Daptomycin Dose-Ranging Evaluation with Single-Dose versus Multidose Ceftriaxone Combinations against Streptococcus mitis/oralis in an Ex Vivo Simulated Endocarditis Vegetation Model

The viridans group streptococci (VGS) are a heterogeneous group of organisms which are important components of the normal human oral flora. Among the VGS, the Streptococcus mitis/oralis subgroup is one of the most common causes of infective endocarditis (IE).

ABSTRACT

The viridans group streptococci (VGS) are a heterogeneous group of organisms which are important components of the normal human oral flora. Among the VGS, the Streptococcus mitis/oralis subgroup is one of the most common causes of infective endocarditis (IE). Daptomycin (DAP) is a potential alternative therapeutic option for invasive S. mitis infections, given high rates of β-lactam resistance and vancomycin tolerance in such strains. However, the ability of these strains to rapidly evolve high-level and durable DAP resistance (DAP-R) is problematic. Recent data suggest that combination DAP-β-lactam therapy circumvents this issue. Human-simulated dose-escalating DAP-alone dose regimens (6, 8, 10, or 12 mg/kg/day times 4 days) versus DAP (6 mg/kg/day) plus ceftriaxone (CRO) (2 g once daily times 4 days or 0.5 g, single dose) were assessed against two prototypical DAP-susceptible (DAP-S) S. mitis/oralis strains (SF100 and 351), as measured by a pharmacokinetic/pharmacodynamic (PK/PD) model of simulated endocardial vegetations (SEVs). No DAP-alone regimen was effective, with regrowth of high-level DAP-R isolates observed for both strains over 96-h exposures. Combinations of DAP-CRO with either single- or multidose regimens yielded significant reductions in log₁₀ CFU/g amounts within SEVs for both strains (∼6 log₁₀ CFU/g) within 24 h. In addition, no DAP-R strains were detected in either DAP-CRO combination regimens over the 96-h exposure. In contrast to prior in vitro studies, no perturbations in two key cardiolipin biosynthetic genes (cdsA and pgsA) were identified in DAP-R SEV isolates emerging from strain 351, despite defective phospholipid production. The combination of DAP-CRO warrants further investigation for treatment of IE due to S. mitis/oralis.

TEXT

The viridans group streptococci (VGS) are important pathogens in invasive diseases, such as infective endocarditis (IE) and intra-abdominal infections (1–3). Among the VGS, the Streptococcus mitis/oralis subgroup is a leading cause of IE (3–5). Antibiotic therapy of invasive S. mitis/oralis infections is difficult, since a significant proportion of these strains are resistant to β-lactams, including the third-generation cephalosporins, and also exhibit in vitro vancomycin tolerance (3, 6–9). The mechanism of β-lactam-resistance in these isolates relates to a reduction of antibiotic affinity to target penicillin-binding proteins (10).

Daptomycin (DAP) has activity against S. mitis/oralis and therefore is a potential alternative for treating these infections (11). Although DAP has demonstrated excellent in vitro activity against S. mitis/oralis, several studies have reported that more than 25% of DAP-susceptible (DAP-S) strains rapidly develop high-level (MICs of >256 μg/ml) and durable DAP resistance (DAP-R) during in vitro exposures to DAP; a similar observation has been made in vivo in experimental IE (2, 12). The mechanism of DAP-R in S. mitis/oralis involves, at least in part, single nucleotide polymorphisms (SNPs) within cdsA (a locus encoding an enzyme responsible for catalyzing the first committed step in cardiolipin biosynthesis) and pgsA, encoding the enzyme responsible for the synthesis of phosphatidylglycerol (13–15). Such SNPs result in essentially no detectable cardiolipin production (14, 15). Cardiolipin serves a number of biologic functions, especially in cell division; it appears to be a pivotal molecule related to the interaction of DAP with target bacteria via mechanisms that are not fully elucidated but are likely to be associated with localization of anionic phospholipid microdomains in the bacterial cell membrane (16–18).

