3
Toxicokinetics
This chapter provides a brief overview of EPA’s evaluation of the toxicokinetics of inhaled formaldehyde in humans and other mammals. This is followed by the committee’s analysis of the use of formaldehyde toxicokinetic data, pharmacokinetic models, and biologically based dose-response (BBDR) models in the 2022 Draft Assessment (EPA, 2022). The committee’s analysis focuses on the toxicokinetic evidence and modeling approaches presented in the 2022 Draft Assessment that were used to support EPA’s key conclusions and derivation of toxicity values. This chapter focuses on inhalation since this route of exposure is the primary focus of the 2022 Draft Assessment.
OVERVIEW OF INHALED FORMALDEHYDE TOXICOKINETICS IN THE 2022 DRAFT ASSESSMENT
The 2022 Draft Assessment provides an extensive evaluation of the toxicokinetics of formaldehyde (Section 1.2.4 and Appendix A, Section A.2). Formaldehyde is a naturally occurring, volatile, water soluble, one-carbon aldehyde. Upon inhalation, formaldehyde rapidly undergoes hydration on contact with the moist mucus layer found in the respiratory tract. Hydration of formaldehyde in water or following addition of alcohol yields methanediol (formaldehyde hydrate), or hemiacetals and acetals, respectively. Once in the respiratory tract epithelium, formaldehyde can undergo metabolism, including enzymatic reactions with the respiratory tissue and nonenzymatic reactions with glutathione and macromolecules, including proteins and DNA.
DISTRIBUTION OF INHALED FORMALDEHYDE
The 2022 Draft Assessment considers factors that influence the distribution of inhaled formaldehyde at the point of entry in the respiratory tract. EPA also reviewed available dosimetry models evaluating the initial delivery of formaldehyde to the upper respiratory tract (URT). A detailed description of dosimetry modeling efforts in humans, monkeys, and rats is provided in Appendix B, Section B.2.2 of the 2022 Draft Assessment.
Factors considered by EPA that could influence the distribution of inhaled formaldehyde include species differences in airway anatomy and physiology, especially the roles of nasal turbinate structure and nasal breathing in rodents versus oronasal breathing in humans. At low inhaled concentrations (<20 ppb), absorption of formaldehyde is nearly complete in the human URT, with limited amounts of formaldehyde reaching the lung. Formaldehyde demonstrates a perpendicular concentration gradient within the epithelium lining the respiratory tract (Overton, 2001). Highest concentrations are anticipated to occur near the airway lumen. As formaldehyde diffuses through the epithelium, a portion of the initially absorbed formaldehyde is lost to hydration, local tissue metabolism, and chemical reactions between formaldehyde and macromolecules. Other factors can also influence the toxicokinetics of formaldehyde. For example, reflex bradypnea occurs in rodents following formaldehyde inhalation, resulting in reduced minute volume (Chang and Barrow, 1984; Chang et al., 1981). This physiologic response to inhaled irritants does not occur in humans or nonhuman primates.
METABOLISM, BINDING, AND REMOVAL OF INHALED FORMALDEHYDE
The 2022 Draft Assessment provides a discussion of the fate of formaldehyde following inhalation. Formaldehyde is metabolized in the URT by glutathione-dependent class III alcohol dehydrogenase (ADH3) and to a lesser extent by S-formyl-glutathione dehydrogenase to formic acid. Formaldehyde can bind noncovalently to glutathione, tetrahydrofolate, or albumin in nasal mucus. It can also bind covalently to macromolecules, forming DNA–protein crosslinks (DPCs); DNA–DNA crosslinks (DDCs); hydroxymethyl–DNA (hm-DNA) adducts; and protein adducts, including N6-formyllysine (Edrissi et al., 2013a,b). Repeated (28-day) exposure of rats to lower formaldehyde concentrations (up to 300 ppb) did not result in formation of either DNA monoadducts (N2-HOMe-dG) or DPCs in the URT or bone marrow (Leng et al. 2019). At higher exposure concentrations, a concentration-dependent increase in DPC formation is observed in the nasal cavity of animals following formaldehyde inhalation.
