a Bristol Centre for Antimicrobial Research and Evaluation, North Bristol NHS Trust, Southmead Hospital, Bristol BS10 5NB; b University of Bristol, Department of Pathology and Microbiology, Bristol, UK
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
We investigated the suitability of the parallel-line bioassay, primarily used to calculate relative drug potencies,5 to determine whether the photodegradation products of ciprofloxacin, levofloxacin, ofloxacin and moxifloxacin possessed antimicrobial activity. We aimed to measure any additional antimicrobial activity in irradiated solutions relative to unirradiated solutions containing the same concentration of parent compound, using this bioassay with Gram-negative indicator organisms to investigate discrepancies in inhibition zone size together with confirmatory high performance liquid chromatography (HPLC) analysis.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Aqueous stock solutions (0.02 mM) of ciprofloxacin, moxifloxacin (Bayer AG, Wuppertal, Germany), ofloxacin and levofloxacin (Hoescht-Marion-Roussel, Hounslow, UK) were prepared and protected from light. A photochemical reaction unit (Beam-Boost; ICT, Frankfurt, Germany) fitted with an 8 W, 254 nm UV lamp (UV Products, Cambridge, UK) was used to photodegrade partially these solutions in a reproducible way by using various flow rates and retention coil lengths. Solutions of ciprofloxacin, levofloxacin, ofloxacin and moxifloxacin, partially photodegraded by between 15 and 89% (as confirmed by HPLC analysis), were obtained using this procedure to produce a range of low, medium and high photodegraded solutions.
Preparation of non-irradiated matched control fluoroquinolone solutions
The residual concentration of parent fluoroquinolone in each irradiated solution was measured by HPLC. A matched control solution was then prepared, containing approximately the same concentration of parent fluoroquinolone as the irradiated sample.
HPLC
Parent fluoroquinolone concentrations were measured by validated isocratic HPLC methods.69
Parallel-line bioassay
Forty-five 8 mm diameter wells were cut into 30 cm square plates of Diagnostic Sensitest (DST) agar (Unipath, Basingstoke, UK), which had been surface seeded with the indicator organisms, Escherichia coli (NCTC 10418), Enterobacter cloacae (clinical isolate) or Klebsiella oxytoca (clinical isolate), using a 107 cfu/mL suspension. The photodegraded solutions and their matched control solutions were assayed neat and diluted 1:2 and 1:4 in sterile water; six replicates of each dilution were then assayed. A random pattern was used to designate sample location on the assay plate and wells were filled until a convex meniscus was visible.10 After overnight incubation at 37°C, the diameters of the zones of growth inhibition were measured using a magnifying zone reader (Luckham Ltd, Burgess Hill, UK).
Statistical analysis
The assays were validated by analysis of variance (ANOVA). Estimated relative potencies (ERP; the ratio of the matched control solution concentration to that of the irradiated solution concentration) were determined as for an asymmetrical, six-point, parallel-line assay.11 Confidence intervals were calculated for the ERP using Fieller's theorem.11 Assays were accepted as valid only if the sample dilutions were linearly related and the test (irradiated) and control (non-irradiated) graphs were parallel.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Exposure of the quinolone solutions to light resulted in a reduction in concentration of parent compound and the appearance of additional chromatographically distinct compounds. Figure 1 (ad) shows chromatograms of 0.02 mM ciprofloxacin, levofloxacin, ofloxacin and moxifloxacin photodegraded by 82, 73, 78 and 83%, respectively, and shows their stable photodegradation products.
|
The Table shows the ERPs of the irradiated solutions compared with the matched control solutions at different percentages of photodegradation.
|
Levofloxacin.
At 15% photodegradation there was no significant difference in ERP with any indicator organism, or at 48% photodegradation with E. cloacae (Table). At 48% photodegradation, however, the ERPs were significantly different (P < 0.05) for E. coli and K. oxytoca, and at 73% photodegradation for all three indicator organisms (P < 0.01). At 87% photodegradation, the ERP was also significantly different (P < 0.01) for E. coli, but the results were inconclusive with E. cloacae and K. oxytoca. Figure 2
illustrates the plots of the zone diameter for E. coli against levofloxacin with 15, 48, 73 and 87% photodegradation. The 95% confidence limits for the test and control solutions overlap at 15 and 48% photodegradation, despite there being a statistically significant difference at 48% (Figure 2a and b
). At 73 and 87%, the 95% confidence limits show no overlap and differences in activity of the test versus control solutions were statistically significant. As with ofloxacin, the ERP increased with increased levels of photodegradation.
|
Moxifloxacin.
There was no significant difference in ERP for irradiated versus control solutions after 30 or 54% photodegradation of the parent drug with any of the indicator organisms, nor after 83% photodegradation with E. coli and K. oxytoca. ERPs were significantly different for 83% photodegradation using E. cloacae (P < 0.01). All plots were linear and parallel as confirmed by ANOVA.