DAP is a cyclic lipopeptide antibiotic that has been broadly utilized for severe Gram-positive infections, including Staphylococcus aureus and enterococci (11). We have previously demonstrated that combinations of DAP plus ceftaroline (CPT) or ceftriaxone (CRO) are synergistic both in vitro and in the simulated endocardial vegetation (SEV) model against two well-characterized DAP-S S. mitis/oralis parental strains (the penicillin-resistant S. mitis/oralis strain, 351, and the penicillin-susceptible strain S. mitis/oralis SF100) (19). This model has the advantage of being able to precisely simulate human antibiotic concentrations to replicate therapeutic regimens used to treat specific infections (19–23). In this SEV model, not only did the combination of standard dose DAP (6 mg/kg) plus either CPT or CRO significantly reduce SEV CFU, but it also prevented the emergence of DAP-R over a 96-h exposure period in both strains (2, 24). In contrast, standard dose DAP-alone therapy was associated with inferior bactericidal effects and was unable to prevent the emergence of high-level DAP-R (19).

The aims of the current study utilizing the SEV model were (i) to determine whether escalating DAP-alone dose regimens (6, 8, 10, and 12 mg/kg), as currently commonly used in clinical practice (25, 26), could improve killing as well as prevent the emergence of DAP-R among S. mitis/oralis strains and (ii) to delineate whether short-term (single-dose) CRO combination therapy would be as effective as longer term (4 days) strategies as measured by these two outcome metrics.

RESULTS

In vitro susceptibility data.

To identify the impact of CRO on the activity of DAP against the S. mitis/oralis strains, we determined the MIC values for DAP and CRO alone and in combination. The MICs for each antimicrobial agent alone, as well as the DAP MICs in the presence of CRO at 0.5× MIC, are summarized in Table 1. MIC values for strain pairs are also presented in Table 1. Interestingly, the DAP MICs were reduced by 2-fold in both S. mitis/oralis 351 and SF100 strains in the presence of CRO at 0.5× its MIC. These data confirm the impact of CRO on DAP susceptibility. The lowering of the DAP MIC with addition of this β-lactam is consistent with our previous work using other cephalosporins against S. mitis/oralis strains (19).

TABLE 1.

MIC values for DAP, CRO, and DAP in the presence of CRO

Strain	MIC (μg/ml)
	DAP	CRO	DAP-CRO
351	0.5	8	0.25
SF100	0.5	0.125	0.25
Strain pairsa
351/D12T96 A1	0.5/>256
SF100/D12T96 B2	0.5/>256

^aFirst strain in each pair is the DAP-S parental strain, and the second is the DAP-R strain.

Pharmacokinetics.

To humanize the dosing of DAP and CRO, we targeted standard and high-dose DAP (6 and 8, 10, and 12 mg/kg/day, respectively) as well as high-dose (2 g every 24 h [q24h]) and single exposure (500 mg once on day 1) CRO. The average maximum concentration of drug in serum (C_max), half-life (t_1/2), and area under the concentration-time curve from 0 to 24 h (AUC_0–24) values for each dose regimen are listed in Table 2.

TABLE 2.

PK data for different doses in SEV model

Dose	Cmax (mg/liter)	AUC0–24 (mg · h/liter)	t1/2 (h)
DAP 6 mg/kg	93.27 ± 0.51	1,029.19 ± 43.73	7.89 ± 0.19 h
DAP 8 mg/kg	124.82 ± 0.72	1,404.865 ± 7.82	7.93 ± 0.03
DAP 10 mg/kg	143.98 ± 0.35	1,759.41 ± 59.90	8.21 ± 0.11
DAP 12 mg/kg	186 ± 3.32	2,251.99 ± 80.31	7.98 ± 0.08
CRO 2 g/kg	278.9 ± 8.41	3,346.35 ± 108.52	7.88 ± 0.01
CRO 0.5 g/kg	66.75 ± 2.41	756.27 ± 7.45	7.86 ± 0.08

A high-performance liquid chromatography (HPLC) assay demonstrated intraday coefficients of variation of between 0.48% and 5.13% for all DAP standards. Bioassay tests for CRO levels demonstrated intraday coefficients of variation of between 0.27% and 7.22% for CRO standards. Our data confirm the ability to achieve the targeted pharmacokinetics of DAP and CRO in the SEV model.

Pharmacodynamics.

