Cell Transformation Assays as Predictors of Human Carcinogenicity

The Report and Recommendations of ECVAM Workshop 391-3

Reprinted with minor amendments from ATLA 27: 745-767

Robert Combes,4 Michael Balls,5 Rodger Curren,6 Michel Fischbach,7 Norbert Fusenig,8 David Kirkland,9 Claude Lasne,10 Joseph Landolph,11 Robert LeBoeuf,12 Hans Marquardt,13 Justin McCormick,14 Lutz Muller,15 Edgar Rivedal,16 Enrico Sabbioni,15 Noriho Tanaka,17 Paule Vasseur13 and Hiroshi Yamasakil9
4FRAME, Russell & Burch House, 96-98 North Sherwood Street, Nottingham NG1 4EE, UK; 5ECVAM, JRC Institute for Health and Consumer Protection, 21020 Ispra (VA), Italy; 6Institute for In Vitro Sciences, 21 Firstfield Road, Suite 220, Gaithersburg, MD 20878, USA; 728/30 Via Trieste, 21030 Brinzio (VA), Italy; 8Division of Carcinogenesis and Differentiation, FSII, 0240, German Cancer Research Centre, im Neuenheimer Feld 280, 69120 Heidelberg, Germany; 9Covance Laboratories, Otley Road, Harrogate, North Yorkshire HG3 lPY, UK; 10Bureau of Chemical Substances and Preparations, Ministry of Land and Country Planning and Environment, 20 Avenue de Ségur, 75302 Paris 07 SP, France; 11USC/Kenneth Norris Jr, Comprehensive Cancer Centre and Hospital, 1441 Eastlake Avenue, P.O. Box 33804, Los Angeles, CA 90033-0804, USA; 12Procter & Gamble, Temselaan 100, 1853 Strombeek-Bever, Brussels, Belgium; 13Department of Toxicology, Hamburg University Medical School, Grindelallee 117, 20146 Hamburg, Germany; 14Carcinogenesis Laboratory, FST Building, Michigan State University, East Lansing, MI 48824-1316, USA; 15Mutagenicity and Carcinogenicity Section, Federal Institute for Drugs and Medical Devices, Seestrasse 10, 13353 Berlin, Germany; 16Institute for Cancer Research, Laboratory for Environmental and Occupational Cancer, The Norwegian Radium Hospital, Montebello, 0310 Oslo, Norway; 17Laboratory of Cell Toxicology, Department of Cellular and Genetic Toxicology, Hatano Research Institute/Food and Drug Safety Centre, 729-5 Ochiai, Hadano, Kanagawa 257, Japan; 13Centre des Sciences de l'Environnment, 1 rue des Récollets, BP 94025, 57040 Metz Cedex 1, France; 19IARC, 150 cours Albert Thomas, 69372 Lyon Cédex 08, France

1ECVAM - The European Centre for the Validation of Alternative Methods. 2This document represents the agreed report of the participants as individual scientists.3The paper by Tsuchiya, T. et al. (ATLA 27, 685-702, 1999) "An international validation study of the improved transformation assay employing Balb/c 3T3 cells: results of a collaborative study on the two-stage cell transformation assay by the Non-genotoxic Carcinogen Study Group" has not been considered in this report.

Address for correspondence: Dr Robert Combes, FRAME, Russell & Burch House, 96-98 North Sherwood Street, Nottingham NG1 4EE, UK.

Address for reprints: ECVAM, TP 580, JRC Environment Institute, 21020 Ispra (VA), Italy

Preface

This is the report of the thirty-ninth of a series of workshops organised by the European Centre for the Validation of Alternative Methods (ECVAM). ECVAM's main goal, as defined in 1993 by its Scientific Advisory Committee, is to promote the scientific and regulatory acceptance of alternative methods which are of importance to the biosciences and which reduce, refine or replace the use of laboratory animals. One of the first priorities set by ECVAM was the implementation of procedures which would enable it to become well-informed about the state-of-the-art of non-animal test development and validation, and the potential for the possible incorporation of alternative tests into regulatory procedures. It was decided that this would be best achieved by the organization of ECVAM workshops on specific topics, at which small groups of invited experts would review the current status of in vitro tests and their potential uses, and make recommendations about the best ways forward (1). In addition, other topics relevant to the Three Rs (reduction, refinement and replacement) concept of alternatives to animal experiments have been considered in several ECVAM workshops.

The workshop on Cell Transformation Assays as Predictors of Human Carcinogenicity was held in Angera, Italy on 12-16 October 1998. The principal aim of the workshop was to seek consensus on ways of increasing the use of mammalian cell transformation assays, especially in human cell systems, for fundamental and applied studies in carcinogenesis, and for the regulatory testing of carcinogens, and to make practical recommendations to facilitate this process. This was endorsed by the fundamental belief of all the participants that the further development, evaluation, routine use and eventual regulatory acceptance of cell transformation assays, in conjunction with other toxicity information, would improve the overall process of safety and risk assessment of carcinogenicity, for the protection of human health. This belief was based on the fundamental premise that the process of in vitro cell transformation closely models the carcinogenic process in vivo.

It should be noted that this report is focused on three rodent cell transformation systems, i.e. those employing primary Syrian hamster embryo cells (SHE cell assay), and two systems based on the mouse fibroblast cell lines, Balb/c 3T3 and C3H/1OT 1/2. This report also covers the recent advances with human cell transformation systems.

Introduction

Cell transformation and carcinogenesis

Carcinogenesis has been shown to be a multi-step process, which involves sequential genetic alterations in a single target cell, which cause subtle alterations in growth control and culminate in cells that are able to form malignant tumours (2, 3; see Figure 1). Genetic changes can result from spontaneous or carcinogen-induced alterations in DNA. Non-genotoxic mechanisms, that are at least initially independent of direct DNA damage, can play a causal role in carcinogenesis (4).

Figure 1: Diagrammatic Representation of Carcinogenesis In Vivo and Cell Transformation In Vitro

a) Carcinogenesis in vivo; b) cell transformation in vitro.

Research to elucidate the mechanisms of carcinogenesis has involved experimental animals, human clinical material, and in vitro cellular and molecular biological methods. In particular, the study of carcinogenesis was greatly facilitated by the discovery of in vitro morphological transformation of mammalian cells in culture some 35 years ago (5, 6). The phenomenon of morphological cell transformation involves changes in the behaviour and growth control of cultured cells, characterized by one or more of the following (depending on the cell system under consideration): alterations in cellular morphology; disorganized patterns of colony growth; and acquisition of anchorage-independent growth (7, 8). Such effects result from changes in the expression of oncogenes and/or tumour suppressor genes.

Cell transformation has been defined as the induction of certain phenotypic alterations in cultured cells that are characteristic of tumorigenic cells (9). These phenotypic alterations can be induced by exposing normal cells to carcinogens, or by expressing activated oncogenes in such cells. Transformed cells that have acquired all the characteristics of malignant cells have the ability to form invasive tumours in susceptible animals. Traditionally, this is assessed by injecting transformed cells into athymic mice or newborn animals of the same species. However, for some transformation systems, such as the SHE cell assay, the ability to exhibit anchorage-independent growth has been shown to be highly predictive of the tumour-forming ability of transformed cells in animals (10). In this report, we use the term "cell transformation" to mean that cells have acquired the properties or characteristics usually considered to be exhibited by tumorigenic cells, and the phrase "malignant transformation" to indicate that the cells are able to form a malignant tumour (cancer) in a susceptible animal host.

Since its discovery, in vitro cell transformation has been shown to be a multistage process which closely models the various stages of in vivo carcinogenesis (11, 12). Several pieces of evidence support this assertion:

  1. induction of cell transformation results from exposure to chemicals which are carcinogenic in whole animals;
  2. malignantly transformed cells injected into susceptible host animals can form tumours, whereas non-transformed cells do not; and
  3. in vitro cell transformation occurs in several stages, as shown by the acquisition of several distinct phenotypes as cells progress from the normal phenotype to the malignant phenotype in transfection experiments with activated-oncogenes, and by the fact that tumour promoting agents can enhance the frequency of cell transformation in cells treated with low concentrations of carcinogens (13, 14).

It should be noted that the number of genetic events required for different types of tumours to develop (15) varies according to the tumour type, thereby complicating the development of in vitro models which begin with a mixed population of cell types. There might, however, be common and early key events, which could be detected in vitro, and these should be sought out to facilitate the development of non-animal, optimised models of carcinogenesis.

Cell transformation assays

The possibility of being able to study carcinogenesis, and to detect carcinogenic hazard to humans, by using cells in culture, has long been recognized. It seems clear that in vitro cell transformation can provide some crucial evidence specific to the tumorigenic potential of a chemical, which cannot be supplied by genotoxicity testing. While several in vitro transformation assays have been developed, their relevance and reliability for identifying genotoxic carcinogens has not been widely accepted. Indeed, there has been more emphasis on the application of genotoxicity tests for this purpose (14), despite their known limited ability to detect all types of carcinogens. Increased interest in the use of cell transformation to detect carcinogens has been prompted by the fact that some cell transformation assays are responsive to chemicals acting via a diverse range of mechanisms, including genotoxic and non-genotoxic mechanisms (11, 16). Also, there is accumulating evidence that at least some of the rodent cell transformation assays exhibit sensitivities and specificities for predicting chemical carcinogenicity, which are comparable to, or better than, those shown by several of the established genotoxicity tests (12).

There are several reasons for the relatively low use of cell transformation assays to predict carcinogenicity, including:

  1. the technical skills required;
  2. the lack of characterization and standardization; and
  3. the fact that the relevance and reliability (in terms of the endpoints measured) of the assays have not been established in independent studies to the point where they can be considered to have been validated for regulatory acceptance and use.

With certain cell transformation assays, concerns have also been raised about dependence on high levels of cytotoxicity to cause increases in transformation frequency. There are also several concerns about the use of cell transformation to identify carcinogens when the process does not model all the stages of carcinogenesis. Thus, although cancer is due to genetic changes in single cells, the characteristic phenotypic changes of the isease are only manifested at the tissue level Cancer cells therefore represent only part of the process of tumorigenesis (17). However, based on the relatively high sensitivity of cell transformation assays for detecting carcinogens, this would indicate that chemical effects at the cellular level can accurately reflect the ability of a chemical to induce the cellular and tissue changes required for neoplastic development.

