Skip to main content
Download PDF

Multi-ancestry phenome-wide association of complement component 4 variation with psychiatric and brain phenotypes in youth

Genome Biology volume 24, Article number: 42 (2023) Cite this article

Abstract

Background

Increased expression of the complement component 4A (C4A) gene is associated with a greater lifetime risk of schizophrenia. In the brain, C4A is involved in synaptic pruning; yet, it remains unclear the extent to which upregulation of C4A alters brain development or is associated with the risk for psychotic symptoms in childhood. Here, we perform a multi-ancestry phenome-wide association study in 7789 children aged 9–12 years to examine the relationship between genetically regulated expression (GREx) of C4A, childhood brain structure, cognition, and psychiatric symptoms.

Results

While C4A GREx is not related to childhood psychotic experiences, cognition, or global measures of brain structure, it is associated with a localized reduction in regional surface area (SA) of the entorhinal cortex. Furthermore, we show that reduced entorhinal cortex SA at 9–10 years predicts a greater number and severity of psychosis-like events at 1-year and 2-year follow-up time points. We also demonstrate that the effects of C4A on the entorhinal cortex are independent of genome-wide polygenic risk for schizophrenia.

Conclusions

Our results suggest neurodevelopmental effects of C4A on childhood medial temporal lobe structure, which may serve as a biomarker for schizophrenia risk prior to symptom onset.

Background

Schizophrenia is a highly heritable neurodevelopmental disorder characterized by the presence of hallucinations, delusions, disorganized behavior, and an array of negative symptoms (e.g., flattened affect, reduced motivation) and cognitive deficits. While clinical diagnosis typically occurs in late adolescence or early adulthood, frank psychosis is often preceded by a prodromal period of weeks to years during which gradual changes in cognition, perception, and motivation occur [1]. An estimated 10% of youth between the ages of 9 and 18 experience subclinical psychosis-like experiences (PLEs), including atypical sensory experiences and delusional beliefs [2], which are associated with significantly greater odds of developing schizophrenia in adulthood [2, 3]. One proposed mechanism through which normative developmental processes may contribute to schizophrenia pathophysiology is through atypical brain development and refinement of synaptic connections across childhood and adolescence [4, 5]. A major goal of modern psychiatric research is to identify the genetic and brain-based antecedents of prodromal psychosis, with the ultimate goal of facilitating early intervention, prevention, and better treatment.

The most statistically significant genetic association with schizophrenia is located within the major histocompatibility complex (MHC) on chromosome 6—a region comprised of genes critical to innate and adaptive immunity [6,7,8,9,10]. Subsequent fine mapping of this locus revealed that a significant proportion of this signal can be attributed to multiallelic copy number variation of the complement component 4A (C4A) gene [11]. The complement system plays a crucial role in the peripheral immune response [12], and in the brain, it also plays an active role in pruning and facilitating microglial-mediated engulfment of synapses during development [13,14,15,16]. These findings have led to the hypothesis that the overactivation of complement genes in the brain may contribute to altered neural development and synaptic pruning, ultimately conferring risk for the phenotypic expression of schizophrenia symptoms.

The C4 locus encodes for C4A and C4B genes, which show common but complex structural variation. Each human haplotype carries a range of C4A and/or C4B copies, present in either long or short forms (i.e., C4AL, C4AS, C4BL, C4BS) depending on the presence of a human endogenous retroviral insertion (HERV) in intron 9 [17]. A higher C4A copy number is associated with a greater risk for schizophrenia, proportional to its effects on C4A expression [11, 18, 19]. C4A copy number also shows sex-dependent effects on schizophrenia risk, with stronger effects in males relative to females [18]. Importantly, while recent neuroimaging work suggests that C4A also affects cognition and brain structure in the general adult population [20], no study to date has examined the impact of C4A on childhood psychosis-related symptoms or brain development.

Here, we performed a comprehensive, phenome-wide association study between genetically regulated expression (GREx) at the C4 locus and childhood psychiatric, behavioral, and brain structural phenotypes to elucidate the role of ­C4 genetic variation on human development (Fig. 1). C4A GREx was imputed into a large, ancestrally diverse sample of 9–10-year-old youth who participated in the Adolescent Brain Cognitive DevelopmentSM (ABCD) Study [21]. We then assessed whether C4A GREx was associated with psychosis-like experiences as well as longitudinal psychiatric and cognitive phenotypes. We find that C4A GREx is not broadly related to childhood psychiatric symptoms or cognition in 9–10-year-old youth but is associated with a reduction in entorhinal cortex surface area. Furthermore, we find that entorhinal cortex surface area at 9–10 years of age is predictive of the number and severity of psychotic-like events at 1- and 2-year follow-up visits. Together, this phenome-wide characterization of the developmental effects of C4A provides evidence for regional effects on temporal lobe structure, which may be an early biomarker for schizophrenia risk.

Fig. 1
figure 1

Study design. A Individual-level genotype data from ABCD subjects was used to impute C4 structural alleles using a multi-ancestry reference panel (Kamitaki et al., [18]). C4A and C4B brain GREx was calculated using previously described weights (Sekar et al., [11]). B Generalized linear models were used to test for the associations between C4A GREx with psychotic-like experiences (PLE) in ABCD youth. Associations were examined in the whole multi-ancestry cohort, as well as within ancestry- and sex-specific subgroups. C We performed a phenome-wide association study (PheWAS) to assess the associations between C4A GREx and a host of developmental, cognitive, and behavioral phenotypes (the “Methods” section). D Linear mixed-effects models tested the association between predicted C4A GREx and neuroimaging measures of cortical thickness, surface area, and volume

Full size image

Results

Genetically regulated expression of C4A in the multi-ancestry ABCD youth cohort

Genetic and behavioral data were obtained from a community sample of sociodemographically and ethnically diverse youth who participated in the ABCD Study, an ongoing longitudinal investigation of child development in the USA [22]. Following quality control, including the removal of subjects with low C4 imputation quality, the sample consisted of 9638 ancestrally diverse youth (1694 African ancestry [AFR]; 5829 European ancestry [EUR]; 2115 Latinx ancestry; Fig. 1A). Imputation accuracy was confirmed by comparing the proportion of youth assigned to each common C4 structural allele to rates previously reported in a large sample of 765 EUR and 250 AFR ancestry individuals, from which the imputation panel was derived [18] (Additional file 1: Table S1). Subsequent analyses were restricted to 7789 youth with the 5 most common C4 structural haplotypes (1327 AFR; 4919 EUR; 1543 Latinx; 47.7% female; Additional file 1: Table S2), although in sensitivity analyses, results were concordant across the full spectrum of haplotypes. C4 structural haplotypes were then used to calculate GREx for C4A and C4B using previously described weights [11]. We validated the accuracy of C4 imputation weights for predicting gene expression in ancestrally diverse cohorts using data from 340 AFR, 863 EUR, and 50 Latinx donors from the PsychENCODE Consortium (Additional file 2: Fig. S1). We further validated that these weights were accurate in the subset of 5–15-year-old samples from PsychENCODE, which match the adolescent period investigated in this study. As expected, imputed expression values for C4A and C4B were negatively correlated (R = − 0.25, P < 2.2e− 16; Additional file 2: Fig. S2). To control for this, downstream statistical analyses included GREx for both C4A and C4B.

Adequacy of statistical power to detect GREx-phenotype associations

Power analyses were conducted in G*Power [23] to determine the range of detectable effects with the current sample size. With an estimated N of 6500, alpha of 0.05, 1 tested predictor, and 11 covariates, we have 95% power to detect effects (Cohen’s f2) as small as 0.002. As previous research in adults suggests that we can expect to see a significant association between C4A GREx and cognitive measures at Cohen’s f2 = 0.056 [20], power calculations indicate that our study is sufficiently powered to detect small but statistically significant effects, equivalent to a Cohen’s d of 0.004.

