ToxSci Advance Access originally published online on March 25, 2007
Toxicological Sciences 2007 98(2):327-331; doi:10.1093/toxsci/kfm068
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Genetic Toxicity Assessment: Employing the Best Science for Human Safety Evaluation Part VI: When Salt and Sugar and Vegetables Are Positive, How Can Genotoxicity Data Serve to Inform Risk Assessment?
Toxicology and Environmental Research and Consulting, The Dow Chemical Company, Midland, Michigan 48674
1 To whom correspondence should be addressed. Fax: (989) 638-9863. E-mail: lpottenger{at}dow.com.
Received February 8, 2007; accepted March 19, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT REFERENCES |
|---|
This opinion piece examines the current approaches in the design and evaluation of genotoxicity data and recommends an alternative that would provide information that could be more useful to human risk assessment. It is suggested that genotoxicity studies, both in vitro and in vivo, be designed similar to other traditional toxicology studies, such that a dose-response relationship is characterized, including identification of a "no-observed-adverse-effect-level" dose. It is further suggested that genotoxicity tests should no longer be designed or interpreted in isolation but should be examined in the context of other available data including toxicokinetics, mechanism of genotoxicity, and relevant exposure information. The answer to improving genetic toxicology testing does not lie in coming up with better, "more sensitive" genotoxicity test systems but rather in the incorporation of contextual improvements in both the experimental design and the interpretation of data collected using the current models. Such a strategy will better position the toxicology and risk assessment communities to cope with the current intellectually uncomfortable dichotomy that directs disproportionate scientific resource to addressing genetic toxicity findings of anthropogenic substances, regardless of dose-exposure context, while at the same time ignoring the plethora and comparatively large amounts of genotoxic and toxic substances that are inescapably present in what are otherwise regarded as healthy foods (salt, sugar, and vegetables).
Key Words: carcinogenesis; genetic toxicology; risk assessment.
Genotoxicity is a conditional property of a chemical, i.e., a chemical may be positive for genotoxicity under certain conditions and negative under a different set of conditions. Accordingly, in genetic toxicology screening research, results are generally reported with a concluding statement that a test material is either positive or negative "under the experimental conditions used". In some studies, negative findings are presented as a "failure" to detect a genotoxic response following examination in the test system of choice. Yet in other reports, negative results are often contextualized with the statement that the test system being used may not be "sensitive enough" to detect a weak genotoxic response. And to further confound the situation, in many instances, especially when a large body of genotoxicity data exists, conflicting positive, negative, and equivocal results are often reported for the same end point by different investigators. All these caveats and contradictions instill a certain degree of ambiguity and uncertainty to the results obtained with genetox tests and to the conclusions drawn from them. This conundrum is further accentuated by the demonstration of genotoxic activity for life-sustaining chemicals such as common table salt and sugar.
These all too common scenarios lead us to question our present approaches for integrating findings from genetic toxicology tests as currently conducted into science-based inputs that inform the risk assessment process. Traditionally, genetox data have been viewed as binary input, with results either "yes, genotoxic" or "no, not genotoxic," Typically, dose-response information is not considered in the interpretation of genetox data, and as a result, there is no formalized guidance encouraging identification of no-observed-adverse-effect-levels (NOAELs) in either in vitro or in vivo genotoxicity studies. This opinion piece proposes that considerations of dose response in genotoxicity studies, particularly when coupled with related dosimetry and toxicity data from either animal or human studies, will provide a more sophisticated and scientifically defensible evaluation of genetox data that is truly informative of potential human health risks associated with real-world exposures to environmental chemicals.
Of particular concern is the universal assumption that positive results from genotoxicity tests will default findings of animal carcinogenicity to linear, no-threshold human risk evaluation models. The consequences of this overriding and dominant influence of genotoxicity testing on risk assessment is enormous, in that many new or existing compounds of potential or actual societal value may be deselected from development or further use based simply on results from standard in vitro genotoxicity tests (and despite availability of other ancillary toxicity evaluations which may otherwise indicate a low order of toxicity).
