Pathomechanisms and Signatures in the Longitudinal Course of Psychosis

13.01.2015

2025-05-20

099_ Shared genetics and symptoms between schizophrenia (SCZ) and Autism Spectrum (Amendment to 090)

Research Question and Aims

Autism spectrum disorder (ASD) is a heterogeneous, heritable neurodevelopmental condition, that affects ~1% of the population across cultures, with a ~4:1 male/female ratio. ASD is characterized by impairments in reciprocal social interaction and communication, as well as restricted, repetitive and stereotyped patterns of behavior, interests and activities (1). Autistic phenotypes, however, transcend diagnostic categories. Different expressions of autistic phenotypes can also be observed in patients with psychiatric diagnoses other than ASD, being schizophrenia (SCZ) one of the psychiatric disorders manifesting ASD features (2). Particularly, the very heterogeneous diagnostic category of SCZ harbors a distinct subgroup of individuals with severe autistic features (2). Moreover, subthreshold deficits in social communication and restricted interests which do not meet formal criteria for an ASD diagnosis can also be found in the general population (3). This evidence supports the continuum dimensional nature of autistic traits.
Both ASD and SCZ present a complex etiology in which genetic factors play a role. The most recent GWAS meta-analysis of ASD has identified 5 genome-wide significant loci for ASD (4) and the latest GWAS meta-analysis on SCZ has identified 287 hits (5). In addition, recent cross-disorder GWAS meta-analyses in psychiatry have shown significant genetic correlations for most pairs of disorders, including ADS and SCZ (rG=0.24-0.26), suggesting a complex, higher-order genetic structure underlying psychopathology, and indicating that a substantial fraction of genetic influences on psychopathology transcend clinical diagnostic boundaries (6). However, little is known about the specific genetic variants with pleiotropic effects in ASD and SCZ, and whether these genetic variants contribute to behavioral phenotypes shared between the two disorders, which would allow clinicians to better diagnose individuals and give more personalized behavioral and pharmacological treatments.
While previous research has established a genetic link between ASD and SCZ, the relationship between ASD polygenic risk scores (ASD-PRS) and autistic phenotypes in SCZ remains unexplored. Autistic traits in SCZ patients can be quantified using the PANSS autism severity score (PAUSS), a measure that captures core ASD domains such as social interaction difficulties, communication impairments, and restricted, repetitive behaviors (2). Notably, ASD-PRS has been associated with autistic phenotypes in the general population (7), but its role within SCZ cohorts has yet to be investigated. Epigenetic factors also play an important role in the complex etiology of SCZ and ASD. DNA methylation is one of the most stable and commonly studied epigenetic marks in psychiatric disorders, and it can be studied at the genome-wide level. Although some studies have tried to identify DNA methylation patterns in ASD and SCZ, results from these epigenome-wide association studies (EWAS) have been largely inconsistent, possibly due to the clinical and biological heterogeneity of these disorders. To our knowledge, no studies have investigated DNA methylation underlying autistic traits in SCZ patients, which would represent a more homogeneous group of individuals.
Taken all together, this study aims to investigate the common underlying aetiology between ASD and SCZ using different approaches. On one side, we will study the polygenic effect of ASD on SCZ diagnosis. On the other side, we will investigate whether the polygenic effect of ASD is associated with autistic phenotypes in SCZ patients. Furthermore, we will explore DNA methylation patterns associated with autistic traits in SCZ individuals through an epigenome-wide association study (EWAS).

Analytic Plan

General Hypotheses:
Our main hypothesis is that studying SCZ and ASD from a trans diagnostic perspective and integrating genomic and epigenomic approaches will improve our understanding of the genetic underpinnings of autism-related phenotypes, providing new insights into the shared biology of ASD and SCZ. We hypothesize that ASD and SCZ share genetic risk factors and that the polygenic effect of shared genomic loci will be associated with autistic phenotypes in SCZ. In addition, we hypothesize that DNA methylation patterns will be associated with autistic traits in SCZ patients.

Participants:
In this study, we will use the summary statistics from the largest GWAS meta-analyses from the Psychiatric Genomics Consortium (PGC) for ASD (38,717 cases and 232,735 controls, unpublished) and SCZ (5) as well as cross-sectional data derived from the PsyCourse dataset.

PsyCourse dataset:
This dataset includes sociodemographic, clinical and genomic data of patients with SCZ (N=647) and controls (N=466). - Phenotype data – Autistic phenotypes:
Based on specific items of the Positive and Negative Syndrome Scale (PANSS) (8), a PANSS autism severity score (PAUSS) will be generated to assess autistic phenotypes (2).

- Genotype data:
Genotyping was conducted based on The Illumina Global Screening Array. Imputation was carried out using the Haplotype Reference Consortium panel in the Michigan Imputation Server. High-quality post-imputation dosage files will be used for Polygenic Score analyses.

