| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Medical Microbiology and Infection Control, Jeroen Bosch Hospital, Nieuwstraat 34, 5211 NL, 's-Hertogenbosch, The Netherlands
2 Laboratory of Microbiology, Wageningen University, Wageningen, The Netherlands
Correspondence
Colin J. Ingham
cingham2{at}mac.com
Received 22 February 2006
Accepted 21 July 2006
Abbreviations: AST, antibiotic sensitivity testing; PI, propidium iodide; SEM, scanning electron microscopy.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Trimethoprim is a diaminopyrimidine antibiotic that inhibits the bacterial dihydrofolate reductase which catalyses the conversion of dihydrofolate to tetrahydrofolate (Hitchins, 1989; Huovinen et al., 1995). Trimethoprim has been reported to cause morphological changes in a few bacterial species only, including Listeria monocytogenes (Minkowski et al., 2001). Antifolate drugs structurally related to trimethoprim have similar effects on Escherichia coli (Nickerson & Webb, 1956). Trimethoprim also affects cell morphology and the synthesis of cell wall precursors in Enterobacter cloacae (Richards & Xing, 1994), and cell wall formation in Enterococcus faecalis (Richards et al., 1995). It has been suggested that trimethoprim exerts its effect on cell morphology by inhibiting DNA replication (Richards & Xing, 1994). Most aspects of the occurrence, mechanism and clinical significance of trimethoprim-mediated filamentation are unknown. Filamentation has been more extensively investigated for other antibiotics (Cudic & Otvos, 2002; Chadfield & Hinton, 2004), including the ß-lactams, for which links to the SOS response and antibiotic survival have been made (Miller et al., 2004).
In this study, Anopore-based growth and in situ microcolony imaging were used to perform rapid AST. We have focused on the response of clinical isolates of the Enterobacteriaceae to trimethoprim, but have obtained additional data that suggest a wider applicability. An advantage of the method was the direct observation of cells exposed to an antibiotic. The value and potential of Anopore as a rapid diagnostic tool are discussed.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
. Vitek 2, API (both bioMérieux) and E-test (AB Biodisk) assays were performed as recommended by the manufacturers, and interpreted according to National Committee for Clinical Laboratory Standards (NCCLS) criteria and the manufacturers' guidelines.
|
Growth, imaging of cells and AST. Anopore chips were placed upon pre-warmed Mueller–Hinton agar plates (Oxoid) for 10 min, then inoculated with 104–105 c.f.u. (2 µl aliquots) of bacteria per compartment. After waiting a few minutes for the liquid to be drawn into the pores, the plates were incubated inverted in a CO2 incubator at 37 °C. These chips were used to perform fast AST using trimethoprim. Trimethoprim or other antibiotics either were present in the Mueller–Hinton agar at a known concentration or were printed directly into the compartment of an Anopore strip which was then incubated on Mueller–Hinton agar lacking the antibiotic. After incubation, staining of cells on the chip surface with fluorescent dyes (through the pores of the Anopore from beneath) was performed as previously described (Ingham et al., 2005). Briefly, this was accomplished by the transfer of the chip to a microscope slide covered with a film of solidified low-melting-point agarose (Sigma) containing 10 µM Syto9 (Invitrogen) and in some cases 40 µM propidium iodide (PI; Invitrogen).
Imaging, processing and statistical analysis. Chips were imaged directly (without the use of any immersion oil or a coverslip) using an Olympus BX41 epifluorescence microscope equipped with UmPlan F1 objective lenses and U-MWIBA and U-M41007 filters (Ingham et al., 2005). Image capture used a Kappa charge-coupled device (CCD) camera controlled by Kappa Image Base software. Raw files of 12 bit images or BMP files of 8 bit images were analysed quantitatively using ImageJ software to implement background correction, median filtration, conversion to a binary image and calculation of colony or cell size (Rasband, 2004). Calculations were performed in Microsoft Excel. Images were merged and displayed using Photoshop 8.0 CS (Adobe). Simple paired comparisons were made with Student's t test. One-way ANOVA with Tukey post-hoc testing (Turner & Thayer, 2001) was used for multiple comparisons. For AST testing on Anopore, MIC90 values were calculated from the mean area of at least 200 microcolonies per data point. Unless stated otherwise, all statistical calculations are the representative result from at least two independent experiments.
