Miniaturized flow cytometry-based in vitro primary human lymphocyte micronucleus assay—validation study
Abstract
Most in vitro mammalian genotoxicity assays show a low specificity (high rate of irrelevant positive results), and therefore, lead to an increase in follow-up in vivo genotoxicity testing. One of the sources of the high rate of in vitro irrelevant positive results that find no confirmation in in vivo studies may be the characteristics of the test system used. It has been shown that cells that are p53 deficient or carry an alteration in DNA repair genes may be more prone to produce high rate of false/irrelevant positive results. Primary human lymphocytes (HuLy) are considered to show a higher specificity in predicting the in vivo genotoxic potential of a tested compound. We recently developed a flow cytometry-based primary human T-lymphocyte micronucleus test (MNT) and showed that the technology is promising and reliable in detecting genotoxic compounds. The purpose of the present work was to develop and validate a miniaturized format of the assay. For validation purposes of the flow cytometry HuLy MNT a wide selection of compounds with different mechanisms of genotoxicity was used. The evaluation covered 30 compounds: 19 commercially available genotoxicants and nongenotoxicants and 11 early pharmaceutical development compounds. Being faster and less tedious than the microscopic analysis, the miniaturized flow cytometry-based methodology showed very promising results. Conveniently, cell division is verified within the same sample as the MN frequency. Moreover analysis of hypodiploid events may provide an indication for a mode of action, i.e. clastogenic versus aneugenic mechanism, for further follow-up testing. Mol. Mutagen. 2012. © 2012 Wiley Periodicals, Inc.
INTRODUCTION
In vitro mammalian genotoxicity tests show low specificity (high rate of irrelevant positive results) and, therefore, increase in follow-up in vivo genotoxicity tests [Kirkland et al.,2007,2008]. As a consequence the development of more accurate in vitro mammalian assays may significantly decrease the number of animals used for the in vivo tests and improve the predictivity of carcinogenicity and genotoxicity for humans. One of the sources of the high rate of irrelevant positive results (in vitro positive result that finds no confirmation in in vivo studies) may be the characteristics of the test system used [Kirkland et al.,2007]. It is believed that the cells that are p53 deficient or carry an alteration in DNA repair genes may be more prone to produce to the high rate of false/irrelevant positive results [Kirkland et al.,2007]. Rodent cell lines exhibit a higher rate of irrelevant positive results than other cell systems. On the contrary, primary human lymphocytes (HuLy) are considered to show more specificity in predicting the results in vivo [Kirkland et al.,2007].
The recently revised draft ICH S2 guideline (2008) accepted the in vitro micronucleus (MN) test as an alternative method for chromosome aberration test. The micronucleus test (MNT) is an assay that detects both aneugenic and clastogenic activity of tested compounds. Moreover, the application of additional follow-up studies like fluorescence in situ hybridization enables the determination of the mode of action and differentiation between clastogenic and aneugenic activity. Currently, the gold standard method for the in vitro MN test is the microscopy-based cytochalasin B (cyto B) HuLy MN test. Cyto B induces cytokinesis block [Carter,1967; Fenech and Morley,1985; Kirkland et al.,2008] and, therefore, enables simultaneous monitoring of cytotoxicity and MN induction in cells that have undergone division. Despite the fact that the HuLy MNT is faster and less tedious to evaluate than the chromosome aberration test, there have been attempts to automate the assay by means of image analysis [Tates et al.,1990; Castelain et al.,1993; Verhaegen et al.,1994; Varga et al.,2004; Decordier et al.,2009] and flow cytometry [Nüsse and Kramer,1984; Viaggi et al.,1995; Wessels and Nusse,1995; Nüsse and Marx,1997]. The advantages of both technologies are the reduction of analysis time and scorer subjectivity. Yet the main challenges for the flow cytometric MN assessment are (1) the elimination of the apoptotic and necrotic debris that may interfere with MN scoring gate and (2) application of the appropriate proliferation assay that shows the comparable sensitivity to cyto B, in the context of primary cells. Recently we published [Lukamowicz et al.,2011] the proof of principle of the novel flow cytometry HuLy MNT. The assay is based on a three-step staining procedure: carboxyfluorescein succinimidyl ester (CFSE) as a proliferation marker, ethidium monazide (EMA) for chromatin of necrotic and late apoptotic cells discrimination and 4,6-diaminodino-2-phenylindole (DAPI) as DNA marker.
