Genetic Analysis in Small for Gestational Age Newborns

UvA-DARE (Digital Academic Repository)

Start small, think big: Growth monitoring, genetic analysis, treatment and quality of life in children with growth disorders

Stalman, S.E.

Publication date 2016 Document Version Final published version

Link to publication

Citation for published version (APA): Stalman, S. E. (2016). Start small, think big: Growth monitoring, genetic analysis, treatment and quality of life in children with growth disorders.

General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

Download date:24 Sep 2021 chapter 5

Genetic Analysis in Small for Gestational Age Newborns

Susanne E. Stalman, Nita Solanky, Miho Ishida, Cristina Alemán-Charlet, Sayeda Abu-Amero, Marielle Alders, Lucas Alvizi, William Baird, Charalambos Demetriou, Peter Henneman, Chela T James, Lia C. Knegt, Lydia J. Leon, Marcel M.A.M. Mannens, Adri Mul, Nicole A. Nibbering, Emma Peskett, Joris A.M. van der Post, Faisal I. Rezwan, Carrie Ris-Stalpers, Gerdine A. Kamp, Frans B. Plötz, Jan M. Wit, Philip Stanier, Gudrun E. Moore, Raoul C. Hennekam

[Submitted] Abstract

Small for gestational age (SGA), defined as a birth weight below the 10th centile, can be a result of fetal growth restriction, associated with perinatal morbidity and mortality. Mechanisms that control prenatal growth are poorly understood. The aim of the present study is to gain more insight into prenatal growth failure and determine whether a combination of genomic analyses forms an effective diagnostic approach in SGA newborns. Array comparative genomic hybridization, genome-wide methylation studies and exome sequencing in 21 SGA newborns with a mean birthweight below the 1st centile demonstrated three CNVs, one systematically disturbed methylation pattern and one sequence variant explaining the SGA. Additional methylation disturbances and sequence variants that potentially contributed to SGA were present 20 patients. In 19 patients, multiple abnormalities were found. Our results confirm the influence of a large number of mechanisms explaining dysregulation of fetal growth.

76 Part II Introduction

The process of human fetal growth starts towards the end of the first trimester at 13 weeks of gestation and is steered by a combination of fetal and maternal genetic factors that affect intrauterine environment to ensure effective nutrient exchange between mother and fetus via the placenta. In affluent countries, 5% to 6% of pregnancies result in small for gestational age (SGA) newborns [1-4]. SGA has been defined either as being <10th centile for weight at a given gestational age or as having a birth length or weight standard deviation score (SDS) of <-2.0 (equivalent to <2.3rd centile) [5]. SGA can be a result of fetal growth restriction (FGR), which is defined as a fetus being unable to reach its individual growth potential [6]. FGR is associated with significant perinatal morbidity and mortality, [7, 8] and babies with FGR can be predisposed to metabolic diseases later in life, including type 2 diabetes, obesity and cardiovascular disease [9]. Thirty to fifty percent of the variation in weight at birth can be explained by genetic or epigenetic causes [10, 11], which include chromosome imbalances, sequence variants and independent epigenetic disturbances. The London Dysmorphology Database con- 5 tains over 400 entities associated with prenatal growth failure [12] and Genome-Wide Association Studies have disclosed less than ten variants associated with fetal growth [13]. Numerous studies on epigenetic influences, especially DNA methylation disturbances, have also been performed [1, 3, 14-19]. Despite this research the mechanisms behind prenatal growth failure are only poorly understood, at least in part due to the heterogeneous nature of growth disturbances. Consequently, an appropriate diagnostic workup for SGA newborns is not well es- tablished, and questions remain what the extent is of the genetic contribution, which genes are important, what the optimal care pathway is for the child, and how we provide adequate counselling to parents. The aim of the present study is to gain further insight into prenatal growth failure and determine whether a combination of different genomic analyses can yield an effective diagnostic approach for SGA newborns. We used array comparative genomic hybridization (array-CGH) to detect copy number variations (CNVs), genome-wide methylation studies to uncover methylation disturbances, and “whole” exome sequencing (WES) to detect sequence variants in a cohort of SGA newborns. We hypothesized that one or more CNVs explaining the SGA may be found, disturbed methylation may be present, in a set of genes known to be aberrantly methylated in low birth weight newborns,

Genetic Analysis in SGA Newborns 77 and that sequence variants may be present in genes targeted because of their known association with SGA.

Methods

Patients We selected 21 small for gestational age (SGA) newborns and their parents from the Baby Bio Bank (BBB) and Moore Cohort (Institute of Child Health, London, UK). The BBB contains biological samples from 2,515 undisturbed and complicated pregnancies, and clinical data from babies and their parents, collected between 2000 and 2014 to be used in research on four main pregnancy complications: growth restriction, preeclampsia, recurrent miscarriages and prematurity [20]. The Moore cohort consists of 319 trio samples collected from newborns and their parents between 1991 and 1994, including a small FGR cohort to investigate genetic causes of intrauterine growth disturbances. 5 th Inclusion criteria for this study included: a weight at birth at or below the 3.4 centile, availability of parental samples and absence of major structural malformations. No pre-eclampsia/HELLP syndrome, systemic disease (e.g. diabetes mellitus), medica- tion use during pregnancy or maternal smoking was present in the mothers, except for one mother (SGA4) who was a moderate smoker during pregnancy and one other mother (SGA3) who had pre-existing essential hypertension at conception for which she received treatment. A control cohort of appropriate for gestational age (AGA) newborns (n = 24) was selected from the Preeclampsia And Non-preeclampsia Database (PANDA) Biobank (Academic Medical Center, Amsterdam, The Netherlands) based on birth weight for GA closest to the 50th centile and an equal distribution of mode of delivery and gestational age in relation to the SGA cases. The PANDA Biobank collected placental biopsies, umbilical cord blood samples and maternal blood samples of 400 women with either preeclampsia or normotensive pregnancies, between 2006 and 2010. The SDS of the weight at birth were calculated using the 1990 British growth references [21] for the British cases and the 1991 reference data by Niklasson et al. for the Dutch controls [22], both based on the LMS method [23]. Descriptive statistics for analysing demographic data was performed in IBM SPSS Statistics, version 22.

78 Part II Ethical approval for all studies was obtained (BBB Research Ethics Committee references: 09/H0405/30 and 09/h0405/30+5, Moore cohort reference: 2001/6029, PANDA Biobank AMC2005_133).

Targeted genes We performed three literature searches on respectively previously reported (1) genes known to be aberrantly methylated in low birthweight newborns (2) genes known to be involved in (regulation of ) DNA methylation and (3) genes in which sequence variants are associated with disorders with a low birthweight as part of the phenotype. Those genes will be referred to as ‘targeted genes’. A list of genes known to be aberrantly methylated in low birthweight newborns was generated as shown in Supplemental Table 1. We searched in PubMed using as search terms: “Birth Weight/genetics”[Mesh], “Infant, Small for Gestational Age”[Mesh], “Infant, Low Birth Weight”[Mesh], “Fetal Growth Retardation/genetics”[Mesh], “DNA Methylation”[Mesh], “DNA Methylation/genetics”[Mesh], “Genomic Imprinting”[Mesh]. Additional articles were selected by hand searching reference lists. 5 Genes known to be involved in (regulation of ) DNA methylation were also listed (Supplemental Table 2). We searched the UniProt resource (Universal Protein Resource, uniprot.org) and Ensembl database (ensebml.org). The search strategy in UniProt was: [go:”dna methylation” AND reviewed:yes AND organism:”Homo sapiens (Human) [9606]”] and [go:”dna demethylation” AND reviewed:yes AND organism:”Homo sapiens (Human) [9606]”]. The GO term “DNA methylation” includes the GO-terms “DNA methylation”, “regulation of DNA methylation”, “DNA methylation on cytosine”, DNA methylation on cytosine within a CG sequence” and “DNA methylation involved in embryo development”. The search term in Ensembl was “DNA methylation” and we used as filters “only Human” and “only Genes”. Several additional genes were derived from the literature. Regarding exome sequencing, we determined genes in which sequence variants are associated with disorders with a low birthweight as part of the phenotype (Supplemental Table 3). We searched in two recent reviews [24, 25] and in the London Dysmorphology Database (LDDB, Oxford University Press, Oxford, UK). Search criteria in LDDB were “birth weight <3rd centile” and/or “short stature, prenatal onset”. Syndromes of which the causative gene was unknown, with intrauterine lethality and with markedly abnormal morphology or visible malformations were excluded. In case

Genetic Analysis in SGA Newborns 79 the severity of dysmorphisms or malformations at birth remained uncertain due to its variability, a syndrome was included.

