Recognition and reprogramming of E3 ubiquitin ligase surfaces by α-helical peptides.

Molecules that induce novel interactions between proteins hold great promise for the study of biological systems and the development of therapeutics, but their discovery has been limited by the complexities of rationally designing interactions between three components, and because known binders to each protein are typically required to inform initial designs. Here, we report a general and rapid method for discovering α-helically constrained (Helicon) polypeptides that cooperatively induce the interaction between two target proteins without relying on previously known binders or an intrinsic affinity between the proteins. We show that Helicons are capable of binding every major class of E3 ubiquitin ligases, which are of great biological and therapeutic interest but remain largely intractable to targeting by small molecules. We then describe a phage-based screening method for discovering "trimerizer" Helicons, and apply it to reprogram E3s to cooperatively bind an enzyme (PPIA), a transcription factor (TEAD4), and a transcriptional coactivator (ß-catenin).

De novo mapping of α-helix recognition sites on protein surfaces using unbiased libraries.

The α-helix is one of the most common protein surface recognition motifs found in nature, and its unique amide-cloaking properties also enable α-helical polypeptide motifs to exist in membranes. Together, these properties have inspired the development of α-helically constrained (Helicon) therapeutics that can enter cells and bind targets that have been considered "undruggable", such as protein-protein interactions. To date, no general method for discovering α-helical binders to proteins has been reported, limiting Helicon drug discovery to only those proteins with previously characterized α-helix recognition sites, and restricting the starting chemical matter to those known α-helical binders. Here, we report a general and rapid screening method to empirically map the α-helix binding sites on a broad range of target proteins in parallel using large, unbiased Helicon phage display libraries and next-generation sequencing. We apply this method to screen six structurally diverse protein domains, only one of which had been previously reported to bind isolated α-helical peptides, discovering 20 families that collectively comprise several hundred individual Helicons. Analysis of 14 X-ray cocrystal structures reveals at least nine distinct α-helix recognition sites across these six proteins, and biochemical and biophysical studies show that these Helicons can block protein-protein interactions, inhibit enzymatic activity, induce conformational rearrangements, and cause protein dimerization. We anticipate that this method will prove broadly useful for the study of protein recognition and for the development of both biochemical tools and therapeutics for traditionally challenging protein targets.

Mechanisms of Pinometostat (EPZ-5676) Treatment-Emergent Resistance in MLL-Rearranged Leukemia.

DOT1L is a protein methyltransferase involved in the development and maintenance of MLL-rearranged (MLL-r) leukemia through its ectopic methylation of histones associated with well-characterized leukemic genes. Pinometostat (EPZ-5676), a selective inhibitor of DOT1L, is in clinical development in relapsed/refractory acute leukemia patients harboring rearrangements of the MLL gene. The observation of responses and subsequent relapses in the adult trial treating MLL-r patients motivated preclinical investigations into potential mechanisms of pinometostat treatment-emergent resistance (TER) in cell lines confirmed to have MLL-r. TER was achieved in five MLL-r cell lines, KOPN-8, MOLM-13, MV4-11, NOMO-1, and SEM. Two of the cell lines, KOPN-8 and NOMO-1, were thoroughly characterized to understand the mechanisms involved in pinometostat resistance. Unlike many other targeted therapies, resistance does not appear to be achieved through drug-induced selection of mutations of the target itself. Instead, we identified both drug efflux transporter dependent and independent mechanisms of resistance to pinometostat. In KOPN-8 TER cells, increased expression of the drug efflux transporter ABCB1 (P-glycoprotein, MDR1) was the primary mechanism of drug resistance. In contrast, resistance in NOMO-1 cells occurs through a mechanism other than upregulation of a specific efflux pump. RNA-seq analysis performed on both parental and resistant KOPN-8 and NOMO-1 cell lines supported two unique candidate pathway mechanisms that may explain the pinometostat resistance observed in these cell lines. These results are the first demonstration of TER models of the DOT1L inhibitor pinometostat and may provide useful tools for investigating clinical resistance. Mol Cancer Ther; 16(8); 1669-79. ©2017 AACR.

