Efficacy of Telavancin Alone and in Combination with Ampicillin in a Rat Model of Enterococcus faecalis Endocarditis

ABSTRACT

We first assessed telavancin (TLV) pharmacokinetics in rats after a single subcutaneous dose of 35 mg/kg of body weight. The pharmacokinetic data were used to predict a TLV dose that simulates human exposure, and the efficacy of TLV was then evaluated using a TLV dose of 21 mg/kg every 12 h against Enterococcus faecalis OG1RF (TLV MIC of 0.06 μg/ml) in a rat endocarditis model with an indwelling catheter. Therapy was given for 3 days with TLV, daptomycin (DAP), or ampicillin (AMP) monotherapy and with combinations of TLV plus AMP, AMP plus gentamicin (GEN), and AMP plus ceftriaxone (CRO); rats were sacrificed 24 h after the last dose. Antibiotics were given to simulate clinically relevant concentrations or as used in other studies. TLV treatment resulted in a significant decrease in bacterial burden (CFU per gram) in vegetations from 6.0 log₁₀ at time 0 to 3.1 log₁₀ after 3 days of therapy. Bacterial burdens in vegetations were also significantly lower in the TLV-treated rats than in the AMP (P = 0.0009)- and AMP-plus-GEN (P = 0.035)-treated rats but were not significantly different from that of the AMP-plus-CRO-treated rats. Bacterial burdens from vegetations in TLV monotherapy and TLV-plus-AMP-and-DAP groups were similar to each other (P ≥ 0.05). Our data suggest that further study of TLV as a therapeutic alternative for deep-seated infections caused by vancomycin-susceptible E. faecalis is warranted.

INTRODUCTION

Enterococci are commensals of the normal human intestinal flora and a known cause of bacterial infective endocarditis (IE). In multiple studies, enterococci have been identified as the third most frequent cause of endocarditis (1–3), with Enterococcus faecalis causing the majority of these infections. A trend toward increasing drug resistance among traditional anti-enterococcal antibiotics (i.e., penicillin, ampicillin [AMP], vancomycin [VAN], and aminoglycosides) poses significant challenges for the treatment of severe infections caused by enterococci, in particular Enterococcus faecium (4, 5). Although most E. faecalis organisms remain susceptible to penicillin and AMP, these agents are usually combined with aminoglycosides or ceftriaxone (CRO) to achieve a bactericidal effect and higher cure rates. Telavancin (TLV) is among the newer agents with in vitro activity against glycopeptide-susceptible E. faecalis. TLV has FDA approvals for complicated skin and skin structure infections (6, 7) and hospital-acquired and ventilator-associated bacterial pneumonia caused by susceptible isolates of Staphylococcus aureus when alternative treatments are not suitable (8). The mechanism of bacterial killing by TLV includes inhibition of bacterial cell wall synthesis by interfering with peptidoglycan synthesis and disruption of membrane homeostasis (6, 7, 9). In vitro, TLV is active against a variety of Gram-positive organisms, including Enterococcus spp., Staphylococcus spp., and Streptococcus spp. (9–13). In vivo efficacy of TLV has been evaluated in several models of staphylococcal infections, including soft-tissue infection (14, 15), murine pneumonia (using a VAN-heteroresistant S. aureus [hVISA] and VAN-intermediate S. aureus [VISA] strains) (16), and rabbit infective endocarditis (methicillin-resistant S. aureus [MRSA] and VISA/glycopeptide-intermediate S. aureus [GISA]) (17, 18). However, the in vivo efficacy of TLV has not been reported against E. faecalis using an endocarditis model. The main goal of this study was to assess the therapeutic efficacy of TLV against a well-characterized strain of E. faecalis (OG1RF, AMP, and VAN susceptible) (19, 20) using the stringent rat infective endocarditis model (21, 22). We first performed a pharmacokinetic study of TLV and then compared the efficacy of this compound as monotherapy and combined with AMP against common therapies used in the treatment of E. faecalis endocarditis, namely, AMP plus CRO or gentamicin (GEN) and daptomycin (DAP). Our results found that TLV (alone and combined with AMP) was highly efficacious in reducing bacterial burdens in infected endocarditis vegetations.

