In Vivo Pharmacodynamics of Cefquinome in a Neutropenic Mouse Thigh Model of Streptococcus suis Serotype 2 at Varied Initial Inoculum Sizes

Abstract

Streptococcus suis serotype 2 is an emerging zoonotic pathogen and causes severe disease in both pigs and human beings. Cefquinome (CEQ), a fourth-generation cephalosporin, exhibits broad-spectrum activity against Gram-positive bacteria such as S. suis. This study evaluated the in vitro and in vivo antimicrobial activities of CEQ against four strains of S. suis serotype 2 in a murine neutropenic thigh infection model. We investigated the effect of varied inoculum sizes (10⁶ to 10⁸ CFU/thigh) on the pharmacokinetic (PK)/pharmacodynamic (PD) indices and magnitudes of a particular PK/PD index or dose required for efficacy. Dose fractionation studies included total CEQ doses ranging from 0.625 to 640 mg/kg/24 h. Data were analyzed via a maximum effect (E_max) model using nonlinear regression. The PK/PD studies demonstrated that the percentage of time that serum drug levels were above the MIC of free drug (%ƒT_>MIC) in a 24-h dosing interval was the primary index driving the efficacy of both inoculum sizes (R² = 91% and R² = 63%). CEQ doses of 2.5 and 40 mg/kg body weight produced prolonged postantibiotic effects (PAEs) of 2.45 to 8.55 h. Inoculum sizes had a significant influence on CEQ efficacy. Compared to the CEQ exposure and dosages in tests using standard inocula, a 4-fold dose (P = 0.006) and a 2-fold exposure time (P = 0.01) were required for a 1-log kill using large inocula of 10⁸ CFU/thigh.

INTRODUCTION

Streptococcus suis is an important pathogen in the swine industry and causes significant economic losses worldwide. Moreover, it is an emerging zoonotic pathogen causing severe infection in people who have close contact with diseased pigs or pork-derived products (1, 2). To date, 35 serotypes based on S. suis capsular antigens have been described. S. suis serotype 2 is the most virulent and the dominant pathogenic serotype. It has been associated with a variety of severe clinical infections such as meningitis, septicemia, pneumonia, arthritis, and endocarditis in both pigs and humans (3, 4). Significant public health concerns were raised due to large outbreaks of S. suis 2 in humans that occurred in China in 1998 and 2005 that caused high morbidity and mortality (5). Because suitable vaccines were not available, the control of S. suis 2 infections depended almost entirely on the use of antimicrobials. However, the occurrence of high levels of resistance of S. suis to certain antimicrobials (e.g., macrolides, lincosamides, tetracyclines, and sulfonamides) has limited the choice of antimicrobial agents for treatment (6). It has previously been reported that the majority of S. suis strains are susceptible to β-lactams (MIC of ≤0.03 μg/ml). This suggested that these drugs may be efficacious in the treatment of S. suis 2 infections (7).

Cefquinome (CEQ) is a fourth-generation cephalosporin developed solely for veterinary use and has been highly effective against a wide variety of Gram-positive and Gram-negative bacteria (8). CEQ has also been tested and was shown to be effective in animals using a standard thigh infection model with both Staphylococcus aureus and Escherichia coli (9, 10).

The pharmacokinetic (PK) and pharmacodynamic (PD) properties used to predict the efficacy of antibiotics are necessary to provide appropriate dosing recommendations to achieve maximum efficient therapy while minimizing bacterial resistance (11). However, data about the PK/PD profiles of CEQ against S. suis 2 are very limited.

The postantibiotic effect (PAE) describes the persistent inhibition of bacterial growth after a brief exposure to antimicrobial agents. It is an important determinant of the optimal dosing since it reduces the frequency of antibiotic dosing (12). β-Lactams induce PAEs lasting approximately 2 h against Gram-positive cocci but have very short or no effects against Gram-negative bacilli (13). However, PAEs have not yet been tested against S. suis 2 in animal models.