To assess the impact of CRO on the efficacy of DAP as well as the suppression of DAP resistance, we evaluated DAP alone and DAP plus CRO in combination, including a single CRO dose on day one. For strain 351 (Fig. 1), the combination of DAP 6 mg/kg plus CRO (2 g once daily) reduced viability down to the detection limit within 8 h (−Δ 6.046 log₁₀ CFU/g), while the combination of DAP 6 mg/kg plus CRO (single dose at 0.5 g simulation) reached the detection limit in 24 h (−Δ 6.06 log₁₀ CFU/g). CRO alone (2 g) killed to near-detection limits at 32 h (−Δ 5.86 log₁₀ CFU/g), with some regrowth of 0.5 log₁₀ CFU/g at 96 h; however, it never reached the preset detection limit of 2 log₁₀ CFU/g. DAP alone at 6 mg/kg caused a modest CFU reduction within 8 h of exposure (−Δ 2 log₁₀ CFU/g), with significant regrowth up to the initial inoculum within 72 h. DAP 8 to 12 mg/kg simulations demonstrated initial bactericidal activity but with significant regrowth to a steady-state log₁₀ CFU/g of between ∼6 and 7.

FIG 1.

Comparison of daptomycin activity with and without CRO in various models in S. mitis/oralis 351. There was an 18-h delay in reaching the detection limit for DAP 6 mg/kg-CRO 0.5 g (single dose) in comparison to that for DAP 6 mg/kg-CRO 2 g.

Both combination regimens of DAP 6 mg/kg plus CRO (2 g and 0.5 g) were superior to all DAP-alone regimens (P < 0.001) when tested with strain 351. These combination regimens were not superior to CRO-alone regimens, despite the MIC of this strain for CRO being 8 mg/liter (P > 0.05). However, the time to reduce the CFU per gram to the detection limit was delayed in comparison to that for DAP-CRO combination regimens.

For strain SF100 (Fig. 2), the SEV killing data were very similar to those for strain 351: (i) none of the DAP-alone regimens had bactericidal activity that reached the assay detection limits; (ii) both DAP plus CRO regimens were highly effective in reducing SEV counts, as was the CRO-alone regimen; and (iii) the DAP plus CRO regimens reached the detection limits more rapidly than the CRO-alone regimens.

FIG 2.

Comparison of daptomycin activity with and without CRO in various models in S. mitis/oralis SF100. DAP 6 mg/kg-CRO 0.5 g (single dose) and DAP 6 mg/kg-CRO 2 g reached the detection limit at the same time, but slight regrowth was observed in DAP 6 mg/kg-CRO 0.5 g (single dose) after 48 h.

For both strains, at 96 h of the SEV model, DAP MICs for SEV isolates for all DAP-alone regimens were 16 to 64 mg/liter (i.e., MIC increases of ≥32-fold compared to that for the parental strains). In contrast, among the DAP 6-mg/kg plus CRO combination regiments, no DAP-R isolates were observed after 96-h exposures. No DAP-R isolates were found in CRO-alone-exposed strains after 96-h exposures. CRO 2-g q12h monotherapy showed activity against both strains (reached CFU/ml detection limit), with no CRO emergence of resistance. The CRO 0.5-g single-dose day one monotherapy regimen did not reduce CFU per milliliter and resulted in the emergence of CRO resistance (MIC of >64 mg/liter) in strain SF100. No CRO resistance was observed in strain 351 for the CRO 0.5-g single-dose regimen.

Cell surface charge.

Prior studies in selected S. aureus and enterococcal strains have shown an association between alterations in the cell surface charge toward more electropositivity and the DAP-R phenotype (27–29). In this context, we investigated the cell surface charges for both current S. mitis/oralis strains.

A representative 351 DAP-R SEV variant (D12T96 A1) demonstrated a significantly enhanced positive surface charge compared to its 351 DAP-S parental strain (P < 0.05), a phenotype observed before with this strain following derivation by serial in vitro DAP passage (28) (Table 3). In contrast, there was no difference in relative surface charge when comparing SF100 DAP-R (D12T96 B2) and its DAP-S parental strain.

TABLE 3.

