Activity of Telavancin against Staphylococcus aureus Isolates, Including Those with Decreased Susceptibility to Ceftaroline, from Cystic Fibrosis Patients

Susceptibility of cystic fibrosis patient-derived MRSA/MSSA strains to telavancin.The CF patient-derived S. aureus strains used in the present study were of either the wild-type or small-colony-variant (SCV) phenotype, identified at the time of culture, and were obtained from three different CF centers. We screened a total of 333 strains; those collected at the Houston Methodist Research Institute, 103 in total, were distributed as 37 MRSA and 66 MSSA strains, while strains originating from UW Health comprised 72 MRSA and 10 MSSA strains. The ones obtained from the Center for Global Infectious Disease Research at Seattle, WA, comprised 148 total S. aureus strains, with 44 being MRSA strains and 104 being MSSA strains. As shown in Table 1, TLV displayed activity against all 333 strains derived from CF patients at three different CF centers from across the United States; the majority of these S. aureus strains were isolated from adult patients, with only 41 strains being found in children. Furthermore, we tested the activity of TLV against 23 MRSA strains; 20 displayed intermediate resistance to ceftaroline (CPT^ir; MICs, 1.5 to 2 μg/ml) and 3 had high-level resistance to ceftaroline (CPT^hr; MIC, 32 μg/ml).

CPT is a new β-lactam antibiotic that specifically targets PBP 2a in MRSA. Although a high level of resistance to CPT remains rare, intermediate resistance is more prevalent in patients with chronic infections. Among all strains, the TLV MIC₉₀ was 0.06 mg/liter, i.e., 8-fold lower than the daptomycin (DAP) and CPT MIC₉₀ and 25-fold lower than the linezolid (LZD) and vancomycin (VAN) MIC₉₀. In the strains with reduced CPT susceptibility, the TLV MIC₉₀ was 0.06 μg/ml for both the CPT^ir TMH 5006 and CPT^hr TMH 5007 strains, showing the absence of cross-resistance between the two antibiotics (Table 1). Of note, although DAP showed in vitro activity against CF patient-derived S. aureus strains, the presence of 1% surfactant resulted in an 8- to 32-fold increase in the DAP MIC (up to 8 μg/ml), while the MICs for the other antibiotics, including TLV, remained unchanged (data not shown). These data support previously documented data showing the inactivation of daptomycin in the lung and provide evidence that other anti-MRSA treatment options retain potency.

Of note, we found for strain AMT-0067-21 the presence of internal colonies for which the TLV Etest MIC was 0.19 μg/ml (the MIC was homogeneous up to 0.047 μg/ml) and the broth microdilution MIC was 0.25 μg/ml. Interestingly, the tested MRSA and MSSA strains were not associated with the VISA or DAP-nonsusceptible phenotype. These data altogether provide evidence that TLV retains potency against CF patient-derived S. aureus strains.

TLV shows bactericidal activity against CF patient-derived MRSA/MSSA isolates.The in vitro effectiveness of TLV was also evaluated by time-kill experiments and was compared to that of DAP, VAN, and CPT. The assay was performed in Muller-Hinton broth supplemented with 0.002% polysorbate 80 for TLV and calcium at 50 μg/ml for DAP, and activity against a representative number of CF patient-derived strains, including a strain with CPT^hr (TMH 5007), was tested (16–18). TLV showed activity against all the tested strains, including CPT^hr strain TMH 5007, in concordance with the TLV MIC values. Moreover, TLV displayed rapid bactericidal activity, with a decrease of more than 3 log₁₀ CFU being achieved during the first 4 h of growth (Fig. 1). Similar activity against the CPT^hr strain was observed, confirming the absence of cross-resistance or reduced activity. The TLV activity profile at a free serum concentration of 8 mg/liter showed that it performed better than VAN (16 mg/liter), LZD (10.4 mg/liter), and CPT (16 mg/liter) (Fig. 1). Together these data support the potential therapeutic application of TLV for the treatment of S. aureus infections in CF patients.

• Open in new tab
• Download powerpoint

FIG 1

Time-kill curves of CF strains AMT 0307-5 (MSSA), AMT 0012-33, WIS 664 (MRSA), and TMH 5007 (MRSA CPT^r), determined using the following human peak free serum drug concentrations: 8 mg/liter for TLV, 16 mg/liter for VAN, 10.4 mg/liter for LZD, and 16 mg/liter for CPT. The limit of detection of that assay was 10.

