Nemonoxacin Enhances Antibacterial Activity and Anti-Resistance Mutation Ability of Vancomycin against Methicillin-Resistant Staphylococcus aureus in an In Vitro Dynamic Pharmacokinetic/Pharmacodynamic Model

ABSTRACT

Reduced susceptibility and emergence of resistance to vancomycin in methicillin-resistant Staphylococcus aureus (MRSA) have led to the development of various vancomycin-based combinations. Nemonoxacin is a novel nonfluorinated quinolone with antibacterial activity against MRSA. The present study aimed to investigate the effects of nemonoxacin on antibacterial activity and the anti-resistant mutation ability of vancomycin for MRSA and explore whether quinolone resistance genes are associated with a reduction in the vancomycin MIC and mutant prevention concentration (MPC) when combined with nemonoxacin. Four isolates, all with vancomycin MICs of 2 μg/mL, were used in a modified in vitro dynamic pharmacokinetic/pharmacodynamic model to investigate the effects of nemonoxacin on antibacterial activity (isolates M04, M23, and M24) and anti-resistant mutation ability (isolates M04, M23, and M25, all with MPCs of ≥19.2 μg/mL) of vancomycin. The mutation sites of gyrA, gyrB, parC, and parE of 55 clinical MRSA isolates were sequenced. We observed that in M04 and M23, the combination of vancomycin (1 g given every 12 h [q12h]) and nemonoxacin (0.5 g once daily [qd]) showed a synergistic bactericidal activity and resistance enrichment suppression. All clinical isolates resistant to nemonoxacin harbored gyrA (S84→L) mutation; gyrA (S84→L) and parC (E84→K) mutations were the two independent risk factors for the unchanged vancomycin MPC in combination. Nemonoxacin enhances the bactericidal activity and suppresses resistance enrichment ability of vancomycin against MRSA, with an MIC of 2 μg/mL. Our in vitro data support the combination of nemonoxacin and vancomycin for the treatment of MRSA infection with a high MIC.

TEXT

Methicillin-resistant Staphylococcus aureus (MRSA) is still recognized as an urgent public health threat, as it is a major cause of health care- and community-associated infections with poor clinical outcomes and high mortality rates (1). In the United States, the reported incidence of MRSA among all S. aureus infections ranges from 7% to 60%, and the mortality rates vary from 5% to 60%, depending on the patient population and site of infection (2). According to the China Antimicrobial Surveillance Network (CHINET) (http://www.chinets.com), the average detection rate of MRSA is 35.3% (3).

Vancomycin, a glycopeptide antibiotic with potent antistaphylococcal activity, is the first-line therapy for the treatment of MRSA infections. A ratio of the area under the curve for 24 h (AUC₂₄) to the MIC of 400 to 600 is considered the optimal pharmacokinetic/pharmacodynamic (PK/PD) target with a better clinical and bacteriological response (4). When the MIC is greater than 1 μg/mL, the probability of achieving an AUC₂₄/MIC target of ≥400 is low with the recommended administration dosage. It has been demonstrated that a vancomycin MIC of ≥1 μg/mL was an independent predictor of mortality in patients with MRSA bacteremia (5). A meta-analysis also found an increase in vancomycin treatment failure and mortality rates for MRSA with a vancomycin MIC of ≥1.5 μg/mL (6). It has been reported that MRSA isolates with a vancomycin MIC of >1 μg/mL are prevalent worldwide (7–9), which brings challenges to the treatment of MRSA infections. Moreover, an increase in vancomycin MIC is associated with the emergence of vancomycin-intermediate S. aureus, heterogeneous vancomycin-intermediate S. aureus, or even vancomycin-resistant S. aureus (10). To limit the emergence of resistant subpopulations of MRSA during treatment with vancomycin, the ratio of AUC₂₄ to the mutant prevention concentration (MPC) should exceed 15 (11). However, the AUC₂₄/MPC with recommended administration dosage in patients of normal renal function would not be able to achieve the ideal target value when the vancomycin MPC is ≥19.2 μg/mL (12). With a limited pool of available antimicrobial agents capable of treating MRSA infection in China, vancomycin-based combination therapy, aiming to reduce the MIC and MPC to inhibit the emergence of antimicrobial resistance and increased antibacterial activity, seems to be an attractive therapeutic option for MRSA infection.

