Genetics and epigenetics of ADHD and comorbid conditions

Applying bioinformatics and statistical genetics methods on population data to understand the mechanisms in ADHD and cocaine-dependence

1. Genetics and epigenetics of attention-deficit/hyperactivity disorder and comorbid conditions Anu Shivalikanjli April 06, 2020 Profs. Bru Cormand and Stephen Faraone

2. HORIZON 2020 Standout 1 – Do Not Delete This Text – Used in Hyperlink This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska -Curie grant agreement No 643051. “Mastering skills in the training network for attention deficit hyperactivity and autism spectrum disorders”

3. Beneficiaries Partners This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska -Curie grant agreement No 643051.

4. WP 2 WP 3 WP 4 PREVALENCE AND DEFINITION GENETICS AND EPIGENETICS MODEL SYSTEMS PREDICTION AND TREATMENT WP 1 TRAINING WP 5 IMPACT WP 6

5. WP 2 WP 3 WP 4 PREVALENCE AND DEFINITION GENETICS AND EPIGENETICS MODEL SYSTEMS PREDICTION AND TREATMENT WP 1 TRAINING WP 5 IMPACT WP 6

6. WP 2 GENETICS AND EPIGENETICS Common and rare genetic variation in ADHD and ASD Epigenetics: methylation, miRNAs Parent-of-origin effects in epigenetics Interplay genome-microbiome

7. Completed stay in UB Setting up the data, high-computing cluster accounts and pipelines Joined the lab Bioinformatic analysis Late 2015 JULY 2017- mid 2018 2016 2018 2019 PUblications 2020 Thesis 2021 India

8. Genetics and epigenetics of attention-deficit/hyperactivity disorder and comorbid conditions Anu Shivalikanjli April 06, 2020 Profs. Bru Cormand and Stephen Faraone

9. Background

10. 10 1 Background Clinical Presentation of ADHD Three core symptoms of ADHD

11. Combined Hyperactive-impulsive Inattentive 2 Background

12. 12 years 3 Background Epidemiology 3.4% adults Persistence

13. Neurobiology of ADHD 13 Background Mid- Sagittal view 4 Executive functioning Attention Reward Outside left hemisphere view

14. Anatomical brain changes in ADHD 14 Background 5

15. Genetic models 15 Mendelian Model Complex Model Variants Disease type Genes Size of cohort Whole exome/whole genome GWAS Technology Family Populations 6 Background

16. Genetic approaches 16 Background 7

17. General heritability 17 Heritability due to GWS SNPs <5% Global heritability estimates of SNPs ~20% SNP-based heritability >70% heritability in ADHD Background 8

18. Family burden Comorbidities ADHD Core symptoms Clinical condition Clinical condition Clinical condition Comorbid disorders e.g. Intellectual disability Anxiety disorders Tic disorders ODD, CD Comorbid disorders e.g. Intellectual disability Anxiety disorders Tic disorders Bipolar disorder Depression ODD, CD Comorbid disorders e.g. Intellectual disability Anxiety disorders Tic disorders Bipolar disorder Depression Substance use disorder Personality disorder ADHD Core symptoms ADHD Core symptoms Adolescence Childhood Adulthood Background 9

19. Cocaine dependence 19 5.2% of adults have tried cocaine Heritability 20% will develop addiction Background 10

20. Cocaine dependence Background Binge/ Intoxication Preoccupation/ Anticipation Withdrawal/ Negative Affect Reward Deficit & Stress Surfeit Executive function deficits Incentive Salience Background 11

21. ADHD and Cocaine dependence No ADHD No cocaine dependence ADHD ADHD + cocaine dependence Cognitive impulsivity Response inhibition Self-medication Impulsivity Risk-taking ADHD Cocaine dependence Conduct disorder Oppositional defiant Drug/alcohol exposure in womb [Endo-] phenotypes (cognitive functioning, impulsivity, dysregulations in neurotransmission) Genetic factors Background 12

22. Implicated Genes and Biomarkers 22 Probably not useful biomarkers Useful biomarkers ADHD Addiction Background 13

23. Epigenetics 23 14 Background

24. Epigenetics Stable changes in gene expression Gene function and brain structure 15 Background

25. DNA Methylation 25 16 Background

26. MicroRNAs 17 Background

27. Objectives

28. Identification of genetic variation that influences brain methylation in ADHD Hypothesis-free + hypothesis driven Identification of common variation in microRNA genes that contribute to ADHD Hypothesis-free + hypothesis driven Genome-wide association meta-analysis of cocaine-dependence: Shared genetics with comorbid conditions Hypothesis-free 28 1 3 2 19 Objectives

