CLOUD with a Silver Lining

When you think of a good weather forecast, you might picture a landscape without a cloud in the sky, but for a bright therapeutic forecast a CLOUD might be exactly what we need. Adding to the search for novel therapies, the Centre of Excellence for Medical Multimedia (CeMM) in Colorado Springs, US, has created a database of FDA-approved drug compounds that can be used to easily compare combinations of drugs: the CeMM Library of Unique Drugs – CLOUD for short.

“At first, we just bought commercially available collections, and used a bioinformatics approach to identify the best of several competing products. But at that stage, we realized that all commercially available drug collections – even when combined – lacked approved compounds for a significant number of target classes,” says Stefan Kubicek, Head of the Proteomics and Metabolomics Facility at CeMM. “We decided to make our own reductionist collection that optimally represents all approved drugs regarding their molecular targets and chemical structures.” The result is a highly condensed library of 308 compounds, which includes features that have been neglected in other libraries, such as the active metabolite of drugs administered as prodrugs.

Why the reductionist approach? “The collection of only 308 compounds is small enough for systematically testing all pairwise combinations,” Kubicek explains. “Carrying out systematic combinatorial screens on the scale of, for example, all compounds in the NIH Chemical Genomics Center collection, would easily overwhelm even the largest industrial screening infrastructures.”

The downscaling of the task allowed CLOUD to be created in just a few simple steps:

1. Extract unique active pharmaceutical ingredients from the Drugs@FDA database.
2. Remove large macromolecules, molecules that don’t operate via protein-ligand interactions, molecules that aren’t used to treat diseases, and molecules that are only found in tropical regions. 
3. Annotate the remaining drugs with their molecular targets, and group them by target class and chemical structure similarity.
4. Combine the list with the 34 drugs that have unknown targets alongside their 35 active forms of prodrugs.

The resulting 308 compounds encompass all FDA-approved chemical entities (including active forms) in a single screening plate. Other aspects of CLOUD creation were more complex. “While databases exist for chemical structures and targets of approved drugs, maximum human plasma concentrations are not systematically annotated and, for many compounds, there is high variability in the numbers reported in the literature,” says Kubicek. “Another challenge was the physical assembly of the collection. Though the majority of these compounds are readily available, a subset was hard to obtain.”

Labs at the CeMM now use CLOUD for the setup and optimization of all their screens. One group has already discovered that a combination of flutamide and phenprocoumon modulates androgen receptor (AR) stability and re-sensitizes AR-mutant prostate cancer cells to flutamide, for example (1). “At concentrations where neither compound affected the viability of these cells, the combination efficiently killed the cancer cells,” says Kubicek. “Based on the known use of the antiandrogen flutamide as a prostate cancer drug, we tested the combination in prostate cancer cells and found an even higher degree of synergy. Thus, we could describe the discovery and molecular characterization of a novel drug synergy that has the potential to clinically benefit patients with resistant prostate cancer.”

The latest discovery showcases the potential of CLOUD – and Kubicek hopes that other screening centers around the world will adopt similar approaches.