Is there something wrong with the existing test systems currently used in genetic toxicology that place the toxicology and risk assessment communities in this position? The answer is "no," the problem does not lie with the test systems themselves. The responses observed in these test systems are real, reproducible, and even in many cases biologically plausible under the experimental conditions and doses being investigated. However, the more important question should not be focused on the quality of current genotoxicity tests but rather how these datasets can be better designed to inform risk assessment. This is obviously a very different and more challenging question but one that is legitimately attracting increasing attention (Dearfield and Moore, 2005
; Müller et al., 2003). Should identification of positive responses, regardless of dose or other context provided by existing ancillary animal toxicity or real-world human exposure information, truly represent the primary objective for in vitro genetox testing? An examination of the validity of this central objective can be aided, in part, by a critical evaluation of genotoxicity studies of common natural substances including salt, sugar, and everyday healthy foods.
Let us examine the case of sodium chloride, common table salt. In vitro results demonstrate that sodium chloride induced DNA damage in Chinese hamster ovary (CHO) (Galloway et al., 1987) and murine kidney cells in culture (Kültz and Chakravarty, 2001), chromosomal aberrations in CHO cells (Galloway et al., 1987; Seeberg et al., 1988), and mutations in a mouse lymphoma cell line (Moore and Brock, 1988; Seidel et al., 2004). These effects were generally seen at concentrations higher than 5 mg/ml, doses that have a profound effect on the osmolality of the treatment medium. In all these studies, there were concentrations of sodium chloride at which no genotoxic effects were observed. These studies are informing us that perturbations in the ionic balance of the cells could lead to DNA damage, and such effects are biologically plausible, albeit only at extremely high salt concentrations, given the importance of ionic strength on various cellular processes. Although the exact mechanism for sodium chloride–induced DNA damage is not known, either physical distortion in chromatin structure or the generation of free radicals (likely secondary to cellular disruption) have been postulated as being responsible for inducing damage to the DNA phosphodiester backbone (Kültz and Chakravarty, 2001). These effects of sodium chloride are not limited to in vitro systems. Eastmond and coworkers have reported that high concentration (2%) of sodium chloride in the diet can induce clastogenicity in the bladder epithelium of the rat (Balakrishnan et al., 2002). From these studies, it is evident that the in vitro test systems have accurately predicted the response of at least one in vivo tissue. This example further illustrates the point that genotoxicity of sodium chloride is a conditional property, in that under certain conditions (high doses leading perturbations in osmotic strength) it can elicit a genotoxic response, while under a different set of conditions (lower dose levels) such a response is not induced nor is it biologically plausible.
Sucrose, an important ingredient of our daily lives, offers another example. It has been shown to be an inducer of chromosome breakage in CHO cells; up to 26% of the cells had aberrations following a short treatment with approximately 110 mg/ml (
0.32M) of sucrose (Galloway et al., 1987). No such effects were observed at concentrations of 85 mg/ml (0.24M). These are indeed high, nonphysiological concentrations; normal human fasting blood glucose levels are 6.5–10.9 mg/ml, whereas the sucrose concentration reported as genotoxic would be well above the range of severe, life-threatening diabetic disease. Although the in vitro effects observed seemed to be specific for clastogenicity, in vivo sucrose was shown to induce 1.8- and 2.3-fold increases in cII mutant frequency in the colon, but not the liver, of Big Blue rats maintained on a diet supplemented with 10 or 30% (wt/wt) of sucrose (Dragsted et al., 2002). The mutant frequency in the colon of rats given 3% (wt/wt) dietary sucrose was not affected. These results seem to suggest that sucrose, under certain conditions, could induce a genotoxic response both in vitro and in vivo. However, an equally informative finding here, based upon the dose-response data, is the existence of thresholds or NOAELs for these genotoxic effects.
How should toxicologists interpret the salt and sugar findings regarding their potential to inform human risk assessment? Although the original purpose of genetic toxicology testing during the 1960's and early 1970's was to identify germ cell mutagens, the seminal publication suggesting "carcinogens are mutagens" (Ames et al., 1973) redirected the field of genetic toxicology toward identifying and/or classifying animal carcinogens. This approach largely remains the operational paradigm today despite the fact that experience of the past decades points to the low specificity of genetic toxicology tests, especially the mammalian cell tests (below 45%), to predict the rodent bioassay results for carcinogenicity (Kirkland et al., 2005). Ames and colleagues have also emphasized the pitfalls inherent in attempting to predict carcinogenicity results from in vitro mutagenicity assays (Ames and Gold, 1997; Ames et al., 1990). We propose that results from genetic toxicology assays must be judged in context of the entirety of available animal and human data, not the least of which should be consideration of its dose/exposure implications. The current operational assumption that genotoxicity observed under any conditions logically infers a carcinogenic liability must be reevaluated, particularly since it is now well appreciated that carcinogenicity is a complex and multistep process involving initiation, promotion, and progression.