No widespread C4A associations with psychotic-like experiences or childhood cognition

Models testing associations between C4A GREx with rates and severity of PLEs were restricted to unrelated individuals who had complete data for the Prodromal Questionnaire Brief Version [24] (PQ-B) and the Kiddie Schedule for Affective Disorders and Schizophrenia [25] (KSADS-5) (N = 6655; 1168 AFR; 4113 EUR; 1374 Latinx). As previous research in clinical high-risk samples indicates that suspiciousness and unusual thought content best predict conversion to psychosis over a 2-year period [26], a binary prodromal psychosis score (PPbin) was computed by categorizing youth who endorsed experiencing at least 3 PLEs reflecting suspicious or unusual thought content and significant distress on the PQ-B and whose parents indicated the presence of hallucinations or delusions on the KSADS-5 (N = 611 cases, N = 6044 controls). We did not observe any significant associations between C4A GREx and psychosis symptoms (i.e., PQ-Bsym, PQ-Bsev, PPbin; Fig. 2, Additional file 1: Table S3, Additional file 2: Figs. S3-S5). As previous work has identified sexually dimorphic effects of C4A on the brain and behavior [18, 19], we further examined a potential interaction between sex and C4A GREx—which was not significant for any of the psychosis variables tested (Additional file 1: Table S3, Additional file 2: Figs. S6-S8).

Fig. 2
figure 2

Phenome-wide association between C4A GREx and cognitive, behavioral phenotypes in the ABCD cohort. A The relationship between C4A GREx and 145 phenotypes was tested in the multi-ancestry cohort (N’s per phenotype ~ 6000). Increased genetically regulated expression of C4A was associated with lower scores on the UPPS-P Impulsive Behavior Scale (Lynam, [27]) lack of premeditation/planning subscale (FDR-corrected P = 0.004). Phenotypes are grouped into broad categories by color. Negative or positive associations with C4A GREx are indicated by the direction of arrows. The FDR-based threshold for phenome-wide significance is indicated by the red dashed line. B Association (absolute 𝛽 ± standard error) between C4A GREx and each of the three psychosis variables investigated (binary case/control, PQ-B number PQ-B severity) in the multi-ancestry cohort (i.e., “ALL” youth [as in top panel]), as well as within each ancestry and sex separately. See Additional file 1: Table S3 for the regression model summary statistics. The results by ancestry and sex are shown in Additional file 2: Figs. S3-S8

Full size image

In adults and schizophrenia patients, genetic variation within the C4 locus, as well as in genes that regulate the complement system, has been associated with poorer cognition and memory [20, 28,29,30]. To assess whether similar effects are present in childhood, we performed a phenome-wide association study (PheWAS) using data from 145 cognitive and psychiatric phenotypes available from the ABCD dataset. In contrast to the findings in adults, C4A GREx was not robustly associated with childhood cognition or psychiatric phenotypes (Fig. 2, Additional file 1: Table S3, Additional file 2: Figs. S3-S8). The only association that surpassed correction for multiple testing in the multi-ancestry cohort was between C4A GREx and the UPPS-P Impulsive Behavior Scale [27] lack of premeditation/planning subscale (𝛽 = − 0.05, 95% CI [− 0.07, − 0.04], FDR-corrected P < 0.004), suggesting that youth with higher C4A GREx show less impulsivity and more careful thinking before taking action. This effect was partially mediated by entorhinal cortex size (indirect effect − 0.014, P = 0.006).

Additional PheWAS were performed using available data from follow-up time points 1 and 2 years later. There were no FDR-significant associations between C4A GREx and PheWAS variables at the 1- or 2-year follow-up (Additional file 1: Tables S4-S5, Additional file 2: Figs. S9-S14). Finally, to assess the convergence between C4 transcriptomic and genetic associations, regression analyses were performed for the joint model assessing the impact of 14 unique C4 allelic combinations on childhood PLEs. The C4 AL:AL haplotype showed nominally significant associations with greater PLEs; however, after Bonferroni correction, the results were no longer significant (Additional file 2: Fig. S15).

Replicable effects of the C4 locus on the entorhinal cortex structure

Our PheWAS results above indicate limited detectable effects of C4 variation on cognitive or psychiatric outcomes in youth. However, in schizophrenia, subtle brain structural changes are known to precede clinical and behavior manifestations [31,32,33]. To determine whether C4A affects childhood brain development, we next tested the association between GREx and measures of cortical thickness (CT), surface area (SA), and volume (VOL) across the whole brain as well as within regional parcellations using the Desikan-Kiliany atlas [34].

In the ancestrally diverse ABCD cohort, C4A GREx was associated with reduced regional SA of the entorhinal cortex (𝛽 = − 10.45, 95% CI [− 13.13, − 7.76], FDR-corrected P = 0.003; Fig. 3A, Additional file 1: Table S6). Furthermore, we found that the relationship between C4A GREx and entorhinal cortex was significant for both female and male youths (female: 𝛽 = − 8.90, 95% CI [− 12.46, − 5.34], P = 0.01; male: 𝛽 = − 11.6, 95% CI [− 15.56, − 7.63], P = 0.003; Fig. 3B). No other brain structure associations were observed after correction for multiple testing (Additional file 1: Tables S7-S9).

Fig. 3
figure 3

C4A GREx is associated with entorhinal cortex surface area and predicts longitudinal psychosis symptoms. A Linear regression analyses were performed in the multi-ancestry cohort to test the association between C4A GREx and regional brain structure (N’s with brain imaging data passing quality control ~ 6500). Brain regions demonstrating a nominally significant association between C4A GREx and surface area (mm2) are shown in color (P < 0.05); only the association with the entorhinal cortex survived FDR correction. The results for all brain regions are provided in Additional File 1: Tables S6-S9. B Female and male youths showed a significant effect of C4A GREx on reduced entorhinal cortex surface area. C Replication of the relationship between C4A GREx and entorhinal cortex (EC) surface area in 8357 adult females and 7790 adult males of European ancestry from the UK Biobank. D Youths falling in the top and bottom quartiles in terms of baseline entorhinal surface area were identified. E Smaller entorhinal cortex surface area (bottom quartile) at baseline predicted greater number and severity of psychosis-like experiences at the 1- and 2-year follow-up time points. The mean number and severity of PLEs are shown inside each bar plot. ~P = 0.05, *P < 0.05, **P < 0.01, ***P < 0.001

Full size image

To assess the robustness and reproducibility of the impact of C4A GREx on entorhinal cortex brain structure, we performed analogous analyses in 16,147 EUR ancestry subjects from the UK Biobank [35]. In the UK Biobank, C4A GREx was also significantly associated with entorhinal cortex SA, with the same direction of effect (full sample: 𝛽 = − 5.01, 95% CI [− 7.95, − 2.06], P = 0.00087; female 𝛽 = − 5.47, 95% CI [− 8.29, − 0.85], P = 0.02; male 𝛽 = − 4.57, 95% CI [− 10.10, − 0.84], P = 0.02; Fig. 3C).

The previous analyses show that C4A GREx is associated with a localized reduction in entorhinal cortex SA. In psychosis spectrum patients, the size of the entorhinal cortex is a predictor of the severity of delusions and negative symptoms [36, 37]. To assess this, we leveraged longitudinal psychosis data available from the ABCD 1- and 2-year follow-up visits. Due to the differences in overall brain size between male and female youths, analyses were performed in each sex separately and focused on comparing youth in the top and bottom quartile for baseline entorhinal cortex SA (Fig. 3D). We observed a significant difference in both the number and severity of reported PLEs at the 1-year follow-up when comparing between youth in the top and bottom quartiles for entorhinal cortex SA; more specifically, female and male youths with smaller entorhinal cortex SA at baseline reported greater frequency of PLEs and more distressing PLEs 1-year later (Wilcoxon Rank Sum Tests r = 0.05–0.09, P = 0.0002–0.04; Fig. 3E, Additional file 1: Table S10). At the 2-year follow-up, female youth showed significant differences in PLEs as a function of baseline entorhinal cortex SA; here again, smaller entorhinal cortex SA at baseline was associated with greater frequency and severity of PLEs (r = 0.11–0.12, P = 0.0002; Fig. 3E, Additional file 1: Table S10). There were no significant associations between entorhinal cortex SA and PLEs in male youth at the 2-year follow-up. Parallel regression analyses modeling entorhinal cortex SA as a continuous predictor were significant in females only, such that smaller entorhinal cortex at baseline predicted greater frequency and severity of PLEs at the 2-year follow-up (PQ-Bsym 𝛽 = − 0.001, 95% CI [− 0.002, − 0.0001], P = 0.03; PQ-Bsev 𝛽 = − 0.002, 95% CI [− 0.003, − 0.0005], P = 0.004).

