Systematically benchmarking peptide-MHC binding predictors: From synthetic to naturally processed epitopes.

A number of machine learning-based predictors have been developed for identifying immunogenic T-cell epitopes based on major histocompatibility complex (MHC) class I and II binding affinities. Rationally selecting the most appropriate tool has been complicated by the evolving training data and machine learning methods. Despite the recent advances made in generating high-quality MHC-eluted, naturally processed ligandome, the reliability of new predictors on these epitopes has yet to be evaluated. This study reports the latest benchmarking on an extensive set of MHC-binding predictors by using newly available, untested data of both synthetic and naturally processed epitopes. 32 human leukocyte antigen (HLA) class I and 24 HLA class II alleles are included in the blind test set. Artificial neural network (ANN)-based approaches demonstrated better performance than regression-based machine learning and structural modeling. Among the 18 predictors benchmarked, ANN-based mhcflurry and nn_align perform the best for MHC class I 9-mer and class II 15-mer predictions, respectively, on binding/non-binding classification (Area Under Curves = 0.911). NetMHCpan4 also demonstrated comparable predictive power. Our customization of mhcflurry to a pan-HLA predictor has achieved similar accuracy to NetMHCpan. The overall accuracy of these methods are comparable between 9-mer and 10-mer testing data. However, the top methods deliver low correlations between the predicted versus the experimental affinities for strong MHC binders. When used on naturally processed MHC-ligands, tools that have been trained on elution data (NetMHCpan4 and MixMHCpred) shows better accuracy than pure binding affinity predictor. The variability of false prediction rate is considerable among HLA types and datasets. Finally, structure-based predictor of Rosetta FlexPepDock is less optimal compared to the machine learning approaches. With our benchmarking of MHC-binding and MHC-elution predictors using a comprehensive metrics, a unbiased view for establishing best practice of T-cell epitope predictions is presented, facilitating future development of methods in immunogenomics.

Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy.

Programmed cell death protein-1 (PD-1) and programmed cell death ligand-1 (PD-L1) checkpoint blockade immunotherapy elicits durable antitumor effects in multiple cancers, yet not all patients respond. We report the evaluation of >300 patient samples across 22 tumor types from four KEYNOTE clinical trials. Tumor mutational burden (TMB) and a T cell-inflamed gene expression profile (GEP) exhibited joint predictive utility in identifying responders and nonresponders to the PD-1 antibody pembrolizumab. TMB and GEP were independently predictive of response and demonstrated low correlation, suggesting that they capture distinct features of neoantigenicity and T cell activation. Analysis of The Cancer Genome Atlas database showed TMB and GEP to have a low correlation, and analysis by joint stratification revealed biomarker-defined patterns of targetable-resistance biology. These biomarkers may have utility in clinical trial design by guiding rational selection of anti-PD-1 monotherapy and combination immunotherapy regimens.

Comprehensive molecular characterization of clinical responses to PD-1 inhibition in metastatic gastric cancer.

Clinical studies support the efficacy of programmed cell death 1 (PD-1) targeted therapy in a subset of patients with metastatic gastric cancer (mGC). With the goal of identifying determinants of response, we performed molecular characterization of tissues and circulating tumor DNA (ctDNA) from 61 patients with mGC who were treated with pembrolizumab as salvage treatment in a prospective phase 2 clinical trial. In patients with microsatellite instability-high and Epstein-Barr virus-positive tumors, which are mutually exclusive, dramatic responses to pembrolizumab were observed (overall response rate (ORR) 85.7% in microsatellite instability-high mGC and ORR 100% in Epstein-Barr virus-positive mGC). For the 55 patients for whom programmed death-ligand 1 (PD-L1) combined positive score positivity was available (combined positive score cut-off value ≥1%), ORR was significantly higher in PD-L1(+) gastric cancer when compared to PD-L1(-) tumors (50.0% versus 0.0%, P value <0.001). Changes in ctDNA levels at six weeks post-treatment predicted response and progression-free survival, and decreased ctDNA was associated with improved outcomes. Our findings provide insight into the molecular features associated with response to pembrolizumab in patients with mGC and provide biomarkers potentially relevant for the selection of patients who may derive greater benefit from PD-1 inhibition.

Differential daptomycin resistance development in Staphylococcus aureus strains with active and mutated gra regulatory systems.