RESULTS

MICs of TLV, comparators, and time-kill experiments.

The MIC of TLV against E. faecalis OG1RF was 0.06 μg/ml, as determined by Sensititre panel CMP4STA (lot number 841738). DAP, AMP, VAN, GEN, and CRO MICs were 2, 0.5, 2, 16, and 256 μg/ml, respectively, and E. faecalis OG1RF was considered susceptible for TLV, DAP, VAN, GEN, and AMP by following CLSI guidelines (23).

Time-kill assays of TLV and DAP alone showed no decrease in the viable bacterial count (CFU per milliliter) from the starting inoculum (time zero) when used at 1× the MICs; 5× the MICs produced a reduction of circa 2 log₁₀ only with DAP. TLV and DAP were not bactericidal when used even at 5× the MICs with AMP (1× MIC) and showed a decrease of ca. 1.9 ± 0.024 log₁₀ and 2.7 ± 0.87 log₁₀, respectively, in the viable bacterial count (CFU per milliliter) from the starting inoculum concentration. TLV showed additive activity when combined with AMP in time-kill assays compared to TLV. However, the DAP (at 1× and 5× the MIC)-plus-AMP combination was not bactericidal or synergistic in in vitro time-kill experiments (reduction of <3 log₁₀ CFU/ml at 24 h); therefore, we did not pursue the DAP-plus-AMP combination in the rat endocarditis model against E. faecalis OG1RF.

Pharmacokinetic analysis.

The serum concentration-time profile observed after a single subcutaneous (s.c.) injection (35 mg/kg of body weight) of TLV is shown in Fig. 1. The best-fit pharmacokinetic parameters are summarized in Table 1. The observed area under the concentration-time curve from 0 to infinity (AUC_0–∞) was 567.26 mg · h/liter (interquartile range [IQR], 514.56 to 687.23 mg · h/liter) (the AUC_0–∞ reported in healthy humans is 747 ± 129 μg · h/ml [24], ∼ 20% higher than the value we observed in rats). As a recent study (25) has indicated that TLV exhibits linear pharmacokinetic disposition in rats, we used our PK analysis to project the desired exposure by increasing the dose of TLV by 20% to approximate the dose of 10 mg/kg used in healthy volunteers. Thus, we used 42 mg/kg daily divided in two doses (21 mg/kg every 12 h [q12h]) subcutaneously for the therapeutic protocol.

FIG 1.

Pharmacokinetic profiles after subcutaneous administration to jugular vein cannulated rats. Means ± standard deviations are shown.

TABLE 1.

Pharmacokinetic parameters of telavancin

Parametera	Median (IQR) value
Cmax (μg/ml)	86.92 (84.85–87.55)
Cmin (μg/ml)	0.400 (0.211–0.637)
ka (h−1)	0.488 (0.405–0.650)
kel (h−1)	0.323 (0.293–0.405)
t1/2 (h)	2.15 (1.79–2.39)
V (liter/kg)	0.156 (0.145–0.195)
AUC0−∞ (mg · h/liter)	567.26 (514.56–687.23)

^aC_max, maximum concentration; C_min, minimum concentration; IQR, interquartile range; k_a, absorption rate constant; k_el, elimination rate constant; t_1/2, elimination half-life; V, volume of distribution; AUC_0−∞, area under the concentration-time curve.

ID₉₀ and efficacy assessment of various antibiotics.

The 90% infectious dose (ID₉₀) of E. faecalis OG1RF in our model was 1.7 × 10⁵ CFU/g. As mentioned above, we used ∼10× the ID₉₀ to infect cardiac valves. Figure 2A and B and Table 2 show the efficacy of various antibiotics (alone and in combination). The mean log₁₀ ± standard deviation (SD) count in the untreated, baseline group (time [t] = 0) was 6.0 ± 0.6 CFU/g (Fig. 2A). TLV, DAP, and VAN monotherapy significantly reduced the bacterial counts in vegetations, with 3.1 ± 1.3, 2.1 ± 1, and 3.6 ± 0.7 mean log₁₀ CFU/g counts, respectively (P ≤ 0.0001 for each antibiotic versus the value at baseline [t = 0]) (Fig. 2A), with no significant difference in bacterial counts between TLV- and DAP-treated animals at the end of therapy (P = 0.1841). Counts (log₁₀ CFU/gram) for both groups are also presented in Table 2. VAN and TLV monotherapy showed nonsignificant differences, while DAP showed a significant reduction in bacterial counts in vegetation (P = 0.0102) (Table 2) compared to VAN. In contrast, AMP monotherapy achieved a slight, but not statistically significant, reduction in bacterial counts compared to the control animals at t = 0 (P = 0.7681) (Fig. 2A). One animal each in the TLV and DAP monotherapy groups were found to have sterile vegetation (Fig. 2A).