Effects of C4A on the entorhinal cortex are independent of polygenic risk for schizophrenia

Schizophrenia is highly polygenic with thousands of genetic variants of small effect size acting additively to confer risk. To better understand the relationship between genetic risk for schizophrenia at the C4 locus and additive polygenic risk across the genome (outside the C4 locus), we computed polygenic risk scores (PRS) in ABCD youth of EUR ancestry (N = 3730 unrelated subjects) using the genome-wide association summary statistics of schizophrenia (40,675 cases, 64,643 controls [10]), excluding variants in the MHC region. Similar to previous findings [38], we did not observe a significant association between schizophrenia PRS and psychotic symptoms in youth—nor did we observe an interaction between C4A GREx and schizophrenia PRS in predicting PQ-Bsym, PQ-Bsev, or PPbin in ABCD youth of EUR ancestry or when examining EUR females and males separately (Additional file 1: Table S11).

Finally, we tested whether schizophrenia PRS—or its interaction with C4A GREx—was a significant predictor of brain structure. Surprisingly, schizophrenia PRS showed no association with entorhinal cortex structure, nor was there an interaction with C4A GREx, indicating a dissociable effect of complement activation and schizophrenia polygenic burden on this brain structure (Fig. 4, Additional file 1: Tables S11-S14). To confirm this, the association between C4A GREx and entorhinal cortex SA remained significant after controlling for schizophrenia PRS (𝛽C4A = − 8.88, 95% CI [− 12.46, − .29], P = 0.01). We also tested whether the interaction between brain surface area polygenic scores and C4A GREx predicted entorhinal cortex SA; the interaction was not significant (P > 0.05). Together, these findings suggest that the effects of C4A GREx on entorhinal cortex SA are specific to the C4 locus and are distinct from the effects of genetic liability for schizophrenia and adult brain size.

Fig. 4
figure 4

Interactions between C4A GREx and polygenic risk for schizophrenia. The interaction between C4A GREx and schizophrenia PRS was not a significant predictor of entorhinal cortex surface area in 9–10-year-old youth of European ancestry (N = 3349); higher C4A GREx was associated with smaller entorhinal cortex surface area across all levels of PRS

Full size image

Discussion

Here, we conduct a comprehensive multi-ancestry, phenome-wide association study of complement component 4 (C4) variation with childhood cognitive, psychiatric, and neuroimaging phenotypes. We find that genetically regulated expression of C4A is not associated with broad psychiatric symptoms or brain structure in 9–10-year-old children, but rather shows localized and replicable effects on the reduced regional surface area of the entorhinal cortex. Importantly, this C4A-mediated reduction in the entorhinal cortex surface area at 9–10 years of age predicted higher rates and greater severity of psychosis-like experiences at follow-up time points 1 and 2 years later, indicating that the effects of C4A on temporal lobe structure may be a biomarker for schizophrenia risk prior to symptom onset (Fig. 5).

Fig. 5
figure 5

Model for the effects of C4A brain expression on childhood brain structure. The effects of increased C4A GREx on reduced entorhinal cortex volume in childhood are dissociated from the effects of genome-wide polygenic risk for schizophrenia. Reduced entorhinal cortex surface area in childhood is associated with higher rates of youth-reported psychotic-like experiences (PLEs) in early adolescence, as well as greater severity of events. Note: Image under “psychotic-like experiences” taken from the Prodromal Questionnaire Brief Version (PQ-B) to assess the severity of distress associated with psychotic-like events

Full size image

Despite having sufficient power to detect associations with small effect sizes, we find that predicted C4A expression is not robustly associated with psychiatric symptoms or cognition in 9–10-year-old youth. These null results are surprising given the strength of the association between C4A and schizophrenia risk [11]. However, these negative findings comport with recent studies assessing the extent to which polygenic risk for schizophrenia is associated with broad psychiatric phenotypes in the ABCD cohort, which have found limited associations with fluid cognition and an index of binary psychosis severity in ABCD youth of European ancestry [39, 40]. Nevertheless, we identified a significant inverse association between predicted C4A expression and lack of planning on the UPPS-P Impulsive Behavior Scale [27]. Higher scores on this subscale are indicative of a tendency not to plan ahead; thus, we find that C4A brain expression is related to less severe symptomatology. The effects of schizophrenia genetic risk at the C4 locus and across the genome may exert more extensive effects on cognition and behavior during later developmental periods (e.g., adolescence and early adulthood), when clinically significant features of psychosis and schizophrenia are more likely to occur. We speculate that the developmental timing and tempo of synaptic pruning may be an important contributor to the emergence of psychiatric symptoms, such that increased pruning during early childhood may be developmentally advantageous until reaching a threshold of “over” pruning later in life.

In addition to the robust genetic association signal observed between schizophrenia and the major histocompatibility complex (MHC) locus, this region also shows associations with depression [41, 42] and to a weaker extent bipolar disorder [7, 43]. However, in both depression and bipolar disorder, dissecting the MHC signal using fine mapping techniques suggests that the observed associations are not driven by genetic variation in C4 haplotypes and arise from other variants in the region [44, 45]. Our null phenome-wide association study (PheWAS) testing the effects of the C4 locus on childhood psychiatric phenotypes agree these findings, suggesting that other sources of genetic variation apart from the C4 locus are likely influencing childhood cognition and behavior. Further large-scale studies will be needed to disentangle the genetic signal between MHC locus and diverse psychiatric disorders, and to examine the effects of other MHC-associated variants on neurodevelopmental trajectories.

In this demographically diverse multi-ancestry cohort, we observed a regionally localized association between C4A GREx and surface area of the entorhinal cortex. Our entorhinal cortex findings are consistent with the extant literature implicating medial temporal lobe structures in the pathophysiology of schizophrenia [46,47,48]. Indeed, in vivo neuroimaging has consistently found atypical volume of medial temporal lobe structures in patients with psychotic disorders relative to controls [49,50,51], and several postmortem studies have found aberrant entorhinal cortex cytoarchitecture in schizophrenia patients, suggestive of either atypical neuronal migration or altered pruning during neurodevelopment [52,53,54,55,56].

The entorhinal cortex serves as a relay for corticohippocampal interactions, serving as the major input and output structure of the hippocampus [57]. Recent high-field magnetic resonance imaging (MRI) research in humans suggests an anteroposterior division of entorhinal cortex functional connectivity, with anterior regions connected to anterior temporal brain areas critical for recognition memory, emotional processing, and social cognition and posterior regions connected to posterior medial brain areas that support episodic memory and spatial navigation [58, 59]. Because the entorhinal cortex receives afferent projections from multiple cortical and subcortical brain regions and is the primary source of information to the hippocampus [60], disruptions to this region during development have the potential to affect both the integration of sensory information and memory [58, 61, 62]. In line with this hypothesis, previous work has shown that smaller entorhinal cortex volume is a predictor of cognition and negative/disorganized symptoms in psychosis spectrum youth [37]. Similarly, we find that reduced entorhinal cortex surface area at 9–10 years of age predicts greater number and severity of psychosis-like events 1 and 2 years later. As brain volume is a function of surface area and cortical thickness, our findings of smaller surface area in youth with high genetic risk for schizophrenia at the C4 locus help to more precisely identify potential morphometric biomarkers predictive of psychosis risk.

Our surface area-specific findings in preadolescent youth contrast with adult neuroimaging studies, which have typically found more widespread reductions in cortical thickness in schizophrenia patients [63] and in clinically high-risk individuals who transition to psychosis [31]--although we did replicate the effects of C4A GREx on entorhinal cortex surface area in adults subjects from the UKBiobank. These contradictory findings may be explained by developmental differences in structural neuroimaging parameters. Recent neuroimaging studies have shown that cortical surface area continues to expand throughout the postnatal period and early childhood [64], while adolescence and early adulthood are characterized by nonlinear decreases in cortical thickness [65], likely driven by developmentally regulated alterations in cortical myelination and pruning. Thus, genetic risk for schizophrenia may have distinct effects on diverse aspects of brain structure at different developmental periods. Importantly, C4A cis-eQTLs (i.e., genomic loci that explain variation in gene expression) have been identified in the fetal brain [66], suggesting that the effects of C4A GREx on brain morphology may begin early in development. We therefore hypothesize that the observed effects of C4A GREx on the cross-sectional entorhinal cortex surface area in children stem from early alterations in brain development and that future studies of the effects of C4A on brain development are likely to observe associations with morphological metrics in flux within a particular developmental window (e.g., adolescent cortical thickness).