A key question addressed by EPA concerns the systemic delivery of inhaled formaldehyde to distant sites. Blood formaldehyde levels of approximately 0.1 mM remain unaltered following formaldehyde inhalation, suggesting that inhaled formaldehyde is not significantly absorbed into blood (Casanova et al., 1988; Heck et al., 1985; Kleinnijenhuis et al., 2013). Because formaldehyde is present in tissues endogenously, EPA also considered how inhaled (exogenous) formaldehyde and endogenous formaldehyde could contribute to adduct formation. Studies (Lu et al., 2010a,b, 2011; Moeller et al., 2011; Yu et al., 2015; Lai et al., 2016) distinguishing DNA monoadducts (e.g., N2-HOMe-dG) or DPCs formed from endogenous or exogenous formaldehyde were used to support EPA’s conclusion that exogenous formaldehyde is not distributed to the bone marrow or other distant tissues.
Finding: EPA concluded that inhaled formaldehyde is not distributed to an appreciable extent beyond the respiratory tract to systemic sites; thus, inhaled formaldehyde is not directly interacting with tissues distal to the portal of entry to elicit effects. EPA’s conclusions regarding systemic delivery of inhaled formaldehyde are based on its expert judgment, with the support of available scientific evidence.
Finding: Despite the lack of evidence regarding systemic delivery of formaldehyde to distant sites, the biological basis for observed systemic effects (described in Chapters 4 and 5) remains unclear. Additional research is needed to address this apparent discrepancy.
DOSIMETRY MODELS
Numerous mathematical models of inhalation dosimetry have been developed for formaldehyde and were examined by EPA for applicability. The models were developed to recapitulate key toxicokinetic observations from experimental animal studies. Several physiologically based pharmacokinetic (PBPK) models describe the disposition of inhaled formaldehyde reacting with upper-respiratory-tract tissue, resulting in the formation (and repair) of DPCs (Conolly et al., 2000; Subramaniam et al., 2008; Klein et al., 2011). Several computational fluid dynamics (CFD) airflow and material transport models account for species differences in airway anatomy and physiology (Hubal et al., 1997; Kimbell et al., 2001). Some CFD models account for the influence of endogenous formaldehyde on the toxicokinetics of inhaled formaldehyde (Schroeter et al., 2014; Campbell et al., 2020). These models predict that the uptake of low concentrations of formaldehyde will be reduced by the presence of endogenous formaldehyde. Other CFD models are coupled with time-dependent PBPK models that describe boundary conditions at the air–tissue interface (Corley et
al., 2015). These boundary conditions account for tissue reactions, first-order and saturable metabolism, and other factors that influence formaldehyde toxicokinetics.
Collectively, EPA used these dosimetry models to show that inhaled formaldehyde is not deposited uniformly throughout the nose; rather, some regions in the nasal cavity receive a higher delivery of formaldehyde compared with other nasal regions. Sites within the nasal epithelium with higher formaldehyde flux to the tissues are also areas where DPCs form and tumors are most likely to arise. Models predict that localized deliveries of formaldehyde to portions of the human nose are comparable to those seen in rats exposed at similar concentrations (Kimbell et al., 2001). The available models suggest further that total nasal deposition is lower in humans or nonhuman primates than in rats, leading to greater penetration of inhaled formaldehyde to the lower respiratory tract.
BBDR models have also been developed for formaldehyde (Conolly et al., 2003, 2004) and were extensively evaluated by EPA. These models incorporate dosimetric and mechanistic data into a single computational model. They have two main subunits: species-specific CFD models to describe formaldehyde delivery in the respiratory tract, and a two-stage clonal growth model for formaldehyde carcinogenesis. The human version of the model (Conolly et al., 2004) incorporates a typical path model (Overton et al., 2001) of the lower respiratory tract that allows for the prediction of formaldehyde delivery to the entire human respiratory tract. A detailed National Research Council (NRC) analysis of the use of these models by EPA is available in an earlier formaldehyde assessment (NRC, 2011).
EPA’S USE OF TOXICOKINETIC DATA AND DOSIMETRY MODELS
The committee first examined EPA’s broad use of toxicokinetic data in the 2022 Draft Assessment. The committee also considered whether toxicokinetic data or toxicokinetic models influenced EPA’s derivation of either its evidence integration judgments for noncancer health effects and the reference concentration (RfC), or its estimation of inhalation unit risk (IUR) for cancer incidence.