29. Results

30. Identification of genetic variation that influences brain methylation in ADHD Identification of common variation in microRNA genes that contribute to ADHD Genome-wide association meta-analysis of cocaine-dependence: Shared genetics with comorbid conditions 30 1 3 2 20 Results

31.  Selection of allele-specific methylation (ASM) SNPs 31 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 21

32. 32 Association results obtained for ASM-variants in ADHD Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 3,896 tagSNPs Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 22

33. 23 Functional follow-up of associated ASM-variants/CpG SNPs as eQTLs for genes Location of CpGs in promoter regions Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 23

34. 34 Functional follow-up of associated ASM-variants and CpGs Location of CpGs in promoter regions 24 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 24

35. Functional follow-up of associated ASM-variants and CpGs Location of CpGs in promoter regions ARTN C2orf82 NEUROD6 PIDD1 RPLP2 GAL Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 25

36. 36 Functional follow-up of associated ASM-variants and CpGs SNPs as eQTLs of the same genes ARTN C2orf82 NEUROD6 PIDD1 RPLP2 GAL 36 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 26

37. ARTN C2orf82 NEUROD6 PIDD1 RPLP2 GAL 37 Functional follow-up of associated ASM-variants and CpGs SNPs as eQTLs of the same genes 37 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 27

38. cg22930187 and cg06207804 38 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD Artemin Ligand of GDNF family survival of sensory and sympathetic peripheral neurons Neuron excitability 38 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 28

39. 39 cg13047596 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD Same direction C2orf82 Proteoglycan transmembrane protein Highly expressed in brain tissues 39 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 29

40. 40 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD cg20225915 P53-Induced Death Domain Protein 1 Cell life regulator gene 40 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 30

41. 41 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD Correlation with methylation levels 60 ASM SNPs cg22930187 cg06207804 cg13047596 cg11554507 cg20225915 cg04464446 3 SNPs 45 SNPs 3 SNP 7 SNPs 2 SNPs ARTN C2orf82 NEUROD6 PIDD1 GAL RPLP2 Within promoter regions of genes eQTLs ARTN C2orf82 NEUROD6 PIDD1 GAL RPLP2 Correlation with brain volumes ↑ NAc ↑ CN ↓ NAc ↓ CN ↓T 41 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 31

42. Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD PGC ADHD GWAS meta-analysis is enriched in ASM-SNPs Better OR for SNPs with low (better) p-values. Significance Threshold N SNPs N ASM SNPs p-value OR 5.00E-08 303 6 1.70-03 4.92 5.00E-07 945 8 4.30E-02 2.08 5.00E-06 2,122 15 3.15E-02 1.74 5.00E-05 6,970 35 1.31E-01 1.23 5.00E-04 25,288 139 4.58E-04 1.35 5.00E-03 115,681 527 6.94E-03 1.12 5.00E-02 651,772 2790 5.54-03 1.05 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 32

43. 43 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD Lookup 43 Results – Chapter 1: Identification of genetic variation that influences brain methylation in ADHD 33 IF: 5.2

44. Identification of genetic variation that influences brain methylation in ADHD Identification of common variation in microRNA genes that contribute to ADHD Genome-wide association meta-analysis of cocaine-dependence: Shared genetics with comorbid conditions 44 Results 1 3 2 34

45. 1355 Intragenic Clustered Singleton miRNAs are <10Kb apart 1761 autosomal miRNA genes 406  135 879 Exonic Clustered Clustered Singleton Intronic Host gene Selection of miRNA regions 45 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 35

46. Exonic Clustered Clustered Singleton 1355 Intragenic Clustered Singleton 1761 autosomal miRNA genes 406  135 879 10kb 10kb 5kb 5kb 10kb Host gene Intronic 10kb Selection of regulatory elements for miRNA genes miRNAs are <10Kb apart Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 36

47. 47 Association results obtained for miRNA common variants in ADHD ~22K tagSNPs 19 tagSNPs Overcoming 5% FDR ~17K tagSNPs Retrieved p-values from summary statistics 12 Associated miRNAs miR-6734 miR-6735 miR-6079 miR-7-1 miR-3135a miR-3666 miR-1273h Chr 1 Chr 3 Chr 7 Chr 9 Chr 16 miR-4655 miR-6506 miR-4271 miR-5193 5’ 3’ phg 5’ 5’ 5’ 5’ 5’ 5’ 5’ 5’ phg 3’ 5’ Variant location 5’: Upstream of miRNA gene 3’ Downstream of miRNA gene phg : Promoter region of host gene 5’ miR-6872 5’ 5’ 5’ 3’ 1761 miRNAs 12 SNPs per region 47 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 37

