In vivo Pharmacokinetic/pharmacodynamic relationship of florfenicol in combination with doxycycline against Riemerella anatipestifer in ducks and the effect upon resistance development

Abstract

Antimicrobial chemotherapy is necessary to control Riemerella anatipestifer (RA), among which florfenicol (FF) is regarded as one of the preferred options. Based on the consideration of drug combination to improve efficacy, the pharmacokinetics and pharmacodynamics of FF combined with doxycycline (DOX) against RA were studied. FF was administered at doses of 20 or 40 mg/kg in combination with DOX (1, 2.5, 5, 10, or 20 mg/kg) via a single intramuscular injection (i.m.). DOX showed slow elimination in ducks with elimination half-life (T_1/2kel) in plasma, lung, and liver of 11.21, 11.53, and 13.01 h, respectively. A single dose of DOX (≥10 mg/kg) combined with FF (20 mg/kg) could exert a bactericidal effect on some tissues (heart, liver, spleen, lungs) in a model of RA strain CVCC3857 (minimum inhibitory concentration (MIC) of FF = 1 µg/mL, MIC of DOX = 2 µg/mL) infection within 24 h, and bactericidal effects (3.01–4.36 log₁₀ CFU/mL) were achieved in various tissues at a FF dose of 40 mg/kg. The AUC_24h/MIC of DOX combined with FF at 20 mg/kg required to produce a drop of 3 Log₁₀CFU/mL was 39.19 h (predicted dose of 25.03 mg/kg) and the value was 19.98 h (predicted dose of 12.76 mg/kg) when the dose of FF was 40 mg/kg. Combination of these two drugs could be used against insensitive strains (RA38 infection model with MIC of FF = 4 µg/mL, MIC of DOX = 2 µg/mL) by administering them twice for 24 h. Continuous passage under antibiotic pressure for 30 days suggested that resistance to FF was delayed in the presence of DOX. Genome resequencing and analyses of single-nucleotide polymorphisms revealed seven mutated genes (fahA, pfam, TonB-dependent receptor gene, proS, porU, RpiB). TonB-dependent receptor genes play a role in bacterial susceptibility. Additionally, both TonB-dependent receptor genes and fahA are involved in bacterial virulence and biofilm formation capabilities. Antimicrobial-treated strains were different from ancestor strains in terms of growth and virulence. Our study provides a data basis for the clinical use of FF and DOX against RA.

Introduction

The rapid development of resistance caused by long-term or irrational use of drugs seriously affects the efficacy of drugs (Liu et al., 2024; Yang et al., 2024). Sometimes, the effective control of disease using a single drug is challenging. Combined use of antibiotics means that drugs with different antibacterial mechanisms or antibacterial spectrum are used to treat the same infection simultaneously. This strategy has the advantage of expanding the antibacterial spectrum, improving antibacterial activity, and delaying the development of bacterial resistance. Studies have shown that a combination of florfenicol (FF) and doxycycline (DOX) has a synergistic effect on Escherichia coli, Pasteurella multocida, and Actinobacillus pleuropneumoniae (Rattanapanadda et al., 2019). In our previous study, the effect on strains with different sensitivity to FF (1–16 µg/mL) was enhanced by using a combination of FF and DOX in vitro (Zhang et al., 2022).

DOX is a tetracycline antibiotic with a bioavailability (F) of 39 % in ducks, strong tissue permeability and long elimination half-life in various animals including ducks (Anadon et al., 1994; Chapuis et al., 2021; Yang et al., 2015; Zhang et al., 2018). The 30S small ribosome subunit of bacteria is the main target of tetracycline drugs. This subunit forms reversible conjugated compounds so that transfer-RNA cannot bind to it and inhibit protein synthesis in bacteria. DOX is a bacteriostatic, and the effect of DOX at low concentrations is reduced because the binding of DOX to ribosomes is uncompetitive (Zhang et al., 2019).

FF is an analog of thiamphenicol. It exerts an antibacterial action by binding to the 50S small ribosome subunit of bacteria. FF possess a strong effect on RA, but FF alone cannot achieve a therapeutic effect if the sensitivity of the strain is reduced (Zhang et al., 2022). Moreover, FF is eliminated rapidly in ducks, even with a half-life of only 1.56 h (Tikhomirov et al., 2019; Zhang et al., 2024). A combination of FF with DOX (with its long half-life) can enhance the antibacterial effect and reduce the number of administrations required.

Some studies have shown that combining drugs can delay the development of bacterial resistance. When Actinobacillus pleuropneumoniae was cultured under the combined stress of FF and thiamphenicol, the frequency of bacterial resistance mutations was very low, similar to that of exposure to the drug alone at sub-inhibitory concentrations (Rattanapanadda et al., 2019). Liu and colleagues showed that a combination of daptomycin and rifampicin prolonged the acquisition of resistance to >20 cycles of each drug (Liu et al., 2020). A combination of two antibiotics can narrow or close the mutation selection window (MSW), which is considered to be one way to delay the development of resistance. Theoretically, the more synergistic the two antimicrobial effects are, the more likely is the combination likely to narrow or close the MSW (Xu et al., 2018; Shi et al., 2023). In addition, changes in the fitness of resistant bacteria affects the enrichment and elimination of resistant bacteria and, thus, affects the development of bacterial resistance (Vogwill and MacLean, 2015).