Notably, the effects of predicted C4A GREx on entorhinal cortex surface area are independent and dissociable from genome-wide polygenic risk for schizophrenia. We provide several lines of evidence to support this conclusion. First, polygenic risk scores (PRS) trained using weights from the largest available GWAS of schizophrenia [10] failed to predict entorhinal cortex surface area in youth of European ancestry. Second, when examining the interaction between C4A GREx and PRS, we find no evidence that schizophrenia PRS moderates the relationship between C4A GREx and entorhinal cortex surface area, and finally, the observed relationship between C4A GREx and entorhinal cortex surface area remains significant even when controlling for the effects of PRS. These findings indicate that the observed effects on the entorhinal cortex surface area are exclusive to genetic risk for schizophrenia at the C4 locus and are distinct from the effects of schizophrenia polygenic risk.

This study has several important limitations. First, our analyses were restricted to youth within the early to mid-adolescent age range. It remains unknown the extent to which C4A GREx may affect neurodevelopment during infancy and very early childhood, as well as during adolescence, when many disorders of mental health—including schizophrenia—begin to emerge. Second, our PheWAS analyses were limited to a single neuroimaging modality, and it is unclear whether genetic risk for schizophrenia on the C4 locus may impact measures of brain functional and structural connectivity, as well as the developmental timing with which these associations may occur. Third, while there are many environmental variables that are likely to influence brain and behavioral development, the current study did not examine the relationship between C4A GREx and environmental factors known to increase the risk for schizophrenia. It will be crucial for future studies to examine the impact of C4A expression on trajectories of brain development using multimodal neuroimaging techniques and incorporating potential interactions with environmental moderators.

Conclusions

In sum, the current study reports the first multi-ancestry phenome-wide association of genetic variation with the C4 locus in children. Our findings indicate that increased C4A brain expression in youth is associated with a localized reduction in the size of the entorhinal cortex that is independent of polygenic risk for schizophrenia. Overall, these data help to identify genetically driven brain-based biomarkers predictive of psychosis risk. Ultimately, a comprehensive understanding of the genetic and environmental precursors to psychosis will require longitudinal indices of brain development across the transition to adolescence and young adulthood, when clinically significant behavioral manifestations of psychiatric disorders often begin to emerge.

Methods

Participants

Data were obtained from the Adolescent Brain and Cognitive Development (ABCD) Study, an ongoing longitudinal investigation of brain and behavioral development in 11,878 children in the United States of America (USA). Youth were recruited primarily from schools within 21 catchment areas chosen to represent the ethnic and sociodemographic diversity of the larger US population [22]. Analyses were performed using de-identified phenotypic data from the baseline time point obtained from the ABCD 3.0 National Data Archive release (DOI 10.15154/1519007). Analyses were restricted to youth with available genetics data. See Additional file 1: Table S1 for subject characteristics.

Genotyping and imputation

Genotyping for the ABCD dataset was performed at the Rutgers University Cell and DNA Repository using the Affymetrix NIDA Smokescreen Array (RUCDR). Quality control was performed by the RUCDR and the ABCD Data Analysis and Informatics Core (DAIC) and included filtering for call signals and rates, and processing through the Ricopili pipeline [67]. Genotyping data was merged with data from the 1000 Genomes Project, and ancestry was assigned using the k-nearest neighbors classification with the first four genetic principal components.

C4 structural alleles were imputed for the ABCD dataset using Beagle4.1 [68] with the latest multi-ancestry C4 imputation reference panel composed of 1265 samples of diverse ancestry [18]. C4A and C4B genetically regulated expression (GREx) was predicted using previously described expression weights [11]. C4 structural allele dosages were multiplied by the following coefficients to compute GREx:

$$C4A\ \textrm{GREx}=\left(0.47\times \textrm{AL}\right)+\left(0.47\times \textrm{AS}\right)+\left(0.20\times \textrm{BL}\right)$$
$$C4B\ \textrm{GREx}=\left(1.03\times \textrm{BL}\right)+\left(0.88\times \textrm{BS}\right)$$

Downstream analyses were restricted to samples of African (AFR), European (EUR), and Latinx ancestries as the imputation panel was composed of those individuals, and C4 imputation accuracy (based on probabilistic dosage) was lower for Southeast Asian and East Asian. Individuals with an average imputed probabilistic dosage < 0.7 were excluded from subsequent analyses. To further ensure the use of high-quality imputed data, analyses were restricted to individuals with the five most common C4 structural haplotypes (AL, AL-AL, AL-BL, AL-BS, BS) resulting in a final sample of 7789 youth (see Additional file 1: Table S2).

For the UK Biobank, chromosome 6 of the phased haplotype data (i.e., ukb_hap_chr6_v2.bgen) was used to impute C4 structural alleles. Beagle4.1, with the multi-ancestry C4 reference panel, was used for imputation, as was done for the ABCD dataset.

Polygenic risk prediction

Polygenic risk scores (PRS) were calculated for EUR subjects. Genotype imputation was performed on the Michigan Imputation Server [69] using the TOPMed reference panel. Quality control included filtering for genotyping rate > 0.99, sample missingness < 0.01, Hardy-Weinberg Equilibrium P > 1 × 10−6, minor allele frequency > 1%, and imputation score > 0.8. GCTA v1.93.31 [70] was used to compute the genetic relationship matrix; a relatedness cutoff of 0.05 was applied. Genetic variants in the major histocompatibility complex (MHC) locus were excluded prior to PRS calculation. PRS were generated using summary statistics from the largest available genome-wide association study of schizophrenia [10] (40,675 cases, 64,643 controls) and adult brain surface area [71] (33,992 subjects) in PRS-CS [72], applying the LD reference panel from the 1000 Genomes Project phase 3 samples and the recommended global shrinkage parameter for highly polygenic traits (phi = 1e−2). Analyses assessing the interaction between C4 expression and PRS were restricted to youth with common C4 structural haplotypes (AL, AL-AL, AL-BL, AL-BS, BS), resulting in a final sample of 3730 EUR youth (female N = 1737; male N = 1993) and 7,715,663 genetic variants.

Neuroimaging data acquisition and preprocessing

Details of the ABCD magnetic resonance imaging (MRI) protocol have been described previously [73]. Briefly, imaging data were collected on either a Siemens Prisma, Phillips, or GE 750 3T scanner using 32-channel head or 64-channel head/neck coils. T1-weighted scan parameters were as follows: Siemens - matrix size 256 × 256, 176 slices, FOV 256 × 256, resolution (mm) 1 × 1 × 1, TR 2500 ms, TE 2.88 ms, flip angle 8°, total scan time 7:12; Phillips - matrix size 256 × 256, 225 slices, FOV 256x240, resolution (mm) 1 × 1 × 1, TR 6.31 ms, TE 2.9 ms, flip angle 8°, total scan time 5:38; and GE - matrix size 256 × 256, 208 slices, FOV 256 × 256, resolution (mm) 1 × 1 × 1, TR 2500 ms, TE 2 ms, flip angle 8°, total scan time 6:09.

Tabulated data representing T1-weighted measurements of subcortical volume (VOL), cortical thickness (CT), and surface area (SA) as derived from the Desikan parcellation atlas [34] in FreeSurfer v5.3 (http://surfer.nmr.mgh.harvard.edu/) were downloaded from the NDA. ABCD-recommended imaging inclusion criteria were followed to ensure the inclusion of high-quality MRI scans, and this excluded participants with serious MR findings and participants whose T1-weighted images or FreeSurfer parcellations failed to pass ABCD DAIC quality control. Regional MRI metrics were averaged across hemispheres.

