Registration Dossier

Administrative data

Key value for chemical safety assessment

Genetic toxicity in vitro

Link to relevant study records

Referenceopen allclose all

Endpoint:
in vitro DNA damage and/or repair study
Type of information:
experimental study
Adequacy of study:
key study
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
study well documented, meets generally accepted scientific principles, acceptable for assessment
Principles of method if other than guideline:
Assessment of DNA damage response to plasma levels of formaldehyde in chicken DT40 and colorectal cancer (RKO) cells with targeted mutations in various DNA repair genes.
GLP compliance:
not specified
Type of assay:
other: DNA damage response
Specific details on test material used for the study:
- Name of test material (as cited in study report): formaldehyde
- Other: 37% aqueous
Species / strain / cell type:
other: Chicken DT40 and colorectal cancer (RKO) cells
Metabolic activation:
not specified
Details on test system and experimental conditions:
Cell viability was determined by the XTT assay.
Cells were exposed to formaldehyde and allowed to divide for 7 days.
Intracellular GSH was determined using a commercially available kit
Species / strain:
other: Chicken DT40 and colorectal cancer (RKO) cells
Additional information on results:
The DT40-derived mutants showed sensitivity to formaldehyde in the following order: FANCD2>>BRCA2>BRCA1=XRCC2=RAD51C=RAD51D=XRCC3=RAD54>RAD52>parent DT40 cells. No correlation was found between GHS concentrations and formaldehyde-induced cell toxicity, revealing that the homologous recombination pathway rather than the NHEJ pathway is involved in repair of DNA-protein crosslinks (DPC). Nucleotide (NER), base pair excision repair (BER) and some other repair mechanisms were shown to not be involved in DPC repair.
The two most sensitive DT40 mutants (FANCD1 (BRCA2) and FANC2 -deficient cells) show sensitivity to formaldehyde at concentrations between 10-15 µmol/L, which lay below or in the very low range of those of endogenous formaldehyde. Therefore, RKO cells and their isogenic cells disrupted in FANCC or FANCG were exposed to formaldehyde and either FANCC or FANCG were hypersensitive to formaldehyde at concentrations of >20 µmol/L or >38 µmol/L, respectively. Thus cells of human origin showed similar hypersensitivity to formaldehyde.
Remarks on result:
other: cells of human origin showed similar hypersensitivity to formaldehyde

.

Endpoint:
in vitro DNA damage and/or repair study
Type of information:
experimental study
Adequacy of study:
key study
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
study well documented, meets generally accepted scientific principles, acceptable for assessment
Principles of method if other than guideline:
Earlier results indicating increased sensitivity to formaldehyde in FANCB-deficient cells were validated by testing a human FA gene-deficient cell line (human B cell line NALM-6). Upstream and downstream components of the FA DNA-repair pathway were tested using mutant chicken DT40 cell lines. An ADH5 (alcohol dehydrogenase 5)-knockout DT40 strain was used to test whether endogenous formaldehyde can be genotoxic (when accumulated).
GLP compliance:
not specified
Type of assay:
other: DNA damage and repair assay, unscheduled DNA synthesis in mammalian cells in vitro
Specific details on test material used for the study:
Name of test material (as cited in study report): Formaldehyde
Species / strain / cell type:
other: human B cell line NALM-6 and Chicken DT40 cells
Details on mammalian cell type (if applicable):
Human B cell line NALM-6: parent cells and FANCB knockout mutants
Metabolic activation:
not specified
Species / strain:
other: human FACB-deficient cells
Remarks on result:
other: cells are very sensitive to formaldehyde

Human FANCB-deficient cells are found to be very sensitive to formaldehyde; this sensitivity correlates with the accumulation of chromatid-type chromosome breakage and radial structure formation. The cells also showed enhanced induction of Ser139 phosphorylation of histone H2AX. Mutations in all components of the FA pathway tested sensitized cells to formaldehyde. Other pathways of DNA repair (tested in chicken DT40 cells) are found te be largely resistant to formaldehyde, with the exception of XPA showing mild sensitivity. However, in contrast to Ridpath et al. (2007) no interstrand crosslinks were found to be involved in formaldehyde induced damage. Chicken DT40 cells deficient in FANCD2 and ADH5 gene stopped growing after 3 days and cell survival improved by removing endogenous formaldehyde with ß-mercaptoethanol.

