Methylene Chloride Toxicology and Cancer Risk

RoseyFields evaluates the toxicology of Methylene Chloride using the following evidence base, drawn from the primary peer-reviewed animal toxicology, mechanistic, and human epidemiological literature. This summary synthesizes findings on toxicokinetics, mode of action, organ-system toxicity, carcinogenicity, and the concordance between animal and human data, and identifies remaining data gaps relevant to chemical-specific risk characterization.

1. Chemical Identity and Physicochemical Properties Relevant to Toxicity

Methylene chloride (dichloromethane, DCM; CASRN 75-09-2) is a low-molecular-weight chlorinated aliphatic hydrocarbon (CH2Cl2) known industrially by synonyms including methylene bichloride, methylene dichloride, Aerothene MM, and Solaesthin. It is a small, volatile, moderately lipophilic solvent that is liquid at room temperature and evaporates readily, properties that govern both its principal route of human exposure (inhalation, with additional potential for oral and dermal contact) and its toxicological behavior.

DCM does not persist in biological tissues and is not considered bioaccumulative; its volatility and relatively rapid metabolic clearance limit tissue retention compared with more lipophilic, poorly metabolized halogenated solvents. Its toxicological reactivity is not intrinsic to the parent compound but arises almost entirely through metabolic biotransformation. Two competing hepatic and extrahepatic metabolic pathways convert DCM into chemically reactive intermediates: an oxidative pathway mediated by cytochrome P450 2E1 (CYP2E1) that yields carbon monoxide (CO) via a formyl chloride intermediate, and a conjugative pathway mediated by glutathione S-transferase theta-1 (GST-T1) that yields a DNA-reactive formaldehyde-glutathione adduct implicated in mutagenicity (Andersen et al., 1991; Green and Graves, 1996; DeMarini et al., 1997). The relative activity of these two pathways differs by species, dose, and tissue, and this toxicokinetic behavior — rather than any direct chemical reactivity of DCM itself — is the central determinant of its toxicity profile.

2. Toxicokinetics: Absorption, Distribution, Metabolism, and Excretion

Absorption and Distribution

DCM is well absorbed following oral, inhalation, and dermal exposure. Once absorbed, it distributes systemically and crosses the placental barrier, entering fetal circulation in rodent models (Anders and Sunram, 1982; Withey and Karpinski, 1985). Because DCM and its principal oxidative metabolite, CO, both cross the placenta, gestational exposure raises the possibility of fetal carboxyhemoglobinemia in addition to direct fetal exposure to the parent compound.

Metabolism

DCM is metabolized through two parallel pathways. The first is an oxidative pathway catalyzed by CYP2E1, generating carbon monoxide and elevating blood carboxyhemoglobin (COHb); this pathway predominates at lower exposure concentrations and is capacity-limited (saturable) at higher doses. The second is a conjugative pathway catalyzed by GST-T1, which produces a reactive, DNA-damaging intermediate and becomes proportionally more important at higher exposure levels once the oxidative pathway saturates (Andersen et al., 1991). Physiologically based pharmacokinetic (PBPK) models developed for the rat (Andersen et al., 1991) and later adapted and refined for mice and humans (Marino et al., 2006; David et al., 2006) have been used to translate applied external doses into internal target-tissue dose metrics — specifically, the rate of DCM metabolized via the GST pathway per liter of liver or lung tissue per day — which form the basis for cross-species extrapolation of both noncancer and cancer effects.

A key toxicokinetic feature relevant to interpreting animal bioassay data is the marked species difference in GST-T1-mediated metabolic capacity. Human hepatic and pulmonary tissues express substantially lower GST-T1 activity than do the tissues of mice, which are the most sensitive species in DCM carcinogenicity bioassays; this difference in metabolic capacity — rather than a difference in the intrinsic hazard of the reactive intermediate — is believed to underlie much of the disparity between rodent tumor incidence and the epidemiological signal in exposed workers (Sherratt et al., 2002). Sex differences in DCM metabolism have also been noted: in some rat oral bioassays, males and females showed differing patterns of hepatocellular tumor response despite receiving comparable doses, and mammary gland tumor responses to inhaled DCM occurred predominantly in females (Nitschke et al., 1988a; Burek et al., 1984).

Excretion

Because DCM’s principal oxidative metabolite is carbon monoxide, a substantial fraction of absorbed dose is eliminated via the respiratory route as CO (measurable indirectly as elevated blood carboxyhemoglobin) or as exhaled unmetabolized parent compound. Blood COHb levels rise with increasing exposure concentration in a dose-related manner but do not increase progressively with repeated-exposure duration, indicating that CO generation reaches a steady state without bioaccumulation over chronic exposure (Nitschke et al., 1988a). The GST pathway, by contrast, yields metabolites that are further processed intracellularly and do not contribute measurably to whole-body elimination kinetics in the same way, but are of primary toxicological interest because of their DNA reactivity.