Relative surface charge and CM fluidity of study strains

Strain pairsa	Surface charge (%)b	CM fluidity (PI value)
351/D12T96 A1	59 ± 9/30 ± 4c	0.31 ± 0.02/0.29 ± 0.03c
SF100/D12T96 B2	24 ± 4/25 ± 5	0.29 ± 0.00/0.30 ± 0.01

^aFirst strain in each pair is the DAP-S parental strain, and the second one is the DAP-R strain.

^bPercentage of bacterial-bound Cyt C calculated from the proportion of unbound Cyt C remaining in the supernatant.

^cP value of <0.05 in 351 DAP-R (D12T96 A1) versus 351 DAP-S S. mitis/oralis parental strain.

Cytoplasmic membrane order (fluidity/rigidity).

It was previously shown that there appears to be an optimal degree of cytoplasmic membrane (CM) order for maximal interaction for distinct CM-targeting cationic antimicrobial peptides, including calcium-DAP (28, 30–32). The 351 DAP-R (D12T96 A1) strain had a significantly increased CM fluidity versus its DAP-S parental strain, mirroring prior reported data for in vitro-derived DAP-R variants of this strain (Table 3) (28). However, there was no substantive difference of CM order in SF100 DAP-R (D12T96 B2) versus its parental strain.

CM phospholipid content.

The CM phospholipid (PL) content is intimately involved in the mechanism of action and resistance of DAP (12, 28, 30, 32–35). As seen before (28), the 351 DAP-R (D12T96 A1) strain exhibited an apparent disappearance of both phosphatidylglycerol and cardiolipin, accompanied by a clear increase in phosphatidic acid content (the initial substrate for cardiolipin biosynthesis), compared to its DAP-S parental strain (Table 4; Fig. 3). In contrast, DAP-R SF100 (D12T96 B2) showed rather similar PL profiles to its DAP-S parental strain.

TABLE 4.

CM phospholipid composition of study strains

Strain pairsa	% of total PLs (mean ± SD)b
	PG	CL	PA
351/D12T96 A1	11 ± 15/0 ± 0c	85 ± 16/0 ± 0c	4 ± 2/100 ± 0c
SF100/D12T96 B2	10 ± 6/20 ± 8	47 ± 5/69 ± 3	15 ± 7/6 ± 4

^aFirst strain in each pair is the DAP-S parental strain, and the second one is the DAP-R strain.

^bPG, phosphatidylglycerol; CL, cardiolipin; PA, phosphatidic acid.

^cP value of <0.001 351 DAP-R (D12T96 A1) strain versus 351 DAP-S S. mitis/oralis parental strain.

FIG 3.

Phospholipid spots shown in 2D-TLC plates were developed with CuSO₄ (100 mg/ml) spray containing 8% phosphoric acid and heated at 180°C. PG, phosphatidylglycerol; CL, cardiolipin; PA, phosphatidic acid; UI, unidentified; phospholipid, glycolipid contaminant seen in all S. mitis/oralis strains (previously detailed [29]).

Role of cdsA and pgsA.

Daptomycin resistance was previously linked to mutations in cdsA and pgsA, two genes that mediate the synthesis of membrane phospholipids (28). For that reason, we compared the sequences of these genes in the DAP-S parental versus DAP-R variants of 351 and SF100. No mutations were present in the open reading frames of either cdsA or pgsA for either of the two DAP-R strains. For the DAP-R strain 351 variant, this indicated that the observed phospholipid changes are likely to be due to other genetic mechanisms impacting the cardiolipin biosynthetic pathway.

DISCUSSION

Several interesting observations emerged in these studies. First, our data demonstrate that addition of CRO to DAP improves the bactericidal activity of DAP both ex vivo and in vivo against two prototypical S. mitis/oralis strains and also prevents the emergence of DAP-R during prolonged DAP exposures in the SEV model. Of note, even a single exposure to low-dose CRO enhanced the bactericidal activity of DAP in this model. These findings open the possibility to use this antibiotic combination for treating S. mitis/oralis infective endocarditis in clinical settings.