Riemerella anatipestifer (RA) is a pathogen that seriously endangers duck breeding. RA infection mainly causes disease and death in ducks aged from 1 week to 8 weeks (Gyuris et al., 2017). We investigated the effect of FF and DOX on RA in vivo and elucidated the effect of a combination of two drugs upon the development of bacterial resistance.

Materials and methods

Materials

The RA strains CVCC3952 and CVCC3857 were obtained from the Chinese Veterinary Microorganism Culture Collection Center (Beijing, China). FF injection (100 mL:10 g) and DOX injection (10 mL:1 g) were provided by Guangdong Dahuanong Animal Health Products (Yunfu, China). The powders of FF (purity ≥98 %) and DOX (purity ≥98 %) were purchased from Aladdin Industrial Corporation (Ontario, CA, USA). Tryptone soy agar (TSA) and tryptone soy broth (TSB; Guangdong Huankai Microbial Technology, Guangzhou, China) supplemented with 5 % newborn calf serum (Guangzhou Ruite Biotechnology, Guangzhou, China) were used to culture RA.

Ethical approval of the study protocol

The protocol for all animal studies was approved (2022A022) by the Animal Ethics Committee of South China Agricultural University (Guangzhou, China).

Animals and an infection model of RA

Seven-day-old shelducks (Tadorna tadorna) weighing 130–150 g were obtained from a commercial farm (Guangxi, China). A model of systemic infection in T. tadorna was established by an intraperitoneal injection of bacteria (10⁹ CFU/mL). The target bacterial load was reached 12 h after inoculation.

Quantification of DOX in plasma and tissues

The DOX concentration in plasma and tissues was measured by high-performance liquid chromatography–tandem mass spectrometry using a system from Agilent Technologies (Santa Clara, CA, USA). A mixture of acetonitrile and phosphate (98:2, v/v) was used as the extractant. Separation was achieved on a C₁₈ column (150 mm × 2.0 mm, 5 µm; Phenomenex, Torrance, CA, USA). The mobile phase comprised solution A (water containing 1/1000 formic acid) and solution B (acetonitrile) at a flow rate of 0.25 mL/min. The elution gradient was: 0 min, 10 % B; 0–1 min, 10 % B; 1–1.5 min, 90 % B; 1.5–4 min, 10 % B; and 10 min, 10 % B. The injection volume was 5 µL. A calibration curve was established with seven DOX concentrations (0.001–0.5 µg/mL).

Pharmacokinetics (PK) of DOX in RA-infected ducks

Ducks were divided randomly into five groups of 72. Twelve hours after RA inoculation, DOX was administered via intramuscular (i.m.) injection into the thigh at doses of 1, 2.5, 5, 10 and 20 mg/kg bodyweight, respectively. Samples of blood, lung tissue, and liver tissue were collected from all the five groups at 0.5, 1, 2, 4, 6, 8, 12, 24 and 36 h after each dose administration. Blood samples (2 mL) were collected from duck heart blood. Heparin was added to each blood sample, followed by centrifugation (3000 rpm, 10 min, 4°C) to separate plasma and stored at −20°C until assayed for doxycycline. After thawing, 0.5 mL of plasma was taken from each sample and mixed evenly. Then, 1 mL was taken for determination of the DOX concentration. For lung and liver tissues, 0.5 g of each sample was weighed and mixed with the homogenate, and 1 g was taken for determination of drug concentrations. The binding of DOX to plasma proteins in vitro was measured as 0.1, 1, and 10 μg/mL with an equilibrium-dialysis method for 72 h at 4 °C. DOX concentration–time data were analyzed using a one-compartmental model with a first-order absorption model by employing WinNonlin 6.1 (Pharsight, Mountain View, Sunnyvale, CA, USA). The formula is as follows:

$$C\left(t\right)=\frac{F•D•K_a}{V_d\left(K_a-k_{el}\right)}\left(e^{-k_{el}t}-e^{-K_{a}t}\right)$$

Where: C(t) is the plasma drug concentration at time t; F is the bioavailability; D is the administered dose; K_a is the absorption rate; K_el is the elimination rate; V_d is the apparent volume of distribution.

E is the antibacterial effect; E_max is the change in the model group (absence of drugs); E₀ is the maximum antibacterial effect; C_e is the PK/PD index; EC₅₀ is the corresponding PK/PD index that produces a 50 % reduction in the maximum antibacterial effect; N is the Hill coefficient.