Findings: The committee found that EPA used toxicokinetic data as follows:
- EPA used these data to support its assumption that “inhaled formaldehyde is not distributed to an appreciable extent beyond the respiratory tract to systemic sites. Thus, EPA assumed that inhaled formaldehyde does not directly interact with tissues distal to the portal of entry to elicit effects. EPA’s conclusions regarding systemic delivery of inhaled formaldehyde are based on its expert judgment, with the support of the available scientific evidence.
- EPA concluded that studies examining potential associations between levels of formaldehyde or formaldehyde by-products (e.g., formate) measured in distal tissues and health outcomes were not relevant to inhaled formaldehyde. This conclusion is consistent with EPA’s state-of-practice methods and supported by the available scientific evidence.
- EPA concluded that formaldehyde toxicokinetics show significant route-to-route difference (e.g., inhalation versus oral). With few exceptions (e.g., genotoxicity), EPA focused solely on inhalation studies. The committee supports this decision, and found it to be consistent with EPA’s state-of-practice methods.
- EPA used toxicokinetic data as a primary consideration informing causality during its evidence synthesis step, as “an explanation for any observed differences in responses
- across route of exposure, other aspects of exposure, species, or life stages.” Toxicokinetic data were used to examine consistency across studies and to evaluate biological gradient/dose-response data. Overall, the committee found that the portions of the 2022 Draft Assessment describing the toxicokinetics of formaldehyde and associated computational models are well organized and extensive. The 2022 Draft Assessment accurately reflects current understanding of the toxicokinetics of inhaled formaldehyde. The literature review in the 2022 Draft Assessment appears to be up to date and includes relevant studies published as of the assessment’s release date. The use of these data in the synthesis step is consistent with EPA’s state-of-practice methods.
EPA used the available models to derive the candidate RfCs (cRfCs) for respiratory tract pathology seen in animals. Dosimetry modeling was used to estimate a human equivalent concentration and account for toxicokinetic differences between animals and humans. These cRfCs were applied to lesion data from two studies (Battelle Columbus Laboratories, 1982; Kerns et al., 1983). In its dosimetry modeling efforts, EPA initially used a CFD model (Kimbell et al., 2001) to determine average flux values in the anterior region of the rat nasal cavity that corresponded to the rat benchmark concentration lower bound (BMCL) derived from the incidence of squamous metaplasia seen in a study. Human CFD models were then used to estimate the exposure concentration at which any region in the human nose (see Appendix B, Section B.1.3) is exposed to this same level of formaldehyde flux using an inspiratory rate of 15 L/min. EPA also applied a second CFD model (Schroeter et al., 2014) to estimate formaldehyde flux at different sites (e.g., squamous epithelium) in the rat nasal cavity. Unlike the Kimbell et al. (2001) model, this revised CFD model considers endogenous formaldehyde production in nasal tissues. EPA estimated the extent to which the results change if flux estimates from Schroeter et al. (2014) are used; See, for example, EPA Summary Table 2-5, “Summary of Derivation of the Point of Departure (POD) for Squamous Metaplasia Based on Observations in F344 Rats,” from Kerns et al. (1983).
Finding: The use of these models is appropriate and consistent with EPA’s state-of-practice methods.
Recommendation 3.1 (Tier 2): To enhance transparency, the summary tables (i.e., Tables 24, 43, and 44) should explicitly identify the models used to derive flux values. Table 44 should clearly indicate whether the BBDR models used here are equivalent to Models 1 and 2 identified in Table 43 and the text. Table 46 should indicate which flux model was used.
EPA also used an alternative method to derive cRfCs for respiratory tract pathology seen in animals (see Chapter 4). This alternative method relied on an assumption of concentration equivalence and was applied to lesion data from Wistar rats (Woutersen et al., 1989). EPA used it because CFD models have not been developed for this strain of rat. In this method, allometric principles are applied. EPA applied additional duration adjustments (6/24) × (5/7) for continuous daily exposure to generate the final human equivalent concentration (HEC). EPA provides extensive discussion of alternative models (e.g., Corley et al., 2015) it considered, and provides its rationale for not using these alternative models in the assessment.
