Predicting the effect of antidepressant treatment on relief from anxiety symptoms

Introduction

Depression and anxiety disorders are among the most prevalent forms of mental illness, with antidepressants frequently used to treat them. Unfortunately, prescription of antidepressant medication is often inexact and relies on a long trial-and-error process.

Objectives

Using machine Learning (ML) algorithms on readily obtainable clinical and demographic data of individuals diagnosed with depression with anxiety symptoms, we hypothesized that we will be able to derive models which will enable a more accurate treatment selection, focusing on relief from anxiety symptoms.

Methods

Patients’ data from the Sequenced Treatment Alternatives to Relieve Depression (START*D) were filtered to include only those who have considerable anxiety symptoms. We then analyzed these patients’ response patterns, focusing on their anxious symptomology. Then, feature selection algorithms were applied to select the most predictive features for anxiety relief. Finally, we trained three ML models for three antidepressants: citalopram, sertraline and venlafaxine, using a training set of participants, and validated them on naïve validation and test datasets. These ML models were then compiled to create a predictive algorithm.

Results

Validating the algorithm on the validation and test sets, our algorithm achieved a balanced accuracy of 64.8% (p<0.001), 79.2% (p<0.001) and 78.03% (p<0.001) for citalopram, sertraline and venlafaxine, respectively.

Conclusions

Our findings support applying ML to accumulating data to achieve an improvement in the treatment of mood disorders. The algorithm we developed may be used as a tool to aid in the choice of antidepressant medication, specifically for depressed patients who exhibit prominent anxiety symptoms.

Disclosure

Dekel Taliaz is the founder and CEO of Taliaz and reports stock ownership in Taliaz. Amit Spinrad and Roni Zoller serve as data scientists in Taliaz.

Optimizing prediction of response to antidepressant medications using machine learning and environmental data

Introduction

Major depressive disorder (MDD) is complex and multifactorial, posing a major challenge of tailoring the optimal medication for each patient. Current practice for MDD treatment mainly relies on trial-and-error, with estimated 42%-53% response rates for antidepressant use.

Objectives

We sought to generate an accurate predictor of response to a panel of antidepressants and optimize treatment selection using a data-driven approach analyzing combinations of clinical and demographic factors.

Methods

We analyzed the response patterns of patients to five antidepressant medications in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study and the Pharmacogenomic Research Network Antidepressant Medication Pharmacogenomic Study (PGRN-AMPS), and employed state-of-the-art machine learning (ML) tools to generate a predictive algorithm. To validate our results and confirm the algorithm’s external generalizability outside of its training groups, we assessed its capacity to predict individualized antidepressant responses on a separate validation and test sets consisting of 1,021 patients overall from both studies.

Results

The algorithm’s ML prediction models achieved an average accuracy of 0.6416 (64.16%, SD 4.4) across the analyzed medications, and a cumulative accuracy of 0.6012 (60.12%), AUC of 0.601, sensitivity of 0.6034 (60.34%) and specificity of 0.599 (59.9%).

Conclusions

These findings support applying ML to accumulating data derived from large studies to achieve a much-needed improvement in the treatment of depression. By an immediate analysis of large amount of combinatorial data at the point of care, such prediction models may support doctors’ prescription decisions, potentially allowing them to tailor the right antidepressant medication sooner.

Disclosure

Dekel Taliaz is the founder and CEO of Taliaz and reports stock ownership in Taliaz. Amit Spinrad and Sne Darki-Morag serve as data scientists in Taliaz.

Uncovering deeply conserved motif combinations in rapidly evolving noncoding sequences

BACKGROUND

Animal genomes contain thousands of long noncoding RNA (lncRNA) genes, a growing subset of which are thought to be functionally important. This functionality is often mediated by short sequence elements scattered throughout the RNA sequence that correspond to binding sites for small RNAs and RNA binding proteins. Throughout vertebrate evolution, the sequences of lncRNA genes changed extensively, so that it is often impossible to obtain significant alignments between sequences of lncRNAs from evolutionary distant species, even when synteny is evident. This often prohibits identifying conserved lncRNAs that are likely to be functional or prioritizing constrained regions for experimental interrogation.

RESULTS

We introduce here LncLOOM, a novel algorithmic framework for the discovery and evaluation of syntenic combinations of short motifs. LncLOOM is based on a graph representation of the input sequences and uses integer linear programming to efficiently compare dozens of sequences that have thousands of bases each and to evaluate the significance of the recovered motifs. We show that LncLOOM is capable of identifying specific, biologically relevant motifs which are conserved throughout vertebrates and beyond in lncRNAs and 3'UTRs, including novel functional RNA elements in the CHASERR lncRNA that are required for regulation of CHD2 expression.

