Scope and Predictive Genetic/Phenotypic Signatures of Bicarbonate (NaHCO3) Responsiveness and β-Lactam Sensitization in Methicillin-Resistant Staphylococcus aureus

Abstract

Addition of sodium bicarbonate (NaHCO₃) to standard antimicrobial susceptibility testing medium reveals certain methicillin-resistant Staphylococcus aureus (MRSA) strains to be highly susceptible to β-lactams. We investigated the prevalence of this phenotype (NaHCO₃ responsiveness) to two β-lactams among 58 clinical MRSA bloodstream isolates. Of note, ∼75% and ∼36% of isolates displayed the NaHCO₃ responsiveness phenotype to cefazolin (CFZ) and oxacillin (OXA), respectively. Neither intrinsic β-lactam MICs in standard Mueller-Hinton broth (MHB) nor population analysis profiles were predictive of this phenotype. Several genotypic markers (clonal complex 8 [CC8]; agr I and spa t008) were associated with NaHCO₃ responsiveness for OXA.

Keywords: antimicrobial susceptibility testing; beta-lactams; methicillin-resistant Staphylococcus aureus; sodium bicarbonate.

FIG 1

Frequency of NaHCO₃ responsiveness among 58 clinical MRSA isolates. (A) Frequency of NaHCO₃ responsiveness to cefazolin (CFZ) and oxacillin (OXA). Responsiveness is defined as a ≥4-fold reduction in MIC in the presence of NaHCO₃ compared to that of medium lacking NaHCO₃. (B) Frequency of coresponsiveness to CFZ and OXA. Coresponsiveness is defined as a strain that is NaHCO₃ responsive to both CFZ and OXA. Correlated phenotypes are those in which a strain is either responsive to both drugs (coresponsive) or responsive to neither drug (nonresponsive). Discordant phenotypes are those in which a strain is responsive to only one drug (either CFZ or OXA). Kappa coefficient of correlation (κ) = 0.25, **, P = 0.008.

FIG 2

Population analysis profiles for NaHCO₃-responsive and nonresponsive strains. (A) Population analysis profiles of CFZ for NaHCO₃-responsive strains in the presence and absence of NaHCO₃. The area under the curve (AUC), calculated by linear approximation, was significantly decreased by exposure to NaHCO₃ in both strains as analyzed by Student’s t test (PB 043-043, ***, P < 0.001; PB 031-038, ****, P < 0.0001). (B) Population analysis profiles of CFZ for NaHCO₃-nonresponsive strains in the presence and absence of NaHCO₃. As analyzed by Student’s t test, there is no significant difference in the AUC for C7 in media ± NaHCO₃; the AUC of RB 067-227 was significantly increased by exposure to NaHCO₃ (*, P < 0.05). (C) Population analysis profiles of CFZ for the methicillin-susceptible S. aureus (MSSA) strain ATCC 25923.

FIG 3

Time kill curves for NaHCO₃-responsive and nonresponsive strains. (A) PB 043-043 (squares). (B) PB 031-038 (circles). (C) C7 (diamonds). (D) RB 067-227 (triangles). Closed symbols represent growth in CA-MHB 100 mM Tris, open symbols represent growth in CA-MHB 100 mM Tris + 44 mM NaHCO₃, solid lines represent no drug control, dashed lines represent exposure to CFZ, and dotted lines represents exposure to OXA. Drug concentrations for PB 043-043 were 16 μg/ml for CFZ and 32 μg/ml for OXA; drug concentrations for all other strains were 64 μg/ml for both CFZ and OXA.

FIG 4

Relationship between intrinsic MIC and responsiveness for CFZ and OXA. (A) CFZ. (B) OXA. Both graphs show low intrinsic MIC/responsive (black), high intrinsic MIC/nonresponsive (white), low intrinsic MIC/nonresponsive (light gray), and high intrinsic MIC/responsive (dark gray). Relationship between intrinsic MIC and responsiveness for CFZ is random (κ = 0.01; 95% confidence interval (CI) = −0.24 to 0.27; Kappa coefficient and McNemar’s test, P > 0.05). Relationship between intrinsic MIC and responsiveness to OXA is also random with significant discordance (κ = 0.02; 95% CI = −0.21 to 0.25; Kappa coefficient, P > 0.05; McNemar’s test *, P = 0.01).