Nemonoxacin is a novel C-8-methoxy nonfluorinated quinolone (NFQ) which displays good in vitro and in vivo activity against MRSA (13, 14). As a category 1 drug (i.e., a drug with an innovative structure that is not listed at home and abroad) of the chemical drug registration classification of China (https://www.nmpa.gov.cn/yaopin/ypggtg/ypqtgg/20200630180301525.html), nemonoxacin malate injection was only first marketed in China in June 2021. It has been shown that nemonoxacin has a lower MIC and a narrower mutation selection window for MRSA than levofloxacin and moxifloxacin (15). Previous studies have demonstrated in vitro synergy against MRSA between vancomycin with ofloxacin (16) or with moxifloxacin (17). Our previous in vitro static checkerboard study also observed in vitro synergy against MRSA with the nemonoxacin and vancomycin combination (18). However, the potential synergetic effect has not been studied in a dynamic time-kill assay simulating humanized dosing of the combination, which is more representative of clinical practice.

Fluoroquinolone resistance is typically attributed to spontaneous point mutations in the quinolone resistance-determining region (QRDR) of genes encoding DNA gyrase (gyrA and gyrB) and topoisomerase IV (parC and parE) (19, 20). These mutations were found to play an important role in resistance to nemonoxacin in Streptococcus pneumoniae (21). However, there are rare reports on the mutations of nemonoxacin resistance in S. aureus. Moreover, whether gene mutations in S. aureus affect the efficacy of a combination regimen with vancomycin and nemonoxacin is unclear.

Therefore, the present study aimed to investigate the effects of nemonoxacin on antibacterial activity and anti-resistant mutation ability of vancomycin for MRSA with a MIC of 2 μg/mL and MPC of ≥19.2 μg/mL in a modified in vitro dynamic PK/PD model and explore whether the quinolone resistance genes are associated with a reduction of vancomycin MIC and MPC when vancomycin is combined with nemonoxacin.

RESULTS

Mutations in DNA gyrase and topoisomerase IV genes.

Overall, mutations in DNA gyrase and topoisomerase IV genes were detected in 42 (76.4%) of the 55 isolates, with concomitant mutations (gyrA with parC, gyrA with parE, and gyrA with parE/parC) in 36 (65.5%) isolates. No mutations were found in the gyrB gene. The mutations in gyrA, parC, and parE are detailed in Table 1. The 13 isolates without any mutations were highly sensitive to nemonoxacin (MIC ≤ 0.06 μg/mL). However, 20 of the 25 isolates with concomitant mutations of gyrA and parC genes were resistant to nemonoxacin (MIC ≥ 2 μg/mL) according to the epidemiological breakpoint of China. Six of the 20 nemonoxacin-resistant isolates with concomitant gyrA S84→L and gyrA E88→G and/or E88→K mutations had MICs of ≥8 μg/mL.

TABLE 1.

Mutation patterns of DNA gyrase and topoisomerase IV genes in 55 clinical isolates of methicillin-resistant Staphylococcus aureus^a

