Abstract

The ability to detect and quantify specific DNA adducts benefits genome stability research, drug development and the evaluation of environmental mutagens. The trapped in agarose DNA immunostaining (TARDIS) assay was developed as a means of detecting and quantifying melphalan and cisplatin DNA adducts at the single-cell level and has since been adapted to quantify topoisomerase–DNA complexes. The method relies on salt–detergent extraction of agarose-embedded cells. Genomic DNA and any covalently attached molecules remain in place in the agarose, while other cellular constituents are removed. Drug–DNA or topoisomerase–DNA complexes are then detected and quantified by sensitive immunofluorescence using adduct-specific antibodies. Here, we give a perspective of the TARDIS assay including a comparison with other methods for quantifying topoisomerase–DNA covalent complexes and provide technical details required to set up and perform the assay.

Original development—quantification of melphalan DNA adducts at the single-cell level

The development of monoclonal antibodies that recognised DNA adducts formed by the anticancer drugs melphalan and cisplatin (1,2) led to the desire to apply these to quantify drug–DNA adducts in individual cells to enable the study of small samples, the assessment of heterogeneity in damage formation and repair within a cell population, and also offered the potential to study spatial distribution of adducts across tissue sections. However, for conventionally fixed cells or tissues, there are two challenges to such analyses. Firstly, access of antibody to DNA adducts could be hindered by DNA-associated proteins. Secondly, antibodies often only recognise adducts in denatured DNA. Therefore, it was essential to achieve efficient denaturation of the DNA without causing loss of material from the microscope slide. These factors raised concerns about the reliability of any conventional approach to immunostaining because differences between cell types and/or changes to the structure of chromatin in response to DNA damage formation could affect the efficiency of any immunostaining procedure. Hence, if conventional immunostaining methods were to be used, one would need to be sure that any differences between cells truly reflected differences in adduct levels and not differences in detection efficiency resulting from differences in accessibility to antibody or because its DNA was more readily denatured (e.g. there was a nearby single-strand break, as discussed below). Another complicating factor is that adduct levels need to be assessed in relation to a reliable measurement of DNA content. Otherwise, one cell could appear as carrying more adducts than another simply because it was in G2 rather than G1 phase of cell cycle or because it was aneuploid rather than diploid.

The technique that was subsequently named the trapped in agarose DNA immunostaining (TARDIS) method was initially conceived as a way to overcome these challenges. The idea for the method came to M.J.T. in 1989 while watching the preparation of agarose sample plugs for use in pulse-field gel electrophoresis in which the intact human DNA molecules are too large to diffuse through agarose and so remained trapped, while the porosity of the agarose was sufficient to readily permit diffusion of proteins such as nucleases. If the agarose could be spread on a microscope slide in a thin layer, it might be possible to retain DNA from individual cells while achieving efficient removal of essentially all cellular proteins. Such pure DNA would then be amenable to rapid alkaline denaturation and subsequent immunofluorescent staining. Also, DNA could be quantified through the use of suitable fluorescent dyes. Finally, the absence of other cellular material could result in lower levels of non-specific staining. Early versions of the comet assay, available at that time, indicated procedures for spreading and retaining agarose onto microscope slides.

Initial attempts to apply this approach focused on using antibody CP9/19 (2) to detect DNA adducts formed by cisplatin. Initially, this worked very well using a standard fluorescence microscope. Immunofluorescence intensity increased in proportion to the cisplatin concentration to which cells had been exposed, from an essentially undetectable level for control cells. It also worked well in the laboratory of Dr P. Twentyman, where analysis was by confocal microscopy. However, subsequent attempts to apply this method were unsuccessful because of persistent high levels of non-specific binding of primary antibody to DNA from control cells. Despite numerous attempts, this problem could not be overcome. Work then focussed on applying the TARDIS assay to measurement of melphalan–DNA adducts using antibody MP5/73 (1) and in this case, the method worked very reliably. It was known from enzyme-linked immunosorbent assay experiments with antibody MP5/73 that to attain maximum adduct recognition, it was necessary to denature the DNA. This property was easily confirmed using the TARDIS method. By having DNA trapped in agarose on a slide, it was easy to investigate the effect of different alkali exposures. Such optimisation was important because over-exposure to alkali can cause chemical conversion of melphalan–guanine adducts to the ring-opened form, which were known not to be recognised by this antibody. In fact, the TARDIS assay was later applied to detect the alkali-induced ring-opened adducts by using a different antibody that does recognise these structures (3). Further studies revealed that optimum detection of melphalan adducts through DNA denaturation could be influenced by the level of strand breaks in the DNA (4). This is explicable, knowing that very large DNA molecules take long times to unwind and fully separate into single-strands because unwinding takes place only from the ends. Until it is fully unwound, DNA can rapidly re-nature upon neutralisation (5). An increase in frequency of strand breaks (single or double) increases the number of points from which DNA can unwind thereby accelerating the overall rate of denaturation. Ionising irradiation increased the overall sensitivity of immunological detection of melphalan–DNA adducts and this was attributed to the introduction of strand breaks. In addition, radiation decreased the extent of inter-cell heterogeneity in immunostaining for melphalan adducts. This indicated that, even in the TARDIS method where DNA is stripped of other macromolecules, differences between cells in immunostaining can result from differences in DNA conformation rather than differences in actual adduct levels. This illustrates the caution needed in interpreting any immunostaining method for DNA adducts where the binding of the antibody used is known to be influenced by DNA conformation. It seems that the same factors will also influence adduct detection in conventionally fixed specimens, but it is only in the TARDIS method that they can be directly investigated.