DNA isolation DNA from the cases was obtained from biopsies from placenta tissue from the fetal side of the placenta near the umbilical cord insertion. DNA from parental blood samples and the cases were extracted using DNEasy Blood and Tissue Kit (Qiagen, CA, USA). DNA from the control samples was biopsied from the maternal site of the placenta and extracted according to the Gentra protocol (Qiagen, CA, USA). To minimize the risk of maternal blood contamination, the placenta biopsies were washed in phosphate- buffered saline and stored in RNAlater. To verify whether no maternal DNA contamination occurred, clustering of male samples and female samples was investigated by principal component analysis.

Array CGH The array comparative genomic hybridization (array-CGH) analysis was performed 5 using the Agilent 180K oligo-array (Amadid 023363) (Agilent Technologies, Inc., Palo Alto, CA), with 13kb overall median probe spacing and the GRCh37/hg19 browser. Standard methods were used for the labelling and hybridization. The samples were hybridized against a pool of 40 healthy sex-matched human reference samples. Data were analysed with the Genomic Workbench 6.5 (Agilent, Santa Clara, CA and Cartagenia (BENCHlab CNV v5.0 (r6643)).

Genome-wide methylation array Bisulfite conversion of genomic DNA for the genome-wide methylation array was performed using the EZ DNA Methylation Kit (Zymo Research, CA, USA). The converted DNA samples were randomized across one batch and hybridized on the Infinium Human Methylation 450K BeadChip array (Illumina, Inc., CA, USA), carried out by a certified Illumina service provider (ServiceXS, Leiden, the Netherlands). The 450K BeadChip applies both the Infinium I and II assays for optimal coverage and exam- ines >450.000 CpG sites across the whole genome. Due to the bisulfite conversion, unmethylated cytosines are converted to uracil, whereas methylated cystosines remain unchanged. The array recognizes these chemically differentiated loci and expresses the degree of methylation in β-values. The β-values correspond to the methylation score for each analysed probe and ranges from 0 (fully unmethylated) to 1 (fully methylated).

80 Part II Quality control of the Illumina 450k assay was performed using MethylAid [26]. Raw data were provided by ServiceXS and used for further statistical analysis. A file containing the β-value methylation data including annotation was produced by GenomeStudio. The methylation data from GenomeStudio and sample phenotype data were exported to the R statistical analysis environment (R version 2.15.2) (http://www.r-project.org), where a single sample analysis [27] was performed. This method allows analysis of genome-wide methylation data in small sample sizes, where each case is individually compared to a control cohort. A sample size of ≥20 controls is recommended for this analysis. The method combines the Illumina Methylation Analyzer (IMA) package (version 3.2.1) [28] and the Crawford-Howell t-test [29]. The IMA package performs a basic quality control and pre-processes the methylation data. Any CpG sites with missing values as well as samples with at least 75% CpG sites having a p-value >0.05, CpG sites where >75% samples have detection p-value >1e-5, probes on the X and Y chromosomes and probes containing SNP(s) were removed. The β-values were converted to M-values by logit transformation [30]. Quantile normalization was used to reduce unwanted technical variation across samples. Peak correction [31] was applied 5 to correct differences between Infinium I and Infinium II type assays. As all cases and controls were hybridized on the same batch, no batch correction was required. The differences between the pre-processed M-values of all single cases and the controls were determined using the Crawford-Howell t-test. Given the large number of significantly differentially methylated probes in our patients resulting from the single sample analysis, a script in Python (version 2.7) (https://www. python.org/) was used for further filtering of this data. Probes with aβ value difference of at least 20%, adjusted p-value <0.05 and a minimum of three differentially methylated probes within 2000 base pairs, allowing for reduction of false positive findings, were selected for hypermethylated and hypomethylated probes, respectively. Probes without a gene annotation were removed from further analysis. Genes found to be hypermethylated and hypomethylated at the same time in the same patient were removed. First, the genome-wide methylation pattern in SGA newborns in the context of the previously reported literature (Supplemental Table 1), was analysed. Second, other genes that were differentially methylated in >5 patients were selected.

Exome sequencing “Whole” Exome Sequencing (WES) was performed by BGI (Hong Kong). A total of 41 samples were analysed using Agilent SureSelect Human All Exon V5 (50M) kit and high

Genetic Analysis in SGA Newborns 81 throughput sequencing technology of Complete Genomics, at a 100x coverage. The 41 samples consist of 10 trio samples from the babies with the lowest birth weights and their parents (SGA1, SGA3, SGA6, SGA15 – SGA21) and 11 singletons of the remaining newborns. For each sample submitted, BGI analysed and provided: reads, results of mappings, and basic bioinformatics analysis (including alignment and assessment, SNP and InDel calling, basic annotation and statistics, SNP validation). At our institu- tion, the data were further annotated, including pathogenicity prediction data, allowing for subsequent filtering of variants. Variants were kept for further examination if: mutation types (SO terms) with “high” and “moderate” impact (Ensembl Variation - Predicted data, ensemble.org), 1K genome minor allele frequency (MAF) <0.05, ExAC allele frequency <0.05, read depth ≥30, quality score ≥30. Variants with known non-pathogenic clinical significance and a combined SIFT and PolyPhen prediction of “tolerated” and “benign” were discarded. Subsequently, we checked the variants in targeted genes known to cause a low birth weight in the context of the previously reported literature (Supplemental Table 3), and determined the likelihood of pathogenicity. At this stage ethnicity-specific MAF were 5 obtained from the 1000 Genome, ExAC and GO-ESP databases. Second, potential de novo variants were selected and verified in IGV (Integrative Genomics Viewer, Broad Institute) in the 10 patients of whom sequencing results from both the newborn and parents were available. Lastly, homozygous and compound heterozygous mutations were analysed. All variants in genes discussed in Results and Discussion have been validated by Sanger sequencing.

Results

Patients All 21 SGA cases (SGA1 – SGA21) had a birth weight (BW) for gestational age (GA) below the 3.4th centile, 19 were below the 2.3nd centile and 14 patients below the 1st centile. Table 1 shows other demographics of the study group and the control samples. Sepa- rate clustering of male cases and control samples from female samples was confirmed, indicating that no maternal DNA contamination was measured (Supplemental Fig. 1).

82 Part II Table 1. Demographics of 21 SGA Newborns and 24 Controls Appropriate for Gestational Age BW Patient ID Gender GA (weeks) (grams) BW (centile) BW (SDS) Ethnicity Mode of delivery Cases SGA1 Female 33.00 1220 0.41 -2.64 Caucasian Caesarean section SGA2 Female 38.00 1980 0.51 -2.57 African Vaginal SGA3 Female 33.71 640 4.7E-5 -4.91 South Caesarean section American SGA4 Female 39.00 2435 3.36 -1.83 Caribbean Caesarean section SGA5 Female 34.00 1350 0.59 -2.52 Asian Caesarean section SGA6 Female 39.57 2120 0.22 -2.85 Caucasian Vaginal SGA7 Male 38.00 2080 0.69 -2.46 South Vaginal American SGA8 Male 38.00 2140 1.04 -2.31 Caucasian Vaginal SGA9 Male 34.43 1543 1.04 -2.31 Caucasian Caesarean section SGA10 Male 39.57 2320 0.62 -2.50 Caucasian Vaginal SGA11 Female 38.57 2180 1.10 -2.29 African Caesarean section SGA12 Female 39.00 2385 2.56 -1.95 African Caesarean section SGA13 Male 38.57 2280 1.36 -2.21 Asian Caesarean section 5 SGA14 Female 37.14 2017 1.83 -2.09 Caribbean Vaginal SGA15 Female 31.71 474 3.14E-4 -4.52 Caucasian Caesarean section SGA16 Male 39.00 2090 0.24 -2.82 African Caesarean section SGA17 Male 22.00 236 0.13 -3.00 Caucasian Termination of pregnancy SGA18 Male 36.00 1600 0.21 -2.86 Caucasian Caesarean section SGA19 Male 37.00 1782 0.26 -2.80 Caucasian Caesarean section SGA20 Male 40.00 1874 0.01 -3.69 Caucasian Vaginal SGA21 Male 40.00 2220 0.20 -2.88 Caucasian Vaginal Mean ± SD - 36.49 ± 4.14 1760 ± 640 0.78 ± 0.88 -2.76 ± 0.77 - - Controls Mean ± SD - 37.48 ± 4.10 2953 ± 926 53.83 ± 15.51 0.10 ± 0.42 - -