Group	QRDR mutation(s) (no. of isolates per group)	No. of isolates per nemonoxacin MIC (μg/mL) of:									Fold change in reduction of vancomycin with nemonoxacin MIC (μg/mL) of:		Fold change in reduction of vancomycin with nemonoxacin MPC (μg/mL) of:
		≤0.06	0.12	0.25	0.5	1	2	4	8	16	0	≥1	0	≥1
All isolates	55 isolates	14	1		1	8	24	1	4	2	18	37	25	27
None (n = 13)	13 isolates	13									4	9	1	12
gyrA alone (n = 1)	S84→L (1)						1				1		1
parC alone (n = 5)	S80→Y (2)		1				1					2	1	1
	E84→K (2)	1			1						1	3	1	1
	S80→Y, E84→K (1)						1					1	1
gyrA with parC (n = 25)	gyrA (S84→L), gyrA (E88→K), and/or gyrA (E88→G), parC (E84→K), and/or parC (S80→Y)					5	13	1	4	2	11	14	16	9
gyrA with parE (n = 4)	gyrA (S84→L), parE (D432→N)					3	1				1	3	1	3

^aQRDR, quinolone resistance-determining region; MPC, mutant prevention concentration.

The reduction of vancomycin MICs and MPCs after combination with nemonoxacin are listed in Table 1. The reduction of vancomycin MPC occurred less frequently in isolates with gyrA plus parC mutations than in those without (76.0% versus 43.3%, χ² = 5.981, P = 0.014). Logistic regression analysis showed that gyrA (S84→L) and parC (E84→K) mutations were the two independent risk factors for the unchanged status of vancomycin MPCs in combination with nemonoxacin (P < 0.05). However, a similar phenomenon was not observed for vancomycin MICs.

Simulated PK profiles and PK/PD parameters.

The concentration-time curves were well simulated by the model. All observed concentrations in each experiment were in the range of 20% of the targeted values. The elimination half-life (t_1/2) values of vancomycin and nemonoxacin were 6.16 ± 0.56 h and 9.56 ± 0.33 h, respectively. The concentration-time curves obtained in the experiments evaluating bactericidal effect and resistance mutation prevention of vancomycin and nemonoxacin are shown in Fig. 1. The area under the concentration-time curve for the free, unbound fraction of vancomycin from 0 to 24 h (fAUC_0–24) (1.0 g given every 12 h [q12h]) was 273.32 ± 12.23.

FIG 1.

Pharmacokinetic (PK) simulation of vancomycin and nemonoxacin in the in vitro PK/PD model. (A) PK simulation monitoring of vancomycin in the experiment on bactericidal effect. (B) PK simulation monitoring of nemonoxacin in the experiment on bactericidal effect. (C) PK simulation monitoring of vancomycin in the experiment on resistance mutation prevention. (D) PK simulation monitoring of nemonoxacin in the experiment on resistance mutation prevention.

PD in the experiment on bactericidal effects.

To study the combined bactericidal effect of vancomycin and nemonoxacin, a PK/PD model was utilized to simulate regimens of vancomycin alone, nemonoxacin alone, and their combination over a 48-h period. Rapid bacterial killing (≥3-log₁₀ reductions in CFU/mL in 12 h) was observed in all the three tested isolates in the 1 g q12h scheme of vancomycin monotherapy, followed by gradual regrowth at 36 h, and maintaining bacterial counts of 10³ to 10⁴ CFU/mL at the 48-h endpoint. For isolates M04 and M23, the combination regimen of vancomycin (1 g q12h) and nemonoxacin (0.5 g once daily [qd]) showed a synergistic effect, and the bacteria counts reached detection limits at 36 and 48 h. The combination regimen also displayed a synergistic effect for M23 and an additive effect for M04 when the vancomycin dose (0.5 g q12h) was halved. However, for isolate M25, the combination regimen only showed a temporary synergistic effect within 12 h, and the bacteria recovered growth after 12 h and rose to 10⁵ CFU/mL at 48 h. The effects of different regimens on the bacterial burden are shown in Fig. 2.

FIG 2.

Antibacterial activity of vancomycin and nemonoxacin alone and in combination against M04 (A), M23 (B), and M25 (C) in the in vitro dynamic pharmacokinetic/pharmacodynamic model.

PD in the experiment on resistance mutation prevention.