The inoculum effect (IE) is also one of the factors influencing the outcome of in vitro susceptibility testing and in vivo therapeutic effects (14–16). The IEs for β-lactams vary, indicating that severe infections characterized by high bacterial densities (e.g., septicemia, endocarditis, and pneumonia) may require higher doses to achieve treatment efficacy (17). Therefore, estimating the treatment effects of CEQ against S. suis 2 at a high bacterial burden can provide meaningful information.

The following studies were designed to characterize the pharmacodynamics of CEQ against S. suis 2. Specifically, we wanted to determine whether inoculum size affects CEQ efficacy in a murine neutropenic thigh infection model. The goals of the experiment were to (i) characterize the in vivo PAEs, (ii) determine the MIC and test in vitro time-killing curves using different inoculum sizes, (iii) determine which PK/PD index is best correlated with efficacy using different inoculum sizes via dose fractionation and pharmacodynamic modeling, and (iv) compare the magnitudes of the index and doses required for efficacy among S. suis 2 strains between those for the standard and large inocula.

MATERIALS AND METHODS

Antibiotics, media, and bacterial strains.

Cefquinome commercial powder (>99.5%) was purchased from Qilu Animal Health Product Company, Ltd., (Guangzhou, China). Tryptic soy broth (TSB) and tryptic soy agar (TSA) were purchased from Guangdong Huankai Microbial Sci. & Tech. Company, Ltd. (Guangzhou, China). Five percent defibrinated sheep blood was provided by Guangzhou Ruite Bio-tec Company, Ltd. (Guangdong, China).

Three S. suis 2 clinical strains (GD110, GF114, and GD4) isolated from pigs and one standard strain, ATCC 43765, were evaluated in this study. The strains were grown, cultured, and quantified in TSB and on TSA plates containing 5% defibrinated sheep blood.

In vitro susceptibility studies.

The MICs of CEQ against S. suis 2 at various inoculum sizes (10⁶ to 10⁸ CFU/ml) were determined by the agar dilution method according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) (18). Tests were conducted in a minimum of three independent runs.

In vitro time-kill curves.

The effects of inoculum size on the in vitro time-kill curve of CEQ against S. suis 2 strain ATCC 43765 were studied using the broth clinical method as previously described (19). Briefly, bacteria were cultured at 37°C in TSB until logarithmic phase and diluted in LB medium to the standard test inoculum size (SI) of 10⁶ CFU/ml or to a high inoculum size (HI) of 10⁸ CFU/ml. Bacteria at these two densities were then exposed for 12 h at 1, 2, 4, 8, 16, and 32 times the MIC and incubated in a shaking water bath at 37°C. Aliquots of 100 μl were removed at 0, 3, 6, 9, and 12 h and serially diluted in sterilized saline and plated on TSA for CFU determinations. Total bacterial counts (CFU/ml) were calculated after 24 h of incubation at 37°C. Bactericidal and bacteriostatic activity were defined as a ≥3-log₁₀ CFU/ml or a <3-log₁₀ CFU/ml reduction at each of the time points in colony count from those of the initial inocula, respectively (19).

Mouse thigh infection model.

Six-week-old specific-pathogen-free female ICR mice weighing 22 to 27 g obtained from the Medical Experimental Animal Center of Guangdong Province, Guangzhou, China, were used for all studies. All animal studies were approved by the animal research committees of South China Agriculture University. Animal housing, feeding, and experimental operations were performed in accordance with the National Research Council recommendations.

The mice were rendered neutropenic by two intraperitoneal injections of cyclophosphamide (Puboxin Biotechnology, Beijing, China) at 4 days preinfection using 150 mg/kg and at 1 day preinfection at 100 mg/kg. Thigh infections were carried out as outlined previously (20). Briefly, S. suis 2 suspensions were prepared from fresh logarithmic-phase subcultures and diluted to the appropriate densities. Subcutaneous (s.c.) thigh injections consisted of 0.1 ml of cells in sterile saline with 10⁷ CFU/ml (standard inoculum, SI) and 10⁹ CFU/ml (high inoculum, HI) into each mouse thigh. Mice in the untreated control group were injected with vehicle only (normal saline). Bacterial densities were confirmed by serial dilutions and plating on TSA for CFU determinations.