Wider current and potential applications

The TARDIS method can, in principle, be applied to any immunologically detectable adduct. Detection of low molecular weight (MW) adducts such as those formed by melphalan are more likely to be affected by DNA conformation. The opposite extreme of adduct size is the covalent binding of proteins to DNA and in such cases, it is seems less likely that immunological detection will be significantly influenced by underlying DNA conformation. The most studied examples of protein adducts are the topoisomerase II complexes that are discussed in detail below. However, it could, in principle, be applied to the detection of other proteins that become covalently bound to the genome such as stalled intermediates of other DNA-modifying enzymes and protein cross-linked to DNA by experimental or therapeutic chemicals or by oxidative damage. Here, it is worthwhile pointing out that the TARDIS assay is currently the only available method that permits measurement specifically of protein–DNA adducts in individual cells since, unlike low MW DNA adducts, it is necessary to efficiently remove all of the normally present but non-covalently bound protein, while retaining all the DNA. Studies referred to above, in which the effect of alkali treatment of melphalan adducts was determined, illustrate the principle that the TARDIS method can permit additional investigations. Since the principle of the method is to allow quite large proteins (antibodies) to access trapped DNA adducts, there is the possibility of carrying out enzymatic as well as chemical alterations to the DNA modifications being investigated. Comparison of immunofluorescent staining with and without such a treatment could provide information about the nature of the target modification such as its susceptibility to specific DNA glycosylases or proteases. For example, enzymatic treatment of comet slides with glycolases has been reported previously as a means of detecting specific DNA lesions (6).

Key factors and limitations

One obvious limitation of the TARDIS method is that it does not permit analysis of cells in the context of tissue architecture in a section. However, analysis of cells microdissected from frozen sections could be feasible. The staining procedures for melphalan adducts all make use of antibodies of the IgG class. This is likely to be an important property because, in addition to such antibodies tending to show higher affinities, the high MW of antibodies of classes such as IgM could hamper penetration into the agarose. In order to detect low levels of DNA modifications, it is, of course, essential that the antibody shows high affinity for the modification and low non-specific binding. However, unlike conventional immunohistochemistry or immunoblotting procedures, TARDIS preparations contain essentially no cellular material other than the DNA and so the main cause of non-specific binding is absent. More of concern is that there should be very low cross-reactivity with normal DNA. Detection of low MW adducts is more likely to be challenged by this factor than detection of large adducts such as proteins. However, efficiency of detection of proteins bound to DNA, particularly if a monoclonal antibody is used, is likely to be influenced by location of the epitope that could be masked it were close to, or within, the DNA-binding site. A range of antibodies have been used for TARDIS to detect topoisomerases. The majority of antibodies that have worked well in TARDIS are rabbit polyclonal antibodies and details are given in the relevant primary publications.

The relative sensitivity of detection for topoisomerase adducts by the TARDIS method is discussed below but, in general, sensitivity is limited by a combination of the following factors: (i) The cross-reactivity and non-specific binding of primary and secondary antibodies. Since the DNA is essentially free of other cellular material, non-specific binding should be very low. (ii) Intensity and stability of excitation light. These factors are both optimised by use of a xenon arc lamp or metal halide illumination (see Microscope specifications given below). (iii) The background light level in the optics of the microscope. This is determined by the efficiency of the excitation and emission filters and also by features such as the glass and adhesives used to manufacture the objectives. (iv) Accuracy of quantification is determined partly by the stability of the light supply but also by the uniformity that the light can be distributed to and collected from all regions of the captured image field. Image correction (described below and in ref. 7) can attempt to correct for non-uniformities in these parameters but this far from perfect.

The development of automated laser scanning microscope-based imaging systems offers the prospect of both reduced background light (as a result of confocal optics), higher excitation intensity and more accurate quantification. Also, automated analysis of large numbers of cells and slides would be a major advantage and would enable high-throughput screening for new compounds targeting topoisomerases or testing procedures for potential topoisomerase-mediated genotoxicity.

Unlike the standard comet assay, the TARDIS method detects specific modifications rather than the presence of general DNA breakage or cross-linkage. It could be of interest to combine TARDIS and comet assays to investigate the distribution of adducts at different points in the DNA comet head and tail. Indeed, immunofluorescent staining for topoisomerase IIα has been combined with comet analysis (8). However, detection sensitivity for adducts in the tail may be limited because detection sensitivity is greatest when DNA concentration is maximised by retaining it in a concentrated spot. In this situation, the ratio of immunostaining to background fluorescence signals is maximal.

Analysis of topoisomerase II—DNA adducts

Following its development to quantify low MW DNA adducts, the TARDIS method was adapted to study topoisomerase II adducts. Human DNA topoisomerase II is the target for a number of anti-cancer drugs including etoposide, mitoxantrone, methyl N-(4′-(9-acridinylamino)-3-methoxy-phenyl) methane sulfonamide (mAMSA), daunorubicin, doxorubicin and idarubicin. These drugs act by stabilising a normally transient DNA–topoisomerase reaction intermediate and are termed topoisomerase poisons (9). Type II DNA topoisomerases generate a transient enzyme-bridged DNA double-strand break through which a second duplex can be passed (strand-passage). At this stage in the reaction cycle, the enzyme is covalently coupled to the DNA through a phosphotyrosine linkage. After strand passage, the break is rejoined (10–12). However, the rejoining step is blocked by topoisomerase poisons, leading to accumulation of stabilised topoisomerase II–DNA complexes. The transient reaction intermediates between DNA and the topoisomerase have been referred to as ‘cleavable complexes’ (9) since addition of sodium dodecyl sulphate (SDS) and proteinase K to these complexes produces a cleaved DNA.