EPA used the available models to derive cRfCs for nasal cancers seen in animals (see Chapter 5). These cRfCs were estimated based on points of departure (PODs) obtained from a pathology study of hyperplasia, labeling studies of proliferating cells, and BBDR modeling results using the Conolly et al. (2003) model (see Section 2.2.1 of the 2022 Draft Assessment). URT cancer risk was extrapolated from the incidence of nasal squamous cell carcinoma in experimental studies
performed on F344 rats. EPA evaluated and compared results from several methods used to model these data, including a BBDR model, statistical time-to-tumor models, and statistical benchmark dose modeling using data on DPCs and formaldehyde flux as dose metrics. Additional analyses and comparisons considered the impact of endogenous formaldehyde concentration on dosimetric estimates. EPA considered these efforts to be secondary or supportive calculations since human data were available. EPA evaluated the BBDR model that includes a CFD component (Conolly et al., 2003) for extrapolation of the rat nasal cancer risk to human exposure scenarios. EPA’s analysis of this model led to the conclusion that the BBDR model for humans (Conolly et al., 2004) was not robust at any formaldehyde exposure concentration. EPA was also concerned that this model presumes that formaldehyde-induced mutagenicity, modeled as proportional to DPC concentration, is not relevant to formaldehyde’s carcinogenicity. EPA used the rat BBDR model (Conolly et al., 2003) to derive multiple PODs and corresponding HECs, but not to extrapolate to human exposure scenarios. DPC tissue concentrations used in the rat BBDR model (Conolly et al., 2003) were calculated using a PBPK model (Conolly et al., 2000). A second set of CFD and PBPK models (Schlosser et al., 2003) were used to predict formaldehyde flux and DPC concentrations in the rat and human nasal cavities. These alternative model estimates are provided in the Assessment Overview of the 2022 Draft Assessment (Table 2-22, “Benchmark Concentrations and Human Equivalents Using Formaldehyde Flux and DPC as Dose Metrics”). As mentioned earlier for squamous metaplasia of the nasal cavity, EPA employed a similar method using CFD models (Kimbell et al., 2001; Schroeter et al., 2014) to predict formaldehyde flux to the rostral portion of the rat nasal cavity (see EPA Summary Tables 2-21 and 2-22).
Finding: The 2022 Draft Assessment generally provides a thorough discussion of the strengths and weaknesses of the available models, including an extensive discussion of how EPA used them in the assessment.
Recommendation 3.2 (Tier 2): To increase transparency, EPA should provide additional clarification regarding its decision not to use the BBDR model to extrapolate rat nasal carcinogenicity to humans. Criteria used by EPA to determine whether the models would be adequately robust for this purpose are not readily available in the 2022 Draft Assessment. Likewise, EPA should provide additional support for its decision not to use the BBDR model (Conolly et al., 2004) for this extrapolation because of the model’s conclusion that formaldehyde-induced mutagenicity, modeled as proportional to DPC concentration, is not relevant to formaldehyde’s carcinogenicity.
Finding: Documentation of the dosimetry methods used is variable in the 2022 Draft Assessment, especially with regard to some summary tables (e.g., Table 2-27). Discussion of the comparison of and basis for unit risk estimates for nasopharyngeal cancer in humans and nasal squamous cell carcinomas in rats mentions the use of CFD and PBPK models without identifying the specific models that were used.
Recommendation 3.3 (Tier 2): To increase transparency, EPA should address these shortcomings by updating tables and text to better document its dosimetry methods.
REFERENCES
Battelle Columbus Laboratories. 1982. A chronic inhalation toxicology study in rats and mice exposed to formaldehyde. Research Triangle Park, NC: Chemical Industry Institute of Toxicology.
Campbell Jr., J. L., P. R. Gentry, H. J. Clewell III, and M. E. Andersen. 2020. A kinetic analysis of DNA-deoxy guanine adducts in the nasal epithelium produced by inhaled formaldehyde in rats-Assessing contributions to adduct production from both endogenous and exogenous sources of formaldehyde. Toxicological Sciences 177(2):325–333.
Casanova, M., H. D. Heck, J. I. Everitt, W. W. Harrington Jr., and J. A Popp. 1988. Formaldehyde concentrations in the blood of rhesus monkeys after inhalation exposure. Food and Chemical Toxicology 26(8):715–716.