Scanning electron microscopy (SEM). Anopore strips cultured with bacteria were glued on a sample holder with conductive carbon cement (Leit-C, Neubauer Chemicalien) and frozen in liquid nitrogen. Samples were transferred under vacuum to the dedicated cryo-preparation chamber (Oxford Cryo-system, CT 1500 HF) onto a sample stage at –90 °C. The samples were freeze-dried for 4 min at –90 °C in a 3x10–7 Pa vacuum to remove water vapour contamination. After the sample surface was sputter-coated with 10 nm platinum, it was transferred to the cold sample stage (–190 °C) inside the Cryo-FESEM (JEOL 6300F Field Emission SEM) and subsequently analysed with an accelerating voltage of 5 kV. Images were digitally recorded (Orion, E.L.I.).
Screening for filamentation. A systematic screening was performed using Anopore chips printed with trimethoprim (0–600 pg mm–2). Over 30 strains were tested on two types of medium (Mueller–Hinton and sheep blood agar), using three dilutions of each sample (inoculations of approximately 105, 104 and 103 c.f.u. over a 10 mm2 area of Anopore) and two incubation times (4–5 and 16 h). After this initial screening, follow-up experiments were performed on strains showing filamentation. These were cultured for 2.5, 5 and 16 h on chips printed with 200 pg trimethoprim mm–2. Filamentation was also tested on Anopore incubated on Mueller–Hinton agar containing 3 µg mitomycin C ml–1 (Sigma).
|
RESULTS AND DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
) and 1468 (resistant) both grew rapidly on Anopore incubated on Mueller–Hinton agar. The doubling times of these strains at 37 °C was calculated from the mean number of cells per microcolony (n>200 at each data point) during the first 2.5 h growth (Ingham et al., 2005). The doubling time during exponential growth on Anopore was 27 min for strain 1499 and 29 min for strain 1468, with lag times of 25–30 min for both strains (all times are the average of two values that differed by less than 20 %). This compared to doubling times of 23 and 24 min, respectively, for strains 1499 and 1468 in shaken Mueller–Hinton broth at the same temperature. With only 25 min culture it was possible to detect growth of strain 1499 by comparing the mean microcolony area of the inoculum (generally 1–2 cells) with that of the cultured population (one-way ANOVA, n=924 for the inoculum, n=774 cultured without trimethoprim, n=567 for culture with trimethoprim; P<0.01 by Tukey HSD). No significant inhibitory effect on growth was found with trimethoprim (P>0.05 by Tukey HSD) within 25 min. However, with 40 min incubation (Fig. 1), trimethoprim reduced growth (n=594 for the inoculum, n=602 cultured without trimethoprim, n=642 cultured with trimethoprim; P<0.01 comparing treated and untreated populations by Tukey HSD). Growth occurred in both the treated and untreated population, as indicated by significant increases in the microcolony areas (mean microcolony area of both cultured populations compared to that of the inoculum; P<0.01 in both cases by Tukey HSD). For the resistant strain 1468 (Fig. 1b), trimethoprim had no significant effect on growth at 40 min (n=604 for the inoculum, n=615 cultured without trimethoprim, n=628 cultured with trimethoprim; one-way ANOVA with Tukey HSD, P>0.05 comparing treated and untreated populations). Taken together, these data suggest that 40 min was close to the minimal time in which it was possible to distinguish a sensitive from a resistant strain by this method.
|
, Table 1). This compared to MICs of 0.25 µg ml–1 by E-test, and <0.5 µg ml–1 by Vitek 2 (Table 1). The resistant strain (1468) had a MIC90 of >10 µg ml–1 by the Anopore method and MICs of >32 µg ml–1 by E-test and >16 µg ml–1 by Vitek 2 (Fig. 2a, Table 1). MIC90 determinations were then extended to other members of the family Enterobacteriaceae and some other Gram-negative rods, such as pseudomonads (Table 1). During the testing period, swarming behaviour by Proteus mirabilis or Proteus vulgaris on Anopore was not observed. There was excellent agreement between established techniques and the Anopore method of AST: resistance (17 strains) and sensitivity (40 strains) to trimethoprim were assigned identically on Anopore when compared to Vitek 2 and/or E-test determinations.