48. 48 Follow-up of significantly associated miRNAs Brain expression 12 Associated miRNAs miR-6734 miR-6735 miR-6079 miR-7-1 miR-3135a miR-3666 miR-1273h Chr 1 Chr 3 Chr 7 Chr 9 Chr 16 miR-4655 miR-6506 miR-4271 miR-5193 miR-6872 Validated target genes Network analysis BrainSpan Atlas miRMine miRIAD Spatio -temporal miRNA profiles in developing human brain (Ziats and Rennert ) Ingenuity Pathway Analysis Ingenuity Pathway Analysis TarBase TargetScan Human miRecords 48 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 38

49. 49 Follow-up of significantly associated miRNAs Brain expression 12 Associated miRNAs miR-6734 miR-6735 miR-6079 miR-7-1 miR-3135a miR-3666 miR-1273h Chr 1 Chr 3 Chr 7 Chr 9 Chr 16 miR-4655 miR-6506 miR-4271 miR-5193 miR-6872 Validated target genes Network analysis BrainSpan Atlas miRMine miRIAD Spatio -temporal miRNA profiles in developing human brain (Ziats and Rennert ) Ingenuity Pathway Analysis Ingenuity Pathway Analysis TarBase TargetScan Human miRecords 49 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 38

50. 50 Follow-up of significantly associated miRNAs Brain expression Cerebellum Cerebellar cortex Cerebellar cortex Primary somato -sensory cortex Primary somato -sensory cortex Ventral parietal cortex 12 Associated miRNAs miR-6734 miR-6735 miR-6079 miR-7-1 miR-3135a miR-3666 miR-1273h Chr 1 Chr 3 Chr 7 Chr 9 Chr 16 miR-4655 miR-6506 miR-4271 miR-5193 miR-6872 Primary somato -sensory cortex Primary visual cortex Cerebellum Whole brain Whole brain Cerebellum Whole brain Cerebellum Whole brain Cerebellum Whole brain Cerebellum Whole brain Whole brain Data unavailable Differentially expressed between PFC and cerebellum –late childhood development 50 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 39

51. 51 Follow-up of significantly associated miRNAs Brain expression 12 Associated miRNAs miR-6734 miR-6735 miR-6079 miR-7-1 miR-3135a miR-3666 miR-1273h Chr 1 Chr 3 Chr 7 Chr 9 Chr 16 miR-4655 miR-6506 miR-4271 miR-5193 miR-6872 Validated target genes Network analysis BrainSpan Atlas miRMine miRIAD Spatio -temporal miRNA profiles in developing human brain (Ziats and Rennert ) Ingenuity Pathway Analysis Ingenuity Pathway Analysis TarBase TargetScan Human miRecords 51 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 39

52. 52 18 9 1 Follow-up of significantly associated miRNAs 12 Associated miRNAs miR-6735 miR-6079 miR-7-1 miR-3135a miR-3666 miR-1273h Chr 1 Chr 3 Chr 7 Chr 9 Chr 16 miR-4655 miR-6506 miR-4271 miR-5193 miR-6872 Validated target genes Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 40

53. 53 18 9 1 Follow-up of significantly associated miRNAs 12 Associated miRNAs miR-7-1 miR-3666 miR-1273h Chr 7 Chr 9 Chr 16 Validated target genes 41 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 41

54. 54 Functions of validated target genes Lithium-responsive BIP Brain thalamus volumes EGFR SNCA Cognitive empathy Depressive episodes in BIP EIF4E Parkinson disease MKNK1 SCZ SLC17A7 (VGLUT1) Brain-specific manner Mediates uptake of gluatamate into synaptic vesicles Cognitive decline, SCZ, MDD and bipolar disorder TAC1 Brain region volumes Intracranial volume Subcortical volumes MEOX2 Encodes neuropeptides – neurotransmitters and induces behavioural response Risk-taking Feeling nervous traits Moderate expression in caudate, NAc , and putamen miR-7-1 miR-3666 54 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 42