In vivo pharmacodynamic (PD) study

The infection model of the CVCC3857 strain of RA (minimum inhibitory concentration (MIC) of FF = 1 µg/mL, MIC of DOX = 2 µg/mL) was used to study the effects of a single dose over 24 h. The study comprised 10 combination groups, four monotherapy groups, and one model group. Each group contained eight ducks. FF (20 or 40 mg/kg) was combined with DOX (1, 2.5, 5, 10, or 20 mg/kg). The four monotherapy groups comprised two FF groups (20 and 40 mg/kg) and two DOX groups (10 and 20 mg/kg). Ducks in the model group were injected with the same volume of drug diluent (glycerol:physiologic (0.9 %) saline = 40:60, v/v). The bacterial load in blood and tissues (heart, liver, spleen, lungs, kidneys, and brain) of the groups stated above was measured 24 h after administration. A series of sample diluents was plated onto a TSA plate containing 5 % calf serum and cultured at 37°C in an atmosphere of 5 % CO₂ for 24 h.

Further studies were conducted using the infection model of the RA38 strain of RA (MIC of FF = 4 μg/mL, MIC of DOX = 2 μg/mL) to investigate the effect of double administration for 24 h. The study comprised four combination groups (FF (20 mg/kg) + DOX (10 mg/kg); FF (20 mg/kg) + DOX (20 mg/kg); FF (40 mg/kg) + DOX (10 mg/kg); FF (40 mg/kg) + DOX (20 mg/kg), two DOX groups (10 and 20 mg/kg), two FF groups (20 and 40 mg/kg), and one model group. Eight ducks were in each group.

PK and PD analyses

The relationship between PK/PD indices and efficacy was investigated using an inhibitory sigmoidal maximum effect (E_max) PD model. Changes in bacterial abundance were obtained in in vivo PK/PD models by a fixed dose of FF in combination with different doses of DOX. The MIC of DOX against RA was determined by a broth-microdilution method. The PK/PD parameters of the area under the concentration–time curve over 24 h (AUC_{24 h}) and %T > MIC (cumulative percentage of time over a 24-h period in which the DOX concentration exceeded the MIC) were calculated using WinNonlin 6.1. The PK/PD parameters required for a bactericidal effect were calculated. The inhibitory sigmoidal E_max model was represented using the following equation:

$$E=E_{\max }-\frac{\left(E_{\max }-E_0\right)\times C_e^N}{E{C}_{50}^N+C_e^N}$$

Where: E is the antibacterial effect; E_max is the change in the model group (absence of drugs); E₀ is the maximum antibacterial effect; C_e is the PK/PD index; EC₅₀ is the corresponding PK/PD index that produces a 50 % reduction in the maximum antibacterial effect; N is the Hill coefficient.

Dose calculations

Based on the AUC_{24 h}/MIC calculated by the inhibitory sigmoid E_max model, the doses required to achieve different magnitudes of efficiency of DOX in the presence of FF could be estimated using the following equation (Toutain and Bousquet-Melou, 2002).

$$Dose\left(perdday\right)=\frac{{\left(AUC/MIC\right)}_{breakpo\text{int}}\times MI{C}_{distribution}\times C{l}_{perhour}}{f_u\times F}$$

where Cl/F is the body clearance scaled by intramuscular bioavailability (0.40 ± 0.08 L/kg/h in this study); AUC_{24 h}/MIC breakpoint is the PK/PD target for a bactericidal effect (39.19 h for FF at 20 mg/kg, 19.98 h for FF at 20 mg/kg); fu is the unbound fraction, which was determined to be 62.64 % using a protein-binding percentage in plasma of 37.36 %.

MIC and mutation preventive concentration (MPC)

RA (10⁵ CFU) was inoculated onto a TSA plate containing FF or DOX (0.125–16 μg/mL, 2-fold dilution). Then, it was inverted in an incubator containing 5 % CO₂ at 37 °C for 24 h. The minimum drug concentration at which no bacterial growth occurs is the MIC. MPC determination was done in reference to the method of Li and colleagues (Li et al., 2016). The suspension was adjusted to 10¹⁰ CFU/mL. Then, 100 μL was placed on a TSA plate containing 1, 2, 4, 8, 16, 32, or 64 MIC of drugs. This action was followed by culture in an incubator in an atmosphere of 5 % CO₂ at 37 °C for 5 days, and observed every 24 h. The minimum drug concentration without bacterial growth was recorded as the MPC. Experiments were carried out in triplicate.

Continuous passage under drug stress

RA in logarithmic growth phase was cultured in medium containing 1 × MIC of FF, 1 × MIC of DOX, or 1 × MIC of FF combined with 1 × MIC of DOX for 8 h. This action was followed by centrifugation (3000 rpm, 10 min, 4°C), resuspension in drug-free TSB culture medium, and culture for 24 h, which was considered to denote one cycle. Continuous passage was undertaken for 30 cycles. The MIC of bacteria against FF and DOX was determined every 24 h. Three replicates were set for each drug concentration (i.e., A1, A2, and A3). After the experiment, the MIC of the bacterial strains on various drugs was measured by a broth-microdilution method.

