3.1 Bacterial strains identification
Among the 31 MABSC isolates in the study set, all 11 isolates designated M. abscessus type 1 by PRA-hsp65 typing were identified as M. abscessus subsp. abscessus by rpoB sequencing. Of the 20 M. abscessus type 2 isolates, eight were identified as M. abscessus subsp. massiliense, nine as M. abscessus subsp. bolletii, and three as M. abscessus subsp. abscessus.
3.2 Minimal inhibitory concentration determination with REMA
By REMA, ATCC 19977T was susceptible to AMK (MIC <4mg/L), intermediate to FOX (64mg/L), resistant to IPM (>128mg/L), and demonstrated inducible resistance to CLR (<0.5mg/L at day 3 and >64mg/L at day 14), Among the 31 clinical isolates the distribution of susceptibility results by microplate testing differed across the antibiotics evaluated and the species of the isolates (Table 2). Clarithromycin demonstrated the most diverse results, with 45% (14/31) of isolates being susceptible, including 100% (8/8) of M. abscessus subsp. massiliense and none (0/9) of M. abscessus subsp. bolletii; 45% (14/31) demonstrated induced resistance and 10% (3/31) were resistant. In contrast, 97% (30/31) of isolates were susceptible to amikacin, 6.4% (2/31) to cefoxitin, and none to imipenem.
3.3 Association of erm(41) and rrl genotypes and clarithromycin susceptibility phenotypes
For all 31 isolates, the genotypes at erm(41) and rrl were compared with the results of CLR microplate susceptibility testing using the Day 14 reading and the subspecies identification (Table 3).
PCR analysis indicated the presence of deletions in erm(41) among 11 of the 14 clarithromycin-susceptible isolates, including all 8 M. abscessus subsp. massiliense. Of the three susceptible isolates without deletions, all were M. abscessus subsp. abscessus that carried the erm(41) T28C point polymorphism25. Among the 17 resistant isolates, PCR indicated that all had an intact WT erm(41). Sequencing of rrl identified only a single isolate with the A2058G mutation; that isolate also had intact erm(41) and was resistant to clarithromycin. Thus, across all 31 MABSC isolates CLR susceptibility by microplate testing was consistent with erm(41) and rrl genotypes.
3.4 Inoculum preparation for MGIT 960/TB eXiST susceptibility testing
Initially, growth of ATCC 19977T was assessed with OADC supplement using dilutions of 10-1, 10-2, 2x10‑4, 10-4, 2x10-5 and 10-8. The BACTEC MGIT 960 system requires that the proportional growth control to reach 400 growth units (GU) after 3.0 days. Growth of ATCC 19977T in MGIT tube with OADC reached that threshold after one day of incubation for all dilutions up to 10-4, and after two and three days of incubation for the 2x10-5 and 10-8 dilutions, respectively. Subsequently, the 10-4 dilution was evaluated with and without OADC enrichment and bacterial growth was detected after two and four days of incubation, respectively. Similar results were obtained with a subset of the clinical isolates. Thus, for MABSC isolates, which have a faster intrinsic growth rate than MTB or slow-growing nontuberculous mycobacteria (NTM), OADC enrichment resulted in accelerated growth rates inconsistent with the timeframes and endpoints specified for the growth control (GC) in the BACTEC MGIT 960 system. Consequently, in all subsequent work OADC was omitted from both growth control and drug testing tubes in MGIT 960, as was already the case in the BMD method.
The appropriate dilution for tubes with drugs was determined using three isolates with different susceptibility profiles for CLR by REMA. The three isolates – ATCC 19977T, which demonstrates inducible resistance, plus two clinical isolates, one fully susceptible (isolate 2566) and one constitutively resistant (isolate 1656) – were evaluated using 10-2 and 10-3 dilutions prepared as described. Only the 10-2 dilution consistently provided the expected susceptibility profiles for all three CLR phenotypes (data not shown) and was subsequently confirmed as satisfactory for the other antibiotics.