FIG 2.

Efficacy of antibiotic therapy for E. faecalis OG1RF in a rat IE model. Rats were treated for 3 days starting 24 h after inoculation (t = 0) and sacrificed 24 h after the last dose. The P values (unpaired t test) shown are between infected, untreated, baseline (t = 0) rats versus infected, treated rats. The data were log transformed, and unpaired t test was performed to obtain P values. Mean values of vegetation (log₁₀ CFU/g) are shown in Table 2, which also shows comparisons between treatment groups. A separate untreated, infected control group sacrificed with a therapy group is also shown to show the end log₁₀ CFU/g value. (A) Monotherapy. Animals were infected with OG1RF and treated with telavancin (TLV), daptomycin (DAP), vancomycin (VAN), and ampicillin (AMP). One animal each died after AMP (60 h posttreatment) and TLV (48 h post treatment) treatment and had 4.3 × 10⁴ and 2 × 10⁴ log₁₀ CFU/g, respectively. Data are included in the graphs and statistical analysis. (B) Combination therapy. Animals were infected with OG1RF and treated with TLV plus AMP, AMP plus CRO, and AMP plus GEN. One animal died in the AMP-plus-CRO treatment group after 60 h and had 3.1 × 10² log₁₀ CFU/g. Data are included in the graphs and were used for statistical analysis.

TABLE 2.

Animals inoculated with E. faecalis OG1RF and treated with antibiotics

Efficacy of antibiotics by treatment groupa				P valueb
Group 1		Group 2
Drug regimen	CFU/g	Drug regimen	CFU/g
TLV 21c	3.1 ± 1.3	AMP 100 + CRO 10c	4.3 ± 1.2	0.1182
		AMP 100 + GEN 2.5	4.8 ± 1.3	0.0353
		AMP 100c	5.9 ± 0.8	0.0009
		VAN 120	3.6 ± 0.7	0.3282
DAP 45.3	2.1 ± 1	TLV 21 + AMP 100	2.5 ± 0.8	0.8111
		TLV 21	3.1 ± 1.3	0.1841
		AMP 100 + CRO 10	4.3 ± 1.2	0.0033
		AMP 100 + GEN 2.5	4.8 ± 1.3	0.0006
		AMP 100	5.9 ± 0.8	<0.0001
		VAN 120	3.6 ± 0.7	0.0102
TLV 21 + AMP 100	2.5 ± 0.8	TLV 21	3.1 ± 1.3	0.1042
		AMP 100 + CRO 10	4.3 ± 1.2	0.0085
		AMP 100 + GEN 2.5	4.8 ± 1.3	0.0022
		AMP 100	5.9 ± 0.8	<0.0001
		VAN 120	3.6 ± 0.7	0.0170
AMP 100 + CRO 10	4.3 ± 1.2	AMP 100 + GEN 2.5	4.8 ± 1.3	0.5383
		AMP 100	5.9 ± 0.8	0.02
		VAN 120	3.6 ± 0.7	0.3683
AMP 100 + GEN 2.5	4.8 ± 1.3	AMP 100	5.9 ± 0.8	0.08
		VAN 120	3.6 ± 0.7	0.1301

^aAll antibiotic doses are in milligrams per kilogram of body weight administered per day for 3 days. The following doses were used: TLV 21, 21 mg/kg s.c., q12h; AMP 100, 100 mg/kg i.m., q8h; GEN 2.5, 2.5 mg/kg i.m., q12h; CRO 10, 10 mg/kg s.c., q8h; DAP 45.3, 50 mg/kg s.c., q24h; VAN 120, 120 mg/kg, s.c., q12h. More efficacious antibiotics are listed first in each grouping. CFU/g indicates mean log₁₀ ± SD CFU/g in vegetation.

