Host-directed antimicrobial drugs with broad-spectrum efficacy against intracellular bacterial pathogens

Abstract

We sought a new approach to treating infections by intracellular bacteria, namely, by altering host cell functions that support their growth. We screened a library of 640 Food and Drug Administration (FDA)-approved compounds for agents that render THP-1 cells resistant to infection by four intracellular pathogens. We identified numerous drugs that are not antibiotics but were highly effective in inhibiting intracellular bacterial growth with limited toxicity to host cells. These compounds are likely to target three kinds of host functions: (i) G protein-coupled receptors, (ii) intracellular calcium signals, and (iii) membrane cholesterol distribution. The compounds that targeted G protein receptor signaling and calcium fluxes broadly inhibited Coxiella burnetii, Legionella pneumophila, Brucella abortus, and Rickettsia conorii, while those directed against cholesterol traffic strongly attenuated the intracellular growth of C. burnetii and L. pneumophila. These pathways probably support intracellular pathogen growth so that drugs that perturb them may be therapeutic candidates. Combining host- and pathogen-directed treatments is a strategy to decrease the emergence of drug-resistant intracellular bacterial pathogens. Importance: Although antibiotic treatment is often successful, it is becoming clear that alternatives to conventional pathogen-directed therapy must be developed in the face of increasing antibiotic resistance. Moreover, the costs and timing associated with the development of novel antimicrobials make repurposed FDA-approved drugs attractive host-targeted therapeutics. This paper describes a novel approach of identifying such host-targeted therapeutics against intracellular bacterial pathogens. We identified several FDA-approved drugs that inhibit the growth of intracellular bacteria, thereby implicating host intracellular pathways presumably utilized by bacteria during infection.

FIG 1

Identification of antibacterial compounds that are nontoxic to host cells. (A) Six hundred forty FDA-approved drugs were first screened for their inhibition of the intracellular growth of bacteria in THP-1 cells. Then, they were screened to eliminate those toxic to THP-1 cells. Finally, host-acting compounds were identified by removing known antimicrobials. (B) Dual-fluorescence images of THP-1 cells treated with representative compounds. (Top) Agents with no effect on the intracellular C. burnetii growth or host cells; (middle) cytotoxic agents; (bottom) agents inhibiting bacterial growth but not cytotoxic. Each image represents the overlay of two fluorescence channels: 590/650 nm to detect mCherry-expressing C. burnetii and 340/380 nm to detect filipin-stained cell membranes. The pseudogreen color was used to visually improve the detection of C. burnetii. Scale bar, 100 µm. (C) Agents causing different degrees of inhibition of C. burnetii intracellular growth and host cell cytotoxicity. Circles, relative prevalence of Coxiella-containing vacuoles (CCVs); triangles, relative mCherry fluorescence; gray bars, relative viability at day 5; RFU, relative fluorescence units. (D) Effect of selected compounds on CCV size (area of individual CCV in each image). (E) Growth of C. burnetii in THP-1 cells treated with selected compounds at 33 µM. Intracellular bacterial abundance was followed over 5 days by the mCherry fluorescence of the bacteria.

FIG 2

Effects of drugs on four intracellular pathogens. (A) Numbers of drugs (in the starting library of 640) that caused >80% inhibition of the intracellular growth of C. burnetii, L. pneumophila, and R. conorii or >50% inhibition of B. abortus. Black bars represent host-targeted compounds, while the gray bars represent pathogen-targeted compounds, including antibiotics, antifungals, antivirals, antiprotozoals, antimalarials, antituberculars, and antipneumocystics. (B) Venn diagram of numbers of overlapping compounds among the 101 most effective host-targeting compounds inhibiting the intracellular growth of four pathogens. The values in the squares give the numbers of compounds that affect the indicated organisms, while the numbers in parentheses in the four sectors of the diagram give the sums of the values for the overlapping compounds for each organism. (C) Heatmap generated using hierarchical clustering of inhibitory efficiencies of all compounds from the library. Blue denotes no inhibition, and red denotes complete inhibition. Labels A to D indicate the representative clusters of hits shown in panel D. Ba, B. abortus; Rc, R. conorii; Cb, C. burnetii; Lp, L. pneumophila. (D) Closeup heatmaps of clusters A to D from panel C. Compound numbers are shown next to each name.

FIG 3

Effect of cholesterol perturbation on the intracellular growth of C. burnetii. (A) HeLa cells treated for 24 h with representative compounds at 33 µM and then fixed and stained for cholesterol with filipin. Grayscale was inverted for better visualization. Scale bar, 100 µm. (B) Effects of 62 cholesterol-disrupting compounds on C. burnetii intracellular growth measured by fluorescence (relative to that of DMSO-treated control cells). (C) Chemical structures of cholesterol and 5 of the 62 compounds that perturbed cholesterol but not C. burnetii growth.

FIG 4

Genetic disruption of cholesterol homeostasis inhibits the intracellular growth of C. burnetii. (A) Images of NPC-1 shRNA-mediated inhibition of intracellular C. burnetii. Panels I and IV show Nomarski images of control (empty-vector shRNA) THP-1 cells and cells transfected with NPC-1 shRNA. Panels II and V show fluorescence signals generated from the shRNA vector. Panels III and VI show intracellular C. burnetii. For direct comparison of fluorescence signals, images were taken at a constant exposure. Scale bars, 100 µm. (B) Growth curve of C. burnetii in THP-1 cells transfected with the control and three different constructs of NPC-1 shRNA. Western blot of NPC-1 levels relative to those in an actin control in cells transfected with NPC-1 shRNA constructs and control shRNA.

FIG 5

Identification of compounds that affect microbial entry. (A) Graphs representing correlations between efficiencies of drug inhibition of the intracellular bacterial growth of C. burnetii, L. pneumophila, and B. abortus when the drug was applied preinfection and postinfection. (B) Heatmaps generated by binning (K-means = 20) hierarchical cluster analysis results for the efficiency of drugs pre- versus postinfection. The color scale ranges from 0 (blue, no inhibition) to 1 (red, complete inhibition). (C) Bar graphs representing compounds from clusters with the highest difference in inhibition between the pre- and postinfection treatments. Inhibition values <1 were set to 1, and values <0 were set to 0.