Whole‑genome resequencing (WGRS)

DNA was extracted using the TIANamp™ bacteria DNA kit according to manufacturer (Tiangen, Beijing, China) instructions. WGRS was done on the NovaSeq™ sequencing platform (paired-end, 2 × 150 bp; Illumina, San Diego, CA, USA). Sequencing data were aligned to the CVCC3857 genome to identify single-nucleotide polymorphisms (SNPs) as well as regions with insertions or deletions. The entire data were analyzed by Shanghai Personalbio Biotechnology (Shanghai, China). The sequence data files have been summitted to the National Centre for Biotechnology Information (NCBI) under BioProject accessions PRJNA1203054 (https://www.ncbi.nlm.nih.gov/bioproject). The Sequence Read Archive (SRA) of R. anatipestifer are publicly accessible under accessions SRR31829543-SRR31829552 (https://www.ncbi.nlm.nih.gov/sra/).

Evaluation of growth performance in vitro

The bacterial solution in logarithmic growth phase was diluted (1:1000) into TSB medium, and cultured at 180 rpm for 24 h at 37 °C. Every 2 h (0–14 h) and 24 h during culture, the bacterial solution (200 μL) was placed into a 96-well flat plate to determine the optical density (OD) at a wavelength of 600 nm. With the culture time as the horizontal coordinate and OD₆₀₀ as the vertical coordinate, a growth curve was drawn. Three parallel experiments were set.

Competitive index

The culture solution of the original strain and induced strain of bacteria were transferred to fresh TSB medium in the same proportion, cultured at 180 rpm for 24 h at 37 °C, and transferred to fresh TSB medium thrice. The cultures were counted at 0, 24, 48 and 72 h on TSA plates, and were also counted on TSA plates containing FF or DOX. Three replicates were set for each group. The competitive index (W) was calculated according to the following equation:

$$W=\frac{\ln (\mathit{RF}/\mathit{RI})}{\ln (\mathit{SF}/\mathit{SI})}$$

Where: RI and SI represent the number of resistant and sensitive bacteria in the mixture at 0 h; RF and SF represent the number of resistant and sensitive bacteria in the mixture at different time points, respectively. If W > 1, then the adaptability of resistant bacteria is stronger than that of sensitive bacteria; if W < 1, then sensitive bacteria dominate; if W = 1, then the competitiveness of bacteria is equal. With culture time as abscissa and competitive index as ordinate, a broken-line graph was drawn.

Virulence evaluation

The larvae of the greater wax moth (Galleria mellonella) were used to detect the toxicity of bacteria. The bacterial solution in logarithmic growth phase was centrifuged (3000 rpm, 10 min, 4°C) and washed thrice with 0.9 % saline, resuspended in 0.9 % saline, and the titer was adjusted to 10⁸ CFU/mL. Larvae stored at 4 °C were resuscitated at 37 °C for 2 h. Larvae of body length 25 ± 5 mm, weight 300 ± 50 mg, that were active with no black spots on their body surface were selected and placed into sterile culture dishes, with 10 larvae in each dish. Next, 10 μL of bacterial solution was injected (using a microsyringe) into the second hindfoot on the right side of the larvae, and the bacterial solution in the larvae was 10⁶ CFU/worm, respectively. The negative control group was injected with 0.9 % saline (10 μL/larva) only. Larvae were incubated at 37 °C for 72 h, and recorded every 12 h. The survival of all G. mellonella larvae in the control group was judged to be valid at all times tested. All strains were tested in triplicate at different times.

Results

PK of DOX in ducks

The detection limits of doxycycline in plasma, lung and liver tissues were all 0.005 μg/mL (or μg/g), and the quantification limits were 0.005 μg/mL, 0.005 μg/g, and 0.01 μg/g, respectively. Calibrated curves were constructed in the concentration range of 0.005–5 μg/mL (or μg/g) in plasma and lung tissue (R² >0.99), and in the concentration range of 0.01–10 μg/g in liver tissue (R² >0.99). Recovery and precision were determined for plasma, lung, and liver tissues containing DOX at 0.1, 1, and 10 μg/mL (or μg/g): the recovery in plasma was 95.02 %–113 %, with within-run relative standard deviations (RSDs) ≤7.83 % (n = 5) and between-run RSDs ≤7.46 % (n = 15); the recovery in lung tissue was 80.08 %–90.80 %, with within-run RSDs ≤5.04 % (n = 5) and between-run RSDs ≤6.79 % (n = 15); the recovery in liver tissue was 83.30 %–96.18 %, with within-run RSDs ≤11.74 % (n = 5) and between-run RSDs ≤8.86 % (n = 15). The concentration–time curves of DOX in RA-infected ducks are shown in Fig. 1, and the main parameters are presented in Table 1. The peak plasma concentrations (C_max) in plasma reached at 2.61 h (T_max, time of peak) after the administration. At a dose of 20 mg/kg, the C_max values in the plasma, lung, and liver were 1.86, 2.48, and 4.88 μg/mL (or μg/g), respectively. The Cl/F (clearance/bioavailability) of DOX in ducks was 0.40 ± 0.08 L/h/kg. The absorption half-time (T_1/2ka) and T_1/2kel of DOX in plasma were 0.60 h and 11.21 h, respectively. The T_1/2kel in lung and liver tissues was 11.53 h and 13.01 h, respectively. The protein-binding fraction of DOX at 0.1, 1, and 10 μg/mL was 37.84 %, 29.92 % and 44.33 %, respectively, with a mean value of 37.36 %.

