In Vivo Activity Of Cotrimoxazole

TITLE: In vivo activity of Cotrimoxazole combined with colistin against Acinetobacter baumannii producing OXA-23 in a Galleria mellonella model

1. Introduction:

Acinetobacter baumannii (A. baumannii) is considered as a critical nosocomial pathogen causing several infections, including bloodstream infections, ventilator- associated pneumonia, urinary tract infections and surgical site infections (Falagas et al., 2015).

It has become a more serious challenge due to their great ability to develop resistance to a wide-range of antimicrobial classes, including carbapenems (Bae et al., 2016).

Limited newer antimicrobial agents as well as restricted therapeutic options for multidrug resistant A. baumannii (MDRAB) infections, render their controlling a great concern.

Such challenge attracted attentions to recall older antibiotics to the treatment arsenal until new options appear [6] (Falagas et al., 2015). In addition, the use of combination therapies has broadened for combating such pathogen (Hornsey et al., 2013).

One lineage of MDRAB known as ‘OXA-23 clone 1’ has been reported as the most prevalent clone among A. baumannii clinical isolates in Egypt With exception of colistin and tigecycline, members of this clone revealed high level of resistance to most of the antimicrobial agents. (Ghaith et al., 2017; Al-Agamy et al., 2014).

In consistence, many reports documented the importance of colistin as an effective therapy for management of MDRAB (Hornsey et al., 2013). Colistin sulphate, is a cyclic polypeptide with detergent-like characteristics. Colistin binds and interacts with lipopolysaccharidesc and phospholipids at the surface of the bacterial outer membrane, disturbing the cytoplasmic membrane permeability. (Vidaillac et al., 2012). This interaction may result in osmotic balance disruption and leakage of the cellular contents (Soon et al. 2011). Colistin effect on the outer membrane can lead to increased permeability to large and/or hydrophobic molecules like cotrimoxazole (Vidaillac et al., 2012)

Several studies have mentioned the potential for synergistic activities of unusual colistin combinations (Wareham et al., 2011; Gordon et al., 2010). Vidaillac et al., 2012). Despite the up growing demand to use old existing antibiotics, the in vivo bactericidal effect of cotrimoxazole and its potential synergistic activity in combination with colistin against MDRAB has not been assigned. Cotrimoxazole is a combination of trimethoprim/sulfamethoxazole (TMP-SMX) in 1/19 concentration ratio to achieve the maximal synergistic activity between both drugs. It inhibits synthesis of bacterial DNA through dihydrofolate pathway inhibition (Goldberg and Bishara 2012). Cotrimoxazole has been widely used, for about half a century, for treatment of both Gram-positive and Gram-negative bacterial infections. Currently, cotrimoxazole has not been a drug of choice for treatment of MDRAB (Karageorgopoulos &Falagas. 2008).

Animal model were greatly used to evaluate the possible efficacy of antimicrobial combinations. For this purpose, mammalian models are usually used as the data obtained are most applicable to human infections. However, mammal infection models are costly and time consuming with real ethical concerns. (Hornsey et al., 2013). Recently, invertebrate infection models as Galleria mellonella, (the larva of the wax moth), has been used for primary confirmation of in-vivo efficacy of antimicrobial agents [Peleg et al., 2009; Antunes et al., 2012] (Hornsey et al., 2013)

The aim of this study was to evaluate the probable synergy and bactericidal activity of colistin/cotrimoxazole combination against MDRAB. Here, a well-identified clinical isolate of MDRAB harbouring blaOXA-23 has been used to study the in vivo activity of colistin /cotrimoxazole combination on a Galleria mellonella infection model.

2. Materials and Methods

2.1. Bacteria and antimicrobial susceptibility testing

Four clinical isolates of A. baumannii (AB1-4) were used as a MDRAB isolate belonging to the widespread epidemic lineage in Egypt; blaOXA-23. Isolates were identified as A. baumannii by conventional microbiological methods and confirmed by species-specific PCR for the blaOXA-51-likegene [Turton et al., 2006]. Descriptive data of the studied isolates are illusterted in table 1. Genes encoding resistance to carbapenems [blaOXA-23-like, blaOXA-24-like, blaOXA-51-like, blaOXA-58-likeand metallo—lactamase (MBL) genes] were identified by PCR and sequencing as previously described [Hornsey et al., 2012]. Escherichia coli ATCC 25922 and A. baumannii ATCC 19606 were used as quality controls in all susceptibility assays.

2.2. Antimicrobial agents.

Colistin sulfate, trimethoprim, and sulfamethoxazole were commercially obtained (Sigma-Aldrich, Saint Quentin Fallavier, France). Each agent was freshly prepared according to the CLSI guidelines in the appropriate solvent (Clinical and Laboratory Standards Institute. 2009).