CEQ was administered s.c. at various time points beginning at 2 h postinfection. The mice were euthanized by CO₂ asphyxiation at 24 h postinjection, and each thigh was immediately removed, homogenized, and suspended in 10 ml of ice-cold sterile saline. Total bacterial counts (CFU/ml) were determined by duplicate plating on TSA as described above. Data are represented as the means (± standard deviations [SD]) of log₁₀ CFU/thigh (4 thighs/time point).

Pharmacokinetic studies.

Pharmacokinetic studies were performed at SI, with S. suis 2 strain ATCC 43765-infected neutropenic mice given single doses of cefquinome at 2.5, 10, 40, 160, and 640 mg/kg s.c. Blood was sampled by retro-orbital puncture using three groups of three mice per dose at 0.083, 0.167, 0.25, 0.5, 0.75, 1, 2, 3, 4, 5, and 6 h after dosing (21). An individual mouse was sampled three or four times. Serum was collected and stored in polypropylene tubes at −80°C until analysis. After being thawed, the samples were centrifuged for 5 min at 6,900 × g, and serum CEQ concentrations were determined by liquid chromatography-tandem mass spectrometry (LC-MS/MS) as previously described (10). The equipment consisted of an Agilent Series 1200 high-performance liquid chromatography (HPLC) system (Agilent, Palo Alto, CA, USA) coupled to an Applied Biosystems API 4000 triple quadrupole mass spectrometer with electrospray ionization interface (Foster City, CA, USA) used in positive mode. The quantitation limit was 0.001 μg/ml. Matrix-matched calibration standards gave linear responses from 0.001 to 1 μg/ml (r > 0.995), and both the intraday and interday variations were less than 7.3%. A value of 92.6% CEQ protein binding was obtained from a previous report and used to calculate free-drug serum concentrations (10).

In vivo postantibiotic effects.

S. suis 2 strain ATCC 43765 was used to produce a thigh infection in mice for determining PAE. Infected mice were injected s.c. with 0.2-ml single doses of CEQ at 2.5 and 40 mg/kg at 2 h postinfection. Untreated control growth was determined at seven sampling times (two mice per time point) over 24 h with sampling at 0, 2, 4, 6, 8, 12, and 24 h. The treated groups (two mice per time point) were sampled nine times over 24 h at 1, 2, 3, 4, 6, 8, 12, and 24 h. The PAE was calculated by subtracting the time it took the pathogens to increase 1 log in the thighs of untreated control animals from the time it took the pathogens to increase the same amount in treated animals after serum drug levels decreased below the MIC of the drug for the pathogens (22). The equation for calculating the PAE was as follows: PAE = T − C, where T is the time for the growth of 1 log₁₀ in treated animals after the free-drug levels in serum had fallen below the MIC and C is the time for the growth of 1 log₁₀ in untreated control animals (23).

PK/PD index determination.

In vivo activity was determined for mice with systemic infection of standard S. suis 2 ATCC 43765 using two inoculum sizes. The therapy started at 2 h postinfection, and 28 dosing regimens were used to determine the effects of dosing level and interval on CEQ efficacy. These regimens consisted of six dosage levels at 4-fold-increments (0.625 to 640 mg/kg/24 h) and consisted of four dosing intervals at 24, 12, 6, and 3 h (i.e., 1, 2, 4, or 8 doses every 24 h). Groups of two mice were involved in each dosing regimen. The mice were euthanized after 24 h, and each thigh was immediately removed and processed for CFU determination. The untreated control mice were sacrificed just before CEQ administration and after 24 h.

