic data, variants of unknown significance clarify the variant interpretation process. When available, data from (VOUS) continue to present a challenge for laboratorians and clinicians large case-control studies may help better understand these differences alike. A variant is more easily interpreted when it impacts a gene that and support a variant's pathogenicity (or lack thereof). has been previously associated with the clinical symptomatology ob- For example, there is conflicting information published in case re- served in the patient. There is substantial evidence to suggest that ports and large case-control studies regarding copy number variation of chromosomal microarray (CMA) is warranted for individuals with KANK1 on chromosome 9p24.3. The KANK1 gene (also called autism spectrum disorder (ASD) (Manning and Hudgins, 2010) and ANKRD15, KN motif and ankyrin repeat domains 1) encodes a protein cerebral palsy (CP) (Oskoui et al., 2015), so this difficulty with variant that regulates actin polymerization and cell motility and has two major interpretation in such patients is increasingly encountered. Moreover, alternately spliced transcripts with different promoters and several these neurodevelopmental sequelae are both nonspecific and highly other possible transcripts (Sarkar et al., 2002; Kent et al., 2002). The heterogeneous, confounding the determination of whether, or to what longest transcript, NM_015158 (termed “variant 1” in the UCSC extent, a variant contributes to the patient's phenotype. Genome Browser build hg19 and referred to in this report as “transcript Clinical case reports may be the first step to understanding a 1”; genomic coordinates chr9:504,695-746,106) is longer by 158 amino

∗ Corresponding author. c/o Lineagen, Inc, 2677 E Parleys Way, Salt Lake City, UT 84109, United States. E-mail addresses: [email protected] (R.J. Vanzo), [email protected] (H. Twede), [email protected] (K.S. Ho), [email protected] (A. Prasad), [email protected] (M.M. Martin), [email protected] (S.T. South), [email protected] (E.R. Wassman). https://doi.org/10.1016/j.ejmg.2018.04.012 Received 3 February 2017; Received in revised form 18 February 2018; Accepted 22 April 2018 Available online 03 May 2018 1769-7212/ © 2018 The Authors. Published by Elsevier Masson SAS. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/). R.J. Vanzo et al. European Journal of Medical Genetics 62 (2019) 15–20 acids and shows tissue specific expression, predominantly in heart and in these individuals, we provide an updated perspective regarding the kidney. One of the shorter, evolutionary conserved transcripts, NM_ potential pathogenicity of CNVs that overlap the KANK1 locus on 153186 (termed “variant 4” in the UCSC Genome Browser build hg19 9p24.3. This includes theoretical assessment of potential transcript- and referred to in this report as “transcript 4”; genomic coordinates specific consequences in 16 patients (15 previously unreported) tested chr9:706,806-746,106) is ubiquitously expressed (Kakinuma et al., in our clinical laboratory. Our review of the data suggests that a critical 2009). Transcript 4 contains 11 exons, and exon 1 is non-coding. As a element for interpretation for this region of terminal 9p may lie in regulator of the actin cytoskeleton, KANK1 has been shown to bind consideration of the specifically impacted KANK1 transcript, which insulin receptor substrate p53 to suppress IRSp53-induced neurite remains to be systematically assessed in the literature. Indeed, despite outgrowth (Roy et al., 2009). Other ‘ankyrin repeat domain’ genes have the high number of individuals with KANK1 CNVs in the Gold Standard also been implicated as disease causing (for example, ANKRD11 in KBG DGV track, relatively few interrupt transcript 4 specifically. syndrome (Sirmaci et al., 2011)). One commonly cited publication on KANK1 from 2005 describes a four-generation family in which nine individuals displayed congenital 2. Patient data/material CP, spastic quadriplegia, and intellectual disability (Lerer et al., 2005). Each of these affected individuals inherited a 225 kb deletion of KANK1 Clinical information on all individuals in our cohort is summarized from unaffected fathers. KANK1 was demonstrated to be expressed from in Table 1 (relatively localized KANK1 deletions) and Table 2 (rela- a single allele in lymphoblastoid cell lines and other tissues in those tively localized KANK1 duplications). All CNVs were detected by clin- unaffected fathers as well as in tissues from normal unrelated in- ical CMA including both copy number responsive and single nucleotide dividuals. This pedigree and expression data led to the authors' con- polymorphic (SNP) probes, passed quality control standards, and were clusion that deletion of KANK1 from the paternal allele (maternal im- reported based on CLIA protocols; because of this