|
The MIC90 determinations described above were made using a defined concentration of the antibiotic in agar plates. Trimethoprim-printed Anopore chips, when placed on Mueller–Hinton agar lacking trimethoprim, could also be used to rapidly distinguish a resistant from a sensitive strain of Ent. aerogenes (Fig. 2b). A dose-dependent response to trimethoprim was observed after 2 h culture. Minimal inhibitory amounts of trimethoprim were 25 pg mm–2 for the sensitive strain (1499) and >600 pg mm–2 for the resistant strain (1468). Anopore chips printed with 200 pg trimethoprim mm–2 were also used to assess the minimal culture time required to distinguish the sensitive from the resistant strain of Ent. aerogenes by calculating the mean area of the microcolonies.
In these assays, trimethoprim was present in the growth medium at a defined concentration or was printed onto Anopore, dried and then allowed to rehydrate. The former method allowed known concentrations of antibiotics to be used, and the agar base may potentially be replaced with a microfluidic underlay or growth matrix in the future, permitting highly multiplexed AST with a wide range of antibiotics, both singly and in combination. Printing the antibiotic directly onto the Anopore means that the breakpoints must be experimentally determined with calibration strains. An inhibition by subnanogram quantities of trimethoprim was observed, possibly because of a localization of antibiotics at a high concentration within the limited volume of the pores in very close proximity to the test bacteria. That small amounts of trimethoprim were bioactive suggests that this approach also has potential for high-throughput screening of compounds from antimicrobial drug libraries which are often limited in quantity. Additionally, antibiotic-printed Anopore was very convenient to store and deploy, and this may facilitate the storage of compound libraries ‘ready to use’.
Detection of a resistant subpopulation
In order to simulate a simple mixed population, a culture of strain 1499 was spiked with strain 1468, so that the latter was in a minority (10 % of the population). After culture for 2 h, this mixed population could be distinguished from the 1499 monoculture by the appearance of microcolonies comprising 12–32 cells, but at a lower frequency than the 1468 monoculture. Comparison of microcolony areas indicated that the spiked population had a mean significantly larger than that of the sensitive population (n=650 for all populations; one-way ANOVA with P<0.01 by Tukey HSD), but smaller than that of the resistant population (P<0.01). After 2 h culture, the resistant subpopulation was still calculated to be in the minority (
30 %). A resistant strain at 1 % of the population could not be reliably detected with 2 h culture. After 3.5 h, a small number of obviously resistant microcolonies (>100 cells) were observed in the spiked population, but not in the 100 % sensitive population. These data suggest that it is possible to resolve mixtures of micro-organisms of clinical significance on Anopore in an analogous way to that possible by the spread-plating of microbes on nutrient agar, but on a far smaller scale. This is an advantage compared to most other methods that rely on culture in liquid media, since in these methods, an antibiotic-resistant subpopulation tends to be detected reliably only when it grows to outnumber antibiotic-sensitive cells. Therefore, the method may be applicable to heterogeneous or mixed populations. In this context, removing the requirement for purification to a single colony has obvious advantages in terms of speed.
Morphological effects of trimethoprim
Direct microscopic observation of AST assays on Anopore resulted in the finding that elongated cells with irregular nucleic acid distributions were a common effect of trimethoprim treatment of sensitive strains. Other antibiotics, including ß-lactams, had a more limited or no obvious effect. Trimethoprim often triggered morphological changes that were quite extreme. The AST assays, backed up by additional screening on Anopore, indicated that filamentation is a common but underreported effect of trimethoprim (Table 1). Filamentation was observed for many trimethoprim-sensitive isolates of the Enterobacteriaceae, including Ent. aerogenes 1499, Pantoea sp. X28, Hafnia alvei 981, Shigella sonnei 2627-1, Shigella flexneri M1, Klebsiella pneumoniae JBZA5, Citrobacter freundii JBZA7 and Citrobacter diversus 711 (Table 1). These are all representatives of species for which trimethoprim has not been previously been reported to induce filamentation. Quantitatively, the results for these strains were generally similar to those of Ent. aerogenes (Fig. 3) in terms of the distribution of cell lengths.