GA=gestational age; BW=birth weight; SDS=standard deviation score

Array CGH The array-CGH yielded abnormalities in three patients. Patient SGA1 showed a mosaic trisomy of chromosome 16 in 70% of the cells (arr[19] 16p13.3q24(64,381-90,163,114) x2~3). Another mosaic imbalance, mosaic Turner syndrome, was seen in patient SGA11

Genetic Analysis in SGA Newborns 83 (arr Xp22.33q28(61,091-155,009,479)x1~2). Patient SGA17 had a deletion of 11p13- p14.1 (arr[h19] 11p14.1p13(29,663,942-33,400,789)x1) causing WAGR syndrome (Wilms tumour, Aniridia, Genital anomalies and mental Retardation).

Genome-wide methylation Quality control of the Illumina 450k assay showed no failed samples for bisulphite conversion, hybridization and overall methylation threshold. Table 2 shows methylation changes in genes known to be aberrantly methylated in low birth weight newborns which we targeted first (see Methods and Supplemental Table 1). Differential methylation was seen in 12 patients of which nine had differential methylation in more than one gene. As patient SGA3 shows an extensively aberrant methylation profile (Fig. 1), the results for this patient are presented separately. Subsequently, all genes found in an untargeted study to be differentially methylated in five or more patients, were analysed (Table 3), showing 28 hypermethylated genes and 6 hypomethylated genes.

5 Table 2. Differential Methylation in Genes Known to Be Aberrantly Methylated in Low Birthweight Newborns Patient Gene Chromosome Control Case No. of Main gene function(s) and influence(s) ID (MapInfo) β-value β-value probes on fetal growtha (mean) (mean) Hypermethylation SGA11 CDKN1C 11 (2905931-2907008) 0.14 0.41 5 Imprinted gene in 11p15.5 region, highly SGA15 11 (2906667-2907073) 0.21 0.48 4 expressed in placenta. Upregulation associated with IUGR placentas, loss of function associated with Beckwith- Wiedemann syndrome, gain of function with Silver-Russel syndrome SGA4 FGF14 13 (103052362- 0.16 0.44 3 Hypomethylation associated with SGA 103052943) or FGR SGA13 GNAS; GNASAS 20 (57414162- 0.62 0.83 3 Hypomethylation of GNASAS associated 57414539) with SGA. Decreased expression of GNAS observed in IUGR placentas SGA1 FOXP1 3 (71631050-71631744) 0.09 0.45 4 Increased methylation associated with SGA7 3 (71631050-71631744) 0.09 0.46 4 FGR

SGA14 NPR3 5 (32710614-32711429) 0.24 0.49 4 Hypermethylation associated with FGR SGA20 5 (32710231-32711517) 0.12 0.39 6

SGA14 NR3C1 5 (142784522- 0.22 0.47 3 Differential methylation in this 142785258) glucocorticoid receptor in placenta correlated with birth weight

84 Part II Table 2. Differential Methylation in Genes Known to Be Aberrantly Methylated in Low Birthweight New- borns (continued) Patient Gene Chromosome Control Case No. of Main gene function(s) and influence(s) ID (MapInfo) β-value β-value probes on fetal growtha (mean) (mean) SGA11 TBX15 1 (119530600- 0.28 0.58 3 Promotor hypomethylation leads to SGA14 119530702) 0.31 0.55 3 TBX15 decrease in FGR placentas SGA15 1 (119530600- 0.28 0.64 3 SGA17 119531093) 0.35 0.58 4 SGA20 1 (119530600- 0.28 0.57 3 SGA21 119530702) 0.30 0.61 4 1 (119530048- 119530932) 1 (119530600- 119530702) 1 (119530600- 119531093) SGA7 WNT2 7 (116963193- 0.18 0.52 5 WNT2 promoter methylation in placenta SGA11 116963502) 0.19 0.51 7 is associated with low birthweight SGA13 7 (116962950- 0.17 0.48 6 SGA16 116964012) 0.17 0.48 6 7 (116962950- 116963502) 7 (116962950- 116963502) 5 SGA10 ZIC1; 3 (147125714- 0.30 0.58 5 Decreased methylation associated with SGA15 ZIC4 147127662) 0.43 0.66 6 SGA or FGR SGA21 3 (147125712- 0.21 0.57 6 147126206) 3 (147126763- 147127662) Hypomethylation SGA17 IGF2AS; INS- 11 (2162406-2162616) 0.44 0.17 5 IGF2 is imprinted and highly expressed IGF2; IGF2 in placenta, hypomethylation of H19/ IGF2 control region is associated with FGR. INS-IGF2 involved in growth and metabolism. IGF2AS is imprinted and expressed in antisense to IGF2 SGA13 KCNQ1; 11 (2721207-2721383) 0.49 0.24 4 Upregulated KCNQ1 and loss of KCNQ1OT1 KCNQ1OT1 associated with IUGR; genetic variants of KCNQ1 associated with Beckwith-Wiedemann syndrome SGA1 TBX15 1 (119526060- 0.68 0.42 4 Promotor hypomethylation leads to 119527377) TBX15 decrease in FGR placentas SGA17 WNT2 7 (116964012- 0.35 0.11 4 WNT2 promoter methylation in placenta 116964802) is associated with low birthweight a For references see Supplemental Table 1

Genetic Analysis in SGA Newborns 85 Hypermethylated probes 8000 Hypomethylated probes 7500 7000 6500 6000 5500 5000 4500 4000 3500 3000 2500

Total differentially methylated probes 2000 1500 1000 500 0

SGA1 SGA2 SGA3 SGA4 SGA6 SGA7 SGA8 SGA10 SGA11PatientSGA13 ID SGA14 SGA15 SGA16 SGA17 SGA18 SGA19 SGA20 SGA21

Figure 1. Total number of differentially methylated probes per patient out of 485,577 interrogated probes, after single case analysis and further probe fi ltering (see Methods). An extensively disturbed methylation profi le is evident in patient SGA3, and patient SGA15 has an increased number of hypermethylated probes in comparison to the other patients.

5 Table 3. Differentially Methylated Genes Present in >5 Patients Gene Chrom. N Main gene function(s) Hypermethylation B3GNT3 19 11 Involved in in L-selectin ligand biosynthesis, lymphocyte homing and lymphocyte traffi cking (ncbi.nlm.nih.gov/gene) PIK3R1 5 11 Associated with insulin resistance and SHORT syndrome [40, 41] Homozygous mutations in mice cause embryonic growth retardation [42] PURA 5 11 Implicated in control of DNA replication and transcription (ncbi.nlm.nih.gov/gene) HSPA1A 6 10 Stabilizes proteins against aggregation, mediates folding of proteins, involved in ubiquitination (ncbi.nlm.nih.gov/gene) TTC23 15 10 Tetratricopeptide repeat domain 23 (ncbi.nlm.nih.gov/gene) CMSS1 3 8 Cms1 ribosomal small subunit homolog (ncbi.nlm.nih.gov/gene) CBR1 21 8 A short-chain reductase, functions as oxidoreductases (ncbi.nlm.nih.gov/gene) HSPA1L 6 8 Stabilizes proteins against aggregation, mediates protein folding (ncbi.nlm.nih.gov/ gene) ACSS3 12 7 Activates acetate allowing use for lipid synthesis or for energy generation (uniprot.org) ANKMY1 2 7 Ankyrin repeat and MYND domain-containing protein 1 (ncbi.nlm.nih.gov/gene) DDX60L 4 7 Probable ATP-dependent RNA helicase DDX60-like (uniprot.org) DIXDC1 11 7 Regulator of Wnt signalling pathway (uniprot.org) [43] ECE1 1 7 Endothelin converting enzyme, associated with M. Hirschsprung, cardiac defects and autonomic dysfunction (ncbi.nlm.nih.gov/gene) ESRRG 1 7 Involved in intrauterine growth restriction and preeclampsia [44]