Second, we studied a range of DAP-alone simulated dose strategies (6, 8, 10, and 12 mg/kg/day) to approximate current clinical regimens for invasive Gram-positive infections (19) in hopes of overriding the propensity of S. mitis/oralis strains to develop DAP-R (2). Our goal was to significantly enhance DAP penetration of the SEVs and thus maximize the amount of drug reaching intra-SEV S. mitis/oralis colonies. We previously demonstrated that a 6-mg/kg/day human-simulated dose equivalent is able to achieve more than 60% (48.2/79.5 μg/ml) DAP penetration into the SEV (36). This would put the concentration of DAP (48 μg/ml) well above the MICs of both parenteral S. mitis/oralis strains used in the present paper. However, despite significant increases in C_max and AUC_0–24 accompanying escalating DAP-alone dose strategies (6, 8, 10, and 12 mg/kg simulations in the SEV model), such regimens only modestly reduced bacterial burdens in SEV and were not able to prevent the evolution of high-level DAP-R in SEVs infected with either study strains. The same findings were observed in the rabbit model of S. mitis experimental endocarditis using 6 and 10 mg/kg of daptomycin (2).

Third, since DAP-R was previously linked to mutations in cdsA and/or pgsA (14, 28), we compared the sequences of these genes in the parental versus their respective DAP-R variants derived from SEVs after 96 h of DAP-alone exposures. Of note, and in contrast to data we have previously published for DAP-R variants of these two study strains following serial in vitro or in vivo DAP exposures in experimental IE (3), there were no SNPs found in either open reading frame (ORF) when comparing parental versus DAP-R variants. These data suggest that the evolution of DAP-R in a distinct environmental milieu (in vitro versus ex vivo versus in vivo) follows different genetic and/or metabolic pathway perturbations and is influenced by the unique characteristics of each microenvironment. For example, SEVs feature platelet-fibrin matrices bathed in artificial growth medium, while in vivo vegetations contain similar matrices in the presence of a variety of host defense cells (e.g., polymorphonuclear leukocytes [PMNs]) and molecules (e.g., antibody, complement, and host defense cationic peptides).

Fourth, for strain 351, its ex vivo-derived DAP-R variant featured an apparent disappearance of phosphatidylglycerol (PG) and cardiolipin (CL) accompanied by increased phosphatidic acid (PA) content; this pattern of phospholipid content is identical to that previously seen with the serial in vitro passage of organisms exposed to DAP alone (28). However, the SNPs in cdsA that accompany this phenotype in the in vitro-derived DAP-R 351 variant were not observed among SEV-derived DAP-R 351 variants. In aggregate, these findings indicate the above-described changes in CM phospholipids and subsequent DAP-R involve different genetic pathways. Such mutations may alter the expression of cdsA or pgsA or, alternatively, may affect membrane biogenesis via other pathways. Whole-genome sequencing comparative analyses are in progress to further adjudicate this issue.

One additional explanation for PG-CL disappearance in DAP-R strain 351, in the absence of perturbations in CL biosynthetic genes, could be an upregulation of release of these phospholipids extracellularly. This paradigm has been recently proposed by the Edwards’ laboratory involving the interaction of Staphylococcus aureus, enterococci, and VGS of the S. mitis group (i.e., S. gordonii) with DAP (37, 38). They documented the capacity of these organisms to release such phospholipids extracellularly in an active process, which could then bind to and inactivate DAP in a “phospholipid decoy” mechanism. However, this mechanism would tend to bind DAP extracellularly and prevent its binding to target CMs; this, however, is the opposite of what has been previously observed in DAP-R S. mitis/oralis strains (29).

Fifth, in contrast to strain 351, no substantive changes in PL profiles were found in DAP-R SF100 variants isolated from SEVs. These outcomes support the notion that the genetic basis of the PG-CL disappearance during emergence of DAP-R in strain 351 under ex vivo conditions involves loci and/or pathways outside the standard CL biosynthetic schema. Moreover, these data sets point to a distinct and strain-specific mechanism(s) of DAP-R emerging ex vivo in variants of SF100, not involving CL biosynthesis.

Sixth, one prevailing concept that has been put forward to explain, at least in part, the DAP-R phenotype in Gram-positive pathogens, including S. mitis/oralis, has been alterations in relative cell surface positive charge, creating a “charge-repulsion” milieu against calcium-DAP and other cationic peptides. Interestingly, and as seen before (28), the in vitro-derived DAP-R 351 variant showed a significant enhancement of positive surface charge versus its DAP-S parental strain. However, this phenotypic perturbation was not seen in the DAP-R SF100 variant. These data do not support charge repulsion as a sole or global explanation for the DAP-R phenotype in these strain pairs and further support the strain specificity of DAP-R mechanisms in S. mitis/oralis strains.

