ToxSci Advance Access originally published online on August 6, 2008
Toxicological Sciences 2008 105(2):233-234; doi:10.1093/toxsci/kfn138
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gene Expression, Dose-Response, and Phenotypic Anchoring: Applications for Toxicogenomics in Risk Assessment
Central Product Safety, Procter & Gamble, Cincinnati, Ohio 45253
1 To whom correspondence should be addressed at Central Product Safety, Procter & Gamble, Cincinnati, OH 45253. E-mail: daston.gp{at}pg.com. Fax: 1-513-627-0323.
Received July 3, 2008; revision received July 3, 2008; accepted July 5, 2008
There are a number of ways in which global analysis of gene expression may be useful in improving the way we assess the risk of chemical agents, especially at the concentrations to which people are likely to be exposed. The methodology for assessing changes in gene expression is quite sensitive: most of the methods in use today have attomolar or even subattomolar sensitivity. That sensitivity, coupled with the observation that virtually all toxic responses are accompanied by changes in gene expression, suggests that it is possible to use gene expression analysis to determine whether responses at the molecular level continue to occur at dose levels below those that produce frankly adverse effects. This kind of information will be extremely important in addressing controversies about the likelihood of occurrence of those adverse effects at ambient exposure levels.
This issue of Toxicological Sciences includes a seminal paper on the use of global gene expression analysis to support conclusions about dose-dependent transitions in toxic responses, in this case, the response of the rodent nasal epithelium to formaldehyde (Andersen et al., 2008
). These investigators correlate histological and gene expression changes in the nose of the formaldehyde-exposed rat. The study design had both time and dose dimensions. The exposure levels were chosen such that the two lowest dosages did not produce discernible effects at the histological level or changes in cell proliferation, nor were they tumorigenic in a cancer bioassay. The lack of effect at a cellular level suggests that there are thresholds for the adverse responses; the gene expression data confirm that the responses are not linear at low doses. This paper provides a novel way to use microarray data in dose-response assessment, the grouping of genes according to gene ontology terms, such that groups of genes that control particular functions are considered together.
Changes in gene expression accompany virtually every toxic response (Nuwaysir et al., 1999). Several years of research in many laboratories has demonstrated that specific mechanisms of toxicity elicit characteristic profiles of gene expression, suggesting that gene expression may be integral to the toxic response or to the cell's attempts to repair toxicant-induced damage. Perhaps the most obvious examples of gene expression driving the toxic response are the responses to ligands of nuclear receptors, as much of the signal transduction for this pathway depends on gene expression (Boverhof et al., 2008; Naciff et al., 2002). Evaluations of the temporal response of the uterus to an estrogenic stimulus illustrate this point: the families of genes responsible for eliciting each stage of the estrous cycle are expressed prior to the appearance of cellular and morphological changes in the tissue; for example, the genes that control cell division and suppress apoptosis are expressed a few hours before the onset of measurable cell proliferation, transcription factors appear a few hours before changes in cellular differentiation, etc. (Naciff et al., 2007).
The ability of gene expression analysis to identify mechanisms of action has primarily been used for predicting potential toxicity, especially in the pharmaceutical industry. Dose-response assessment is another application of gene expression analysis in toxicology and risk assessment. The sensitivity of the methodology is good enough that it is possible to detect small effects at low levels of exposure, if they exist. The global nature of microarray analysis (i.e., all genes in the genome can be evaluated) means that these experiments can be unbiased. If there are pleiotropic effects at lower exposure levels that would elicit a different profile of gene expression, those genes would not go unnoticed.
The ability to detect effects at levels lower than the lowest-observed effect level for traditional toxicity endpoints makes it possible to address questions about the linearity of the dose-response curve at these lower exposure levels. From a mechanistic standpoint, a comparison of gene expression results at doses at or above the LOAEL with those below the LOAEL can address hypothesized dose-dependent transitions in toxic response. Dose-dependent transitions are inflection points in the dose-response curve that occur when the concentration of the exogenous agent is sufficiently high to alter normal physiological function (Slikker et al., 2004). There is evidence for a dose-dependent transition for formaldehyde, in that inflammatory and proliferative responses in the nasal epithelium do not occur at doses lower than those that are tumorigenic (Conolly et al., 2003). However, it is possible that there were effects produced at lower levels that were simply not detected by the methods used previously.