86 Part II

Page 1 Table 3. Differentially Methylated Genes Present in >5 Patients (continued) Gene Chrom. N Main gene function(s) GPR135 14 7 Probable G-protein coupled receptor 135 (uniprot.org) MECOM 3 7 Transcriptional regulator and oncoprotein involved amongst others in cell differentiation and proliferation (ncbi.nlm.nih.gov/gene) MOSC2 1 7 Component of the benzamidoxime prodrug-converting complex (uniprot.org) PRKCDBP 11 7 Possible tumour suppressor function (ncbi.nlm.nih.gov/gene) TBX15a 1 6 Methylation correlates with weight and stature of newborn[45] [45]FAM221A 7 6 Family with sequence similarity 221 member A (ncbi.nlm.nih.gov/gene) GGT1 22 6 Associated with growth retardation in mice [46] HLA-F 6 6 Belongs to HLA class I heavy chain paralogues (ncbi.nlm.nih.gov/gene) MICA 6 6 Associated with susceptibility to psoriasis 1 and psoriatic arthritis (ncbi.nlm.nih.gov/ gene) MT1E 16 6 Binds various heavy metals (uniprot.org) NAT15 16 6 Uncharacterized LOC100609520 (ncbi.nlm.nih.gov/gene) NOL3 16 6 Anti-apoptotic protein, regulates enzyme activities of caspase 2, caspase 8 and p53 (ncbi. nlm.nih.gov/gene) PAK6 15 6 Expression linked to prostate cancer (ncbi.nlm.nih.gov/gene) PPP1R2P1 6 6 Inhibitor of protein-phosphatase 1 (uniprot.org) 5 Hypomethylation NAPRT1 4 9 Catalyses conversion of nicotinic acid to NA mononucleotide (uniprot.org) GUCY1B3 4 8 Receptor for ligands such as nitric oxide, oxygen and nitrovasodilator drugs (ncbi.nlm. nih.gov/gene) SDHAP3 5 7 Succinate dehydrogenase complex fl avoprotein subunit A pseudogene 3 (ncbi.nlm.nih. gov/gene) C11orf87 11 6 Uncharacterized protein C11orf87 (uniprot.org) FGF8 10 6 Involved in regulation of embryonic development, cell proliferation, cell differentiation and cell migration (uniprot.org) GDA 9 6 May play a role in microtubule assembly (ncbi.nlm.nih.gov/gene) aTBX15 is the only targeted gene differentially methylated in >5 patients

SGA3 showed differential methylation in 26 targeted genes known to be aberrantly methylated in low birth weight newborns (Supplemental Table 4). A possible explanation for this extensively disturbed methylation pattern in SGA3 would be an alteration in a gene known to be involved in (the regulation of ) DNA-methylation (see Methods and Supplemental Table 2). Of these, four were hypermethylated and 11 hypomethylated (Supplemental Table 5). Additionally, WES data were checked for sequence variants in genes involved in regulating DNA-methylation (Supplemental Table 5), showing a

Genetic Analysis in SGA Newborns 87 heterozygous missense mutation in MPHOSPH8 (p.Asp460Tyr). The same variant, with a known minor allele frequency (MAF) of 2-3% (rs75390100), was found in two other patients (SGA2 and SGA15). Due to an administrative error, three samples (SGA5, SGA9 and SGA12) could not be included in the genome-wide methylation analysis.

Exome sequencing Exome sequencing without filtering yielded over 70,000 single nucleotide variants and ~5.000 InDel variants in the 21 patients studied. After filtering (see Methods) we first evaluated sequence variants in genes that if mutated are known to be associated with disorders in which a low birthweight is part of the phenotype (Supplemental Table 3). This targeted analysis yielded potentially pathogenic heterozygous variants in 32 genes, one homozygous variant and two compound heterozygous variants (Table 4). In this targeted gene panel, no de novo variants were identified in newborns of whom sequencing data of the parents were available. In patients in which no WES was performed in their parents, variants were sequenced by Sanger in parents and showed inheritance of 5 all variants from one or both parents. Second, de novo variants in untargeted genes were analysed in silico (see Methods). Two de novo single nucleotide variants were predicted to be potentially pathogenic (Table 4). Third, we analysed all WES data for homozygous variants in untargeted genes, and found three homozygous missense mutations of potential interest (Table 4). Lastly, we evaluated data for compound heterozygous mutations in untargeted genes and found one compound heterozygous variant (Table 4). All variants described have been validated by Sanger sequencing.

Discussion

In the present study we investigated 21 SGA newborns using a combination of array- CGH, genome wide methylation array and exome sequencing. In four patients (19%), we found a genetic abnormality that likely contributes to their low birth weight. In these and 16 other patients (95%), abnormalities were found that potentially can influence fetal growth but require further functional analyses. There were three CNVs (in 3 patients), 37 variants (in 18 patients) in targeted genes known to be associated with a low birthweight, and disturbed methylation in 30 targeted genes (in 13 patients) known

88 Part II Probably damaging Probably damaging Unknown Possibly damaging Possibly damaging Probably damaging Probably damaging Probably damaging Benign Benign* Probably damaging Probably damaging Unknown PolyPhen2 Deleterious Deleterious Deleterious Deleterious Tolerated Deleterious Tolerated Deleterious Deleterious Deleterious Deleterious Deleterious Deleterious SIFT rs141305290 rs1799802 - rs142000963 - rs35221558 rs144516367 - - rs6413484 - - rs374082762 SNP ID 0.001 0.005-0.007 3.01e-5 0.001-0.002 - none-0.001 0.001 0-0.005 0-0.009 none - - 6.0e-5-1.1e-4 MAF 5 a Associated disorder (inheritance) Opsismodysplasia (AR) Xeroderma pigmentosum/Cockayne syndrome (AR) Osteogenesis imperfecta (AD) Hutchinson-Gilford Progeria (AD) De la Chapelle syndrome (neonatal osseous dysplasia) (AR) Chromosome 12q14 microdeletion KBG syndrome (AD) Hurler syndrome (AR) Bartter syndrome (AR) GH insensitivity (Laron syndrome) (AR) Microcephaly + chorioretinopathy (AR) Seckel syndrome (AR) SED Congenita; Hypochondrogenesis; Achondrogenesis type 2 (AD) Alt. C>T C>T G>T C>T G>A G>T G>A G>T C>T G>T C>T C>T G>A Position 11:71943374 16:14028081 17:48270172 1:156108510 5:149360723 12:65564283 16:89347372 4:995634 1:16359674 5:42699970 22:50664782 21:47801666 12:48369268 Gene INPPL1 ERCC4 COL1A1 LMNA SLC26A2 LEMD3 ANKRD11 IDUA CLCNKA GHR TUBGCP6 PCNT COL2A1 Sequence Variants in Targeted and Untargeted Genes in Targeted 4. Sequence Variants Table Patient ID SGA9 SGA9 SGA9 SGA8 SGA7 SGA6/M SGA6/ M SGA5 SGA5 SGA4 SGA3/ M SGA2 Targeted genes Targeted Missense variants SGA1/ M