PK/PD parameter magnitude studies.

In experiments similar to those described above, mice were infected with three additional strains of S. suis 2 (GD110, GF114, and GD4). The mice were then treated with CEQ doses administered every 6 h for 24 h, and doses varied from 0.625 to 160 mg/kg and from 0.625 to 640 mg/kg for the SI and HI groups, respectively. Two mice were used for each dosing regimen. At the end of the study, the mice were euthanized, and each thigh was removed and processed as outlined above.

PK and PD analysis.

The pharmacokinetic parameters of each dosage regimen were calculated using the WinNonlin, version 5.2, program (Pharsight, Mountain View, CA, USA). A one-compartment model was used to calculate pharmacokinetic constants, including elimination half-life (t_1/2), area under the concentration-time curve (AUC), and maximum concentration of drug in serum (C_max). The PK/PD parameters for the treatment doses not specifically studied were linearly extrapolated from the values of the nearest studied doses. The sigmoid effect (E_max) derived from the Hill model for data smoothing was used to characterize the relationship between antibiotic efficacy and three PK/PD parameters: percentage of time that serum drug levels are above the MIC of free drug (%ƒT_>MIC, where the f prefix used with the parameter indicates free drug), the ratio of the maximum concentration of the free drug in serum to the MIC (ƒC_max/MIC), and the ratio of the area under the concentration-time curve from 0 to 24 h to the MIC (ƒAUC_0–24/MIC). The final equation used was E = (E_max × C^N)/(EC_50^N + C^N), where E is the effect and, in our case, equal to the log change in CFU counts per thigh between treated and untreated control mice after the 24-h period of the study, E_max is maximum effect, C is the PK/PD parameter being examined (%ƒT_>MIC, ƒC_max/MIC, and ƒAUC/MIC) as well as the 24-h total dose in this study, EC₅₀ is the C value that results in 50% of the E_max, and N is a sigmoid factor that controls the slope of the curve. Nonlinear regression analysis was used to determine which PK/PD index best correlated with the number of CFU/thigh at 24 h at two inoculum sizes. The coefficients of determination (R²) were used to estimate the variance associated with regression modeling with each of the PK/PD indices. Dosing calculated by the sigmoid E_max model for the static effect and 1-log killing at different inocula and the rate of dose between SI and HI (IE index) were compared by t tests.

RESULTS

In vitro susceptibility testing and time-kill curves.

The MICs of CEQ against the studied strains at two inoculum sizes are listed in Table 1, and they ranged from 0.03 to 0.24 μg/ml. We initiated this study by evaluating the effects of inoculum size on MIC determination. The higher inoculum size resulted in MICs that were only 2- to 4-fold higher than those with the SI and ranged from 0.06 to 0.5 μg/ml (Table 1). These results were not judged to be significant (P = 0.25). Additionally, the MICs for the pre- and posttreatment isolates at the SI and HI exposed to CEQ therapy in the thighs of treated mice were monitored. However, the MICs for these organisms were not changed, indicating that preexisting and/or occurring mutant isolations did not happen during the experiment.

TABLE 1.

MIC of CEQ against S. suis 2 strains

S. suis 2 strain	MIC (mg/liter) at an inoculum of:
	106 CFU/ml	108 CFU/ml
ATCC 43765	0.03	0.06
GD110	0.03	0.06
GD114	0.12	0.5
GD4	0.24	0.5

We next evaluated the time-dependent killing action of CEQ against S. suis 2 ATCC 43765. The organism grew under control conditions with a mean of 2.6 ± 0.16 (SI) and 0.86 ± 0.09 (HI) log₁₀ CFU/ml after 12 h. The onset rates of killing between the two inoculum sizes were similar during the first 6 h. After this period, 2× MIC yielded bacteriostatic activity for the SI, and 4× MIC achieved maximal killing, with a higher concentration providing little added benefit (Fig. 1). However, in the HI group, the CEQ bactericidal activity was markedly attenuated, and even increasing the concentration to 32× MIC could not achieve bactericidal activity (Fig. 1).