3. Mechanism(s) of Action / Mode of Action

DCM toxicity operates through two largely distinct, dose- and pathway-dependent mechanisms tied to its two metabolic routes.

The CYP2E1 oxidative pathway generates carbon monoxide, which binds hemoglobin to form carboxyhemoglobin, reducing oxygen-carrying capacity. This mechanism accounts for the neurological and other systemic effects reported in acutely and occupationally exposed humans and is consistent with the reversible neurobehavioral effects (impaired attention and reaction time) described in retired workers with historical exposures in the range of 100–200 ppm (Lash et al., 1991) and in earlier occupational cohorts (Cherry et al., 1981).

The GST-T1 conjugative pathway is central to DCM’s carcinogenic mode of action. Metabolism via GST-T1 produces a short-lived, highly reactive intermediate capable of forming DNA adducts and inducing mutations. Supporting mechanistic evidence includes: enhanced bacterial mutagenicity of DCM in Salmonella strains engineered to express mammalian GST activity, with an absence of mutagenicity in GST-deficient strains (Thier et al., 1993; Oda et al., 1996; Simula et al., 1993); positive mutagenicity in the CHO/HPRT mammalian cell assay when GST activity is present (Graves and Green, 1996; Graves et al., 1996); and dose-related increases in DNA single-strand breaks, sister chromatid exchange, and micronucleus formation in mouse liver and lung tissue following in vivo inhalation exposure (Allen et al., 1990; Sasaki et al., 1998; Graves et al., 1994b, 1995). Notably, these genotoxic effects were tissue-specific, occurring predominantly in liver and lung — the two sites of tumor formation in mouse bioassays — while assays of bone marrow and other tissues with lower GST-T1 activity were largely negative (Westbrook-Collins et al., 1990; Sheldon et al., 1987; Gocke et al., 1981), consistent with a tissue-localized mutagenic mechanism rather than systemic genotoxicity. Immunohistochemical work further localized GST-T1 expression to bile duct epithelial cell nuclei in human liver tissue, providing a plausible mechanistic link to the biliary tract cancer signal observed in some occupational cohorts (Sherratt et al., 2002).

Because the reactive GST-T1 metabolite is short-lived, mechanistic data support the hypothesis that it exerts its genotoxic effects locally, within the tissue where it is generated, rather than after systemic distribution — a consideration that shaped the choice of tissue-specific (rather than whole-body) internal dose metrics in the PBPK-based dose-response modeling used to derive cancer risk estimates from the mouse bioassay data (Marino et al., 2006; David et al., 2006).

4. Animal Toxicological Evidence by Organ System / Health Outcome

Hepatic Effects

The liver is the most consistently affected noncancer target organ across species and routes of exposure. In a chronic 2-year drinking-water bioassay in F344 rats (85/sex/dose, plus 135 controls) at mean doses of 0–235 (males) and 0–263 (females) mg/kg-day, dichloromethane produced a dose-related increase in hepatic foci/areas of cellular alteration, first observed after 78 weeks of treatment and progressing through week 104; incidences were significantly elevated at all but the lowest dose group, with a NOAEL of 5 mg/kg-day and LOAEL of 50 mg/kg-day (Serota et al., 1986a). Reductions in several clinical chemistry parameters (alkaline phosphatase, creatinine, blood urea nitrogen, total protein, cholesterol) were also observed at the highest dose in both sexes. A supporting 90-day subchronic drinking-water study in F344 rats and mice found significant increases in hepatocyte vacuolation and necrosis at doses between roughly 166 and 1,469 mg/kg-day depending on sex (Kirschman et al., 1986).

By the inhalation route, a 2-year bioassay in Sprague-Dawley rats (90/sex/dose) exposed to 0, 50, 200, or 500 ppm DCM (6 hours/day, 5 days/week) found statistically significant increases in nonneoplastic hepatic lesions — lipid vacuolation and multinucleated hepatocytes — in females at 500 ppm, with a non-significant trend toward vacuolation in males at the same concentration (Nitschke et al., 1988a). A comparable inhalation bioassay in Sprague-Dawley rats exposed to 500–3,500/3,000 ppm reported hepatic vacuolation correlated with fatty change at concentrations of 500 ppm and above, corroborating the Nitschke findings across a broader exposure range (Burek et al., 1984). Hepatic steatosis has also been described in dogs exposed to DCM (Haun et al., 1971).