The purpose of the study presented in this manuscript was to develop and validate a miniaturized format of the assay based on a 96-multiwell plate to fit a high throughput platform of the flow cytometry. Miniaturization of the technology is in particular advantageous for the early phase drug development due the low sample volumes and limited availability of test compounds. For the purposes of the flow cytometry-based HuLy MNT validation a wide selection of compounds with different mechanisms of genotoxicity was used. The evaluation covered 30 compounds: 19 commercially available genotoxicants and nongenotoxicants and 11 early pharmaceutical development compounds.
MATERIALS AND METHODS
Cells and Culture Medium
Blood was obtained from eight healthy, nonsmoking donors aged between 20 and 47 years. Peripheral blood mononuclear cells (PBMC) were isolated using Vacutainer CPT tubes (Becton Dickinson, Franklin Lakes, NJ) using 8 ml per Vacutainer CPT tube. Approximately 15 ml were required to prepare 96-multiwell plate of cell culture. The cultures were prepared using blood from a single donor per experiment. The isolated PBMC were washed three times with phosphate buffered saline (Oxoid, Hampshire, England).
The flow cytometry-based HuLy MNT technology has already been published elsewhere [Lukamowicz et al.,2011]. Briefly, PBMCs before cultivation were stained for 10 min at room temperature with 5 µM carboxyfluorescein diacetate succinimidyl ester (CFDA-SE) purchased from Fluka (Steinheim, Germany) in a final density of 50 × 106 cells/ml. CFDA-SE diffuses through cell membranes, where it is cleaved by a nonspecific cellular esterase forming CFSE. CFSE binds to cellular proteins and is equally divided between daughter cells upon cell division and therefore enables cell proliferation monitoring [Lyons and Parish,1994; Lyons,1999; Hasbold and Hodgkin,2000]. The intensity of CFDA-SE staining was controlled with a set of controls: (1) cellular autofluorescence (cells were cultivated after isolation and washing step, omitting the CFDA-SE staining) and (2) “nondivided” cells (a culture of nonstimulated PBMCs stained with CFDA-SE was prepared using culture medium without phytohemaglutinin).
T-lymphocyte cultures at a concentration of 0.7 × 106 cells/ml were incubated in 37°C in a humid atmosphere with 5% CO2. The culture medium consisted of RPMI 1640 + GlutaMAX I + 25 mM HEPES (Invitrogen, Merelbeke, Belgium) supplemented with 15% (v/v) of heat inactivated fetal bovine serum (Invitrogen, Merelbeke, Belgium), 1% (v/v) Penicillin-Streptomycin (Invitrogen, Merelbeke, Belgium) and 3.75% Phytohemaglutinin (Invitrogen, Merelbeke, Belgium). The cultures were maintained in 96 U-bottom plates, 300 µl of culture volume per well. These conditions were selected as the most efficient of those experimentally analyzed.
Compounds and Treatment Schedules
The compounds were selected and classified based on ECVAM recommended list of genotoxic (positive and weak positive) and nongenotoxic (false positive and negative) chemicals for the assessment of the performance of new or improved genotoxicity tests [Kirkland et al.,2008] and from list for validation from Annex 3 of the OECD guideline 487 [OECD Guideline,2010] (Table I). The validation included also the 11 selected internal Novartis early development compounds (A–K).