Human cells contain two topoisomerase II isoforms, topoisomerase IIα and topoisomerase IIβ. In vitro cleavage assays showed that both isoforms form drug-stabilised cleavable complexes (13,14). To study the isoform-specific effects of drugs in cells, we adapted the TARDIS assay to allow isoform-specific detection of topoisomerase II DNA adducts (15), an outline of the methodology is shown in Figure 1. This study showed for the first time that both topoisomerase isoforms could form DNA complexes when human cells were exposed to etoposide (15). The stabilised complexes increased in a dose-dependent manner and were detectable after a 2-h treatment with 1 μM etoposide. The rate of disappearance of etoposide-stabilised topoisomerase II complexes after drug removal was analysed and the half-life found to be 30 min for topoisomerase IIα and 15 min for topoisomerase IIβ. The TARDIS assay has subsequently been used to assess topoisomerase II adducts levels in human and murine cell lines exposed to a number of compounds listed in Table I. Both topoisomerase II isoforms formed stabilised complexes with DNA when cells were treated with mAMSA or mitoxantrone. The half-life of the mAMSA-stabilised complexes following drug removal was 15 min for both the alpha and beta isoforms. In contrast, the half-life of complexes following mitoxantrone removal was 6 h for beta and 10 h for alpha. Only topoisomerase IIα complexes could be detected following exposure to anthracyclines such as doxorubicin, daunorubicin and idarubicin. Following removal of idarubicin from the cell culture media, the levels of complexes continued to rise for at least 48 h. Two alkylating anthraquinones, Alchemix and ZP275 preferentially stabilised topoisomerase IIα complexes (30), as did DACA and TAS-103 (28). NK314 a synthetic benzo[c]phenanthridine alkaloid targets topoisomerase IIα preferentially (18). A number of naturally occurring compounds including selenite, curcumin, digitoxin and dietary flavonoids stabilise topoisomerase II complexes and have been analysed using the TARDIS assay (21–26) (see Table I). The TARDIS assay has also been used to detect topoisomerase I complexes in response to camptothecin (27) and other compounds (21–25,29). Other groups have used the TARDIS assay (31–33).

Table I

Examples of TARDIS analyses for topoisomerase II (T2) α or β poisoning or catalytic inhibition and for topoisomerase I

Drugs Cell lines TARDIS data
 
References 
T2α poison T2β poison T2 catalytic inhibition Topo I 
Etoposide CEM, MEFs Yes Yes   (15–17
NK314 MEFs Yes Yes   (18
Mitoxantrone CEM, MEFs Yes Yes   (16,17,19
mAMSA MEFs Yes Yes   (16,17
Idarubicin K562 Yes Yes Yes  (20
Idarubicinol K562 Yes Yes   (20
Selenite K562 Yes Yes  Yes (21
Digitoxin K562 Yes Yes Yes  (22,23
Curcumin K562 Yes Yes  Yes (24
Ferulic acid K562 No No  No (24
dibenzoylmethane K562 No No  No (24
Genistein K562 Yes Yes  No (25
luteolin K562 Yes Yes  No (25
Quercetin K562 Yes Yes  No (26
Apigenin K562 Yes Yes  No (26
Fisetin K562 No No Yes No (26
Myricetin K562 Yes Yes  Yes (26
EGCG K562 Yes Yes  Yes Unpublished resultsa 
H202 K562 Yes Yes  Yes Unpublished resultsa 
Camptothecin CEM    Yes (27
DACA CEM Yes Yes  No (28
TAS-103 CEM Yes Yes  Yes (28
XR5944 K562 Yes Yes  Yes (29
XR11576 K562 Yes Yes  Yes (29
ICRF193 MEFs, K562 Yes Yes Yes  Unpublished resultsb 
Aclarubicin MEFs   Yes  Unpublished resultsb 
Merbarone MEFs   Yes  Unpublished resultsb 
Intoplicine MEFs   Yes  Unpublished resultsb 
Novobiocin MEFs   No  Unpublished resultsb 
Suramin MEFs   No  Unpublished resultsb 
Drugs Cell lines TARDIS data
 
References 
T2α poison T2β poison T2 catalytic inhibition Topo I 
Etoposide CEM, MEFs Yes Yes   (15–17
NK314 MEFs Yes Yes   (18
Mitoxantrone CEM, MEFs Yes Yes   (16,17,19
mAMSA MEFs Yes Yes   (16,17
Idarubicin K562 Yes Yes Yes  (20
Idarubicinol K562 Yes Yes   (20
Selenite K562 Yes Yes  Yes (21
Digitoxin K562 Yes Yes Yes  (22,23
Curcumin K562 Yes Yes  Yes (24
Ferulic acid K562 No No  No (24
dibenzoylmethane K562 No No  No (24
Genistein K562 Yes Yes  No (25
luteolin K562 Yes Yes  No (25
Quercetin K562 Yes Yes  No (26
Apigenin K562 Yes Yes  No (26
Fisetin K562 No No Yes No (26
Myricetin K562 Yes Yes  Yes (26
EGCG K562 Yes Yes  Yes Unpublished resultsa 
H202 K562 Yes Yes  Yes Unpublished resultsa 
Camptothecin CEM    Yes (27
DACA CEM Yes Yes  No (28
TAS-103 CEM Yes Yes  Yes (28
XR5944 K562 Yes Yes  Yes (29
XR11576 K562 Yes Yes  Yes (29
ICRF193 MEFs, K562 Yes Yes Yes  Unpublished resultsb 
Aclarubicin MEFs   Yes  Unpublished resultsb 
Merbarone MEFs   Yes  Unpublished resultsb 
Intoplicine MEFs   Yes  Unpublished resultsb 
Novobiocin MEFs   No  Unpublished resultsb 
Suramin MEFs   No  Unpublished resultsb 
a

