A nonlinear dosimetric model for hemoglobin adduct formation by the neurotoxic agent acrylamide and its genotoxic metabolite glycidamide.

Hemoglobin (Hb) adducts, formed by the neurotoxic agent acrylamide (AA) and its genotoxic metabolite glycidamide (GA), were measured in the rat by means of a method for simultaneous determination of the adducts formed to cysteine. A novel, nonlinear dosimetric model was developed to describe Hb adduct formation. This model incorporates the saturable kinetics of the metabolic conversion in vivo of AA to GA. The pharmacokinetic parameters Vmax and Km and the first-order rates of elimination, k1 and k2, for AA and GA from all processes except conversion of AA to GA, were estimated directly from Hb adduct data to 19 M hr-1, 66 microM, 0.21 hr-1, and 0.48 hr-1, respectively. At low concentrations, approximately 60% of AA was metabolized to GA. The nonlinear dosimetric model for adduct formation has potential general applicability in high-to-low-dose extrapolation of genotoxic effects.

The development of analytical methods for quantitating adducts formed by electrophilic substances with macromolecules such as hemoglobin (Hb) and DNA has made significant progress over the past decades, and such adducts have been proposed as biomarkers for use in epidemiological studies aiming to monitor human exposure to toxic agents. In contrast, mathematical models for adduct formation, crucial for the use ofthese biomarkers for human risk estimation, have been relatively unsophisticated and essentially limited to the linear model. Our work with Hb adduct formation by the neurotoxic agent acrylamide (AA) and its genotoxic metabolite glycidamide (GA) prompted us to develop a nonlinear mathematical model for adduct formation applicable also at the high exposures where capacity-limited processes come into play and toxic effects are commonly observed in experimental organisms. Humans are exposed to AA mainly in the chemical industry, molecular biology laboratories, and grouting operations (1). Acrylamide has for a long time been known to be neurotoxic in animals and man, and more recently its genotoxic and cancerinitiating properties have attracted increasing attention. Despite ' the fact that AA is carcinogenic in mice when followed by treatment with phorbol esters and is mutagenic in mammalian organisms, it is negative in the Ames test both in the absence and in the presence ofS9 mix. Although epidemiological studies have not demonstrated increased rates oftumors in exposed workers, it is classified by the International Agency for Research on Cancer as a possible human carcinogen based on animal bioassays in the mouse and in the rat (2).
Because human exposure to AA in most occupational settings is believed to occur through a combination ofdermal and inhalational routes (1), there is a need for a biomarker of exposure, such as Hb adducts, that is indicative ofthe total exposure. Initial work on the use ofadducts formed by cysteine residues in Hb as biomarkers for AA exposure was performed by Bailey et al. (3), who developed a mass fragmentographic technique to determine such adducts in rats given 0-5 mg AA/kg body weight. In this range of concentrations, Hb adduct formation was found to be a convex function of the injected concentration. This unusual nonlinear relationship between Hb adduct formation and injected concentration, as well as several studies showing an impact ofphenobarbital induction on the neurotoxicity of AA in rats, indicated to us that AA underwent further metabolism in a cytochrome P-450-mediated reaction. In an attempt to confirm the hypothesis that AA was metabolically oxidized to the epoxide GA, we analyzed Hb hydrolysates from rats treated with AA and were able to identify S-(2-carboxy-2-hydroxyethyl)-cysteine, the adduct formed by GA, by means of GC-MS (4) (Fig. 1).
In contrast, adduct formation by either AA or its metabolite GA in rats treated with an initial concentration, [AA]o, of AA in the range 0-100 mg/kg body weight was inconsistent with a linear model. Thus, while adduct formation by AA was convex (3), approaching linearity at high concentrations, adduct formation by°G A was a concave function of the injected concentration, presumably reflecting a saturable Michaelis-Menten type of kinetics for the metabolic conversion of AA to GA. The following kinetic scenarios for AA and GA, respectively, were thus assumed: To evaluate the role of GA in the induction of toxic effects associated with AA exposure, a series of experiments was undertaken to study its toxicological properties and quantitative relevance as a metabolite. It was thus found that AA itself was primarily responsible for causing peripheral neuropathy as judged from the inability of GA to cause neuropathological damage to sciatic nerves or have an effect in the hindlimb splay test in treated animals (5, Deng et al., in preparation). On the other hand, GA had a strong impact on body weight, sperm cell viability, and epididymal daily sperm production in male rats (5), implicating a role of GA in the induction of male reproductive toxicity in rodents given AA. Already prior to its discovery as a metabolite ofAA, GA had been shown to give a positive response in the Ames test (6). It was thus hypothesized (4) that GA was the agent responsible for the cancer-initiating properties associated with AA exposure in experimental animals.
To be able to measure tissue doses (7), DAA and DGA, of both AA and GA, for risk assessment of neurotoxic (DAA) or reproductive and genotoxic effects (DGA), respectively, we developed a GC-MS technique that allows us to simultaneously determine Hb adducts to cysteine formed by both electrophiles (8) (Fig. 1). Using this techique, the concentration of adducts formed by GA-(GAHb) in rats was found to be linearly dependent Equations 5 and 6 describe tissue doses and are thus proportional to Hb adduct formation (Equation 1) by a parent compound or a metabolite, respectively, undergoing a saturable process. As capacity-limited processes are commonplace in the metabolism oftoxic agents and Equations S and 6are useful also for modeling DNA adduct formation, they merit consideration for incorporation into general models for high-to-low-dose extrapolation of risks associated with exposure to genotoxic/cancer-initiating agents.
In summary, we have developed a method to simultaneously determine Hb adduct formation by AA and its genotoxic metabolite GA in treated rats. Based on the observed nonlinearities in the Hb adduct formation by the two agents, we have developed a novel mathematical model (10) that describes adduct formation in vivo by agents undergoing saturable metabolic processes and is potentially useful for high-to-low-dose extrapolation of genotoxic effects. By means ofthis model the values ofthe pharmacokinetic parameters Vmax, Kin, k,, and k2 were estimated in the rat. Studies are now under way to determine Hb adducts by AA and GA in humans exposed to AA industrially and in molecular biology laboratories as a basis for a quantitative risk estimation of this substance. This manuscript was presented as a poster at the Conference on Biomonitoring and Susceptibility Markers in Human Cancer: Applications in Molecular Epidemiology and Risk Assessment that was held in Kailua-Kona, Hawaii, 26 October-l November 1991.
This work was financially supported by Cooperative Agreement CR-816768 with the U.S. EPA. Although the research described in this article has been supported by the U.S. Environmental Protection Agency, it has not been subjected to Agency review and therefore does not necessarily reflect the views of the Agency, and no official endorsement should be inferred.