The exact MPCs of M04, M23, and M24 were 19.2, 25.6, and 22.4 μg/mL, respectively. Figure 3 illustrates the time courses of killing and enrichment of resistant mutant bacterial population with the drug-free control, vancomycin monotherapy, and combination of vancomycin and nemonoxacin. After 7 days of exposure to vancomycin alone, three isolates maintained total bacterial counts of 10⁹ to 10¹⁰ CFU/mL with a growth pattern similar to that of the control. The mutations were observed in all three isolates. On 2×MIC vancomycin-containing agar plates, the resistant subpopulations increased to 10⁴ CFU/mL at 120, 96, and 72 h for M04, M23, and M24, respectively. The combination of vancomycin and nemonoxacin exhibited sustained bactericidal activity against isolates M04 and M23, with a >3-log CFU/mL reduction in the total bacterial counts. No mutations were observed for these two isolates during the 7-day experimental period. However, bactericidal activity was not observed for M24 in the combination regimens; the total and mutant bacterial populations were similar to that with vancomycin monotherapy. For all isolates, the mutant bacterial counts on 4×MIC vancomycin-containing agar plates were both below the limit of reliable detection.

FIG 3.

Anti-resistant mutation ability of vancomycin and nemonoxacin alone and in combination against M04 (A), M23 (B), and M24 (C) in the in vitro dynamic pharmacokinetic/pharmacodynamic model.

The MICs of the resistant subpopulations on 2×MIC vancomycin-containing agar plates were all 8 μg/mL. The enrichment times of M04, M23, and M24 on the 2×MIC agar plates were 120, 96, and 72 h, respectively, in the regimen of vancomycin alone. The increase in MIC did not occur during the 7-day experiment.

DISCUSSION

In the present study, we evaluated vancomycin, nemonoxacin, and their combination in an in vitro dynamic PK/PD model. We observed that the combination exhibited a bactericidal effect and suppressed resistance with simulating humanized dosing. To our knowledge, this is the first report to study the effect of nemonoxacin on vancomycin antibacterial activity and anti-resistant mutation ability.

Previous data have suggested that vancomycin is a time-dependent antibiotic with a postantibiotic effect, and a good bactericidal effect is achieved against MRSA when the ratio of AUC to MIC is ≥400. In the present study, the vancomycin MICs of the four isolates were all 2 μg/mL and AUC_0–24/MIC of vancomycin after 1 g q12h ranged from 257.95 to 286.13. In the vancomycin monotherapy, in spite of a rapid bacterial killing that was observed at first, the bacterial counts regrew at 10³ to 10⁴ CFU/mL within 48 h, which is consistent with previous studies (22, 23). Thus, our study further indicated that vancomycin monotherapy of 1 g q12h could not completely kill (without regrow) MRSA with a vancomycin MIC of 2 μg/mL.

Zhu (11) and colleagues first observed that an AUC_0–24/MPC ratio of >15 was a straightforward way to restrict the acquisition of resistance in a reference S. aureus strain ATCC 43300 in a rabbit infection model. In the present study, vancomycin monotherapy 1 g q12h produced an AUC/MPC ranging from 20.00 to 29.40, and AUC_0–24/MPC above 15 was still unable to restrict the emergence of the mutants for clinical isolates. There are possible reasons for this inconsistency. First, the PK process in humans could be well simulated in the in vitro models without considering species differences between humans and animals (24). Second, the reference strain and clinical isolates might also be different in the development of resistance. Based on our data, we propose that a higher vancomycin AUC_0–24/MPC ratio might be needed to restrict the acquisition of resistance in the clinical MRSA isolates.