55. 55 Follow-up of significantly associated miRNAs Brain expression 12 Associated miRNAs miR-6734 miR-6735 miR-6079 miR-7-1 miR-3135a miR-3666 miR-1273h Chr 1 Chr 3 Chr 7 Chr 9 Chr 16 miR-4655 miR-6506 miR-4271 miR-5193 miR-6872 Validated target genes Network analysis BrainSpan Atlas miRMine miRIAD Spatio -temporal miRNA profiles in developing human brain (Ziats and Rennert ) Ingenuity Pathway Analysis Ingenuity Pathway Analysis TarBase TargetScan Human miRecords Using all miRNAs as input – focal miRNAs 43 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD

56. 56 miR-5193 IL9 miR-7-1-3p miR-3135a miR-7a-5p miR-610 miR-6734-3p miR-6735-3p miR-1273h-3p miR-4271 UTS2R GPR65 GPR174 ADGRA3 ADGRE2 RGR ADRB2 GPR26 XCR1 ADRA2A HTR1D YWHAG Adrenoreceptor Beta arrestin PTGER1 GPR78 OR10H2 GPR161 ADGRF4 CCKAR HTR4 GPR25 Network analysis ADGRG7 C3AR1 ADGRG4 44 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD

57. 57 CCKAR SCZ YWHAG Moderate language laterization Superior frontal grey matter volumes SCZ risk allele ADGRE2 MDD GPR26 Encodes a protein distantly related to Serotonin Expressed exclusively in brain tissue 45 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD

58. 58 Follow-up of significantly associated miRNAs Brain expression 12 Associated miRNAs miR-6734 miR-6735 miR-6079 miR-7-1 miR-3135a miR-3666 miR-1273h Chr 1 Chr 3 Chr 7 Chr 9 Chr 16 miR-4655 miR-6506 miR-4271 miR-5193 miR-6872 Data unavailable Validated target genes Orthologs Network analysis Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 46

59. 59 Overlap of associated miRNA regions and GWS-ADHD loci Demontis et al ., 2019 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 47

60. 60 Overlap of associated miRNA regions and GWS-ADHD loci Demontis et al ., 2019 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD 48

61. 61 Lookup 49 Results - Chapter 2: Role of common variation in micro-RNA genes in ADHD

62. Identification of genetic variation that influences brain methylation in ADHD Identification of common variation in microRNA genes that contribute to ADHD Genome-wide association meta-analysis of cocaine-dependence: Shared genetics with comorbid conditions 62 Results 1 3 2 50

63. 63 Results – Chapter 3: GWA meta-analysis of cocaine dependence Sample sets and pipeline N cases = 2,085 N controls= 4,293 EUR Sample 2 Sample 3 Sample 4 468 N controls 609 504 504 N cases 1284 410 1190 1409 Illumina HumanOmni1-Quad_v1-0_B Illumina ILMN_Human-1 Illumina HumanOmni1-Quad_v1-0_B Illumina Human660Q-Quad_v1_A QC PCA Imputation SNP-based analysis Sample 1 RICOPILI - R apid I mputation and CO mputational PI pe LI ne Gene-based analysis METAL MAGMA 9,290,362 markers GWAS 51

64. 64 Results – Chapter 3: GWA meta-analysis of cocaine dependence Results from the GWAS meta-analysis QQ plot SNP-based analysis 23 lead SNPs = 22 genomic risk loci = 112 genes  Suggestive Associations SNP to Gene 447 - brain eQTLs - 12 genes   BTN3A2 ,  HIST1H2AK ,  ZSCAN31 ,  PRSS16,   ZNF184   Chromosome 6 Schizophrenia-associated 2 lead SNPs GWAS  P  = 3.1e-06 and 3.4e-06 458 nominally associated SNPs 77 genes  52

65. 65 Results – Chapter 3: GWA meta-analysis of cocaine dependence Circo -plot of chromosome 6 genomic risk locus 0.8 0.6 0.4 r2 for the SNPs in the GRL <=0.2 Top SNPs in the risk locus Chromosome ring Chromatin interactions eQTLs Manhattan plot SNPs   P  < 0.05 Risk locus 53

66. 66 Results – Chapter 3: GWA meta-analysis of cocaine dependence Gene-based analysis SNP-based analysis 10% FDR threshold 54

67. 67 Results – Chapter 3: GWA meta-analysis of cocaine dependence Gene-based analysis Enriched in immune system and histone-related genes Five SNPs nominally associated with cocaine dependence ( P  < 1e-04) and been associated with SCZ and BIP: the risk allele is the same in all studies. rs17693963 - reported in five studies - brain eQTL for: PRSS16 ,  ZSCAN9, ZNF184  and  ZSCAN31 Top differentially expressed genes HIST1H2BD ,  HIST1H2BC ,  HIST1H2BH ,  HIST1H2BG  and  HIST1H4K - lymphoblastoid cell lines in SCZ 10% FDR threshold 55

