Drug screening to identify compounds to act as co-therapies for the treatment of Burkholderia species

Burkholderia pseudomallei is a soil-dwelling organism present throughout the tropics. It is the causative agent of melioidosis, a disease that is believed to kill 89,000 people per year. It is naturally resistant to many antibiotics, requiring at least two weeks of intravenous treatment with ceftazidime, imipenem or meropenem followed by 6 months of orally delivered co-trimoxazole. This places a large treatment burden on the predominantly middle-income nations where the majority of disease occurs. We have established a high-throughput assay for compounds that could be used as a co-therapy to potentiate the effect of ceftazidime, using the related non-pathogenic bacterium Burkholderia thailandensis as a surrogate. Optimization of the assay gave a Z’ factor of 0.68. We screened a library of 61,250 compounds and identified 29 compounds with a pIC50 (-log10(IC50)) greater than five. Detailed investigation allowed us to down select to six “best in class” compounds, which included the licensed drug chloroxine. Co-treatment of B. thailandensis with ceftazidime and chloroxine reduced culturable cell numbers by two orders of magnitude over 48 hours, compared to treatment with ceftazidime alone. Hit expansion around chloroxine was performed using commercially available compounds. Minor modifications to the structure abolished activity, suggesting that chloroxine likely acts against a specific target. Finally, an initial study demonstrates the utility of chloroxine to act as a co-therapy to potentiate the effect of ceftazidime against B. pseudomallei. This approach successfully identified potential co-therapies for a recalcitrant Gram-negative bacterial species. Our assay could be used more widely to aid in chemotherapy to treat infections caused by these bacteria.

ATP levels in an untreated culture of B. thailandensis was quantified with Bactiter Glo reagent, in a series of two-fold dilutions with media. Signal from the Bactiter Glo was converted to an ATP concentration using an ATP standard curve in the same media. Results are the mean of three replicates.
Error indicates 95% confidence intervals. Z' for 0.8 to 0.4 = 0.61. Upper: Standard curve created from gDNA dilutions. qPCR of the BTH_II0730 gene from B.

-
thailandensis with the SYBR green probe showed a detectable difference in 10-fold dilutions over a good dynamic range from 10 -2 to 10 3 ng/mL. A best fit of these data gives y = -1.5ln(x) + 26; R 2 = 0.99.
Lower: Calculated DNA quantities from a dilution series of B. thailandensis. No significant differences are observed between different dilutions.

Plasmid encoded fluorescent protein
This approach used a strain of B. thailandensis modified with a plasmid containing red fluorescent protein (RFP) whose expression was driven by the groS promoter. The rationale for this approach was that the constitutive expression should give a consistent signal for surviving cells; and the fluorescence should give both sensitivity and a good dynamic range. Despite having a significantly slowed metabolism, antibiotic tolerant cells are still able to produce proteins. Whilst this approach gave excellent resolution in detecting a two-fold difference in seeded cell numbers and displaying a suitable dynamic range (Fig 3), the assay was not suitable for determining the effects of additional compounds on cells. This was due to the expressed fluorophore accumulating in solution and inhibiting the detection of cell number reductions (Fig 4). Fluorescence of a B. thailandensis culture expressing a plasmid encoding RFP in the presence and absence of ceftazidime. A culture was grown to OD600 = 1.6 before being harvested and resuspended in fresh LB containing 750 µg/ml chloramphenicol. The culture was adjusted to 8 x 10 8 cfu/mL (pretreatment) with the same media or media supplemented with ceftazidime to 730 M (antibiotic treated).
Cells were grown for 20 h at 37 °C and fluorescence read at ex 588 nm and em 635 nm using an Infinite M200 Pro (Tecan) plate reader. Antibiotic treated and blank media were similarly measured for comparison.

LIVE/DEAD cell viability staining
LIVE/DEAD® BacLight™ Cell viability staining is a two-color fluorescence assay, combining a membrane soluble green nucleic acid stain (SYTO9), and a red nucleic acid stain (propidium iodide) that does not penetrate membranes. The ratios of dyes used was optimized for B. thailandensis. Initially, when LB media was used, this method did not give an acceptable signal to noise ratio (data not shown).
However, the use of M9 minimal media reduced background fluorescence. Further optimization included incubation with a breathable membrane. These steps improved the assay to give LIVE / DEAD staining a suitable dynamic range ( Figure 5) and differentiating ability. As this stain is more expensive than PrestoBlue, it was not selected for the high throughput screen.