^bThe data were log transformed, and unpaired t test was performed to obtain P values.

^cOne animal died in each of the AMP and AMP-plus-CRO treatment groups after 60 h and had 4.3 × 10⁴ and 3.1 × 10² log₁₀ CFU/g, respectively, in aortic valve vegetation samples. One animal was found dead in the TLV treatment group after 48 h and had 2 × 10⁴ log₁₀ CFU/g in aortic valve vegetation. All of these data were included in the graphs or were used in statistical analysis.

Figure 2B and Table 2 show the results of combination therapies. All combinations produced a statistically significant decrease in bacterial counts in vegetations compared to the baseline controls (t = 0). The mean counts (log₁₀ ± SD CFU/gram) of vegetation were 2.5 ± 0.8, 4.8 ± 1, and 4.3 ± 1 for TLV plus AMP, AMP plus GEN, and AMP plus CRO, respectively (all P values were <0.05 versus results for the controls at t = 0) (Fig. 2B and Table 2). A statistically significant decrease in bacterial counts was detected with the TLV plus AMP group compared to the animals treated with AMP plus GEN (P = 0.0022) and AMP plus CRO (P = 0.0085).

Comparisons between monotherapy and combination regimens are shown in Table 2. Both TLV and DAP treatment regimens achieved a statistically significant greater reduction in bacterial CFU per gram than the AMP-plus-GEN regimen (P value of 0.0353 and 0.0006, respectively). DAP alone was also more efficacious in reducing bacterial burdens in vegetations than AMP plus CRO (P = 0.0033), while TLV reduced the CFU per gram by 1.1 ± 0.7 log₁₀ more than the AMP-plus-CRO combination (nonsignificant). Of note, high-dose TLV (42 mg/kg, s.c., q12h, i.e., 84 mg/kg/day) resulted in a count of 3.7 ± 0.4 mean log₁₀ CFU/g in vegetations and did not confer any additional benefit in eradication of bacteria from vegetations either alone or combined with AMP compared to the lower-dose TLV treatment regimen.

One animal died in each of the AMP and AMP-plus-CRO treatment groups after 60 h (received 2.5 days of treatment) and had 4.3 × 10⁴ and 3.1 × 10² log₁₀ CFU/g, respectively, in aortic valve vegetation samples. One animal was found dead in the TLV treatment group after 48 h (received 2 days of treatment) and had 2 × 10⁴ log₁₀ CFU/g in aortic valve vegetation. All of these data were included in the graphs or were used for statistical analysis. The count for the untreated control group left with the therapy group was ∼1.3 log₁₀ CFU/g higher than that for the untreated control group at baseline (t = 0) (Fig. 2A and B).

DISCUSSION

Enterococcal endocarditis is a challenging disease that requires bactericidal therapy, yet these organisms are frequently tolerant to antibiotics. The standard treatment of E. faecalis endocarditis relies on the use of a combination regimen that consists of a β-lactam (AMP or penicillin) plus an aminoglycoside (GEN or streptomycin) or double-β-lactam therapy of AMP plus CRO. Moreover, VAN is an alternative in patients with β-lactam allergy. Clinical data on the use of these regimens suggest that patients relapse in 3 to 5% of cases, and the use of aminoglycosides may be associated with renal toxicity requiring treatment withdrawal (26, 27).

Previous studies using in vitro and in vivo models of TLV against multidrug-resistant S. aureus (including methicillin-resistant and VAN-intermediate strains) have established that TLV has equivalent or superior efficacy compared to that of VAN (17, 18, 28, 29). Madrigal et al. and Miro et al. reported that TLV monotherapy (at humanized therapeutic doses) was sufficient to significantly reduce bacterial counts in vegetations in a methicillin-resistant S. aureus rabbit model of aortic valve endocarditis, leading to sterilization of vegetations in some animals (17, 18). Clinical reports have also illustrated the successful use of TLV in cases of infective endocarditis caused by Corynebacterium striatum and S. aureus refractory to VAN or DAP (30–34). However, in vivo and clinical data on the use of TLV against deep-seated or endovascular enterococcal infections are lacking.