Chang, J. C., and C. S. Barrow. 1984. Sensory irritation tolerance and cross-tolerance in F-344 rats exposed to chlorine or formaldehyde gas. Toxicology and Applied Pharmacology. 76(2):319–27. https://doi:10.1016/0041-008x(84)90013-9.
Chang, J. C., W. H. Steinhagen, and C. S. Barrow. 1981. Effect of single or repeated formaldehyde exposure on minute volume of B6C3F1 mice and F-344 rats. Toxicology and Applied Pharmacology 61, no. 3: 451–459. https://10.1016/0041-008x(81)90368-9.
Conolly, R. B., J. S. Kimbell, D. Janszen, P. M. Schlosser, D. Kalisak, J. Preston, and F. J. Miller. 2003. Biologically motivated computational modeling of formaldehyde carcinogenicity in the F344 rat. Toxicological Sciences 75(2):432–447.
Conolly, R. B., J. S. Kimbell, D. Janszen, P. M. Schlosser, D. Kalisak, J. Preston, and F. J. Miller. 2004. Human respiratory tract cancer risks of inhaled formaldehyde: Dose-response predictions derived from biologically-motivated computational modeling of a combined rodent and human dataset. Toxicological Sciences 82(1):279–296.
Conolly, R. B., P. D. Lilly, and J. S. Kimbell. 2000. Simulation modeling of the tissue disposition of formaldehyde to predict nasal DNA-protein cross-links in Fischer 344 rats, rhesus monkeys, and humans. Environmental Health Perspectives 108(Suppl 5):919–924.
Corley, R. A., S. Kaliban, A. P. Kuprat, J. P. Carson, R. E. Jacob, K. R. Minard, J. G. Teeguarden, C. Timchalk, S. Pipavath, R. Glenny, and D. R. Einstein. 2015. Comparative risks of aldehyde constituents in cigarette smoke using transient computational fluid dynamics/physiologically based pharmacokinetic models of the rat and human respiratory tracts. Toxicological Sciences 146(1):65–88.
Edrissi, B., K. Taghizadeh, and P. C. Dedon. 2013a. Quantitative analysis of histone modifications: Formaldehyde is a source of pathological n(6)-formyllysine that is refractory to histone deacetylases. PLoS Genetics 9(2):e1003328.
Edrissi, B., K. Taghizadeh, B. C. Moeller, D. Kracko, M. Doyle-Eisele, J. A. Swenberg, and P. C. Dedon. 2013b. Dosimetry of N(6)-formyllysine adducts following [(1)(3)C(2)H(2)]-formal-dehyde exposures in rats. Chemical Research in Toxicology 26(10):142–143.
EPA (U.S. Environmental Protection Agency). 2022. IRIS Toxicological Review of Formaldehyde-Inhalation, External Review Draft. Washington, DC. https://iris.epa.gov/Document/&deid=248150 (accessed September 18, 2023).
Heck, H. D., M. Casanova-Schmitz, P. B. Dodd, E. N. Schachter, T. J. Witek, and T. Tosun. 1985. Formaldehyde (CH2O) concentrations in the blood of humans and Fischer-344 rats exposed to CH2O under controlled conditions. American Industrial Hygiene Association Journal 46(1):1–3.
Hubal, E. A., P. M. Schlosser, R. B. Connelly, and J. S. Kimbell. 1997. Comparison of inhaled formaldehyde dosimetry predictions with DNA-protein cross-link measurements in the rat nasal passages. Toxicology and Applied Pharmacology 143(1):47–55.
Kerns, W. D., K. L. Pavkov, D. J. Donofrio, E. J. Gralla, and J. A. Swenberg. 1983. Carcinogenicity of formaldehyde in rats and mice after long-term inhalation exposure. Cancer Research 43(9):4382-4392.
Kimbell, J. S., R. P. Subramaniam, E. A. Gross, P. M Schlosser, and K. T. Morgan. 2001. Dosimetry modeling of inhaled formaldehyde: Comparisons of local flux predictions in the rat, monkey, and human nasal passages. Toxicological Sciences 64(1):100–110.
Klein, M. D., B. K. Sinha, and R. P. Subramaniam. 2011. Statistical inferences from formaldehyde DNA-protein cross-link data: Improving methods for characterization of uncertainty. Journal of Biopharmaceutical Statistics 21(1):42–55.
