Mutagenesis, Vol. 16, No. 2, 179-182,
March 2001
© 2001 UK Environmental Mutagen Society/Oxford University Press
Industrial Genotoxicology Group (IGG): Cytotoxicity In Vitro,Royal Society of Medicine, London, UK, 6 December 1999
1 AstraZeneca, Loughborough and 2 Covance Laboratories, Harrogate, UK
 |
Introduction |
|---|
 Top Introduction MechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
Some chemicals are not detectable in in vitro genotoxicity assays unless the concentrations tested are cytotoxic. However, excessive toxicity often does not allow a proper evaluation of the relevant genetic end point (CPMP/ICH, 1995
). At very low survival levels in mammalian cells mechanisms other than direct genotoxicity per se can lead to `positive' results that are related to cytotoxicity and not genotoxicity. These considerations are conflicting. They have been balanced by defining a desired level of toxicity for the top concentrations of cytotoxic compounds in in vitro genotoxicity tests. However, recently the appropriate level of toxicity at the upper limit of testing in the in vitro chromosomal aberration test has been challenged (Galloway, 2000). The objective of the meeting was to consider how cells die and measurement of cytotoxicity in relation to genotoxicity testing.
|
Mechanisms |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
Kevin Chipman (University of Birmingham, Birmingham, UK) distinguished between necrosis and apoptosis and highlighted target molecules for toxicants. He pointed out that mechanisms of toxicity to a cell may differ depending on the concentration of the test substance. For instance, while low doses may stimulate protective systems, increased concentrations may cause cells to proliferate or to die by apoptosis as an extreme tissue protective response. Even higher concentrations, particularly involving ATP depletion, will eventually cause necrotic cell death by one or a combination of several mechanisms. A range of assays is available to distinguish the mechanism of toxicity. Some implications of loss of cell viability on DNA damage and mutation were introduced. DNA damage may contribute to toxicity and, conversely, toxicity may indirectly affect DNA integrity. In particular, secondary genotoxicity through glutathione depletion associated with toxicity was considered.
Colin Arlett (MRC Cell Mutation Unit, Brighton, UK) pointed out that the use of survival curves in genetic toxicology is in comparisons of dose but may be useful to determine the relative significance of specific DNA lesions. Cell death may be estimated on the basis of short-term tests applied immediately or during treatment using dye exclusion by live cells or release of a radioactive metabolite from dead cells. Such tests tend to underestimate the proportion of dead cells and do not take into account the ability of cells to recover or repair damage. While clonal cell survival may be considered to be the most appropriate measure of response, it can only be used in situations where there is an acceptable level of cloning efficiency. Surrogate assays of cloning such as the MTT assay have been proposed as practical alternatives, with the advantages of speed and automation. However, two examples, based on the response of human fibroblasts to cis-platin or UVB radiation, indicate that the MTT assay may be up to 1000 times less sensitive than the clonal assay.
An understanding of the mechanism of the increased sensitivity to UV radiation of human T cells in comparison with fibroblasts is achieved by measurement of specific DNA lesions. Fibroblasts are killed by the 6–4 photoproduct while T cells are killed by a combination of this lesion and its Dewar isomer. Since unstimulated human T cells may not be subject to apoptosis and since these cells do not undergo necrosis, then an alternative pathway to lethality must be sought. The relationship between lethal events and chromosome damage serve as a reminder that mutations to the cancerous state are only relevant to cells which remain capable of replication.
Michael Green (University of Brighton, Brighton, UK) observed that mammalian cells have a finite lifespan, possibly linked to the shortening of telomeres with age. While Dolly the sheep has short telomeres, human cells that express telomerase become immortal. He suggested that senescence was a third way by which cells die. Senescence in cells resembles the effects of DNA damage and vice versa. Thus DNA damage accelerates telomere loss and may explain the phenomenon of delayed clastogenicity, several generations after exposure to ionizing radiation. If DNA damage causes senescence, it is possible that cell survival in viable treated cells may be different from the untreated population, for example by dividing more unevenly and thus appearing as uneven sized colonies. Since end points such as gene mutation are only measured in cells that survive, only a certain level of sensitivity is achievable, and it is not known whether the proportion of mutants is the same in surviving, as opposed to dead, cells. Although mutation is a highly relevant end point for cancer, measuring mutation has an inbuilt error. He speculated as to what mechanism of DNA damage might be specific to a >80% reduction in viability in the mouse lymphoma assay (MLA) and whether this may equate to scoring necrotic cells in the Comet assay.
|
Overview of how cytotoxicity is measured in each of the tests |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
According to the International Conference on Harmonisation (ICH), for the bacterial reverse mutation test a reduction in the number of revertants or a clearing or dimunition of the background lawn is desirable. For in vitro cytogenetic tests (IVCA), for a cell line a >50% reduction in cell number or confluency or for lymphocyte cultures an inhibition of mitotic index (MI) by >50% is considered sufficient. In mammalian cell mutation tests (MLA) ideally the highest concentration should produce at least 80–90% toxicity (no more than 10–20% survival).