As the standard of treatment for enterococcal endocarditis includes the use of penicillin or AMP plus GEN or CRO (35), we compared the effectiveness of TLV (monotherapy and in combination with AMP) against these standard regimens using a rat endocarditis model. As expected, AMP plus GEN or CRO significantly decreased bacterial titers compared to results at t = 0, consistent with the recommendations for the use of these combinations in enterococcal endocarditis (36, 37). Interestingly, our results show that TLV monotherapy resulted in a significantly greater reduction in bacterial burdens in vegetations than the standard combination of AMP plus GEN but not compared to AMP plus CRO. Moreover, the addition of AMP to TLV resulted in better efficacy in decreasing bacterial CFU per gram in vegetations compared to these two commonly used regimens (AMP plus GEN and AMP plus CRO); however, it is important to point out that while our doses for these agents are similar to what others have used to study enterococcal IE, they likely are underdosed (38) relative to what would be achieved in humans. In the recent guidelines for the treatment of endocarditis, the American Heart Association endorsed the above-described combinations and indicated that, in the presence of high-level resistance to aminoglycosides, AMP plus CRO would be the therapy of choice (35). Our results suggest that TLV alone or the combination of TLV plus AMP is worth additional studies to explore the feasibility of their use as possible alternative options for patients infected with E. faecalis strains when standard therapy is not effective, considering that the successful use of TLV in deep-seated bacterial infections is supported by only a few case reports of Corynebacterium striatum and S. aureus. Thus, further studies would be needed to establish a place in therapy of infective endocarditis (30–34). Moreover, due to the possibility of renal toxicity with the use of TLV (31), it is unclear if such toxicity would limit its clinical utility when longer courses of therapy are likely to be needed. Thus, clinical studies would be important to clarify these issues.

TLV is a glycopeptide antibiotic that has in vitro bactericidal activity against Gram-positive pathogens and has shown a MIC₉₀ of 0.12 μg/ml against VAN-susceptible E. faecalis (12). In vitro, TLV remains active against enterococcal isolates harboring vanB due to its enhanced binding to the peptidoglycan precursors and reduced induction compared to isolates carrying the vanA gene cluster (13, 39). Unfortunately, VanA is the most common type of vancomycin resistance found in clinical enterococcal isolates, and VanA⁺ strains exhibit higher MICs of TLV (MIC₉₀ of >1 μg/ml) (13, 39, 40). Moreover, the use of TLV against VanB⁺ isolates may be problematic due to the ability of these organisms to increase the level of expression of the van genes and potentially switch from inducible to constitutive status, which would decrease the efficacy of TLV.

Our study is not without limitations. Serum concentrations of AMP, GEN, and CRO were not determined in this study; rather, we administered the antibiotics at doses that were utilized in previous experimental endocarditis models (41, 42). The efficacy of these combinations (AMP plus GEN and AMP plus CRO) were confirmed in our model, despite potential suboptimal dosing compared to humans, and the activity of AMP plus CRO was not statistically significantly different from that of TLV treatment alone. However, extrapolation of the results of this study to the clinical field may be difficult until further studies validating these findings with human-like pharmacokinetics (PK) of AMP, GEN, and CRO are done.

In summary, 3 days of TLV therapy exhibited activity in vivo against E. faecalis strain OG1RF compared to standard regimens. These results suggest that TLV warrants further study as a potential alternative for the treatment of endovascular enterococcal infections caused by ampicillin- and vancomycin-susceptible E. faecalis strains.

MATERIALS AND METHODS

Bacterial strains and growth media used for in vitro and in vivo experiments.

AMP- and VAN-susceptible E. faecalis strain OG1RF (a rifampin- and fusidic acid-resistant strain we have extensively used for IE studies and whose genome has been sequenced and closed) (19, 20) was used in the rat endocarditis model. E. faecalis OG1RF was grown either on brain heart infusion agar (BHI) (Difco Laboratories) for in vitro experiments or in BHI plus 40% horse serum (BHIS) for in vivo experiments. Enterococcosel agar (EA) (Becton Dickinson) supplemented with rifampin (RIF) at 100 μg/ml was used to plate tissue homogenates for bacterial recovery.

MIC determination and time-kill experiments.