Genetic Analysis in SGA Newborns 89 Probably damaging Probably damaging* Probably damaging Probably damaging Probably damaging Probably damaging Possibly damaging Benign Possibly damaging Probably damaging Probably damaging Probably damaging Possibly damaging PolyPhen2 Deleterious Deleterious Tolerated Deleterious Deleterious Tolerated Deleterious - Deleterious Deleterious Deleterious Deleterious Tolerated SIFT rs145191978 rs7958709 rs62624978 - - - - rs140080433 rs114025919 rs4253227 - rs3087482 SNP ID -4 e b b none none 6.06E-05 none-1.2e-4 0.002-0.005 2.0e-4 4.6e-5 9.6e-5 1.5e-5 - 7.5e-5-2.3e-4 0.003-0.004 0.005 9.4e-5-2.2 MAF 5 a Pallister–Hall syndrome; Pfeiffer syndrome (AD) Meier-Gorlin syndrome (AR) Mucolipidosis II alpha/ beta (I-cell disease) (AR) Revesz (Coats plus) syndrome (AR) Spondyloepimetaphyseal dysplasia type aggrecan; SED type Kimberley (AR, AD) Acrocapitofemoral dysplasial (AR) Chromosome 12q14 – microdeletion Noonan-like syndrome (AD) Kabuki syndrome 1 (AD) SERKAL syndrome (AR) GH insensitivity (Laron syndrome) (AR) Cockayne syndrome type B (AR) Associated disorder (inheritance) Neurofibromatosis (-Noonan) syndrome (AD) G>A T>G T>G C>G C>T C>T G>A C>A G>A G>A G>T C>T Alt. G>A 8:38275843 1:52861899 12:102164255 17:8138569 15:89400692 2:219920120 12:65633925 11:119170384 12:49434925 1:22446860 5:42689094 10:50667229 Position 17:29546083 FGFR1 ORC1 GNPTAB CTC1 ACAN IHH LEMD3 CBL KMT2D WNT4 GHR ERCC6 Gene NF1 Sequence Variants in Targeted and Untargeted Genes (continued) in Targeted 4. Sequence Variants Table SGA11 SGA13 SGA13 SGA12 SGA12 SGA17/ M SGA11 SGA17/ P SGA11 SGA15/ M SGA11 SGA14 Patient ID SGA10

90 Part II Benign - Probably damaging Possibly damaging - - - - Unknown Benign Probably damaging PolyPhen2 Deleterious - Tolerated Tolerated - - - - Tolerated Deleterious Deleterious SIFT rs5217 - rs187864727 rs185933892 rs11064480 rs141565234 - - - rs150727525 rs201880233 SNP ID none - 0.07 0.08 0.04 8.98e-5 - - - 0.002-0.004 0.003-0.004 MAF 5 a Transient neonatal diabetes Transient mellitus type 3 (AD) Spondylometaphyseal dysplasia – type Maroteaux (AD) Bowen-Conradi syndrome (AR) Noonan syndrome (AD) Coffin-Siris syndrome (AD) Primary microcephaly 1(AR) Multiple epiphyseal dysplasia 3 (AD) Associated disorder (inheritance) Cockayne syndrome type A (AR) Chromosome 3p14.3 microdeletion Alt. C>T C>T C>A G>C C>G C>T T>C C>A G>C C>G C>T Position 11:17409055 11:17409622 12:110240859 12:110230568 12:7080187 2: 39216456 19:11097681 8:6301941 20:61448920 5:60195517 3:63264392 Gene KCNJ11 TRPV4 EMG1 SOS1 SMARCA4 MCPH1 COL9A3 ERCC8 SYNPR Sequence Variants in Targeted and Untargeted Genes (continued) in Targeted 4. Sequence Variants Table Patient ID Untargeted genes De novo missense variants SGA4 Compound heterozygous variants SGA3 Homozygous variant SGA8 Splice acceptor variant SGA3/F Splice donor variant SGA14 Stop gained variant SGA21/ M Missense variant & splice region SGA15/ M SGA21/ P SGA18

Genetic Analysis in SGA Newborns 91 Possibly damaging Probably damaging Probably damaging Probably damaging PolyPhen2 Unknown Benign - Tolerated Deleterious Deleterious Deleterious SIFT Tolerated Deleterious - rs7961645 rs12076815 - - SNP ID rs147049237 - - 0.16- 0.19 0.09- 0.12 - - 0.001- 0.002 - - MAF and may [63]

5 [64] [14] a be associated with fetal hypotrophy Plays a role in folate metabolism Fibroblast growth factor, Fibroblast growth factor, involved in embryonic development and cell growth (uniprot.org) Appears to play a role in cell recognition and regulation of cell behaviour (omim.org) Negative regulation of cell proliferation; positive regulation of cell death (uniprot.org) Cooperates with AGTR2 to Cooperates with AGTR2 inhibit ERK2 activation and cell proliferation (uniprot.org) may play an importantPHLDA1 role in the anti-apoptotic effects of insulin-like growth factor-1 Associated disorder (inheritance) Alt. T>C T>A C>T C>T G>T C>G T>G Position 14:64892539 12:4543487 1:18703328 10:102766749 8:17503628 12:76425137 12:76424368 Gene MTHFD1 FGF6 IGSF21 LZTS2 MTUS1 PHLDA1 For references of targeted genes see Supplemental 3. Table Alt.=Alteration; M=maternal; P=paternal; AD=autosomal dominant; AR=autosomal recessive; Sequence Variants in Targeted and Untargeted Genes (continued) in Targeted 4. Sequence Variants Table a MAF=minor allele frequency (1000 Genome, ExAC and GO-ESP). MAF=minor allele frequency (1000 Genome, ExAC Patient ID

SGA12 SGA11 Homozygous variants SGA2 SGA20 SGA15 Compound heterozygous variants SGA11

92 Part II to be aberrantly methylated in low birth weight newborns. In evaluating the probability of pathogenicity of present findings we reasoned that the likelihood increased if an abnormality was present in a targeted gene, so a gene of which we had determined in advance that a low birthweight is part of the phenotype. In addition, homozygous, compound heterozygous and de novo variants were detected in other, untargeted genes, that did not result from our literature search, as well as abnormally methylated genes, present in more than five patients, that were not included in the targeted analyses. Three CNVs (14%) were detected in the present cohort, a relatively higher number compared to previous reports in patients with SGA or short stature [32-35]. Mosaic trisomy 16 is known to lead to a high risk of prenatal abnormalities [36], frequently including SGA, and can thus be considered a valid explanation for SGA in patient SGA1. Patient SGA11 showed mosaicism for monosomy X. About 50% of individuals with Turner syndrome have a mosaic karyotype [37], and it typically includes a low birth weight. In patient SGA17 an 11p14.1-p13 deletion, as seen in WAGR syndrome, was identified. This syndrome features reduced intrauterine growth as a known phenotype [38, 39]. As hypothesized in advance, methylation disturbances in several genes known to be 5 aberrantly methylated in low birth weight newborns, were found. In general, more hypermethylation than hypomethylation was found in the present cohort. A methylation abnormality potentially involved in SGA was detected in 13 patients, of which five showed differential methylation in several imprinted genes from the 11p15.5 imprinted region associated with fetal growth restriction, including CDKN1C, KCNQ1, IGF2AS, INS and IGF2 [14]. Methylation disturbances in untargeted genes were considered if the disturbance was present in at least 5 patients, and these were detected in 34 genes. Seven of these genes (PIK3R1, DIXDC1, ESRRG, TBX15, GGT1 and FGF8) appeared to be of specific interest, in view of their gene functions. PIK3R1 plays a role in the metabolism of insulin and is associated with SHORT syndrome [40, 41] and embryonic growth retardation in mice [42]. DIXDC1 is involved in the Wnt signalling pathway, which plays a major role in regulating embryonic development [43]. ESRRG is known to be involved in intrauterine growth restriction and preeclampsia [44]. Differential methylation of TBX15 in placenta is associated with variation in neonatal weight and stature [45]. GGT1 is associated with growth retardation in mice [46] and FGF8 is a member of the fibroblast growth factors, which play an important role in the regulation of embryonic development, cell proliferation, cell differentiation [47] and bone formation [48, 49].