Lastly, in general, alterations in CM fluidity have been closely related to resistance to killing by calcium-DAP and other cationic peptides among DAP-R Gram-positive pathogens (1–8). These phenotypes are believed to link to an altered ability of DAP to interact with and/or insert within target CMs at extremes of CM order. It is presumed that there are certain CM order set points for the optimal interaction of target CMs with cationic peptides such as calcium-DAP, rendering strains more or less susceptible to such agents (1–8). The DAP-R 351 (but not the DAP-R SF100) variant demonstrated significant increases in CM fluidity compared to that of its DAP-S parental strain. The enhancement of CM fluidity in the 351 DAP-R strain may partly explain the mechanism of DAP-R evolving within SEVs ex vivo, but this phenotype is clearly strain specific.

In conclusion, the mechanism(s) of DAP-R in S. mitis/oralis emerging when the organism is growing within a platelet-fibrin ex vivo matrix is most likely to be multifactorial, strain specific, and distinct from those evolving during in vitro or in vivo passage, although there appear to be definite overlapping scenarios. Furthermore, it is highly probable that other compensatory adaptive metabolic changes are critically important in dictating DAP-R in S. mitis/oralis. Importantly, CM phenotypic perturbations may well be identifiable in specific ex vivo-derived DAP-R S. mitis/oralis strains (e.g., DAP-R 351) and correlate with signature phenotypic consequences (i.e., changes in PLs, relative surface charge, and/or CM fluidity). We recognize that one major limitation in our study is only having investigated two S. mitis/oralis strains, which restricts the generalizability of our data. In this regard, additional investigations that include more S. mitis/oralis strains and other β-lactams with various exposure regimens are in- progress in our laboratories. In addition, our SEV models were terminated after only 96 h. Thus, longer antibiotic exposure times in this SEV model should be explored in future studies to more closely mirror the in vivo circumstance in IE.

Notwithstanding these limitations, we have now demonstrated that S. mitis/oralis strains are apparently capable of developing DAP-R by more than one pathway. Moreover, given the high level of homology of S. mitis/oralis to other pathogenic bacterial species, it is likely that these findings are broadly applicable. In addition, the 96-h SEV model can clearly present the trend of CFU per milliliter reduction or emergence of resistance in different regimens; however, as noted above, longer antibiotic exposures with this model are warranted to determine the impact of these dosing regimens on the emergence of resistance.

MATERIALS AND METHODS

Bacterial isolates.

Strains 351 and SF100 were identified to the species level previously as S. oralis based on whole-genome sequencing. The two respective DAP-R variants were obtained from the current SEV model following 96 h of high-dose DAP (12 mg/kg/day) simulated treatment Table 1.

Antimicrobial agents and media.

DAP and CRO were obtained from Merck & Co., Inc. (Whitehouse Station, NJ). Mueller-Hinton broth II (MHB; Difco, Detroit, MI) with 50 mg/liter of calcium and 12.5 mg/liter magnesium was used for susceptibility testing. In addition, for in vitro susceptibility testing, the medium was supplemented with 5% lysed horse blood.

Susceptibility testing.

MICs were determined in duplicates using the broth microdilution method according to the Clinical and Laboratory Standards Institute (CLSI) guidelines for viridans group streptococci (39). Combination MIC values for DAP were determined by adding in CRO at 0.5× MIC to assess the ability of CRO to reduce the DAP MIC. Daptomycin MIC fold reductions from baseline were calculated by dividing single drug DAP MICs by DAP MICs in the presence of CRO.

Surface charge.

The binding profile of cytochrome c, a highly polycationic molecule (Cyt C), is used to mimic the interactions of cationic antimicrobial peptides, such as calcium-DAP, with charged bacterial surfaces (12, 28, 30, 32–34). This assay was used to measure the relative positive cell surface charges of our two study strains as well as their DAP-R variants emerging from the SEV model described below (12, 28, 33). Briefly, each strain was grown in brain heart infusion (BHI) medium to an optical density at 600 nm (OD₆₀₀) of ∼1.0 and then incubated with 0.5 mg/ml of Cyt C for 10 min; the residual amount of Cyt C remaining in the bacterial supernatant was then determined spectrophotometrically (12, 28, 32, 33). An increase in the quantity of Cyt C binding (i.e., less cation in the supernate) correlates to a relatively less positively charge bacterial surface (12, 28, 30, 32, 33). These data are expressed as the mean (± standard deviation [SD]) of bound Cyt C and represent data from a minimum of four independent experiments for both strain pairs on distinct days.