Fig. 1.

Concentration–time curves of doxycycline in plasma, lung, and liver tissues following a single intramuscular injection (μg/mL or μg/g).

Table 1.

Pharmacokinetic parameters of doxycycline following a single intramuscular injection of various doses in the plasma, lung, and liver tissues of R. anatipestifer-infected ducks.

Dose (mg/kg)	1		2.5	5	10	20	Mean
Plasma
AUC24 (μg·h/mL)		3.48	8.11	12.82	22.28	37.11	-
T1/2ka (h)		0.63	0.82	0.86	0.35	0.34	0.60 ± 0.20
T1/2kel (h)		9.22	11.58	11.24	11.48	12.54	11.21 ± 0.99
Cl/F (L/h/kg)		0.29	0.31	0.39	0.45	0.54	0.40 ± 0.08
Tmax (h)		2.62	3.38	3.45	1.80	1.80	2.61 ± 0.66
Cmax (μg/mL)		0.22	0.40	0.64	1.21	1.86	-
Lung
AUC24 (μg·h/g)		4.18	10.04	18.92	29.49	36.25	-
T1/2ka (h)		0.50	0.65	0.36	0.28	0.32	0.42 ± 0.14
T1/2kel (h)		11.40	12.25	13.40	11.54	9.04	11.53 ± 1.43
Tmax (h)		2.35	2.91	1.94	1.56	1.60	2.07 ± 0.51
Cmax (μg/g)		0.22	0.48	0.88	1.61	2.48	-
Liver
AUC24 (μg·h/g)		10.02	19.58	36.79	65.72	125.58	-
T1/2ka (h)		0.55	0.47	0.67	0.34	0.32	0.47 ± 0.13
T1/2kel (h)		10.34	12.25	13.06	12.93	16.48	13.01 ± 1.99
Tmax (h)		2.45	2.31	3.03	1.84	1.86	2.30 ± 0.44
Cmax (μg/g)		0.57	0.97	1.66	3.19	4.88	-

AUC, area under the concentration–time curve; T_1/2ka, absorption half-time; T_1/2kel, elimination half-life; Cl, clearance; F, bioavailability; T_max, time of peak; C_max, peak concentration.

In vivo PD of DOX combined with FF against RA in a single dose

The bacterial quantity in blood and tissues of the model group increased by 0.04–0.93 log₁₀ CFU/mL at 36 h, compared with that at 12 h, after infection. When DOX (20 mg/kg) or FF (40 mg/kg) were used alone, the maximum reduction of bacterial count was 1.24 log₁₀ CFU/mL and 1.66 log₁₀ CFU/mL, respectively (Fig. 2a). The effect of different doses of DOX combined with FF (20 mg/kg) is shown in Fig. 2b. When DOX was administered at >10 mg/kg, the bacterial count in brain and kidney tissues decreased by ≥2 log₁₀ CFU/mL within 24 h of interaction with bacteria; a bactericidal effect (3.16–3.35 log₁₀ CFU/mL) could be achieved in heart, liver, spleen, and lung tissues. When combined with FF (40 mg/kg), bactericidal effects (3.01–4.36 log₁₀ CFU/mL) were achieved in various tissues at a DOX dose >10 mg/kg (Fig. 2c).

Fig. 2.

In vivo pharmacodynamics of florfenicol and doxycycline against R. anatipestifer following a single intramuscular injection in 24 h (n = 8). a, Florfenicol alone or doxycycline alone. b, Florfenicol (20 mg/kg) combined with doxycycline at different doses. c, Florfenicol (40 mg/kg) combined with doxycycline at different doses.

PK/PD analyses

The E_max relationships of three PK/PD parameters versus the antibacterial effect are depicted in Fig. 3. When the dose of FF was 20 mg/kg, the correlation coefficient (R²) of AUC_{24 h}/MIC, C_max/MIC, and %T > MIC with antibacterial effects was 0.918, 0.914, and 0.750, respectively. These results suggested that the effect of DOX (combined with FF) against R. anatipestifer was concentration-dependent and the efficacy was driven by AUC_{24 h}/MIC and C_max/MIC. The AUC_{24 h}/MIC and C_max/MIC for eliciting a reduction of 3 log₁₀CFU/mL were 58.56 h and 15.10, respectively (Table 2). When the FF dose was 40 mg/kg, the AUC_{24 h}/MIC and C_max/MIC required to decrease the bacterial population by 3 log₁₀ CFU/mL were 19.98 h and 2.2, respectively (Table 3).

Fig. 3.

