We have adapted our methodology for assessment of susceptibility or resistance to quinolones [14–16] to develop a simple and rapid procedure to determine susceptibility or resistance to antibiotics that act at the cell wall. This objective may have great application since these are by far the most frequent antibiotics employed all over the world .
To assess quinolone activity, the lysing conditions of bacteria were designed to remove the cell wall and membranes from all the cells equally, leaving the cell wall affected or intact [14, 15]. In the case of antibiotics that act at the cell wall, the lysis must be adapted to only affect to those bacteria whose cell wall had been damaged by the antibiotic. These bacteria would have a debilitated wall, which would be much more sensitive to the lysing conditions designed to such an effect. The lysis affects the cells differentially, depending on the integrity of the wall. If the bacterium is susceptible, the weak cell wall is affected by the lysing solution so that the nucleoid of DNA contained inside the bacterium is released and spread. On the other hand, if the bacterium is resistant to the antibiotic, it would be virtually unaffected by the lysis solution and does not liberate the nucleoid, remaining essentially with its usual morphological appearance.
The present work describes a logical sequence of experiments to achieve the objective of developing a simple and rapid procedure to determine susceptibility or resistance to antibiotics that act at the cell wall. Firstly, it was necessary to demonstrate the ability of the procedure to discriminate susceptible, intermediate and resistant strains. This was confirmed in clinical E. coli strains. As a consequence of the images obtained and to provide an adequate interpretation, the nature of the microgranular-fibrilar extracellular background observed in the preparations was recognized. The influence of culture conditions and incubation time on the observed effect was explored, allowing a detailed dose-effect analysis of the β-lactam, establishing categories of cell wall damage. Finally, the utility of the methodology was demonstrated and extended to clinically relevant gram-negative and gram-positive microorganisms. To our knowledge, there are no references on our work to discuss, given the novelty of the technique.
The procedure was able to distinguish E. coli strains that were susceptible, intermediate and resistant to amoxicillin/clavulanic acid. Susceptible strains appeared lysed releasing the nucleoid after the cut-off dose point of susceptibility (8/4 μg/ml), whereas intermediate strains only were affected by the threshold dose of resistance (32/16 μg/ml). Intermediate strains were only lysed after this latter dose. From the clinical point of view, besides the control 0 dose, the assay with the breakpoint dose of susceptibility could be enough to discriminate susceptible from non-susceptible strains. This may make the analysis of lots of strains very accessible with the procedure. In fact, the important fact for the therapeutic decision is the differentiation between susceptible or non-susceptible. Intermediate strains should not be treated with the antibiotic, being preferable to use an alternative one to which they are totally susceptible.
The growing stage of the bacterial population must influence the efficacy of the antibiotic, affecting the kinetics of action. In fact, cells that are not growing or in stationary phase extraordinarily decrease the susceptibility to β-lactams . Moreover, the effect at the cell wall of the antibiotic is a dynamic process, concentration and time-dependent.
From the experiment of incubation time, it is deduced that to discriminate with accuracy the susceptible strains from the rest it is enough, in a practical clinical approach, to assess the control 0 dose and the CLSI cut-off dose for susceptibility, incubating with the antibiotic for 60 min in case of cultures growing 24 h in agar plate, as usual in the standard clinical microbiology laboratory. If the cultures were exponentially growing in liquid medium, the incubation time with the antibiotic may be decreased for 30 min. We have observed that the greater the ageing of the culture in agar plate, or when the culture is achieving the stationary phase of growth, the longer the incubation time necessary to observe the effect of the antibiotic, even several hours. To evaluate clinical strains using the technique to assess the integrity of the cell wall, it is mandatory to simultaneously process a sensitive, an intermediate and a resistant strain as controls of the activity of the antibiotic and the efficacy of the technique.