C. A. Austin and M. Lopez-Lazaro.

b

C. A. Austin and F. Errington.

Fig. 1

Basis of TARDIS analysis (A) Schematic diagram of the steps involved in processing cells for TARDIS analysis: (1) cells are embedded in agarose on a glass slide; (2) after extraction with SDS and salt, most cellular components are removed but genomic DNA and covalent adducts remain; (3) adducts are detected by immunofluorescent staining using specific primary antibodies and FITC-labelled secondary antibody; (4) areas occupied by genomic DNA are identified by Hoechst fluorescence; (5) FITC fluorescence is quantified within the areas defined as containing DNA. (B) Representative images recorded for etoposide-treated K562 cells. Immunofluorescence was carried out with anti-topoisomerase IIα and FITC-conjugated second antibody. (C) Scattergram representation of the topoisomerase IIα signal derived from etoposide-treated K562 cells. Each filled circle represents the integrated fluorescence value of a single nucleus. The horizontal lines represent the median values.

Fig. 1

Basis of TARDIS analysis (A) Schematic diagram of the steps involved in processing cells for TARDIS analysis: (1) cells are embedded in agarose on a glass slide; (2) after extraction with SDS and salt, most cellular components are removed but genomic DNA and covalent adducts remain; (3) adducts are detected by immunofluorescent staining using specific primary antibodies and FITC-labelled secondary antibody; (4) areas occupied by genomic DNA are identified by Hoechst fluorescence; (5) FITC fluorescence is quantified within the areas defined as containing DNA. (B) Representative images recorded for etoposide-treated K562 cells. Immunofluorescence was carried out with anti-topoisomerase IIα and FITC-conjugated second antibody. (C) Scattergram representation of the topoisomerase IIα signal derived from etoposide-treated K562 cells. Each filled circle represents the integrated fluorescence value of a single nucleus. The horizontal lines represent the median values.

Variations

Embedding suspension cells or trypsinised adherent cell culture cells in agarose on a glass slide allows the use of harsh extraction conditions which strips away essentially all non-covalently attached cellular material, leaving genomic DNA and covalent adducts for analysis. This protocol, employed in the TARDIS assay, results in low background signals in untreated cells and corresponding high sensitivity. One variation described by Agostinho et al. (34) utilises less harsh extraction conditions to remove most topoisomerase II not covalently attached or trapped in closed clamp form on the genomic DNA from adherent cells grown on glass slides or coverslips. Under these conditions, extracted cells and their genomic DNA remain attached without the need for agarose embedding. In our hands, this method suffers from higher background levels of topoisomerase fluorescence in untreated cells making it less suitable than the standard TARDIS assay for quantitative analysis. However, it has been used successfully by Agostinho et al. and ourselves to examine the distribution in the nucleus of topoisomerase II–DNA complexes induced by topoisomerase poisons or the topoisomerase inhibitor ICRF193.

Other methods to study topoisomerase II DNA adducts

There are a number of other techniques that have been used to detect and quantify topoisomerase—DNA complexes; these include the SDS–K precipitation method (35), alkaline elution (36), in vivo complex of enzyme (ICE) bioassay (37) and band depletion (38). The SDS–K precipitation and alkaline elution methods are the most sensitive; however, they lack specificity for topoisomerases as any protein–DNA complexes can give a signal and they both also rely on the use of radioactive labelling. The TARDIS assay, ICE bioassay and band depletion are more specific for topoisomerase, and since they use an antibody detection step, they can each be used to detect specific topoisomerase isoforms. The TARDIS and ICE bioassays are comparable in terms of their sensitivity, and both are much more sensitive than band depletion. Band depletion in K562 cells requires >200 μM etoposide to obtain significant depletion of topoisomerase II, whereas the TARDIS assay gives an above background signal with 1 μM for topoisomerase IIα and a robust signal for both isoforms from 10 μM. ICE bioassay gave good signals at 100 μM (18,37). For topoisomerase I, 25 μM camptothecin gave a 50% reduction in band depletion (29), yet only 1 μM was needed to give a reproducible signal in the TARDIS assay (27,29). Ten micromolar topotecan gave a strong ICE bioassay signal (37).

The TARDIS assay is the only one of these assays that detects complexes in individual cells and it requires orders of magnitude fewer cells per assay. This makes it applicable for use on samples from patients during therapy where cell numbers are likely to be limited. For example, acute myeloid leukemia cells treated ex vivo with camptothecin showed a dose-dependent increase in topoisomerase I complexes (27). Samples from patients after treatment with etoposide have also been analysed (C. A. Austin and E. Willmore, unpublished results). The TARDIS assay has also been used to study drug resistant cells lines and to compare the parental sensitive line with the selected drug resistant line. This showed a reduction in topoisomerase II complexes in lines resistant to topoisomerase II drugs (C. A. Austin and E. Willmore, unpublished results).

The TARDIS assay is based on immunofluorescent detection of protein (or DNA-reactive drug) adducts that are covalently coupled to DNA. Such complexes are formed with topoisomerase poisons such as etoposide, mAMSA or mitoxantrone as discussed above. In these cases, the topoisomerase complexes are reversible upon removal of drug. However, topoisomerases (and other chromatin proteins) can also become cross-linked to genomic DNA by compounds that generate reactive oxygen species (ROS). These cross-linked complexes are effectively irreversible in the cell. Thus, the propensity of a compound to generate stabilised topoisomerase–DNA complexes through enzyme poisoning or lead to ROS-mediated cross-linked topoisomerase II–DNA complexes can be distinguished by reversibility studies and the inclusion of antioxidants such as N-acetylcysteine (24).