Kleinnijenhuis, A. J., Y. C. Staal, E. Duistermaat, R. Engel, and R. A. Woutersen, 2013. The determination of exogenous formaldehyde in blood of rats during and after inhalation exposure. Food and Chemical Toxicology 52:105–112.
Lai, Y., R. Yu, H. J. Hartwell, B. C. Moeller, W. M. Bodnar, and J. A. Swenberg. 2016. Measurement of endogenous versus exogenous formaldehyde-induced DNA-protein crosslinks in animal tissues by stable isotope labeling and ultrasensitive mass spectrometry. Cancer Research 76(9):2652–2661.
Leng, J., C. W. Liu, H. J. Hartwell, R. Yu, Y. Lai, W. M. Bodnar, K. Lu, and J. A. Swenberg. 2019. Evaluation of inhaled low-dose formaldehyde-induced DNA adducts and DNA-protein cross-links by liquid chromatography-tandem mass spectrometry. Archives of Toxicology 93(3):763–773.
Lu, K., L. B. Collins, H. Ru, E. Burmdez, and J. A. Swenberg. 2010a. Distribution of DNA adducts caused by inhaled formaldehyde is consistent with induction of nasal carcinoma but not leukemia. Toxicological Sciences 116(2):441–451.
Lu, K., W. Ye, L. Zhou, L. B. Collins, X. Chen, A. Gold, L. M. Ball, and J. A. Swenberg. 2010b. Structural characterization of formaldehyde-induced cross-links between amino acids and deoxynucleosides and their oligomers. Journal of the American Chemical Society 132(10):3388–3399.
Lu, K., B. Moeller, M. Doyle-Eisele, J. McDonald, and J. A. Swenberg. 2011. Molecular dosimetry of N2-hydroxymethyl-dG DNA adducts in rats exposed to formaldehyde. Chemical Research in Toxicology 24(2):159–161.
Moeller, B. C., K. Lu, M. Doyle-Eisele, J. McDonald, A. Gigliotti, and J. A. Swenberg. 2011. Determination of N2-hydroxymethyl-dG adducts in the nasal epithelium and bone marrow of nonhuman primates following 13CD2-formaldehyde inhalation exposure. Chemical Research in Toxicology 24(2):162–164.
NRC (National Research Council). 2011. Review of the Environmental Protection Agency’s draft IRIS assessment of formaldehyde. Washington, DC: The National Academies Press.
Overton, J. H. 2001. Dosimetry modeling of highly soluble reactive gases in the respiratory tract. Inhalation Toxicology 13(5):347–357.
Overton, J. H., J. S. Kimbell, and F. J. Miller. 2001. Dosimetry modeling of inhaled formaldehyde: The human respiratory tract. Toxicological Sciences 64(1):122–134.
Schlosser, P. M., P. D. Lilly, R. B. Connelly, D. B. Janszen, and J. S. Kimbell. 2003. 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 Analysis 23(3):473–487.
Schroeter, J. D., J. Campbell, J. S. Kimbell, R. B. Connelly, H. J. Clewell, and M. E. Andersen. 2014. Effects of endogenous formaldehyde in nasal tissues on inhaled formaldehyde dosimetry predictions in the rat, monkey, and human nasal passages. Toxicological Sciences 138(2):412–424.
Subramaniam, R. P., C. Chen, K. S. Crump, D. Devoney, J. F. Fox, C. J. Portier, P. M. Schlosser, C. M. Thompson, and P. White. 2008. Uncertainties in biologically-based modeling of formaldehyde-induced respiratory cancer risk: Identification of key issues. Risk Analysis 28(4):907–923.
Woutersen, R. A., A. van Garderen-Hoetmer, J. P. Bruijntjes, A. Zwart, and V. J. Feron. 1989. Nasal tumours in rats after severe injury to the nasal mucosa and prolonged exposure to 10 ppm formaldehyde. Journal of Applied Toxicology 9(1):39–46.
Yu, R., Y. Lai, H. J. Hartwell, B. C. Moeller, M. Doyle-Eisele, D. Kracko, W. M. Bodnar, T. B. Starr, and J. A. Swenberg. 2015. Formation, accumulation, and hydrolysis of endogenous and exogenous formaldehyde-induced DNA damage. Toxicological Sciences 146(1):170–182.