Genetic toxicity in vivo

Link to relevant study records

Referenceopen allclose all

Endpoint:
genetic toxicity in vivo, other
Remarks:
gene expression
Type of information:
experimental study
Adequacy of study:
key study
Reliability:
1 (reliable without restriction)
Rationale for reliability incl. deficiencies:
test procedure in accordance with generally accepted scientific standards and described in sufficient detail
Principles of method if other than guideline:
Concentration and exposure duration transitions in FA mode of action (MOA) were examined with pharmacokinetic (PK) modeling for tissue formaldehyde acetal (FAcetal) and glutathione (GSH) and with histopathology and gene expression in nasal epithelium from rats exposed to 0, 0.7, 2, 6, 10, or 15 ppm FA 6 h/day for 1, 4, or 13 weeks.
GLP compliance:
not specified
Specific details on test material used for the study:
Name of test material (as cited in study report): formaldehyde
Species:
rat
Strain:
Fischer 344
Sex:
male
Details on test animals and environmental conditions:
TEST ANIMALS
- Source: Charles River Breeding Laboratories, Inc. (Wilmington, MA or Kingston, NY).
- Age at study initiation: 6-7 weeks
- Acclimation period: ca 2 weeks
Route of administration:
inhalation: vapour
Details on exposure:
Two 13-week inhalation exposure studies were conducted. Rats were whole-body exposed to target vapor concentrations of 0 (control), 0.7, 2, 6, 10, or 15 ppm FA in either 8 m3 (first 13-week study) or 1 m3 (second 13-week study) stainless steel and glass chambers. Controls were exposed to filtered air only. Chambers were provided with air at a flow rate of ~12 to 15 air changes per hour, and the airflow rate was monitored and recorded.
Test atmospheres were generated by slowly vaporizing solid paraformaldehyde in stainless steel pans of various sizes in sealed stainless steel canisters in an oven. Nitrogen flowed through the stainless steel canister and carried vaporized FA into the charcoal-HEPA-filtered air supplying each chamber. Atmosphere concentrations were monitored every 30 min during the 6-h exposure period, using a calibrated infrared (IR) analyzer .
Duration of treatment / exposure:
6 h/day, for 1, 4, or 13 weeks
Frequency of treatment:
5 days/week
Dose / conc.:
0.7 ppm (analytical)
Dose / conc.:
2 ppm (analytical)
Dose / conc.:
6 ppm (analytical)
Dose / conc.:
10 ppm (analytical)
Dose / conc.:
15 ppm (analytical)
No. of animals per sex per dose:
For the cell proliferation and histopathology study: 8 per dose per time period.
For the gene expression, cytokines, hematology, and bone marrow evaluation: 15 per dose per time point.
Control animals:
yes
Tissues and cell types examined:
Nasal epithelium
Genotoxicity:
other: Dose dependencies in MOA, high background FAcetal, and nonlinear FAcetal/GSH tissue kinetics indicate that FA concentrations < 1 or 2 ppm would not increase risk of cancer in the nose or any other tissue or affect FA homeostasis within epithelial cells.

Cell proliferation: Clear dose-response trends at all three exposure durations with increases seen at 6, 10, and 15 ppm but not at the two lower exposure concentrations.

 

Histopathology: Treatment-related nasal lesions (predominantly inflammation, squamous cell metaplasia, and epithelial hyperplasia) were found in the respiratory/transitional epithelium in rats exposed to 2 ppm FA or higher.

At higher exposure concentrations >6 ppm and especially the 1-week exposure, there was erosion or ulceration of respiratory epithelia and/or necrosis of underlying structures. The incidence of the lesions was related to the inhaled concentration and inversely related to the exposure duration; with longer exposure, the more robust squamous epithelium showed less erosion than did the respiratory epithelium it had replaced.

 

Gene Expression Profiling: The total number of genes that was significantly altered across all concentrations and durations was 2197, but patterns of gene expression varied with exposure concentration and duration. Although at 13 weeks the numbers of genes significantly up- and downregulated were higher than at 1-week and 4-week exposure, no grouping of genes appeared to be uniquely associated with this longer duration, as had been observed at shorter exposures.

Enrichment analysis was performed on the highest three concentrations at all exposure durations and for the up- and downregulated gene groupings at 1 and 4 weeks, and indicated a diverse suite of enriched pathways. The top 10 of these pathways includes Wnt, TGF-beta, Erbb and Hedgehog signalling, as well as pathways related to DNA repair and cell cycle. At 10 ppm, 8 of 10 of the top pathways were cell cycle related with the two others related to DNA damage and Erbb signaling. At 15 ppm, cell cycle and DNA damage were represented; however, cell adhesion and immune response pathways were also present.