The examples of salt and sugar underscore the importance of understanding the full dose-response relationship for any effect, including genotoxicity. Characterization of dose response is a fundamental tenet of toxicology, and toxicology researchers legitimately spend considerable time and resources to design toxicology studies such that a clear NOAEL will be identified at the end of a complicated study. Furthermore, it is important to remember that dose influences not only the effect but also the mechanism responsible for the observed effect (Counts and Goodman, 1995; Slikker et al., 2004a,b) or lack of an effect. Unfortunately, there has been little parallel emphasis placed in the past on identifying a NOAEL dose in genetic toxicology studies. To the contrary, a significant emphasis is placed on optimizing conditions for identifying an effect regardless of important contextual factors such as dose and other ancillary animal toxicity responses (including such extremes as lethality!). Such emphasis has resulted in a myriad of genetic toxicology tests and complex testing strategies which all too often yield apparently contradictory results.
The argument that all toxicity tests, including genetic toxicity tests, have limited statistical power relative to their desired intent to predict potential population-level response is traditionally used as the primary justification for testing chemicals at extremely high doses regardless of context to known toxicity or exposure characteristics of the tested agent. However, this argument is no longer tenable in that it mandates the assumption that biological responses must be linear over the entire range of dose input, an assumption that modern biology and toxicology have shown repeatedly as unequivocally false. This historical emphasis on characterization of high-dose genotoxicity responses, which are often unwarranted in and of themselves if considered in context of animal toxicity or human exposures, has nonetheless also diverted needed attention from evaluation of the low end of the dose-response curve including identification of NOAELs. It should be noted that other branches of toxicology, e.g., reproductive and developmental toxicity, have always embraced the concept of experimentally defined NOAEL values, in spite of the limitations of sample sizes, and have adopted methodologies such as benchmark dose approaches to refine such NOAEL determinations.
The existence of NOAEL doses in genetic toxicology is a controversial topic. Certainly, everyone is familiar with the traditional position that a single DNA lesion in a single cell could result in a mutation, which in turn could lead to the initiation of a normal cell into a cancerous cell. The underpinnings for this thinking are rooted in the single-hit theory of mutation induction for ionizing radiation and the reported linearity for dose response for DNA adduct formation. This conventional thinking has dominated the field of genetic toxicology for the past several decades, i.e., there is no safe dose level for those chemicals that are capable of interacting with DNA, and thus many genetic toxicologists have ignored the concept of identifying a NOAEL in the design of their experiments.
The state-of-the-science of mutagenicity no longer supports an unquestioned adherence to the no-threshold concept for genotoxicity responses and its associated implication for consideration (or lack thereof) of dose response. Understanding of the molecular mechanisms of genotoxicity has revealed the presence of multiple and varied steps intervening between the initial genetic insult (e.g., formation of DNA adducts) and the fixation of a mutation, the key event of a heritable change in genetic information. In addition, evidence of multiple redundancies present in biological systems provides alternate pathways for the damaged cell to return to a normal state or remove itself through cell death (apoptosis). It is recognized that redundancy of genetic code and tolerance toward some amino acid substitutions in the normal functioning of a protein allows for some flexibility to cope with certain DNA sequence changes and provides additional protection and stability to the system. Recent studies with DNA-reactive chemicals consistently suggest that even direct-acting genotoxic substances may exhibit clear NOAEL responses, i.e., doses that do not induce mutagenic effects different from background. Specific examples include the induction of sex-linked recessive lethal mutations in Drosophila by exposure to ethylene oxide or propylene oxide (Nivard et al., 2003), induction of mutations in the lac I transgene of Big Blue rats dosed with 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (Hoshi et al., 2004), micronucleus induction in mice by mitomycin C or 1-ß-D-arabinofuranosylcytosine (Asano et al., 2006), induction of gene mutations and micronucleus formation in vitro by methylmethanesulfonate (MMS) (Jenkins et al., 2005), and induction of mutations in the in vitro mouse lymphoma cell assay with MMS or methylnitrosourea (Pottenger et al., 2006).