Objectives of the workshop

In an International Agency for Research on Cancer (IARC) review on cell transformation assays (18), it was concluded that "refinements of cell transformation tests, including the use of human cells, human xenobiotic metabolism and appropriate tissue-specific target cells, are needed". Since that time, there has been an interest in developing and improving cell transformation assays for carcinogen identification, and several recent developments with rodent cell transformation assays might now make them suitable for use in regulatory testing. In particular, modifications to the SHE cell assay protocol, which make it more reproducible and more predictive of carcinogenicity (12, 19, 20), have prompted recommendations for rodent cell transformation assays to be seriously considered for regulatory testing purposes. This changing situation was reflected in a recent IARC publication, which endorsed the potential value of results from cell transformation assays in reviewing the carcinogenic potential of chemicals (21).

In the light of these developments, the principal objectives of the workshop were:

  1. to assess the current scientific status and relevance of cell transformation assays for detecting carcinogens, and for studying the phenomenon of carcinogenesis at the cellular level;
  2. to decide on the most appropriate role that the currently available cell transformation assays, either as routine tests or specific research tools, could play in detecting carcinogens;
  3. to assess the feasibility of developing transformation assays based on human cells to improve the detection of carcinogens of relevance to humans;
  4. to discuss possible ways in which transformation assays could be improved with respect to their relevance, reliability and level of protocol standardization; and
  5. to formulate a set of recommendations to facilitate the use of the most promising assays for carcinogen detection.

Rodent Cell Transformation Assays

Rodent cell transformation assays involve either finite life-span cells, such as the SHE cell assay, or immortalized fibroblast cell lines, such as the Balb/c 3T3 or C3H/1OT 1/2 transformation systems. The theoretical advantage of using finite life-span cells is that the individual changes that the cells must undergo to become malignant can be studied. In practice, however, quantitative tests have usually been developed for one step in the carcinogenic process, and this forms the basis of the transformation assay. It can be difficult to maintain cell lines in culture when they exhibit low frequencies of spontaneous transformation. Another limitation is that the genetic and/or non-genetic changes that give rise to these cell lines are unknown. Moreover, it is only possible to study those changes which still must be acquired by the immortalised cells for them to become malignantly transformed. Morphological transformation of cell colonies and focus formation on a monolayer are the most commonly used endpoints, regardless of the type of system used.

Primary culture assays

One of the earliest assays which used primary cells was the SHE cell transformation assay (5, 22-24), which uses the formation of transformed colonies as the endpoint. The cells employed in the assay are diploid, and are used a few passages after the isolation of mixed populations of embryonic cells which are at various stages of differentiation. As such, SHE cells have a limited life-span in culture, and rarely become tumorigenic, unless exposed to a carcinogen.

Transformation assays with immortalized cell lines

Balb/c 3T3 (25) and C3H/1OT 1/2 cells are used in two rodent cell transformation assays involving immortalized mouse fibroblast cell lines, in which the number of foci in a monolayer are scored (26). These cell transformation systems are based on the use of established immortal cell lines which have an aneuploid karyotype.

Cells are exposed to a carcinogen in culture, and then assessed for their ability to form foci of morphologically transformed cells against a confluent monolayer of cells. A positive result is where there is a significant increase in the number of morphologically transformed foci. Different types of foci have been recognised according to their morphological appearance and staining characteristics under the microscope (11).

Cell transformation can be assayed simultaneously in the same cultures with other endpoints; for example, gene mutation, sister-chromatid exchange formation, chromosomal damage, micronucleus induction and aneuploidy (11, 27, 28). This approach can be useful in providing additional information that can be useful in interpreting transformation data; for example, in deciding whether the transforming activity of a chemical is due to a genotoxic or a non-genotoxic mechanism.

Other potential mechanisms involved in the production of transformed cells in vitro, apart from mutation, include recombination aneuploidy, selection (for example, of preexisting genetically altered cells such as H19-deficient SHE cells), and modification of gene expression (for example, over-induction of ornithine decarboxylase in SHE cells). In the C3H/1OT 1/2 cell transformation system, some 300 potential changes in gene expression have been detected with differential display techniques. Of these, it is thought that 8-10 changes represent key early events which are crucial for transformation to occur and proceed to completion (29, and J. Landolph, personal communication). These events need to be identified and characterised further.

The ultimate verification of a fully transformed phenotype can be achieved by inoculating transformed cells into susceptible host animals and noting any tumorigenesis. Ideally, this process should be undertaken whenever a new agent is shown to induce cell transformation. This is rarely done on a routine basis, due to the cost of the tumorigenicity assays, and is considered unnecessary when using a recognised assay. The fact that this can be achieved in principle with the cells being used, is taken as sufficient verification of the relevance of the morphologically transformed phenotype scored.

Protocols for Cell Transformation Assays

Early studies showed that cell proliferation and transformation frequencies are influenced by several different protocol parameters, including pH, nature of serum, cell culture method, cell passage number, cell seeding density, and the mode of isolation and handling of cells, especially in the case of the SHE cell assay (11, 23, 30). Other criteria for consideration in optimising protocols include the treatment period, the cell division time required for fixation and expression of the transformation event, the phenotypic properties of transformed colonies, the in vivo tumorigenicity of transformed cells, and the responsiveness of the cells to known carcinogens and non-carcinogens. From this work, some recommended protocols for these rodent cell-based transformation assays, which have been in widespread use, were presented by Kakunaga & Yamasaki (31) and Dunkel et al. (32).

More recently, there have been further attempts to define optimised protocols for the assays, particularly the Balb/c 3T3 assay (33-36), and the SHE cell assay (37, 38).

The Balb/c 3T3 assay

The history of the development of the Balb/c 3T3 assay has been reviewed by Schechtman (30). Several sub-clones of the original clone A31 have been generated, and some of these were compared for their suitability for cell transformation. One such sub-clone, A31-714, was identified as possessing the desired properties, and was used extensively to develop an optimised protocol for the basic transformation assay. The use of this clone has been superseded, and the clones most extensively used are clone A31-1 and clone A31-1-13. These clones differ substantially, however, both in their capacities to endogenously metabolise test compounds, and in their spontaneous transformation frequencies. It was shown that clones exhibiting low frequencies appeared to exhibit reduced sensitivity to certain chemical carcinogens. In fact, it has been suggested that clone A31-113 is promotion-proficient; its high spontaneous cell transformation frequency could reflect this (39).

Major protocol modifications to the standard transformation assay with Balb/c 3T3 cells have included changes to exposure and post-exposure conditions, for highly cytotoxic and short-lived chemicals, and the addition of a variety of forms of endogenous metabolic activation. An amplification transformation (Level II) assay has also been developed, in an effort to increase the responsiveness of the system to a broader range of chemical carcinogens. Such studies parallel the modifications which also resulted in a Level II assay for the C3H/10T 1/2 transformation system. The principal change to the standard assay involved harvesting and re-plating the indicator cells at confluence. This Level II assay was proposed as a reliable and sensitive cell transformation method with Balb/c 3T3 cells, despite concerns over the random nature of the appearance of transformed foci (30).

Notwithstanding the implementation of the above modifications to the protocol, there are still a number of drawbacks with these assays, notably:

  1. some 4-6 weeks of cell culture are required for the development of unequivocally transformed foci,
  2. the frequency of transformation is relatively low; and
  3. the assay is labour-intensive.

A new protocol has been developed to address some of these problems, and especially to shorten the length of the assay (36). This protocol involves clone A31-1-1 cells and a modified medium, in conjunction with a new treatment regimen. This resulted in a substantial improvement in transformation frequency the formation of clearly transformed foci at an earlier stage of incubation, and enhanced responsiveness of the assay to several carcinogens, as well as a new protocol suitable for investigating promoting activity (40).

This new protocol is currently the subject of an extensive validation study being conducted in Japan. The results of the preliminary validation stage are encouraging, and confirm the increased sensitivity of the protocol, its suitability for detecting both initiators and promoters, with an overall reduction in cultivation time from 5 weeks to 3 weeks, and low levels of interlaboratory and intralaboratory variability in results (N. Tanaka, personal communication). A second collaborative trial is under way, involving the use of more initiating and promoting agents than in the first trial, together with positive and negative control chemicals. It should be emphasised that the main rationale for developing this new protocol is not only to shorten the length of the assay, but also to increase the transformation frequency, and to permit the use of a decreased concentration of fetal calf serum in the culture medium.

The C3H/1OT 1/2 assay

The history behind the development of the C3H/1OT 1/2 assay was presented and reviewed by Landolph (11). This assay involves the use of contact-inhibited fibroblasts with a low saturation density and an extremely low spontaneous transformation frequency. The standard assay takes 6 weeks to perform, and the results obtained are dependent on the use of suitable batches of fetal calf serum. The scoring and recognition of the various types of transformed foci require training by personnel skilled in the use of the assay, and various suggestions for improving the scoring of transformed foci have been presented (11). Such studies have shown that protocols for this assay system result in the susceptibility of the cells to a wide variety of chemical carcinogens. The cells should first be synchronized to detect short-lived alkylating agents, and then treated during the S-phase or the GAS boundary phase of the cell cycle. A protocol for the assay has also been developed for detecting both initiating and promoting events (11).

There have been attempts to increase the sensitivity of the assay by treating synchronous cells 5 days after seeding, or seeding increased numbers of cells. With regard to the use of exogenous activation systems, the rat liver S9 fraction has been shown to enhance the yield of colonies of morphologically transformed C3H/1OTt/2 cells exposed to aflatoxin B1 and cyclophosphamide (41). Also, the length of exposure to test chemical can be increased by employing two sequential 3-day exposures. In the Level II assay referred to earlier, 10,000 cells are seeded, and then re-seeded after 4 weeks, by which time they have become confluent (42). More research into this issue needs to be conducted. Nevertheless, the C3H/1OT 1/2 assay still requires further development and independent evaluation. It should be noted that, due to the strong dependence of transformation frequency on surviving cell number, transformation frequency is not calculated in this assay, but rather transformation yield is expressed as numbers of type II and type III foci per number of dishes treated, the number of type III foci per total dishes treated the number of dishes with type III foci, and the yield of dishes with type II or type III foci (11, 43, 44).

The SHE cell assay

The protocol for the SHE cell assay has undergone several modifications since it was developed by Berwald & Sachs (5), the most extensive of which concern the recent changes introduced by LeBoeuf et al. (12). Early versions of the assay included the use of an S9 metabolizing system for the identification of procarcinogens (45).