2.3. Susceptibility testing.

Minimum inhibitory concentrations (MICs) for cotrimoxazole (at a ratio of 1:19) and colistin were determined in duplicate by the Clinical and Laboratory Standards Institute (CLSI) reference broth microdilution method [Clinical and Laboratory Standards Institute, 2014]. Susceptibility was determined using CLSI breakpoints. Results of colistin susceptibility were interpreted in accordance with CLSI criteria (susceptible, ≤2 mg/L;resistant, ≥4 mg/L). As neither EUCAST nor CLSI have breakpoints for trimethoprim or sulfamethoxazole versus A. baumannii, we applied the CLSI resistance breakpoints for Enterobacteriaceae (for trimethoprim [susceptible, 8 g/ml; resistant, 16 g/ml] and sulfamethoxazole [susceptible, 256 g/ml; resistant, 512 g/ ml]) (European Committee on Antimicrobial Susceptibility Testing. 2015, Livermore et al., 2014, Nepka et al., 2016).

2.4. Synergy testing by the chequerboard assay

Synergy between cotrimoxazole (1/19) and colistin was assessed using the microtitre plate checkerboard assay as described previously [Hornsey et al., 2012]. Briefly, 96-well microtitre plates were set-up with increasing concentrations of cotrimoxazole(1/19) at concentrations varying from 0.25 to 128 g/ml in the horizontal wells and colistin (0–4 mg/L) in the vertical wells and were inoculated with 105 CFU/mL of A. baumannii prepared in LB broth. Plates were visually assessed after 24 h of incubation at 35°C for turbidity to determine growth. Synergy was assessed by calculation of the fractional inhibitory concentration index (FICI) and the susceptible breakpoint index (SBPI) as previously described [Milne et al; 2010]. FIC values were interpreted according to the following criterion: the potential for bacteriostatic effect, meaning the potential for synergy when FICI of ≤0.5 and a SBPI of >2(11). FICI = (MIC of A in combination/MIC of A) + (MIC of B in combination/MIC of B)The indices were interpreted as: an FICI of ≤0.5 =synergy; an FICI of≥ .0.5 and ≤4.0 = no interaction; and an FICI of ˃.4 =antagonism.7 SBPI = (susceptible breakpoint A/MIC of A in combination) + (susceptible breakpoint B/MIC of B in combination) Milne and Gould, 2009. All experiments were carried out in triplicate, and the results were described as the mode values.

2.5. Time–kill assays and bactericidal activity

Time–kill assays were conducted for each strain using cotrimoxazole (1/19) alone, colistin alone and a combination of both drugs according to a previously described methodology [Phee et al., 2013]. In brief, colistin was added at a final concentration of 1 mg/L and cotrimoxazole (1/19) at 10 mg/L. The concentration of colistin and cotrimoxazole (1/19) was selected as one that is achievable as the steady-state plasma concentration when pharmacokinetically optimized dosing is used [Markou et al., 2008; Dvorchik et al., 2010]. Antimicrobial regimens consisted of multiples of the MIC (0.25 and 0.5 MIC) of each agent alone or in combination were used.LB broth was inoculated with 5 ×105 CFU /mL of the A. baumannii strain and was incubated at 37°C. Colony counts were obtained at 0, 2, 4, 8 and 24 h to determine the viable CFU/mL. For all time-kill experiments, aliquots (100 µl) were serially diluted in cold and sterile normal saline. Bacterial counts were determined by plating three spots of 10 µl of appropriate dilutions on MHA plates and incubating them at 35°C for 18 to 24 h. Time-kill curves were then designed by plotting mean colony counts (log 10 CFU/ml) versus time. The bactericidal activity was defined as = 3 log10 CFU/ml reduction in the colony count relative to the initial inoculum. Synergy was interpreted as ≥ 2 log10 decrease in CFU/ml by the drug combination when compared with its most active constituent, and = 2 log10 decrease in the CFU/ml below the initial inoculum, at any time point (Principe et al., 2009).

2.6. Galleria mellonella treatment assays

The G. mellonella infection model for A. baumannii was adapted from that proposed by Peleg et al. [Peleg et al., 2009]. Batches of G. mellonella caterpillars (KaideRuixin Co., Ltd., Tianjin, China) in their final instar stage were were stored in the dark on wood shavings at 15 ◦C prior to use. Caterpillar masses varied slightly but were typically 250 mg and this value was used to calculate treatment doses. To establish the inoculum required to kill G. mellonella over 24–96h, 16 caterpillars were inoculated with 10 µL of bacterial suspensions containing 5x 105CFU/larva of bacteria in phosphate-buffered saline (PBS). Bacteria were injected into the haemocoels through the last left proleg using a 50 µL Hamilton syringe (Hamilton, Shanghai, China). Caterpillars were incubated at 37°C and were observed daily for 4 days. Antibiotics were administered via 10 µL injections into the last right proleg within 2 h of inoculation. The following doses, simulating human doses, were used: 2.5 mg/kg for colistin and 4 mg/kg for cotrimexazol. Treatment was given only once. Sixteen uninoculated and mock-inoculated (sterile PBS) caterpillars were used as controls. The caterpillars were observed for survival every 24 h for 4 days. Experiments were performed three times on separate occasions.

2.5. Statistics

All statistical analyses were performed using GraphPad Prismv.5.04 (GraphPad Software Inc., San Diego, CA). Survival curveswere analysed using the log-rank test, with a P-value of ≤0.05considered statistically significant.