stringency they were printing) is pathogenic for CP, spastic quadriplegia, type 2. To our not validated by a second methodology. knowledge, only one other publication reports a KANK1 deletion in a There are sixteen individuals (12 males/4 females; Patients 1–16) in patient with CP and ID, which was inherited from a healthy father our cohort with deletions of KANK1. One of these deletions also in- (Segel et al., 2015). cludes DOCK8 (Patient 1). Age at time of testing ranged from 0.9 years However, Vanzo et al. subsequently reported a three-generation to 16.7 years. According to medical record review, all patients exhibited family with a distinct neurodevelopmental phenotype and inheritance ASD and/or DD/ID. Two patients also exhibited epilepsy (Patients 1 pattern (Vanzo et al., 2013). In this family, an individual with a pa- and 11, Table 1). ternally inherited KANK1 deletion (192 kb) manifested clinical features There are twelve individuals (9 males/3 females; Patients 17–28) in of ASD, motor delay, and intellectual disability (ID) while the proband's our cohort with duplications that contain part of KANK1, including one two brothers (who also harbored the paternally-inherited deletion) pair of siblings (Patients 25 and 26) within KANK1. Eight of these also were entirely unaffected. Other unrelated reports question the presence partially include DOCK8 (Patients 8–24). Age at time of testing ranged of imprinted genes on chromosome 9, including individuals with ma- from 1.3 to 12.4 years. All patients exhibited ASD and/or DD/ID. Two ternal uniparental disomy of chromosome 9 who are either unaffected patients also exhibited seizures (Patient 17 and 27, Table 2). or have clinical features attributed solely to an “unmasked” autosomal Only two of these patients' medical records indicated or suggested a recessive condition (Castanet et al., 2010; Sulisalo et al., 1997; Tiranti diagnosis of CP. Patient 11 had a reportedly normal pregnancy and a et al., 1999; Bjorck et al., 1999). salmonella infection resulting in sepsis and hospitalization at 2 weeks of Another body of data suggests that KANK1 deletions are not clini- age; he had further medical complication during infancy. Patient 17 cally relevant for CP, ASD, or any other phenotype. The Database of was born at 27-weeks' gestation and remained in the NICU for three Genomic Variants (DGV; dgv.tcag.ca) (MacDonald et al., 2013), which months due to multiple medical complications. includes a recently published case-control study by Coe et al. (2014), lists numerous individuals with CNVs including KANK1 (these can be viewed in UCSC Genome Browser by applying the “Developmental 3. Methods Delay” track housed under “Phenotype and Literature”; genome.ucsc. edu). According to the Coe et al. publication, deletions of KANK1 were We searched our internal laboratory database of genomic findings not significantly enriched in a primarily pediatric cohort with mixed from 22,054 patients who underwent clinical CMA testing from neurodevelopmental features including ASD, developmental delay September 2010 through December 2017 for CNVs overlapping (DD), and/or intellectual disability (ID) versus those without similar genomic region chr9:470,294-746,106 (hg19) that encompass KANK1. features (p = 0.911). Of note, the DGV contains a “Gold Standard There were 67 total cases, but we refined this cohort to focus our as- Variants track” (hg19_dgvgold), which is a curated, dynamic set of sessment on KANK1 clinical relevance and to limit confounding vari- variants from select studies within DGV that meet stringent criteria. ables (Tables 1 and 2). Patients were eliminated if they harbored one or This cohort illustrates that KANK1 CNVs affect different parts of this more additional clinically reported CNV (n = 10) (Table 3), if their gene (see Fig. 1a and b below). Upon deeper assessment of more KANK1 CNV contained genes in addition to KANK1 and neighboring stringent criteria, it is clear that KANK1 CNVs impacting the coding DOCK8 (n = 16) (Table 4), or patients with both (n = 13) (Table 4). region of transcript 4 are far less common in controls than currently This resulted in 28 patients with a single reported CNV relatively lo- held notions suggest. calized to KANK1: 16 with deletions and 12 with duplications. Patient 5 Without a deeper assessment of more refined control datasets such (deletion) was previously reported by Vanzo et al. (2013) and the re- as Gold Standard DGV, a laboratorian or clinician may reasonably maining 27 are novel to published literature. conclude that KANK1 CNVs are not sufficient to cause either CP or ASD. The platform primarily utilized in this dataset is an enhanced whole However, this information is neither readily searchable or available in genome chromosome microarray (CMA) known as FirstStepDx PLUS many clinical interpretation contexts, nor