Prediction of Effectiveness and Toxicities of Immune Checkpoint Inhibitors Using Real-World Patient Data.

PURPOSE: Although immune checkpoint inhibitors (ICIs) have improved outcomes in certain patients with cancer, they can also cause life-threatening immunotoxicities. Predicting immunotoxicity risks alongside response could provide a personalized risk-benefit profile, inform therapeutic decision making, and improve clinical trial cohort selection. We aimed to build a machine learning (ML) framework using routine electronic health record (EHR) data to predict hepatitis, colitis, pneumonitis, and 1-year overall survival. METHODS: Real-world EHR data of more than 2,200 patients treated with ICI through December 31, 2018, were used to develop predictive models. Using a prediction time point of ICI initiation, a 1-year prediction time window was applied to create binary labels for the four outcomes for each patient. Feature engineering involved aggregating laboratory measurements over appropriate time windows (60-365 days). Patients were randomly partitioned into training (80%) and test (20%) sets. Random forest classifiers were developed using a rigorous model development framework. RESULTS: The patient cohort had a median age of 63 years and was 61.8% male. Patients predominantly had melanoma (37.8%), lung cancer (27.3%), or genitourinary cancer (16.4%). They were treated with PD-1 (60.4%), PD-L1 (9.0%), and CTLA-4 (19.7%) ICIs. Our models demonstrate reasonably strong performance, with AUCs of 0.739, 0.729, 0.755, and 0.752 for the pneumonitis, hepatitis, colitis, and 1-year overall survival models, respectively. Each model relies on an outcome-specific feature set, though some features are shared among models. CONCLUSION: To our knowledge, this is the first ML solution that assesses individual ICI risk-benefit profiles based predominantly on routine structured EHR data. As such, use of our ML solution will not require additional data collection or documentation in the clinic.

My Cancer Genome: Coevolution of Precision Oncology and a Molecular Oncology Knowledgebase.

PURPOSE: The My Cancer Genome (MCG) knowledgebase and resulting website were launched in 2011 with the purpose of guiding clinicians in the application of genomic testing results for treatment of patients with cancer. Both knowledgebase and website were originally developed using a wiki-style approach that relied on manual evidence curation and synthesis of that evidence into cancer-related biomarker, disease, and pathway pages on the website that summarized the literature for a clinical audience. This approach required significant time investment for each page, which limited website scalability as the field advanced. To address this challenge, we designed and used an assertion-based data model that allows the knowledgebase and website to expand with the field of precision oncology. METHODS: Assertions, or computationally accessible cause and effect statements, are both manually curated from primary sources and imported from external databases and stored in a knowledge management system. To generate pages for the MCG website, reusable templates transform assertions into reconfigurable text and visualizations that form the building blocks for automatically updating disease, biomarker, drug, and clinical trial pages. RESULTS: Combining text and graph templates with assertions in our knowledgebase allows generation of web pages that automatically update with our knowledgebase. Automated page generation empowers rapid scaling of the website as assertions with new biomarkers and drugs are added to the knowledgebase. This process has generated more than 9,100 clinical trial pages, 18,100 gene and alteration pages, 900 disease pages, and 2,700 drug pages to date. CONCLUSION: Leveraging both computational and manual curation processes in combination with reusable templates empowers automation and scalability for both the MCG knowledgebase and MCG website.

Functional and structural similarity of human DNA primase [4Fe4S] cluster domain constructs.

The regulatory subunit of human DNA primase has a C-terminal domain (p58C) that contains a [4Fe4S] cluster and binds DNA. Previous electrochemical analysis of a p58C construct revealed that its affinity for DNA is sensitive to the redox state of the [4Fe4S] cluster. Concerns about the validity of this conclusion have been raised, based in part on differences in X-ray crystal structures of the p58C272-464 construct used for that study and that of a N-terminally shifted p58C266-456 construct and consequently, an assumption that p58C272-464 has abnormal physical and functional properties. To address this controversy, a new p58C266-464 construct containing all residues was crystallized under the conditions previously used for crystallizing p58C272-464, and the solution structures of both constructs were assessed using circular dichroism and NMR spectroscopy. In the new crystal structure, p58C266-464 exhibits the same elements of secondary structure near the DNA binding site as observed in the crystal structure of p58C272-464. Moreover, in solution, circular dichroism and 15N,1H-heteronuclear single quantum coherence (HSQC) NMR spectra show there are no significant differences in the distribution of secondary structures or in the tertiary structure or the two constructs. To validate that the two constructs have the same functional properties, binding of a primed DNA template was measured using a fluorescence-based DNA binding assay, and the affinities for this substrate were the same (3.4 ± 0.5 µM and 2.7 ± 0.3 µM, respectively). The electrochemical properties of p58C266-464 were also measured and this p58C construct was able to engage in redox switching on DNA with the same efficiency as p58C272-464. Together, these results show that although p58C can be stabilized in different conformations in the crystalline state, in solution there is effectively no difference in the structure and functional properties of p58C constructs of different lengths.