| Compound | Cas. number | Providera | Classification | HuLy MNT | Hypodiploid classification | |
|---|---|---|---|---|---|---|
| 1b | Cytosine arabinoside | 147–94–4 | Fluka | Potent clastogen | Positive | Potent aneugen |
| 2 | Cisplatin | 15,663–27–1 | Aldrich | Potent clastogen | Positive | Potent aneugen |
| 3 | Bleomycin | 9,041–93–4 | Sigma | Potent clastogen | Positive | Potent clastogen |
| 4b | Mitomycin C | 50–07–7 | Sigma | Potent clastogen | Positive | Potent clastogen |
| 5 | Etoposide | 33,419–42–0 | Sigma | Potent clastogen | Positive | Potent clastogen |
| 6b | Colchicine | 64–86–8 | Sigma | Potent aneugen | Positive | Potent aneugen |
| 7 | Carbendazim | 10,605–21–7 | Aldrich | Potent aneugen | Positive | Potent aneugen |
| 8b | Vinblastine | 143–67–9 | Sigma | Potent aneugen | Positive | Potent aneugen |
| 9 | Vincristine | 2,068–78–2 | Sigma-Aldrich | Potent aneugen | Positive | Potent aneugen |
| 10 | N-Ethyl-N-nitrosourea | 759–73–9 | Sigma | Weak genotoxicant | Positive | Potent clastogen |
| 11 | 5-Fluorouracil | 51–21–8 | Sigma | Weak genotoxicant | Positive | Potent clastogen |
| 12 | Hydroxyurea | 127–07–1 | Sigma | Weak genotoxicant | Positive | Potent clastogen |
| 13b | Di-(2-ethylhexyl)phthalate | 17–81–7 | Fluka | Nongenotoxicant | Negative | Negative |
| 14b | Nalidixic acid | 389–08–2 | Fluka | Nongenotoxicant | Negative | Negative |
| 15b | Pyrene | 129–00–0 | Aldrich | Nongenotoxicant | Negative | Negative |
| 16 | Dexamethasone | 50–02–2 | Sigma | Apoptogenic | Negative | Negative |
| 17 | Benzyl alcohol | 100–51–6 | Sigma-Aldrich | False positive | Negative | Negative |
| 18 | Tertiary-butylhydroquinone | 1,948–33–0 | Fluka | False positive | Negative | Negative |
| 19 | Phthalic anhydride | 85–44–9 | Fluka | False positive | Negative | Negative |
The solvent selected for all test chemicals was dimethyl sulfoxide (Sigma-Aldrich, Steinheim, Germany).
Following the Novartis standard operating procedures treatment was performed 24 hr after culture initiation for 44 hr with no recovery time. Due to the cell cycle alterations induced by the 5-fluorouracil the treatment was performed 24 hr after culture initiation for 20 hr followed by 24 hr recovery time.
Flow Cytometry-Based HuLy MNT
The flow cytometry analysis was performed with a FACS LSR II (using the Diva 6.1.2 software) cytometer equipped with argon (488 nm) and UV (355 nm) lasers.
The staining procedure was in detail described elsewhere [Lukamowicz et al.,2011] and was modified to the 96-multiwell miniaturized platform. These modified conditions were selected as the most efficient of the experimentally analyzed. Briefly, at harvest cell cultures were centrifuged at 300g for 7 min. In the first step, cells were stained with 60 µl of EMA solution (covalently binds the DNA of cells with disrupted membrane integrity—dead/dying cells) [Avlasevich et al.,2006]. As follows cells were lysed in a two-step procedure and stained with DAPI (100 µl of lysis solution 1 + 100 µl of lysis solution 2). The specimens were measured immediately after the staining using the BD high throughput sampler (HTS: BD 96-well plate automated sample acquisition handler) device that provide the automated sample delivery for the flow cytometry analysis. The analysis was performed using 100 µl of sample volume, 1.5 µl/sec, sample flow speed, that showed to be the most efficient of all tested. Moreover, the system was washed with 200 µl between the wells. Each sample was mixed two times with an application of 100 µl of mixing volume with the speed of 180 µl/sec.
Ten thousand gated, divided nuclei (EMA negative, DAPI positive, CFSE indicating cell division) per well were analyzed (eight cultures per concentration for standard compounds, four cultures per concentration for Novartis early development compounds).