Neurological Effects

Gross neurological impairment has not been observed in chronic rodent inhalation bioassays at concentrations as high as 4,000 ppm (Mennear et al., 1988; Nitschke et al., 1988a; Burek et al., 1984). A dedicated subchronic neurotoxicity study in F344 rats exposed to concentrations up to 2,000 ppm for 13 weeks found no treatment-related effects on an observational battery, hind-limb grip strength, evoked potentials, or histopathology of brain, spinal cord, or peripheral nerve tissue when tests were conducted 65 hours after the final exposure (Mattsson et al., 1990). A multidisciplinary neurobehavioral screening study in female F344 rats did report some altered neurological function measures (Moser et al., 1995), although standardized batteries in the chronic bioassays did not detect overt neurological impairment.

Reproductive and Developmental Effects

A one-generation oral reproductive toxicity study in Charles River CD rats found no effects on fertility index, litter size, or pup survival (General Electric Company, 1976), and an oral developmental toxicity study in F344 rats found no effects on pup survival, resorption rates, or pup weight at maternal doses up to 450 mg/kg-day on gestation days 6–19, a dose sufficient to depress maternal weight gain (Narotsky and Kavlock, 1995). By inhalation, a two-generation reproductive study in F344 rats found no effects on fertility index, litter size, neonatal survival, growth rates, or reproductive histopathology at concentrations up to 1,500 ppm, though exposure was not continuous throughout gestation and nursing, limiting its sensitivity (Nitschke et al., 1988b). A separate inhalation study of male Swiss-Webster mice exposed for 6 weeks prior to mating reported a reduction in fertility index at 150 and 200 ppm, though the statistical significance of this finding varied depending on the analytical method used (Raje et al., 1988).

Single-dose developmental inhalation studies used relatively high exposure concentrations. Long-Evans rats exposed to 4,500 ppm for 14 days prior to mating and during gestation (or during gestation alone) showed decreased offspring birth weight and altered behavioral habituation to novel environments (Bornschein et al., 1980; Hardin and Manson, 1980). Swiss-Webster mice and Sprague-Dawley rats exposed to 1,250 ppm on gestation days 6–15 showed increased delayed ossification of the sternebrae but no major teratogenic effects (Schwetz et al., 1975). No oral or inhalation studies incorporating dedicated neurobehavioral testing batteries in offspring were identified, representing a recognized data gap given DCM’s placental transfer and the relatively high CYP2E1 activity in the developing human fetal brain (Withey and Karpinski, 1985; Anders and Sunram, 1982; Hines, 2007; Johnsrud et al., 2003; Brzezinski et al., 1999).

Immunological Effects

A 4-week inhalation study in rats found no effect of DCM exposure, at concentrations up to 5,000 ppm, on antibody responses to sheep red blood cells, suggesting an absence of systemic immunosuppression under repeated exposure conditions (Warbrick et al., 2003). In contrast, a single 3-hour acute inhalation exposure to 100 ppm DCM in CD-1 mice produced evidence of pulmonary immunosuppression using a functional host-defense assay measuring susceptibility to Streptococcal pneumonia and clearance of Klebsiella bacteria (Aranyi et al., 1986); a related methodological approach applied to other chlorinated solvents supports the validity of this assay (Selgrade and Gilmour, 2010). The apparent discordance between the acute, localized pulmonary immunosuppression seen after a single high-concentration exposure and the absence of systemic immune effects after weeks of exposure suggests a portal-of-entry effect in the lung rather than a systemic immunotoxic mechanism, though chronic low-dose immunotoxicity data remain limited.

Hematological Effects

In the chronic oral rat bioassay, mean hematocrit, hemoglobin, and red blood cell counts were increased in both sexes at dietary/drinking-water doses of 50 mg/kg-day and above after 52 and 78 weeks of treatment (Serota et al., 1986a), a pattern consistent with a compensatory erythropoietic response to chronic low-level carboxyhemoglobinemia from CYP2E1-mediated CO generation.

Renal and Endocrine Effects

No treatment-related renal histopathological changes or significant urinary abnormalities were observed in the chronic oral rat bioassay (Serota et al., 1986a). No significant effects on organ weights, clinical chemistry, or plasma hormone levels were found in the chronic inhalation rat bioassay (Nitschke et al., 1988a), and no dedicated endocrine-toxicity studies specific to DCM were identified in the reviewed database.

5. Animal Carcinogenicity Bioassay Findings

DCM carcinogenicity has been evaluated in multiple chronic bioassays across rats, mice, and hamsters, by both the oral and inhalation routes.