In vitro selection of TLV-resistant (TLV^r) mutants.To determine the fate of mutation selection that can be projected by the potential prolonged use of TLV in CF patients, we investigated the ease of in vitro mutation selection in three representative clinical strains from CF patients: the AMT 0114-48, WIS 664, and TMH 5007 (CPT^hr) strains. S. aureus ATCC 25923 was included as a non-CF patient-derived control strain. These strains were serially passaged in the presence of subinhibitory concentrations of TLV starting at 0.03 μg/ml and escalating up to 3 μg for 40 days. After 15 days, the strains showed an increase in the TLV MIC from 0.064 μg/ml up to 0.25 to 1 μg/ml, followed by a progressive increase up to 3 μg/ml after 40 days of exposure (Table 2). The enhanced MICs of the mutants were stable and unaltered after 10 passages in the absence of TLV. As shown in Table 2, there was a 3- to 4-fold increase in the VAN MIC, which went from 1.5 to 6 μg/ml, and an 8- to 10-fold increase in the DAP MIC, suggesting potential cross-resistance between TLV, VAN, and DAP antibiotics. Moreover, the in vitro-derived TLV^r mutants grew at lower rate than the parent strains and were defective in growth, requiring 48 h to obtain normal-size colonies on tryptic soy agar (TSA) blood agar plates (data not shown). Important observations were taken from these results: (i) the ease of mutant selection observed in S. aureus ATCC 25913 control strains led us to conclude that TLV mutant resistance is independent of the CF patient background of the strains, and (ii) the likelihood that TLV strains with increased TLV MICs, which was observed for only one strain resistant in vivo, AMT-0067-21 (MIC, 0.19 μg/ml), will occur seems rare, considering the fact that this strain represented only 0.3% of the total of 333 strains tested.

Genes associated with in vivo and in vitro TLV resistance.To investigate the genetic mutations associated with the in vivo (MRSA AMT-0067-21) or in vitro TLV-resistant mutants, the full genomes of all the strains were sequenced and compared with the sequence of reference strain S. aureus N315 (PATRIC accession number 158879.11). Mutated genes were categorized by function to identify themes of bacterial physiology that may contribute to reduced susceptibility to TLV (Table 3). The most common nonsynonymous single nucleotide polymorphisms (SNP) found by comparing the sequence of each strain with that of its counterpart parental strain were found in cell wall-associated genes, corresponding to sdrCDE (encoding a cell wall-associated genes), tcaA (encoding a transmembrane protein associated with teicoplanin resistance), and dltD (encoding a d-alanine transfer protein). In addition, we found additional nonsynonymous SNPs in various cell wall-associated genes: spa, clfA, clfB, sdrE, and srdD (Table 3). Moreover, additional SNPs were found in the in vivo-derived strain MRSA AMT-0067-21, against which TLV displayed reduced activity, notably, murE (a stop codon at position Q), dltA, and vraG (encoding a bacitracin export permease). Gene deletion occurred for isdB (encoding a cell surface receptor for hemoglobin), mutL (encoding a DNA mismatch protein), and fnB (encoding a fibrinogen binding protein) and in the yycFG regulon (encoding a two-component regulatory system) (Table 3). These results may suggest that a decrease in susceptibility to TLV is mainly associated with changes in cell wall-related genes.

In vitro TLV mutants are associated with reduced virulence.In order to determine whether TLV resistance acquired in vitro may impact virulence traits in S. aureus, we used Galleria mellonella as an in vivo model as it possesses an immune system with reasonable homology to that of vertebrates, and numerous enzymatic cascades akin to complement fixation and blood coagulation occur in the hemolymph, resulting in hemolymph clotting and melanin production as key defense mechanisms against invading microbes. These tissue types are similar to those encountered by S. aureus during invasive infections in humans (19).

Groups of larvae (10/group) were inoculated with a bacterial suspension containing parent strains AMT 0114-48 and WIS 664 and control strain ATCC 25913 and their corresponding TLV-resistant mutants derived in vitro, AMT 0114-48 TLV^r, WIS 664 TLV^r, and ATCC 25913 TLV^r (10⁶ bacteria/worm), as previously described (19). An uninfected control group received phosphate-buffered saline (PBS) treatment to control for multiple injections. Worms were monitored daily, and any deaths that occurred over the next 10 days were recorded. Worms injected with PBS showed 100 to 90% survival at day 8 (Fig. 2), but groups injected with the parent strains (e.g., AMT 0114-48) displayed low survival rates (≤50 to 0%, day 6; Fig. 2). In contrast, groups of worms infected with TLV^r strains (e.g., AMT 0114-48 TLV^r) had a survival rate of 90% at day 6, followed by a survival rate of 60% at day 8. A similar trend was observed for worms injected with the ATCC 25923 strain, although the survival rate was higher (40%) than that for worms injected with CF patient-derived parent S. aureus strains (0 to 20%), while the survival rates for worms injected with the TLV^r strains were comparable. These results may suggest that TLV^r is associated with changes in virulence fitness.

• Open in new tab
• Download powerpoint

FIG 2

Comparison of virulence of S. aureus strains and their isogenic TLV-resistant mutants in G. mellonella model of infection. The worms were infected at a dose of 10⁶ bacteria per worm. Survival was monitored for 10 days.