68. * nominal result 2.17 23.63 68 Polygenic architecture of cocaine dependence Results – Chapter 3: GWA meta-analysis of cocaine dependence Partitioned heritability analysis by functional annotations on LD Score Regression (LDSC) 55

69.  Cocaine dependence and shared genetic factors with comorbid conditions (including ADHD) Results – Chapter 3: GWA meta-analysis of cocaine dependence 56

70.  Cocaine dependence and shared genetic factors with comorbid conditions Results – Chapter 3: GWA meta-analysis of cocaine dependence Error bars: 95% confidence limits Significance threshold: P < 7.1e-03 Genetic correlation based on LD Score (LDSC) regression analysis Cocaine dependence shares genetic risk factors with several comorbid traits -> Horizontal (biological) pleiotropy 57

71.  Cocaine dependence and shared genetic factors with comorbid conditions Results – Chapter 3: GWA meta-analysis of cocaine dependence Best fit results from Polygenic Risk Score (PRS) analysis for each tested phenotype Values: p-value for significance for the most predictive models Significance threshold: P < 5.7e-04 58

72. 72 Lookup Results – Chapter 3: GWA meta-analysis of cocaine dependence The largest GWAS meta-analysis on cocaine dependence in European ancestry individuals Identified susceptibility regions on chromosome 6 – HIST1H2BD IF 4.3 59

73. Conclusions

74. Common genetic risk variants for ADHD identified in a previous genome-wide association study (GWAS) that included 20,000 cases and 35,000 controls are enriched in SNPs that correlate with levels of DNA methylation. Eight Allele-Specific Methylation tagSNPs are significantly associated with ADHD and correlate with differential methylation at six CpG sites in cis in different brain areas. These six CpG sites are located at possible promoter regions of six genes expressed in brain: ARTN , C2orf82 , NEUROD6 , PIDD1 , RPLP2 and GAL . For three of these six genes, SNPs associated with ADHD and correlating with methylation levels are eQTLs in brain. Consistently, methylation and gene expression show opposite directions: ARTN and PIDD1 (reduced methylation, increased expression), C2orf82 (increased methylation, reduced expression). ADHD risk alleles are associated with increased brain expression of ARTN and PIDD1 and with decreased brain expression of C2orf82 . SNPs in C2orf82 correlate with changes in brain volumes. Our study highlights the ARTN , C2orf82 and PIDD1 genes as potential contributors to ADHD susceptibility. 74 Conclusions – Chapter 1 60

75. We have performed a case-control association study to investigate the contribution to ADHD of common genetic variation in 1,761 autosomal miRNAs using pre-existing GWAS data from 20,000 cases and 35,000 controls. We have identified 19 significant associations of SNPs with ADHD that highlight 12 miRNA genes, all located within protein-coding genes. The associated variants are located in the putative regulatory regions of the miRNA genes or in the promoter region of the host protein-coding gene. The highlighted miRNAs are expressed in different brain tissues, specifically in cerebellum Three of the highlighted miRNAs - miR-3666, miR-7-1 and miR-1273h - have validated target mRNAs. Pathway analysis of ADHD-associated miRNAs revealed two biological pathways. One of the pathways involves miRNA-mediated regulation of serotonin receptor genes and it is likely to be involved in neurological functions and diseases. 75 Conclusions – Chapter 2 60

76. We have performed the largest cocaine dependence GWAS meta-analysis in individuals of European ancestry, including 2,100 cases and 4,300 controls. Although the SNP-based analysis revealed no genome-wide significant associations with cocaine dependence, probably due to limited sample size, the gene-based analysis identified the HIST1H2BD gene, previously associated with schizophrenia. The estimated SNP-based heritability of cocaine dependence is approximately 30%. A significant genetic correlation has been observed between cocaine dependence and ADHD, schizophrenia, major depressive disorder and risk-taking behaviour, suggesting a shared genetic basis across pathologies and traits. Polygenic risk score (PRS) analysis shows that all the comorbid features analysed (ADHD, schizophrenia, major depressive disorder, aggressiveness, antisocial personality or risk-taking behaviour) can predict cocaine dependence 76 Conclusions – Chapter 3 60

77. Genetics and epigenetics of attention-deficit/hyperactivity disorder and comorbid conditions Anu Shivalikanjli April 06, 2020 Profs. Bru Cormand and Stephen Faraone