The selection of an antibiotic combination is often based on the fractional inhibitory concentration index (FICI), and the combination showing synergy (FICI ≤ 0.5) is preferentially selected. MIC determination is actually based on monitoring viable bacteria at a constant antibiotic concentration, which may not fully reflect the interaction between the bacteria and in vivo dynamic process of antibiotics. We noted that the vancomycin MICs of M04 and M23 (FICI > 0.5 with vancomycin and nemonoxacin) were both decreased to 1 μg/mL in combination with an AUC_0–24/MIC of >400. Moreover, in the dynamic PK/PD model, the combination regimen with vancomycin and nemonoxacin showed a synergistic effect with a rapid bacterial clearance. Therefore, more attention should be given to a vancomycin MIC decrease rather than only focusing on FICI when choosing an antibiotic in combination with vancomycin to achieve PK/PD target.

As an NFQ, nemonoxacin has antibacterial activity against MRSA due to the replacement of the fluorine atom in the R-6 position with hydrogen and the insertion of methoxy at position C-8. Several studies comparing NFQs and the corresponding fluoroquinolones have reported that removing the fluorine at the 6-position systematically lowers the level of toxicity (14, 25–27). Moreover, phase I to III clinical trials and postmarketing clinical studies of nemonoxacin (28–31) have demonstrated that adverse events associated with nemonoxacin are mild, and some adverse events, like tendinitis, tendon ruptures, and irreversible peripheral neuropathy, listed in the box warning of the FDA for other fluoroquinolones, have not been observed. Our previous study showed that 43.7% of MRSA isolates with a vancomycin MIC of ≥2 μg/mL were sensitive to nemonoxacin, and vancomycin MIC and MPC values could be significantly decreased by nemonoxacin (18). In the present study, the dynamic time-kill studies showed the combination of vancomycin (1.0 g q12h) and nemonoxacin (0.5 g qd) not only produced a rapid and complete bactericidal effect but also obviously inhibited the mutation. These results indicate a potential of the combination for clinical isolates with high MIC (i.e., ≥2 μg/mL) and MPC (i.e., ≥19.2 μg/mL) values.

For fluoroquinolones, mutations in gyr and par genes are the key determinant of resistance in S. aureus (32). Nemonoxacin is a newly designed compound of NFQ. To our knowledge, there is only one study (33) on the mutations of genes in S. aureus responsible for resistance to nemonoxacin. It has been indicated that the hot spot mutations of QRDR in parC in Gram-positive cocci did not cause a decrease in the antibacterial activity of NFQ (25) and that S. aureus resistance to fluoroquinolones is dominated by gyrA mutations (20), but the mutations in gyrA have not been reported in nemonoxacin-resistant S. aureus strains. In the present study, we found gyrA mutations at positions of S84→L and E88→G or E88→K in clinical resistant isolates, suggesting that gyrA mutation may also be responsible for S. aureus resistance to nemonoxacin.

In the present study, we preliminarily explored the relationship between mutations in resistance genes and vancomycin MIC or MPC reduction caused by nemonoxacin. We observed that specific mutations of gyrA gene (S84→L) and parC gene (E84→K) were two independent risk factors for the unchanged vancomycin MPCs in the combination of vancomycin and nemonoxacin. This finding indicates that mutations of gyrA and parC not only decrease the sensitivity of MRSA to nemonoxacin but also reduce the effect of nemonoxacin on the anti-resistant mutation ability of vancomycin.

There are several limitations in the present study. First, the in vitro PK/PD infection model utilized in our study does not reflect the effect of the host immune system, which may affect the exposure-response relationship. Second, for isolates with lower MPCs, it is unknown whether mutations may not arise at all in the combination of vancomycin and nemonoxacin. We assume that the mutations can be suppressed entirely for the isolates with a lower MPC (<19.2 μg/mL) under the conventional dosing regimen of vancomycin with an AUC/MPC of >15 based on the study by Zhu et al. (11). However, we found that vancomycin monotherapy of 1 g q12h that produced an AUC/MPC above 15 was still unable to restrict the emergence of the mutants for clinical isolates. Thus, it remains unknown whether the combinations inhibit the mutations entirely in the isolates with lower MPCs. Third, despite the structural advantages with a low incidence of adverse events, as a newly approved drug only in China, the clinical application of nemonoxacin is still limited. More studies to further assess the safety and efficacy of combined vancomycin and nemonoxacin are warranted.