Genotoxicity induced by mechanisms that do not involve the reactivity of a chemical/its metabolites with DNA are conceptually and experimentally further along in being recognized as having a nonlinear dose-response curve with a NOAEL (and may well reflect the findings with salt and sugar). Indirect genotoxicity can be the result of interaction/interference of a chemical with the function of enzymes involved in DNA replication (e.g., topoisomerase II) or repair or other normal cellular replication processes (e.g., spindle formation). For certain chemicals, the mutagenicity is considered to be the consequence of disruption of normal defense mechanisms, such as glutathione depletion, or induction of oxidative stress secondary to chemical metabolism or other cellular cytotoxic responses. Such effects are often limited to high doses that saturate detoxification pathways or disrupt homeostatic mechanisms (Slikker et al., 2004a,b). A key recognition from these observations is that dose can influence mechanism of toxicity, and for toxicity data, including genotoxicity data, to be useful to risk assessment, it must be anchored to relevant doses, relevant routes of exposure, and relevant dose rates (Conolly et al., 1999).
We propose that, in addition to more thoughtful determination of top dose selection in genotoxicity tests (avoiding artifactual false positives), attention to identification of clear NOAELs for genotoxicity should be factored into the design of genetic toxicology studies. Such data will likely prove extremely valuable in the risk assessment process. Risk assessments are conducted in the context of exposure potential, and thus it makes good scientific sense that the NOAELs identified for various genetic toxicology end points, when examined collectively, should at least inform the risk assessor whether or not an adverse event is likely to occur at a particular exposure level.
Targeting incorporation of both NOAEL and appropriate high dose selection considerations into the study design of both in vitro and in vivo genetox testing is likely to provide valuable dose-response perspective to future chemical risk assessments. These approaches will be particularly useful when coupled with continuing refinements in both internal dose characterization of animal toxicity tests and real-world human and biomonitoring exposure assessments. Making in vitro results informative for risk assessment entails some understanding of how the in vitro doses used for genotoxicity testing compare to internal doses from in vivo studies, doses that actually reach (target) tissues in vivo, and how such information can be translated across dose, route of administration, species, and other pharmacokinetic considerations. Thus, the time is ripe to use the full range of in vivo toxicity findings to provide important toxicological context for interpretation of in vitro genetox data. For example, positive genetox responses identified only at very high in vitro concentrations, not otherwise attainable in the whole animal due to toxicity limitations, should logically imply a lack of relevance of this high-dose response for risk assessment, i.e., the toxicity response itself would be the overriding health concern. In addition, recent advances in analytical and modeling sciences now provide readily available and reasonably inexpensive tools to develop dosimetry information that can identify the likelihood of high-dose–specific, nonlinear responses, as well as provide important context to potential real-world human exposures. For example, methods have been developed that should significantly enhance the ability to quantitate key internal dose parameters in animals as part of a standard repeat dose toxicity evaluation of chemicals (Saghir et al., 2006).
We have stated at the outset of this opinion piece that genotoxicity is a conditional property, i.e., a chemical may be positive for genotoxicity under certain conditions and negative under a different set of conditions. The response is dependent upon factors such as the target cells (e.g., DNA repair and metabolic competency), dose, and the end point evaluated. It is conceivable that most, if not all, chemicals will be genotoxic under some specific set of conditions, such as using the S-9 fraction from a different species or different tissues, use of a different inducing agent than the commonly used Aroclor-induced rat liver S-9 for in vitro studies, or using a genetically engineered cell line or rodent model with exquisite "sensitivity" to detect genotoxins. The fact that salt, sugar, and even juices from commonly consumed "healthy" vegetables and food crops have been shown to induce genotoxicity in cell cultures (Charles et al., 2002; Kassie et al., 1996) must serve as a signal reminder that findings from genotoxicity tests cannot and should not be simplistically translated, as is often currently done, to inferences of adverse human health outcomes.