Genetic Analysis in SGA Newborns 93 Although promoter methylation is generally associated with reduced gene expression and methylation of a gene itself typically with increased gene expression [50, 51], DNA methylation differs pre- and postnatally both quantitatively and in terms of CpG versus non-CpG methylation in utero. RNA samples are not available in the present series, but RNA expression studies in similar studies should provide insight in the consequences of the hyper- and hypomethylation detected in the present candidate genes. In contrast to the generally more frequent hypermethylation profile in the present cohort, patient SGA3 showed a predominant hypomethylation pattern and an extensively disturbed methylation profile. First, an external epigenetic influence, such as tobacco smoke or infectious pathogens, could be the cause of this observation [52]. The essential hypertension of SGA3’s mother may be of importance in this respect [53]. Second, a mutation in genes regulating DNA methylation, such as the DNMTs and TETs [54- 57], could theoretically cause widespread DNA methylation disturbances. In SGA3, a heterozygous mutation in MPHOSPH8 was found, a gene important in mediating epigenetic control. However, the relatively high minor allele frequency (MAF) argues against such role. Also maternal mutations in so called ‘maternal-effect genes’ such 5 as NLRP5, NLRP7 I and KHDC3L [58] can cause multi-locus imprinting disturbances in their offspring, usually resulting in hypomethylation at multiple loci and seen primar- ily in female offspring [59]. Even though SGA3 had a predominant hypomethylation pattern and is a female, we were unable to detect such mutation. Lastly, disturbed methylation of 15 genes known to be involved in (regulation of ) DNA methylation was present in SGA3. Especially the abnormal methylation of DNMT1, DNMT3B, TET1, UHRF1 and ZFP57 may be of interest as abnormal methylation of one of these may have had a subsequent more extensive effect. The evaluation of exome sequencing, targeted for genes associated with disorders in which a low birthweight is part of the phenotype, uncovered 37 sequence variants in 35 genes. Each may be considered as being potentially pathogenic and, if mutated, these genes can cause malformation syndromes, skeletal dysplasias and endocrine disorders. Our analyses showed a splice acceptor variant in SOS1 (c.3347-1G>A), a variant that is described previously in patients with Noonan-syndrome [60], and in combination with the low reported MAF considered pathogenic. Seven other missense variants in targeted genes (ACAN, CBL, COL1A1, COL2A1, FGFR1, GRH, NF1) may well be pathogenic, given their associated disorders and inheritance patterns, low MAF and prediction programs indicating the variant as being pathogenic and additional features in several variants as described below. Our analyses uncovered a c.484G>T variant of GHR (rs6413484)

94 Part II in FGF4. The same rsID is assigned to a G>A change at this position (c.484G>A), and this specific heterozygote variant has been associated with short stature [61]. However, the fact that most patients with growth hormone deficiency and insensitivity are born with a normal weight and length casts doubt on its role in SGA. The other heterozygous variants in targeted genes are less likely to be causative due to their function, because of their association with recessive disorders, high MAF, and all prediction programs indicating the variant as likely to be benign or tolerated. We found de novo mutations in two untargeted genes. MTUS1 is a tumour suppressor gene controlling cell proliferation, however no function interfering with fetal growth is known [62] so the meaning of this variant remains uncertain. The protein encoded by LZTS2 acts as a tumour suppressor and is involved in regulating embryonic development by the Wnt signalling pathway [43]. Homozygous mutations were found in four genes. For three of these (EMG1, FGF6, IGSF21) the ethnicity-specific MAFs of the patients are between 4.5% and 17% and high numbers of homozygotes are reported in the Exome Aggregation Consortium (ExAC). Therefore, we consider these to be non-pathogenic variants. One patient was homozy- 5 gous for the p.Val316Ala variant in MTHFD1. MTHFD1 is important for folate metabolism [63] and embryonic development and a mutation in this gene has been associated with fetal hypotrophy [64]. Given the function of the gene, the type of mutation, and because the variant is not a known polymorphism, this variant may be pathogenic. Compound heterozygous variants of interest were found in two targeted genes and one untargeted gene. Two variants in TRPV4 show a high MAF therefore are considered unlikely pathogenic. The variants in KCNJ11, associated with transient neonatal diabetes and a low birth weight [65], is considered a likely pathogenic variant. PHLDA1 is suggested to be involved in the anti-apoptotic effects of insulin-like growth factor-1 and associated with embryo developmental competence [66, 67] and may contribute to the SGA phenotype. Our results do not indicate the presence of a single, unifying theme explaining the dysregulation of fetal growth, and confirms previous findings that growth in utero is influenced by a large number of genes [68, 69]. We demonstrate that there is no predominant type of genetic abnormality present in SGA newborns: copy number variations, methylation disturbances and sequence variants may all contribute in part to the phenotype. In 19 patients, combinations of a CNV, (multiple) sequence variants and (multiple) methylation disturbances are present. Table 5 shows an overview of the genetic abnormalities detected in each patient, classified into the probability of their contribution to SGA. The polygenic nature of SGA indicates interactions between the

Genetic Analysis in SGA Newborns 95 Table 5. Overview of Genetic Abnormalities Classified into the Probability of their Contribution to SGA Sample Finding(s) ID Likely contributing Probably contributing Possibly contributing SGA1 70% trisomy 16 mosaicism COL2A1 variant FOXP1 hypermethylation; PIK3R1 hypermethylation; ESRRG hypermethylation; GGT1 hypermethylation SGA2 PIK3R1 hypermethylation; DIXDC1 hypermethylation; ESRRG hypermethylation; GGT1 hypermethylation SGA3 Extensively disturbed methylation pattern; SOS1 variant SGA4 KCNJ11 variant PIK3R1 hypermethylation; FGF14 hypermethylation; GGT1 hypermethylation; GHR variant SGA5 SGA6 PIK3R1 hypermethylation; DIXDC1 hypermethylation SGA7 FOXP1 hypermethylation; WNT2 hypermethylation; PIK3R1 hypermethylation; GGT1 hypermethylation 5 SGA8 PIK3R1 hypermethylation; GGT1 hypermethylation SGA9 COL1A1 variant SGA10 NF1 variant DIXDC1 hypermethylation; FGF8 hypomethylation; ZIC1 hypermethylation SGA11 46% monosomy X FGFR1 variant; CDKN1C TBX15 hypermethylation; WNT2 mosaicism hypermethylation; PHLDA1 hypermethylation variant SGA12 ACAN variant; MTHFD1 compound heterozygote variant SGA13 KCNQ1/KCNQ10T1 GNAS/GNASAS hypermethylation; WNT2 hypomethylation hypermethylation SGA14 NPR3 hypermethylation; NR3C1 hypermethylation; ESRRG hypermethylation; TBX15 hypermethylation; GGT1 hypermethylation; FGF8 hypomethylation SGA15 CDKN1C hypermethylation PIK3R1 hypermethylation; DIXDC1 hypermethylation; ESRRG hypermethylation; TBX15 hypermethylation; FGF8 hypomethylation; MTUS1 de novo variant; ZIC1 hypermethylation SGA16 WNT2 hypermethylation; PIK3R1 hypermethylation; DIXDC1 hypermethylation; ESRRG hypermethylation SGA17 11p14.1-p13 deletion CBL variant; IGF2AS/INS-IGF2/ DIXDC1 hypermethylation; ESRRG IGF2 hypomethylation hypermethylation; TBX15 hypermethylation

96 Part II Table 5. Overview of Genetic Abnormalities Classified into the Probability of their Contribution to SGA (continued) Sample Finding(s) ID Likely contributing Probably contributing Possibly contributing SGA18 PIK3R1 hypermethylation; DIXDC1 hypermethylation SGA19 PIK3R1 hypermethylation; FGF8 hypomethylation SGA20 NPR3 hypermethylation; PIK3R1 hypermethylation; ESRRG hypermethylation; TBX15 hypermethylation; FGF8 hypomethylation; LZTS2 de novo variant SGA21 TBX15 hypermethylation; FGF8 hypomethylation; ZIC1 hypermethylation

various genetic abnormalities, and the pathogenic mechanisms explaining this are likely extremely complex. A separate study in each patient will be needed to evaluate this in detail. Our results mirror a similar study in children with postnatal growth failure [70], using a similar approach in evaluating variants in genes known to be as- 5 sociated with short stature as well as studying variants in other, untargeted genes [70]. The latter authors highlight the multitude of genetic causes for short stature and the complexity of interpretation of variants and their pathogenicity, which resembles the observations in the present study. We acknowledge limitations of the present study. The size of our cohort is small and the power to draw general conclusions is limited. The genetic heterogeneity within the present cohort also appears high. We therefore used an individual-based data analysis approach to enable suitable data analysis. Using filtering strategies based on population allele frequencies as in the present study limits the ability to determine the pathogenicity of WES variants, since such data lack individual phenotypic data. Furthermore, we had no access to clinical follow-up data of the presently studied cohort. The large number of potentially pathogenic variants inhibits investigating each individual variant extensively; ideally, each variant would require a separate, detailed study. We conclude that copy number variations, methylation disturbances and sequence variants all contribute to prenatal growth failure. Such genetic workup can be an effective diagnostic approach in SGA newborns. The results of such studies in individual patients may have important consequences for patient care and counselling of patients and their families.