FIG 1.

In vitro time-kill course of CEQ against S. suis 2 ATCC 43765 at both standard (left) and high (right) initial inocula.

Pharmacokinetics.

The single-dose pharmacokinetics of CEQ are shown in Fig. 2. A one-compartment model was used to depict the CEQ concentration-versus-time profile. The elimination half-life ranged from 0.25 to 0.28 h. The kinetics of escalating doses were linear for both C_max and AUC and ranged from 1.96 to 606.71 mg/liter and from 1.58 to 551.40 mg h/liter, respectively (Table 2). This significant linearity correlation allowed us to extrapolate the PK/PD parameters for other doses which were not specifically studied.

FIG 2.

Mean serum drug concentration-time profiles of CEQ following single subcutaneous doses in infected mice.

TABLE 2.

Single-dose pharmacokinetics of CEQ in infected mice

Dose (mg/kg)	Cmax (mg/liter)	AUC (mg · h/liter)	Tmax (h)a	t1/2 (min)
2.5	1.96	1.58	0.25	0.33
10	8.08	6.44	0.25	0.32
40	29.55	25.26	0.27	0.34
160	136.11	120.37	0.28	0.35
640	606.71	551.40	0.28	0.38

^aTime to maximum concentration.

Postantibiotic effects in vivo.

The effect of single doses of CEQ at 2.5 or 40 mg/kg on the in vivo killing and regrowth of S. suis 2 ATCC 43765 are shown in Fig. 3. Dose-dependent killing of organisms was observed at two drug concentrations, with prolonged growth inhibition. The organism grew 1.28 ± 0.56 log₁₀ CFU/thigh over 24 h in untreated animals, and the growth point of 1 log₁₀ CFU/thigh occurred at 7.15 h. Two doses reduced the bacterial load by 0.22 ± 0.03 and 0.99 ± 0.09 log₁₀ CFU/thigh in 2 and 5 h, respectively. Based upon serum PK determinations, the serum CEQ concentrations following a single dose of 2.5 or 40 mg/kg remained above the MIC for 2.35 or 3.78 h, respectively. Finally, 9.6 h and 15.7 h were required for the number of CFU in the thighs of treated mice to increase 1 log₁₀ CFU/thigh at the time serum levels had fallen below the MIC, which resulted in PAE values in vivo of 2.45 and 8.55 h, respectively.

FIG 3.

In vivo postantibiotic effects (PAEs) of CEQ after two single subcutaneous doses against S. suis 2 ATCC 43765 using a neutropenic mouse thigh model. The filled bars represent the time that serum CEQ concentrations remained above the MIC for the infecting organism.

PK/PD indices.

The relationships between the three PK/PD indices and CEQ efficacy against the control strain ATCC 43765 in the thighs of neutropenic mice with SI and HI infections are illustrated in Fig. 4. At the start of therapy, the mice had 6.15 ± 0.35 and 8.18 ± 0.12 log₁₀ CFU/thigh with the SI and HI infections, respectively. Bacterial density increased by a mean of 1.60 (SI) and 0.36 (HI) log₁₀ CFU/thigh after 24 h in untreated control mice. Escalating doses of CEQ resulted in the time-dependent killing of both inocula. For the SI group, the changes in log CFU values had the strongest correlation with %ƒT_>MIC (R² = 91%). Correlations with ƒAUC/MIC (R² = 58%) and the ƒC_max/MIC ratio (R² = 37%) were both poor (Fig. 4).

FIG 4.

Relationship between three PK/PD indices and changes in the log₁₀ CFU/thigh with both standard (filled circles) and high (open circles) initial inocula of S. suis 2 ATCC 43765 in the thighs of neutropenic mice.