- Methylation data:
Methylation profiling will be assessed in 285 selected individuals from blood DNA by using the Infinium HumanMethylationEPICv2 BeadChip Kit (Illumina), which interrogates the DNA methylation profile of over 1,000,000 CpG loci covering a wide range of genomic regions, including all the known genes and additional regulatory elements.

Analytic plan:
1. Shared genetic etiology between ASD and SCZ
We will run a cross-trait PRS framework to study shared etiology between ASD and SCZ. For this, the PRS-CS tool (9) will be used to infer posterior SNP effect sizes of the largest ASD-GWAS meta-analysis of ASD (unpublished). Subsequently, ASD-PRSs will be calculated in the PsyCourse sample using Plink v1.9 (10). The associations between the ASD-PRS and SCZ status will be tested including sex, age and the first 10 ancestry principal components (PCs) as covariates.

2. Polygenic effects on autistic phenotypes in SCZ
The ASD-PRS calculated from the largest ASD-GWAS meta-analysis from the PGC (unpublished) will be tested for association with autistic phenotypes, measured by means of the PAUSS scale (2), in schizophrenia patients from the PsyCourse cohort including sex, age and the first 10 ancestry principal components (PCs) as covariates. As a second step, and with the aim to improve- the prediction of autistic phenotypes in SCZ, we will identify overlapping variants with concordant and discordant effect direction in the largest ASD-GWAS meta-analysis from the PGC (unpublished) and SCZ (5), as previously described (11). PRSs will be calculated based on the summary statistics of the largest ASD-GWAS meta-analysis separately for the concordant and discordant variants using the PRS-CS tool (9) and plink (10). The obtained PRSs will be tested for association with autistic phenotypes in SCZ patients from the PsyCourse cohort including sex, age and the first 10 ancestry principal components (PCs) as covariates. Since the prevalence and the manifestation of autistic traits differ in males and females, sex will be included as a covariate in all analyses. Sex-stratified analyses will also be performed.

3. DNA methylation patterns in autistic traits
3.1. Extreme group selection based on autistic traits:
Extreme group selection based on autistic traits: Based on the PAUSS scores, we will identify extreme groups (autistic and non-autistic SCZ individuals) from the PsyCourse cohort. Extreme groups will be primarily selected on the basis of the PAUSS severity score and will be matched by age and sex. For this, 190 individuals will be selected based on their PAUSS score (n=95 low PAUSS and n=95 high-PAUSS) at visit 1. PAUSS scores from subsequent visits will be considered in order to select individuals based on trait rather than state, ensuring that the selection reflects enduring autistic traits rather than temporary states that may fluctuate over time. This approach will allow us to select SCZ patients without manifestation of autistic symptoms and patients manifesting high severity of those. Moreover, a group of n=95 healthy controls will be selected (matched by age and sex) as controls to account for potential confounding effects of age, sex, and other non-disease related variables. The inclusion of healthy controls will also enable us to determine the specificity of methylation changes observed in SCZ individuals with varying autistic traits.

3.2. Methylation profiling:
Genome-wide DNA methylation profiling for all samples (n=95 SCZ low-PAUSS, n=95 SCZ high-PAUSS, n=95 healthy controls) will be assessed in an external platform (Life & Brain GmbH, Bonn, Germany) by using the Infinium HumanMethylationEPICv2 BeadChip Kit (Illumina), which interrogates the DNA methylation profile of over 1,000,000 CpG loci covering a wide range of genomic regions, including all the known genes and additional regulatory elements.

3.3. Differentially methylated positions (DMPs) and regions (DMRs):
The processing of raw Illumina microarray data will be done with ChAMP (12) following standard protocols. Blood-cell type proportions and smoking status for each subject will be estimated with ChAMP (12) and EpiSmokEr (13), respectively. We will perform a DNA methylome-wide association study (MWAS) to identify differentially methylated positions (DMPs) and regions (DMRs) between low-PAUSS and high-PAUSS subjects. Blood cell counts, smoking score, genetic ancestry PCs, sex, and age will be used as covariates. For both, DMPs and DMRs, Benjamini-Hochberg multiple-testing correction will be used to correct the false discovery rate (FDR). We will investigate an over representation of DMP in some biological networks using the FEM package (14). This package performs a systems-level integrative analysis of DNA methylation and seeks modules of functionally related genes.
As this part of the project will be supported by existing funding, timely approval is crucial: the allocated budget must be used before October 2025, and ideally, samples should be submitted for processing in September 2025. DNA methylation profiling will be conducted at the Life & Brain platform in Bonn.

