Assessing the effects of antipsychotic medications on schizophrenia functional analysis: a postmortem proteome study

A characteristic cerebellar biosignature for bipolar disorder, identified with fully automatic machine learning

Background

Transcriptomic profile differences between patients with bipolar disorder and healthy controls can be identified using machine learning and can provide information about the potential role of the cerebellum in the pathogenesis of bipolar disorder.With this aim, user-friendly, fully automated machine learning algorithms can achieve extremely high classification scores and disease-related predictive biosignature identification, in short time frames and scaled down to small datasets.

Method

A fully automated machine learning platform, based on the most suitable algorithm selection and relevant set of hyper-parameter values, was applied on a preprocessed transcriptomics dataset, in order to produce a model for biosignature selection and to classify subjects into groups of patients and controls. The parent GEO datasets were originally produced from the cerebellar and parietal lobe tissue of deceased bipolar patients and healthy controls, using Affymetrix Human Gene 1.0 ST Array.

Results

Patients and controls were classified into two separate groups, with no close-to-the-boundary cases, and this classification was based on the cerebellar transcriptomic biosignature of 25 features (genes), with Area Under Curve 0.929 and Average Precision 0.955. The biosignature includes both genes connected before to bipolar disorder, depression, psychosis or epilepsy, as well as genes not linked before with any psychiatric disease. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed participation of 4 identified features in 6 pathways which have also been associated with bipolar disorder.

Conclusion

Automated machine learning (AutoML) managed to identify accurately 25 genes that can jointly - in a multivariate-fashion - separate bipolar patients from healthy controls with high predictive power. The discovered features lead to new biological insights. Machine Learning (ML) analysis considers the features in combination (in contrast to standard differential expression analysis), removing both irrelevant as well as redundant markers, and thus, focusing to biological interpretation.

Canonical and Non-Canonical Antipsychotics' Dopamine-Related Mechanisms of Present and Next Generation Molecules: A Systematic Review on Translational Highlights for Treatment Response and Treatment-Resistant Schizophrenia

Schizophrenia is a severe psychiatric illness affecting almost 25 million people worldwide and is conceptualized as a disorder of synaptic plasticity and brain connectivity. Antipsychotics are the primary pharmacological treatment after more than sixty years after their introduction in therapy. Two findings hold true for all presently available antipsychotics. First, all antipsychotics occupy the dopamine D2 receptor (D2R) as an antagonist or partial agonist, even if with different affinity; second, D2R occupancy is the necessary and probably the sufficient mechanism for antipsychotic effect despite the complexity of antipsychotics' receptor profile. D2R occupancy is followed by coincident or divergent intracellular mechanisms, implying the contribution of cAMP regulation, β-arrestin recruitment, and phospholipase A activation, to quote some of the mechanisms considered canonical. However, in recent years, novel mechanisms related to dopamine function beyond or together with D2R occupancy have emerged. Among these potentially non-canonical mechanisms, the role of Na²⁺ channels at the dopamine at the presynaptic site, dopamine transporter (DAT) involvement as the main regulator of dopamine concentration at synaptic clefts, and the putative role of antipsychotics as chaperones for intracellular D2R sequestration, should be included. These mechanisms expand the fundamental role of dopamine in schizophrenia therapy and may have relevance to considering putatively new strategies for treatment-resistant schizophrenia (TRS), an extremely severe condition epidemiologically relevant and affecting almost 30% of schizophrenia patients. Here, we performed a critical evaluation of the role of antipsychotics in synaptic plasticity, focusing on their canonical and non-canonical mechanisms of action relevant to the treatment of schizophrenia and their subsequent implication for the pathophysiology and potential therapy of TRS.

Antipsychotic drug use complicates assessment of gene expression changes associated with schizophrenia

Recent postmortem transcriptomic studies of schizophrenia (SCZ) have shown hundreds of differentially expressed genes. However, the extent to which these gene expression changes reflect antipsychotic drug (APD) exposure remains uncertain. We compared differential gene expression in the prefrontal cortex of SCZ patients who tested positive for APDs at the time of death with SCZ patients who did not. APD exposure was associated with numerous changes in the brain transcriptome, especially among SCZ patients on atypical APDs. Brain transcriptome data from macaques chronically treated with APDs showed that APDs affect the expression of many functionally relevant genes, some of which show expression changes in the same directions as those observed in SCZ. Co-expression modules enriched for synaptic function showed convergent patterns between SCZ and some of the APD effects, while those associated with inflammation and glucose metabolism exhibited predominantly divergent patterns between SCZ and APD effects. In contrast, major cell-type shifts inferred in SCZ were primarily unaffected by APD use. These results show that APDs may confound SCZ-associated gene expression changes in postmortem brain tissue. Disentangling these effects will help identify causal genes and improve our neurobiological understanding of SCZ.

Differential vulnerability of anterior cingulate cortex cell types to diseases and drugs

In psychiatric disorders, mismatches between disease states and therapeutic strategies are highly pronounced, largely because of unanswered questions regarding specific vulnerabilities of different cell types and therapeutic responses. Which cellular events (housekeeping or salient) are most affected? Which cell types succumb first to challenges, and which exhibit the strongest response to drugs? Are these events coordinated between cell types? How does disease and drug effect this coordination? To address these questions, we analyzed single-nucleus-RNAseq (sn-RNAseq) data from the human anterior cingulate cortex-a region involved in many psychiatric disorders. Density index, a metric for quantifying similarities and dissimilarities across functional profiles, was employed to identify common or salient functional themes across cell types. Cell-specific signatures were integrated with existing disease and drug-specific signatures to determine cell-type-specific vulnerabilities, druggabilities, and responsiveness. Clustering of functional profiles revealed cell types jointly participating in these events. SST and VIP interneurons were found to be most vulnerable, whereas pyramidal neurons were least. Overall, the disease state is superficial layer-centric, influences cell-specific salient themes, strongly impacts disinhibitory neurons, and influences astrocyte interaction with a subset of deep-layer pyramidal neurons. In absence of disease, drugs profiles largely recapitulate disease profiles, offering a possible explanation for drug side effects. However, in presence of disease, drug activities, are deep layer-centric and involve activating a distinct subset of deep-layer pyramidal neurons to circumvent the disease state's disinhibitory circuit malfunction. These findings demonstrate a novel application of sn-RNAseq data to explain drug and disease action at a systems level.