The first-in-class lipopeptide antibiotic daptomycin (DAP) is highly active against Gram-positive pathogens including ß-lactam and glycopeptide resistant strains. Its molecular mode of action remains enigmatic, since a defined target has not been identified so far and multiple effects, primarily on the cell envelope have been observed. Reduced DAP susceptibility has been described in S. aureus and enterococci after prolonged treatment courses. In line with its pleiotropic antibiotic activities, a unique, defined molecular mechanism of resistance has not emerged, instead non-susceptibility appears often accompanied by alterations in membrane composition and changes in cell wall homeostasis. We compared S. aureus strains HG001 and SG511, which differ primarily in the functionality of the histidine kinase GraS, to evaluate the impact of the GraRS regulatory system on the development of DAP non-susceptibility. After extensive serial passing, both DAPR variants reached a minimal inhibitory concentration of 31 µg/ml and shared some phenotypic characteristics (e.g. thicker cell wall, reduced autolysis). However, based on comprehensive analysis of the underlying genetic, transcriptomic and proteomic changes, we found that both strains took different routes to achieve DAP resistance. Our study highlights the impressive genetic and physiological capacity of S. aureus to counteract pleiotropic activities of cell wall- and membrane-active compounds even when a major cell wall regulatory system is dysfunctional.

Antibacterial small molecules targeting the conserved TOPRIM domain of DNA gyrase.

To combat the threat of antibiotic-resistant Gram-negative bacteria, novel agents that circumvent established resistance mechanisms are urgently needed. Our approach was to focus first on identifying bioactive small molecules followed by chemical lead prioritization and target identification. Within this annotated library of bioactives, we identified a small molecule with activity against efflux-deficient Escherichia coli and other sensitized Gram-negatives. Further studies suggested that this compound inhibited DNA replication and selection for resistance identified mutations in a subunit of E. coli DNA gyrase, a type II topoisomerase. Our initial compound demonstrated weak inhibition of DNA gyrase activity while optimized compounds demonstrated significantly improved inhibition of E. coli and Pseudomonas aeruginosa DNA gyrase and caused cleaved complex stabilization, a hallmark of certain bactericidal DNA gyrase inhibitors. Amino acid substitutions conferring resistance to this new class of DNA gyrase inhibitors reside exclusively in the TOPRIM domain of GyrB and are not associated with resistance to the fluoroquinolones, suggesting a novel binding site for a gyrase inhibitor.

Dual-Targeting Small-Molecule Inhibitors of the Staphylococcus aureus FMN Riboswitch Disrupt Riboflavin Homeostasis in an Infectious Setting.

Riboswitches are bacterial-specific, broadly conserved, non-coding RNA structural elements that control gene expression of numerous metabolic pathways and transport functions essential for cell growth. As such, riboswitch inhibitors represent a new class of potential antibacterial agents. Recently, we identified ribocil-C, a highly selective inhibitor of the flavin mononucleotide (FMN) riboswitch that controls expression of de novo riboflavin (RF, vitamin B2) biosynthesis in Escherichia coli. Here, we provide a mechanistic characterization of the antibacterial effects of ribocil-C as well as of roseoflavin (RoF), an antimetabolite analog of RF, among medically significant Gram-positive bacteria, including methicillin-resistant Staphylococcus aureus (MRSA) and Enterococcus faecalis. We provide genetic, biophysical, computational, biochemical, and pharmacological evidence that ribocil-C and RoF specifically inhibit dual FMN riboswitches, separately controlling RF biosynthesis and uptake processes essential for MRSA growth and pathogenesis. Such a dual-targeting mechanism is specifically required to develop broad-spectrum Gram-positive antibacterial agents targeting RF metabolism.

Affinity Selection-Mass Spectrometry Identifies a Novel Antibacterial RNA Polymerase Inhibitor.