Julie Clements (Covance Laboratories, Harrogate, UK) defined the measures of cytotoxicity used in the MLA: relative survival (RS) is the relative cloning efficiency (CE) at the end of treatment; relative suspension growth (RSG) is a measure of suspension growth during the expression period; relative total growth (RTG) is a combined measure of suspension growth and cloning efficiency at the end of the expression period. Current issues for the microwell version of the assay include: (i) should the RS be adjusted to take account of loss of cells during the treatment period; (ii) should the RTG include the treatment period (3 and 24 h)? At an informal meeting of microwell MLA users it was recently agreed that RTG should be measured from the start of the treatment period. As a follow-up to the International Workshop on Genotoxicity Test Procedures (IWGTP) for the MLA in Washington, 1999, data were gathered to compare the measures of toxicity used in the MLA with a view to recommending one measure. (Retrospective note: at the second IWGTP in New Orleans, 2000, it was decided that the single recommended measure of toxicity would be the RTG and that this should include the treatment period.)
Gill Clare (AstraZeneca, Loughborough, UK) reviewed the main ways by which cytotoxicity is measured in the IVCA and in vitro micronucleus test (IVMNT). The method of measurement tends to depend upon whether lymphocytes or cell lines are used and if the cell line is grown as a monolayer (when subjective measurements of cell density are possible).
Counting cells at mitosis or interphase is universally applicable. As the MI is derived from the same population of cells at the same time as the analysis of chromosome damage, this measure of viability is directly relevant to this end point. Qualitative information on the appearance and number of cells and the morphology of the chromosomes can also be gained. However, the measure is time-dependent, providing information on surviving cells only for the time when the metaphase-arresting agent is present. This means that toxicity may be underestimated if there is substantial cell loss before the MI is measured and that it may be disproportionately influenced by compounds that interfere with a specific stage of the cell cycle. Also, it tends to be variable.
Cell counts provide an indirect measure of either cell loss or lack or slowing of division, but cannot distinguish between them. The efficacy of the measurement depends on sufficient time having elapsed between the start of exposure and counting. The majority of cells also need to divide and survive intact in the negative control population. In short-term lymphocyte cultures total cell counts are insensitive, as only a proportion of lymphocytes from the population of white blood cells divide and early stages of culture are characterized by net loss of white cells.
The rate of cell division can be measured for the period when bromodeoxyuridine is present which, for example, will show if compounds interfere with a specific stage of the cell cycle.
For the IVMNT a significant proportion of nuclei must have divided once. If a method such as cell counting is used it is important that the vast majority of cells are dividing at a rate that is known. Cytochalasin B will permit nuclear, but not cytoplasmic, cleavage. This makes the IVMNT analogous to that in vivo, since cells that have divided once following exposure to the compound can be identified and analysed for micronuclei. However, there are a few disadvantages associated with the use of cytochalasin B (Fenech, 1997) and there are those who consider that it holds no advantages for cell lines (Kalweit et al., 1999).
V79 cells at metaphase round up and detach from the flask. At an interval after exposure to the substance mitotic cells are shaken off, allowed to divide, then prepared and analysed for micronuclei. V79 cells have also been cultured as single cells on slides and then, when the slides are prepared, the number of cells in each cluster is used to calculate the number of cell divisions.
Brian Burlinson (Glaxo Wellcome) reviewed the cytotoxicity measurements and requirements for the unscheduled DNA synthesis (UDS) and Comet assays. For the former, the initial measurement was for viability of the cells, by trypan blue exclusion, after isolation from the liver. In the absence of compound-related cytotoxicity mean viability values of ~75% have been reported in hepatocytes taken from the rat (Ashby et al., 1985). Other measures of cytotoxicity relied upon the subjective assessment of the morphology of the cells both prior to and during scoring. Those slides showing overt cytotoxicity, i.e. an increased number of apoptotic and/or pyknotic cells, are discarded and for those kept only morphologically normal cells are scored.