The standard, historical assay involves 7-day treatment of the cells at clonal density with test chemical in Dulbecco's modified Eagle's medium, at pH 7.1-7.3 arid containing fetal bovine serum. Colonies are then fixed, stained and scored for morphological transformation. Conducted according to historical procedures, the SHE cell assay has several characteristics which have discouraged its use in routine carcinogen screening, including:

  1. low frequencies of morphological transformation following exposure to carcinogens, necessitating the scoring of large numbers of colonies to allow statistical procedures to be applied;
  2. wide discrepancies in the ability of different cell isolates and serum lots to support chemically induced morphological transformation; and
  3. difficulties with scoring and identifying colonies of transformed cells.

However, these difficulties have not been experienced equally in all laboratories conducting the assay, and are dependent on modifications of test compound treatment protocol, medium pH, serum and cell source.

Several modifications to the standard approach have been developed:

  1. 48-hour exposure to test chemical in mass culture, followed by re-plating of cells to allow clonal expression in medium free of the test chemical;
  2. 24-hour exposure to one chemical, followed by exposure to a second chemical for the remainder of the expression period, as an assay for tumour promoters;
  3. 24-hour exposure, followed by a chemical-free expression period to assess the stability of the transformed phenotype;
  4. renewal of medium containing test chemical 2 days prior to fixation of colonies to remove transformation-inhibiting factors; and
  5. renewal of serum factors for expression of transformed phenotype.

Some of these modifications are based on the observation that expression of morphologically transformed SHE cell colonies depends on factors in serum, as well as on factors produced by the cells (46).

The use of a reduced pH assay, involving Dulbecco's modified Eagle's medium at pH 6.7 for cell culture, originally developed to enhance clonal SHE cell growth, offers substantial benefits (12, 38). For example, the use of the reduced pH protocol has been shown to result in a 5-10-fold increase in transformation frequencies, a decrease in susceptibility to fluctuations in serum quality, and less ambiguity in the scoring of the transformed phenotype. There are likely to be several mechanisms responsible for the enhancing effects of reduced pH on the performance of the assay (20). These include changes in cell physiology, which result in an increase in clonal cell growth, an extended cell life-span, and maintenance of a higher proportion of cells in the cultures which are transformable, compared to cultures under higher pH conditions (20). These properties of low pH cultures are thought to be due to a retardation of cellular differentiation in the "susceptible" cells (R. LeBoeuf, personal communication).

Relative performance of the assays for predicting carcinogenicity

There have been several collaborative trials to investigate and compare the performance of rodent cell transformation assays for predicting carcinogenicity (12-14, 33-35 47-49). Consideration of the results obtained is complicated because a range of protocols and overall designs have been used in these studies. The results have revealed problems in interlaboratory and intralaboratory reproducibility, which were not necessarily related to lack of protocol standardization (14). Predictive performance varied according to the assay and the laboratory conducting the test, with several carcinogens being negative in some of the transformation assays, and some non-carcinogens inducing cell transformation.

In their literature survey of data on the C3H/1OT 1/2 and Balb/c 3T3 assays, Kuroki & Sasaki (47) concluded that, of the 104 chemicals tested variously in both assays, the sensitivities recorded for the C3H/1OT 1/2 and Balb/c 3T3 systems were reasonably high, being 81Clc and 68%, respectively. The respective specificities for the two assays were 60% and 67%. A reasonably good concordance between transformation and in vivo carcinogenicity was obtained for polycyclic aromatic hydrocarbons, and the Balb/c 3T3 assay appeared to be responsive to some metal carcinogens (50-53) and aromatic amines. Since the publication of this study, the C3H/1OT 1/2 assay has been shown to be responsive to carcinogenic arsenic, nickel and chromium compounds (11, 16). Nevertheless, the standard protocol for this assay system was relatively insensitive to a number of alkylating agents. However, the use of synchronised cultures, or exposing the cells 5 days after seeding made C3H/1OT 1/2 cells sensitive to alkylating agents, as shown in several laboratories (11). Also, it is noteworthy that some so-called non-genotoxic carcinogens, including arsenic, asbestos, benzene and diethylstilboestrol, were detected by rodent cell transformation systems, especially by the SHE cell assay (14, 27).

Data are now available for 472 chemicals which have been tested in the standard (high pH) SHE cell assay protocol (49). Rodent bioassay data are available for 213 of these chemicals (177 carcinogens, and 36 non-carcinogens). The overall concordance for transformation and rodent bioassay data was 80%, with a sensitivity of 82%, and a specificity of 69%. A high degree of interlaboratory agreement was recorded (49).

Analysis of a database comprising 48 chemicals tested in the reduced pH protocol assay showed 85% concordance between transformation and in vivo rodent carcinogenicity, with values for sensitivity and specificity of 87% and 83%, respectively (10). Since these reviews were published, more data have become available under the US National Toxicology Program (NTP) collaborative study and additional studies, to enable an analysis of data on a total of 75 chemicals tested at pH 6.7 to be undertaken (12, with extra information added by using unpublished data [R. LeBoeuf, personal communication]). The results reveal an overall concordance of 83%, a sensitivity of 83% and a specificity of 82% (Table I). This performance was greatly superior to the performance of the Salmonella assay with the same set of chemicals (12).

Table I:Summary of Validation Studies for Rodent Cell Transformation Assays and Predictivity of Rodent Carcinogenicity, Using the Syrian Hamster Embryo (SHE) Cell Assay at pH 6.7 (Low pH) and the Balb/c 3T3 Cell Assay

Assay Number of chemicals tested Overall concordance (%) Sensitivity (%) Specificity (%) Positive predictivity (%) Negative (%) predictivity
Overall Results
SHE (low pH)a,b 75 83 83 82 89 79
Balb/c 3T3c 147 71 80 60 70 71
Chemicals mutagenic in Salmonella
SHE (low pH)a,b 27 93 95 88 95 88
Balb/c 3T3d 69 73 94 30 73 70
Chemicals non mutagenic in Salmonella
SHE (low pH)a,b 48 77 75 80 84 70
Balb/c 3T3e 75 69 64 74 66 72

Comparisons valid only when the same chemicals have been tested.

aData from references 10, 38, 84 and R. LeBoeuf (personal communication).

bConcordance between results from the Salmonella assay and the bioassay in rodents for the same data set = 55% (41/75).

cData from reference 55 (Table 6).

dData from reference 55 (Table 9).

eData from reference 55 (Table 10).

One concern with the data sets discussed above is an imbalance between the number of carcinogens and non-carcinogens tested. For example, in the study on 472 chemicals 177 out of 213 chemicals for which rodent bioassay data exist were carcinogens, while only 36 were classified as non-carcinogens. In the data set comprising 70 chemicals, there were 43 carcinogens and 27 non-carcinogens. More non-carcinogens should be evaluated, to increase confidence in the specificity of these assays. It should be noted that data from several different studies have been pooled for comparative purposes.

An analysis of data for 161 chemicals obtained by using an optimised Balb/c 3T3 assay protocol, has been published by Matthews et al. (33-35, 54, 55). In this case, there were 84 carcinogens and 77 non-carcinogens, and the overall concordance was 71%, with a sensitivity of 80% and a specificity of 60C%. Comparison with corresponding Salmonella data revealed that both the SHE cell assay and the Balb/c 3T3 assay can detect non-mutagenic carcinogens, with the SHE cell assay system having a higher sensitivity for this class of chemical and a higher specificity for mutagenic non-carcinogens, when compared with the Balb/c 3T3 assay. However, caution must be exercised in direct comparisons of the methods, as not all of the chemicals tested in the two systems were the same.

It is clear that the databases for the SHE cell assay and the Balb/c 3T3 transformation cell assay are more extensive and more recent than those available for the C3H/1OT 1/2 cell transformation assay. It was noted that, despite problems with these assays, and the fact that the protocols used were not necessarily optimised, the overall ability to identify carcinogens is encouraging. However, it is uncertain how reliable and useful the information obtained from the Balb/c 3T3 system really is, because some of the data were obtained by using clone A31-1-13, which has been judged to be unsuitable for routine use (as noted earlier). Nevertheless, this clone is used in laboratories throughout the world, and it will be interesting to see the data from the on-going Japanese validation study by using an optimised protocol with a different clone.

It is also worth noting that surveys of the published in vitro activities of 31 chemical entities, classified as Group I human carcinogens by the IARC, showed that the use of several different cell transformation assays (in rodent cells and human cells) exhibited a high level of predictivity (56, and R.D. Combes, personal communication). Thus, 81% of the 26 chemicals tested in transformation assays based on rodent cells were positive, 90% of the ten chemicals tested in human cells induced transformation, and 85% of the 27 chemicals tested in animal cells and/or human cells were positive. Of the nine chemicals tested in both types of system, the majority (67%) were positive, none of the chemicals were active in rodent cells and inactive in human cells, or negative in both systems. Only one chemical, bis(chloromethyl) ether, specifically transformed human cells. In addition, cell transformation tests yielded low numbers of either equivocal or inconsistent data. These results are encouraging for those wishing to further develop cell transformation assays, particularly if these methods are to have a role in regulatory testing and in identifying carcinogens.

Human Cell-based Transformation Systems

Clearly, an ideal transformation assay would be one which utilised human cells, since most transformation assays are conducted to determine the risk of human exposure to potential carcinogens. Moreover, it is important that the mechanisms responsible for transformation in an ideal assay are well understood, so that the relevance of the endpoint being measured in the assay to carcinogenesis in humans is clear. Such a transformation system would need to be subjected to a validation study to assess its reliability in various laboratories, and to assess its capacity to discriminate between carcinogens and non-carcinogens. However, the validation of human cell transformation assays will have to be undertaken with reference to data on the restricted number of known definite human carcinogens (for example, as classified by the IARC) and to those animal data which are likely to be of relevance to human hazard.

So far, it has proved to be extremely difficult to develop cell transformation systems by using human cells in culture with tumorigenicity as the ultimate endpoint. The main problem seems to be that, unlike animal cells in culture, cultured human cells do not spontaneously give rise to immortalized cells. Since the carcinogenic process involves the sequential selection of cells with appropriate mutations in oncogenes and/or tumour suppressor genes, non-immortalised cells senesce before they have acquired all the genetic changes necessary for tumorigenicity. Thus, the human cell transformation systems currently available involve the use of genetically altered cell lines, rather than primary cultures, in which the cells have acquired an immortalized phenotype (17, 57).

Two main human cell transformation cell systems have been developed and studied, namely, the HaCaT keratinocyte cell transformation model (17), and the MSU-1 human fibroblast cell transformation model (58, 59).