Phenotypic data

Psychotic experiences were assessed with the Prodromal Questionnaire Brief Version [24] (PQ-B) and the Kiddie Schedule for Affective Disorders and Schizophrenia [25] (KSADS-5). The PQ-B is a 21-item child-report questionnaire designed to measure the presence and severity of psychotic-like experiences (PLEs) in childhood. PQ-B symptom scores (PQ-Bsym) were calculated as the total number of endorsed PLEs; severity score (PQ-Bsev) was calculated as the child-reported level of distress (range 1–5) across all endorsed items. In line with previous research indicating that unusual thought content and suspiciousness are most predictive of psychosis risk [26], a binary score (PPbin) was developed by defining prodromal psychosis “cases” as youth who endorsed experiencing at least 3 PLEs reflecting suspicious or unusual thought content (i.e., PQ-B items 1, 4, 5, 7, 8, 11, 12, 13–18), significant distress related to these experiences (i.e., PQ-B distress scores > 6 across items 1, 4, 5, 7, 8, 11, 12, 13–18), and whose parents indicated the child experienced hallucinations, delusions, and associated psychotic symptoms on the KSADS-5 assessment; youth not meeting these criteria served as the “control” group. Cognitive and psychiatric variables from the reported C4 phenome-wide association study (PheWAS) can be found in Additional file 1: Table S3.

Rates and severity of childhood PLEs were non-normally distributed, with the majority of youth reporting zero or few PLE symptoms (PQ-Bsym; M = 2.55, SD = 3.52) and endorsing low levels of PLE severity (PQ-Bsev; M = 6.09, SD = 10.39; Fig. 1B).

Statistical analyses

Generalized linear models with gamma family and log link were run using the glm package in R version 3.6.3 to test the association between C4 GREx and continuous measures of psychotic experiences (i.e., PQ-Bsym, PQ-Bsev), as these variables demonstrated a right-skewed distribution. The association between C4 GREx and PPbin was tested via logistic regression with binomial family using the glm package in R. Fixed effects included predicted C4A and C4B GREx, age, socioeconomic status (SES; taken as the average of parent education and income at the baseline time point), four genetic principal components, sex, and site ID. Statistical models run in related individuals using family ID as a random effect failed to converge; thus, analyses were restricted to one randomly selected subject from each family.

Models testing multivariable phenome-wide associations were also restricted to one individual per family. For normally distributed variables, linear models were fitted, while for non-normally distributed variables, generalized linear models with gamma distribution and log link were fitted. C4A and C4B GREx, age, SES, four genetic principal components, sex, and site ID were specified as fixed effects. Benjamini-Hochberg false discovery rate (FDR) correction was used to account for the number of variables tested at each time point (i.e., baseline, 1-year follow-up, 2-year follow-up). The significance of covariates of interest was assessed using the likelihood ratio test.

To test the association between C4 GREx and neuroimaging indices of brain volume, cortical thickness, and surface area, linear mixed models were run via the lme4 package in R. C4A and C4B GREx, age, four genetic principal components, SES, and sex were entered as fixed effects; MRI device serial number and family ID were entered as random effects. To control for global brain effects, whole brain volume, mean cortical thickness, or mean surface area were also included as covariates for respective models. The significance of covariates of interest was assessed using the likelihood ratio test, and FDR correction was used to account for the number of variables tested within each imaging measure (i.e., VOL, CT, SA).

Sex differences in the effects of C4 GREx on brain and behavioral phenotypes were tested by including an interaction term between C4A and C4B GREx and sex in statistical models. Follow-up analyses were performed in each sex separately.

The total number of participants in each analysis (i.e., those with complete PQ-Bsym, PQ-Bsev, PPbin, phenotype, and/or brain-imaging data) is provided in Additional file 1.

Plots were generated using R packages ggpubr, ggplot, and ggesg. PheWAS results were plotted in R by modifying publicly available code from Dr. Yoonjung Joo (https://rpubs.com/helloyjjoo/pheWAS_connectome).

Replication in the UK Biobank

The UK Biobank resource was accessed through application number 38029. Brain imaging data was acquired for 18,361 unrelated individuals with European ancestry. Haplotypes at the C4 locus were imputed as described for the ABCD dataset. After further restricting to individuals with the five most common C4 haplotypes (AL, AL-AL, AL-BL, AL-BS, BS), a total of 16,147 individuals remained in the final analyses. A linear model was then fitted between entorhinal cortex surface area and C4A GREx, using C4B GREx, sex, age at scan, average cortex surface area, location of head in the scanner (data fields 25756 to 25759), first four genotype PCs, and scan sites as covariates. A null model was also fitted with only covariates but not C4A GREx. The statistical significance of the association was evaluated by performing a likelihood ratio test. The analysis was performed in females (N = 8357) and males (N = 7790) separately, without using sex as a covariate.

Validation of expression weights for C4 genes in PsychENCODE

Previously harmonized genotype array and frontal cortex RNA-seq data from PsychENCODE [19, 74] consisted of uniformly processed data from six studies: BipSeq, LIBD_szControl, CMC_HBCC, CommonMind, BrainGVEX, and UCLA-ASD. We imputed C4 alleles in six studies from PsychENCODE separately in the same manner as the ABCD dataset using Beagle4.1 with the latest multi-ancestry C4 imputation reference panel. This analysis consisted of 340 AFR, 863 EUR, and 50 Latinx samples. C4 expression was then predicted using previously described expression weights [11], which was associated with observed C4 gene expression (Additional file 2: Fig. S1). Observed C4 gene expression was strongly associated with corresponding predicted gene expression, except for C4B gene expression in AFR samples (P > 0.05); C4A/C4B expression weights accurately predicted childhood and adolescent C4A expression in EUR PsychENCODE samples between the ages of 5-15 (P < 0.05).

Availability of data and materials

The datasets supporting the conclusions of this article are available through the Adolescent Brain Cognitive Development study [21] on the National Data Archive (https://doi.org/10.15154/1519007) and with application to the UK Biobank (https://www.ukbiobank.ac.uk/) [35]. Genotype array and RNA sequencing data were obtained from PsychENCODE (https:// doi.org/10.7303/syn12080241) [74]. The codes used to perform statistical analyses are available under the MIT license on GitHub (https://github.com/leamhernandez/C4A-ABCD) [75] and have also been deposited in Zenodo (https://doi.org/10.5281/zenodo.7542574) [76].

References

  1. Klosterkötter J, Hellmich M, Steinmeyer EM, Schultze-Lutter F. Diagnosing schizophrenia in the initial prodromal phase. Arch Gen Psychiatry. 2001;58(2):158–64.

    Article  PubMed  Google Scholar 

  2. Poulton R, Caspi A, Moffitt TE, Cannon M, Murray R, Harrington H. Children’s self-reported psychotic symptoms and adult schizophreniform disorder: a 15-year longitudinal study. Arch Gen Psychiatry. 2000;57(11):1053–8.

    Article  CAS  PubMed  Google Scholar 

  3. Healy C, Brannigan R, Dooley N, Coughlan H, Clarke M, Kelleher I, et al. Childhood and adolescent psychotic experiences and risk of mental disorder: a systematic review and meta-analysis. Psychol Med. 2019;49(10):1589–99.

    Article  PubMed  Google Scholar 

  4. Feinberg I. Schizophrenia: caused by a fault in programmed synaptic elimination during adolescence? J Psychiatr Res. 1982;17(4):319–34.

    Article  PubMed  Google Scholar 

  5. McGlashan T, Hoffman R. Schizophrenia as a disorder of developmentally reduced synaptic connectivity. Arch Gen Psychiatry. 2000;57(7):637–48.

    Article  CAS  PubMed  Google Scholar 

  6. Stefansson H, Ophoff RA, Steinberg S, Andreassen OA, Cichon S, Rujescu D, et al. Common variants conferring risk of schizophrenia. Nature. 2009;460(7256):744–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Purcell SM, Wray NR, Stone JL, Visscher PM, O’Donovan MC, Sullivan PF, et al. Common polygenic variation contributes to risk of schizophrenia and bipolar disorder. Nature. 2009;460(7256):748–52.

