An unbiased cell morphology-based screen for new, biologically active small molecules

Abstract

We have implemented an unbiased cell morphology-based screen to identify small-molecule modulators of cellular processes using the Cytometrix (TM) automated imaging and analysis system. This assay format provides unbiased analysis of morphological effects induced by small molecules by capturing phenotypic readouts of most known classes of pharmacological agents and has the potential to read out pathways for which little is known. Four human-cancer cell lines and one noncancerous primary cell type were treated with 107 small molecules comprising four different protein kinase-inhibitor scaffolds. Cellular phenotypes induced by each compound were quantified by multivariate statistical analysis of the morphology, staining intensity, and spatial attributes of the cellular nuclei, microtubules, and Golgi compartments. Principal component analysis was used to identify inhibitors of cellular components not targeted by known protein kinase inhibitors. Here we focus on a hydroxyl-substituted analog (hydroxy-PP) of the known Src-family kinase inhibitor PP2 because it induced cell-specific morphological features distinct from all known kinase inhibitors in the collection. We used affinity purification to identify a target of hydroxy-PP, carbonyl reductase 1 (CBR1), a short-chain dehydrogenase-reductase. We solved the X-ray crystal structure of the CBR1/hydroxy-PP complex to 1.24 A resolution. Structure-based design of more potent and selective CBR1 inhibitors provided probes for analyzing the biological function of CBR1 in A549 cells. These studies revealed a previously unknown function for CBR1 in serum-withdrawal-induced apoptosis. Further studies indicate CBR1 inhibitors may enhance the effectiveness of anticancer anthracyclines. Morphology-based screening of diverse cancer cell types has provided a method for discovering potent new small-molecule probes for cell biological studies and anticancer drug candidates.

Figure 1. Cell Morphology–Based Screen for Biologically Active Small Molecules

(A) Steps in the drug-screening process. Five human cell types, including one primary and four cancer cell lines, were treated for 24 h with the screening library that included compounds of known function and related analogs. The Cytometrix (TM) data analysis package was used to analyze microscopy data for each treatment condition. (B) PCA plot of the phenotypic attributes. Colored spheres represent a single compound at one concentration (ranging from 6 nM to 40 μM by 3-fold increases); lines connecting the spheres indicate a single compound's effects over a range of concentrations. Spheres are colored as follows: known protein kinase inhibitors (blue), paclitaxil (green), and novel compounds structurally related to the protein kinase inhibitors (red). The PCA provides aggregate variables termed “components” made up of multiple independent variables, each with a “loading factor.” These values are provided in Table S2. Structures of each compound are given in Figure S1. (C) Structures of the known kinase inhibitors PP and PP2 (blue), as well as the novel “hit compound” hydroxy-PP (red), are shown. Linker analogs of PP (PP-L) and hydroxy-PP (hydroxy-PP-L) that were used to ascertain the functional tolerance of replacing the t-butyl substituent at N-1 of the “hit compound” hydroxy-PP are shown. (D) Morphological attribute tabulation for cells treated with 129.4 μM PP2 (blue lines) or 0.4 μM hydroxy-PP (red lines) in each of five cell types. Data for the x-axis is grouped by the probe used (α-tubulin antibody, Hoechst dye, and lectin stain). Each of 14 attributes contributing to the magnitude of the response (y-axis) is shown as a red-filled square. (E) Visual morphology of A549 cells treated with hydroxy-PP or PP2. Hoechst dye or α-tubulin antibody was used to stain cells. The PP2-treated cells are more elongated and have more a condensed nuclear structure as compared with hydroxy-PP-treated cells.

Figure 2. Affinity Purification and Identification of Human CBR1

(A) Reactigel beads appended with hydroxy-PP or PP (control) were used for affinity purification of hydroxy-PP protein targets. (B) Hydroxy-PP-binding proteins in A549 cell lysates. Cytosolic fractions of A549 cell lysate (1.7 mg protein each) were incubated with the indicated affinity resin, and bound proteins were resolved by SDS-PAGE (12% acrylamide gel) followed by silver staining. Untreated beads and PP-control resin samples (lanes 1 and 2) indicate little nonspecific binding. Lanes 3, 4, 5, and 6 were loaded using the hydroxy-PP resin incubated with cell lysate and the indicated competitor. Vehicle or competitor compounds (200 μM) were added to the lysate 30 min before incubation with beads (lanes 4–6). Protein of bands B1–B3 did not bind hydroxy-PP beads when pretreated with hydroxy-PP (lane 5). (C) MS/MS peptide sequencing. Two tryptic peptides from bands B1–B3 were used to identify human CBR1.

Figure 3. IC₅₀ Values against CBR1 and Fyn Kinase Are Tabulated for PP Derivatives

Figure 4. Co-Crystal Structures of CBR1–Hydroxy-PP and Hck–PP1

(A, C, and E) show the binding mode of hydroxy-PP in co-crystals with CBR1. The inhibitor is oriented with its t-butyl group partially exposed to solvent and points toward the surface of the protein. The phenolic moiety of the inhibitor binds deeply within the substrate-binding pocket and makes close contacts to Ser193 and Tyr139 of the catalytic triad and the bound cofactor NADP. (B, D, and F) show the binding mode of the kinase inhibitor PP1 in complex with Hck. PP1 occupies the ATP-binding pocket as an adenosine analog. Although both protein structures show different folds (A and B), the morphology of CBR1- and Hck-binding sites are similar, and inhibitors hydroxy-PP and PP1 bind to these sites with similar shape complementarity (C and D). Key H-bond interactions between hydroxy-PP and the Ser193 and Tyr139 of CBR1 are indicated (E). The exocyclic amine of PP1 in complex with Hck makes essential H-bonds with the main-chain carbonyl oxygen of Glu339 and the side-chain oxygen of Thr338 (F). Disruption of this key H-bonding interaction by derivatization of the exocyclic amine destroys kinase affinity. The figure was prepared using the PyMol 2002 graphics system (DeLano Scientific, San Carlos, California, United States).

Figure 5. CBR1 Inhibitors Enhance Daunorubicin-Mediated A549-Cell Killing, yet Prevent Apoptosis in Serum-Starved Cells

(A) Cell viability as a function of drug treatment. DMSO, PP-L (8 μM), and hydroxy-PP-Me (8 μM) do not have a pronounced effect on cell viability when used alone. Daunorubicin (DR) alone induces a moderate decrease in cell viability that is accentuated by concomitant treatment with hydroxy-PP-Me. (B) Cell viability decreases dose dependently with concomitant daunorubicin (DR) treatment. Daunorubicin treatment induces a moderate decrease in cell viability when used alone. Hydroxy-PP-Me (1–8 μM) induces a dose-dependent decrease in cell viability with concomitant daunorubicin treatment. (C) PI staining of dead cells is appreciably decreased in serum-starved cells treated with CBR1 inhibitors. A high number of cells were PI stained 65 h following serum starvation in both control and PP-L treated conditions (top). Cells treated with the CBR1 inhibitors hydroxy-PP-L or hydroxy-PP-Me during serum starvation show appreciably less staining (bottom). (D) Quantification of PI-stained cells by fluorescence measurement 65 h following serum starvation. Hydroxy-PP-L and hydroxy-PP-Me induce a dose-dependent decrease in PI staining; whereas PP-L does not.