The HaCaT cell line

The HaCaT cell line was derived by spontaneous immortalization of normal human keratinocytes, most probably due to mutations in the p53 gene, and the consequent loss of senescence genes (60, 61). These cells can be propagated indefinitely, have an aneuploid karyotype with specific chromosomal changes, but have maintained most of the phenotypic properties of normal human keratinocytes (62). The non-tumorigenic phenotype of the cells has been maintained in long-term culture (61). However, HaCaT cells can be transformed to tumorigenic variants, which exhibit both benign and malignant phenotypes, by ras oncogene activation and by several other interventions (63, 64).

Malignant HaCaT-ras clones are characterised by the loss of a copy of chromosome 15 a characteristic also found in cell lines derived from skin carcinomas. Chromosome transfer and transfection of the thrombospondin-1 gene (located on chromosome 15) indicated tumour suppressor gene function, which seems to be associated with inhibition of tumour vascularisation (65). Malignant HaCaT-ras cells also exhibit progressive abrogation of transforming growth factor-β (TGF-β1) and epidermal growth factor (EGF) growth control in vitro, although this was not associated with an increased potential for anchorage-independent growth, or with loss of differentiation (64).

There are no defined in vitro criteria for distinguishing between tumorigenic and non-tumorigenic HaCaT cells. Nevertheless, this immortalized cell line is a convenient model for studying tumour progression by various carcinogenic agents (64). The cells also maintain the expression of epidermal cytochrome P450 isozymes in culture (N.E. Fusenig, personal communication). HaCaT cells can be grown in vitro as organotypic cultures which exhibit reduced differentiation and tissue organization capacities. However, a well-structured and differentiated squamous epithelium develops when the cells are co-cultured with fibroblasts (66). The invasive or non-invasive behaviour of malignant or non-tumorigenic cells is easily identified in a surface transplantation assay. However, this discrimination has not been reproduced in organotypic co-cultures (64).

The MSU-1 cell line

Since it has proved impossible to generate malignantly transformed human fibroblasts from non-immortalised human fibroblasts, Morgan et al. (67) have developed diploid immortalized human fibroblasts to model the process which causes human soft tissue sarcomas. In this model, the v-myc oncogene was transfected into normal human fibroblasts. Clonal cells which express the v-myc protein were then selected. These cells grew normally and then went into crisis and senescence. However, after the senescing cells had detached and floated away, a small number of cells still growing in the flask were found to be diploid, and to have experienced more than 125 divisions after crisis, suggesting that they had become immortalised. These cells (MSU-1.0 cells) were almost certainly clonal in origin, and genetically identical to the parent cells. A rapidly growing variant arose in the population of MSU-1.0 cells, and was used to establish a new cell line (MSU-1.1). MSU-1.1 cells carry two unique marker chromosomes and have 45 chromosomes per cell (indicative of their clonal origin). The MSU-1.1 cells are chromosomally stable over more than 100 population doublings, and have the wild-type p53 gene. Both MSU-1.0 and MSU-1.1 cells are non-tumorigenic in athymic mice.

MSU-1.0 cells have never been transformed to tumorigenicity, although MSU-1.1 cells have been converted to malignant cells by expression of the H-ras oncogene (68). Exposure of MSU-1.1 cells to (+/-)-7β-8a-dihydroxy-9a,10a- epoxy-7,8,9-tetrahydroxybenzo[a]pyrene, closely related to the procarcinogen benzo[a]pyrene, gave a dose-related increase in transformed foci (69). About half of the cell lines that were developed from the cells forming foci induced sarcomas when injected into athymic mice. This is the first example of carcinogen-induced malignant transformation of human cells in culture. Exposing MSU-1.1 cells to ionising radiation also resulted in transformed foci (70). A more definitive study by O'Reilly et al. (59) presented data which demonstrated that focus formation was a linear function of ionising radiation dose. Moreover, it was shown that the loss of p53 gene expression was associated with malignant transformation, although this was not sufficient to cause tumorigenicity. All the focus-derived tumorigenic cell lines exhibited loss of wild-type p53 gene expression, but some non-tumorigenic cell lines had also lost wild-type p53 gene expression. All tumorigenic cell lines derived from the MSU-1.1 cells exhibited a loss (or at least a marked reduction) of the expression of the p16 protein (71).

Other studies still in progress have demonstrated that MSU-1.1 cells exhibit a dose-dependent increase in transformed foci, when exposed to ultraviolet radiation or to either of two chemical carcinogens which do not require metabolic activation (methylnitrosourea and ethylnitrosourea). In each case, some of the focus-derived cell lines were tumorigenic, as previously noted (J. McCormick, personal communication).

It should be possible to adapt this cell transformation assay to be responsive to pro-carcinogens, by co-culturing the cells with metabolically active feeder cells, as demonstrated by Aust et al. (72). Aust et al. found that cytotoxicity and mutagenicity were caused by benzo[a]pyrene in human diploid fibroblasts, after co-culturing these cells with human kidney carcinoma cells. It has also been demonstrated that, when human diploid fibroblasts are incubated with pro-carcinogens and an appropriate S9 fraction, cytotoxicity and mutagenicity result (J. McCormick, personal communication). Thus, this should prove to be a useful alternative method of carcinogen activation.

Status of human cell systems as transformation assays

The work described with the HaCaT and MSU-1.1 human cell lines is encouraging. The MSU-1.1 cells form foci after carcinogen treatment in a manner similar to that described for the rodent cell transformation assays. This human cell-based system has the potential to be a reliable human cell transformation assay. A test of the ability of different laboratories to obtain similar results with this system, by using standardised protocols with a suitable set of carcinogenic and non-carcinogenic chemicals, should be undertaken. In the meantime, other human cell-based transformation systems should be developed to facilitate the investigation of species-specific, tissue-specific and mechanism-specific aspects of carcinogenesis.

Additional Points

An exogenous metabolising system is used to detect chemicals requiring activation, particularly with cell line assays (11). There is evidence that the cell types employed in transformation assays possess the capacity to metabolically activate several classes of procarcinogens. Thus, C3H/1OT 1/2 cells can activate polycyclic aromatic hydrocarbons, and aflatoxin B1 to some extent. Also, several carcinogens which require metabolic activation are positive in the SHE cell transformation assay, without the addition of exogenous metabolic activation systems. Indeed, SHE cells have a functional and inducible cytochrome P450 system, which can still cause metabolic activation, while being different to rat liver in terms of the various isozyme profiles (73). Nevertheless, more research is required on the necessity of using exogenous enzyme sources, since the standard procedures for adding such systems (for example, post-mitochondrial supernatant, microsomes and intact hepatocytes) to genotoxicity assays do not appear to have had the same success when used in cell transformation systems. This could be due to the toxicity of some preparations of aroclor-induced S9 preparations (24, 41, 48).

Current and Potential Uses of Cell Transformation

Non-regulatory use

Cell transformation assays have long been used in mechanistic studies on carcinogenesis and to elucidate the possible mechanisms of action of known or suspected carcinogens. Those assays which permit the processes of initiation and promotion to be investigated have been more widely used. Transformation assays have also been widely deployed in early decision-making for prioritizing chemicals to be tested in animals for carcinogenicity. A further reason for the early use of cell transformation assays is to screen chemicals for potential chemo-preventive activities (74).

For carcinogen-screening purposes, the ideal test needs to be relatively inexpensive and quick to perform, preferably with the results being available within days, or after a few weeks at most. The screen should require only small amounts of test chemical and, very importantly, should have a low probability of generating false negative data. This latter property is necessary, since positive data can be further evaluated in subsequent tests, whereas a false negative result could mean that a carcinogen is not detected which could lead to a potential human health risk, or a waste of resources spent on further compound development. As indicated above, database analysis suggests that only the SHE cell transformation assay has been shown to fulfil most of these criteria, especially the latter requirement of low false negative predictions.

The SHE cell transformation assay, conducted under commercial conditions according to a Good Laboratory Practice protocol involving initial toxicity determination, dose-selection and exposure for 24 hours and 7 days to detect short-term and longer-term effects, takes 3-5 months to perform. The cost per chemical tested under contract is approximately US$40,000. This figure is comparable to the amount required for a small battery of conventional genotoxicity tests. The SHE cell transformation assay is considerably cheaper and faster than a carcinogenicity bioassay and any of the new transgenic rodent assays. The duration of the SHE assay is longer than the ideal time for a screen, but it might be feasible to reduce this time by automating the assay, for example, by using image analysis techniques (75). A computer-controlled system is available for scoring transformed colonies on the basis of colour, density and morphological features, such as area, texture and pattern (D. Kirkland, personal communication). This program can be used to train investigators, and its implementation reduces subjectivity in scoring transformants.

The regulatory use of cell transformation assays

Cell transformation as an endpoint appears in some, but not all, published testing schemes and strategies for the mutagenicity/carcinogenicity evaluation of chemicals (76). However, such tests are recommended only as supplements to core genotoxicity tests for carcinogens. This is mainly because the molecular basis of the transformed phenotype and its relationship to cancer in vivo is not fully understood. Such views have arisen mainly because of a lack of understanding of the underlying mechanisms involved in transformation, the alleged subjectivity of the in vitro endpoints scored, the lack of regulatory test data on cell transformation, and the absence of adequate formal validation studies.

The above limited use of cell transformation assays for regulatory toxicity purposes is reflected in various regulatory and non-regulatory guidelines for carcinogenicity testing. Thus, for human pharmaceuticals, the positions of the International Conference on Harmonisation(ICH; 77), the US Food and Drug Administration (FDA) and the European Union (EU), as stated in the Federal Register (78) and by the ICH (79), are that "data from in vitro assays, such as cell transformation, can be useful at the compound selection stage" (i.e. during the prioritization of chemicals for animal testing). The FDA also recognises that additional, scientifically justifiable, short-term in vitro tests can be required for the full evaluation of genetic toxicity, and it has recommended the use of cell transformation assays on a case-by-case basis in a draft report (80). The use of the Balb/c 3T3 assay was suggested, since data on more than 200 chemicals tested under identical conditions were available when these draft guidelines were being prepared, and the assay can detect non-mutagenic carcinogens. This latter point is compatible with the recommendations of a multi-disciplinary European Commission study group, which concluded that certain cell transformation assays have a potentially important role in the detection of non-genotoxic carcinogens (81). Thus, the application of these cell transformation assays seems to be beneficial for carcinogen detection, where completely negative genotoxicity test data have been obtained (14).

Cell transformation data could also be of value in a weight-of-evidence approach, when the only positive result was in a genotoxicity test with lower specificity, such as the mouse lymphoma assay.