Sensitive strains from gram-negative bacteria assayed showed a background of extracellular microgranular-fibrilar material, its concentration being dose and time dependent. This material corresponded to DNA fragments released by the bacteria, since it was digested by DNase I and hybridized with a specific whole genome probe, being clearly visualized with high sensitive DNA dyes, i.e. SYBR Gold. It is interesting to note that this background of DNA fragments was practically undetectable in gram-positive strains, despite being susceptible to β-lactams or vancomycin. Moreover, it was also undetected in the same bacteria after quinolone treatment in susceptible strains, as evidenced in our previous works with the procedure [15, 16]. This fact suggests that the release of DNA fragments could be specific to cell wall directed antibiotics or β-lactams at least.
This interesting phenomenon requires a deeper study in future works, to address the mechanisms and kinetics of production. DNA fragmentation must be a secondary effect, after cell wall damage. It could be a passive result of attack by DNases or reactive species of oxygen (ROS) liberated in the affected bacteria, or it could be active, a consequence of an apoptotic-like process triggered after cell wall damage. Considering to the first possibility, it has been recently reported that, unlike bacteriostatic antibiotics, β-lactams induce the formation of ROS in gram-negative and gram-positive microorganisms . Hydroxyl radicals should attack proteins and DNA, possibly inducing DNA breakages, resulting in death of the bacteria. This response was also found with other bactericidal antibiotics, like fluoroquinolones. Possibly, the increased permeability of the cell wall that would result after impairment of peptidoglycan biosynthesis by the β-lactams, would allow the release of DNA fragments to the medium. Nevertheless, the DNA fragments that result by the particular effect of quinolones through trapping of DNA gyrase and topoisomerase IV on chromosomal DNA and/or by possible ROS attack [18, 19], could not be released out of the bacterium since the cell wall would be intact initially, at least. In the case of gram-positive bacteria, it should be analyzed more confidently if DNA fragmentation is produced after β-lactam treatment, although more delayed than in gram-negative. If this is the case, despite of the effect, the thicker cell wall of gram-positive bacteria would also prevent the release of DNA fragments.
From the practical point of view, the background of DNA fragments was visualized without the necessity of incubation in lysing solution or any manipulation, so it could be used for a rapid determination of sensitivity or resistance, in liquid cultures. Nevertheless, the presence of the background could be indicative of susceptibility only in gram-negative bacteria, in those here assayed at least. Furthermore, the dilution of the culture modifies the density of the background, and different bacteria and different strains may show important differences in the amount of extracellular DNA fragments. A more confident discrimination between sensitive and resistant strains is achieved when also evaluating the cell wall response to the specific lysing solution.
The dose-response study shows that the β-lactam induces a progressive effect with increasing dose on the cell wall. This effect is evident even before the MIC dose, although it is very weak and seems not prevent growth of most of bacteria after removing the antibiotic. The cell wall damage is not homogeneous among cells, although a predominant level is observed for each dose. This level is more intense as dose increases. The heterogeneity in the effect on the cell wall is not mainly dependent on the growing stage since the cultures were exponentially growing when exposed to ampicillin. The background of DNA fragments appears to be observed at the MIC dose, and increases as dose increases, within the range of doses assayed.
The methodology has been confirmed as a rapid and simple procedure to distinguish susceptible and resistant strains of eight gram-negative and four gram-positive species, assaying four different β-lactams and vancomycin. The results were reproducible and accurate, in the 46 clinical strains. Although preliminary, the results are encouraging. Expanded work analysing many more strains is in progress. For example, links have been established between glycopeptide resistance and cell wall thickening in vancomycin-intermediate Staphylococcus aureus (VISA), as well as between macrolides and thickened cell walls in S. aureus [20, 21]. These are interesting strains to be tested. Furthermore, the examination of the slides is going to be automated using a microscopy platform coupled with image capture and digital image analysis. This will allow a high-throughput examination of large numbers of microorganisms, for a rapid identification of resistant and susceptible strains to antibiotics that interfere with peptidoglycan biosynthesis.