Comparisons with other methods of assessing genotoxicity indicates that the TARDIS assay has a similar level of sensitivity to the micronucleus assay, comet assay and phosphorylated histone H2AX assay (39). A modified comet assay has been reported where the cells were probed with topoisomerase II antibodies to determine if there were any DNA adducts in the comet tail (8).

There are a range of other uses of the TARDIS assay. It can be used to analyse adduct removal mechanisms in vitro. We have used topoisomerase II adducts on genomic DNA within agarose on slides as a substrate; addition of enzymes to the slides and incubation has been used to determine which proteins are able to remove the toposiomerase II protein. Covalent modifications of toposiomerase could be investigated. Preliminary experiments detected sumoylated proteins (29) and it could be used to detect phosphorylated toposiomerase II or acetylated topoisomerase II if appropriate antibodies were available.

We give details of the preparation of slides for TARDIS analysis, equipment and software requirements and the method of data analysis in sections below.

TARDIS equipment and procedure

Preparing slides

No specialist equipment is required for the preparation of slides for analysis. A detailed protocol is given in the supplementary material, available at Mutagenesis Online. Briefly, cultured cells are plated into six-well plates and exposed to the drug under study. At the end of the exposure period, the cells are washed and suspended in molten low melting point agarose at 37°C. This cell suspension is then spread as a thin layer onto glass slides. The agarose-embedded cells are then extracted with ionic detergent and salt to remove proteins and other molecules not covalently bound to DNA. After washing to remove salt and detergent, slides are processed for immunofluorescence for the protein or molecule of interest. Counterstaining with Hoechst allows the localisation of extracted nuclei on the slide. The TARDIS protocol is outlined in Figure 1A. Use of a conventional immunofluorescence microscope has proved adequate for many studies; however, to attain useful detection sensitivity and accuracy, certain equipment details need to be considered (see microscope and camera specifications given below).

Typically, the immunofluorescent stage employs an unlabelled primary antibody and fluorescently labelled secondary antibody. We have traditionally used fluorescein isothiocyanate (FITC)-labelled second antibodies but other fluorochromes such as AlexaFluor488 (Invitrogen) would also be suitable. Monochrome images are recorded for the Hoechst and FITC channels using appropriate narrow-band filter sets for each field of cells. Examples of topoisomerase II TARDIS images collected for etoposide-treated K562 cells are shown in Figure 1B.

Image analysis

Image analysis consists of three stages. Firstly, images are background and shade corrected to correct for stray light and background originating in the camera electronics and for unevenness in the intensity of illumination, respectively. Secondly, the area corresponding to each nucleus is identified using the Hoechst image to generate a mask around each nucleus. This mask is applied to the FITC image and parameters including number of pixels and average pixel intensity are measured for each masked area (nucleus). In the third stage, these data are exported into a spreadsheet, typically to calculate total FITC and Hoechst intensities measured for each nucleus and this data is represented graphically. An example of the type of output obtained is shown on Figure 1C. We currently use Volocity software (Perkin Elmer) for image management and analysis. A flow diagram for the image analysis is given in Figure 2.

Fig. 2

Flow diagram for TARDIS image capture and analysis.

Fig. 2

Flow diagram for TARDIS image capture and analysis.

The following points need to be considered carefully for successful quantitative immunofluorescence: (i) Exposure time. Exposure times must be sufficiently long to record a robust fluorescent image from the brightest samples, typically in the range of 100 msec to 2 sec for topoisomerase TARDIS, depending on the sample, camera and optics; but the brightest images in a set must not saturate the recording device. This can be checked by taking the sample with the brightest signal or a bright control (for example, cells treated with 100 μM etoposide for topoisomerase II TARDIS) and recording a consecutive series of images at increasing exposures. Plotting the median fluorescence per nucleus should give a linear response against time. Choose an exposure within the linear part of the response. (ii) Bleaching. The above discussion does not take into account fluorescent bleaching of the sample, which may result in lower fluorescent values than expected at longer exposures. Once an optimum exposure is selected in (i) above, record two consecutive images from a single field of cells at that exposure. Both images will give equivalent median fluorescence values unless significant bleaching has occurred. If the second exposure results in significantly reduced fluorescence, consider reducing excitation intensity or use of a more photostable fluorochrome. (iii) Background and shade correction. It is important to minimise the effect of background signal arising from camera electronics, stray light and background fluorescence. This is achieved by recording FITC and Hoechst images from a blank slide or preferably a region of the sample slide containing no cells. This ‘Dark Reference’ image is recorded using the same settings including exposure time as the sample slides. This allows subtraction of the background image from the sample images. Any unevenness in the illumination of the sample is minimised using a shade correction. For this, an image of an evenly fluorescent sample is recorded. This can be achieved using fluorescent plastic slides or solutions of 3 μM fluorescein and 200 μM 4-methylumbeliferone in 20 μm-deep observation chambers (7). These images are called ‘Bright Reference’ images. The intensity of each pixel in the bright reference image is divided by the mean of all of the pixels in the image. The resulting ratios corresponding to each pixel in the reference image are used to correct the corresponding pixels in each sample image. In practice, software such as Volocity allows the user to assemble combined background and shade corrections for each channel (FITC and Hoechst) using the dark and bright reference images and apply this correction to the sample images in one go. The means of achieving this will differ between software packages. (iv) Internal standards and controls. It is useful to include a positive control or internal standard in each set of slides to facilitate comparison between individual experiments. We generally include 100 μM etoposide-treated cells as a positive control for topoisomerase II TARDIS. An alternative approach would be to use fluorescent beads that are available from suppliers such as Invitrogen.