The FDA not only accepts cell transformation data for regulatory purposes, but also requests that such information on new pharmaceuticals is made available for review. The resulting data could be useful for mechanistic studies, and negative data might also be used in an overall hazard assessment, to permit the limited use of a drug, depending on its nature and other toxicity information. Industry is using cell transformation assays more widely for these purposes, with the SHE cell assay now being the most frequently employed transformation system Thus, some AIDS drugs have been marketed for therapeutic use on the basis of negative cell transformation results, before the outcome of carcinogenicity bioassays has become available. Although it is considered that cell transformation assays cannot yet be used to replace the rodent bioassay, it should be noted that cell transformation is an integral part of US regulatory guidelines for medical device testing.

Information on hazard data submitted to the German Federal Institute for Drugs and Medical Devices for new chemical entities related to new pharmaceuticals over the period 1990-1997, concerning 335 active principles, has been analysed. Of these chemicals, 32, i.e. approximately 10%, have been tested in an in vitro cell transformation assay (L. Muller, personal communication). Reasons given for undertaking cell transformation for these chemicals include:

  1. a need for further knowledge, in addition to a positive genotoxicity test battery, in the absence of cancer bioassay data (for some nucleoside analogues, an alkylating agent and a growth factor);
  2. a need for further information to supplement a negative genotoxicity test battery (for chemicals which traditionally are not tested in the cancer bioassay); and
  3. a need for interim data before the final outcome of an on-going cancer bioassay.

Of the 32 chemicals tested for cell transformation, nine were positive. The database consisted of 35 tests for the 32 chemicals tested, comprising 21 Balb/c 3T3 assays, nine SHE cell assays, four tests with the C3H/10T 1/2 assay, and one test with the baby hamster kidney assay. The positive data obtained for three nucleoside analogues, an alkylating agent and a growth factor are compatible with basic knowledge on the potential mechanisms of cell transformation involved in the assays used. However, the evidence for the cell transforming activity of the test compound acting as an inhibitor of microtubule polymerization might have been unexpected.

The cell transformation data obtained for these chemicals did not conflict with the outcome of genotoxicity testing (82). Comparison of the cell transformation, genotoxicity and cancer bioassay data for the above chemicals is complicated by lack of information. Thus, for almost half (15 out of 32 chemicals) of the chemicals analysed, no carcinogenicity bioassay data were available. In the case of the microtubule inhibitor, which was positive in the Balb/c 3T3 assay, two life-time cancer bioassays were negative. Two of the three antiviral nucleoside analogues tested which were also positive in the Balb/c 3T3 assay, were non-carcinogenic in vivo, and one was carcinogenic. Those chemicals in the database which were negative for both genotoxicity and cell transformation were variously carcinogenic or non-carcinogenic in bioassays. Lastly, for three chemicals, the conduct of a cell transformation assay was considered an essential element in the toxicity testing programme for eventual registration. The application of a cell transformation assay was used to argue against the need to conduct a bioassay.

Where there are no conflicting data, cell transformation results can be used to corroborate genotoxicity test information, in a weight-of-evidence approach. Therefore, it is important to consider the results of cell transformation assays in combination with other data in assessing the possible carcinogenicity of test chemicals. For example, a positive result in the SHE cell transformation assay, in combination with a positive result in the Salmonella assay, should be considered to be strong evidence of rodent carcinogenicity for a chemical. Similarly analyses of the available databases suggest that, where there is no evidence for mutagenicity or activity in the SHE cell assay, and where no structural alerts are apparent for the test chemical, the probability of rodent carcinogenicity would be about 10% for data based on the SHE cell assay, and 20% when Balb/c 3T3 assay data were used (12).

In the face of such evidence, and earlier information showing high predictive performance for carcinogenicity, it is justifiable to question the basis of the low priority given by regulatory bodies to cell transformation assays, in comparison with the high reliance placed on genotoxicity testing. Moreover, as some transformation systems are responsive to non-genotoxic carcinogens, a non-mutagenic chemical inducing cell transformation should be considered as having the potential to induce tumours in rodents.

As noted above, one major reason why the policy for using cell transformation data has not been widely adopted relates to the relative lack of understanding of the mechanistic basis for in vitro cell transformation, compared with knowledge of the mechanisms underlying standard genotoxicity tests. For example, the ability of certain chemicals, such as some steroids, oestrogens and peroxisome proliferators, to induce cell transformation has so far gone unexplained. Moreover, tumour formation in an experimental animal is considered by some to be a biologically more plausible and more relevant endpoint than cell transformation for predicting carcinogenic hazard in humans. There are several reasons why this view might not be justified, however, particularly if in vitro data could be supplemented with information from animal studies of intermediate duration, and from absorption, distribution, metabolism and excretion (ADME) investigations, coupled with pharmacokinetic modelling (83). Furthermore, the ability to use human cell-based transformation assays, once they become available, would reduce uncertainties over data extrapolation due to species differences.

There is no doubt that there will need to be a substantial change in the attitudes of regulators toward in vitro cell transformation assays, if these methods are to assume greater importance in regulatory testing schemes for carcinogens. This is because the main endpoint currently used in traditional rodent bioassays is a positive diagnosis of a benign or malignant tumour. The important question to address is the extent to which information from cell transformation assays can be used to reduce the uncertainty that human exposure to a chemical under a given set of conditions will pose a carcinogenic risk. It is recognised that this risk will vary according to the chemical and to the conditions concerned. Therefore, it is recommended that there should be a case-by-case approach to the regulatory use of cell transformation data, rather than a single strategy, such as one which limits this use to compound selection.

Other potential uses

The SHE cell assay could be particularly useful when genotoxicity assays exhibit poor predictability with certain classes of chemicals. An example of such a chemical class concerns the aromatic amines, for which the Salmonella assay has a high false positive rate. In a group of ten Salmonella-positive aromatic amines, the SHE cell assay accurately detected the five chemicals that were carcinogenic in rodents (12, 84).

There are also certain circumstances under which the application of parts of the standard genotoxicity test battery is inappropriate; for example, the testing of antimicrobial drugs and chemicals acting on certain mammalian cell receptors, and where in vivo testing is compromised by lack of compound absorption (85). In such circumstances, one option is to expand mammalian cell testing by using other genotoxicity assays. However, the use of a cell transformation assay could also provide useful data. Such information might be useful in overcoming known problems where high incidences of positive data, that might not be biologically relevant, are generated by tests such as the mouse Iymphoma and chromosomal aberration tests.

Cell transformation assays might also have a potentially important role in providing useful hazard data to supplement the information derived from conventional genotoxicity test batteries, in situations where in vivo testing is difficult or impossible. Such situations include the testing of cosmetic ingredients and tobacco products, which is prohibited by law in some countries. Situations can also be envisaged in which the application of cell transformation tests, such as the SHE cell assay, might be preferable on scientific grounds; for example, photocarcinogenicity testing, where the relevance of current in vivo animal testing methods, and the interpretation of the data obtained for human hazard, are obscure. Moreover, the use of a photochemical cell transformation assay to complement photochemical mutagenicity testing, which is now required by some regulatory agencies, would be much more desirable from an animal welfare point of view. This is because photocarcinogenicity testing usually involves the use of severe procedures, in which animals are exposed to tumour-inducing doses of ultraviolet radiation.

Prospects for the wider utilization of cell transformation assays

The wider use of cell transformation assays has also been limited because some laboratories have experienced variable success and logistical problems with the available protocols. The isolation and use of primary cell cultures in a standardized way is more difficult to achieve and define than is the use of cell lines. It is for this reason that some investigators have encountered problems with the isolation of transformable batches of SHE cells for use in the high pH assay (86). There have also been problems with scoring endpoints, and uncertainties over the relevance of the development and behaviour of transformed cells in vitro to the process of neoplasia and tumorigenesis in vivo.

However, it should be remembered that not all in vitro assays will necessarily need to be simple and straightforward and readily applicable in all laboratories. It is more important that the necessary expertise and training are acquired by a sufficient number of laboratories to ensure that these assays can be available whenever they are needed. If this is recognized, there should be no problems and no major barriers to the interlaboratory transfer of a method. However, such an assertion needs to be agreed by all concerned.

In the past, small protocol changes have been shown to affect the response of SHE cells to carcinogens, although this problem has been alleviated by the introduction of methodological improvements, such as the low pH assay. Nevertheless, there is a general need to develop protocol modifications for all the cell transformation assays, so that their performance is less influenced by changes in serum, cell source, and other variables such as excessive selection pressure on cells in culture, due, for example, to high levels of cytotoxicity.

A further, very important, limitation of cell transformation assays concerns a lack of optimised and standardized protocols, particularly with regard to the use of exogenous metabolizing systems for detecting pro-carcinogens. A thorough investigation into this requirement is needed. There have also been several reports of variations among laboratories in the way cells grow, and in the appearance of normal and morphologically transformed colonies, which can lead to subjectivity in scoring an endpoint. Once again, appropriate training of the personnel involved in conducting the assays, possibly together with the implementation of automated techniques, should minimise variation in data, as discussed earlier.

Suggestions for advancing the use of cell transformation

There is still considerable and widespread scepticism about the increased regulatory use of cell transformation assays for detecting carcinogens for two principal reasons:

  1. the genotoxicity tests currently employed are judged by many to be sufficient for detecting carcinogens which pose a risk to humans; and
  2. the protocols for the conduct of the assays are technically demanding, and not widely standardized, making their broad implementation difficult at this time.

Evidence was presented during the workshop to show that, at least in the case of one rodent cell-based system, cell transformation can be highly predictive of carcinogenicity. This predictivity exceeds that provided by the standard battery of genotoxicity tests.

In the case of the rodent cell transformation systems discussed, immediate priority should be given to the organization of a prevalidation study of assay protocols which meet internationally agreed criteria for test development (87). This will permit the definition of performance criteria to ensure that a valid test system is available for each assay, before it is broadly used. Human cell-based transformation systems require further development, in the expectation that, in the longer term, they will provide the most relevant and most reliable approach to the identification of human carcinogens.