|
, strains 2364 and 1170 of Ent. cloacae and L. monocytogenes 1444). Not all isolates of a species showed the same response. E. coli 2657, H. alvei 5636 and L. monocytogenes 620 were less prone to trimethoprim-mediated elongation than other clinical isolates of the same species (Table 1), but were not trimethoprim resistant. Nor was trimethoprim-induced filamentation found in all species tested (Table 1). Given the variation in trimethoprim-mediated filamentation between isolates we cannot conclude that Serratia marcescens or Salmonella typhimurium do not filament (Table 1), merely that this response is a common but not universal one for trimethoprim-sensitive Enterobacteriaceae.
The distribution of filamented cells over the surface of the Anopore was most often in discrete microcolonies, generally clusters of <50 cells. There was great variation in both the length and in the permeability to PI within a microcolony (data not shown). A moderate degree of broadening and flattening of the more highly damaged cells was, however, observed (Fig. 4). This is consistent with reports that trimethoprim interferes with the supply of cell-wall precursors in Ent. cloacae (Richards & Xing, 1994; Richards et al., 1995). A wide range of responses was observed in what must (in most cases) be the product of a single cell, or pair of cells, generating diversity during the course of the assay in the presence of trimethoprim above the minimal inhibitory amount. Such microcolonies did not develop into visible colonies on Anopore with extended incubation for 24 h, and only trimethoprim-sensitive isolates could be recovered from these experiments. This suggested that trimethoprim-resistant mutants were not being generated. Antibiotic-refractory subpopulations (e.g. ‘persister’ cells) have been shown to exist within populations of bacteria (Balaban et al., 2001). The phenotypes of such subpopulations do not appear to be determined by any stable genetic change. Assessing heterogeneity within microcolonies may be a relatively simple way of looking at this phenomenon. Extreme variation in the response to trimethoprim was also seen by SEM for Ent. aerogenes 1499 (Fig. 4). In addition, variation in cell width and in the intactness of the cell envelope was observed. In extreme cases, cells seemed to be collapsing, showing both flattening and visible lesions in the cell wall (Fig. 4c). It was possible to see the remains of these lysed and fragmented cells in close proximity to apparently intact ones, with fragments preserved on the Anopore support in a way that would not have been possible with liquid culture. Given the possibility of withdrawing antibiotics by transfer of an Anopore chip to a new medium, this form of in situ imaging may have further potential in the observation of highly damaged and also of recovering cells in the study of antibiotic effects.
|
). The resulting population was similar to the trimethoprim-treated population in terms of the length and heterogeneity of DNA distribution and PI permeability. A more extensive screening suggested that the induction of the SOS response leading to filamentation was general, and included strains that did not show obvious filamentation with trimethoprim (data not shown). The SOS response can result in the inhibition of cell division, but still allows growth leading to filamentation (Neidhardt et al., 1990). Given the widespread presence of the SOS pathway in Eubacteria (Friedberg et al., 1995), this is an obvious explanation of why trimethoprim-mediated filamentation was so frequently observed in this study. The SOS response has also been implicated in mediating the filamentation of E. coli in response to ß-lactam antibiotics, via a periplasmic binding protein and a two-component regulatory system (Miller et al., 2004). The SOS response is also known to facilitate the horizontal dissemination of resistance genes for a range of antibiotics, including trimethoprim (Beaber et al., 2003). The induction of filamentation by mitomycin C suggests that activation of the SOS response can indeed trigger filamentation in the strains we have studied, and the effect on microcolony heterogeneity was similar to that of trimethoprim. However, this does not explain why filamentation was not a universal response in trimethoprim-sensitive strains. If the SOS response is involved, the relationship may not be a simple one.
