Samuli Ripatti and colleagues report the results of a genome-wide association study for circulating lipid levels based on 1000 Genomes Project imputation. Their results implicate several new loci, refine the association signals at many established loci and highlight the impact of low-frequency variants on lipid traits. Using a genome-wide screen of 9.6 million genetic variants achieved through 1000 Genomes Project imputation in 62,166 samples, we identify association to lipid traits in 93 loci, including 79 previously identified loci with new lead SNPs and 10 new loci, 15 loci with a low-frequency lead SNP and 10 loci with a missense lead SNP, and 2 loci with an accumulation of rare variants. In six loci, SNPs with established function in lipid genetics (CELSR2, GCKR, LIPC and APOE) or candidate missense mutations with predicted damaging function (CD300LG and TM6SF2) explained the locus associations. The low-frequency variants increased the proportion of variance explained, particularly for low-density lipoprotein cholesterol and total cholesterol. Altogether, our results highlight the impact of low-frequency variants in complex traits and show that imputation offers a cost-effective alternative to resequencing.

Due to relatively high costs and labor required for experimental profiling of the full target space of chemical compounds, various machine learning models have been proposed as cost-effective means to advance this process in terms of predicting the most potent compound-target interactions for subsequent verification. However, most of the model predictions lack direct experimental validation in the laboratory, making their practical benefits for drug discovery or repurposing applications largely unknown. Here, we therefore introduce and carefully test a systematic computational-experimental framework for the prediction and pre-clinical verification of drug-target interactions using a well-established kernel-based regression algorithm as the prediction model. To evaluate its performance, we first predicted unmeasured binding affinities in a large-scale kinase inhibitor profiling study, and then experimentally tested 100 compound-kinase pairs. The relatively high correlation of 0.77 (p < 0.0001) between the predicted and measured bioactivities supports the potential of the model for filling the experimental gaps in existing compound-target interaction maps. Further, we subjected the model to a more challenging task of predicting target interactions for such a new candidate drug compound that lacks prior binding profile information. As a specific case study, we used tivozanib, an investigational VEGF receptor inhibitor with currently unknown off-target profile. Among 7 kinases with high predicted affinity, we experimentally validated 4 new off-targets of tivozanib, namely the Src-family kinases FRK and FYN A, the non-receptor tyrosine kinase ABL1, and the serine/threonine kinase SLK. Our sub-sequent experimental validation protocol effectively avoids any possible information leakage between the training and validation data, and therefore enables rigorous model validation for practical applications. These results demonstrate that the kernel-based modeling approach offers practical benefits for probing novel insights into the mode of action of investigational compounds, and for the identification of new target selectivities for drug repurposing applications. Significant efforts have been devoted in recent years to the development of machine learning models to support different stages of drug development process. Given the enormous size of the chemical universe, such models could offer a complementary and cost-effective means to experimental determination of drug-target interactions, toward prioritization of the most potent ones for further verification in the laboratory. In order to demonstrate the benefits of the prediction models in practical application cases, we carefully evaluated the predictive power of a well-established machine learning model in filling the gaps in existing profiling studies and prediction of target interactions for a new drug candidate. As a specific case study, we focused on kinase inhibitors, which form the largest class of new drugs approved for cancer treatment, but are also known to have wide multi-target activities contributing both to their therapeutic and toxic responses. The high agreement observed between the predicted and experimentally-measured drug-target bioactivities under the implemented rigorous validation setup demonstrates the potential of the machine learning approach, not only for filling the gaps in existing drug-target interaction maps, but also toward off-target interaction prediction for investigational drugs, and finding potential new uses for already approved drugs (drug repurposing).

The molecular pathways that drive cancer progression and treatment resistance are highly redundant and variable between individual patients with the same cancer type. To tackle this complex rewiring of pathway cross-talk, personalized combination treatments targeting multiple cancer growth and survival pathways are required. Here we implemented a computational–experimental drug combination prediction and testing (DCPT) platform for efficient in silico prioritization and ex vivo testing in patient-derived samples to identify customized synergistic combinations for individual cancer patients. DCPT used drug–target interaction networks to traverse the massive combinatorial search spaces among 218 compounds (a total of 23,653 pairwise combinations) and identified cancer-selective synergies by using differential single-compound sensitivity profiles between patient cells and healthy controls, hence reducing the likelihood of toxic combination effects. A polypharmacology-based machine learning modeling and network visualization made use of baseline genomic and molecular profiles to guide patient-specific combination testing and clinical translation phases. Using T-cell prolymphocytic leukemia (T-PLL) as a first case study, we show how the DCPT platform successfully predicted distinct synergistic combinations for each of the three T-PLL patients, each presenting with different resistance patterns and synergy mechanisms. In total, 10 of 24 (42%) of selective combination predictions were experimentally confirmed to show synergy in patient-derived samples ex vivo. The identified selective synergies among approved drugs, including tacrolimus and temsirolimus combined with BCL-2 inhibitor venetoclax, may offer novel drug repurposing opportunities for treating T-PLL. Significance: An integrated use of functional drug screening combined with genomic and molecular profiling enables patient-customized prediction and testing of drug combination synergies for T-PLL patients. Cancer Res; 78(9); 2407–18. ©2018 AACR.