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Dilutions of irradiated and control solutions containing the same parent drug concentrations were assayed and if the plots of zone diameter versus logarithmic concentration were linear, parallel and significantly different (with the test zones consistently larger than the control zones), the presence of microbiologically active products was indicated. Three Gram-negative organisms were tested in this study but if used for screening, a panel of both Gram-negative and Gram-positive organisms could be used.
Six of the 42 assay results (Table) were invalid due to non-parallelism and were not investigated further. However, there was a large difference between test and control zones in these experiments (data not shown) and it is interesting that the invalid results were obtained only with ofloxacin and levofloxacin, which are different isomeric forms of the same compound. Non-parallel lines could indicate the presence of active compounds differing greatly in molecular size or with a different mode of action from the parent. These findings are worthy of further work but, for the purposes of this study, indicate that the requirement for parallelism might cause some microbiologically active breakdown products to be overlooked.
Although no significant differences between control and test curves were seen with ciprofloxacin, statistically significant differences were seen with the other quinolones tested. Moxifloxacin showed a significant difference only after substantial degradation (83%) and with only one organism (E. cloacae). With both ofloxacin and levofloxacin, however, a difference was seen with all organisms. The difference was quantitative, in as much as the ERP increased with increased photodegradation.
Our findings suggest that this approach to testing partially photodegraded compounds can identify the presence of microbiologically active products. Such products were most clearly demonstrable in those solutions that had undergone >70% degradation, but were not always detected by all three indicator strains tested. Testing a single substantially degraded solution against a panel of organisms is probably the best way to screen for such activity.
Although the assay protocol invalidated curves that were not parallel, the invalid assays obtained with levofloxacin and ofloxacin strongly implied that active breakdown products were present. It is suggested that any solutions showing significantly different zone sizes between test and control are worth further investigation, whether the curves are parallel or not.
![]() |
Acknowledgments |
---|
![]() |
Notes |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 . Tiefenbacher, E. M., Haen, E., Przybilla, B. & Kurz, H. (1994). Photodegradation of some quinolones used as antimicrobial therapeutics. Journal of Pharmaceutical Sciences 83, 4637.[ISI][Medline]
3 . Ferguson, J. & Johnson, B. E. (1990). Ciprofloxacin-induced photosensitivity: in-vitro and in-vivo studies. British Journal of Dermatology 123, 920.[ISI][Medline]
4 . Phillips, G., Johnson, B. E. & Ferguson, J. (1990). The loss of antibiotic activity of ciprofloxacin by photodegradation. Journal of Antimicrobial Chemotherapy 26, 7839.[Abstract]
5 . Hewitt, W. & Vincent, S. (1989). Assay design and evaluation. In Theory and Application of Microbiological Assay, 1st edn, (Hewitt, W. & Vincent, S., Eds), pp. 185211. Academic Press Ltd, London.
6 . Gau, W., Ploschke, K., Schmidt, K. & Weber, B. (1985). Determination of ciprofloxacin (BAY 0 9867) in biological fluids by high performance liquid chromatography. Journal of Liquid Chromatography 8, 48597.[ISI]
7
.
Tobin, C. M., Sunderland, J., White, L. O. & MacGowan, A. P. (1999). A reverse-phase isocratic high-performance liquid chromatography assay for levofloxacin. Journal of Antimicrobial Chemotherapy 43, 4345.
8 . White, L. O., MacGowan, A. P., Lovering, A. M., Reeves, D. S. & MacKay, I. G. (1987). A preliminary report on the pharmacokinetics of ofloxacin, des-methyl ofloxacin and ofloxacin N-oxide in patients with chronic renal failure. Drugs 34, Suppl. 1, 5661.[ISI][Medline]
9
.
Tobin, C. M., Sunderland, J., White, L. O., MacGowan, A. P. & Reeves, D. S. (1998). An isocratic high performance liquid chromatography (HPLC) assay for moxifloxacin, a new 8-methoxyquinolone. Journal of Antimicrobial Chemotherapy 42, 2789.
10 . Andrews, J. M. (1998). Microbiological assays. In Clinical Antimicrobial Assays, (Reeves, D. S., Wise, R., Andrews, J. M. & White, L. O., Eds), pp. 3544. Oxford University Press, Oxford.
11 . Finney, D. J. (1978). Statistical Methods in Biological Assay, 3rd edn, pp. 6988. Griffin, London.
12 . Quintiliani, R., Nightingale, C. H. & Tilton, R. (1984). Comparative pharmacokinetics of cefotaxime and ceftizoxime and the role of desacetylcefotaxime in the antibacterial activity of cefotaxime. Diagnostic Microbiology and Infectious Disease 2, Suppl., 63S70S.[Medline]
13 . Korner, R. J., McMullin, C. M., Bowker, K. A., White, L. O., Reeves, D. S. & MacGowan, A. P. (1994). The serum concentrations of desmethyl ofloxacin and ofloxacin N-oxide in seriously ill patients and their possible contributions to the antibacterial activity of ofloxacin. Journal of Antimicrobial Chemotherapy 34, 3003.[ISI][Medline]
Received 19 October 2000; accepted 5 December 2000