Conclusions and clinical potential
This is a new application of an unusual but appropriate material in culture-based diagnostics. The data presented demonstrate rapid and accurate AST testing in a few hours, rather than overnight. This speed was made possible by the ability to stain and image microcolonies in situ with little redistribution of organisms on the Anopore surface. Images of fields of microcolonies could be rapidly captured, digitized and compared. This was done using a type of microscope available in most microbial diagnostics laboratories, with data analysis performed using publicly available software (Rasband, 2004). This study has concentrated on a single antibiotic, trimethoprim, and strains within the family Enterobacteriaceae, but we have demonstrated examples of AST with other Gram-negative bacteria and antibiotics, suggesting that this is a generally applicable technique. Clearly, this approach needs to be tested more widely, focusing on the important issues of intermediate resistance, the interactions between subpopulations, and inducible resistance mechanisms. However, there are obvious advantages in the speed and imaging simplicity of the method that should make this effort worthwhile. The early detection of growth may prove even more advantageous when applied to slower growing organisms. Preliminary results for drug testing of Mycobacterium tuberculosis on Anopore (A. Ben Ayad, and others, unpublished results) suggest that this difficult bacterium can be tested for rifampicin and isoniazid resistance with only a few days culture. Given the slow growth and relatively easy imaging of many fungi by microscopy, these may also present attractive subjects for this method.
In terms of clinical applications, only a few mm2 of Anopore (costing a few Euro cents) is required to test an antibiotic. Fluorescent stains can be expensive, but are used in small quantities. Well-characterized and relatively low-cost stains, such as acridine orange, are also usable. Taken together, these considerations suggest that the method can be economic, particularly if a large number of tests are combined on a single strip of Anopore. The ability to distribute controlled and variable amounts of compound into very many defined areas of the substrate may also allow break-point testing, and synergistic or antagonistic interactions by means of ‘micro’ chessboard titrations. In terms of workflow, streamlining the staining method by incorporating low-toxicity fluorescent stains into the growth medium will be advantageous in reducing hands-on time. Non-fluorescent stains or contrast-enhancing techniques for conventional light microscopy could be used to adapt the method to the developing world, where conventional light microscopes are more common than those capable of fluorescence (Olympus Europe, personal communication). For analysis, publicly available image-processing software (Rasband, 2004) is relatively easily tailored to microcolony analysis, with only minor scripting or macro development. Finally, we suggest that there is greater potential for the automation of culture with Anopore than with conventional microbial growth methods such as agar-based culture.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. E. Budding, C. J. Ingham, W. Bitter, C. M. Vandenbroucke-Grauls, and P. M. Schneeberger The Dienes Phenomenon: Competition and Territoriality in Swarming Proteus mirabilis J. Bacteriol., June 15, 2009; 191(12): 3892 - 3900. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. Ingham, M. Beerthuyzen, and J. van Hylckama Vlieg Population Heterogeneity of Lactobacillus plantarum WCFS1 Microcolonies in Response to and Recovery from Acid Stress Appl. Envir. Microbiol., December 15, 2008; 74(24): 7750 - 7758. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. Ingham, A. Sprenkels, J. Bomer, D. Molenaar, A. van den Berg, J. E. T. van Hylckama Vlieg, and W. M. de Vos From the Cover: The micro-Petri dish, a million-well growth chip for the culture and high-throughput screening of microorganisms PNAS, November 13, 2007; 104(46): 18217 - 18222. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. M. W. den Besten, C. J. Ingham, J. E. T. van Hylckama Vlieg, M. M. Beerthuyzen, M. H. Zwietering, and T. Abee Quantitative Analysis of Population Heterogeneity of the Adaptive Salt Stress Response and Growth Capacity of Bacillus cereus ATCC 14579 Appl. Envir. Microbiol., August 1, 2007; 73(15): 4797 - 4804. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
) and -resistant (1468, 
) strains of Ent. aerogenes cultured on Anopore for 2 h. Error bars show SD of the mean. (a) Trimethoprim present in the Mueller–Hinton agar. Additionally, strain 1468 was tested up to 10 µg trimethoprim ml–1 with no inhibition of growth. (b) Trimethoprim printed on Anopore. Strain 1468 was tested up to 600 pg trimethoprim mm–2 with no growth inhibition.