In conclusion, nemonoxacin has an enhancement effect on bactericidal activity and suppresses resistant enrichment ability of vancomycin for MRSA with an MIC of 2 μg/mL. Our in vitro data support the combination of nemonoxacin and vancomycin for treating MRSA infection with a high vancomycin MIC.

MATERIALS AND METHODS

Antimicrobials and media.

The vancomycin reference standard (purchased from the Control of Pharmaceutical and Biological Products, Beijing, China; no.130360-201302) and nemonoxacin reference standard (provided by Zhejiang Medicine Co., Ltd., Hangzhou, China, no. E0) were used in susceptibility testing. The reference standards were also used as the standards, respectively, for measuring vancomycin and nemonoxacin using high-performance liquid chromatography (HPLC), described below. To better simulate the medication of clinical patients, vancomycin active pharmaceutical ingredient (API) (no. 310190106) and nemonoxacin API (no. 205MP181204), both kindly supplied by Zhejiang Medicine Co., Ltd., were used for PK/PD experiments. Mueller-Hinton broth (MHB) and Mueller-Hinton agar (MHA) (Oxoid, Hampshire, England) were used for susceptibility testing and PK/PD experiments.

All drug solutions were freshly prepared and filtered using a 0.22-μm sterile filter membrane immediately before use.

Bacterial strains.

Fifty-five no-duplicate clinical isolates of MRSA with a vancomycin MIC of 2 μg/mL identified by Vitek 2 automated system (bioMérieux, Lyon, France) were obtained from patients of 20 tertiary general hospitals in China, and their MICs and MPCs were previously reported (18). All isolates were stored at −70°C with lyophilized stock culture to ensure parallel experimentation and subcultured on blood agar plates at 35°C for 24 h prior to each experiment.

According to our previous study (18), four isolates (M04, M23, M24, and M25) with a vancomycin MIC of 2 μg/mL confirmed by the standard agar dilution method (34) were selected in a modified in vitro dynamic PK/PD model. The MICs of nemonoxacin alone in these four isolates were 1 (M04), 2 (M23), 2 (M24), and 2 (M25) μg/mL, respectively. Specifically, three isolates, M04, M23, and M25 with a fractional inhibitory concentration index (FICI) of vancomycin plus nemonoxacin of 0.504, 1, and 2, were selected in the experiment on the bactericidal effects, which showed partial synergy (0.5 < FICI < 1), additive (FICI = 1), and indifferent (1 < FICI < 4) effects, respectively (35). Three isolates, M04, M23, and M24, all with a vancomycin exact MPC of ≥19.2 μg/mL (all with MPC of 32 μg/mL) but with different degrees to reduce the MPCs of vancomycin in combination, were selected in the experiment on resistance mutation prevention. The MPCs of vancomycin in combination with nemonoxacin for M04, M23, and M24 were 4, 16, and 32 μg/mL, respectively.

Screening for mutations in DNA gyrase and topoisomerase IV genes.

Genomic DNA of 55 clinical isolates was extracted using the bacteria genomic DNA extraction kit (Tiangen, Beijing, China) following the manufacturer’s protocol (36, 37). The primers of the target genes (gyrA, gyrB, parC, and parE) were designed using the Primer Premier 5.0 software based on the S. aureus strain Newman gene sequence deposited in the GenBank database (GenBank accession no. NC_009641) (Table S1 in the supplemental material). Genes were amplified by PCR under the following conditions: 35 cycles of denaturation at 96°C for 20 s, annealing at 52°C for 30 s, and extension at 72°C for 30 s, with a final elongation step at 72°C for 10 min. PCR products were purified using the magnetic bead PCR purification kit (Ensure Biologicals, Shanghai, China) and sequenced using an ABI 3730xl DNA sequencer (Applied Biosystems, Foster City, USA). The isolates were classified into different groups based on the QRDR mutations present in their DNA gyrase and topoisomerase IV genes.