Relationships of florfenicol in combination with doxycycline for PK/PD parameters versus antibacterial effect after inhibitory sigmoid E_max simulation. a, b, c Florfenicol dose = 20 mg/kg. d, e, f Florfenicol dose = 40 mg/kg.

Table 2.

Estimation of PK/PD parameters of the combinations of florfenicol (20 mg/kg) and doxycycline against R. anatipestifer derived from the E_max model.

PK/PD parameters of doxycycline	AUC24h/MIC (h)	Cmax/MIC	T% > MIC
Emax (Log10 CFU/mL)	0.53	0.54	0.00
EC50	8.83	0.57	4.47
E0 (Log10 CFU/mL)	−3.98	−4.34	−3.82
Hill's slope	0.86	0.88	0.53
R2	0.918	0.914	0.750
Decrease in 3 Log10 CFU/mL	39.19	1.72	51.66

Table 3.

Estimation of PK/PD parameters of the combinations of florfenicol (40 mg/kg) and doxycycline against R. anatipestifer derived from the E_max model.

PK/PD parameters of doxycycline	AUC24h/MIC (h)	Cmax/MIC	T% > MIC
Emax (Log10 CFU/mL)	0.06	0.06	0.04
EC50	16.97	0.46	13.78
E0 (Log10 CFU/mL)	−4.76	−3.79	−5.70
Hill's slope	0.63	0.89	0.03
R2	0.912	0.913	0.566
Decrease in 3 Log10CFU/mL	19.98	2.20	-

In vivo PD of DOX combined with FF against RA administered twice in 24 H

The effect of FF and DOX against RA following intramuscular injection twice in 24 h is shown in Fig. 4. DOX alone reduced bacterial counts by ≤1 log₁₀ CFU/mL. When FF was administered (20 or 40 mg/kg), the maximum reduction in tissues was 1.63 log₁₀ CFU/mL and 2.55 log₁₀ CFU/mL, respectively. FF (40 mg/kg) combined with DOX (10 mg/kg) reduced the bacterial count in all tissues by >2 log₁₀ CFU/mL. FF (40 mg/kg) combined with DOX (20 mg/kg) reduced the bacterial count in all tissues by >3 log₁₀ CFU/mL except brain tissue.

Fig. 4.

In vivo pharmacodynamics of florfenicol and doxycycline against the R. anatipestifer strain RA38 following two intramuscular injections in 24 h (n = 8). a, Florfenicol alone or doxycycline alone. b, Florfenicol in combination with doxycycline.

MIC and MPC Of FF and DOX against RA

The MIC and MPC of FF against the RR strain CVCC3857 were 1 μg/mL and 4 μg/mL, respectively. The MIC and MPC of DOX against the RA strain CVCC3857 were 1 μg/mL and 8 μg/mL, respectively.

Susceptibility changes of continuous passage

After 30 days of continuous passage under drug pressure, the MIC of the FF group and DOX group increased 4-fold compared with that of the original strain, and the MIC of the FF combined with DOX group increased 4-fold compared with the that of the original strain, while the MIC of FF remained unchanged (Fig. 5). Simultaneously, the MIC of the strain against chloramphenicol and thiamphenicol increased 2-fold, but the MIC against tetracycline and chlortetracycline remained unchanged (Table 4).

Fig. 5.

MIC changes of florfenicol and doxycycline against the R. anatipestifer CVCC3857 strain in serial passages during 30 days.

Table 4.

MICs of R. anatipestifer after serial passage for 30 days treated with different antibacterial agents (μg/mL).

Strain	Florfenicol	Thiamphenicol	Chloramphenicol	Doxycycline	Tetracycline	Chlortetracycline
FF-1	4	8	4	1	8	8
FF-2	4	8	4	1	8	8
FF-3	4	8	4	1	8	8
DOX-1	1	4	2	4	8	8
DOX-2	1	4	2	4	8	8
DOX-3	1	4	2	4	8	8
FF+DOX-1	1	8	4	4	8	8
FF+DOX-2	1	8	4	4	8	8
FF+DOX-3	1	8	4	4	8	8
3857(original)	1	4	2	1	8	8

Mutated genes of RA after serial passage for 30 days treated with different antibacterial agents

The genomes of nine strains induced for 30 generations and the primary strain were extracted, and SNP analyses were undertaken. Six mutant genes were identified in the genomes of nine antimicrobial-induced strains (fahA, pfam, TonB-dependent receptor gene, proS, porU and RpiB) encoding products involved in enzyme synthesis, antimicrobial sensitivity, virulence and biofilm synthesis (Table 5, Table 6).

Table 5.

Mutated genes of R. anatipestifer after serial passage for 30 days treated with different antibacterial agents.

Group	Mutated genes
	RA1	RA2	RA3
FF	fahA, pfam, TonB-dependent receptor gene	fahA, pfam, TonB-dependent receptor gene	fahA, pfam, TonB-dependent receptor gene
DOX	fahA, pfam, TonB-dependent receptor gene, RpiB	fahA, pfam, TonB-dependent receptor gene, proS, porU	fahA, pfam, TonB-dependent receptor gene, proS, porU
FF+DOX	fahA, pfam, TonB-dependent receptor gene	fahA, pfam, TonB-dependent receptor gene	fahA, pfam, TonB-dependent receptor gene, RpiB

Table 6.