The importance of cytotoxicity measurements in the Comet assay was highlighted at the IWGTP. As DNA damage is associated with cell death, it is critical that the highest dose tested does not induce excessive cytotoxicity and that cytotoxicity is evaluated concurrently with each Comet experiment. Although no particular method for measuring cytotoxicity was recommended, the expert panel agreed that the maximum acceptable reduction in viability should be ~30% (Tice et al., 2000). The panel also agreed that cytotoxicity must be considered when interpreting any results and comment made upon any potential confounding factors related to cytotoxicity.
|
Practical experience |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
The following issues were discussed, with reference to data:
- comparability of data when different methods are used to measure cytotoxicity;
- issues relating to measurement of cytotoxicity and setting the top concentration in an assay;
- how conflicting results from more than one method of measuring cytotoxicity during an assay have been resolved;
- whether cytotoxicity is a cause as well as a consequence of genetic damage;
- whether one assay is more likely than another to give positive results, because of the method for measuring cytotoxicity.
Gill Clare presented the results of a questionnaire that had been circulated among members. Sixteen replies had been received from 13 companies. The majority agreed that high cytotoxicity interfered with the interpretation of genotoxicity data and had examples of compounds that induced principally gaps and breaks or gene mutations at cytotoxic concentrations only. For selection of the highest concentration in the in vitro chromosome aberration test the reduction in the MI aimed for was just over 50%, but acceptable values varied from just over 50 to 80%. An additional comment was that there should not be a large reduction in cell numbers. If this was the case, then a reduction in MI to 40% of control values was acceptable. Two respondees used cell counts. They considered that reductions of 80–90% were acceptable. Subjective assessments of the likelihood of finding 100 metaphases to score on a slide were also made. Other methods of selecting the top dose included [3H]thymidine labelling, the quality of the metaphases and cell cycle delay.
Most laboratories tried to get as close to the recommended upper level of cytotoxicity as practicably possible. Tests would be repeated with a narrower range of concentrations, spaced by less than 2-fold dilutions. Up to about 10 concentrations would be used. The number of repeat tests varied between laboratories and depended on the toxicity profile of the compound. It was common to use more than one method to measure cytotoxicity during an assay. For the MLA, RS, CE and RTG were used. For the IVCA, MI, cell counts, effects on chromosome morphology and proliferative index were used. For the IVMNT, the proportion of multinucleate to mononucleate cells and cell counts were used. There were many examples of a conflict in results. These included differences in sensitivity between RS, RTG and CE, MI increased when there was a cell cycle delay in M phase and a normal MI but cell numbers greatly reduced after 3 h treatment. Examples given of experience with new or alternative cytotoxicity assays included the MTT (or XTT) and Alamar Blue assays, incorporation of [3H]thymidine, counting lymphocytes using magnetic particles to isolate T cells and flow cytometry.
|
IVCA results |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
Leslie Akhurst (Huntingdon Life Sciences Ltd, Alconbury, UK) compared cell counts and MI using CHL cells. For one study a cell count that was 40% of the control value was obtained, but the MI was reduced to only 9% of the control value. Examples were given using data from two surfactants, one enzyme and one pharmaceutical intermediate. One of the surfactants only gave a positive response at dose levels that reduced the MI by
50%. No increase was seen at intermediate dose levels that were moderately toxic. The enzyme was only positive when using a continuous treatment and showed a response at the highest dose level which reduced the MI to 42% of the control value, but not at the intermediate dose level which reduced the MI to 50%. The pharmaceutical intermediate was mainly positive at dose levels that reduced the MI to 30–40% of control values. There had subsequently been a discussion about selection of the top dose with the sponsor, as one of the concentrations had reduced the MI to 54% of the control value. When the audience were asked how readily they would choose a top concentration that reduced the MI to <50% of control values, one person responded that a regulator had not accepted data where the highest concentration showed a 49% reduction! Richard Marshall (Covance Laboratories, Harrogate, UK) had evaluated 44 data sets from assays with CHO or CHL cells where mitotic inhibition and relative cell number at harvest had been analysed as measures of toxicity. The method of choice for selecting the maximum concentration for analysis with these cells was cell number. However, on 12 occasions numbers of mitotic cells limited the maximum dose that could be analysed. On 16 occasions a higher concentration would have been selected if MI rather than cell number had been used as the criterion and on the remaining 16 cell number and mitotic inhibition agreed. Examples of comparative data were presented.
The merit of using relative population doubling from the start of treatment as an alternative to cell number at harvest, as proposed by Hema Murli and Sheila Galloway, was discussed. This approach generally results in the selection of lower concentrations for analysis and may prevent high toxicity leading to incorrect conclusions for clastogenicity.