Blood-based identification of non-responders to anti-TNF therapy in rheumatoid arthritis.

BACKGROUND: Faced with an increasing number of choices for biologic therapies, rheumatologists have a critical need for better tools to inform rheumatoid arthritis (RA) disease management. The ability to identify patients who are unlikely to respond to first-line biologic anti-TNF therapies prior to their treatment would allow these patients to seek alternative therapies, providing faster relief and avoiding complications of disease. METHODS: We identified a gene expression classifier to predict, pre-treatment, which RA patients are unlikely to respond to the anti-TNF infliximab. The classifier was trained and independently evaluated using four published whole blood gene expression data sets, in which RA patients (n = 116 = 44 + 15 + 30 + 27) were treated with infliximab, and their response assessed 14-16 months post treatment according to the European League Against Rheumatism (EULAR) response criteria. For each patient, prior knowledge was used to group gene expression measurements into disease-relevant biological signaling mechanisms that were used as the input features for regularized logistic regression. RESULTS: The classifier produced a substantial enrichment of non-responders (59 %, given by the cross validated test precision) compared to the full population (27 % non-responders), while identifying nearly a third of non-responders. Given this classifier performance, treatment of predicted non-responders with alternative biologics would decrease their chance of non-response by between a third and a half, substantially improving their odds of effective treatment and stemming further disease progression. The classifier consisted of 18 signaling mechanisms, which together indicated that higher inflammatory signaling mediated by TNF and other cytokines was present pre-treatment in the blood of patients who responded to infliximab treatment. In contrast, non-responders were classified by relatively higher levels of specific metabolic activities in the blood prior to treatment. CONCLUSIONS: We were able to successfully produce a classifier to identify a population of RA patients significantly enriched in anti-TNF non-responders across four different patient cohorts. Additional prospective studies are needed to validate and refine the classifier for clinical use.

An algorithm for score aggregation over causal biological networks based on random walk sampling.

BACKGROUND: We recently published in BMC Systems Biology an approach for calculating the perturbation amplitudes of causal network models by integrating gene differential expression data. This approach relies on the process of score aggregation, which combines the perturbations at the level of the individual network nodes into a global measure that quantifies the perturbation of the network as a whole. Such "bottom-up" aggregation relates the changes in molecular entities measured by omics technologies to systems-level phenotypes. However, the aggregation method we used is limited to a specific class of causal network models called "causally consistent", which is equivalent to the notion of balance of a signed graph used in graph theory. As a consequence of this limitation, our aggregation method cannot be used in the many relevant cases involving "causally inconsistent" network models such as those containing negative feedbacks. FINDINGS: In this note, we propose an algorithm called "sampling of spanning trees" (SST) that extends our published aggregation method to causally inconsistent network models by replacing the signed relationships between the network nodes by an appropriate continuous measure. The SST algorithm is based on spanning trees, which are a particular class of subgraphs used in graph theory, and on a sampling procedure leveraging the properties of specific random walks on the graph. This algorithm is applied to several cases of biological interest. CONCLUSIONS: The SST algorithm provides a practical means of aggregating nodal values over causally inconsistent network models based on solid mathematical foundations. We showed its utility in systems biology, where the nodal values can be perturbation amplitudes of protein activities or gene differential expressions, while the networks can be models of cellular signaling or expression regulation. Since the SST algorithm is based on general graph-theoretical considerations, it is scalable to arbitrary graph sizes and can potentially be used for performing quantitative analyses in any context involving signed graphs.

Quantitative assessment of biological impact using transcriptomic data and mechanistic network models.