The template for analysis uses several parameters, including forward scatter, side scatter, DAPI, and EMA fluorescence, to exclude potential contamination of the MN and nucleus regions with misclassified events like late apoptotic/necrotic chromatin or debris. The detailed gating strategy is described by Lukamowicz et al. [2011].
Cytotoxicity Assessment
Top concentration should aim at 55% ± 5% of cytotoxicity for the in vitro micronucleus assay [OECD Guideline,2010] The Replicative Index RI proposed by Kirsch-Volders et al. [2004] and recommended by OECD guideline 487 [Fellows et al.,2008; Lorge et al.,2008; OECD Guideline,2010] was used as a cytoxicity parameter with the 60% cytotoxicity (40% of RI) cut-off value. If no cytotoxicity was observed the top concentration was 1 mM or 0.5 mg/ml whichever was the lower in accordance with the recommendations (ICH draft guideline S2(R1), 2008).
The CFSE-based results were calculated as follows. The number of nuclei classified with CFSE as nondivided was divided by the factor 1 (A). The number of nuclei classified with CFSE as 1 division is divided by the factor 2 (B). The number of nuclei classified with CFSE as 2 divisions was divided by the factor 4 (C), 3 divisions was divided by the factor 8 (D) and 4 divisions were divided by the factor 16 (E).
The final RI for was calculated as follows:
-
A = Number of nuclei (0 division)/1
-
B = Number of nuclei (1 division)/2
-
C = Number of nuclei (2 divisions)/4
-
D = Number of nuclei (3 divisions)/8
Hypodiploid
It is suggested that compounds that show aneugenic activity form an increased rate of hypodiploid events, possibly due to a nondisjunction and chromosome loss processes [Bryce et al.,2010,2011]. In order to determine the MOA of the tested compounds, the frequency of hypodiploid events was analyzed.
The hypodiploid gate was defined as the events within the range of 1/3 to 2/3 of G1 peak of DAPI fluorescence and the size of G1 peak (Fig. 1). The hypodiploid events were selected from the population that was accepted by the gating strategy based on the FSC, SSC, DAPI fluorescence, and EMA fluorescence [in detail described by Lukamowicz et al.,2011—plots A–F]. The hypodiploid frequency was calculated as a ratio of the number of hypodiploid events to the number of divided nuclei and expressed as fold increase over concurrent solvent control. Compounds that induced a more than or equal to threefold increase of hypodiploid events over solvent control values were considered as potent aneugens.
Hypodiploid gating strategy based on FSC-A and DAPI-fluorescence. The plots represent negative control sample (top panel), colchicine (0.018 µM), and mitomycin C (0.03 µg/ml) treated sample (bottom panel). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Experiment Validity Criteria
-
The MN frequency of the solvent control was within the range of the solvent control historical values (MN% = 0.46 ± 0.26, n = 38 cultures; valid min = 0.20%, valid max. = 0.72).
-
The MN frequency of the positive control (MMC 0.03 µg/ml) was within the range of positive control historical values (MN% = 2.91 ± 0.79, n = 16 cultures; valid min. = 2.12%, valid max. = 3.70).
Positivity Criteria for MN Test
-
The MN frequency was above the maximum value of the historical negative control range and
-
The frequency of MN showed at least a twofold increase over the solvent control.
RESULTS
The miniaturization and validation of the flow cytometry HuLy MNT was performed using 30 compounds: 19 standard genotoxic or nongenotoxic compounds and 11 internal early development Novartis compounds.
The standard compounds were selected based on literature recommendations [Kirkland et al.,2008; OECD Guideline,2010]. Early development pharmaceutical compounds classification was based on results obtained from microscopic MN test.
All negative compounds (di-(2-ethylhexyl)phthalate, nalidixic acid, pyrene, compound I, compound J, compound K) showed no MN induction up to the highest test concentration (1 mM, solubility limitations or toxicity) (Figs. 2 and 3, Tables I and II). No induction of hypodiploid events was observed for all negative compounds (Tables I and II).
