An improved strategy for drug screening may soon pave the way for novel medications for diseases that have previously seen little therapeutic innovation. A team of researchers, including scientists from Sema4, found that patient-derived cells offer a more effective approach for assessing drug response than conventional models, demonstrating the value of a precision medicine approach to drug screening.
The scientists from Icahn School of Medicine at Mount Sinai, Sema4, and Eli Lilly focused their study on schizophrenia, a severe neuropsychiatric disorder with a strong genetic component. Symptoms of schizophrenia can, in around one-third of individuals, be controlled with drugs that modulate dopamine activity. The remaining two-thirds of patients, however, do not respond to or have only a partial response to these medications, and drug discovery has been limited due to a lack of useful models for screening candidate treatments.
The team reasoned that, as drug response is thought to be a heritable component of schizophrenia, screening needs to take into account neuropsychiatric genetics and be carried out in a biologically-relevant model. At the outset of the study, the researchers used in silico prioritization to select drugs with predicted or demonstrated interactions with schizophrenia-related biology, allowing them to focus on a smaller, more refined set of drugs than is typically used in high-throughput screens. The resulting 135 candidate drugs were then used to treat neural progenitor cells from 12 schizophrenia patients and 12 healthy controls, along with eight generic cancer cell lines.
Following drug treatment, the scientists searched for drug-induced gene expression changes and found differential responses in schizophrenia biology-associated genes in the patient-derived cells compared with the control-derived cells and cancer cell lines. In some cases, certain drugs reversed the gene expression signatures associated with schizophrenia. The proof-of-concept study, published in Nature Communications, demonstrates that patient-derived cells yield more disease-relevant information than generic cell lines and establishes the feasibility of expression-based drug screens of patient-derived neural cells.
“This study is one of the first instances of transcriptomic drug screening, whereby we profiled the global response of neural cells following treatment,” said Dr. Kristen Brennand, Associate Professor of Neuroscience, Psychiatry, and Genetics and Genomic Sciences at Mount Sinai and senior author of the paper. “Our approach portends an entirely new way of screening drugs that doesn’t rely on synaptic assays, which are technically difficult to conduct, and demonstrates that there is tremendous value in gene expression-based drug screening using patient-derived cells because it can generate results that are more reflective of disease biology.”
Even with the initial in silico drug prioritization, the study was complex, requiring the use of an innovative high-throughput gene expression profiling technology from Genometry to optimize the number of expression signatures generated. Interpretation of the resulting massive dataset required a multinational effort, led by first author Benjamin Readhead in collaboration with Sinai colleague Gabriel Hoffman and Brian Eastwood of Eli Lilly in the United Kingdom.
“Collaborations are critical in the field of therapeutics discovery and development for the integration and holistic interpretation of increasingly granular and diverse data,” said senior author Dr. Radoslav Savić, Director of Scientific Collaborations at Sema4, Associate Professor of Genetics and Genomic Sciences at Mount Sinai and corresponding author of the paper. “This study demonstrates how unmet research needs can be addressed by connecting complementary forces, from experts based in academia, health systems, and industry, to the individuals who contribute their cells and without whom the work would not be possible.”
The results of the study show that neural cells from patients can respond differently to drugs than those from healthy controls, and should have immediate value in improving drug discovery, not only for schizophrenia but also for other diseases currently lacking biologically relevant screening models. “The next step is to expand the work into neurons, the cell type relevant to schizophrenia,” says Dr. Brennand. “It will be important to gather data such as this across more cell donors and more cell types, so we can discover the diversity of drug responses and refine our precision medicine approach further.”