Shared transethnic genetic basis of panic disorder and psychiatric and related intermediate phenotypes

Discrimination between healthy participants and people with panic disorder based on polygenic scores for psychiatric disorders and for intermediate phenotypes using machine learning

OBJECTIVE

Panic disorder is a modestly heritable condition. Currently, diagnosis is based only on clinical symptoms; identifying objective biomarkers and a more reliable diagnostic procedure is desirable. We investigated whether people with panic disorder can be reliably diagnosed utilizing combinations of multiple polygenic scores for psychiatric disorders and their intermediate phenotypes, compared with single polygenic score approaches, by applying specific machine learning techniques.

METHODS

Polygenic scores for 48 psychiatric disorders and intermediate phenotypes based on large-scale genome-wide association studies (n = 7556-1,131,881) were calculated for people with panic disorder (n = 718) and healthy controls (n = 1717). Discrimination between people with panic disorder and healthy controls was based on the 48 polygenic scores using five methods for classification: logistic regression, neural networks, quadratic discriminant analysis, random forests and a support vector machine. Differences in discrimination accuracy (area under the curve) due to an increased number of polygenic score combinations and differences in the accuracy across five classifiers were investigated.

RESULTS

All five classifiers performed relatively well for distinguishing people with panic disorder from healthy controls by increasing the number of polygenic scores. Of the 48 polygenic scores, the polygenic score for anxiety UK Biobank was the most useful for discrimination by the classifiers. In combinations of two or three polygenic scores, the polygenic score for anxiety UK Biobank was included as one of polygenic scores in all classifiers. When all 48 polygenic scores were used in combination, the greatest areas under the curve significantly differed among the five classifiers. Support vector machine and logistic regression had higher accuracy than quadratic discriminant analysis and random forests. For each classifier, the greatest area under the curve was 0.600 ± 0.030 for logistic regression (polygenic score combinations N = 14), 0.591 ± 0.039 for neural networks (N = 9), 0.603 ± 0.033 for quadratic discriminant analysis (N = 10), 0.572 ± 0.039 for random forests (N = 25) and 0.617 ± 0.041 for support vector machine (N = 11). The greatest areas under the curve at the best polygenic score combination significantly differed among the five classifiers. Random forests had the lowest accuracy among classifiers. Support vector machine had higher accuracy than neural networks.

CONCLUSIONS

These findings suggest that increasing the number of polygenic score combinations up to approximately 10 effectively improved the discrimination accuracy and that support vector machine exhibited greater accuracy among classifiers. However, the discrimination accuracy for panic disorder, when based solely on polygenic score combinations, was found to be modest.

Higher polygenic risk scores for anxiety disorders are associated with reduced area in the anterior cingulate gyrus

Anxiety disorders are heterogeneous, show a moderate genetic contribution and are associated with inconsistent cortical structure alterations. Here, we investigated whether genetic factors for anxiety disorders contribute to cortical alterations by conducting polygenic risk score (PRS) analyses. We calculated PRSs for anxiety disorders at several P value thresholds (from P_T ≤ 5.0 × 10^-8 to P_T ≤ 1.0) based on the latest large-scale genome-wide association study of anxiety disorders from the UK biobank (25,453 cases; 58,113 controls) in an independent sample of psychiatrically and physically healthy subjects (n = 174). Using regression after adjusting for confounding factors, we tested whether these PRSs were associated with the surface area and cortical thickness in 34 bilateral brain regions extracted using FreeSurfer. A higher PRS for anxiety disorders at P_T ≤ 1.0 was significantly associated with a reduced right caudal anterior cingulate area (beta = -0.25, p_uncorrected = 9.51 × 10^-4, p_corrected = 0.032). PRSs based on more common SNPs, especially from P_T ≤ 0.01 to P_T ≤ 1.0, were associated with the right caudal anterior cingulate area (a maximum at P_T ≤ 0.5: R² = 0.066, beta = -0.27, p_uncorr = 3.81 × 10^-4, p_corr = 0.013). Furthermore, individuals in the highest quartile for anxiety disorder PRS had lower surface area and volume in the right anterior cingulate gyrus than those in the lowest quartile. We suggest a shared genetic etiology between anxiety disorders and structural features of the anterior cingulate gyrus, possibly contributing to the pathogenesis of anxiety disorders via emotional dysregulations. Our findings suggest the potential usefulness of PRS to reduce pathological heterogeneity among anxiety disorders.

Shared genetic basis between reproductive behaviors and anxiety-related disorders

