Background: The spread of multi-drug resistant tuberculosis (MDR-TB) is a leading global public-health challenge. Because not all biological mechanisms of resistance are known, culture-based (phenotypic) drug-susceptibility testing (DST) provides important information that influences clinical decision-making. Current phenotypic tests typically require pre-culture to ensure bacterial loads are at a testable level (taking 2-4 weeks) followed by 10-14 days to confirm growth or lack thereof.
Methods and findings: We present a 2-step method to obtain DST results within 3 days of sample collection. The first involves selectively concentrating live mycobacterial cells present in relatively large volumes of sputum (~2-10mL) using commercially available magnetic-nanoparticles (MNPs) into smaller volumes, thereby bypassing the need for pre-culture. The second involves using microchannel Electrical Impedance Spectroscopy (m-EIS) to monitor multiple aliquots of small volumes (~10μL) of suspension containing mycobacterial cells, MNPs, and candidate-drugs to determine whether cells grow, die, or remain static under the conditions tested. m-EIS yields an estimate for the solution "bulk capacitance" (Cb), a parameter that is proportional to the number of live bacteria in suspension. We are thus able to detect cell death (bactericidal action of the drug) in addition to cell-growth. We demonstrate proof-of-principle using M. bovis BCG and M. smegmatis suspended in artificial sputum. Loads of ~ 2000-10,000 CFU of mycobacteria were extracted from ~5mL of artificial sputum during the decontamination process with efficiencies of 84% -100%. Subsequently, suspensions containing ~105 CFU/mL of mycobacteria with 10 mg/mL of MNPs were monitored in the presence of bacteriostatic and bactericidal drugs at concentrations below, at, and above known MIC (Minimum Inhibitory Concentration) values. m-EIS data (ΔCb) showed data consistent with growth, death or stasis as expected and/or recorded using plate counts. Electrical signals of death were visible as early as 3 hours, and growth was seen in < 3 days for all samples, allowing us to perform DST in < 3 days.
Conclusion: We demonstrated "proof of principle" that (a) live mycobacteria can be isolated from sputum using MNPs with high efficiency (almost all the bacteria that survive decontamination) and (b) that the efficacy of candidate drugs on the mycobacteria thus isolated (in suspensions containing MNPs) could be tested in real-time using m-EIS.