Cytotoxicity and MN induction after treatment with standard genotoxic and nongenotoxic compounds. *More than or equal to twofold increase over concurrent solvent control; na- concentration not analyzed due to elevated cytotoxicity. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
Cytotoxicity and MN induction after treatment with early stage development compounds. *More than or equal to twofold increase over concurrent solvent control; na- concentration not analyzed due to elevated cytotoxicity. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
| Compound | Microscopy MNT | Flow cytometry MNT | Hypodiploid MOA classification |
|---|---|---|---|
| A | Positive | Positive | Potent aneugen |
| B | Positive | Positive | Potent aneugen |
| C | Positive | Positive | Potent aneugen |
| D | Positive | Positive | Potent aneugen |
| E | Positive | Positive | Potent aneugen |
| F | Positive | Positive | Potent aneugen |
| G | Weak positive | Weak positive | Potent clastogen |
| H | Weak positive | Positive | Potent aneugen |
| I | Negative | Negative | Negative |
| J | Negative | Negative | Negative |
| K | Negative | Negative | Negative |
- The compounds were analyzed by microscopy and flow cytometry. The MOA classification is based on flow cytometry quantification of hypodiploid events. Compounds that induced more than or equal to threefold increase of hypodiploid events over solvent control values were considered as potent aneugens.
Compounds that were expected to induce an increase in MN frequency were positive with the flow cytometry analysis (Figs. 2 and 3, Tables I and II). All tested compounds classified as potent aneugens (colchicine, carbendazim, Vincristine, vinblastine) induced an increase in hypodiploid events (i.e. showed an aneugenic signature) (Table I). An aneugenic signature was noted as well for compounds A and B that was confirmed in follow-up analysis with fluorescence in situ hybridization (data not shown). Clastogens included in this validation showed no increase in hypodiploid frequency, except for two compounds: cytosine arabinoside and cisplatin. However, for these two compounds the increase in hypodiploid events was observed, only at highly cytotoxic concentrations (cytosine arabinoside 43% RI, cisplatin 51% RI). Moreover, for cytosine arabinoside and cisplatin the accumulation of cells in the G1 phase of the cell cycle after the first division was observed.
All compounds that were classified as false positive—compounds that potentially may produce an irrelevant positive result (benzyl alcohol, tertiary-butylhydroquinone, and phthalic anhydride) were classified as negative. Tertiary-butylhydroquinone was tested up to the 58% of RI. A higher concentration induced excessive cell death that made the analysis impossible in two independent experiments. No clastogenic/aneugenic signature was observed for all tested false-positive compounds.
Dexamethasone (DEXA), an apoptogenic agent, did not induce an increase in MN frequency. Moreover no hypodiploid induction was noted.
DISCUSSION AND CONCLUSIONS
The purpose of this study was to miniaturize and validate the previously described flow cytometry-based primary human lymphocyte micronucleus test [Lukamowicz et al.,2011]. Moreover, the determination of the potential mode of action (clastogenic/aneugenic activity) of a test compound was conducted.
MN Induction
The micronuclei are expressed only in the population of cells that have completed mitosis. Therefore, for an in vitro test approach, the demonstration of cell division is mandatory. Currently a “gold standard” for in vitro primary MNT is a Cytochalasin B (Cyto B) micronucleus test. Cyto B is an inhibitor of actin polymerization, which induces cytokinesis block [Carter,1967; Fenech and Morley,1985; OECD Guideline,2010] and therefore enables proliferation monitoring. The main challenge for the flow cytometry automation of HuLy MNT was to develop a reliable and sensitive proliferation assay. In our study we present the miniaturization and validation of the previously described flow cytometry-based HuLy MNT assay [Lukamowicz et al.,2011]. In this novel technology CFSE is used as a proliferation marker. CFSE is a fluorescent product of cellular carboxyfluorescein diacetate succinimidyl ester (CFDA-SE) cleavage that freely diffuses through cell membranes. CFSE fluorescence is halved with each generation and therefore enable proliferation monitoring in whole cells [Lyons and Parish,1994; Lyons,1999] and isolated nuclei [Hasbold and Hodgkin,2000]. In our approach EMA is used as an apoptotic/necrotic cell discriminating dye [Avlasevich et al.,2006] and DAPI as a DNA dye.