Substrate Binding Regulates Redox Signaling in Human DNA Primase.

Generation of daughter strands during DNA replication requires the action of DNA primase to synthesize an initial short RNA primer on the single-stranded DNA template. Primase is a heterodimeric enzyme containing two domains whose activity must be coordinated during primer synthesis: an RNA polymerase domain in the small subunit (p48) and a [4Fe4S] cluster-containing C-terminal domain of the large subunit (p58C). Here we examine the redox switching properties of the [4Fe4S] cluster in the full p48/p58 heterodimer using DNA electrochemistry. Unlike with isolated p58C, robust redox signaling in the primase heterodimer requires binding of both DNA and NTPs; NTP binding shifts the p48/p58 cluster redox potential into the physiological range, generating a signal near 160 mV vs NHE. Preloading of primase with NTPs enhances catalytic activity on primed DNA, suggesting that primase configurations promoting activity are more highly populated in the NTP-bound protein. We propose that p48/p58 binding of anionic DNA and NTPs affects the redox properties of the [4Fe4S] cluster; this electrostatic change is likely influenced by the alignment of primase subunits during activity because the configuration affects the [4Fe4S] cluster environment and coupling to DNA bases for redox signaling. Thus, both binding of polyanionic substrates and configurational dynamics appear to influence [4Fe4S] redox signaling properties. These results suggest that these factors should be considered generally in characterizing signaling networks of large, multisubunit DNA-processing [4Fe4S] enzymes.

A Polymerase With Potential: The Fe-S Cluster in Human DNA Primase.

Replication of DNA in eukaryotes is primarily executed by the combined action of processive DNA polymerases δ and ɛ. These enzymes cannot initiate synthesis of new DNA without the presence of a primer on the template ssDNA. The primers on both the leading and lagging strands are generated by DNA polymerase α-primase (pol-prim). DNA primase is a DNA-dependent RNA polymerase that synthesizes the first ~10 nucleotides and then transfers the substrate to polymerase α to complete primer synthesis. The mechanisms governing the coordination and handoff between primase and polymerase α are largely unknown. Isolated DNA primase contains a [4Fe-4S]2+ cluster that has been shown to serve as a redox switch modulating DNA binding affinity. This discovery suggests a mechanism for modulating the priming activity of primase and handoff to polymerase α. In this chapter, we briefly discuss the current state of knowledge of primase structure and function, including the role of its iron-sulfur cluster. This is followed by providing the methods for expressing, purifying, and biophysically/structurally characterizing primase and its iron-sulfur cluster-containing domain, p58C.

Response to Comments on "The [4Fe4S] cluster of human DNA primase functions as a redox switch using DNA charge transport".

Baranovskiy et al and Pellegrini argue that, based on structural data, the path for charge transfer through the [4Fe4S] domain of primase is not feasible. Our manuscript presents electrochemical data directly showing charge transport through DNA to the [4Fe4S] cluster of a primase p58C construct and a reversible switch in the DNA-bound signal with oxidation/reduction, which is inhibited by mutation of three tyrosine residues. Although the dispositions of tyrosines differ in different constructs, all are within range for microsecond electron transfer.

The [4Fe4S] cluster of human DNA primase functions as a redox switch using DNA charge transport.

DNA charge transport chemistry offers a means of long-range, rapid redox signaling. We demonstrate that the [4Fe4S] cluster in human DNA primase can make use of this chemistry to coordinate the first steps of DNA synthesis. Using DNA electrochemistry, we found that a change in oxidation state of the [4Fe4S] cluster acts as a switch for DNA binding. Single-atom mutations that inhibit this charge transfer hinder primase initiation without affecting primase structure or polymerization. Generating a single base mismatch in the growing primer duplex, which attenuates DNA charge transport, inhibits primer truncation. Thus, redox signaling by [4Fe4S] clusters using DNA charge transport regulates primase binding to DNA and illustrates chemistry that may efficiently drive substrate handoff between polymerases during DNA replication.

Molecular basis for manganese sequestration by calprotectin and roles in the innate immune response to invading bacterial pathogens.

The S100A8/S100A9 heterodimer calprotectin (CP) functions in the host response to pathogens through a mechanism termed "nutritional immunity." CP binds Mn(2+) and Zn(2+) with high affinity and starves bacteria of these essential nutrients. Combining biophysical, structural, and microbiological analysis, we identified the molecular basis of Mn(2+) sequestration. The asymmetry of the CP heterodimer creates a single Mn(2+)-binding site from six histidine residues, which distinguishes CP from all other Mn(2+)-binding proteins. Analysis of CP mutants with altered metal-binding properties revealed that, despite both Mn(2+) and Zn(2+) being essential metals, maximal growth inhibition of multiple bacterial pathogens requires Mn(2+) sequestration. These data establish the importance of Mn(2+) sequestration in defense against infection, explain the broad-spectrum antimicrobial activity of CP relative to other S100 proteins, and clarify the impact of metal depletion on the innate immune response to infection.