does it allow for a more de- built upon the ultra-high resolution Affymetrix platform with additional tailed assessment of involved portions or distinct transcripts of genes. probe content designed to optimize testing for individuals with DD/ID Therefore, in clinical practice, the sheer amount of “control” in- and/or ASD (Karen et al., 2016; Hensel et al., 2017). Three cases in our dividuals with CNVs impacting a gene coupled with case reports be- final cohort of 28 were analyzed with unenhanced Affymetrix arrays. come the default lines of evidence for KANK1 variant interpretation. Most tests were performed due to the presence of neurodevelopmental While part of the complexity in interpreting this region may be due disability (including ASD) and/or congenital anomalies. to genetic modifiers, such as point mutations or epigenetic alterations

16 R.J. Vanzo et al. European Journal of Medical Genetics 62 (2019) 15–20

Fig. 1. (a) (above). DGV showing KANK1 transcripts 1 and 4 at top (transcript 4 shown in yellow highlight), Gold Standard DGV track at middle, and Standard DGV track at bottom (losses in red, gains in blue). The majority of CNVs impacting KANK1 in this control database impact transcript 1 as opposed to transcript 4. Note that exon 1 from KANK1 transcript 4 is non-coding. Database accessed in January 2018 to create image. (b) (above). Zoomed in image of Fig. 1(a) which illustrates that far fewer control individuals have CNVs that impact KANK1 transcript 4 when looking solely at DGV Gold Standard track. This includes four entries for a total of 22 individuals (gssvL128556 – 4 deletions; gssvL129171 – 4 deletions; gssvG38188 – 12 duplications; gssvG38187 – 2 duplications). Entry gsvL128612 along with all smaller deletions at the bottom of the ﬁgure do not overlap the coding region of transcript 4. Database accessed in January 2018 to create image. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the Web version of this article.)

Fig. 2. (above). Comparison of our current deletion cohort as well as the deletion reported by Lerer et al. (n = 17; shown in red) & our duplication cohort (n = 12; shown in blue) as they impact the longer KANK1 transcript (“transcript 1”) and the KANK1 alternate/shorter transcript (“transcript 4”); shown with light blue highlight/background. Also note the number of regulatory elements congregating around the beginning of transcript 4 (Layered H3K27Ac track in UCSC genome browser, at bottom). (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the Web version of this article.)

17 R.J. Vanzo et al. European Journal of Medical Genetics 62 (2019) 15–20

4. Results 2008). It is therefore unclear what role DOCK8 plays in neurodevelopment. Nevertheless, since CNVs impacting both KANK1 and DOCK8 Fig. 2 illustrates the size and position of all CNVs as viewed in UCSC are commonly observed in public databases as well as our cohort, we Genome Browser. included them in this series. The sixteen deletions range in size from 24 kb–225 kb. In all cases, There are two limitations of the Coe et al. study as it relates to our KANK1 appears to be partially included in the deletion. In one case, data. Firstly, in the paper and supplemental materials, it is difficult if DOCK8 is also partially included. Based on CMA data, fourteen dele- not impossible to readily assess CNVs and corresponding p-values with tions impact only transcript 1, while two patients (Patients 5 and 16) respect to the specific genes impacted or contained within the CNV. For have deletions that encompass both main KANK1 transcripts (tran- example, it is not discernible how many of the “Signature” deletions scripts 1 and 4). and “Control” deletions under 10 Mb impacted KANK1 only or KANK1 The twelve duplications range in size from 59 kb–363 kb. Patient 17 and neighboring DOCK8. This type of dissection would allow a more has two non-overlapping duplications isolated to KANK1 and DOCK8 direct comparison of our data to Coe's case-control data for determi- which are 111 kb and 321 kb in size (in Fig. 2, these duplications are nation of potential clinical significance with respect to localized CNVs. shown on one line). Based on CMA data, in all cases, KANK1 appears to Secondly, the “control” cohort (including the Wellcome Trust Case be partially included in the duplication. In eight cases, DOCK8 is also Control Consortium) contains newborns who, based on the age-de- partially included. Five patients (Patients 25 and 26, who are siblings, pendent phenotype, would not likely not have been adequately assessed and Patients 17, 21, and 28) have a duplication that encompasses both relative to an ASD phenotype. This would skew the statistical analysis. KANK1 transcripts, while the other duplications involve only transcript Although it is clear that KANK1 is deleted or duplicated in both 1. cases and controls, there is