The data presented in this study demonstrate that the flow cytometry-based HuLy MNT assay [Lukamowicz et al.,2011] can successfully be miniaturized. Miniaturization was performed on the level of cell cultivation, treatment, staining, and analysis that provided the significant reduction of the required compound amount. Moreover, the flow cytometry measurement was performed with the application of the HTS device that significantly improved the assay performance, reduced the technical assistance, and analysis time.
Our data show that all tested compounds classified as positive or weak positive induced an increase in MN frequency. Positive MNT results performed with MMC and VB are in agreement with the results presented by Lukamowicz et al. [2011]. DEXA induced no increase in the frequency of MN, in agreement with previously published data [Lukamowicz et al.,2011]. In the presented study all negative compounds were correctly classified. Compounds that were regarded as false positives by ECVAM workshop [Kirkland et al.,2008] showed no MN induction and, therefore, the potential reduction of false-positive result was shown. The data presented at the ECVAM workshop [Kirkland et al.,2008], showed a positive result in association with a high concentration or elevated toxicity. Therefore, it may be hypothesized that application of a sensitive cytotoxicity parameter (in our study: CFSE division marker in combination with EMA staining) with a limited concentration range as recommended by the draft ICH guideline show a way for reduction of in vitro false-positive results.
Mode of Action (MOA) of MN Induction
MN as a genotoxicity endpoint can be induced by two main groups of compounds depending on the MOA: aneugens and clastogens [Mateuca et al.,2006; Kirsch-Volders et al.,2011]. Aneugens interact with the elements of the mitotic apparatus leading to chromosome or chromatid loss or missegregation. Clastogens, on the contrary, may induce double strand (ds) or single strand (ss) DNA breaks that are converted into ds DNA breaks after DNA replication, resulting in the formation of MNs. Aneugens show a characteristic steep dose–response with a threshold [Elhajouji et al.,1995; Aardema et al.,1998; Elhajouji et al.,2011] and MNs are expected to be bigger than those induced by clastogens [Hashimoto et al.,2010]. Moreover, it was suggested that aneugens form an increased rate of hypodiploid events, possibly due to a nondisjunction and chromosome loss processes [Bryce et al.,2010,2011]. Our study shows that the hypodiploid analysis approach described by Bryce et al [2010,2011] is applicable to the assay presented by Lukamowicz et al. [2011]. All tested aneugens showed an aneugenic signature, which is in agreement with published data for CHO-K1 cells [Bryce et al.,2010,2011]. For all tested clastogens, except for cytosine arabinoside and cisplatin, no indication for an aneugenic activity was noted. For cytosine arabinoside and cisplatin, an increase in hypodiploid events was observed only in conjugation with elevated cytotoxicity. Moreover, the analysis of the cell cycle changes induced by cytosine arabinoside and cisplatin showed an accumulation of cells in G1 phase in contradiction with the expected G2 arrest after aneugen exposure [Elhajouji et al.,1998; Freidberg et al.,2005]. Therefore, we speculate that the observed increase in hypodiploid events caused by cytosine arabinoside and cisplatin treatment is a result of an increased cell degradation associated with the high toxicity of the concentrations used.
In conclusion, we have shown that the flow cytometry-based primary human T-lymphocyte MNT method previously described by our laboratory is a reliable technology. Being faster and less tedious than the microscopic analysis, the miniaturized flow cytometry-based methodology showed very promising results especially for early screening purposes. Conveniently, the division tracking is scored within the same sample as the MN frequency. Moreover, analysis of hypodiploid events may provide an indication for a mode of action for further follow-up testing.