Oral (Drinking Water) Bioassays

In male B6C3F1 mice exposed to DCM in drinking water for 2 years at doses up to approximately 234 mg/kg-day, statistically significant increases in combined hepatocellular adenomas and carcinomas were observed at all but the lowest dose, with a marginally significant overall trend test (Serota et al., 1986b; Hazleton Laboratories, 1983). Female mice in the same study did not show increased proliferative hepatocellular lesions. In the parallel F344 rat drinking-water bioassay, no increase in liver tumors was seen in males, and the pattern of hepatocellular carcinoma/neoplastic nodule incidence in females was irregular (non-monotonic) across dose groups, leading the study authors to conclude the rat data did not support a treatment-related increase in liver tumor incidence (Serota et al., 1986a).

Inhalation Bioassays

Chronic inhalation exposure of B6C3F1 mice to 2,000 or 4,000 ppm DCM produced clear, concentration-related increases in both hepatocellular tumors and bronchoalveolar (lung) tumors in male and female mice (Mennear et al., 1988; NTP, 1986); lung tumor incidence was elevated even after only 26 weeks of exposure followed by an extended exposure-free period, and liver and lung tumor elevations were also seen after just 52 weeks of continuous exposure (Kari et al., 1993; Maronpot et al., 1995). In F344 rats exposed by inhalation to concentrations of 50–4,000 ppm across several independent bioassays, no consistent increase in liver tumors was observed, but statistically significant increases in benign mammary tumors (adenomas/fibroadenomas) occurred in females, with corroborating findings across studies conducted at 50–500 ppm (Nitschke et al., 1988a) and 500–3,500 ppm (Burek et al., 1984); low rates of mammary-region sarcoma/fibrosarcoma were also noted in males in two studies (Nitschke et al., 1988a; NTP, 1986). The Nitschke et al. (1988a) rat bioassay additionally reported an increased incidence of astrocytoma/glioma (mixed glial cell) tumors, a rare tumor type in rats, providing some biological plausibility for the brain cancer signal later observed in occupational epidemiology. In Syrian golden hamsters exposed to concentrations as high as 3,500 ppm for 2 years, no statistically significant increase in tumor incidence at any site was observed (Burek et al., 1984), indicating a marked species difference in tumor susceptibility consistent with lower hamster hepatic GST-T1 activity.

Gavage Studies

A gavage bioassay in Sprague-Dawley rats and Swiss mice, terminated early (at 64 weeks) due to high treatment-related mortality, found a non-statistically-significant increase in malignant mammary adenocarcinomas in high-dose (500 mg/kg-day) female rats and a non-statistically-significant, dose-related increase in pulmonary adenomas in male mice (Maltoni et al., 1988); the early termination and high mortality limit the interpretability of this study.

6. Comparison of Animal Findings to Human Epidemiological Data

Human occupational cohort and case-control studies provide evidence of associations between DCM exposure and several of the cancer sites implicated in the animal bioassays, though generally with weaker and less precise effect estimates. Two long-follow-up occupational cohorts reported modestly elevated standardized mortality ratios for brain cancer, with wide confidence intervals reflecting small numbers of cases (Tomenson, In Press; Hearne and Pifer, 1999), and a case-control study using detailed lifetime job-history exposure reconstruction found statistically significant positive trends with increasing exposure probability, duration, and intensity, and a markedly elevated odds ratio in workers with both high-intensity and long-duration (>20 years) exposure (Heineman et al., 1994). A separate case-control study of female brain cancer cases using less detailed occupational data found only a weak overall association with no clear exposure-response trend (Cocco et al., 1999).

For liver and biliary tract cancer, the strongest signal comes from a triacetate fiber production cohort with comparatively high historical exposures, which reported an elevated standardized mortality ratio based on a small number of cases, predominantly biliary tract cancers, with the estimate becoming less extreme (though still elevated) as follow-up time lengthened and the number of expected cases increased (Lanes et al., 1990, 1993). No other cohort with comparable exposure intensity has been identified, limiting confidence in this finding. Three case-control studies conducted in Germany, Italy, and the United States found consistent, though individually imprecise, elevated risks of non-Hodgkin lymphoma associated with solvent exposure that included DCM (Seidler et al., 2007; Miligi et al., 2006; Barry et al., 2011; Wang et al., 2009), and an additional study reported an association with multiple myeloma (Gold et al., 2011).