The present study correlates the gene expression changes that occur at three dose levels and over three time points following short-term exposure to formaldehyde. The three dose levels were similar to the lowest levels used in the chronic bioassay. The highest of the three, 6 ppm, was the lowest tumorigenic concentration, a dose that also induces cell proliferation. The gene expression analysis indicated no changes in gene expression at the lowest exposure concentration. Fifteen genes were changed at the intermediate concentration level after 5 days of exposure, but the effect had resolved by day 15 of the experiment and no genes were changed versus control. Approximately twice as many genes were changed in expression on day 5 in the 6 ppm group and about four times as many on day 15. These results clearly indicate dose- and time-dependent changes in gene expression and support the notion of a dose-dependent transition for cell proliferation, thought to be the precursor event for formaldehyde-induced tumorigenesis. The nature of the gene expression changes is consistent with the phenotypic effects at the various dose levels.
These investigators derived dose-response curves and estimated benchmark doses for sets of affected genes that were based on gene ontology terms for cellular/extracellular location. This grouping of genes for dose-response analysis acknowledges that genes typically do not change in isolation and that one would expect any change that is causally related to toxicity to be in a set of related genes rather than a single gene. It is interesting to note that excluding the lowest benchmark dose, all were in the range of 3.1–4.1 ppm, comparable to the benchmark dose for formaldehyde-induced proliferation (4.9 ppm) reported previously (Schlosser et al., 2003), and even the lowest BMD of 1.6 ppm, for genes associated with basal plasma membrane, were only three-fold lower than the BMD for cell proliferation. This is good evidence for a tight correlation in gene expression changes and changes at higher levels of biological organization.
The results of this study should be useful for the risk assessment for formaldehyde. At present, there is still uncertainty about the mode of action for formaldehyde and the nature of the dose-response relationship at exposure levels below the experimental LOAEL. The research presented by Andersen et al. (2008) supports the notion that there is a clear, dose-dependent transition in the effects of formaldehyde at the level of transcription, as well as at the cellular and histopathological levels. One hopes that this information is taken into consideration in establishing regulatory guidance for formaldehyde.
More importantly, Andersen et al. (2008) have provided a model for the use of gene expression in dose-response analysis. The use of gene ontology terms as a means of grouping genes for dose-response analysis is innovative. It should be noted that these experiments are neither simple nor inexpensive. Dose-response studies cannot be done without multiple dose levels. To this was added multiple time points, a necessary consideration because the histopathological changes to formaldehyde occur over a period of time. Enough animals needed to be evaluated, and microarrays run, at each dose- and time-point to provide reasonable statistical power. Still, given the societal importance of a less uncertain risk assessment for formaldehyde, the results more than justify the expense.
REFERENCES
Andersen ME, Clewell HJ, Bermudez E, Willson GA, Thomas RS. Genomic signatures and dose-dependent transitions in nasal epithelial responses to inhaled formaldehyde in the rat. Toxicol. Sci. (2008) Advance Access published on May 21, 2008.10.1093/toxsci/kfn097.
Boverhof DR, Burgoon LD, Williams KJ, Zacharewski TR. Inhibition of estrogen-mediated uterine gene expression responses by dioxin. Mol. Pharmacol. (2008) 73:82–93.
Conolly RB, Kimbel JS, Janszen D, Schlosser PM, Kalisak D, Preston J, Miller FJ. Biologically motivated computational modeling of formaldehyde carcinogenicity in the F344 rat. Toxicol. Sci. (2003) 75:432–437.
Naciff JM, Jump ML, Torontali SM, Carr GJ, Tiesman JP, Overmann GJ, Daston GP. Gene expression profile induced by 17a-ethinyl estradiol, bisphenol A, and genistein in the developing female reproductive system of the rat. Toxicol. Sci. (2002) 68:184–199.
Naciff JM, Overmann GJ, Torontali SM, Carr GJ, Khambatta ZS, Tiesman JP, Richardson BD, Daston GP. Uterine temporal responses to acute exposurto 17alpha-ethinyl estradiol in the immature rat. Toxicol. Sci. (2007) 97:467–490.
Nuwaysir EF, Bittner M, Trent J, Barrett JC, Afshari CA. Microarrays and toxicology: the advent of toxicogenomics. Mol. Carcinog. (1999) 24:153–159.[CrossRef][Web of Science][Medline]
Schlosser PM, Lilly PD, Conolly RB, Janszen DB, Kimbell JS. Benchmark dose risk assessment for formaldehyde using airflow modeling and a single-compartment, DNA-protein cross-link dosimetry model to estimate human equivalent doses. Risk Anal. (2003) 23:473–478.[CrossRef][Web of Science][Medline]
Slikker WJ, Andersen ME, Bogdanffy MS, Bus JS, Cohen SD, Connolly RB, David RM, Doerrer NG, Dorman DC, Gaylor DW, et al. Dose-dependent transitions in mechanisms of toxicity. Toxicol. Appl. Pharmacol. (2004) 201:203–225.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||