In vitro dynamic PK/PD model.

The hollow-fiber infection model (HFIM) described in detail by our previous study (38) was modified in the following aspects in the present study. First, two peristaltic pump-controlling drug reservoirs were used for the combination model. Second, drugs and bacterial suspension were injected into the central reservoirs, and samples were taken from the central reservoirs to eliminate the drug concentration lag between the two compartments. Third, a computer control program was used for automatic sampling to collect samples conveniently and precisely. The volume of the central reservoirs was 400 mL. The fresh MHB was pumped into and removed from the central reservoir via two peristaltic pumps at the appropriate rate according to the PK parameters of antibiotics. In the combination model, the flow rate of the peristaltic pump controlling the eliminated reservoir was based on the antibiotic with shorter t_1/2. Corresponding drug reservoirs supplemented the antibiotic with longer t_1/2s. The entire system, except the computer, was placed in an incubator at 37°C. A starting inoculum of ∼10⁶ CFU/mL was targeted for the experiment on the bactericidal effects. Considering that mutant subpopulations are present at low frequencies with 10⁻⁶ to 10⁻⁸ (39), a bacterial suspension of 10⁸ CFU/mL was then injected into the central compartment and incubated until the bacteria reached a concentration of approximately 10⁹ CFU/mL, which was used as the initial inoculum in the experiment on the resistance mutation prevention. The central and eliminated reservoirs were shaken continuously to ensure proper mixing of the drug.

Simulated PK profiles.

In the experiment on the bactericidal effects, a total of four regimens, including (i) vancomycin alone (1 g q12h, 1 h infusion), (ii) nemonoxacin alone (0.5 g q24h, 1.5 h infusion), (iii) vancomycin (1g q12h, 1 h infusion) combined with nemonoxacin (0.5 g q24h, 1.5 h infusion), and (iv) vancomycin (0.5 g q12h, 1 h infusion) combined with nemonoxacin (0.5 g q24h, 1.5 h infusion), were evaluated on each of the three clinical isolates over a 48-h treatment period. In the experiment on the resistance mutation prevention, two regimens, including (i) vancomycin alone (1 g q12h, 1 h infusion), and (ii) vancomycin (1 g q12h, 1 h infusion) combined with nemonoxacin (0.5 g q24h, 1.5 h infusion), were evaluated on each of the isolates over a 168-h treatment period since an extended time period is required for mutation enrichment (38, 40, 41). The drug-free group was used as the control group.

Monoexponential profiles that mimicked q12h administration (1 h infusion) of vancomycin with an elimination half-life (t_1/2) of 6.0 h and q24h administration (1.5 h infusion) of nemonoxacin with a t_1/2 of 10.0 h were simulated. The serum protein-binding rates of vancomycin and nemonoxacin were 50% and 16%, respectively. The simulated free peak concentration (fC_max) of vancomycin (1 g) and nemonoxacin (0.5 g) were 18.00 μg/mL and 6.01 μg/mL, respectively (42, 43).

PK samples of vancomycin were collected at 0 (predose), 1, 2, 3, 4, 6, 8, and 12 h after the first dose, and before (trough concentrations) and after (peak concentrations) administration for each following dose. PK samples of nemonoxacin were collected at 0 (predose), 1, 1.5, 2, 3, 4, 6, 8, 12, and 24 h after the first dose before (trough concentrations) and after (peak concentrations) administration for each following dose. All samples were stored at −80°C until analysis. All observed concentrations should fall within the acceptable range of 80% to 120% of targeted values.

Samples analysis.