Genetic Analysis in SGA Newborns 97 Acknowledgments We would like to thank Jessica Glebbeek, Patricia Schmidt, Sander de Jong, Femke van Sinderen and Jet Bliek for their help in the laboratory. The present study has been made possible by grants of the following institutions: Tergooi Grant for Support of Scientific Research, KNAW Ter Meulen Fund, Jo Kolk Study Fund, ZonMW Rare Disease Network Grant, the Baby Bio Bank, Wellbeing of Women and Biomedical Research Center Grant. All sponsors are gratefully acknowledged. Sponsors had no involvement in any stage of the study design, data collection, data analyses or interpretation of the data or decision to publish study results.

98 Part II References

1. Bourque DK, Avila L, Penaherrera M, von Dadelszen 12. Oxford Medical Databases: London Dysmorphology P, Robinson WP. Decreased placental methylation at and Dysmorphology Photo Library Version 3.0 . the H19/IGF2 imprinting control region is associated (Oxford, Oxford University Press, 2001). with normotensive intrauterine growth restriction 13. Freathy RM, Mook-Kanamori DO, Sovio U, but not preeclampsia. Placenta 2010;31:197-202. Prokopenko I, Timpson NJ, Berry DJ, et al. Variants 2. Demetriou C, Abu-Amero S, Thomas AC, Ishida M, in ADCY5 and near CCNL1 are associated with fetal Aggarwal R, Al-Olabi L, et al. Paternally expressed, growth and birth weight. Nat Genet 2010;42:430-5. imprinted insulin-like growth factor-2 in chorionic 14. Cordeiro A, Neto AP, Carvalho F, Ramalho C, Doria villi correlates significantly with birth weight. PLoS S. Relevance of genomic imprinting in intrauterine One 2014;9:e85454. human growth expression of CDKN1C, H19, IGF2, 3. Moore GE, Ishida M, Demetriou C, Al-Olabi L, Leon KCNQ1 and PHLDA2 imprinted genes. J Assist LJ, Thomas AC, et al. The role and interaction of Reprod Genet 2014;31:1361-8. imprinted genes in human fetal growth. Philos Trans 15. Hillman SL, Finer S, Smart MC, Mathews C, Lowe R Soc Lond B Biol Sci 2015;370:20140074. R, Rakyan VK, et al. Novel DNA methylation profiles 4. Lee PA, Chernausek SD, Hokken-Koelega AC, associated with key gene regulation and transcription Czernichow P. International Small for Gestational pathways in blood and placenta of growth-restricted Age Advisory Board consensus development confer- neonates. Epigenetics 2015;10:50-61. ence statement: management of short children born 16. Lambertini L, Lee TL, Chan WY, Lee MJ, Diplas A, small for gestational age, April 24-October 1, 2001. Wetmur J, et al. Differential methylation of imprinted Pediatrics 2003;111:1253-61. genes in growth-restricted placentas. Reprod Sci 5. Clayton PE, Cianfarani S, Czernichow P, Johannsson 2011;18:1111-7. 5 G, Rapaport R, Rogol A. Management of the child 17. Leeuwerke M, Eilander MS, Pruis MG, Lendvai A, born small for gestational age through to adulthood: Erwich JJ, Scherjon SA, et al. DNA Methylation a consensus statement of the International and Expression Patterns of Selected Genes in Societies of Pediatric Endocrinology and the Growth First-Trimester Placental Tissue from Pregnancies Hormone Research Society. J Clin Endorinol Metab with Small-for-Gestational-Age Infants at Birth. Biol 2007;92:804-10. Reprod 2016;94:37. 6. ACOG. Intrauterine growth restriction - Clinical 18. Oostdijk W, Grote FK, SM dMK-S, Wit JM. Diagnostic management guidelines for obstetrician–gynecolo- approach in children with short stature. Horm Res gists. Int J Gynaecol Obstet 2001;72:85-96. 2009;72:206-17. 7. Alexander GR, Kogan M, Bader D, Carlo W, Allen 19. Tobi EW, Heijmans BT, Kremer D, Putter H, M, Mor J. US birth weight/gestational age-specific Delemarre-van de Waal HA, Finken MJ, et al. DNA neonatal mortality: 1995-1997 rates for whites, methylation of IGF2, GNASAS, INSIGF and LEP and hispanics, and blacks. Pediatrics 2003;111:e61-6. being born small for gestational age. Epigenetics 8. McMinn J, Wei M, Schupf N, Cusmai J, Johnson EB, 2011;6:171-6. Smith AC, et al. Unbalanced placental expression 20. Abu-Amero S, Thomas A, White S, Rogers K, of imprinted genes in human intrauterine growth Miranda AMP, Solanky N, et al. The Baby Bio Bank restriction. Placenta 2006;27:540-9. - A Legacy for Researchers Worldwide into Common 9. Barker DJ. The developmental origins of chronic Complications of Pregnancy. J Gen Pract 2014;2:158. adult disease. Acta Paediatr Suppl 2004;93:26-33. 21. Freeman JV, Cole TJ, Chinn S, Jones PR, White EM, 10. Clausson B, Lichtenstein P, Cnattingius S. Genetic Preece MA. Cross sectional stature and weight influence on birthweight and gestational length reference curves for the UK, 1990. Arch Dis Child determined by studies in offspring of twins. BJOG 1995;73:17-24. 2000;107:375-81. 22. Niklasson A, Ericson A, Fryer JG, Karlberg J, 11. Lunde A, Melve KK, Gjessing HK, Skjaerven R, Lawrence C, Karlberg P. An update of the Swedish Irgens LM. Genetic and environmental influences on reference standards for weight, length and head birth weight, birth length, head circumference, and circumference at birth for given gestational age gestational age by use of population-based parent- (1977-1981). Acta Paediatr Scand 1991;80:756-62. offspring data. Am J Epidemiol 2007;165:734-41. 23. Cole TJ. The LMS method for constructing