At the higher inocula, the %ƒT_>MIC value still held the best correlation (R² = 63%) although the nonlinear curve was shifted somewhat to the top, indicating a diminished effect at the bactericidal efficacy. As with the SI group, there was no correlation between log CFU values and the other indices (Fig. 4). These data demonstrated that the %ƒT_>MIC value was the most closely linked to the treatment efficacy with both the SI and HI even though the HI reduced the value of the %ƒT_>MIC.

Magnitudes of %ƒT_>MIC and doses required for efficacy.

To determine whether the lengths of exposure and the sizes of the doses required for efficacy were similar for the various inocula, the antibiotic activities of the CEQ 6-h dosing regimens against three additional strains of S. suis 2 were determined. At the start of therapy, mice had 6.16 ± 0.26 and 8.12 ± 0.22 log₁₀ CFU/thigh, with a growth rate of 1.22 ± 0.14 and 0.34 ± 0.12 log₁₀ CFU/thigh after 24 h in untreated control mice infected with the SI and HI, respectively.

The degree of exposure (%ƒT_>MIC) necessary to achieve a net static effect and 1-log₁₀ killing against multiple organisms is given in Table 3. The %ƒT_>MIC values were not statistically different in the degree of CEQ exposure required to achieve a net bacteriostatic effect between the SI and HI groups (P = 0.19). There was, however, roughly a 2-fold increase in the 1-log killing for the HI group (P = 0.01).

TABLE 3.

The %ƒT_>MIC required for static effect and 1-log₁₀ killing with CEQ treatment in vivo at the standard and high inocula

S. suis strain	%ƒT>MIC of CEQ required with an inoculum of:
	106 CFU/ml			108 CFU/ml
	Emax	Static effect	1-log10 killing	Emax	Static effect	1-log10 killing
ATCC 43765	5.05	18.01	35.03	3.66	15.52	48.65
GD110	3.76	25.62	43.94	4.6	37.43	59.96
GD114	4.64	19.21	34.83	4.07	31.43	54.75
GD4	4.37	20.11	34.47	3.56	27.12	44.74
Mean ± SD	4.45 ± 0.54	20.74 ± 3.36	37.06 ± 4.58a	3.97 ± 0.47	27.86 ± 9.26	52.03 ± 6.71a

^aP < 0.05, comparing the %ƒT_>MIC values for the standard and high inocula.

The CEQ dosages required for stasis and 1-log killing are shown in Table 4. The IE index for the static effect was low (1.59 ± 0.67), and for 1-log killing it increased to 4.47 ± 1.21. The HI showed a statistically significant effect on the magnitude of dosage required for 1-log killing efficacy and was almost 4-fold higher than that for the SI (P = 0.006).

TABLE 4.

Inoculum effect and doses required for static effect and 1-log₁₀ killing of each organism

S. suis strain	Dose (mg/kg/day) of CEQ required to achieve:
	Static effect			1-log10 killing
	SI	HI	IE index	SI	HI	IE index
ATCC 43765	0.7	0.7	1	8.6	28.02	3.28
GD110	0.4	0.8	2	2.2	12.61	5.73
GD114	0.6	1.4	2.33	3.2	16.8	5.25
GD4	1.4	2.1	1.06	5.8	21.02	3.6
Mean ± SD	0.78 ± 0.43	1.25 ± 0.64	1.59 ± 0.67	4.95 ± 2.87	19.61 ± 6.57	4.47 ± 1.21a

^aP < 0.01, comparing the inoculum effect (IE) index values for the standard (SI) and high (HI) inocula.

DISCUSSION

S. suis 2 is most often associated with severe invasive disease in both pigs and humans in many countries. β-Lactams are the gold standard for therapy in treating these infections although resistance to penicillin and third-generation cephalosporins is becoming more common. Infections with resistant isolates can be traced directly to treatment failures (24).

S. suis is considered the paradigm for the intersections between animal and human resistomes (23). Therefore, the prudent use of β-lactams for S. suis 2 infections in pigs is essential to maintain therapeutic efficacy and to minimize selection of resistant S. suis strains.