Conclusions and Recommendations

General

  1. Several lines of evidence indicate that in vitro cell transformation assays are relevant for detecting animal carcinogens:
    1. the parallel nature of the molecular and cellular properties induced by carcinogens, and the behaviour in tissue culture of transformed cells and of cells derived from naturally occurring or carcinogen-induced tumours;
    2. the fact that a significant proportion of physical and chemical agents which cause tumours in whole animals, also induce cell transformation in fibroblasts in vitro; and
    3. the observation that progeny cells from a morphologically transformed colony or focus (both endpoints scored in transformation assays) have a significant probability of inducing tumours when they are injected into susceptible host animals.
  2. The databases on the three principal rodent cell transformation assays (those employing primary SHE cells, and two mouse cell line systems, Balb/c 3T3 and C3H/lOT 1/2) contain different amounts of information. Nevertheless, it can be concluded that positive rodent cell transformation assay data should, in general, be considered to be indicative of a high probability of rodent carcinogenicity, while negative results are indicative of non-carcinogenicity.
  3. Cell transformation assays might also have a role to play in providing additional useful information on chemicals for which the biological significance of the bioassay results are uncertain (for example, where tumorigenic effects are seen only at very high dose levels). Cell transformation data can also be useful where genotoxicity data for a particular chemical class are known to have limited predictive value (for example, aromatic amines in the Salmonella assay), and to help clarify the meaning of positive results from genotoxicity assays, when the relevance of the data to carcinogenic activity is uncertain.
  4. A study of the process of cell transformation can provide information on the molecular mechanisms involved in carcinogenesis, and the phenotypic properties conferred on cells by activation of oncogenes or loss of tumour suppressor gene activity. Cell transformation assays can also be used to investigate the activity of chemicals with chemopreventive activity.
  5. Cell transformation assays could provide information which, in combination with data from other testing methods, could be useful for identifying the carcinogenic potential of physical and chemical agents in humans.

State of development of cell transformation assays and potential roles of current assays

  1. There is evidence to suggest that a better prediction of potential human or rodent carcinogenicity could be obtained by using cell transformation assays in conjunction with selected genotoxicity tests, rather than by using genotoxicity tests alone. Cell transformation assays can identify carcinogens which are not typically detected by conventional genotoxicity assays, while maintaining or improving the accurate identification of non-carcinogens. Cell transformation assays have the potential to detect various types of carcinogens, including those that are thought to act via genotoxic and non-genotoxic mechanisms.
  2. A more extensive database on the use of cell transformation assays for screening purposes should be set up, alongside the standard genotoxicity assays (for comparative purposes), by using chemicals with known activities in rodent bioassays. In the longer term, such information should be used to add at least one of the established rodent cell transformation assays (SHE, Balb/c 3T3 or C3H/10T 1/2) to standard carcinogenicity screening packages, at the very least as a secondary screen. Such an assay might eventually replace one of the primary genotoxicity tests (for example, the mouse Iymphoma assay or detection of chromosomal aberrations), since colony transformation or focus formation can arise as a result of exposure of cells to genotoxic or non-genotoxic agents.
  3. The inclusion of cell transformation assays in a photocarcinogenicity screen would be worthwhile. This could be pursued in conjunction with current activities in developing in vitro tests for photochemical toxicity and photochemical genotoxicity.

Prevalidation and validation status of cell transformation assays

  1. Optimal testing strategies, combining both genotoxicity and transformation endpoints, should be developed, in conjunction with optimising study protocols and improving the transferability of the technology (for example, through the provision of a common repository of recommended cells and cell lines for transformation assays), and improving the quality and coverage of the databases available for assessing the relative predictive performances of cell transformation assays for predicting carcinogenicity.
  2. ECVAM should undertake a comprehensive and up-to-date review of the literature on cell transformation, to provide a basis for comparison of the various cell transformation systems available, with respect to their strengths, deficiencies, status of the databases, and predictive performances for rodent carcinogenicity. The results of this review should be used to produce a summary of the current status of the development of the various assays, and their readiness for prevalidation and validation.
  3. The outcome of this review should also be used to consider the need to organise a focused interlaboratory study involving one or more of the rodent cell-based transformation assays, once they are considered to be ready according to the ECVAM criteria for test development. The objective of the study will be to determine the performance of the assays for their ability to predict the carcinogenicity of a carefully selected set of about 20 chemicals, and to systematically evaluate their interlaboratory transferability and reproducibility. The test chemical set should include genotoxic carcinogens, non-genotoxic carcinogens and non-carcinogens. It could be based on the collection of chemicals being used in the current validation of transgenic rodent models for carcinogenicity, organised by the International Life Science Institute.
  4. Although the currently available rodent cell transformation assays can provide useful information, further improvements to all these assays are required to optimise their performance and reliability, in terms of consistency of response in the same and in different laboratories, for detecting carcinogens of relevance to human health. Any further development of the Balb/c 3T3 assay should, however, await the outcome of the current Japanese validation study. A thorough investigation is needed to ascertain the necessity of using exogenous metabolic activation systems in the various cell transformation assays.
  5. The suitability of the currently available rodent cell transformation assay protocols for independently managed interlaboratory prevalidation studies should be established by ECVAM as a matter of urgency, in the light of information arising from this workshop, the results of on-going studies, and the outcome of a further detailed analysis of the published literature. Human cell-based systems
  6. Recent successes in reproducibly transforming human cells suggest that a reliable human cell transformation assay system will eventually be developed. However, a comprehensive review of the status of all the currently available human cell transformation systems should be commissioned by ECVAM.
  7. There is a need to identify the mechanistic basis for the higher resistance of human cells, comparison to rodent cells, to undergo spontaneous and carcinogen-induced immortalization.
  8. Although the mechanisms responsible for rodent and human carcinogenicity are not yet fully understood, current knowledge suggests that the sequence of processes for the transformation of rodent and human cells are similar, or nearly identical. However, there are species differences between rodents and humans which could be important with regard to cancer development. These include:
    1. the fact that humans are a long-lived species with cancer arising with increasing frequency after 35 years of age, whereas rodents live for only 2-3 years;
    2. species differences in xenobiotic metabolism; and
    3. the much higher frequencies of spontaneous and induced immortalization and transformation of rodent cells in culture, compared with human cells.
  9. The development of human cell transformation assays by using immortalized human cells can be expected to contribute to the understanding of such problems, and to provide information on the molecular mechanisms of cell transformation.
  10. Since no test system with immortalized human cells has yet been standardized further research is required to improve the currently available systems, before they can be considered ready for assessment as potential test methods for detecting carcinogens In the case of the MSU-1.1 cell system, research is required to enable suitable carcinogen metabolising systems to be included in the transformation assay. In addition the robustness of the assay needs to be established through its transfer to other laboratories. Reliable in vitro criteria and easy detection of cell transformation are still needed for the assay. Thus far, HaCaT cells have not been transformed to focus-forming cells or to cells with some unique in vitro characteristic which can be related to tumorigenicity. This problem needs to be overcome, if this cell line is to form the basis of a useful cell transformation assay.
  11. Human cell transformation assays should include the use of endogenous and exogenous rat or human liver S9 preparations, such as are used in genotoxicity assays. An ultimate goal would be to create transformable cell lines expressing human cytochrome P450 isozymes, as is the case with the HaCaT cell line.
  12. More effort should be directed toward the development of cell transformation assays based on human cells, especially epithelial cells, from a variety of tissues, to identify human carcinogens with tissue-specific and/or species-specific effects.
  13. Information obtained from improved human cell-based transformation systems could also be very useful in the development of new chemotherapy treatments for human cancers.