Exposure to biologically active substances such as therapeutic drugs or environmental toxicants can impact biological systems at various levels, affecting individual molecules, signaling pathways, and overall cellular processes. The ability to derive mechanistic insights from the resulting system responses requires the integration of experimental measures with a priori knowledge about the system and the interacting molecules therein. We developed a novel systems biology-based methodology that leverages mechanistic network models and transcriptomic data to quantitatively assess the biological impact of exposures to active substances. Hierarchically organized network models were first constructed to provide a coherent framework for investigating the impact of exposures at the molecular, pathway and process levels. We then validated our methodology using novel and previously published experiments. For both in vitro systems with simple exposure and in vivo systems with complex exposures, our methodology was able to recapitulate known biological responses matching expected or measured phenotypes. In addition, the quantitative results were in agreement with experimental endpoint data for many of the mechanistic effects that were assessed, providing further objective confirmation of the approach. We conclude that our methodology evaluates the biological impact of exposures in an objective, systematic, and quantifiable manner, enabling the computation of a systems-wide and pan-mechanistic biological impact measure for a given active substance or mixture. Our results suggest that various fields of human disease research, from drug development to consumer product testing and environmental impact analysis, could benefit from using this methodology.

Assessment of network perturbation amplitudes by applying high-throughput data to causal biological networks.

BACKGROUND: High-throughput measurement technologies produce data sets that have the potential to elucidate the biological impact of disease, drug treatment, and environmental agents on humans. The scientific community faces an ongoing challenge in the analysis of these rich data sources to more accurately characterize biological processes that have been perturbed at the mechanistic level. Here, a new approach is built on previous methodologies in which high-throughput data was interpreted using prior biological knowledge of cause and effect relationships. These relationships are structured into network models that describe specific biological processes, such as inflammatory signaling or cell cycle progression. This enables quantitative assessment of network perturbation in response to a given stimulus. RESULTS: Four complementary methods were devised to quantify treatment-induced activity changes in processes described by network models. In addition, companion statistics were developed to qualify significance and specificity of the results. This approach is called Network Perturbation Amplitude (NPA) scoring because the amplitudes of treatment-induced perturbations are computed for biological network models. The NPA methods were tested on two transcriptomic data sets: normal human bronchial epithelial (NHBE) cells treated with the pro-inflammatory signaling mediator TNFα, and HCT116 colon cancer cells treated with the CDK cell cycle inhibitor R547. Each data set was scored against network models representing different aspects of inflammatory signaling and cell cycle progression, and these scores were compared with independent measures of pathway activity in NHBE cells to verify the approach. The NPA scoring method successfully quantified the amplitude of TNFα-induced perturbation for each network model when compared against NF-κB nuclear localization and cell number. In addition, the degree and specificity to which CDK-inhibition affected cell cycle and inflammatory signaling were meaningfully determined. CONCLUSIONS: The NPA scoring method leverages high-throughput measurements and a priori literature-derived knowledge in the form of network models to characterize the activity change for a broad collection of biological processes at high-resolution. Applications of this framework include comparative assessment of the biological impact caused by environmental factors, toxic substances, or drug treatments.

Scaffold number in yeast signaling system sets tradeoff between system output and dynamic range.

Although the proteins comprising many signaling systems are known, less is known about their numbers per cell. Existing measurements often vary by more than 10-fold. Here, we devised improved quantification methods to measure protein abundances in the Saccharomyces cerevisiae pheromone response pathway, an archetypical signaling system. These methods limited variation between independent measurements of protein abundance to a factor of two. We used these measurements together with quantitative models to identify and investigate behaviors of the pheromone response system sensitive to precise abundances. The difference between the maximum and basal signaling output (dynamic range) of the pheromone response MAPK cascade was strongly sensitive to the abundance of Ste5, the MAPK scaffold protein, and absolute system output depended on the amount of Fus3, the MAPK. Additional analysis and experiment suggest that scaffold abundance sets a tradeoff between maximum system output and system dynamic range, a prediction supported by recent experiments.