CEQ was chosen in this study due to its broad antibacterial spectrum and because it represents a reliable a choice for the initial empirical treatment in the case of a critical illness. We used PK/PD modeling as a scientific tool in an attempt to establish optimal dosing regimens and provide guidelines for empirical therapy. To the best of our knowledge, the current study is the first to evaluate the in vivo pharmacodynamic activity of CEQ against S. suis 2. The PK/PD analysis determined the following: (i) that CEQ produced a significant and prolonged PAE, (ii) that the variable %ƒT_>MIC was most closely linked to efficacy with both inoculum sizes although this was most effective at the lower standard dose, and (iii) that the in vitro time-kill study and in vivo experiments demonstrated that efficacy of CEQ was weakened with increasing bacterial density.

PAEs were demonstrated with CEQ against S. suis 2 since the persistent suppression of bacterial regrowth was observed. Meanwhile, we noticed that the regrowth did not happen immediately after the drug levels had decreased below the MIC, which was probably due to postantibiotic subinhibitory concentrations (postantibiotic sub-MIC effect [PA-SME]) (25). In the in vivo situation, a suprainhibitory concentration of a drug would always be followed by subinhibitory concentrations (sub-MICs). Persisting sub-MICs remaining in the thighs of infected mice were difficult to separate from true in vivo PAEs in an animal model, resulting in an increase in the in vivo PAE (26). The presence of prolonged PAEs would prevent the bacterial regrowth immediately between the doses, which would then allow more widely spaced dosing intervals without loss of efficacy (27). CEQ induced a prolonged PAE that ranged from 2.45 to 8.55 h and that was directly correlated with dose. This phenomenon was similar to the long PAE times of β-lactams used against Staphylococcus spp. (28). However, these data conflict with earlier studies that found nonexistent or short PAEs for β-lactams against Streptococcus pneumoniae in vivo (29). This difference may be due to differences in growth rates between S. suis and S. pneumoniae. S. pneumoniae multiplies too slowly in vivo to reach the time required for a 1-log₁₀ increase.

Using the standard inocula, the mean values of %ƒT_>MIC required for stasis and 1-log killing against the four strains of S. suis 2 were 20.74% and 37.06%, respectively. These data were similar to the %ƒT_>MIC requirement for the antibiotic activities of other cephalosporins against Staphylococcus spp.; however, they were lower than those against S. pneumoniae (28). Furthermore, Andes and Craig found that %ƒT_>MIC mean values of 26% to 39% were sufficient to achieve a static effect and 1-log killing of a new cephalosporin (PPI-0903) against methicillin-resistant Staphylococcus aureus (MRSA) that produced a long PAE with PPI-0903. Higher mean values of 39% and 43% were necessary for S. pneumoniae that presented no or only a modest PAE (22). These results agree with our observations indicating that the prolonged PAE in S. suis 2 resulted in CEQ administration at long intervals between doses.

The inoculum effect was first reported using a sulfa compound against Streptococcus haemolyticus in 1940 (30). Subsequently, many studies assessed this phenomenon on β-lactams against various pathogens in vivo or in vitro (28). These studies demonstrated that the IE varied between β-lactam subclasses to various degrees (19, 31). Eng et al. divided 10 β-lactam antibiotics into three groups (poor, slow, and rapid) for their bactericidal effects based upon how their antibiotic activity on Pseudomonas aeruginosa was affected by different inoculum sizes (16). Furthermore, Lee et al. have even suggested that IE should be included in animal efficacy studies when new antibiotics are tested (32).

In the present study, in vitro time-kill data showed that maximal killing was attained with the SI at 4× MIC. However, only a bacteriostatic effect was observed with the HI at various multiples of the MICs of drug. This may account for the difference in the kill numbers between the inoculum sizes. In previous studies, we found that growth rate was static and reduced after 12 h in culture medium, probably due to the lack of nutrient and the cellular damage caused by H₂O₂, O₂⁻, and OH˙ which were produced by S. suis itself (33). However, these values were sufficient to demonstrate a 12-h killing effect, so this was chosen as the endpoint for the current study.