an overrepresentation of CNVs impacting transcript 4 in the Coe et al. case cohort as compared to the control 5. Discussion cohort. When the corresponding phenotype track is visible in the UCSC genome browser, it appears that there are 39 deletions in cases versus 2 − Laboratorians and clinicians experience difficulty when determining in controls that impact transcript 4 (p = 1.92 × 10 6); similarly, there the clinical impact of CNVs if published case studies appear to disagree are 48 duplications in cases versus 12 in controls that impact transcript with each other, with robust case-control data, or with publicly avail- 4 (p = 0.0007) (p values were calculated via inference for two pro- able databases. CNVs involving KANK1 clearly illustrate this challenge. portions method). This significant difference could reflect the fact that We identified 28 out of 22,054 patients in our database who harbor a CNVs impacting transcript 4 are rarer than those impacting transcript 1, CNV relatively localized to KANK1. This represents an incidence of 1-in- possibly due to its smaller size. Alternatively, it may suggest that de- 788 in a cohort of patients with some form of neurodevelopmental letion/disruption of transcript 4 is in fact clinically relevant (or more disability (this number rises to 1-in-580 if using patients from Tables severe phenotypically) but this effect is masked in large case-control 1–3 in our dataset). Given their relative frequency, the disagreement studies that do not discern between included transcripts. This point is between clinical case reports and case-control data regarding clinical made more interesting by the fact that standard DGV contains over 100 significance, we felt that further review was warranted. Specifically, we individuals with CNVs impacting any part of KANK1, while Gold evaluated CMA data with specific focus on whether CNVs impacting the Standard DGV contains only 22 individuals with CNVs impacting longer KANK1 transcript (transcript 1), the alternate KANK1 transcript KANK1 transcript 4. (transcript 4), or both could theoretically be correlated with clinical Of the seven patients in our series whose CNV impacts transcript 4, consequences that result from KANK1 aberration. While deletions and two have deletions (Patients 5 and 16) and four have duplications duplications may lead to divergent phenotypes (for example, a loss of [Patients 25 and 26 (siblings, inheritance unknown), 17 (maternally function vs. a gain of function of a certain gene or genes), this issue is inherited), 21, and 28]. Patient 16 could possibly have one of the more complex. If a gene is only partially included in a duplication (as is the severe ASD-related phenotypes within our deletion cohort based on case for our entire duplication cohort presented here), the mechanism regression (see Table 1 for clinical features/testing indications); this of pathogenicity may actually be loss of function due to gene disruption could be random and based on the phenotypic spectrum/variable ex- (Newman et al., 2015). pressivity of features, could be ascertainment due to detail provided in Interestingly, a recent case-control study found that duplications submitted medical records, or could support the hypothesis that dele- involving DOCK8 are significantly more common in individuals with tion/disruption of transcript 4 is more likely clinically relevant and/or autism or various psychiatric disorders when compared to controls: more severe than if transcript 1 alone is impacted. ASD (p = 0.00384), schizophrenia (p = 0.008), and depression Interestingly, there appears to be an alternate start site beyond the (p = 0.00731) (Glessneret al, 2017). These were not significantly en- typical start site for transcript 4. This results in smaller transcripts that riched in individuals with ADHD (p = 0.0899). Across all cohorts, may warrant additional considerations for interpretation. According to DOCK8 exonic duplications did reach genome-wide significance the BrainSpan database (www.brainspan.org), the last three exons of − (p = 7.5 × 10 6). Exonic duplications in KANK1 reached suggestive KANK1 show increased levels of expression in the postnatal brain. − significance (p = 3.45 × 10 5) in the meta-analysis across all cohorts. Independent mutation analysis suggests that these exons also have a However, the authors did not discuss a mechanism as to how these low mutation burden in a normal population, indicating selective duplications may cause these conditions and the same association was pressure to maintain their sequence integrity for normal neurodeve- not seen with deletions of DOCK8. lopmental function (Uddin et al, 2014, 2016). Furthermore, examina- While we chose to limit the number of additional genes impacted by tion of the KANK1 genomic structure across species suggests that the CNVs, we did include cases of