Microscope specifications

We have used two quite different conventional microscope platforms for quantitative immunofluorescence. The main specific requirements are as follows: (i) good quality planar apochromatic high NA ×10 and/or ×20 objectives with low levels of autofluorescence associated with the glass and adhesives used to assemble the lenses, (ii) a bright stable fluorescent light source ideally a xenon arc lamp or metal halide arrangement such as an X-cite system (EXFO), suitable narrow band filter sets for FITC (or other fluorochrome) and Hoechst fluorescence and sensitive monochrome cooled CCD camera (see below). Systems can be motorised or fully manual, though considerable time can be saved during image acquisition through partial automation with a motorised system. Specifications for the original system used are described by Frank et al. (7) and we also use a similar system based on a Leica DMLB microscope. Recently, we have also employed an Olympus IX81 motorised microscope (see Table II).

Table II

Microscope specifications used for topoisomerase TARDIS

System Leica DMLB Olympus IX81 
Operation Manual Motorised/automated 
Illumination 75 W Xenon Metal Halide X-cite120Q 
Filters Narrow band pass ‘filter cubes’ Narrow band pass filter cubes 
Objectives (suitable for TARDIS) ×10 HC PL APO NA 0.4 ×10 VPLA FL NA 0.3. 
×20 HC PL APO NA 0.7 
Camera Photometris Coolsnap HQ2 Hamamatsu Orca-AG 
System Leica DMLB Olympus IX81 
Operation Manual Motorised/automated 
Illumination 75 W Xenon Metal Halide X-cite120Q 
Filters Narrow band pass ‘filter cubes’ Narrow band pass filter cubes 
Objectives (suitable for TARDIS) ×10 HC PL APO NA 0.4 ×10 VPLA FL NA 0.3. 
×20 HC PL APO NA 0.7 
Camera Photometris Coolsnap HQ2 Hamamatsu Orca-AG 

Camera specifications

Fluorescence arising from DNA adducts is faint compared to typical immunofluorescence of cellular structures in normally fixed cells, so a sensitive monochrome camera system is required. In the system described by Frank et al. (4) for the detection of melphalan adducts, images were captured using a cooled slowscan CCD with 16-bit (64 000 greyscale) output and 770 × 1152 pixels (Astrocam). A similarly specified recent model is the Hamamatsu Orca II BT 1024 (see Table II). For quantification of topoisomerase DNA complexes as well as the above-mentioned system, we have used both Hamamatsu ORCA-AG and Photometrics Coolsnap HQ2 cameras. Both of these cameras employ cooled CCDs with 6.45 μm photosites and 12/14-bit AD conversion. Other required characteristics for TARDIS or indeed any quantitative immunofluorescence analysis include: high quantum efficiency, large photosite full-well capacity and low dark current. These attributes relate to efficiency of converting photons to electrons at the photosite, number of electrons to ‘fill’ the photosite and accumulation of electrons at photosites in the absence of illumination (thermal noise), respectively. Camera models are continually being updated by manufacturers, but some values for the above attributes are given in Table III as a guide. Suitable image sensors contain about a million photosites (1 megapixel) and are operated in 2 × 2 binning mode, whereby 2 × 2 squares of photosites acts as a single entities.

Table III

Camera characteristics (data from manufacturer’s datasheets)

 Orca-BTII 1024 (Hamamatsu) Orca-AG (Hamamatsu) Coolsnap HQ2 (Photometrics) 
Photosite size 13 × 13 μm 6.45 × 6.45 μm 6.45 × 6.45 μm 
No of pixels 1024 × 1024 1344 × 1024 1393 × 1040 
AD conversion 16 bit 12 or 16 bit 14 bit 
Full-well capacity 80 K e 18 K e 16 K or 30 K e binned 
Dark current 0.01 e/pixel/sec 0.03 e/pixel/sec 0.001 e/pixel/sec (−30°C) 
 Orca-BTII 1024 (Hamamatsu) Orca-AG (Hamamatsu) Coolsnap HQ2 (Photometrics) 
Photosite size 13 × 13 μm 6.45 × 6.45 μm 6.45 × 6.45 μm 
No of pixels 1024 × 1024 1344 × 1024 1393 × 1040 
AD conversion 16 bit 12 or 16 bit 14 bit 
Full-well capacity 80 K e 18 K e 16 K or 30 K e binned 
Dark current 0.01 e/pixel/sec 0.03 e/pixel/sec 0.001 e/pixel/sec (−30°C) 

Supplementary data

Supplementary material is available at Mutagenesis Online.

Funding

Leukaemia and Lymphoma Research (07038).

Conflict of interest statement: None declared.