The basic science of genetic toxicology has progressed far beyond the simple binary approach to interpretation of testing results. The answer to improving genetic toxicology testing does not lie in coming up with better, "more sensitive" test systems but rather in the incorporation of contextual improvements in both the experimental design and the interpretation of data collected using the current models. Such a strategy will better position the toxicology and risk assessment communities to cope with the current intellectually uncomfortable dichotomy that directs disproportionate scientific resource to addressing genetic toxicity findings of anthropogenic substances, regardless of dose-exposure context, while at the same time ignoring the plethora and comparatively large amounts of genotoxic and toxic substances that are inescapably present in what are otherwise regarded as healthy foods (Mattsson, 2007). Serious consideration of such broader perspectives may well drive evolution of more rational and scientifically defensible genotoxicity testing and interpretation strategies in the future. Genetic toxicologists are learning that in vitro positive findings on such innocuous chemicals as salt and sugar are "false positives," and that it is appropriate to limit in vitro test concentrations so as to not unduly affect osmotic strength. Ample opportunities exist to improve genetox study design well beyond such simple limit dose concepts, and by continually questioning the conventional wisdom, making genotoxicity data more useful to inform human risk assessment.
|
REFERENCES |
|---|
|
TOPABSTRACTREFERENCES |
|---|
Ames BN, Durston WE, Yamasaki E, Lee FD. Carcinogens are mutagens: A simple test system combining liver homogenates for activation and bacteria for detection. Proc. Natl. Acad. Sci. U.S.A. (1973) 70:2281–2285.
Ames BN, Gold LS. Environmental pollution, pesticides, and the prevention of cancer: Misconceptions. FASEB J. (1997) 11:1041–1052.[Abstract]
Ames BN, Profet M, Gold LS. Nature's chemicals and synthetic chemicals: Comparative toxicology. Proc. Natl. Acad. Sci. U.S.A. (1990) 87:7782–7786.
Asano N, Torous DK, Tometsko CR, Dertinger SD, Morita T, Hayashi M. Practical threshold for micronucleated reticulocyte induction observed for low doses of mitomycin C, Ara-C and colchicine. Mutagenesis (2006) 21:15–20.
Balakrishnan S, Uppala PT, Rupa DS, Hasegawa L, Eastmond DA. Detection of micronuclei, cell proliferation and hyperdiploidy in bladder epithelial cells of rats treated with o-phenylphenol. Mutagenesis (2002) 17:89–93.
Charles GD, Linscombe VA, Tornesi B, Mattsson JL, Gollapudi BB. An in vitro screening paradigm for extracts of whole foods for detection of potential toxicants. Food Chem. Toxicol. (2002) 40:1391–1402.[CrossRef][Web of Science][Medline]
Conolly RB, Beck BD, Goodman JI. Stimulating research to improve the scientific basis of risk assessment. Toxicol. Sci. (1999) 49:1–4.
Counts JL, Goodman JI. Principles underlying dose selection for, and extrapolation from, the carcinogen bioassay: Dose influences mechanism. Regul. Toxicol. Pharmacol. (1995) 21:418–421.[CrossRef][Web of Science][Medline]
Dearfield KL, Moore MM. Use of genetic toxicology information for risk assessment. Environ. Mol. Mutagen. (2005) 46:236–245.[CrossRef][Web of Science][Medline]
Dragsted LO, Daneshvar B, Vogel U, Autrup HN, Wallin H, Risom L, Moller P, Molck AM, Hansen M, Poulsen HE, et al. A sucrose-rich diet induces mutations in the rat colon. Cancer Res. (2002) 62:4339–4345.
Galloway SM, Deasy DA, Bean CL, Kraynak AR, Armstrong MJ, Bradley MO. Effects of high osmotic strength on chromosome aberrations, sister-chromatid exchanges and DNA strand breaks, and the relation to toxicity. Mutat. Res. (1987) 189:15–25.[CrossRef][Web of Science][Medline]
Hoshi M, Morimura K, Wanibuchi H, Wei M, Okochi E, Ushijima T, Takaoka K, Fukushima S. No-observed effect levels for carcinogenicity and for in vivo mutagenicity of a genotoxic carcinogen. Toxicol. Sci. (2004) 81:273–279.
Jenkins GJ, Doak SH, Johnson GE, Quick E, Waters EM, Parry JM. Do dose response thresholds exist for genotoxic alkylating agents? Mutagenesis (2005) 20:389–398.