Considering that β-lactams are cell wall-active agents, their killing effects are dependent on active pathogen growth (31). Therefore, the lower growth rate of the pathogen in the HI group than that in the SI group (0.91 ± 0.09 versus 2.8 ± 0.16 log₁₀ CFU/ml) most likely caused the lower killing efficacy of CEQ in the HI group due to the lack of very active pathogen growth.

Although the influences on in vitro killing activity of β-lactams are demonstrated, the impact of the IE on in vivo experiments remains controversial. Maglio et al. evaluated cefepime against Escherichia coli using TEM-derived extended-spectrum beta-lactamases (ESBLs) and found that no differences in exposure times and static doses were required for efficacy using various inoculum sizes (34). Lee et al. reached the same conclusions with their work using S. pneumoniae (32). However, in that same study, opposite results were observed for MRSA. There was a reduction in killing effect with vancomycin and ceftobiprole using large inocula. Furthermore, two other studies that reported that IE exerts a significant influence suggested that higher β-lactam doses would be necessary for efficacy when dense bacterial infections were encountered (15, 19). Our results are in agreement with these studies.

Using dose fractionation experiments, the %ƒT_>MIC value was associated with efficacy, and the HI group showed a lower correlation than the SI group. In the HI infection model, there was a significant relationship between efficacy and %ƒT_>MIC in the 0 to 40 range. However, achieving bactericidal activities was correlated not only to %ƒT_>MIC but also to the drug concentration. When this was extended to three additional strains, we again observed that higher doses (4-fold) and longer exposure times (%ƒT_>MIC) were necessary for a 1-log kill.

This phenomenon was previously demonstrated by Williamson and Tomasz, who studied penicillin binding protein (PBP) acylation using benzylpenicillin against S. pneumoniae. The PBPs reached a distinct degree of saturation that was not correlated to antibiotic concentration but to the antibiotic dose when S. pneumoniae was exposed to various multiples of drug MICs (35). Moreover, we could not find mutant isolates pre- and postexperiment, indicating that preexistent mutant populations in the HI group were not a reason why higher doses of CEQ are needed to achieve the experimental endpoints in the high-inoculum samples. Although the potential mechanisms for the inoculum effect are unclear, Stevens et al. proposed that the IE for β-lactams was caused by a gradual decrease in the expression of PBPs 1 and 4 in Streptococcus pyogenes at later growth phases (36).

In conclusion, CEQ exhibits concentration-independent killing as well as a prolonged PAE against S. suis 2. The most important pharmacodynamic parameter was %ƒT_>MIC for describing the in vivo activity. Meanwhile, the killing effect of CEQ was dramatically influenced by inoculum size. A high inoculum raised the magnitude of dose and the length of exposure associated with efficacy for four study strains. These data suggest that animal dosage regimens should supply a %ƒT_>MIC of CEQ for 20 to 40% of the interval for S. suis 2. Based on a previous pharmacokinetic study of CEQ in pigs (37), MIC distribution (38), and the value of %ƒT_>MIC target indices, a 5,000-subject Monte Carlo simulation was performed using Crystal Ball Professional, version 11.1.2.4, software to predict a reasonable dosage regimen. The probability of target attainment (PTA) of >100% could be achieved for a static effect and 1-log₁₀ killing effect for a MIC of ≤0.13 μg/ml under the recommended clinical dose of 2 mg/kg/24 h, indicating that the current recommended CEQ dosages are sufficient in treating infectious diseases of pig caused by S. suis 2. However, a 4-fold higher dose and a 2-fold %ƒT_>MIC of the drug are required for efficacy in the treatment of severe infections. The results in this study may provide a more rational basis for determining optimal doses for treatment regimens of β-lactams in human and animal infections of S. suis 2.