CM fluidity/rigidity.

Strains were grown to stationary phase at 37°C, pelleted by centrifugation, and then washed with phosphate-buffered saline (PBS). A whole-cell suspension of the study strains was assayed to quantify the degree of CM order (fluidity/rigidity) at an OD₆₀₀ of 1.0 (∼10⁸ CFU ml⁻¹). The assay was performed by using and measuring the three-dimensional CM orientation of the fluorescent probe, 1,6-diphenyl-1,3,5-hexatriene (DPH) by polarizing spectrofluorimetry (excitation and emission wavelengths of 360 nm and 426 nm, respectively). The descriptive method for quantifying DPH integration into CMs and the calculations of the degree of fluorescence polarization (polarization index [PI]) are described elsewhere (28, 30, 32). An inverse relationship occurs between PI values and CM fluidity (i.e., lower PI equates to a greater extent of CM fluidity) (28, 30, 32). These data included a minimum of three independent experiments on different days.

CM PL composition.

The lipid extraction procedures have been previously described elsewhere (28, 30, 32, 34). The S. mitis/oralis strain’s major PLs, PG, CL, and PA, were separated by utilizing two-dimensional thin-layer chromatography (2D-TLC) in a unique solvent system as described before (12, 28, 30, 33, 34). All TLC plates showed a copurified glycolipid (identified as digalactosyldiacylglycerol by mass spectrometry [28]) found in all our study S. mitis/oralis strains to date. The well-separated PL spots were scraped from the TLC plates and digested at 180°C for 3 h with 0.3 ml 70% perchloric acid, and absorbance profiles were measured spectrophotometrically at OD₆₆₀ to quantify the phosphorus content. The TLC plates were expose to iodine vapors and sprayed with cupric sulfate (CuSO₄; 100 mg/ml) containing 8% (vol/vol) phosphoric acid and heated at 180°C (12, 28, 30, 31, 34) for accurate identification and confirmation of all spots on the TLC plate.

cdsA and pgsA amplification and sequence analysis.

Genomic DNA was used for PCR amplification of cdsA and pgsA, and the amplicons were isolated from the designated S. mitis strains as previously described (28). Vent DNA polymerase was used for standard PCR amplification. Generated PCR products encompassing the entire open reading frame for the designated gene were purified using a Qiagen PCR purification kit and sequenced by Sequetech (Mountain View, CA). Sequence reads were assembled and aligned using CLC Sequence Viewer (CLC Bio, Aarhus, Denmark).

Ex vivo PK/PD model.

A well-characterized, simulated endocardial vegetation (SEV) pharmacokinetic/pharmacodynamic (PK/PD) model was used for all antibiotic experiments (40, 41). SEVs were prepared as previously described (42), by mixing cryoprecipitate as the source of fibrin, a platelet suspension (American Red Cross), aprotinin (Sigma-Aldrich), and an organism suspension (final inoculum, 10⁹ CFU/0.5 g). This methodology results in SEVs that contain approximately 3 to 3.5 g/dl of albumin and 6.8 to 7.4 g/dl of total protein (42). The resultant SEV mixture was then clotted using bovine thrombin in siliconized microcentrifuge tubes (Pfizer, New York, NY, USA). SEVs were removed from the tubes and suspended in the model via a sterile monofilament line. Fresh medium was continuously added and removed from the model along with the drug via a peristaltic pump (Masterflex; Cole-Parmer Instrument Company, Chicago, IL, USA) at a rate which was set to simulate the half-lives of the antibiotics of interest. For the SEV model, due to the dependency of DAP on calcium for antimicrobial activity and calcium loss from the medium due to calcium binding to albumin, the medium was supplemented to a concentration of 75 mg/liter of calcium as previously described (23).