At the 1-week 15 ppm exposure concentration, a widespread activation of various immune response pathways likely associated with an inflammatory response following cytotoxicity was observed. Benchmark doses for significantly enriched pathways were lowest at 13 weeks. Seven genes, in previous studies found to be upregulated genes at lower exposure concentrations, were combined in a ‘‘Sensitive Response Genes’’-grouping (SRG) and had the lowest BMD of 1 ppm.

 

PK analysis: The PK analysis showed that the lower two inhaled FA concentrations (0.7 and 2 ppm) would be characterized by only minor changes in cellular GSH and intracellular FAcetal. Above 4 ppm, FAcetal increases with a steeper dose response and free GSH is reduced to much more significant degree.

Transcriptional and histological changes corresponded to the dose ranges in which the PK model predicted significant reductions in free GSH and increases in FAcetal. Genomic changes at 0.7–2 ppm likely represent changes in extracellular FAcetal and GSH. DNA replication stress, enhanced proliferation, squamous metaplasia, and stem cell niche activation appear to be associated with FA carcinogenesis. At 2 ppm, sensitive response genes (SRGs)—associated with cellular stress, thiol transport/reduction, inflammation, and cell proliferation—were upregulated at all exposure durations. Dose dependencies in MOA, high background FAcetal, and nonlinear FAcetal/GSH tissue kinetics indicate that FA concentrations below 1 or 2 ppm would not increase risk of cancer in the nose or any other tissue or affect FA homeostasis within epithelial cells.

Endpoint:
in vivo mammalian cell study: DNA damage and/or repair
Type of information:
experimental study
Adequacy of study:
key study
Reliability:
1 (reliable without restriction)
Rationale for reliability incl. deficiencies:
test procedure in accordance with generally accepted scientific standards and described in sufficient detail
Principles of method if other than guideline:
Fischer rats were exposed by nose-only inhalation for 6 h to 0.7, 2, 5.8, and 9.1 ppm formaldehyde and protein adducts were determined using liquid chromatography-coupled tandem mass spectrometry.
GLP compliance:
not specified
Type of assay:
other: formation of DNA adducts
Specific details on test material used for the study:
Name of test material (as cited in study report): formaldehyde
Species:
rat
Strain:
Fischer 344
Sex:
male
Details on test animals and environmental conditions:
- Age at study initiation: 6 weeks
Route of administration:
inhalation: vapour
Details on exposure:
Nose-only inhalation
Duration of treatment / exposure:
6 h
Frequency of treatment:
Single exposure
Dose / conc.:
0.7 ppm (analytical)
Dose / conc.:
2 ppm (analytical)
Dose / conc.:
5.8 ppm (analytical)
Dose / conc.:
9.1 ppm (analytical)
No. of animals per sex per dose:
3
Control animals:
yes
Tissues and cell types examined:
Nasal epithlium, distant tissues (lung, liver and bone marrow)
Genotoxicity:
other: Exposure to inhaled FA was not observed to affect endogenous protein adduct formation.
Additional information on results:
Exposure to inhaled FA was not observed to affect endogenous adduct formation. Levels of endogenous adducts (ranging from 2−4 FLys per 1E4 lysines), were similar in tissues of rats exposed to the highest dose (9.1 ppm) of [13C2H2]-FA and control rats. Exposure-dependent formation of exogenous FLys was only detected in the nasal epithelium. In distant tissues of the lung, liver and bone marrow, the exogenous adducts did not increase beyond the natural isotope abundance level. In addition to total protein, the analysis of protein in cytosolic, membrane and nuclear compartments revealed exposure-dependent formation of exogenous FLys only in the nasal epithelium. A clear exposure−response relationship was observed for lysine N6-formylation, with exogenous adducts in total protein rising from <3% of endogenous adducts to >40% for a ∼10-fold increase in FA exposure (0.7 to 9.1 ppm). At each of these FA exposures, the ratios were in the order cytoplasmic ≈ membrane > soluble nuclear > chromatin protein bound formaldehyde, indicating a decrease in the exogenous FA-Lys concentration from the cytoplasmic to the nuclear proteins.
Endpoint:
in vivo mammalian cell study: DNA damage and/or repair
Type of information:
experimental study
Adequacy of study:
key study
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
study well documented, meets generally accepted scientific principles, acceptable for assessment
Principles of method if other than guideline:
Both endogenous and exogenous N2-hydroxymethyl-dG adducts in nasal DNA of rats exposed to formaldehyde for 6 h were quantified by a highly sensitive nano-UPLC-MS/MS method.
GLP compliance:
not specified
Type of assay:
other: DNA adduct formation
Specific details on test material used for the study:
Name of test material (as cited in study report): formaldehyde
Species:
rat
Strain:
not specified
Sex:
not specified
Route of administration:
inhalation: gas
Duration of treatment / exposure:
6 h
Frequency of treatment:
Single exposure
Post exposure period:
2 h
Dose / conc.:
0.7 ppm
Dose / conc.:
2 ppm
Dose / conc.:
5.8 ppm
Dose / conc.:
9.1 ppm
Dose / conc.:
15.2 ppm
No. of animals per sex per dose:
5 rats per group
Tissues and cell types examined:
Nasal respiratory epithelium, bone marrow
Genotoxicity:
other: The number of exogenous N2-hydroxymethyl-dG adducts induced was highly non-linear
Genotoxicity:
other: In bone marrow, exogenous formaldehyde adducts were below the detection limit of 20 µmol, whereas endogenous N2-hydroxymethyl dG adducts were ∼15 adducts/1E7 dG.
Additional information on results:
The number of exogenous N2-hydroxymethyl-dG adducts induced was highly non-linear: 0.039 ± 0.019, 0.19 ± 0.08, 1.04 ± 0.24, 2.03 ± 0.43, and 11.15 ± 3.01 adducts/1E7 dG for 0.7, 2.0, 5.8, 9.1, and 15.2 ppm [13CD2]-formaldehyde exposure for 6 h, respectively. In contrast, the amount of endogenous N2-hydroxymethyl-dG did not exhibit a concentration-dependent effect, as 3.62 ± 1.33, 6.09 ± 3.03, 5.51 ± 1.06, 3.41 ± 0.46, and 4.24 ± 0.92 adducts/1E7 dG were present at 0.7, 2.0, 5.8, 9.1, and 15.2 ppm [13CD2]-formaldehyde exposures, respectively.

