The antibiotics pipeline is insufficient to address the growing public health threat posed by Gram-negative ESKAPE pathogens—which are among the hardest to treat—and by other antibiotic-resistant bacteria. Of the 42 antibiotics currently in development, only 16 have the potential to target Gram-negative pathogens, and only one of those represents a novel class, according to an analysis by The Pew Charitable Trusts. To stay ahead of resistance, more novel antibiotics targeting Gram-negative bacteria are needed.
The brain drain associated with major pharmaceutical companies exiting the market—and challenges in sharing information across the shrinking community of researchers focused on developing new antibiotics—has exacerbated the already difficult problem of understanding how Gram-negative pathogens so adeptly evade antibiotics. To make matters worse, decades of discovery data are either unpublished or scattered across the literature. And because the methodologies that researchers use are not always standardized, the results are difficult to compare, past mistakes are often repeated, and progress is stymied.
To help close this research gap, Pew in September 2018 launched a free, publicly available tool to collect, standardize, and share Gram-negative research. The Shared Platform for Antibiotic Research and Knowledge (SPARK) houses previously unreleased data as well as published data, all of which are curated and standardized by expert microbiologists, enabling users to generate new insights into how antibiotics penetrate and stay inside Gram-negative bacteria. By focusing specifically on open access to previously unpublished results and data standardization, SPARK complements other antibiotic resistance information-sharing initiatives. These include the Global Antibiotic Research & Development Partnership’s REVIVE, the Innovative Medicines Initiative’s Antimicrobial Resistance (AMR) Accelerator Programme, AntibioticDB, and the Joint Programming Initiative on Antimicrobial Resistance’s Virtual Research Institute—all of which have various goals, from funding and streamlining research to connecting scientists in the AMR field.
What data are on SPARK?
SPARK currently contains 70,000 minimum inhibitory concentration (MIC) and 1,800 50 percent enzymatic inhibitory concentration (IC50) curated data points. Some of the data have been extracted from published research, but most have been contributed by biopharmaceutical companies. For example, one dataset comes from a discontinued LpxC research program that explored a novel way of attacking Gram-negative bacteria. Although no active human clinical studies for LpxC-targeting antibiotics exist, that could change as scientists around the world build upon the program’s critical findings.
Novartis also shared previously unavailable data from its discontinued programs, namely LpxADK and DNA gyrase, as well as target-agnostic efflux panel screenings. The efflux panel data include results from the screening of nearly 200 antibiotics, representing several classes and including compounds previously licensed for human use, against 22 strains (both wild type and efflux mutant strains) of predominantly Gram-negative ESKAPE pathogens.
As data are added to SPARK, curators supplement the data with annotated methodologies noting any deviations from Clinical & Laboratory Standards Institute protocols to facilitate comparisons of results from laboratories around the world.
How can researchers use SPARK?
Registered users can search SPARK data and apply computational analysis tools to easily conduct comparative analyses of MIC and IC50 data. Users can filter searches by bacterial species (e.g., A. baumannii), accumulation phenotype (e.g., efflux deficient, hyperpermeable), and/or compound properties (e.g., logD, number of rotatable bonds). They can also draw functional groups or chemical structures to use as search criteria. With an initial set of search outputs, researchers can apply SPARK’s visualization tools to rapidly identify interesting data clusters or trends—for example, based on physicochemical properties—for further analyses.
SPARK’s modeling tool allows users—before conducting resource-intensive wet-lab research—to generate predictions of which compounds are more likely to enter and accumulate inside Gram-negative pathogens. SPARK allows scientists to easily share results, request feedback on hypotheses, pose questions, and refine their models. Now, with a more targeted list of promising compounds, a microbiologist in one lab can collaborate with a medicinal chemist on another continent to synthesize chemical compounds and test in assays to generate additional MIC and IC50 data.
Pew continues to expand SPARK by pursuing additional data from current and discontinued programs in addition to extracting more data from published research. In tandem, Pew will continue to promote SPARK’s use by scientists around the world and seek feedback to improve the platform’s utility. Pew will share lessons from SPARK’s design, development, and growth to facilitate the launch of similar knowledge-sharing initiatives.