KANK1 aberration that also impact transcript 4 and the additional, shorter transcripts (not specified in DOCK8, which occurred in a significant portion of our cohort (nine of RefSeq) are closer in evolutionary origin than transcript 1. Taken to- 28; with one deletion and eight duplications). Homozygous loss-of- gether, these data suggest an important neurodevelopmental role for function mutations in DOCK8 lead to autosomal recessive hyper-IgE transcript 4 and the other shorter KANK1 transcripts. If this is the case it recurrent infection syndrome that is not characterized by neurodeve- could potentially explain the more severe ASD-related phenotype in lopmental disabilities (DD/ID, ASD, or CP) in either affected patients or Patients 5 and 16. The fact that Patient 5, who also has intellectual their carrier parents (Zhang et al., 2009; Renner et al., 2004). However, disability, has two unaffected siblings with the same deletion raises one publication describes two unrelated individuals with ID who have questions as to its actual causative role (in preparing this review, we rearrangements disrupting DOCK8 (one deletion including DOCK8 and reached out to this family and found that they remain healthy and one with a translocation breakpoint within DOCK8)(Griggs et al., developmentally normal at 16 years and 4 years old); however, this is

18 R.J. Vanzo et al. European Journal of Medical Genetics 62 (2019) 15–20 consistent with widely published observations that the manifestation of information/standards in the face of new data. autism is subject to reduced penetrance/variable expression among family members carrying identical CNVs (Heil and Schaaf, 2013). Contributors There have been several publications supporting the presence of an ASD susceptibility locus on distal 9p24 (Vinci et al., 2007; Abu-Amero RV, AP, MM, and HT are employees of and stock option holders in et al, 2010). Eighteen out of our 28 (64%) patients had a diagnosis of Lineagen. SS and ERW are consultants for Lineagen. ASD documented in their medical records. Of the ten that did not, one patient was less than a year old (making ASD a potential diagnosis for Web resources the future), nine patients had some other sort of developmental delay, and six had speech deficits in addition to their developmental delay. Database of Genomic Variants – http://dgv.tcag.ca/dgv/app/home. Thus, it is clear that 27 out of 28 individuals have neurodevelopmental University of California Santa Cruz Genome Browser – https:// phenotypes. genome.ucsc.edu. A widely cited publication on KANK1 abnormalities in a single fa- Lineagen, Inc. and FirstStepDx PLUS – https://lineagen.com/home/. mily suggests a role in spastic quadriplegic cerebral palsy (CP) (Lerer Angel's Hands Foundation - https://angelshands.org/. et al., 2005). Similarly, a second publication identified a paternally- inherited KANK1 deletion in a patient with spastic quadriplegia and ID Funding (Segel et al., 2015). Only two of the patients in this series has any in- dication of CP, (Patient 11 - spastic diplegia, Patient 15 - unspecified), This work was supported by Lineagen, Inc. and both patient had significant perinatal complications which were likely contributory factors (Drougia et al., 2007). Thus, our present Acknowledgements series does not support a role for KANK1 in the pathogenesis of CP. It is important to consider, however, the vast phenotypic and genotypic The authors wish to thank the entire dedicated and talented team at variability observed among individuals with CP; further molecular Lineagen, as well as laboratory staff at Fullerton Genetics Laboratory, analysis focused on the various CP subtypes may reveal patterns not Columbia University Medical Center, Taueret Laboratories, LLC, and currently evident in un-delineated cohorts. This type of evaluation is ARUP Laboratories. We would also like to thank the Angel's Hands under exploration (Zarrei et al., 2017) and, despite this current body of Foundation (South Jordan, UT), as funding by this organization made data, further supports the performance of CMA in individuals with CP. testing for one of these patients possible. We wish to commend the In summary, the increased use of high resolution CMA technology ordering providers for advocating and caring for their patients and, shows that the region around KANK1 is impacted more frequently in finally, to recognize the families represented herein for their extra- neurodevelopmental disabilities than was once believed (around 1-in- ordinary love and resilience. 