ATP measurement
A culture of B. thailandensis was grown to a cell density of OD600 1.6, equivalent to 1.6 x10 9 CFU/ml.
Cells were harvested and resuspended in an equal amount of LB supplemented with 730 M of ceftazidime. Samples were incubated statically for 24 hours at 37 °C. ATP concentration was determined against an ATP ladder produced from a 10 µM stock solution of ATP in LB which was serially diluted in 10-fold dilutions in a 96 well plate. Aliquots of 100 µl for all samples; t0, t24 and ATP standards were added to wells of an opaque, white walled 96 well plate (Corning, #3917) and 100 µl BacTiter-Glo reagent (Promega, #G8230) added. Plates were mixed using an orbital shaker and incubated at room temperature for 5 minutes before luminescence was read using an Infinite M200 Pro (Tecan) plate reader.

qPCR evaluation of cell numbers
The experiment was designed according to the minimum information for publication of quantitative realtime PCR experiments (MIQE) guidelines 4 . Primers were designed for BTH_II0730 using Applied Biosystems Primer Express 3.0 software for a 300-400 bp amplicon. Genomic DNA (gDNA) was extracted from a stationary culture of B. thailandensis using a GeneJET Genomic DNA Purification Kit (ThermoFisher, #K0491) and quantified using a NanoDrop 2000c spectrophotometer. gDNA was diluted to 2 ng/µl and 10-fold dilutions were made in distilled water for a standard curve. Two-fold dilutions of B. thailandensis from a stationary culture in LB were produced from OD600 1.6 to 0.2. qPCR was carried out using SYBR Green Real-Time PCR Master Mix (ThermoFisher) on a Step One Real-Time PCR System (Applied Biosystems).

Plasmid encoded fluorescent protein
The plasmid pBHR4-groS-RFP 5 was conjugated into B. thailandensis strain E264. A culture was grown to OD600 1.6 before being harvested and resuspended in fresh LB containing 750 µg/ml of chloramphenicol. The culture was adjusted to 8 x 10 8 cfu/mL and two-fold dilutions made in a blackwalled 96 well assay plate (Corning, in LB media. To test with antibiotic, cells were harvested by centrifugation and resuspended in LB media supplemented with 730 µM of ceftazidime. Fluorescence was read at ex 588 nm and em 635 nm using an Infinite M200 Pro (Tecan) plate reader.

LIVE/DEAD cell viability staining
The LIVE/DEAD BacLight bacterial cell viability kit (Invitrogen, #L7012) was used for this assay. A culture was grown to OD600 1.6 before being harvested and resuspended in fresh LB or M9 media containing 730 µM of ceftazidime. Two and ten-fold dilutions were made with the same media, and 100 L added to a black walled 96-well plate (Corning, #3904), and the samples incubated statically for 24 hours at 37 °C. A master mix of equal volume SYTO9 to propidium iodide was prepared and 3 µl added to wells and mixed thoroughly. Plates were incubated at room temperature in the dark for 15 minutes and fluorescence was read at ex 480 / em 500 nm for SYTO9 stain and ex 490 / em 635 nm for propidium iodide.

Cytotoxicity assay
The cytotoxicity assay was performed using the SH-SY5Y human neuroblastoma cell line. 10,000 cells/well were seeded into a 96 well tissue culture plate and grown overnight in 100 L DMEM supplemented with fetal calf serum at 37 °C in 5% CO2. 10 L of 30 M chloroxine in culture medium with 0.1% DMSO was added to test wells and incubated for 4 or 24 h as above. Cytotoxicity was determined using an LDH cytotoxicity assay kit (Thermo, #88953), following the manufacturer's instructions. 100% cytotoxicity was determined by adding 10 L of the kit cell lysis reagent to control cells and incubating at 37 °C for one hour. 0% cytotoxicity was determined using culture medium. 10 L of water was added to the control cells one hour before readings were taken to provide identical volume to test samples. Statistics were performed using R.