CONCLUSIONS

We expect that LncLOOM will become a broadly used approach for the discovery of functionally relevant elements in the noncoding genome.

Maternal Type-I interferon signaling adversely affects the microglia and the behavior of the offspring accompanied by increased sensitivity to stress

Viral infection during pregnancy is often associated with neuropsychiatric conditions. In mice, exposure of pregnant dams to the viral mimetic poly(I:C), serves as a model that simulates such pathology in the offspring, through a process known as Maternal Immune Activation (MIA). To investigate the mechanism of such effect, we hypothesized that maternal upregulation of Type-I interferon (IFN-I), as part of the dam's antiviral response, might contribute to the damage imposed on the offspring. Using mRNA sequencing and flow cytometry analyses we found that poly(I:C) treatment during pregnancy caused reduced expression of genes related to proliferation and cell cycle in the offspring's microglia relative to controls. This was found to be associated with an IFN-I signature in the embryonic yolk sac, the origin of microglia in development. Neutralizing IFN-I signaling in dams attenuated the effect of MIA on the newborn's microglia, while systemic maternal administration of IFNβ was sufficient to mimic the effect of poly(I:C), and led to increased vulnerability of offspring's microglia to subsequent stress. Furthermore, maternal elevation of IFNβ resulted in behavioral manifestations reminiscent of neuropsychiatric disorders. In addition, by adopting a "two-hit" experimental paradigm, we show a higher sensitivity of the offspring to postnatal stress subsequent to the maternal IFNβ elevation, demonstrated by behavioral irregularities. Our results suggest that maternal upregulation of IFN-I, in response to MIA, interferes with the offspring's programmed microglial developmental cascade, increases their susceptibility to postnatal stress, and leads to behavioral abnormalities.

A Unique Microglia Type Associated with Restricting Development of Alzheimer's Disease

Alzheimer's disease (AD) is a detrimental neurodegenerative disease with no effective treatments. Due to cellular heterogeneity, defining the roles of immune cell subsets in AD onset and progression has been challenging. Using transcriptional single-cell sorting, we comprehensively map all immune populations in wild-type and AD-transgenic (Tg-AD) mouse brains. We describe a novel microglia type associated with neurodegenerative diseases (DAM) and identify markers, spatial localization, and pathways associated with these cells. Immunohistochemical staining of mice and human brain slices shows DAM with intracellular/phagocytic Aβ particles. Single-cell analysis of DAM in Tg-AD and triggering receptor expressed on myeloid cells 2 (Trem2)^-/- Tg-AD reveals that the DAM program is activated in a two-step process. Activation is initiated in a Trem2-independent manner that involves downregulation of microglia checkpoints, followed by activation of a Trem2-dependent program. This unique microglia-type has the potential to restrict neurodegeneration, which may have important implications for future treatment of AD and other neurodegenerative diseases. VIDEO ABSTRACT.

Microglia development follows a stepwise program to regulate brain homeostasis

Microglia, the resident myeloid cells of the central nervous system, play important roles in life-long brain maintenance and in pathology. Despite their importance, their regulatory dynamics during brain development have not been fully elucidated. Using genome-wide chromatin and expression profiling coupled with single-cell transcriptomic analysis throughout development, we found that microglia undergo three temporal stages of development in synchrony with the brain--early, pre-, and adult microglia--which are under distinct regulatory circuits. Knockout of the gene encoding the adult microglia transcription factor MAFB and environmental perturbations, such as those affecting the microbiome or prenatal immune activation, led to disruption of developmental genes and immune response pathways. Together, our work identifies a stepwise microglia developmental program integrating immune response pathways that may be associated with several neurodevelopmental disorders.

Breaking immune tolerance by targeting Foxp3(+) regulatory T cells mitigates Alzheimer's disease pathology

Alzheimer's disease (AD) is a neurodegenerative disorder in which chronic neuroinflammation contributes to disease escalation. Nevertheless, while immunosuppressive drugs have repeatedly failed in treating this disease, recruitment of myeloid cells to the CNS was shown to play a reparative role in animal models. Here we show, using the 5XFAD AD mouse model, that transient depletion of Foxp3(+) regulatory T cells (Tregs), or pharmacological inhibition of their activity, is followed by amyloid-β plaque clearance, mitigation of the neuroinflammatory response and reversal of cognitive decline. We further show that transient Treg depletion affects the brain's choroid plexus, a selective gateway for immune cell trafficking to the CNS, and is associated with subsequent recruitment of immunoregulatory cells, including monocyte-derived macrophages and Tregs, to cerebral sites of plaque pathology. Our findings suggest targeting Treg-mediated systemic immunosuppression for treating AD.