The concentrations of nemonoxacin were analyzed using the Shimadzu HPLC system (Kyoto, Japan). The liquid chromatographic separation process was conducted on a Shim-pack GIST C₁₈ column (4.6 by 250 mm; 5 μm; made in Japan). The mobile phase composition was 0.05 mol/liter sodium dihydrogen phosphate (adjusted to pH 5.0 by formic acid) and methanol with a ratio of 57:43 (vol/vol). The flow rate was 1.2 mL/min, and the detection wavelength was 295 nm. A 300-μL sample was mixed with 60 μL of the internal standard (sitafloxacin, 1.0 mg/mL) and then extracted with 240 μL precipitate solution (99:1 [vol/vol] ratio of methanol to formic acid) and centrifuged for 10 min at 15,000 × g. A total of 30 μL of the top layer was injected into the HPLC system. The standard calibration curves with good linearity were built within the concentration range of 0.6 to 30.0 μg/mL (y = 0.654214x − 0.00838102 [R² = 0.9999]). The lower limit of quantification was 0.06 μg/mL. The precision in intra- and interday assays was within 0.40% to 2.48%.

Vancomycin concentrations were quantified using two-dimensional HPLC (LC-20A; Shimadzu, Japan) as described in our previous study (44). The calibration curve was linear in the range of 1 to 40 μg/mL (y = 121156x − 16394 [R² = 0.9999]). The precision in intra- and interday assays was within 1.47% to 5.81%.

PD analysis.

The initial MPC values of the three strains were 32 μg/mL, which was derived from our previously published literature using the agar dilution method (18). The exact MPC was determined as described by Blondeau et al. (45). Briefly, bacterial cultures were enriched in broth to a concentration of 3 × 10¹⁰ CFU/mL. One hundred microliters of the bacterial suspension were evenly plated on a series of MHA plates containing a vancomycin concentration of 1×MPC (32.0 μg/mL), 9/10×MPC (28.8 μg/mL), 4/5×MPC (25.6 μg/mL), 7/10×MPC (25.6 μg/mL), 3/5×MPC (19.2 μg/mL), and 1/2×MPC (16 μg/mL) and incubated for 72 h at 35°C. The exact MPC was defined as vancomycin concentration with no visible bacterial growth on agar plates after incubation for 72 h (46).

In the experiment on the bactericidal effects, serial samples were obtained at 0, 1, 2, 3, 4, 6, 8, 12, 24, 36, and 48 h for quantification of bacteria. In the experiment on the resistant mutation prevention, samples were collected at 0, 6, 12, 24, 48, 72, 96, 120, 144, and 168 h. Total bacterial populations were quantified by inoculating the 10-fold serially diluted samples (100 μL) onto the blood plates using an automatic spiral plater (easySpiral; Interscience, Cantal, France) and a colony counter (Scan 300; Interscience) after incubation at 37°C for 18 to 20 h. Subpopulations with reduced susceptibility were quantified by plating 100-μL samples onto agar plates containing vancomycin at 2×MIC and 4×MIC in the experiment on the resistance mutation prevention. The limit of reliable detection was 50 CFU/mL, and each experiment was performed in duplicate.

Bactericidal, bacteriostatic, and effective activities were defined when CFU/mL decreased by ≥3 log, between ≥2 log and <3 log, and between ≥1 log and <3 log, respectively. Synergistic and addictive effects were defined when the CFU/mL decreased by ≥2 log and between ≥1 log and <2 log, respectively, in the combination of the two drugs. Mutation was defined when there was a ≥2-log increase in CFU/mL present on agar plates containing vancomycin at 2×MIC.

Statistical analysis.

The PK data were analyzed using Phoenix WinNonlin 6.0 software (Pharsight Co. Ltd., Missouri, USA). Comparisons of gene mutations were tested by the chi-square test. Possible risk factors of the unchanged MIC and MPC of vancomycin in combination with nemonoxacin were analyzed using logistic regression. All statistical analyses were performed using R version 3.6.2 (R Foundation for Statistical Computing, Vienna, Austria; http://www.R-project.org). A P value of <0.05 was considered statistically significant.