3.5 Demonstration of the BACTEC MGIT 960/TB eXiST susceptibility test
The TB eXiST software monitors each tube for the growth of the organisms over time. For those tubes designated as representing a specific isolate with different concentrations of a single antibiotic plus the growth control, the results are automatically plotted together on one graph whose vertical axis is growth units and horizontal axis, the number of days of incubation. Figure 1 displays the results for an assay of ATCC 19977T with CLR. The solid blue line is the growth control (the 10‑4 dilution), which reached 400 GU at 3.5 days. The dotted and dashed blue lines represent 10‑3 and 10-2 dilutions, respectively, which reached 400 GU in <72 hours and, as noted above, would not be valid controls. The other colored lines represent the growth curves in the presence of different concentrations of CLR. The black vertical line (a mix dots and dashes) to the far right marks the endpoint of the assay, which is prespecified in the TB eXiST software as seven days after the GC tube reaches 400 GU; in the example shown, GC reached 400 GU at 3.5 days, and so the assay endpoint is 10.5 days.
The interpretation of the BACTEC MGIT 960 system is based on the incubation time at which growth at the breakpoint (also referred to as the critical concentration) for the antibiotic being tested (e.g., 2 mg/L for CLA) reaches 100 GU relative to the time the GC reaches 400 GU and the time of the assay endpoint. Specifically, the isolate is considered resistant if growth in the presence of antibiotic achieves 100 GU before the GC reaches 400 GU; intermediate, if it reaches 100 GU after the GC gets to 400 GU, but before assay endpoint; and susceptible, if it fails to grow at all or reaches 100 GU only after the assay endpoint. In the assay shown in Figure 1, ATCC 19977T in the presence of CLR 2 mg/L (green line) achieved 100 GU at 5.6 days incubation, and the isolate is therefore assessed as intermediate susceptibility for CLR. Of note, the organism met criteria for intermediate susceptibility in the presence of CLR across all the concentrations tested (0.5 through 8 mg/L), consistent with the inducible resistance phenotype demonstrated in REMA.
The results of CLR susceptibility testing for M. abscessus subsp. abscessus isolates 1656 and 2566 using the BACTEC MGIT 960 system are shown in Figures 2 and 3, respectively. At the breakpoint, isolate 1656 achieved 100 GU at 5.25 days, at least 1 full day before the GC curve reached 400 GU, and thus meets criteria for resistant. Isolate 2566 demonstrates susceptibility to CLR, with no growth observed in the presence of any drug concentration tested. In media alone isolate 2566 achieved 400 GU at 4.8 days indicating a valid assay. For both isolates, the MGIT 960 results were concordant with MIC testing.
Using the BACTEC MGIT 960 system, ATCC 19977T was assessed as susceptible to AMK, intermediate resistant to FOX, and resistant to IPM (curves not shown). All three results were concordant with MIC testing.
3.6 Application of BACTEC MGIT 960/TB eXiST susceptibility test
Using the protocol developed as detailed above, the 31 study isolates were tested for susceptibility to CLR, AMK, FOX, and IPM by BACTEC MGIT 960/TB eXiST system (Table 4) and the results compared to those obtained with MIC susceptibility testing using the REMA protocol (Table 5).
For amikacin, there was 100% concordance with 30 isolates susceptible and one resistant in both systems. For imipenem, 26 isolates were resistant by both systems. The remaining five isolates were also resistant by MIC testing, but were intermediate by MGIT 960, and thus represented minor errors.
For cefoxitin, both methods gave the same results for 18 isolates, which included one susceptible, two resistant, and 15 intermediate. All the discrepancies were minor errors, including 12 isolates assessed as intermediate by REMA, but resistant by BACTEC MGIT 960 system and one isolate that was susceptible by REMA, but intermediate by MGIT.
Comparison of the two methods for CLR requires consideration of the phenomenon of inducible resistance, which, using the REMA method, was identified as no growth at 3 days, but growth at some later timepoint up to 14 days. By that technique, all 14 (45.2%) isolates assessed as susceptible (that is, no growth thru Day 14), were also susceptible by MGIT 960. Microplate identified 3 (9.6%) isolates as resistant (i.e., growth at Day 3), with 14 (45.2%) demonstrating inducible resistance (i.e., no growth at Day 3, with growth when monitored at Day 5, 7, 10, or 14). Among those 17 isolates, the MGIT 960 protocol, without modifications, reported 9 (29.0%) as resistant and 8 (25.8%) as intermediate. The specific observations in both methods for the 14 inducible resistant isolates are detailed in Table 6. There was no apparent correlation between the duration of incubation required to detect inducible resistance by REMA and whether the isolate was reported as intermediate or resistant by BACTEC MGIT 960. Accepting the results for intermediate or resistant by MGIT as equivalent to resistant or inducible-resistant by BMD, then the overall concordance rate for CLR was 100% (31/31).