Changes in amino acids of the mutated genes of Escherichia coli after serial passage for 30 days treated with different antibacterial agents.

Mutated gene	Mutation type	Mutated position	Function	Base change	Change in amino acids
fahA	stoploss	35548	fumarylacetoacetase	T1099G	X367E
pfam	nonsynonymous SNV	132148	peptidase M14	C406A	P136T
TonB-dependent receptor gene	nonsynonymous SNV	146518	outer membrane beta-barrel family protein	T109G	C37G
RpiB	stoploss	63023	ribose 5-phosphate isomerase B	T184G	X62G
proS	nonsynonymous SNV	37020	proline-tRNA ligase	A250G	T84A
porU	nonsynonymous SNV	180252	type IX secretion system sortase PorU	A682G	N228D

Growth curves

Compared with the original strains, the growth rate of induced strains was similar within 12 h, but the bacterial count of some strains was higher after 24 h (Fig. 6).

Fig. 6.

Growth curve of test strains.

In vitro relative fitness of antibiotic-induced strains

Within 72 h of culture, the competitive index of strains in the DOX group and FF combined with DOX group showed a decreasing trend (Fig. 7). The final competitive index of nine induced strains (FF-1, FF-2, FF-3, DOX-1, DOX-2, DOX-3, FF+DOX-1, FF+DOX-2, FF+DOX-3) was 0.92, 1.08, 0.95, 0.99, 0.29, 0.29, 0.28, 0.33, and 0.28, respectively.

Fig. 7.

Relative fitness of R. anatipestifer strains.

Survival curve of g. mellonella larvae

Within 72 h of the test, the mortality of each group is shown in Fig. 8. The percentage of larvae surviving in the FF-1 group was 93 %, and all larvae survived in FF-2 and FF-3 groups. The percentage of larvae of three strains surviving in the DOX group was 67 %, 73 % and 60 %, respectively. The percentage of larvae of three strains in the FF combined with DOX group was 83 %, 93 % and 100 %, respectively. Thirty percent of larvae of the original strain (CVCC3857) survived. None of the G. mellonella larvae in the saline control group died.

Fig. 8.

Survival curve of G. mellonella larvae (n = 30, per group).

Discussion

In veterinary practice, the correct use of PK/PD parameters can optimize the dose and interval of antimicrobial agents, reduce the drug dose while achieving an optimal therapeutic effect, and delay the emergence of resistance. For concentration-dependent drugs, the “ideal” concentration in plasma or tissue should be 10–12-times the MIC (Yan et al., 2017). Time-dependent drugs can be administered repeatedly to maintain drug concentrations above the MIC but, to avoid drug toxicity, reasonable dosing intervals should be designed. It is difficult to design dosing intervals and doses for time-dependent drugs with short T_1/2kel (Fair and Tor, 2014). It is safer to give more toxic antibiotics ≥8-h apart (Lappin et al., 2017).

In our previous study, FF was eliminated rapidly in ducks with a T_1/2kel of 1.67 h, whereas it has a longer of 8.34 h T_1/2kel in chickens after oral administration (Anadón, et al., 2008; Zhang et al., 2024) DOX has a longer half-life of 11.21 h in ducks than 6.03 h in chickens (Anadon, et al., 1994) Studies indicated that the presence of DOX results in slower metabolism of FF in pigs by inhibiting the metabolic enzymes CYP3A29 and competing for binding to metabolic sites (Xu et al., 2022). DOX slightly slowed down FF metabolism by inhibiting CYP3A24 in goats and prolonged the withdrawal period of FF in the kidney from 17.18 days to 18.61 days (Wang et al., 2021). CYP 3A also played a key role in the PK of FF in chickens (Wang et al., 2018). In this study, DOX may affect the metabolism of FF in ducks. For pharmacodynamics, the maximum effect of a single dose against RA in the present study within 24 h was 1.24 log₁₀ CFU/mL (Fig. 2a). In vitro studies have shown that FF combined with DOX can enhance the antibacterial effect (Zhang et al., 2022). In our study, FF combined with DOX could achieve a bactericidal effect within 24 h after a single dose. The mechanism may be that the combination of DOX and FF enhances the antibacterial effect, and the persistence of DOX (which has a longer T_1/2kel) inhibits the re-growth of bacteria during FF below the effective concentration. In addition, the significance of combination is to expand the antibacterial spectrum, reduce the dose and delay the development of bacterial resistance.