Kassie F, Parzefall W, Musk S, Johnson I, Lamprecht G, Sontag G, Knasmuller S. Genotoxic effects of crude juices from Brassica vegetables and juices and extracts from phytopharmaceutical preparations and spices of cruciferous plants origin in bacterial and mammalian cells. Chem. Biol. Interact. (1996) 102:1–16.[CrossRef][Web of Science][Medline]
Kirkland D, Aardema M, Henderson L, Müller L. Evaluation of the ability of a battery of three in vitro genotoxicity tests to discriminate rodent carcinogens and non-carcinogens I. Sensitivity, specificity and relative predictivity. Mutat. Res. (2005) 584:1–256.[Web of Science][Medline]
Kültz D, Chakravarty D. Hyperosmolality in the form of elevated NaCl but not urea causes DNA damage in murine kidney cells. Proc. Natl. Acad. Sci. U.S.A. (2001) 98:1999–2004.
Mattsson J. Mixtures in the real world: The importance of plant self-defense toxicants, mycotoxins, and the human diet. Toxicol. Appl. Pharmacol. (2007) doi: 10.1016/j.taap.2006.12.024.
Moore MM, Brock KH. High concentrations of sodium chloride induce a "positive" response at the TK locus of L5178Y/TK+/– mouse lymphoma cells. Environ. Mol. Mutagen. (1988) 12:265–268.[Web of Science][Medline]
Müller L, Blakey D, Dearfield KL, Galloway S, Guzzie P, Hayashi M, Kasper P, Kirkland D, MacGregor JT, Parry JM, et al. Strategy for genotoxicity testing and stratification of genotoxicity test results—Report on initial activities of the IWGT Expert Group. Mutat. Res. (2003) 540:177–181.[Web of Science][Medline]
Nivard MJM, Czene K, Segerback D, Vogel EW. Mutagenic activity of ethylene oxide and propylene oxide under XPG proficient and deficient conditions in relation to N-7-(2-hydroxyalkyl)guanine levels in Drosophila. Mutat. Res. (2003) 529:95–107.[Web of Science][Medline]
Pottenger LH, Zhang F, Schisler MR, Bartels MJ, Gollapudi BB. Dose-responses and apparent thresholds for DNA adducts and in vitro mutagenicity in mouse lymphoma cells treated with MMS and MNU. In: Environ. Mol. Mutagen. P-2 (Abstract), Environmental Mutagenesis Society Annual Meeting, Vancouver, Canada (2006).
Saghir SA, Mendrala AL, Bartels MJ, Day SJ, Hansen SC, Sushynski JM, Bus JS. Strategies to assess systemic exposure of chemicals in subchronic/chronic diet and drinking water studies. Toxicol. Appl. Pharmacol. (2006) 211:245–260.[CrossRef][Web of Science][Medline]
Seeberg AH, Mosesso P, Forster R. High-dose-level effects in mutagenicity assays utilizing mammalian cells in culture. Mutagenesis (1988) 3:213–218.
Seidel SD, Sparrow BR, Kan HL, Stott WT, Schisler MR, Linscombe VA, Gollapudi BB. Profiles of gene expression changes in L5178Y mouse lymphoma cells treated with methyl methanesulfonate and sodium chloride. Mutagenesis (2004) 19:195–201.
Slikker W Jr,, Andersen ME, Bogdanffy MS, Bus JS, Cohen SD, Conolly RB, David RM, Doerrer NG, Dorman DD, Gaylor DW, et al. Dose-dependent transitions in mechanisms of toxicity. Toxicol. Appl. Pharmacol. (2004a) 201:203–225.[CrossRef][Web of Science][Medline]
Slikker W Jr,, Andersen ME, Bogdanffy MS, Bus JS, Cohen SD, Conolly RB, David RM, Doerrer NG, Dorman DD, Gaylor DW, et al. Dose-dependent transitions in mechanisms of toxicity: Case-studies. Toxicol. Appl. Pharmacol. (2004b) 201:226–294.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. K. Elespuru, R. Agarwal, A. H. Atrakchi, C. A. H. Bigger, R. H. Heflich, D. R. Jagannath, D. D. Levy, M. M. Moore, Y. Ouyang, T. W. Robison, et al. Current and Future Application of Genetic Toxicity Assays: The Role and Value of In Vitro Mammalian Assays Toxicol. Sci., June 1, 2009; 109(2): 172 - 179. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||