    Article  CAS  PubMed  Google Scholar 

  8. Ripke S, Neale BM, Corvin A, Walters JTR, Farh KH, Holmans PA, et al. Biological insights from 108 schizophrenia-associated genetic loci. Nature. 2014;511(7510):421–7.

    Article  CAS  PubMed Central  Google Scholar 

  9. Ripke S, Sanders AR, Kendler KS, Levinson DF, Sklar P, Holmans PA, et al. Genome-wide association study identifies five new schizophrenia loci. Nat Genet. 2011;43(10):969–76.

    Article  CAS  Google Scholar 

  10. Pardiñas AF, Holmans P, Pocklington AJ, Escott-Price V, Ripke S, Carrera N, et al. Common schizophrenia alleles are enriched in mutation-intolerant genes and in regions under strong background selection. Nat Genet. 2018;50(3):381–9.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Sekar A, Bialas AR, de Rivera H, Davis A, Hammond TR, Kamitaki N, et al. Schizophrenia risk from complex variation of complement component 4. Nature. 2016;530:177–83.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Janeway CA, Travers P, Walport M, Shlomchik MJ. The complement system and innate immunity. In: Immunobiology: the immune system in health and disease. 5th ed. New York: Garland Science; 2001.

    Google Scholar 

  13. Stevens B, Allen NJ, Vazquez LE, Howell GR, Christopherson KS, Nouri N, et al. The classical complement cascade mediates CNS synapse elimination. Cell. 2007;131(6):1164–78.

    Article  CAS  PubMed  Google Scholar 

  14. Schafer DP, Lehrman EK, Kautzman AG, Koyama R, Mardinly AR, Yamasaki R, et al. Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner. Neuron. 2012;74(4):691–705.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Comer AL, Jinadasa TI, Sriram BI, Phadke RA, Kretsge LN, Nguyen TPH, et al. Increased expression of schizophrenia-associated gene C4 leads to hypoconnectivity of prefrontal cortex and reduced social interaction. PLoS Biol. 2020;18(1):e3000604.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Yilmaz M, Yalcin E, Presumey J, Aw E, Ma M, Whelan CW, et al. Overexpression of schizophrenia susceptibility factor human complement C4A promotes excessive synaptic loss and behavioral changes in mice. Nat Neurosci. 2021;24(2):214–24.

    Article  CAS  PubMed  Google Scholar 

  17. Dangel AW, Mendoza AR, Baker BJ, Daniel CM, Carroll MC, Wu L-C, et al. The dichotomous size variation of human complement C4 genes is mediated by a novel family of endogenous retroviruses, which also establishes species-specific genomic patterns among Old World primates. Immunogenetics. 1994;40(6):425–36.

    Article  CAS  PubMed  Google Scholar 

  18. Kamitaki N, Sekar A, Handsaker RE, de Rivera H, Tooley K, Morris DL, et al. Complement genes contribute sex-biased vulnerability in diverse disorders. Nature. 2020;582(7813):577–81.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Kim M, Haney JR, Zhang P, Hernandez LM, Wang L, Perez-Cano L, et al. Brain gene co-expression networks link complement signaling with convergent synaptic pathology in schizophrenia. Nat Neurosci. 2021;24(6):799–809.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. O’Connell KS, Sønderby IE, Frei O, van der Meer D, Athanasiu L, Smeland OB, et al. Association between complement component 4A expression, cognitive performance and brain imaging measures in UK Biobank. Psychol Med. 2021;52(15):1–11.

  21. Volkow ND, Koob GF, Croyle RT, Bianchi DW, Gordon JA, Koroshetz WJ, et al. The conception of the ABCD study: from substance use to a broad NIH collaboration. Dev Cognit Neurosci. 2018;32:4–7. https://doi.org/10.15154/1519007.

    Article  Google Scholar 

  22. Garavan H, Bartsch H, Conway K, Decastro A, Goldstein RZ, Heeringa S, et al. Recruiting the ABCD sample: design considerations and procedures. Dev Cognit Neurosci. 2018;32:16–22.

    Article  CAS  Google Scholar 

  23. Faul F, Erdfelder E, Buchner A, Lang A-G. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behav Res Methods. 2009;41(4):1149–60.

    Article  PubMed  Google Scholar 

  24. Karcher NR, Barch DM, Avenevoli S, Savill M, Huber RS, Simon TJ, et al. Assessment of the Prodromal Questionnaire-Brief Child Version for measurement of self-reported psychoticlike experiences in childhood supplemental content. JAMA Psychiatry. 2018;75(8):853–61.

    Article  PubMed  PubMed Central  Google Scholar 

  25. Kaufman J, Birmaher B, Axelson D, Perepletchikova F, Brent D, Ryan N. Kiddie schedule for affective disorders and schizophrenia present and lifetime version 2013. Working Draft (KSADS-PL); 2013.

    Google Scholar 

  26. Cannon TD, Yu C, Addington J, Bearden CE, Cadenhead KS, Cornblatt BA, et al. An individualized risk calculator for research in prodromal psychosis. 2016;173(10):980–8.

  27. Lynam DR, Smith GT, Whiteside SP, Cyders MA. The UPPS-P: Assessing five personality pathways to impulsive behavior. West Lafayette, IN: Purdue University. 2006;10.

  28. Athanasiu L, Giddaluru S, Fernandes C, Christoforou A, Reinvang I, Lundervold AJ, et al. A genetic association study of CSMD1 and CSMD2 with cognitive function. Brain Behav Immun. 2017 Mar;1(61):209–16.

    Article  Google Scholar 

  29. Donohoe G, Walters J, Hargreaves A, Rose EJ, Morris DW, Fahey C, et al. Neuropsychological effects of the CSMD1 genome-wide associated schizophrenia risk variant rs10503253. Genes Brain Behav. 2013;12(2):203–9.

    Article  CAS  PubMed  Google Scholar 

  30. Zhang C, Lv Q, Fan W, Tang W, Yi Z. Influence of CFH gene on symptom severity of schizophrenia. Neuropsychiatr Dis Treat. 2017;13:697.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Cannon TD, Chung Y, He G, Sun D, Jacobson A, van Erp TGM, et al. Progressive reduction in cortical thickness as psychosis develops: a multisite longitudinal neuroimaging study of youth at elevated clinical risk. Biol Psychiatry. 2015;77(2):147–57.

    Article  PubMed  Google Scholar 

  32. Chung Y, Haut KM, He G, van Erp TGM, McEwen S, Addington J, et al. Ventricular enlargement and progressive reduction of cortical gray matter are linked in prodromal youth who develop psychosis. Schizophr Res. 2017;189:169–74.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Chung Y, Allswede D, Addington J, Bearden CE, Cadenhead K, Cornblatt B, et al. Cortical abnormalities in youth at clinical high-risk for psychosis: findings from the NAPLS2 cohort. NeuroImage: Clin. 2019;23:101862.

    Article  PubMed  Google Scholar 

  34. Desikan RS, Ségonne F, Fischl B, Quinn BT, Dickerson BC, Blacker D, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage. 2006;31(3):968–80.

    Article  PubMed  Google Scholar 

  35. Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Prasad KMR, Patel AR, Muddasani S, Sweeney J, Keshavan MS. The entorhinal cortex in first-episode psychotic disorders: a structural magnetic resonance imaging study. Am J Psychiatry. 2004;161(9):1612–9.

    Article  PubMed  Google Scholar 

  37. Roalf DR, Quarmley M, Calkins ME, Satterthwaite TD, Ruparel K, Elliott MA, et al. Temporal lobe volume decrements in psychosis spectrum youths. Schizophr Bull. 2017;43(3):601–10.

    PubMed  Google Scholar 

  38. Olde Loohuis LM, Mennigen E, Ori APS, Perkins D, Robinson E, Addington J, et al. Genetic and clinical analyses of psychosis spectrum symptoms in a large multiethnic youth cohort reveal significant link with ADHD. Transl Psychiatry. 2021;11(1):1–10.

    Article  Google Scholar 

  39. Loughnan R, Palmer CE, Makowski C, Thompson WK, Barch DM, Jernigan TL, et al. Unique prediction of developmental psychopathology from genetic and familial risk. J Child Psychol Psychiatry. 2022. https://doi.org/10.1111/jcpp.13649.