Genetic Analysis in SGA Newborns 99 normalized growth standards. Eur J Clin Nutr 35. Zahnleiter D, Uebe S, Ekici AB, Hoyer J, Wiesener 1990;44:45-60. A, Wieczorek D, et al. Rare Copy Number Variants 24. Baron J, Savendahl L, De Luca F, Dauber A, Phillip M, Are a Common Cause of Short Stature. PLoS Genet Wit JM, et al. Short and tall stature: a new paradigm 2013;9:e1003365. emerges. Nat Rev Endocrinol 2015;11:735-46. 36. Chareonsirisuthigul T, Worawichawong S, Parinayok 25. Wit JM, Oostdijk W, Losekoot M, van Duyvenvoorde R, Promsonthi P, Rerkamnuaychoke B. Intrauterine HA, Ruivenkamp CA, Kant SG. Novel genetic causes growth retardation fetus with trisomy 16 mosaicism. of short stature. Eur J Endocrinol 2016;174:R145-73. Case Rep Genet 2014;2014:739513. 26. van Iterson M, Tobi EW, Slieker RC, den Hollander 37. Hennekam RCM, Krantz ID, Allanson JE. “Turner W, Luijk R, Slagboom PE, et al. MethylAid: visual syndrome” Gorlin’s syndromes of the head and neck. and interactive quality control of large Illumina (Oxford, Oxford University Press: 2010). 450k datasets. Bioinformatics (Oxford, England). 38. Tezcan B, Rich P, Bhide A. Prenatal Diagnosis 2014;30:3435-7. of WAGR Syndrome. Case Rep Obstet Gynecol 27. Rezwan FI, Docherty LE, Poole RL, Lockett GA, 2015;2015:928585. Arshad SH, Holloway JW, et al. A statistical method 39. Bremond-Gignac D, Gerard-Blanluet M, Copin H, for single sample analysis of HumanMethylation450 Bitoun P, Baumann C, Crolla JA, et al. Three patients array data: genome-wide methylation analysis of with hallucal polydactyly and WAGR syndrome, patients with imprinting disorders. Clin Epigenetics including discordant expression of Wilms tumor in 2015;7:48. MZ twins. Am J Med Genet A 2005;134:422-5. 28. Wang D, Yan L, Hu Q, Sucheston LE, Higgins MJ, 40. Schroeder C, Riess A, Bonin M, Bauer P, Riess O, 5 Ambrosone CB, et al. IMA: an R package for high- Dobler-Neumann M, et al. PIK3R1 mutations in throughput analysis of Illumina’s 450K Infinium SHORT syndrome. Clin Genet 2014;86:292-4. methylation data. Bioinformatics 2012;28:729-30. 41. Winnay JN, Solheim MH, Dirice E, Sakaguchi M, 29. Crawford JR, Howell DC. Comparing an Individual’s Noh HL, Kang HJ, et al. PI3-kinase mutation linked Test Score Against Norms Derived from Small to insulin and growth factor resistance in vivo. J Clin Samples. Clin Neuropsychol 1998;12:482-6. Invest 2016;126:1401-12. 30. Du P, Zhang X, Huang CC, Jafari N, Kibbe WA, Hou 42. Foukas LC, Claret M, Pearce W, Okkenhaug K, Meek L, et al. Comparison of Beta-value and M-value meth- S, Peskett E, et al. Critical role for the p110alpha ods for quantifying methylation levels by microarray phosphoinositide-3-OH kinase in growth and analysis. BMC Bioinformatics 2010;11:587. metabolic regulation. Nature 2006;441:366-70. 31. Dedeurwaerder S, Defrance M, Calonne E, 43. Nayeem SB, Arfuso F, Dharmarajan A, Keelan JA. Denis H, Sotiriou C, Fuks F. Evaluation of the Role of Wnt signalling in early pregnancy. Reprod Infinium Methylation 450K technology. Epigenomics Fertil Dev 2016;28:525-44. 2011;3:771-84. 44. Poidatz D, Dos Santos E, Duval F, Moindjie H, 32. Canton APM, Costa SS, Rodrigues TC, Bertola Serazin V, Vialard F, et al. Involvement of estrogen- DR, Malaquias AC, Correa FA, et al. Genome-wide related receptor-gamma and mitochondrial content screening of copy number variants in children born in intrauterine growth restriction and preeclampsia. small for gestational age reveals several candidate Fertil Steril 2015;104:483-90. genes involved in growth pathways. Eur J Endocrinol 45. Chelbi ST, Doridot L, Mondon F, Dussour C, 2014;171:253-62. Rebourcet R, Busato F, et al. Combination of 33. van Duyvenvoorde HA, Lui JC, Kant SG, Oostdijk W, promoter hypomethylation and PDX1 overexpression Gijsbers ACJ, Hoffer MJV, et al. Copy number variants leads to TBX15 decrease in vascular IUGR placentas. in patients with short stature. Eur J Hum Genet Epigenetics 2011;6:247-55. 2013;22:602-9. 46. Yamada K, Tsuji T, Kunieda T. Phenotypic 34. Wit JM, van Duyvenvoorde HA, van Klinken JB, characterization of Ggt1(dwg/dwg) mice,a mouse Caliebe J, Bosch CAJ, Lui JC, et al. Copy Number model for hereditary gamma-glutamyltransferase Variants in Short Children Born Small for Gestational deficiency. Exp Anim 2013;62:151-7. Age. Horm Res Paediatr 2014;82:310-8. 47. Powers CJ, McLeskey SW, Wellstein A. Fibroblast

100 Part II growth factors, their receptors and signaling. Endocr female-dominant phenomenon? J Hum Genet 2016; Relat Cancer 2000;7:165-97. 61:765-9. 48. Lin JM, Callon KE, Lin JS, Watson M, Empson V, 60. Justino A, Dias P, Joao Pina M, Sousa S, Cirnes L, Tong PC, et al. Actions of fibroblast growth factor-8 Berta Sousa A, et al. Comprehensive massive parallel in bone cells in vitro. Am J Physiol Endocrinol Metab DNA sequencing strategy for the genetic diagnosis of 2009;297:E142-50. the neuro-cardio-facio-cutaneous syndromes. Eur J 49. Valta MP, Hentunen T, Qu Q, Valve EM, Harjula Hum Genet 2015;23(3):347-53. A, Seppanen JA, et al. Regulation of osteoblast 61. Sanchez JE, Perera E, Baumbach L, Cleveland WW. differentiation: a novel function for fibroblast growth Growth hormone receptor mutations in children factor 8. Endocrinology 2006;147:2171-82. with idiopathic short stature. J Clin Endocrinol 50. Lim DHK, Maher ER. DNA methylation: a form Metab 1998;83:4079-83. of epigenetic control of gene expression. Obstet 62. Seibold S, Rudroff C, Weber M, Galle J, Wanner C, Gynaecol 2010;12:37-42. Marx M. Identification of a new tumor suppressor 51. Dean W, Lucifero D, Santos F. DNA methylation in gene located at chromosome 8p21.3-22. Faseb J mammalian development and disease. Birth Defects 2003;17:1180-2. Res C Embryo Today 2005;75:98-111. 63. Carroll N, Pangilinan F, Molloy AM, Troendle J, 52. Ho SM, Johnson A, Tarapore P, Janakiram V, Zhang Mills JL, Kirke PN, et al. Analysis of the MTHFD1 X, Leung YK. Environmental epigenetics and its promoter and risk of neural tube defects. Hum Genet implication on disease risk and health outcomes. Ilar 2009;125:247-56. J 2012;53:289-305. 64. Lorenc A, Seremak-Mrozikiewicz A, Barlik M, Wolski 53. Wang J, Gong L, Tan Y, Hui R, Wang Y. Hypertensive H, Drews K. The role of 401a>G polymorphism 5 epigenetics: from DNA methylation to microRNAs. J of methylenetetrahydrofolate dehydrogenase Hum Hypertens 2015;29:575-82. gene (MTHFD1) in fetal hypotrophy. Ginekol Pol 2014;85:494-9. 54. Hamidi T, Singh AK, Chen T. Genetic alterations of DNA methylation machinery in human diseases. 65. Anderson de la Llana S, Klee P, Santoni F, Stekelen- Epigenomics 2015;7:247-65. burg C, Blouin JL, Schwitzgebel VM. Gene Variants Associated with Transient Neonatal Diabetes 55. Kaneda M, Okano M, Hata K, Sado T, Tsujimoto N, Li Mellitus in the Very Low Birth Weight Infant. Horm E, et al. Essential role for de novo DNA methyltrans- Res Paediatr 2015;84:283-8. ferase Dnmt3a in paternal and maternal imprinting. Nature 2004;429:900-3. 66. Hamel M, Dufort I, Robert C, Leveille MC, Leader A, Sirard MA. Identification of follicular marker 56. Monk D. Germline-derived DNA methylation and genes as pregnancy predictors for human IVF: new early embryo epigenetic reprogramming: The evidence for the involvement of luteinization process. selected survival of imprints. Int J Biochem Cell Biol Mol Hum Reprod 2010;16:548-56. 2015;67:128-38. 67. Nagai MA. Pleckstrin homology-like domain, family 57. Strogantsev R, Krueger F, Yamazawa K, Shi H, Gould A, member 1 (PHLDA1) and cancer. Biomed Rep P, Goldman-Roberts M, et al. Allele-specific binding 2016;4:275-81. of ZFP57 in the epigenetic regulation of imprinted and non-imprinted monoallelic expression. Genome 68. Roberts DF. The genetics of human fetal growth. Biol 2015;16:112. Postgrad Med J 1978;54 Suppl 1:107-16. 58. Docherty LE, Rezwan FI, Poole RL, Turner CL, 69. Tower C, Baker P. The Genetics of fetal growth Kivuva E, Maher ER, et al. Mutations in NLRP5 are restriction: Implications for management. Rev associated with reproductive wastage and multilocus Gynaecol Perinat Prac 2006;6:99-105. imprinting disorders in humans. Nat Commun 70. Wang SR, Carmichael H, Andrew SF, Miller TC, 2015;6:8086. Moon JE, Derr MA, et al. Large-scale pooled 59. Sano S, Matsubara K, Nagasaki K, Kikuchi T, next-generation sequencing of 1077 genes to identify Nakabayashi K, Hata K, et al. Beckwith-Wiedemann genetic causes of short stature. J Clin Endo Metab syndrome and pseudohypoparathyroidism type Ib in 2013;98:E1428-37. a patient with multilocus imprinting disturbance: a

Genetic Analysis in SGA Newborns 101