788 of those with neurodevelopmental disabilities and 1-in-580 if considering those with KANK1 aberrations along with a second CNV). Appendix A. Supplementary data Our data could indicate that KANK1 is a dosage-sensitive locus for developmental delay with or without ASD especially when the shorter Supplementary data related to this article can be found at http://dx. KANK1 transcript (termed “transcript 4” or “variant 4”) is involved. The doi.org/10.1016/j.ejmg.2018.04.012. suggestion in the early literature that abnormalities of KANK1 is causative for CP (based on one family) is not supported by subsequent References literature or our experience. The potential contribution of KANK1 CNVs, particularly those impacting transcript 4, to ASD pathogenicity Abu-Amero, et al., 2010. A De novo marker chromosome derived from 9p in a patient should be seriously considered when found in the diagnostic work-up of with 9p partial duplication syndrome and autism features: genotype-phenotype fi correlation. BMC Med. Genet. 11, 135. such individuals despite the lack of de nitive association. Importantly, Bjorck, E.J., Anderlid, B.M., Blennow, E., 1999. Maternal isodisomy of chromosome 9 because this includes retrospective assessment of CMA data and the with no impact on the phenotype in a woman with two isochromosomes: i(9p) and i extent of KANK1 transcript inclusion in the CNVs, we wish to clarify (9q). Am. J. Med. Genet. 87 (1), 49–52. Castanet, M., Mallya, U., Agostini, M., Schoenmakers, E., Mitchell, C., Demuth, S., that experimental data were not generated to directly access neural Raymond, F.L., Schwabe, J., Gurnell, M., Chatterjee, V.K., 2010. Maternal isodisomy RNA or protein expression. Furthermore, while the CMA platforms for chromosome 9 causing homozygosity for a novel FOXE1 mutation in syndromic clinically utilized in this cohort provide good coverage of the KANK1 congenital hypothyroidism. J. Clin. Endocrinol. Metab. 95 (8), 4031–4036. fi region, limitations to CMA technology prevent certainty in determining Coe, B.P., Witherspoon, K., Eichler, E.E., et al., 2014. Re ning analysis of copy number variation identifies specific genes associated with developmental delay. Nat. Genet. specific transcript inclusion in CNVs. 46 (10), 1063–1071. Further studies to assess potential modifiers of the impact of Drougia, A., et al., 2007 Aug. Incidence and risk factors for cerebral palsy in infants with – KANK1-related CNVs is warranted and may help explain the apparent perinatal problems: a 15-year review. Early Hum. Dev. 83 (8), 541 547. fi Glessner, et al., 2017 Nov 30. Copy number variation meta-analysis reveals a novel du- confusion as to the signi cance of this gene/region. Sequence analysis plication at 9p24 associated with multiple neurodevelopmental disorders. Genome of the alternate KANK1 allele, identification of other genomic regions Med. 9 (1), 106. http://dx.doi.org/10.1186/s13073-017-0494-1. that could represent a two-hit model, or functional analysis will all Griggs, B.L., et al., 2008 Feb. Dedicator of cytokinesis 8 is disrupted in two patients with mental retardation and developmental disabilities. Genomics 91 (2), 195–202. contribute to our understanding. Additionally, it may be worthwhile to Heil, K.M., Schaaf, C.P., 2013. The genetics of autism spectrum disorders – a guide for assess epigenetic modifications and other regulatory elements as they clinicians. Curr. Psychiatr. Rep. 15 (1), 334. relate to this region based on the prior suggestion of random mono- Hensel, C., et al., 2017 Feb 27. Analytical and clinical validity study of FirstStepDx PLUS: a chromosomal microarray optimized for patients with neurodevelopmental condi- alleleic expression (Segel et al., 2015) and the fact that the UCSC tions. PLoS Curr. 9. genome browser (“Layered H3K27Ac track”) shows a concentrated Kakinuma, N., Zhu, Y., Wang, Y., Roy, B.C., Kiyama, R., 2009. Kank proteins: structure, presence of active enhancers upstream of the start site for transcript 4 functions and diseases. Cell. Mol. Life Sci. 66, 2651–2659. fi Karen, S. Ho, Hope, Twede, Rena, Vanzo, et al., 2016. Clinical performance of an ultra- (Fig. 2). It is important in evaluating clinical signi cance of a CNV high resolution chromosomal microarray optimized for neurodevelopmental dis- based on large case-control studies to have adequate phenotypes and orders. BioMed Res. Int. 2016, 3284534. http://dx.doi.org/10.1155/2016/3284534. condition-relevant age distributions in both groups, as well as more 7 pages. detailed evaluation of the relevant genomic architecture and transcript Kent, W.J., Sugnet, C.W., Furey, T.S., Roskin, K.M., Pringle, T.H., Zahler, A.M., Haussler, D., 2002 Jun. The human genome browser at UCSC. Genome Res. 12 (6), 996–1006. utilization in relevant tissues. Our analysis of this complex locus Lerer, I., Sagi, M., Meiner, V., Cohen, T., Zlotogora, J., Abeliovich, D., 2005. Deletion of highlights both the limits and dynamic nature of genomic under- the ANKRD15 gene at 9p24.3 causes parent-of-origin-dependent inheritance of fa- – standing; this reinforces the importance of challenging “accepted” milial cerebral palsy. Hum. Mol. Genet. 14, 3911 3920.