Drug combinations are becoming a standard treatment of many complex diseases due to their capability to overcome resistance to monotherapy. In the current preclinical drug combination screening, the top combinations for further study are often selected based on synergy alone, without considering the combination efficacy and toxicity effects, even though these are critical determinants for the clinical success of a therapy. To promote the prioritization of drug combinations based on integrated analysis of synergy, efficacy and toxicity profiles, we implemented a web-based open-source tool, SynToxProfiler (Synergy-Toxicity-Profiler). When applied to 20 anti-cancer drug combinations tested both in healthy control and T-cell prolymphocytic leukemia (T-PLL) patient cells, as well as to 77 anti-viral drug pairs tested in Huh7 liver cell line with and without Ebola virus infection, SynToxProfiler prioritized as top hits those synergistic drug pairs that showed higher selective efficacy (difference between efficacy and toxicity), which offers an improved likelihood for clinical success. High-throughput combinatorial screening is an established approach to identify candidate drug combinations to be further developed as safe and effective treatment options for many diseases, such as various types of cancers, bacterial, malarial, and viral infections. The selection of top performing drug combinations for further development is an important step for the success of the screen, where not only the synergy but also selective efficacy and potential toxicity of the drug pairs should be critically assessed. Currently, there is no method available for this; therefore, we developed SynToxProfiler tool, which was demonstrated in two different application cases to prioritize synergistic drug pairs with higher efficacy and lower toxicity as top hits, providing thus an increased likelihood for their clinical success.

T-cell prolymphocytic leukemia (T-PLL) is a rare and poor-prognostic mature T-cell leukemia. Recent studies detected genomic aberrations affecting JAK and STAT genes in T-PLL. Due to the limited number of primary patient samples available, genomic analyses of the JAK/STAT pathway have been performed in rather small cohorts. Therefore, we conducted—via a primary-data based pipeline—a meta-analysis that re-evaluated the genomic landscape of T-PLL. It included all available data sets with sequence information on JAK or STAT gene loci in 275 T-PLL. We eliminated overlapping cases and determined a cumulative rate of 62.1% of cases with mutated JAK or STAT genes. Most frequently, JAK1 (6.3%), JAK3 (36.4%), and STAT5B (18.8%) carried somatic single-nucleotide variants (SNVs), with missense mutations in the SH2 or pseudokinase domains as most prevalent. Importantly, these lesions were predominantly subclonal. We did not detect any strong association between mutations of a JAK or STAT gene with clinical characteristics. Irrespective of the presence of gain-of-function (GOF) SNVs, basal phosphorylation of STAT5B was elevated in all analyzed T-PLL. Fittingly, a significant proportion of genes encoding for potential negative regulators of STAT5B showed genomic losses (in 71.4% of T-PLL in total, in 68.4% of T-PLL without any JAK or STAT mutations). They included DUSP4, CD45, TCPTP, SHP1, SOCS1, SOCS3, and HDAC9. Overall, considering such losses of negative regulators and the GOF mutations in JAK and STAT genes, a total of 89.8% of T-PLL revealed a genomic aberration potentially explaining enhanced STAT5B activity. In essence, we present a comprehensive meta-analysis on the highly prevalent genomic lesions that affect genes encoding JAK/STAT signaling components. This provides an overview of possible modes of activation of this pathway in a large cohort of T-PLL. In light of new advances in JAK/STAT inhibitor development, we also outline translational contexts for harnessing active JAK/STAT signaling, which has emerged as a ‘secondary’ hallmark of T-PLL.

Cancer cells with heterogeneous mutation landscapes and extensive functional redundancy easily develop resistance to monotherapies by emerging activation of compensating or bypassing pathways. To achieve more effective and sustained clinical responses, synergistic interactions of multiple druggable targets that inhibit redundant cancer survival pathways are often required. Here, we report a systematic polypharmacology strategy to predict, test, and understand the selective drug combinations for MDA-MB-231 triple-negative breast cancer cells. We started by applying our network pharmacology model to predict synergistic drug combinations. Next, by utilizing kinome-wide drug-target profiles and gene expression data, we pinpointed a synergistic target interaction between Aurora B and ZAK kinase inhibition that led to enhanced growth inhibition and cytotoxicity, as validated by combinatorial siRNA, CRISPR/Cas9, and drug combination experiments. The mechanism of such a context-specific target interaction was elucidated using a dynamic simulation of MDA-MB-231 signaling network, suggesting a cross-talk between p53 and p38 pathways. Our results demonstrate the potential of polypharmacological modeling to systematically interrogate target interactions that may lead to clinically actionable and personalized treatment options.