Several factors likely contribute to the qualitative but imprecise concordance between rodent bioassay findings and human epidemiological signals. First, the rare tumor types implicated (brain cancer, biliary tract cancer, certain hematopoietic malignancies) are inherently difficult to study with adequate statistical power in occupational cohorts of modest size, and most available studies report wide confidence intervals spanning or approaching the null. Second, toxicokinetic differences in GST-T1 metabolic capacity — substantially lower in human liver and lung tissue than in mouse tissue, the most sensitive species in the bioassays — mean that a given external exposure concentration produces a proportionally smaller internal dose of the DNA-reactive metabolite in humans than in mice, providing a biologically plausible explanation for why the human epidemiological signal, while directionally consistent with a mutagenic carcinogen, is weaker in magnitude than the mouse bioassay results would suggest on an external-dose basis (Sherratt et al., 2002). This species difference is precisely why cross-species extrapolation in DCM risk assessment relies on PBPK-model-derived internal (metabolized) dose metrics rather than simple external-dose scaling (Andersen et al., 1991; Marino et al., 2006; David et al., 2006). Third, exposure misclassification is a recurring limitation in the human studies, particularly those relying on death-certificate-based occupational classification, which would be expected to bias associations toward the null (Cantor et al., 1995; Rothman and Greenland, 1998). Confounding by co-exposure to other solvents is also a concern in some cohorts, such as one aircraft-maintenance cohort in which DCM-associated risks for breast cancer were of similar magnitude to associations with other co-occurring solvents (Radican et al., 2008).

For noncancer endpoints, occupational studies of neurological function among DCM-exposed workers (Cherry et al., 1981; Lash et al., 1991) generated candidate reference values broadly similar in magnitude to the value derived from the rodent hepatic bioassay data, supporting the biological plausibility of DCM’s neurological effects in humans via the CO-generating oxidative metabolic pathway, even though the liver was ultimately selected as the more sensitive and better-characterized critical effect across species. Elevated serum bilirubin was reported in one occupational cohort (Ott et al., 1983), but this was not replicated in other hepatic enzyme measures or in a later analysis of the same cohort (Soden, 1993), and epidemiological studies in general have limited sensitivity to detect subclinical hepatic changes of the type identified in the more sensitive rodent bioassays.

7. Data Gaps and Areas of Ongoing Research

Several limitations in the DCM toxicology database bear on the confidence of hazard characterization. No two-generation reproductive toxicity study by the oral route is available; the existing oral reproductive and developmental studies are one-generation or single-dose in design (General Electric Company, 1976; Narotsky and Kavlock, 1995), and the available inhalation two-generation study did not maintain continuous exposure through gestation and nursing (Nitschke et al., 1988b), limiting its ability to fully characterize reproductive and developmental toxicity.

No study — oral or inhalation — has incorporated a standardized neurobehavioral or neurodevelopmental testing battery in offspring following gestational DCM exposure, despite documented placental transfer of DCM (Withey and Karpinski, 1985; Anders and Sunram, 1982), relatively high CYP2E1 activity in the developing human fetal brain compared with the fetal liver (Hines, 2007; Johnsrud et al., 2003; Brzezinski et al., 1999), and demonstrated behavioral effects following single-dose developmental inhalation exposure in rats (Bornschein et al., 1980; Hardin and Manson, 1980). This represents one of the more significant gaps affecting confidence in the susceptible-population characterization for DCM.

Chronic, repeated-exposure immunotoxicity data at low concentrations are limited. The available subchronic study found no systemic immunosuppression (Warbrick et al., 2003), but this contrasts with evidence of acute pulmonary immunosuppression following a single low-concentration exposure (Aranyi et al., 1986); whether repeated low-level exposure produces a comparable localized pulmonary effect has not been directly tested.

The mechanistic basis for DCM-associated mammary gland tumors in rats remains incompletely characterized. Unlike the liver and lung tumor responses in mice, which are well supported by tissue-specific genotoxicity data, the mammary tumor findings are driven predominantly by benign lesions, dose-response patterns are less clear, and mammary tissue has not been directly examined in genotoxicity or GST-T1 expression studies, leaving the relevant dose metric and mode of action for this tumor type uncertain.

Finally, direct measurement of DCM-induced mutagenic events following oral exposure is sparse; much of the mechanistic genotoxicity database derives from inhalation exposure studies, and the linkage between oral bioassay tumor findings and the GST-T1 mutagenic mode of action has relied substantially on PBPK-based pharmacokinetic modeling to bridge across the two exposure routes (Marino et al., 2006) rather than on direct oral-route genotoxicity data. Continued refinement of PBPK models, additional chronic low-dose immunotoxicity and neurodevelopmental studies, and larger, better-exposure-characterized occupational cohorts with greater statistical power for rare cancer outcomes would meaningfully reduce the remaining uncertainty in DCM’s toxicological hazard characterization.

Bibliography