Peter Jenkinson (Safepharm Laboratories, Shardlow, UK) compared cell counts versus MI data using CHL cells. He said that for many chemicals there is little convergence between these two measures of toxicity. In the example, depending on the particular exposure group, either cell count, MI or neither method was appropriate for dose selection. If an inappropriate measure was used, then a valid positive response could easily be overlooked. The conclusion was that dose selection was improved by the use of multiple measures of toxicity and that no one single method was appropriate for all cases.
Andreas Czich (RCC Cytotest Cell Research, Rossdorf, Germany) said that of 98 compounds tested between June 1997 and January 1999, five were positive without toxicity, 33 were positive with toxicity, 30 were negative without toxicity and 30 were negative with toxicity. He compared and contrasted the use of the XTT assay, MI and relative cell numbers as methods for measuring cytotoxicity in the chromosome aberration assay. He noted that the XTT assay was poorly comparable with a microscopic evaluation. One of the issues for debate is the use of precipitating concentrations. In the context of selection of the top dose for analysis, the ICH recommend that toxicity, rather than precipitation, is the criterion for selection of the top dose, except where a precipitate interferes with scoring of the test. Data were presented using precipitating concentrations.
Eryl Jones (AstraZeneca CTL, Alderley Park, UK) reviewed 160 chromosome aberration tests. Of these, the top concentration was limited by precipitate (5%), MI (6%), a combination of MI and insufficient metaphases on the slide (20%), insufficient metaphases on the slide (31%) or to either 5 mg/ml or 10 mM (38%). The same concentrations of mitomycin C and cyclophosphamide had been used for many tests. The MI varied widely, from ~20 to 120% relative to the negative controls. There was no clear relationship between the MI and the percentage of cells showing chromosome damage.
|
IVMNT results |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
Mike O'Donovan (AstraZeneca, Loughborough, UK) presented data on long-term lymphocytes, where cells had been exposed to 1.25, 2.5, 5 or 10 µg/ml hycanthone methanesulphonate for 3 h, counted and divided into aliquots with or without cytochalasin B. After 24 h incubation the percentage of multinucleate cells and the percentage of binucleates with micronuclei had been counted from cultures exposed to cytochalasin B and cells had been counted from the other cultures. Cytotoxicity was expressed as the per cent increase in control cell numbers or the proportions of multinucleate to mononucleate cells. There was an increased frequency of micronuclei after exposure to all concentrations. There was no decrease in the percentage of binucleate to mononucleate cells (~58% in control and exposed cultures). The mean increase in cell numbers relative to the control decreased from 89% after 1.25 µg/ml to 45% after 10 µg/ml. In this case binucleate incidence was not a good measure of cytotoxicity, probably because cell loss before addition of cytochalasin B was not taken into account.
|
MLA results |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
Robert Durward (Safepharm Laboratories, Shardlow, UK) showed that, although in the majority of cases there is good agreement, in some examples there can be a large discrepancy between toxicity of a chemical as measured by RS or RTG. However, the calculation of RS includes a correction factor for cells lost during the exposure phase of the experiment. If this correction factor is included in the RTG calculation, then the correlation between the two measures of toxicity is improved.
Mike O'Donovan (AstraZeneca, Loughborough, UK) presented data on colchicine, showing that it was negative even at a survival well below 10%. Similar toxicities were seen using RS and RSG, with slightly higher toxicities expressed as RTG. Ethidium bromide has been suggested as a compound for which RS and RTG give very different estimates of toxicity. This is true, with similar RTG seen at levels three times lower than RS or RSG. Most of this difference is because much lower cloning efficiency was seen after 48 h expression than immediately after the end of treatment. His opinion was that examples such as ethidium bromide should not be used to make decisions about the most appropriate measure of cytotoxicity.
Mick Fellows (Covance Laboratories, Harrogate, UK) presented a review of microwell MLA studies performed over the last 12 months. The database included approximately 50 studies, the majority consisting of two experiments in the absence and presence of S9 and many with at least one 24 h treatment experiment in the absence of S9. The database included toxic and non-toxic compounds. Experience suggests that the toxicity measures RS and RTG often give contrasting results. In general, RTG gives a higher estimate of toxicity. This is especially evident if the RTG calculation includes growth/loss of cells during the treatment period. Does this mean that if RS was used as the sole indicator of toxicity we would often be testing to excessively high doses or is RTG over-reliant on suspension growth?