The growing prevalence of drug resistant bacteria is a significant global threat to human health. The antibacterial drug rifampin, which functions by inhibiting bacterial RNA polymerase (RNAP), is an important part of the antibacterial armamentarium. Here, in order to identify novel inhibitors of bacterial RNAP, we used affinity-selection mass spectrometry to screen a chemical library for compounds that bind to Escherichia coli RNAP. We identified a novel small molecule, MRL-436, that binds to RNAP, inhibits RNAP, and exhibits antibacterial activity. MRL-436 binds to RNAP through a binding site that differs from the rifampin binding site, inhibits rifampin-resistant RNAP derivatives, and exhibits antibacterial activity against rifampin-resistant strains. Isolation of mutants resistant to the antibacterial activity of MRL-436 yields a missense mutation in codon 622 of the rpoC gene encoding the RNAP ß' subunit or a null mutation in the rpoZ gene encoding the RNAP ω subunit, confirming that RNAP is the functional cellular target for the antibacterial activity of MRL-436, and indicating that RNAP ß' subunit residue 622 and the RNAP ω subunit are required for the antibacterial activity of MRL-436. Similarity between the resistance determinant for MRL-436 and the resistance determinant for the cellular alarmone ppGpp suggests a possible similarity in binding site and/or induced conformational state for MRL-436 and ppGpp.

Epitopes and Mechanism of Action of the Clostridium difficile Toxin A-Neutralizing Antibody Actoxumab.

The exotoxins toxin A (TcdA) and toxin B (TcdB) are produced by the bacterial pathogen Clostridium difficile and are responsible for the pathology associated with C. difficile infection (CDI). The antitoxin antibodies actoxumab and bezlotoxumab bind to and neutralize TcdA and TcdB, respectively. Bezlotoxumab was recently approved by the FDA for reducing the recurrence of CDI. We have previously shown that a single molecule of bezlotoxumab binds to two distinct epitopes within the TcdB combined repetitive oligopeptide (CROP) domain, preventing toxin binding to host cells. In this study, we characterize the binding of actoxumab to TcdA and examine its mechanism of toxin neutralization. Using a combination of approaches including a number of biophysical techniques, we show that there are two distinct actoxumab binding sites within the CROP domain of TcdA centered on identical amino acid sequences at residues 2162-2189 and 2410-2437. Actoxumab binding caused the aggregation of TcdA especially at higher antibody:toxin concentration ratios. Actoxumab prevented the association of TcdA with target cells demonstrating that actoxumab neutralizes toxin activity by inhibiting the first step of the intoxication cascade. This mechanism of neutralization is similar to that observed with bezlotoxumab and TcdB. Comparisons of the putative TcdA epitope sequences across several C. difficile ribotypes and homologous repeat sequences within TcdA suggest a structural basis for observed differences in actoxumab binding and/or neutralization potency. These data provide a mechanistic basis for the protective effects of the antibody in vitro and in vivo, including in various preclinical models of CDI.

Chemical Genetic Analysis and Functional Characterization of Staphylococcal Wall Teichoic Acid 2-Epimerases Reveals Unconventional Antibiotic Drug Targets.

Here we describe a chemical biology strategy performed in Staphylococcus aureus and Staphylococcus epidermidis to identify MnaA, a 2-epimerase that we demonstrate interconverts UDP-GlcNAc and UDP-ManNAc to modulate substrate levels of TarO and TarA wall teichoic acid (WTA) biosynthesis enzymes. Genetic inactivation of mnaA results in complete loss of WTA and dramatic in vitro ß-lactam hypersensitivity in methicillin-resistant S. aureus (MRSA) and S. epidermidis (MRSE). Likewise, the ß-lactam antibiotic imipenem exhibits restored bactericidal activity against mnaA mutants in vitro and concomitant efficacy against 2-epimerase defective strains in a mouse thigh model of MRSA and MRSE infection. Interestingly, whereas MnaA serves as the sole 2-epimerase required for WTA biosynthesis in S. epidermidis, MnaA and Cap5P provide compensatory WTA functional roles in S. aureus. We also demonstrate that MnaA and other enzymes of WTA biosynthesis are required for biofilm formation in MRSA and MRSE. We further determine the 1.9Å crystal structure of S. aureus MnaA and identify critical residues for enzymatic dimerization, stability, and substrate binding. Finally, the natural product antibiotic tunicamycin is shown to physically bind MnaA and Cap5P and inhibit 2-epimerase activity, demonstrating that it inhibits a previously unanticipated step in WTA biosynthesis. In summary, MnaA serves as a new Staphylococcal antibiotic target with cognate inhibitors predicted to possess dual therapeutic benefit: as combination agents to restore ß-lactam efficacy against MRSA and MRSE and as non-bioactive prophylactic agents to prevent Staphylococcal biofilm formation.

TarO-specific inhibitors of wall teichoic acid biosynthesis restore ß-lactam efficacy against methicillin-resistant staphylococci.