  40. Karcher NR, Paul SE, Johnson EC, Hatoum AS, Baranger DA, Agrawal A, et al. Psychotic-like experiences and polygenic liability in the ABCD Study®. Biol Psych: Cogn Neurosci Neuroimaging. 2022;7(1):45–55.

    Google Scholar 

  41. Howard DM, Adams MJ, Shirali M, Clarke T-K, Marioni RE, Davies G, et al. Genome-wide association study of depression phenotypes in UK Biobank identifies variants in excitatory synaptic pathways. Nat Commun. 2018;9(1):1470.

    Article  PubMed  PubMed Central  Google Scholar 

  42. Wray NR, Ripke S, Mattheisen M, Trzaskowski M, Byrne EM, Abdellaoui A, et al. Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet. 2018;50(5):668–81.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Ruderfer DM, Fanous AH, Ripke S, McQuillin A, Amdur RL, Gejman PV, et al. Polygenic dissection of diagnosis and clinical dimensions of bipolar disorder and schizophrenia. Mol Psychiatry. 2013;19(9):1017–24.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Glanville KP, Coleman JRI, Hanscombe KB, Euesden J, Choi SW, Purves KL, et al. Classical human leukocyte antigen alleles and C4 haplotypes are not significantly associated with depression. Biol Psychiatry. 2020;87(5):419–30.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Mullins N, Forstner AJ, O’Connell KS, Coombes B, Coleman JRI, Qiao Z, et al. Genome-wide association study of more than 40,000 bipolar disorder cases provides new insights into the underlying biology. Nat Genet. 2021;53(6):817–29.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Arnold SE. The medial temporal lobe in schizophrenia. J Neuropsychiatry Clin Neurosci. 2006;9(3):460–70.

    Google Scholar 

  47. Shenton ME, Dickey CC, Frumin M, McCarley RW. A review of MRI findings in schizophrenia. Schizophr Res. 2001;49(1–2):1–52.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Bakhshi K, Chance SA. The neuropathology of schizophrenia: a selective review of past studies and emerging themes in brain structure and cytoarchitecture. Neuroscience. 2015;303:82–102.

    Article  CAS  PubMed  Google Scholar 

  49. Honea R, Crow TJ, Passingham D, Mackay CE. Regional deficits in brain volume in schizophrenia: a meta-analysis of voxel-based morphometry studies. Am J Psychiatry. 2005;162(12):2233–45.

    Article  PubMed  Google Scholar 

  50. Mathew I, Gardin TM, Tandon N, Eack S, Francis AN, Seidman LJ, et al. Medial temporal lobe structures and hippocampal subfields in psychotic disorders: findings from the Bipolar-Schizophrenia Network on Intermediate Phenotypes (B-SNIP) study. JAMA Psychiatry. 2014;71(7):769–77.

    Article  PubMed  Google Scholar 

  51. van Erp TGM, Hibar DP, Rasmussen JM, Glahn DC, Pearlson GD, Andreassen OA, et al. Subcortical brain volume abnormalities in 2028 individuals with schizophrenia and 2540 healthy controls via the ENIGMA consortium. Mol Psychiatry. 2016;21(4):547–53.

    Article  PubMed  Google Scholar 

  52. Jakob H, Beckmann H. Prenatal developmental disturbances in the limbic allocortex in schizophrenics. J Neural Transm. 1986;65(3-4):303–26.

    Article  CAS  PubMed  Google Scholar 

  53. Arnold SE, Hyman BT, Van Hoesen GW, Damasio AR. Some cytoarchitectural abnormalities of the entorhinal cortex in schizophrenia. Arch Gen Psychiatry. 1991;48(7):625–32.

    Article  CAS  PubMed  Google Scholar 

  54. Arnold SE, Ruscheinsky DD, Han LY. Further evidence of abnormal cytoarchitecture of the entorhinal cortex in schizophrenia using spatial point pattern analyses. Biol Psychiatry. 1997;42(8):639–47.

    Article  CAS  PubMed  Google Scholar 

  55. Akil M, Lewis DA. Cytoarchitecture of the entorhinal cortex in schizophrenia. Brief Reports. 1997;154(7):1010–2.

    CAS  Google Scholar 

  56. Krimer LS, Herman MM, Saunders RC, Boyd JC, Hyde TM, Carter JM, et al. A qualitative and quantitative analysis of the entorhinal cortex in schizophrenia. Cereb Cortex. 1997;7(8):732–9.

    Article  CAS  PubMed  Google Scholar 

  57. Muñoz M, Insausti R. Cortical efferents of the entorhinal cortex and the adjacent parahippocampal region in the monkey (Macaca fascicularis). Eur J Neurosci. 2005;22(6):1368–88.

    Article  PubMed  Google Scholar 

  58. Schröder TN, Haak KV, Jimenez NIZ, Beckmann CF, Doeller CF. Functional topography of the human entorhinal cortex. eLife. 2015;4:e06738.

    Article  Google Scholar 

  59. Ranganath C, Ritchey M. Two cortical systems for memory-guided behaviour. Nat Rev Neurosci. 2012;13(10):713–26.

    Article  CAS  PubMed  Google Scholar 

  60. Garcia AD, Buffalo EA. Anatomy and function of the primate entorhinal cortex. Annu Rev Vis Sci. 2020;6:411–32.

    Article  PubMed  Google Scholar 

  61. Eichenbaum H, Yonelinas AP, Ranganath C. The medial temporal lobe and recognition memory. Annu Rev Neurosci. 2007;30:123–52.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. Moser EI, Moser M-B. Grid cells and neural coding in high-end cortices. Neuron. 2013;80(3):765–74.

    Article  CAS  PubMed  Google Scholar 

  63. van Haren NEM, Schnack HG, Cahn W, van den Heuvel MP, Lepage C, Collins L, et al. Changes in cortical thickness during the course of illness in schizophrenia. Arch Gen Psychiatry. 2011;68(9):871–80.

    Article  PubMed  Google Scholar 

  64. Gilmore JH, Knickmeyer RC, Gao W. Imaging structural and functional brain development in early childhood. Nat Rev Neurosci. 2018;19(3):123–37.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  65. Tamnes CK, Herting MM, Goddings A-L, Meuwese R, Blakemore S-J, Dahl RE, et al. Development of the cerebral cortex across adolescence: a multisample study of inter-related longitudinal changes in cortical volume, surface area, and thickness. J Neurosci. 2017;37(12):3402–12.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  66. O’Brien HE, Hannon E, Hill MJ, Toste CC, Robertson MJ, Morgan JE, et al. Expression quantitative trait loci in the developing human brain and their enrichment in neuropsychiatric disorders. Genome Biol. 2018;19(1):1–13.

    Article  Google Scholar 

  67. Lam M, Awasthi S, Watson HJ, Goldstein J, Panagiotaropoulou G, Trubetskoy V, et al. RICOPILI: Rapid Imputation for COnsortias PIpeLIne. Bioinformatics. 2020;36(3):930–3.

    Article  CAS  PubMed  Google Scholar 

  68. Browning BL, Browning SR. Genotype imputation with millions of reference samples. Am J Hum Genet. 2016;98(1):116–26.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  69. Das S, Forer L, Schönherr S, Sidore C, Locke AE, Kwong A, et al. Next-generation genotype imputation service and methods. Nat Genet. 2016;48(10):1284–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  70. Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet. 2011;88(1):76–82.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  71. Grasby KL, Jahanshad N, Painter JN, Colodro-Conde L, Bralten J, Hibar DP, et al. The genetic architecture of the human cerebral cortex. Science. 2020;367(6484).

  72. Ge T, Chen CY, Ni Y, Feng YCA, Smoller JW. Polygenic prediction via Bayesian regression and continuous shrinkage priors. Nature. Communications. 2019;10(1).

  73. Casey BJ, Cannonier T, Conley MI, Cohen AO, Barch DM, Heitzeg MM, et al. The Adolescent Brain Cognitive Development (ABCD) study: imaging acquisition across 21 sites. Dev Cognit Neurosci. 2018;32:43–54.