Phil Clay (AstraZeneca, Alderley Park, UK) reviewed 94 studies performed over the last 3 years, of which 16 were positive. Dose levels were chosen on the basis of RS, with the highest dose at 10–20% RS (not scored if <10%). The RTG was calculated retrospectively for those studies that were positive. Most studies showed no difference between RS and RTG, but four studies showed significant reproducible differences. For all four studies RTG was less than RS. He concluded that using RTG would not have affected the outcome of the studies.
|
New/alternative ways of measuring cytotoxicity |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
Julie Hayes (AstraZeneca, Alderley Edge, UK) discussed the advantages and disadvantages of the Alamar Blue (resazurin) assay. CHL cells were exposed to mitomycin C or cyclophosphamide. Cytotoxicty was expressed in terms of the measured concentration of Alamar Blue, which is converted from a blue non-fluorescent dye to a pink fluorescent product in the presence of living cells. Micronuclei were measured after exposure to concentrations of the positive control compounds that were known to induce chromosome damage in this test system.
Victor Oreffo (AstraZeneca, Loughborough, UK) presented results on gene expression analysis following in vivo or in vitro exposure to the oxidative stress agents menadione and diquat. Livers from CD-1 mice and primary hepatocyte cultures were compared. Transcription profiling was carried out using Affymetrix GeneChip analysis. Results suggested that a substantial percentage of genes were similarly affected following exposure to the two related redox cyclers. Up-regulation of acute phase proteins and down-regulation of metabolic enzymes were found to be a consistent feature of in vivo toxicity. Analysis of in vitro samples revealed a markedly different pattern of gene expression. Individual gene expression changes are being assessed using a branched DNA amplification technique. Initial results indicate that haemoxygenase induction (a sensitive marker of oxidative stress) is compound- and cell type-dependent. Differences in expression pattern are likely to be attributable to differences in toxic response.
|
Summary |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
The importance of cytotoxicity in relation to genotoxicity tests cannot be overstated.
One of the main determinants in setting doses in short-term genotoxicity assays is cytotoxicity. At present the methods recommended for the particular assay should be followed. However, different measures of cytotoxicty and ways of measuring it can generate different results. This raises the question as to whether the different measures and methods of measurement are equally acceptable.
For the IVCA, a comparison of the sensitivities of the various methods for measuring cytotoxicity, development of methods for reliably counting the target lymphocytes and agreement on standards for scoring MI would be useful.
Since mean values of 75% viability have been reported from hepatocytes taken from rats for the UDS assay in vivo, it would seem reasonable that this level of viability should be achievable in vitro.
High levels of cytotoxicity can generate false positive results. The relationship between cytotoxicity and genotoxicity may not always follow simple predictable rules. Appropriate and thorough measurements of cytotoxicty, coupled with an understanding of mechanisms, will help to delineate relevant from irrelevant genotoxicity findings in vitro. One way forward might be to gather further data on reference toxins that induce damage by an indirect mechanism of action.
|
References |
|---|
|
TopIntroductionMechanismsOverview of how cytotoxicity...Practical experienceIVCA resultsIVMNT resultsMLA resultsNew/alternative ways of...SummaryReferences |
|---|
-
Ashby,J., Lefevre,P.A., Burlinson,B. and Penman,M.G. (1985) An assessment of the in vivo rat hepatocyte DNA repair assay. Mutat. Res., 156, 1–18.[Web of Science][Medline]
CPMP/ICH (1995) Note for guidance on genotoxicity: guidance on specific aspects of regulatory genotoxicity tests for pharmaceuticals, paper presented to the International Conference on Harmonisation, Topic S2A. CPMP/ICH/141/95.
Fenech,M. (1997) The advantages and disadvantages of the cytokinesis-block micronucleus method. Mutat. Res., 362, 11–18.
Galloway,S.M. (2000) Cytotoxicity and chromosome aberrations in vitro: experience in industry and the case for an upper limit on toxicity in the aberration assay. Environ. Mol. Mutagen., 35, 191–201.[Web of Science][Medline]
Kalweit,S., Utesch,D., von der Hude,W. and Madle,S. (1999) Chemically induced micronucleus formation in V79 cells—comparison of three different test approaches. Mutat. Res., 439, 183–190.[Web of Science][Medline]
Tice,R., Agurell,E., Anderson,D., Burlinson,B., Hartman,A., Kobayashi,H., Miyamae,Y., Rojas,E., Ryu,J.C. and Sasaki,Y.F. (2000) The single cell gel/comet assay: guidelines for in vitro and in vivo genetic toxicology testing. Environ. Mol. Mutagen., 35, 206–221.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||