The widespread emergence of methicillin-resistant Staphylococcus aureus (MRSA) has dramatically eroded the efficacy of current ß-lactam antibiotics and created an urgent need for new treatment options. We report an S. aureus phenotypic screening strategy involving chemical suppression of the growth inhibitory consequences of depleting late-stage wall teichoic acid biosynthesis. This enabled us to identify early-stage pathway-specific inhibitors of wall teichoic acid biosynthesis predicted to be chemically synergistic with ß-lactams. We demonstrated by genetic and biochemical means that each of the new chemical series discovered, herein named tarocin A and tarocin B, inhibited the first step in wall teichoic acid biosynthesis (TarO). Tarocins do not have intrinsic bioactivity but rather demonstrated potent bactericidal synergy in combination with broad-spectrum ß-lactam antibiotics against diverse clinical isolates of methicillin-resistant staphylococci as well as robust efficacy in a murine infection model of MRSA. Tarocins and other inhibitors of wall teichoic acid biosynthesis may provide a rational strategy to develop Gram-positive bactericidal ß-lactam combination agents active against methicillin-resistant staphylococci.

Staphylococcus aureus Survives with a Minimal Peptidoglycan Synthesis Machine but Sacrifices Virulence and Antibiotic Resistance.

Many important cellular processes are performed by molecular machines, composed of multiple proteins that physically interact to execute biological functions. An example is the bacterial peptidoglycan (PG) synthesis machine, responsible for the synthesis of the main component of the cell wall and the target of many contemporary antibiotics. One approach for the identification of essential components of a cellular machine involves the determination of its minimal protein composition. Staphylococcus aureus is a Gram-positive pathogen, renowned for its resistance to many commonly used antibiotics and prevalence in hospitals. Its genome encodes a low number of proteins with PG synthesis activity (9 proteins), when compared to other model organisms, and is therefore a good model for the study of a minimal PG synthesis machine. We deleted seven of the nine genes encoding PG synthesis enzymes from the S. aureus genome without affecting normal growth or cell morphology, generating a strain capable of PG biosynthesis catalyzed only by two penicillin-binding proteins, PBP1 and the bi-functional PBP2. However, multiple PBPs are important in clinically relevant environments, as bacteria with a minimal PG synthesis machinery became highly susceptible to cell wall-targeting antibiotics, host lytic enzymes and displayed impaired virulence in a Drosophila infection model which is dependent on the presence of specific peptidoglycan receptor proteins, namely PGRP-SA. The fact that S. aureus can grow and divide with only two active PG synthesizing enzymes shows that most of these enzymes are redundant in vitro and identifies the minimal PG synthesis machinery of S. aureus. However a complex molecular machine is important in environments other than in vitro growth as the expendable PG synthesis enzymes play an important role in the pathogenicity and antibiotic resistance of S. aureus.

Chemical Genomics-Based Antifungal Drug Discovery: Targeting Glycosylphosphatidylinositol (GPI) Precursor Biosynthesis.

Steadily increasing antifungal drug resistance and persistent high rates of fungal-associated mortality highlight the dire need for the development of novel antifungals. Characterization of inhibitors of one enzyme in the GPI anchor pathway, Gwt1, has generated interest in the exploration of targets in this pathway for further study. Utilizing a chemical genomics-based screening platform referred to as the Candida albicans fitness test (CaFT), we have identified novel inhibitors of Gwt1 and a second enzyme in the glycosylphosphatidylinositol (GPI) cell wall anchor pathway, Mcd4. We further validate these targets using the model fungal organism Saccharomyces cerevisiae and demonstrate the utility of using the facile toolbox that has been compiled in this species to further explore target specific biology. Using these compounds as probes, we demonstrate that inhibition of Mcd4 as well as Gwt1 blocks the growth of a broad spectrum of fungal pathogens and exposes key elicitors of pathogen recognition. Interestingly, a strong chemical synergy is also observed by combining Gwt1 and Mcd4 inhibitors, mirroring the demonstrated synthetic lethality of combining conditional mutants of GWT1 and MCD4. We further demonstrate that the Mcd4 inhibitor M720 is efficacious in a murine infection model of systemic candidiasis. Our results establish Mcd4 as a promising antifungal target and confirm the GPI cell wall anchor synthesis pathway as a promising antifungal target area by demonstrating that effects of inhibiting it are more general than previously recognized.