MICs of TLV, GEN, DAP, and vancomycin were determined using the TREK methodology and Sensititre panel CMP4STA (lot number 841738; Theravance Biopharma Antibiotics, Inc.) by following guidelines provided by the manufacturer. MICs of AMP and CRO were performed by broth microdilution using cation-adjusted Mueller-Hinton II broth by following criteria stipulated by the Clinical and Laboratory Standards Institute (CLSI) (23). TLV (lot number CL2-434) was provided by Theravance Biopharma Antibiotics, Inc. DAP (lot number MCB2009) was provided by Cubist Pharmaceuticals, Inc. (Lexington, MA), and CRO, AMP, and GEN were obtained from Sigma-Aldrich Chemicals Co. (St. Louis, MO). These antibiotics were reconstituted as recommended by the manufacturers.

The bactericidal activity of DAP and TLV was evaluated by time-kill curves alone or in combination with AMP (1× MIC) against E. faecalis OG1RF. Bacteria were grown in flasks in a final volume of 20 ml of Mueller Hinton II broth (supplemented with 50 mg calcium/ml for DAP) with a starting inoculum concentration of ∼10⁷ CFU/ml from an overnight-grown culture. DAP and TLV were added at concentrations of 1× MIC and 5× MIC with or without AMP (1× MIC). Viable counts were determined at 0, 6, and 24 h by plating appropriate dilutions of the cultures on BHI agar plates. Antibiotic carryover was eliminated by centrifuging 1-ml samples of the culture and resuspending the pelleted bacteria in the same volume of 0.9% NaCl prior to plating. Time-kill and synergism studies were performed three times on different days, and results of a representative experiment are presented. The level of detection was 200 CFU/ml. Bactericidal activity was defined as a ≥3-log₁₀ decrease in the number of CFU per milliliter between 0 and 24 h. Synergy was defined as a ≥2-log₁₀ decrease in colony count at 24 h by the combination compared to that of the most active single agent.

Rat PK analysis.

The PK and IE studies were approved by the University of Texas Health Science Center at Houston Animal Welfare Committee (HSC-AWC-14-036). A total of three male Sprague-Dawley rats (weight, ∼200 g) with cannulated jugular veins (Harlan Laboratories, Houston, TX) were used to facilitate blood sampling. Each animal was given a single subcutaneous (s.c.) TLV dose of 35 mg/kg. Nine blood samples were obtained from each animal at 0 h (prior to TLV dosing) and at 0.083, 0.25, 0.5, 1, 2, 4, 8, and 24 h after TLV dosing. Animals were not anesthetized during the blood collection process and were sacrificed after the last blood sample was collected by following the approved protocol. Blood samples of ∼0.2 ml per time point were collected from the animals and transferred to K2 EDTA Vacutainer tubes and inverted repeatedly to mix. The tubes were immediately placed in a wet ice bath after mixing and then centrifuged at 1,500 to 2,000 × g for 10 min at 4°C. Plasma samples were aliquoted into newly labeled, ice-chilled tubes and were immediately frozen at −80°C. Frozen plasma samples were analyzed by liquid chromatography with tandem mass spectrometry (LC-MS) by the manufacturer at Theravance Biopharma Antibiotics, Inc. The concentrations (in micrograms per milliliter) detected at each time point were averaged, and a one-compartment model was fit to the mean concentration-time profile to derive the best-fit PK parameters using ADAPT II (computational modeling platform). Drug exposures were expressed as area under the concentration-time curves from time zero to infinity (AUC_0–∞). Terminal half-life (t_1/2), highest plasma concentration observed (C_max), and time to C_max (T_max) were calculated. The TLV dosing regimen for the actual treatment regimen against OG1RF-inoculated rats was guided by the results of the PK analysis (described below).

Rat endocarditis model.

