Biologically Motivated Computational Modeling: Contribution to Risk Assessment

M. E. (Bette) Meek1

Safe Environments Programme, Health Canada, Address Locator 0801C2, Tunney's Pasture, Ottawa, Ontario K1A 0L2, Canada

Received September 15, 2004; accepted September 17, 2004

The article highlighted in this issue is "Human Respiratory Tract Cancer Risks of Inhaled Formaldehyde: Dose Response Predictions Derived from Biologically Motivated Computational Modeling of a Combined Rodent and Human Dataset," by Rory Conolly, Julia Kimbell, Derek Janszen, Paul Schlosser, Darin Kalisak, Julian Preston, and Frederick Miller.

In the featured article, Conolly et al. describe the development of the human component of a biologically motivated computational model to predict exposure response at levels of formaldehyde less than those associated with squamous cell carcinomas (SCC) observed in Fischer 344 rats exposed by inhalation. The article addresses extension of the computational model to the entire respiratory tract of humans, complementing a previous description, which presented modeling for the nasal airways of rats (Conolly et al., 2003Go). Extension to the entire respiratory tract is relevant for prediction of risk associated with oronasal breathing of humans, as occurs at higher exertion levels characteristic of those likely in the occupational environment.

The computational model incorporates regenerative cell proliferation as a required step in the induction of tumors by formaldehyde and a contribution from mutagenicity that has greatest impact at low exposures through modeling of complex functional relationships for cancer due to actions of formaldehyde on mutation, cell replication, and exponential clonal expansion. Species variations in dosimetry are taken into account by computational fluid dynamics modeling of formaldehyde flux in various regions of the nose and a single path model for the lower respiratory tract of humans.

Specifically, the animal model includes an anatomically realistic three-dimensional computational fluid dynamics model (CFD) of rat nasal airflow and site-specific flux of formaldehyde into the tissue in which the nasal SCC developed. Flux is the relevant dose metric for the two relevant noncancer effects in the tissues: formation of DNA-protein crosslinks (DPX) and cytolethality/regenerative cellular proliferation (CRCP). A two-stage clonal growth model links the modes of action with mutation accumulation and tumor formation.

In the human component of the model described in this issue, predictions of regional dosimetry are based on human versions of the CFD model and a linked typical path model for the lower respiratory tract. Regional formation of DPX driven by the CFD-predicted flux of formaldehyde into tissue is predicted by a human DPX model based on scale up from rat and rhesus monkey DPX models. CRCP data were those for the rat. Baseline parameter values for the human clonal growth model were calibrated against human lung cancer incidence data (Peto et al., 1992Go; SEER, 2003).

The body of chemical specific and more generic biological information on which the model is based is extensive. Model structure is also consistent with the outcome of consideration of weight of evidence for mode of action of tumor induction in a formal framework (Liteplo and Meek, 2003Go; Meek et al., 2003Go). The database on which this weight of evidence analysis and the computational exposure response model on formaldehyde draws is impressive, including a specifically designed inhalation bioassay in which extensive dose-response data on intermediate endpoints were collected (Monticello et al., 1991Go, 1996Go). The clonal expansion model also draws upon an impressive more generic body of work on relative roles of mutation and cell proliferation in cancer induction dating back over 20 years. The three-dimensional computational fluid dynamics models contribute considerably not only to formaldehyde-specific but generic understanding of site specificity as a function of airflow resulting from the complex anatomy of the nasal passages and lung of rats and humans.

The authors clearly describe uncertainties of the model structure and parameter values which include the lack of direct correlation between DPX and mutation (dose-response data for other potentially mutagenic lesions associated with formaldehyde not being available). Formal sensitivity analysis and Monte-Carlo based analysis of variability and uncertainty were also not conducted precluding identification of model parameters with greatest impact or presentation of confidence intervals around model generated risk predictions. However, in all cases, in the absence of relevant information, the authors have made conservative assumptions, thereby increasing confidence that action taken on the basis of resulting estimated risks would be health protective.