References

1.
Tilby
MJ
Styles
JM
Dean
CJ
Immunological detection of DNA damage caused by melphalan using monoclonal antibodies
Cancer Res.
 , 
1987
, vol. 
47
 (pg. 
1542
-
1546
)
2.
Tilby
MJ
Johnson
C
Knox
R
Cordell
J
Roberts
JJ
Dean
CJ
Sensitive detection of DNA modifications induced by cisplatin and carboplatin in vitro and in vivo using a monoclonal antibody
Cancer Res.
 , 
1991
, vol. 
51
 (pg. 
123
-
129
)
3.
Tilby
MJ
McCartney
H
Cordell
J
Frank
AJ
Dean
CJ
A monoclonal antibody that recognises alkali-stabilised melphalan-DNA adducts and its application in immunofluorescence microscopy
Carcinogenesis
 , 
1995
, vol. 
16
 (pg. 
1895
-
1901
)
4.
Frank
AJ
Tilby
MJ
Quantification of DNA adducts in individual cells by immunofluorescence: effects of variation in DNA conformation
Exp. Cell Res.
 , 
2003
, vol. 
283
 (pg. 
127
-
134
)
5.
Ahnstrom
G
Erixon
K
Radiation induced strand breakage in DNA from mammalian cells. Strand separation in alkaline solution
Int. J. Radiat. Biol.
 , 
1973
, vol. 
23
 (pg. 
285
-
289
)
6.
Smith
CC
O'Donovan
MR
Martin
EA
hOGG1 recognizes oxidative damage using the comet assay with greater specificity than FPG or ENDOIII
Mutagenesis
 , 
2006
, vol. 
21
 (pg. 
185
-
190
)
7.
Frank
AJ
Proctor
SJ
Tilby
MJ
Detection and quantification of melphalan-DNA adducts at the single cell level in hematopoietic tumor cells
Blood
 , 
1996
, vol. 
88
 (pg. 
977
-
984
)
8.
Chen
S
Gomez
SP
McCarley
D
Mainwaring
MG
Topotecan-induced topoisomerase IIα expression increases the sensitivity of the CML cell line K562 to subsequent etoposide plus mitoxantrone treatment
Cancer Chemother. Pharmacol.
 , 
2002
, vol. 
49
 (pg. 
347
-
355
)
9.
Liu
LF
DNA topoisomerase poisons as antitumor drugs
Annu. Rev. Biochem.
 , 
1989
, vol. 
58
 (pg. 
351
-
375
)
10.
Liu
LF
Liu
CC
Alberts
BM
Type II DNA topoisomerases: enzymes that can unknot a topologically knotted DNA molecule via a reversible double-strand break
Cell
 , 
1980
, vol. 
19
 (pg. 
697
-
707
)
11.
Wang
JC
DNA topoisomerases
Annu. Rev. Biochem.
 , 
1996
, vol. 
65
 (pg. 
635
-
692
)
12.
Austin
CA
Marsh
KL
Eukaryotic DNA topoisomerase IIβ
Bioessays
 , 
1998
, vol. 
20
 (pg. 
215
-
226
)
13.
Cornarotti
M
Tinelli
S
Willmore
E
Zunino
F
Fisher
LM
Austin
CA
Capranico
G
Drug sensitivity and sequence specificity of human recombinant DNA topoisomerases IIα (p170) and IIβ (p180)
Mol. Pharmacol.
 , 
1996
, vol. 
50
 (pg. 
1463
-
1471
)
14.
Marsh
KL
Willmore
E
Tinelli
S
Cornarotti
M
Meczes
EL
Capranico
G
Fisher
LM
Austin
CA
Amsacrine-promoted DNA cleavage site determinants for the two human DNA topoisomerase II isoforms α and β
Biochem. Pharmacol.
 , 
1996
, vol. 
52
 (pg. 
1675
-
1685
)
15.
Willmore
E
Frank
AJ
Padget
K
Tilby
MJ
Austin
CA
Etoposide targets topoisomerase IIα and IIβ in leukemic cells: isoform-specific cleavable complexes visualized and quantified in situ by a novel immunofluorescence technique
Mol. Pharmacol.
 , 
1998
, vol. 
54
 (pg. 
78
-
85
)
16.
Errington
F
Willmore
E
Leontiou
C
Tilby
MJ
Austin
CA
Differences in the longevity of topo IIα and topo IIβ drug-stabilized cleavable complexes and the relationship to drug sensitivity
Cancer Chemother. Pharmacol.
 , 
2004
, vol. 
53
 (pg. 
155
-
162
)
17.
Errington
F
Willmore
E
Tilby
MJ
Li
L
Li
G
Li
W
Baguley
BC
Austin
CA
Murine transgenic cells lacking DNA topoisomerase IIβ are resistant to acridines and mitoxantrone: analysis of cytotoxicity and cleavable complex formation
Mol. Pharmacol.
 , 
1999
, vol. 
56
 (pg. 
1309
-
1316
)
18.
Toyoda
E
Kagaya
S
Cowell
IG
, et al.  . 
NK314, a topoisomerase II inhibitor that specifically targets the alpha isoform
J. Biol. Chem.
 , 
2008
, vol. 
283
 (pg. 
23711
-
23720
)
19.
Padget
K
Quantitative analysis and drug sensitivity of human topoisomerase II α and β
PhD Thesis
 , 
1998
Newcastle upon Tyne, UK
Department of Biochemistry and Genetics, University of Newcastle upon Tyne
20.
Willmore
E
Errington
F
Tilby
MJ
Austin
CA
Formation and longevity of idarubicin-induced DNA topoisomerase II cleavable complexes in K562 human leukaemia cells
Biochem. Pharmacol.