Resources needed

Socio-Demographics:
v1_id
v1_sex
v1_age
v1_school
v1_ed_status
v1_center
v2_age
v3_age
v4_age

Ethnicity:
v1_cntr_brth
v1_cntr_brth_m
v1_cntr_brth_f
v1_cntr_brth_gmm
v1_cntr_brth_gfm
v1_cntr_brth_gmf
v1_cntr_brth_gff

Diagnosis (personal and familiar):
v1_fam_hist
v1_scid_dsm_dx_cat
v1_stat
v1_scid_dsm_dx
v1_scid_age_MDE
v1_scid_no_MDE
v1_scid_age_mania
v1_scid_no_mania
v1_scid_age_hypomania
v1_scid_ever_psyc

Psychiatric history:
v1_dur_illness

Psychiatric Treatment:
v1_Antidepressants
v1_Antipsychotics
v1_Mood_stabilizers
v1_Tranquilizers
v1_Other_psychiatric
v1_lith
v1_lith_prd
v2_Antidepressants
v2_Antipsychotics
v2_Mood_stabilizers
v2_Tranquilizers
v2_Other_psychiatric
v2_lith
v2_lith_prd
v3_Antidepressants
v3_Antipsychotics
v3_Mood_stabilizers
v3_Tranquilizers
v3_Other_psychiatric
v3_lith
v3_lith_prd
v4_Antidepressants
v4_Antipsychotics
v4_Mood_stabilizers
v4_Tranquilizers
v4_Other_psychiatric
v4_lith
v4_lith_prd

Neuropsychological Evaluation:
v1_nrpsy_mwtb

Substance abuse:
v1_ever_smkd
v1_age_smk
v1_no_cig
v1_alc_pst12_mths
v1_alc_5orm
v1_lftm_alc_dep
v1_sti_cat_evr
v1_can_cat_evr
v1_opi_cat_evr
v1_kok_cat_evr
v1_hal_cat_evr
v1_inh_cat_evr
v1_tra_cat_evr
v1_var_cat_evr

Somatic diseases:
v1_ang_pec
v1_heart_att
v1_stroke
v1_diabetes
v1_hyperthy
v1_hypothy
v1_osteopor
v1_asthma
v1_copd
v1_allerg
v1_neuroder
v1_psoriasis
v1_autoimm
v1_cancer
v1_stom_ulc
v1_kid_fail
v1_stone
v1_epilepsy
v1_migraine
v1_parkinson
v1_liv_cir_inf
v1_tbi
v1_beh
v1_eyear
v1_inf

Psychiatric symptoms:
v1_panss_g9
v1_panss_g10
v1_panss_g11
v1_panss_g12
v1_panss_g13
v1_panss_g14
v1_panss_g15
v1_panss_g16
v1_panss_sum_gen
v1_panss_sum_tot
v2_panss_p1
v2_panss_p2
v2_panss_p3
v2_panss_p4
v2_panss_p5
v2_panss_p6
v2_panss_p7
v2_panss_sum_pos
v2_panss_n1
v2_panss_n2
v2_panss_n3
v2_panss_n4
v2_panss_n5
v2_panss_n6
v2_panss_n7
v2_panss_sum_neg
v2_panss_g1
v2_panss_g2
v2_panss_g3
v2_panss_g4
v2_panss_g5
v2_panss_g6
v2_panss_g7
v2_panss_g8
v2_panss_g9
v2_panss_g10
v2_panss_g11
v2_panss_g12
v2_panss_g13
v2_panss_g14
v2_panss_g15
v2_panss_g16
v2_panss_sum_gen
v2_panss_sum_tot
v3_panss_p1
v3_panss_p2
v3_panss_p3
v3_panss_p4
v3_panss_p5
v3_panss_p6
v3_panss_p7
v3_panss_sum_pos
v3_panss_n1
v3_panss_n2
v3_panss_n3
v3_panss_n4
v3_panss_n5
v3_panss_n6
v3_panss_n7
v3_panss_sum_neg
v3_panss_g1
v3_panss_g2
v3_panss_g3
v3_panss_g4
v3_panss_g5
v3_panss_g6
v3_panss_g7
v3_panss_g8
v3_panss_g9
v3_panss_g10
v3_panss_g11
v3_panss_g12
v3_panss_g13
v3_panss_g14
v3_panss_g15
v3_panss_g16
v3_panss_sum_gen
v3_panss_sum_tot
v4_panss_p1
v4_panss_p2
v4_panss_p3
v4_panss_p4
v4_panss_p5
v4_panss_p6
v4_panss_p7
v4_panss_sum_pos
v4_panss_n1
v4_panss_n2
v4_panss_n3
v4_panss_n4
v4_panss_n5
v4_panss_n6
v4_panss_n7
v4_panss_sum_neg
v4_panss_g1
v4_panss_g2
v4_panss_g3
v4_panss_g4
v4_panss_g5
v4_panss_g6
v4_panss_g7
v4_panss_g8
v4_panss_g9
v4_panss_g10
v4_panss_g11
v4_panss_g12
v4_panss_g13
v4_panss_g14
v4_panss_g15
v4_panss_g16
v4_panss_sum_gen
v4_panss_sum_tot

Genetic data:
Raw GSA genotypes to calculate PCs.
Imputed GSA genotypes for the calculation of PRSs.

Biomaterial:
DNA from visit 1 for the selected individuals’ samples (n=95 SCZ low-PAUSS, n=95 SCZ
high-PAUSS, n=95 healthy controls) for methylation profiling.