Twenty-four hours after DOX (10 and 20 mg/kg) had been administered, the bacterial count in blood and tissues decreased by 0.59–1.86 log₁₀ CFU/mL and 0.00–2.59 log₁₀ CFU/mL, respectively. Zhang and colleagues reported the C_max of DOX in pigs infected with Haemophilus parasuis to be 4.31 μg/mL, and the AUC _{24 h}/MIC corresponding to the predicted bactericidal level was 98 h (Zhang et al., 2018). DOX reduced the number of M. hyopneumoniae by 1 log₁₀ CFU/mL with AUC_{24 h}/MIC and C_max/MIC of 164 h and 9.89, respectively (Zhang et al., 2019). The low plasma concentration of DOX at a dose of 10 mg/kg prevents it from achieving the suggested PK/PD parameters for Streptococcus species (Chapuis et al., 2021). At a dose of 20 mg/kg, the C_max/MIC of DOX against RA in plasma was 1.86 μg/mL and AUC_{24 h}/MIC was 37.11 h in this study, respectively, suggesting that it was difficult to achieve the PK/PD targets reported. DOX (10 mg/kg) combined with FF (20 mg/kg) could reduce the bacterial count in lung tissue by 3.35 log₁₀ CFU/mL in 24 h. DOX (10 mg/kg) combined with FF (40 mg/kg) could achieve a bactericidal effect in heart, liver, spleen, lung, kidney, and brain tissues. The characteristics of DOX against RA in the presence of FF tended to be dependent upon the concentration. When DOX was combined with FF (20 or 40 mg/kg), the AUC_{24 h}/MIC required to reduce the bacterial population by 3 log₁₀ CFU/mL was 39.19 h and 19.98 h, which corresponded to predicted doses of 25.03 mg/kg and 12.76 mg/kg, respectively.

The effect of drug combination on the development of bacterial resistance was studied by exposing RA to FF and DOX for 30 days to observe changes in bacterial susceptibility and the generation of resistance genes. The MIC was increased 4-fold in DOX and FF groups, from generation 4 and generation 2, respectively. There was no change in sensitivity to FF in the combination group. Drug combinations can limit the evolution of resistance (Rodriguez de Evgrafov et al., 2015). TonB-dependent receptor genes have been reported to influence bacterial susceptibility by affecting the uptake of essential nutrients, including iron. Iron carriers bound to antimicrobial molecules can overcome the low permeability of the outer membrane of Gram-negative bacteria (McPherson et al., 2012; van Delden et al., 2013; Luscher et al., 2018; Li et al., 2021). TonB-dependent receptor genes affect bacterial virulence from the aspects of growth rate, spread speed, and biofilm-formation ability (Abdollahi, Rasooli, and Mousavi Gargari, 2018). fahA encodes the metabolic enzyme fumarylacetoacetase, which may play a part in bacterial motility and biofilm formation (Laganenka et al., 2020; Jiang et al., 2021). Mutations in fahA, pfam, and TonB-dependent receptor genes were found in all groups, and no other reported reasons related to bacterial resistance were found. The development of bacterial resistance is a complex process and is regulated by multiple genes. In addition, a large number of mutant genes with no demonstrated function require further study and analysis.

Growth curves, competitive index, and virulence were used to study the fitness of bacteria. There was no obvious difference between the induced strain and original strain in terms of the growth rate; the final bacterial count of some strains was higher than that of the original strain. The competitive index of the DOX group and combination group dropped to 0.29 and 0.28, respectively, at 72 h, indicating that there was a fitness cost in the induced strains, and the competitive ability with sensitive bacteria was reduced in the process of bacterial evolution. G. mellonella larvae were used to evaluate the virulence of RA (Liu et al., 2019), and all the strains induced by the drug showed a decrease in virulence.

Conclusions

The PK of DOX and PD of DOX combined with FF against RA were studied. DOX possessed a long T_1/2kel (11.21 h) in plasma, and the C_max was 1.86 μg/mL. At a single 24-h dose, combination with FF (20 mg/kg) reduced the bacterial count by 3 log₁₀ CFU/mL with an AUC_{24 h}/MIC of 39.19 h; combination with FF (40 mg/kg) reduced the bacterial count by 3 log₁₀ CFU/mL with an AUC_{24 h}/MIC of 19.98 h. The results suggested that the therapeutic dosing strategies of FF 20 mg/kg combined with DOX 25.03 mg/kg or FF 40 mg/kg combined with DOX 12.76 mg/kg could achieve bactericidal effect in infections caused by R. anatipestifer in field conditions. A combination of these two drugs administered twice over 24 h could be used against less susceptible strains. Continuous passage for 30 days under antibiotic pressure revealed mutations in fahA, pfam, and TonB-dependent receptor genes in all groups. TonB-dependent receptor genes were reported to influencebacterial susceptibility and fahA was associated with virulence and the biofilm-formation ability. The presence of DOX could delay the development of resistance to FF. The decrease in the competition index and virulence of drug-induced strains means that the competitive ability with sensitive bacteria was reduced in the process of bacterial evolution.

Funding

This work was supported by Biological and Medical Sciences of Applied Summit Nurturing Disciplines in Anhui Province (Anhui Education Secretary Department [2023]13).

Declaration of competing interest

The authors declare no conflicts of interest.