Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry for Combined Species Identification and Drug Sensitivity Testing in Mycobacteria

Abstract

Species identification and drug susceptibility testing (DST) of mycobacteria are important yet complex processes traditionally reserved for reference laboratories. Recent technical improvements in matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) has started to facilitate routine mycobacterial identifications in clinical laboratories. In this paper, we investigate the possibility of performing phenotypic MALDI-based DST in mycobacteriology using the recently described MALDI Biotyper antibiotic susceptibility test rapid assay (MBT-ASTRA). We randomly selected 72 clinical Mycobacterium tuberculosis and nontuberculous mycobacterial (NTM) strains, subjected them to MBT-ASTRA methodology, and compared its results to current gold-standard methods. Drug susceptibility was tested for rifampin, isoniazid, linezolid, and ethambutol (M. tuberculosis, n = 39), and clarithromycin and rifabutin (NTM, n = 33). Combined species identification was performed using the Biotyper Mycobacteria Library 4.0. Mycobacterium-specific MBT-ASTRA parameters were derived (calculation window, m/z 5,000 to 13,000, area under the curve [AUC] of >0.015, relative growth [RG] of <0.5; see the text for details). Using these settings, MBT-ASTRA analyses returned 175/177 M. tuberculosis and 65/66 NTM drug resistance profiles which corresponded to standard testing results. Turnaround times were not significantly different in M. tuberculosis testing, but the MBT-ASTRA method delivered on average a week faster than routine DST in NTM. Databases searches returned 90.4% correct species-level identifications, which increased to 98.6% when score thresholds were lowered to 1.65. In conclusion, the MBT-ASTRA technology holds promise to facilitate and fasten mycobacterial DST and to combine it directly with high-confidence species-level identifications. Given the ease of interpretation, its application in NTM typing might be the first in finding its way to current diagnostic workflows. However, further validations and automation are required before routine implementation can be envisioned.

Keywords: MALDI-TOF; MBT-ASTRA; drug susceptibility testing; mycobacteria.

FIG 1

MALDI-TOF MS-based DST of M. tuberculosis (A) and nontuberculous mycobacteria (B). Shown are box plots of RG values of the first sample point at which the AUC passed the threshold of 0.015. Three technical repeats were analyzed per sample, which are listed on the x axis. The whiskers extend to data points that are less than 1.5 times the interquartile range away from 1st/3rd quartile. The dotted line indicates the chosen RG threshold value of 0.5, above which a resistant strain is defined in this study.

FIG 2

MALDI-TOF MS-based MIC testing of clinical M. tuberculosis strains. Strains were incubated in the present of rising concentrations of INH (A) and RIF (B). Their resistance profile, determined by standard methods, is indicated as INH susceptible (INH-S; MIC, <0.1 μg/ml), INH low resistance (INH-LR) (MIC, <0.4 μg/ml), INH high resistance (INH-HR) (MIC, >0.4 μg/ml), rifampin susceptible (RIF-S; MIC, <1 μg/ml), or rifampin resistant (RIF-R; MIC, >1 μg/ml). The same parameters detailed in Fig. 1 are applied for the MBT-ASTRA analyses. The numbers on the x axis represent the National Reference Center identification (ID) numbers.