Endocarditis was produced in male Sprague-Dawley rats weighing ∼200 g using previous methodologies with slight modifications (21, 22). Briefly, animals were anesthetized with isoflurane for placement of intravascular catheters and the right carotid artery exposed. A sterile polyethylene catheter (e.g., Intramedic PE 10; Clay Adams, Parsippany, NJ) was inserted through a small incision and advanced to 4 cm into the left ventricle. The inoculum that infected 90% of rats (ID₉₀) for E. faecalis OG1RF (in saline suspension) was determined by injecting various inocula (ranging from 10³ to 10⁷ CFU/rat) via the catheter 15 min after catheter placement. The ID₉₀ was determined by the method of Reed and Muench (43) by scoring infected versus noninfected vegetations. All therapy experiments used an inoculum that represented ∼10 times the ID₉₀ of the E. faecalis OG1RF strain, estimated by the optical density at 600 nm (OD₆₀₀), and we determined the actual inoculum by CFU count. The catheter was ligated immediately after the bacterial inoculation and left in place for the duration of the experiment.

Antimicrobial therapy.

The TLV dose scheme was derived from the PK study. Other antibiotics were given to simulate clinically relevant concentrations in humans and/or have been used previously to evaluate efficacy in experimental endocarditis (21, 41, 42, 44–47). The CFU of bacteria per gram of vegetation of untreated baseline controls were determined 24 h after bacterial inoculation (t = 0), which was the time of therapy initiation in the treatment groups, by sacrificing 2 to 3 animals per experiment and plating homogenized aortic valves containing vegetations onto EA plus RIF (100 μg/ml). The following antibiotic regimens were used: (i) TLV, 42 mg/kg/day (21 mg/kg, s.c., q12h); (ii) AMP (100 mg/kg, q8h, intramuscularly [i.m.]) (42); (iii) AMP (100 mg/kg, q8h, i.m.) plus TLV (21 mg/kg, s.c., q12h); (iv) AMP (100 mg/kg, q8h, i.m.) plus GEN (2.5 mg/kg, q12h, i.m.) (42), (v) AMP (100 mg/kg, q8h, i.m.) plus CRO (10 mg/kg, q8h, s.c.) (41), and (vi) DAP (45.3 mg/kg, q24h, s.c.) (21, 45) and VAN (120 mg/kg, q12h, s.c.) (47, 48). A group of infected but untreated control animals was left along with the therapy group to show the end log₁₀ CFU per gram versus the baseline (t = 0) untreated controls.

Per current recommendations, E. faecalis endocarditis requires combination therapy as a first-line approach (such as β-lactam plus aminoglycoside or another β-lactam) or higher-than-FDA-approved doses to achieve successful therapy (e.g., DAP) (35). We tested regimens of high-dose TLV consisting of 42 mg/kg s.c. q12h as monotherapy and in combination with AMP (100 mg/kg, q8h, i.m.) to evaluate the efficacy of these strategies for sterilization of vegetation. Animals were treated for 3 days and sacrificed between 15 h (AMP, CRO, and GEN) and 24 h (DAP, TLV, and VAN) after the last antibiotic dose. Aortic valve vegetations and surrounding tissues were aseptically removed from sacrificed animals, weighed, and homogenized in 1 ml of 0.9% saline. Serial dilutions of homogenized tissues were carried out, and the entire volume of each dilution, including the undiluted sample, was plated onto EA-plus-RIF plates to enumerate bacteria. Animals were included in the final analysis only if they survived the first 24 h of therapy and if catheters were found across the aortic valve in the left ventricle (21, 49). Bacterial CFU per gram of vegetation were obtained per treatment group and compared to those at initiation of therapy (untreated controls at t = 0), to an untreated control group (monitored for 4 days), and among all other treatment groups using geometric mean log₁₀ CFU per gram ± standard deviations. The minimum detection limit was 10 CFU/g of aortic valve containing vegetation. Pulsed-field gel electrophoresis (SmaI digestion) was performed to confirm the identity of inoculated bacteria and of randomly picked bacterial colonies recovered from vegetations.

Data analyses.

The geometric means of the bacterial CFU per gram found in rats were calculated in each group. To negate the effect of large positive skewing of recovery values, the bacterial log₁₀ CFU was log transformed prior to performing unpaired t test for statistical calculation and to obtain P values (50). Cultures yielding no growth were scored as sterile and were assigned a value of 1 CFU for statistical analysis and to obtain geometric mean CFU per gram from all animals. Data/graphs were generated using Prism for Windows (version 4.00; GraphPad Software). Overall, differences were considered significant at a P level of <0.05.