Important in the consideration of the robustness of any biologically motivated computational model and its acceptance is peer input and review, which necessarily must address complexity and integration of a vast array of varying types of information. In this context, it is noteworthy that the computational model described in the highlighted article draws upon earlier input from a steering group of representatives of relevant agencies from two governments, industry, and a consulting group, and incorporates revisions suggested by an external peer review workshop, representing a broad range of expertise including genetic toxicology, cancer biology, epidemiology, modeling, and statistics (Report of Health Canada/U.S. EPA External Peer Review Workshop on Formaldehyde, Health Canada, 1998Go).

In addition to peer input and review, however, it is effective communication of the complex content of biologically motivated computational exposure response models that is likely the critically essential component of their acceptance. Presentation in the highlighted article is worthy of comment in this context. The authors have effectively and efficiently communicated the essential features of a highly complex model including the nature of the information on which the model draws, its limitations, assumptions made in the absence of relevant data, and the comparison of resulting risk estimates with those for approaches incorporating less biological data, though formal sensitivity analysis was not conducted.

While all biologically motivated case specific models entail simplification of cancer biology, which requires selection of a number of parameters and use of simplifying assumptions, they provide insight into the biological basis of responses observed experimentally and are the preferred basis for the extrapolation of data outside of the range of observation. Where underlying databases are robust and the models adequately developed taking into account peer input and review, they are clearly preferred over default methodology for dose-response consideration in risk assessment (It is notable in this context that default is assumed but not necessarily proven to be protective.) The potential adverse impact on advancing understanding of the modes of action of toxic chemicals and their resulting impacts on health of deterrent to those producing data relevant to the development of biologically motivated case-specific models resulting from lack of their application in risk assessment in regulatory context is significant. Avoiding limitations of understanding resulting from inadequate communication is critical in this context and Conolly et al. make significant contribution in clearly, concisely, and defensibly presenting their important work in the highlighted article.

NOTES

1 For correspondence via fax: (613) 954-2486. E-mail: Bette_Meek{at}hc-sc.gc.ca.

REFERENCES

Conolly, R. B., Kimbell, J. S., Janszen, D., Schlosser, P. M., Kalisak, D., Preston, J., and Miller, F. J. (2003). Biologically motivated computational modeling of formaldehyde carcinogenicity in the F344 rat. Toxicol. Sci. 75, 432–447.[Abstract/Free Full Text]

Health Canada. (1998). Report of Health Canada/U.S. EPA External Peer Review Workshop on Formaldehyde. Health Canada, Ottawa, Ontario (unpublished).

Liteplo, R. G., and Meek, M. E. (2003). Inhaled formaldehyde: Exposure estimation, hazard characterization and exposure-response analysis. J. Toxicol. Environ. Health Part B 6, 85–114.[CrossRef][ISI]

Meek, M. E., Bucher, J. R., Cohen, S. M., DeMarco, V., Hill, R. N., Lehman-McKeeman, L. D., Longfellow, D. G., Pastoor, T., Seed, J., and Patton, D. E. (2003). A framework for human relevance analysis of information on carcinogenic modes of action. Crit. Rev. Toxicol. 33, 591–654.[ISI][Medline]

Monticello, T. M., Miller, F. J., and Morgan, K.T. (1991). Regional increases in rat nasal epithelial cell proliferation following acute and subchronic inhalation of formaldehyde. Toxicol. Appl. Pharmacol. 111, 409–421.[ISI][Medline]

Monticello, T. M., Swenberg, J. A., Gross, E. A., Leininger, J. R., Kimbell, J. S., Seilkop, S., Starr, T. B., Gibson, J. E., and Morgan, K. T. (1996). Correlation of regional and nonlinear formaldehyde-induced nasal cancer with proliferating populations of cells. Cancer Res. 56, 1012–1022.[Abstract]

Peto, R., Lopez, A. D., Borcham, J., Thun, M., and Heath, C. (1992). Mortality from tobacco in developed countries: Indirect estimation from national vital statistics. Lancet 339, 1268–1278.[ISI][Medline]

Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov). SEER Stat Database: Incidence – SEER 9 Regs Public-Use, Nov 2002 Sub (1973–2000), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003, based on November 2002 submission.