At assigned times, SEVs were removed from the chamber model, weighed, homogenized, and quantitatively cultured to determine S. mitis/oralis counts. Tryptic soy agar supplemented with 5% sheep’s blood (Difco) was used for quantitative subculturing from the SEV models. Plates were incubated in an anaerobic chamber for 18 to 24 h at 37°C before performing colony counts. Bacterial counts were graphed versus time over 96 h to evaluate bactericidal activity of single and combination regimens (Fig. 1 and 2). The limit of detection for these samples is 2 log₁₀ CFU/g. Results were plotted (log₁₀ CFU/g versus time [h]) for determining antibacterial activity.

DAP was administered once daily via an injection port. The simulated DAP regimens with a targeted t_1/2 of 8 h were 6, 8, 10, and 12 mg/kg/day, with peak concentrations of 93.9, 123.3, 141.1, and 183.7 mg/liter, respectively. The DAP regimens were tested alone and in combination with CRO 2 g (once daily over 96 h) or 0.5 g (single dose on day one), at corresponding CRO peaks of 257 and 64.25 mg/liter, respectively. All models were performed in duplicates to ensure reproducibility.

PD analyses.

For PD evaluations, two SEVs (clots) were aseptically removed from each model (total of 4 samples) at 0-, 4-, 8-, 24-, 32-, 48-, 72-, and 96-h time points. After weighing, each SEV was homogenized with trypsin, serially diluted with cold normal saline, and plated. To minimize the antibiotic carryover, the SEV samples were diluted appropriately before plating. Bactericidal and bacteriostatic activities for DAP alone or DAP-CRO combinations were defined as ≥3 log₁₀ and ≤3 log₁₀ CFU/g reductions in counts compared to that for untreated control SEVs. The impact of DAP-CRO combinations was defined as “enhanced” if the combination reduced SEV counts by ≥2 log₁₀ CFU/g versus the DAP-alone or CRO-alone regimens. Reductions in bacterial counts by 1 log₁₀ to 2 log₁₀ CFU/g compared with the most active single agent were considered “indifferent.”

PK analyses.

Sampling for PK assays was performed by removing 1 ml of medium (through the injection port) from each of the two parallel SEV models at 0, 2, 4, 6, 8, 24, 48, 72, and 96 h in duplicates as previously described (43). DAP PK evaluations were confirmed by a validated high-performance liquid chromatography assay that conforms to the guidelines set forth by the College of American Pathologists (44). A bioassay method was applied for assessing concentrations of CRO using Kocuria rhizophila (formerly known as Micrococcus luteus) ATCC 9341. For combination regimens, CRO concentrations were determined with Escherichia coli ATCC 25922 as the test organism. All assays and standards were performed in duplicates.

Sterile 0.25-in. paper disks were placed on agar (antibiotic medium 11 Difco agar plate containing 0.5 McFarland suspension of this organism). Each blank paper disk was filled with 20 μl of either a standard CRO concentration or the unknown chamber model samples from different time points (SEV PK samples). A standard curve was then created using inhibition zone size versus known concentrations; the inhibition zone sizes at each sample time point were plotted against this standard curve for the calculation of PK parameters. All samples were examined in duplicates. The CRO peak, AUC_0–24, and half-life metrics were determined by the trapezoidal method, utilizing standard pharmacokinetic modeling software (PK analyst version 1.1; MicroMath Scientific Software, Salt Lake City, UT).

Emergence of DAP-R in the SEV model.

The emergence of DAP-R over 96-h DAP exposures within the SEV model was evaluated by determining DAP MICs of isolates recovered from 96-h SEVs when parallel plated on antibiotic-free and DAP-containing plates (3× the baseline MIC of the parental DAP-S strain). The DAP MICs of colonies detected on DAP-containing plates were then formally determined using the broth microdilution technique via CLSI guidelines (39).

Statistical analysis.

The two-tailed Student’s t test was used for statistical analysis of quantitative data related to cell surface charge, fluidity, and phospholipid content. One-way analysis of variance (ANOVA) with Tukey’s post hoc test was applied to compare PD outcomes in the SEV model regarding DAP, CRO, and DAP-CRO combination impacts upon log₁₀ CFU per gram comparative results. P values of ≤0.05 were considered significant.