 , 
2002
, vol. 
63
 (pg. 
1807
-
1815
)
21.
Lopez-Lazaro
M
Willmore
E
Elliott
SL
Austin
CA
Selenite induces topoisomerase I and II-DNA complexes in K562 leukemia cells
Int. J. Cancer
 , 
2008
, vol. 
123
 (pg. 
2217
-
2221
)
22.
Lopez-Lazaro
M
Pastor
N
Azrak
SS
Ayuso
MJ
Cortes
F
Austin
CA
Digitoxin, at concentrations commonly found in the plasma of cardiac patients, antagonizes etoposide and idarubicin activity in K562 leukemia cells
Leuk. Res.
 , 
2006
, vol. 
30
 (pg. 
895
-
898
)
23.
Lopez-Lazaro
M
Pastor
N
Azrak
SS
Ayuso
MJ
Austin
CA
Cortes
F
Digitoxin inhibits the growth of cancer cell lines at concentrations commonly found in cardiac patients
J. Nat. Prod.
 , 
2005
, vol. 
68
 (pg. 
1642
-
1645
)
24.
Lopez-Lazaro
M
Willmore
E
Jobson
A
Gilroy
KL
Curtis
H
Padget
K
Austin
CA
Curcumin induces high levels of topoisomerase I- and II-DNA complexes in K562 leukemia cells
J. Nat. Prod.
 , 
2007
, vol. 
70
 (pg. 
1884
-
1888
)
25.
Lopez-Lazaro
M
Willmore
E
Austin
CA
Cells lacking DNA topoisomerase IIβ are resistant to genistein
J. Nat. Prod.
 , 
2007
, vol. 
70
 (pg. 
763
-
767
)
26.
Lopez-Lazaro
M
Willmore
E
Austin
CA
The dietary flavonoids myricetin and fisetin act as dual inhibitors of DNA topoisomerases I and II in cells
Mutat. Res.
 , 
2010
, vol. 
696
 (pg. 
41
-
47
)
27.
Padget
K
Carr
R
Pearson
AD
Tilby
MJ
Austin
CA
Camptothecin-stabilised topoisomerase I-DNA complexes in leukaemia cells visualised and quantified in situ by the TARDIS assay (trapped in agarose DNA immunostaining)
Biochem. Pharmacol.
 , 
2000
, vol. 
59
 (pg. 
629
-
638
)
28.
Padget
K
Stewart
A
Charlton
P
Tilby
MJ
Austin
CA
An investigation into the formation of N- [2-(dimethylamino)ethyl]acridine-4-carboxamide (DACA) and 6-[2-(dimethylamino)ethylamino]- 3-hydroxy-7H-indeno[2, 1-C]quinolin-7-one dihydrochloride (TAS-103) stabilised DNA topoisomerase I and II cleavable complexes in human leukaemia cells
Biochem. Pharmacol.
 , 
2000
, vol. 
60
 (pg. 
817
-
821
)
29.
Jobson
AG
Willmore
E
Tilby
MJ
Mistry
P
Charlton
P
Austin
CA
Effect of phenazine compounds XR11576 and XR5944 on DNA topoisomerases
Cancer Chemother. Pharmacol.
 , 
2009
, vol. 
63
 (pg. 
889
-
901
)
30.
Pors
K
Paniwnyk
Z
Teesdale-Spittle
P
Plumb
JA
Willmore
E
Austin
CA
Patterson
LH
Alchemix: a novel alkylating anthraquinone with potent activity against anthracycline- and cisplatin-resistant ovarian cancer
Mol. Cancer Ther.
 , 
2003
, vol. 
2
 (pg. 
607
-
610
)
31.
Marchion
DC
Bicaku
E
Turner
JG
Daud
AI
Sullivan
DM
Munster
PN
Synergistic interaction between histone deacetylase and topoisomerase II inhibitors is mediated through topoisomerase IIβ
Clin. Cancer Res.
 , 
2005
, vol. 
11
 (pg. 
8467
-
8475
)
32.
Lynch
A
Harvey
J
Aylott
M
Nicholas
E
Burman
M
Siddiqui
A
Walker
S
Rees
R
Investigations into the concept of a threshold for topoisomerase inhibitor-induced clastogenicity
Mutagenesis
 , 
2003
, vol. 
18
 (pg. 
345
-
353
)
33.
Lu
HR
Zhu
H
Huang
M
Chen
Y
Cai
YJ
Miao
ZH
Zhang
JS
Ding
J
Reactive oxygen species elicit apoptosis by concurrently disrupting topoisomerase II and DNA-dependent protein kinase
Mol. Pharmacol.
 , 
2005
, vol. 
68
 (pg. 
983
-
994
)
34.
Agostinho
M
Rino
J
Braga
J
Ferreira
F
Steffensen
S
Ferreira
J
Human topoisomerase IIα: targeting to subchromosomal sites of activity during interphase and mitosis
Mol. Biol. Cell
 , 
2004
, vol. 
15
 (pg. 
2388
-
2400
)
35.
Rowe
TC
Grabowski
D
Ganapathi
R
Isolation of covalent enzyme-DNA complexes
Methods Mol. Biol.
 , 
2001
, vol. 
95
 (pg. 
129
-
135
)
36.
Kohn
KW
Principles and practice of DNA filter elution
Pharmacol. Ther.
 , 
1991
, vol. 
49
 (pg. 
55
-
77
)
37.
Subramanian
D
Furbee
CS
Muller
MT
ICE bioassay. Isolating in vivo complexes of enzyme to DNA
Methods Mol. Biol.
 , 
2001
, vol. 
95
 (pg. 
137
-
147
)
38.
Kaufmann
SH
Svingen
PA
Immunoblot analysis and band depletion assays
Methods Mol. Biol.
 , 
1999
, vol. 
94
 (pg. 
253
-
268
)
39.
Watters
GP
Smart
DJ
Harvey
JS
Austin
CA
H2AX phosphorylation as a genotoxicity endpoint
Mutat. Res.
 , 
2009
, vol. 
679
 (pg. 
50
-
58
)

Supplementary data