References

  1. Anon. (1994). ECVAM News & Views. ATLA 22: 7-11.
  2. Maronpot R.R. (1991). Chemical carcinogenesis. In Handoook of Toxicologic Pathology, pp. 91-129. New York, USA: Academic Press.
  3. Barrett, J.C. (1993). Mechanisms of multistep carcinogenesis and carcinogen risk assessment. Environmental Health Perspectives 1OO: 9-20.
  4. Combes, R.D. (1997). Detection of non-genotoxic carcinogens: major barriers to replacement of the rodent assay. In Animal Alternatives, Welfare and Ethics (ed. L.F.M. van Zutphen & M. Balls), pp. 627-634. Amsterdam, The Netherlands: Elsevier.
  5. Berwald, Y. & Sachs, L. (1965). In vitro transformation of normal cells to tumour cells by carcinogenic hydrocarbons. Journal of the National Cancer Institute 35: 641-661.
  6. Meyer, A.L. (1983). In vitro transformation assays for chemical carcinogens. Mutation Research 115: 323-338.
  7. Barrett, J.C. & Fletcher, W.F. (1987). Cellular and molecular mechanisms of multi-step carcinogenesis in cell culture models. In Mechanisms of Environmental Carcinogenesis, Vol:2, Multistep Models of Carcinogenesis, pp. 73-116. Boca Raton, FL, USA: CRC Press.
  8. Yuspa,S.H. & Poirier, M.C. (1988). Chemical carcinogenesis: from animal models to molecular models in one decade. Advances in Cancer Research 50: 25-70.
  9. Barrett, J.C., Kakanuga, T., Kuroki, T., Neubert, D., Trosko, J.E., Vasilieu, J.M., Williams, G.M. & Yamasaki, H. (1986). Mammalian cell transformation in culture. In Long and Short Term Test Assays for Carcinogens: A Critical Appraisal. IARC Scientific Publication No. 83 (ed. R. Monsanto H. Bartsch, H. Vainio, J. Wilbourn & H. Yamasaki), pp. 267-286. Lyon, France: IARC.
  10. LeBoeuf, R.A., Kerckaert, G.A., Aardema, M.J., Gibson, D.P., Brauninger, R. & Isfort, R.J. (1996). The pH 6.7 Syrian hamster embryo cell transformation assay for assessing the carcinogenic potential of chemicals. Mutation Research 356: 85-127.
  11. Landolph, J.R. (1985). Chemical transformation in C3H1OT 1/2 CC118 mouse embryo fibroblasts: historical background, assessment of the transformation assay, and evolution and optimization of the transformation assay protocol. In Transformation Assay of Established Cell Lines: Mechanisms and Application, IARC Scientific Publications No. 67 (ed. T. Kakunaga & H. Yamasaki), pp. 185-203. Lyon. France: IARC.
  12. LeBoeuf, R.A., Kerckaert, K.A., Aardema, M.J. & Isfort, R J. (1999). Use of Syrian hamster embryo and Balb/c 3T3 cell transformation for assessing the carcinogenicity potential of chemicals. In The Use of Short- and Medium-term Tests for Carcinogenic Hazard Eualuation. IARC Scientific Publications No. 146 (ed. D.B. McGregor, J.M. Rice & S. Venitt), pp. 409-425. Lyon, France: IARC.
  13. Heidelberger, C., Freeman, A.E., Pienta, R.J., Sivak, A., Bertram, J.S., Casto, B.C., Dunkel, V.C., Francis, M.W., Kakunaga, T., Little, J.B. & Schechtman, L.M. (1983). Cell transformation by chemical agents -- a review and analysis of the literature. A report of the US Environmental Protection Agency Gene-Tox program. Mutation Research 114: 283-385.
  14. Muller, L., Miltenburger, H.G., Marquardt, H. & Madle, S. (1993). The significance of in vitro cell transformation assays for routine testing. In Current Issues in Genetic Toxicology, pp. 102-113. Munich, Germany: Medizin-Verlag.
  15. Sugimura, T., Terada, M. Yokota, J., Hirohashi S. & Wakabayashi, K. (1992). Multiple genetic alterations in human carcinogenesis. Environmental Health Perspectives 98: 5-12.
  16. Landolph, J.R., Dews, P.M., Ozbun, L. & Evans, D.P. (1996). Metal-induced gene expression and neoplastic transformation. In Toxicology of Metals (ed. L. Magos & T. Suzuki), pp. 321-329. Boca Raton, FL, USA: CRC Press.
  17. Fusenig, N.E. & Boukamp, P. (1994). Carcinogenesis studies of human cells: reliable in vitro methods. In Cell Culture in Pharmaceutical Research (ed. N.E. Fusenig & H. Graf), pp. 79-102. Berlin, Germany: Springer.
  18. Swierenga, S.H.H. & Yamasaki, H (1992). Performance of tests for cell transformation and gapjunctional intercellular communication for detecting non-genotoxic carcinogenic activity. In Mechanisms of Carcinogenesis in Risk Identification. IARC Scientific Publications No. 116 (ed. H. Vainio, P.N. McGee, D.B. McGregor & A.J. McMichael), pp. 165-193. Lyon, France: IARC.
  19. Isfort, R.J. & LeBoeuf, R.A. (1995). The Syrian hamster embryo (SHE) cell transformation system: a biologically relevant in vitro model -- with carcinogen predicting capabilities Ñ- of in vivo multistage neoplastic transformation. Critical Reuiews in Oncogenesis 6: 251-260.
  20. Isfort, R.J. & LeBoeuf, R.A. (1996). Application of in vitro cell transformation assays to predict the carcinogenic potential of chemicals. Mutation Research 365: 161-173.
  21. McGregor, D.B., Rice, J.M. & Venitt, S., eds (1999) Consensus report. In The Use of Shortand Medium-term Tests for Carcinogenic Hazard Eualuation. IARC Scientific Publications No. 146, pp. 1-18. Lyons France: IARC.
  22. Di Paolo, J.A. (i980). Quantitative in uitro transformation of Syrian golden hamster embryo cells with the use of frozen stored cells. Journal of the National Cancer Institutes 64, 1485-1489.
  23. Lubet, R.A., Nims, R.W., Kiss, E., Kouri, R.E., Putman, D.L. & Schechtman, L.M. (1986). Influence of various parameters on benzo[a]pyrene enhancement of adenovirus SA7 transformation of Syrian hamster embryo cells. Environmental Mutagenesis 8: 533-542.
  24. Pienta, R.J., Poiley, J.A. & Lebherz, W.B. (1977). Morphological transformation of early passage golden Syrian hamster embryo cells derived from cryopreserved primary cultures as a reliable in vitro assay for identifying diverse carcinogens. International Journal of Cancer 19: 642-655.
  25. Kakunaga, T. (1973). A quantitative system for assay of malignant transformation by chemical carcinogens using a clone derived from Balb 3T3. International Journal of Cancer 12: 463-473.
  26. Reznikoff, C.A., Bertram, J.S., Brankow, D.W. & Heidelberger, C. (1973). Quantitative and qualitative studies on chemical transformation of cloned C3H mouse embryo cells, sensitive to post-confluence inhibition of cell division. Cancer Research 33: 3239-3249.
  27. Tsutsui, T., Hayashi, N., Maizumi, H., Huff, J. & Barrett, J.C. (1997). Benzene-, catechol-, hydroquinone- and phenol-induced cell transformation, gene mutations, chromosome aberrations, aneuploidy, sister-chromatic exchanges and unscheduled DNA synthesis in Syrian hamster embryo cells. Mutation Research 373: 113-123.
  28. Fritzenschaf, H. Kohlpoth, M., Rusche, B. & Schiffmann, D. (1993). Testing of known carcinogens and noncarcinogens in the Syrian hamster embryo (SHE) micronucleus test in vitro: correlations with in vivo micronucleus formation and cell transformation. Mutation Research 319: 47-53.
  29. Shuin, T., Billings, P.C., Lillehaug, J.R., Patierno, S., Roy-Burman, P. & Landolph, J.R. (1986). Enhanced expression of c-myc and decreased expression of c-fos proto-oncogenes in chemically and radiation-transformed C3H1OT1/2 C1 8 mouse embryo cell lines. Cancer Research 46: 5302-5311.
  30. Schechtman, L.M. (1985). Balb/c 3T3 cell transformation: protocols, problems and improvements. In Transformation Assay of Established Cell Lines: Mechanisms and Application. IARC Scientific Publications No. 67 (ed. T. Kakunaga & H. Yamasaki), pp. 165-184. Lyon, France: IARC.
  31. Kakunaga, T. & Yamasaki, H. (1985). Experimental protocols by the working group. In Transformation Assay of Established Cell Lines: Mechanisms and Application. IARC Scientific Publications No. 67 (ed. T. Kakunaga & H. Yamasaki), pp. 207-219. Lyon, France: IARC.
  32. Dunkel, V.C., Rogers, C., Swierenga, S.H.H., Brillinger, R.L., Gilman, J.P.W. & Nestman, E.R. (1991). Recommended protocols based on a survey of current practice in genotoxicity testing laboratories. III. Cell transformation in C3H1OT~/2 mouse embryo cell, Balb/c 3T3 mouse fibroblast and Syrian hamster embryo cell cultures. Mutation Research 246: 285-300.
  33. Matthews, E.J. (1993). Transformation of Balb/c 3T3 cells. I. Investigation of experimental parameters that influence detection of spontaneous transformation. Environmental Health Perspectives 101: Suppl. 2, 277-291.
  34. Matthews, E.J. (1993). Transformation of Balb/c 3T3 cells. II. Investigation of experimental parameters that influence detection of benzo[a]pyrene-induced transformation. Environmental Health Perspectives 101: Suppl. 2, 293-310.
  35. Matthews, E.J. (1993). Transformation of Balb/c 3T3 cells. III. Development of a co-culture clonal survival assay for quantification of chemical cytotoxicity in high-density cell cultures. Environmental Health Perspectives 101: Suppl. 2, 311-318.
  36. Tsuchiya, T. & Umeda, M. (1995). Improvement in the efficiency of the in vitro transformation assay method using Balb/3T3 A31-1-1 cells. Carcinogenesis 8: 1887-1894.
  37. Aardema, M.J., Isfort, R.J., Thompson, E.D. & LeBoeuf, R.A. (1996). The low pH Syrian hamster embryo (SHE) cell transformation assay: a revitalized role in carcinogen prediction. Mutation Research 356: 5-9.
  38. Kerckaert, G.A., Isfort, R.J., Carr, G.J., Aardema M.J. & LeBoeuf, R.A (1996). A comprehensive protocol for conducting the Syrian hamster embryo cell transformation assay at pH 6.70. Mutation Research 356: 65-84.
  39. Yamasaki, H., Enomoto, T., Shiba, Y., Kanno, Y. & Kakunaga, T. (1985). Intercellular communication capacity as a possible determinant of transformation sensitivity of Balb/c 3T3 clonal cells. Cancer Research 45: 637-641.
  40. Tsuchiya, T. & Umeda, M. (1997). Relationship between exposure to TPA and appearance of transformed cells in MNNG-initiated transformation of Balb/c 3T3 cells. International Journal of Cancer 73: 271-276.
  41. Billings, P.C., Heidelberger, C. & Landolph, J.R. (1985). S-9 metabolic activation enhances aflatoxin-mediated transformation of C3H1OT1/2 cells. Toxicology and Applied Pharmacology 77: 58-65.
  42. Schechtman, L.M., Kiss, F., McCarvill, J., Gallagher, M., Kouri R.E. & Lubet, R.A. (1982). A method for amplification of sensitivity of the C3H1OT1/2 cell transformation assay. Proceedings of the American Association for Cancer Research 23: 74.
  43. Miura, T., Patierno, S.R., Sakuramoto, T. & Landolph, J.R. (1989). Morphological and neoplastic transformation of C3H/1OT 1/2 C1 8 mouse embryo cells by insoluble carcinogenic nickel compounds. Environmental and Molecular Mutagenesis 14: 65-78.