Reproductive behaviors are associated with risks for psychiatric disorders. Reproductive phenotypes are moderately heritable and have genetic overlaps with risks for psychiatric disorders. However, the genetic and causal relationships between anxiety-related disorders or specific anxiety disorders and reproductive phenotypes remain unknown. We utilized large-scale genome-wide association study (GWAS) results (n = 9537-542,901) for five reproductive phenotypes [age at menarche, age at first sexual intercourse (AFS), age at first birth (AFB), number of children ever born (NEB), and age at menopause] and five anxiety-related disorders [panic disorder, anxiety disorders from the ANGST and the UK biobank (UKBB), posttraumatic stress disorder (PTSD) and obsessive-compulsive disorder (OCD)]. To assess genetic correlations and causal associations, linkage disequilibrium score regression and Mendelian randomization analyses, respectively, were performed. We found that AFS and AFB were negatively correlated with anxiety disorders ANGST (AFS: r_g ± SE = -0.28 ± 0.08, p = 6.00 × 10^-4; AFB: -0.45 ± 0.11, p = 3.26 × 10^-5), anxiety disorders UKBB (AFS: -0.18 ± 0.03, p = 9.64 × 10^-9; AFB; -0.25 ± 0.03, p = 2.90 × 10^-13) and PTSD (AFS: -0.42 ± 0.12, p = 4.00 × 10^-4; AFB: -0.44 ± 0.12, p = 2.00 × 10^-4) and positively correlated with OCD (AFS: 0.25 ± 0.05, p = 2.46 × 10^-6; AFB: 0.25 ± 0.05, p = 3.92 × 10^-7). Conversely, NEB was negatively correlated with OCD (-0.28 ± 0.08, p = 6.00 × 10^-4). We revealed bidirectional effects between earlier AFS and AFB and anxiety disorders (odds ratios: OR_{earlier AFS→Anxiety} = 1.64, p = 2.27 × 10^-8; OR_{earlier AFB→Anxiety} = 1.15, p = 2.28 × 10^-3; OR_{Anxiety→earlier AFS} = 1.02, p = 6.62 × 10^-8; OR_{Anxiety→earlier AFB} = 1.08, p = 1.60 × 10^-4). In contrast, we observed unidirectional effects of later AFS and AFB on OCD (OR_{later AFS→OCD} = 2.18, p = 2.16 × 10^-6; OR_{later AFB→OCD} = 1.22, p = 0.016). We suggest that those who have earlier sexual debut and childbirth are prone to risk for anxiety disorders and vice versa, while those who have later sexual debut and childbirth are genetically prone to risk for OCD. Our findings further support revising the diagnostic criteria (DSM-5) such that OCD is independent from anxiety disorders.

Polygenic Risk Scores Differentiating Schizophrenia From Bipolar Disorder Are Associated With Premorbid Intelligence in Schizophrenia Patients and Healthy Subjects

BACKGROUND

Impairments in intelligence are more severe in patients with schizophrenia (SCZ) than in patients with bipolar disorder (BD) despite clinical and genetic similarities between the disorders. Genetic loci differentiating SCZ from BD, that is, SCZ-specific risk, have been identified. Polygenetic [risk] scores (PGSs) for SCZ-specific risk are higher in SCZ patients than in healthy controls (HCs). However, the influence of genetic risk on impaired intelligence is poorly understood. Here, we investigated whether SCZ-specific risk could predict impairments in intelligence in SCZ patients and HCs.

METHODS

Large-scale genome-wide association study datasets related to SCZ vs BD, childhood intelligence (CHI), and adulthood intelligence (n = 12 441-282 014) were utilized to compute PGSs. PGSs derived from the genome-wide association studies were calculated for 130 patients with SCZ and 146 HCs. Premorbid and current intelligence and the decline were measured in SCZ patients and HCs. Correlations between PGSs and intelligence functions were investigated.

RESULTS

High PGSs for SCZ-specific risk were correlated with low premorbid intelligence in SCZ patients and HCs (β = -0.17, P = 4.12 × 10-3). The correlation was still significant after adjusting for diagnostic status (β = -0.13, P = .024). There were no significant correlations between PGSs for SCZ-specific risk and current intelligence or intelligence decline (P > .05). PGSs for CHI were lower in SCZ patients than in HCs (R2 = 0.025, P = .025), while the PGSs for CHI were not significantly correlated with premorbid and current intelligence, the decline, or the PGSs for SCZ-specific risk (P > .05).

CONCLUSIONS

These findings suggest that genetic factors differentiating SCZ from BD might affect the pathogenesis of SCZ and/or pathological differences between SCZ and BD via the impairment of premorbid intelligence, that is, crystallized intelligence, while genetic factors for CHI might affect the pathogenesis of SCZ but not via impairments in intelligence.

Comparison of Emotional Dysregulation Features in Cyclothymia and Adult ADHD

Background and Objectives: Emotional dysregulation is central to the problem of the overlap between attention-deficit/hyperactivity disorder (ADHD) and cyclothymia. The aim of the study was to evaluate comorbidity rates between ADHD and cyclothymic disorder and to explore demographic and clinical differences among the groups, focusing on affective temperament and emotional dysregulation. Materials and Methods: One hundred sixty-five outpatients attending the Second Psychiatry Unit at the Santa Chiara University Hospital (Pisa) were consecutively recruited: 80 were diagnosed with ADHD, 60 with cyclothymic disorder, and 25 with both conditions. Temperament Evaluation of Memphis, Pisa, Paris, and San Diego (TEMPS-M) and the 40-item version of Reactivity, Intensity, Polarity, and Stability questionnaire (RI-PoSt-40) were administered. Results: Cyclothymic patients were more frequently female and older with respect to the ADHD groups. Both comorbid and non-comorbid ADHD patients showed significantly lower educational attainment and more frequently had substance use disorders. Panic disorder was common in non-comorbid cyclothymic patients, who showed significantly higher rates of familial panic disorder, major depressive disorder and suicide attempts in comparison with patients only diagnosed with ADHD. Cyclothymic patients without ADHD were also characterized by fewer hyperthymic temperamental traits, higher depressive and anxious dispositions, and a greater negative emotionality. No significant differences among groups were observed for cyclothymic temperament and overall negative emotional dysregulation, but comorbid patients with both conditions scored the highest in these subscales. This group also showed significantly higher affective instability with respect to ADHD patients without cyclothymia and was less frequently diagnosed with bipolar disorder type II than patients from both the other groups. Conclusions: ADHD and cyclothymia often co-occur and show similar levels of emotional dysregulation. However, cyclothymic patients may be more prone to negative emotionality in clinical settings. Subjects with “sunny” cyclothymic features might escape the attention of clinicians unless ADHD is present.