    Article  CAS  Google Scholar 

  74. Gandal MJ, Zhang P, Hadjimichael E, Walker RL, Chen C, Liu S, et al. Transcriptome-wide isoform-level dysregulation in ASD, schizophrenia, and bipolar disorder. Science. 2018;362(6420):eaat8127. https://doi.org/10.7303/syn12080241.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  75. Hernandez LM, Kim M, Zhang P, Bethlehem R, Hoftman G, Loughnan R, Smith D, Bookheimer SY, Fan CC, Bearden CE, Thompson W, Gandal M. Multi-ancestry phenome-wide association of complement component 4 variation with psychiatric and brain phenotypes in youth. GitHub. (2023). https://github.com/leamhernandez/C4A-ABCD.

  76. Hernandez LM, Kim M, Zhang P, Bethlehem R, Hoftman G, Loughnan R, Smith D, Bookheimer SY, Fan CC, Bearden CE, Thompson W, Gandal M. Multi-ancestry phenome-wide association of complement component 4 variation with psychiatric and brain phenotypes in youth. Zenodo. (2023). https://doi.org/10.5281/zenodo.7542574.

Download references

Acknowledgements

Not applicable.

Review history

The review history is available as Additional file 3.

Peer review information

Veronique van den Berghe was the primary editor of this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.

Funding

This work was supported by the National Institute of Mental Health (R01MH121521 to MJG; R01MH123922 to MJG; P50HD103557 to MJG, R01MH107250, and U01MH081902 to CEB; K00MH119663 to LMH; R01MH120025 to CCF; T32MH073526 and F30MH125523 to MK), SFARI Bridge to Independence Award to MJG, National Center for Advancing Translational Science (UL1TR001881 to GDH), National Institute of General Medical Sciences (T32GM008042 to MK), UCLA Friends of the Semel Institute Research Scholar Award to GDH, Karen Seykora NARSAD Young Investigator Grant from the Brain & Behavior Research Foundation to GDH, and the Harvey L. and Maud C. Sorensen Foundation Fellowship to GDH.

Data used in the preparation of this article were obtained from the Adolescent Brain Cognitive DevelopmentSM (ABCD) Study (https://abcdstudy.org), held in the NIMH Data Archive (NDA). This is a multisite, longitudinal study designed to recruit more than 10,000 children aged 9–10 and follow them over 10 years into early adulthood. The ABCD Study® is supported by the National Institutes of Health and additional federal partners under award numbers U01DA041048, U01DA050989, U01DA051016, U01DA041022, U01DA051018, U01DA051037, U01DA050987, U01DA041174, U01DA041106, U01DA041117, U01DA041028, U01DA041134, U01DA050988, U01DA051039, U01DA041156, U01DA041025, U01DA041120, U01DA051038, U01DA041148, U01DA041093, U01DA041089, U24DA041123, and U24DA041147. A full list of supporters is available at https://abcdstudy.org/federal-partners.html. A listing of participating sites and a complete listing of the study investigators can be found at https://abcdstudy.org/consortium_members/. ABCD consortium investigators designed and implemented the study and/or provided data but did not necessarily participate in the analysis or writing of this report. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIH or ABCD consortium investigators.

Author information

Authors and Affiliations

  1. Department of Psychiatry, Semel Institute, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, CA, 90095, USA

    Leanna M. Hernandez, Minsoo Kim, Pan Zhang, Gil Hoftman, Susan Y. Bookheimer, Carrie E. Bearden & Michael J. Gandal

  2. Department of Human Genetics, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, CA, 90095, USA

    Minsoo Kim & Michael J. Gandal

  3. Program in Neurobehavioral Genetics, Semel Institute, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, CA, 90095, USA

    Minsoo Kim

  4. University of Cambridge, Department of Psychiatry, Cambridge Biomedical Campus, Cambridge, CB2 0SZ, UK

    Richard A. I. Bethlehem

  5. Population Neuroscience and Genetics Lab, University of California, San Diego, San Diego, CA, 92093, USA

    Robert Loughnan, Diana Smith, Chun Chieh Fan & Wesley K. Thompson

  6. Department of Psychiatry, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA

    Michael J. Gandal

  7. Lifespan Brain Institute at Penn Med and the Children’s Hospital of Philadelphia, Philadelphia, PA, USA

    Michael J. Gandal

  8. Department of Genetics, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA

    Michael J. Gandal

Authors
  1. Leanna M. Hernandez
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Minsoo Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Pan Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Richard A. I. Bethlehem
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Gil Hoftman
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Robert Loughnan
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Diana Smith
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Susan Y. Bookheimer
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Chun Chieh Fan
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Carrie E. Bearden
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Wesley K. Thompson
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Michael J. Gandal
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

This study was designed by LMH and MJG. SYB, CCF, and WKT contributed to the ongoing project design and management of the Adolescent Brain Cognitive Development Study. LMH performed the data analysis, with support from MK, PZ, CCF, WKT, and MJG. LMH, MK, PZ, RAI, GH, RL, DS, SYB, CCF, CEB, WKT, and MJG contributed to the data interpretation. LMH, MK, PZ, and MJG wrote the manuscript, with critical input from all authors. The authors read and approved the final manuscript.

Authors’ Twitter handles

@LM__Hernandez (Leanna M. Hernandez); @mikejg84 (Michael J. Gandal)

Corresponding authors

Correspondence to Leanna M. Hernandez or Michael J. Gandal.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Table S1.

Subject characteristics as assessed at the baseline timepoint in youth with the 5 most common C4 structural haplotypes. Table S2. Count and proportion of C4 alleles in the ABCD cohort (1,694 AFR; 5,829 EUR; 2,115 Latinx). Table S3. Summary of PheWAS results in the ABCD cohort. Table S4. Summary of 1-year follow-up PheWAS in the ABCD cohort. Table S5. Summary of 2-year follow-up PheWAS in the ABCD cohort. Table S6. Summary of C4 GREx associations with regional brain surface area in the ABCD cohort. Table S7. Summary of C4 GREx associations with whole brain volume (WBV), mean surface area (SA), and mean cortical thickness (CT). Table S8. Summary of C4 GREx associations with regional brain cortical thickness in the ABCD cohort. Table S9. Summary of C4 GREx associations with subcortical brain volume in the ABCD cohort. Table S10. Summary of Wilcoxon Rank Sum Tests comparing PQ-B symptoms and PQ-B severity at 1- and 2-year follow-up timepoints in youth in the top/bottom quartile for entorhinal cortex surface area at baseline. Table S11. Summary of schizophrenia PRS associations in youth of EUR ancestry. Table S12. Summary of schizophrenia PRS associations with regional cortical thickness in youth of EUR ancestry. Table S13. Summary of schizophrenia PRS associations with regional surface area in youth of EUR ancestry. Table S14. Summary of schizophrenia PRS associations with subcortical volume in youth of EUR ancestry.

Additional file 2: Fig. S1.

Relationship between predicted C4A/C4B expression and observed C4A/C4B expression across ancestries. Fig. S2. C4 descriptives by ancestry. Fig. S3. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in youth of AFR ancestry. Fig. S4. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in youth of Latinx ancestry. Fig. S5. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in youth of EUR ancestry. Fig. S6. Sex-differences in the association between predicted C4A gene expression and behavioral phenotypes in the ABCD cohort. Fig. S7. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in female youth. Fig. S8. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in male youth. Fig. S9. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in the ABCD cohort at the 1-year follow-up. Fig. S10. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in female youth at the 1-year follow-up. Fig. S11. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in male youth at the 1-year follow-up. Fig. S12. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in the ABCD cohort at the 2-year follow-up. Fig. S13. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in female youth at the 2-year follow-up. Fig. S14. Phenome-wide association between predicted C4A gene expression and behavioral phenotypes in male youth at the 2-year follow-up. Fig. S15. Relationship between C4 haplotypes and baseline psychosis-like experiences in the multi-ancestry ABCD sample. Fig. S16. Factors influencing entorhinal cortex surface area at baseline.

Additional file 3.

Review history.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Hernandez, L.M., Kim, M., Zhang, P. et al. Multi-ancestry phenome-wide association of complement component 4 variation with psychiatric and brain phenotypes in youth. Genome Biol 24, 42 (2023). https://doi.org/10.1186/s13059-023-02878-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s13059-023-02878-0

Keywords

Genome Biology

ISSN: 1474-760X

Contact us