  44. Patierno, S.R., Banh, D. & Landolph, J.R. (1988). Transformation of C3H/1OT 1/2 mouse embryo cells to focus formation and anchorage independence by insoluble lead chromate but not soluble calcium chromate: relationship to mutagenesis and internalization of lead chromate particles. Cancer Research 48: 5280-5288.
  45. Pienta, R.J. (1979). A hamster model system for identifying carcinogens. In Carcinogens: Identification and Mechanism of Action (ed. A.C. Griffin & C.R. Shaw), pp. 121-141. New York, USA: Raven Press.
  46. Rivedal, E. & Haddeland, U. (1996). Role of serum in the morphological transformation of Syrian hamster embryo cells. Characterisation and partial purlfication of protein factors in fetal bonne serum. Toxicology in Vitro 10: 217-227.
  47. Kuroki, T. & Sasaki, K. (1985). Relationship between in vitro cell transformation and in vivo carcinogenesis based on available data on the effects of chemicals. In Transformation Assay of Established Cell Lines: Mechanisms and Application. IARC Scientific Publications No. 67 (ed. T. Kakunaga & H. Yamasaki), pp. 93 118. Lyon, France: IARC.
  48. McGregor, D. & Ashby, J. (1985). Summary report on the performance of the cell transformation assays. In Progress in Mutation Research, Vol. 5, Eualuation of Short-term Tests for Carcinogens (ed. J. Ashby, F.J. de Serres, M. Draper M. Ishidate, Jr, B.H. Margolin, B.E. Matter & M.D. Shelby), pp. 103-115. Amsterdam, The Netherlands: Elsevier Science.
  49. Isfort, R.J., Cody, D.B., Kerckaert, G.A., Tycko, B. & LeBoeuf, R.A. (1997). Role of the H19 gene in Syrian hamster embryo cell tumorigenicity. Molecular Carcinogenesis 20: 189-193.
  50. Bertolero F., Pozzi, G., Sabbioni, E. & Saffiotti U. (1987). Cellular uptake and metabolic reduction of pentavalent arsenic as determinants of cytotoxicity and morphological transformation. Carcinogenesis 8: 803-808.
  51. Sabbioni, E., Pozzi, G., Pintar, A., Casella, L. & Garattini, S. (1991). Cellular retention, cytotoxicity and morphological transformation by vanadium (IV) in Balb/3T3 cell line. Carcinogenesis 12: 47-52.
  52. Salbioni, E., Fischbach, M., Pozzi, G., Pietra, R., Gallonni, M. & Piette, J.L. (1991). Cellular retention, toxicity and carcinogenic potential of seafood arsenic. I. Lack of cytotoxicity and transforming activity of arsenobetaine in the Balb/3T3 cell line. Carcinogenesis 12: 1287-1291.
  53. Sabbioni, E. & Balis, M. (1995). Use of cell culture and nuclear and radioanalytical techniques for environmental, occupational and biomedical metal toxicology research at the JRC-Ispra. In Alternative Methods in Toxicology and the Life Sciences, Vol. 11, The World Congress on Alternatives and Animal Use in the Life Sciences: Education, Research, Testing (ed. A.M. Goldberg & L.F.M. van Zutphen), pp. 101-108. New York, USA: Mary Ann Liebert.
  54. Matthews, E.J., Spalding, J.W. & Tennant R.W. (1993). Transformation of BALB/c 3T3 cells. IV. Rank-ordered potency of 24 chemical responses detected in a sensitive new assay procedure. Environmental Health Perspectives 101: 347-482.
  55. Matthews, E.J., Spalding, J.W. & Tennant, R.W. (1993). Transformation of BALB/c 3T3 cells. V. Transformation responses of 168 chemicals compared with mutagenicity in Salmonella and carcinogenicity in rodent bioassays. Environmental Health Perspectives 101: 347-482.
  56. IARC (1987). IARC Monographs on the Eualuation of Carcinogenic Risks to Humans, Suppl. 6. Genetic and Related Effects: An Updating of Selected IARC Monographs from Volumes 1 to 42, 729pp. Lyon, France: IARC.
  57. McCormick, J.J. & Maher, V.M. (1989). Malignant transformation of mammalian cells in culture, including human cells. Environmental and Molecular Mutagenesis 14: 105-113.
  58. McCormick, J.J. & Maher, V.M. (1994). Analysis of the multistep process of carcinogenesis using human fibroblasts. Risk Analysis 14: 257-263.
  59. O'Reilly, S., Walicka, M., Kohler, S.K., Dunstan, R., Maher, V.M. & McCormick, J.J. (1998). Dosedependent transformation of cells of human fibroblast cell strain MSU-1.1 by cobalt-60 gamma radiation and characterization of the transformed cells. Radiation Research 150: 577-584.
  60. Boukamp, P., Dzarlieva-Petrusevska, R.T., Breitkreutz, D., Hornung, J., Markham, A. & Fusenig, N.E (1988). Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line. Journal of Cell Biology 106: 761-771.
  61. Boukamp, P., Popp, S., Altmeyer, S., Hulsen, A., Fasching, C., Cremer, T. & Fusenig, N.E. (1997). Sustained non-tumorigenic phenotype correlates with a largely stable chromosome content during long-term culture of the human keratinocyte cell line HaCaT. Genes Chromosomes and Cancer 19: 201-214.
  62. Breitkreutz, D., Schoop, V.M., Mirancea, N., Baur, M., Stark, Ha. & Fusenig, N.E. (1998). Epidermal differentiation and basement membrane formation by HaCaT cells in surface transplants. European Journal of Cell Biology 75: 273-286.
  63. Boukamp, P., Peter, W., Pascheberg, U., Altmeier, S., Fasching, C., Stanbridge, E.J. & Fusenig, N.E. (1995). Step-wise progression in human skin carcinogenesis in vitro involves mutational inactivation of p53, ras-H oncogene activation and additional chromosome loss. Oncogene 11: 961-969.
  64. Fusenig, N.E. & Boukamp, P. (1998). Multiple stages and genetic alterations in immortalization, malignant transformation, and tumor progression of human skin keratinocyes. Molecular Carcinogenesis 23: 144-158.
  65. Bleuel, K., Popp, S. Fusenig, N.E., Stanbridge, E.J. & Boukamp, P. (1999). Tumor suppression in human skin carcinoma cells induced by chromosome 15 transfer of TSP-1 overexpression is caused by halting tumor vascularization. Proceedings of the National Academy of Sciences, USA 96: 2065-2070.
  66. Schoop, V., Mirancea, N. & Fusenig N E. (19991. Epidermal organization and differentiation of HaCaT keratinocytes in organotypic co-culture with human dermal fibroblasts. Journal of Inuestigative Dermatology 112: 343-353.
  67. Morgan, T.L., Yang, D., Fry, D.G., Hurlin, P.J., Kohler, S.K., Maher, V.M. & McCormick, J.J. (1991). Characteristics of an infinite life span diploid human fibroblast cell strain and a near diploid strain arising from a clone of cells expressing a transfected v-myc oncogene. Environmental Cell Research 197: 125-136.
  68. Hurlin, P.J., Maher, V.M. & McCormick, J.J. (1989). Malignant transformation of human fibroblasts caused by expression of a transfected T24 HRAS oncogene. Proceedings of the National Academy of Sciences, USA 86: 187-191.
  69. Yang, D., Louden, C., Reinhold, D.S., Kohler, S.K., Maher, V.M. & McCormick, J.J. (1992). Malignant transformation of human fibroblast cells strain MSU-1.1 by (+/-)-7β,8a-dihydroxy-9a,10a-epcxy-7,8,9,10-tetrahydrobenzo[a]pyrene. Proceedings of the National Acadamy of Sciences, USA 89: 3237-3241.
  70. Reinhold, D.S., Walicka, M., Elkassaby, M., Milam, L.D., Kohler, S.K., Dunstan, R.W. & McCormick, J.J. (1996). Malignant transformation of human fibroblasts by ionising radiation. International Journal of Radiation Biology 69: 707-715.
  71. Taggie-Hoffman, A.A., Maher, V.M. & J.J. McCormick (1999). Decreased p16lNK4a expression in tumorgenic fibroblastic cell lines derived from MSU-1.1 cells. Proceedings of the American Association for Cancer Research 40: 684.
  72. Aust, A.E., Falahee, K.J., Maher, V.M. & McCormick, J.J. (1980). Human cell-mediated benzo[a]pyrene cytotoxicity and mutagenicity in human diploid fibroblasts. Cancer Research 40: 4070-4075.
  73. Stuard, S.B. Kerckaert, G.A. & Lehman-McKeeman, L.D. (1999). Characterisation of the metabolic capacity of Syrian hamster embryo (SHE) cells. Toxicological Sciences 48: 366.
  74. Steele, V.E., Sharma, S., Mehta, R., Elmore, E., Redpath, J.L., Rudd, C., Bagheri, D., Sigman, C.C. & Kelloff, G.J. (1996). Use of in vitro assays to predict the efficacy of chemopreventive agents in whole animals. Journal of Cellular Biochemistry 265: 29-53.
  75. Ridder, G.M., Stuard, S.B, Kerckaert, G.A., Cody, D.B., LeBoeuf, R.A. & Isfort, R.J. (1997). Computerized image analysis of morphologically transformed and nontransformed Syrian hamster embryo (SHE) cell colonies: application to objective SHE cell transformation assay scoring. Carcinogenesis 18: 1965-1972.
  76. Gatehouse, D.G. (1994). Mutagenicity. In lnternational Pharmaceutical Product Registration - Aspects of Quality, Safety and Efficacy (ed. A.C. Cartwright & B.R. Mathews), pp. 474-552. New York, USA: Ellis Horwood.
  77. ICH (1999). ICH Guidance SIB: Testing for Carcinogenicity of Pharmaceuticals. Note 1. Web site http://www.ifpma.org.
  78. Federal Register (1998). Federal Register February 23, Vol. 63, No. 35.
  79. ICH (1998). Genotoxicity: a standard battery of genotoxicity testing of pharmaceuticals. In The Proceedings of the Fourth International Conference on Harmonisation (ed. P.F. D'Arcy & D.W.G. Harron), pp. 987-998. Antrim, UK: Greystone Books.
  80. FDA (1993). Food and Drug Administration Draft: Toxicological Principles for the Safety Assessment of Direct Food Additines and Color Additives Used in Food. Washington, DC: US Food and Drug Administration Centre for Food Safety and Applied Nutrition.
  81. Yamasaki, H., Ashby, J., Bignami, M., Jongen, W. Linnainmaa, K., Newbold, R.F., Bguyen-Ba, G., Parodi, S., Rivedal, E., Schiffmann, D., Simons, J.W.I.M. & Vasseur, P. (1996). Nongenotoxic carcinogens: development of detection methods based on mechanisms. Mutation Research 353: 47-63.
  82. Muller, L. & Kasper, P. (1999). Human biological relevance and the use of threshold arguments in regulatory genotoxicity assessment: experience with pharmaceuticals. Mutation Research, in press.
  83. Blaauboer, B.J., Bayliss, M.K., Castell, J.V., Evelo, C.T.A., Frazier, J.M., Groen, K., Gulden M., Guillouzo, A., Hissink, A.M., Houston, J.B., Johanson, G., de Jongh, J., Kedderis, G.L., Reinhardt, C.A., van de Sandt, J.J.M. & Semino, G. (1996). The use of biokinetics and in vitro methods in toxicological risk evaluation. The report and recommendations of ECVAM workshop 15. ATLA 24: 473-497.
  84. Kerckaert, G.A., LeBoeuf, R.A. & Isfort, R.J. (1998). Assessing the predictiveness of the Syrian hamster embryo cell transformation assay for determining rodent carcinogenic potential of single ring aromatic/nitroaromatic amine compounds. Toxicological Sciences 41: 189-197.
  85. ICH (1999). ICH Guidance S2B: A Standard Battery for Genotoxicity Testing of Pharmaceuticals. Web site http://www.ifpma.org/ichl.html.
  86. Muller, K., Kasper, P. Thiele, N., Henke, M., Madle, S. & Muller, L. (1990). A critical comment on the SHE cell transformation assay. Mutagenesis 5: 628-629.
  87. Balls, M. (1998). Why is it proving to be so difficult to replace animal tests? Laboratory Animals 27: 44-47.