19 R.J. Vanzo et al. European Journal of Medical Genetics 62 (2019) 15–20

MacDonald, J.R., Ziman, R., Yuen, R.K., Feuk, L., Scherer, S.W., 2013 Oct 29. The da- intellectual disability, skeletal malformations, and macrodontia. Am. J. Hum. Genet. tabase of genomic variants: a curated collection of structural variation in the human 89 (2), 289–294. genome. Nucleic Acids Res. 42, D986–D992 PubMed PMID: 24174537. Sulisalo, T., Makitie, O., Sistonen, P., Ridanpaa, M., el-Rifai, W., Ruuskanen, O., de la Manning, M., Hudgins, L., Professional Practice and Guidelines Committee, 2010. Array- Chapelle, A., Kaitila, I., 1997. Uniparental disomy in cartilage-hair hypoplasia. Eur. J. based technology and recommendations for utilization in medical genetics practice Hum. Genet. 5 (1), 35–42. for detection of chromosomal abnormalities. Genet. Med. 12 (11), 742–745. Tiranti, V., Lamantea, E., Uziel, G., Zeviani, M., Gasparini, P., Marzella, R., Rocchi, M., Newman, S., et al., 2015 Feb 5. Next-generation sequencing of duplication CNVs reveals Fried, M., 1999. Leigh syndrome transmitted by uniparental disomy of chromosome that most are tandem and some create fusion genes at breakpoints. Am. J. Hum. 9. J. Med. Genet. 36 (12), 927–928. Genet. 96 (2), 208–220. Uddin, M., et al., 2014. Brain-expressed exons under purifying selection are enriched for Oskoui, M., et al., 2015. Clinically relevant copy number variations detected in cerebral de novo mutations in autism spectrum disorder. Nat. Genet. 46 (7), 742–747. palsy. Nat. Commun. 6, 7949. Uddin, M., et al., 2016. Indexing eﬀects of copy number variation on genes involved in Renner, E.D., et al., 2004. Autosomal recessive hyperimmunoglobulin E syndrome: a developmental delay. Sci. Rep. 6, 28663. distinct disease entity. J. Pediatr. 144 (1), 93–99. Vanzo, R., Martin, M., Sdano, M., South, S., 2013 May. Familial KANK1 deletion does not Roy, B.C., Kakinuma, N., Kiyama, R., 2009. Kank attenuates actin remodeling by pre- follow expected imprinting pattern. Eur. J. Med. Genet. 56 (5), 256–259. venting interaction between IRSp53 and Rac1. J. Cell Biol. 184 (2), 253–267. Vinci, G., et al., 2007. Association of deletion 9p, 46,XY gonadal dysgenesis and autistic Sarkar, S., Roy, B.C., Hatano, N., Aoyagi, T., Gohji, K., Kiyama, R., 2002. A novel ankyrin spectrum disorder. Mol. Hum. Reprod. 13 (9), 685–689. repeat-containing gene (Kank) located at 9p24 is a growth suppressor of renal cell Zarrei, M., et al., 2017 Aug 3. De novo and rare inherited copy-number variations in the carcinoma. J. Biol. Chem. 277, 36585–36591. hemiplegic form of cerebral palsy. Genet. Med. http://dx.doi.org/10.1038/gim. Segel, R., et al., 2015 Apr 21. Copy number variation in cryptogenic cerebral palsy. 2017.83. Neurology 84 (16), 1660–1668. Zhang, Q., et al., 2009. Combined immunodeﬁciency associated with DOCK8 mutations. Sirmaci, A., et al., 2011. Mutations in ANKRD11 cause KBG syndrome, characterized by N. Engl. J. Med. 361 (21), 2046–2055.