In bone marrow, exogenous formaldehyde adducts were below the detection limit of 20 µmol, whereas endogenous N2-hydroxymethyl dG adducts were ∼15 adducts/1E7 dG. The authors conclude that taking into consideration the sensitivity of the method and the amounts of endogenous adducts that less than 1 exogenous adduct would be present in 1500 endogenous adducts at 15.2 ppm. It was considered highly implausible that this one additional exogenous adduct could induce malignant transformation in the bone marrow if 1500 endogenous adducts do not.
Endpoint:
genetic toxicity in vivo
Type of information:
experimental study
Adequacy of study:
key study
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
study well documented, meets generally accepted scientific principles, acceptable for assessment
Principles of method if other than guideline:
Rats were exposed (whole-body) to 0, 0.5, 1, 2, 6, 10 or 15ppm FA for four weeks (5 days/week, 6 h/day). The frequencies of micronuclei (MN), histopathological changes and cell proliferation were determined. The micronucleus test (MNT) was performed with nasal epithelial cells prepared from six animals per group. Two thousand cells per animal were analysed for the presence of MN. The other six rats per group were subcutaneously implanted with osmotic pumps containing 5-bromo-2-deoxyuridine (BrdUrd), three days prior to necropsy. Paraffin sections were made from the nasal cavity (four levels) of these animals for histopathology and cell-proliferation measurements.
GLP compliance:
yes
Specific details on test material used for the study:
Name of test material (as cited in study report): formaldehyde
Species:
rat
Strain:
Fischer 344/DuCrj
Sex:
male
Details on test animals and environmental conditions:
-Age: Nine weeks old at the beginning of the experiment
Route of administration:
inhalation
Details on exposure:
Whole body exposure.
Duration of treatment / exposure:
6 h/day.
Frequency of treatment:
4 weeks, 5 days/week
Dose / conc.:
0.51 ppm (analytical)
Remarks:
± 0.05 ppm
Dose / conc.:
1 ppm (analytical)
Remarks:
± 0.11 ppm
Dose / conc.:
2.02 ppm (analytical)
Remarks:
± 0.15 ppm
Dose / conc.:
6.12 ppm (analytical)
Remarks:
± 0.34 ppm
Dose / conc.:
10 ppm (analytical)
Remarks:
± 0.39 ppm
Dose / conc.:
14.96 ppm (analytical)
Remarks:
± 0.05 ppm
No. of animals per sex per dose:
12
Control animals:
yes
Positive control(s):
Standard positive control used for the in vivo erythrocyte MNT: six male rats were treated twice orally, with an interval of 24 h, with 10 mg/kg b.w. cyclophosphamide (CP) and analysed 24 h after the last treatment.
No appropriate positive control has been established for the MNT with nasal epithelial cells. Therefore, a follow-up study with CP has been performed. Groups of six rats were treated with CP (20 mg/kg) once by gavage and sacrificed 3, 7, 14 and 28 days later.
Tissues and cell types examined:
Nasal epithelial, blood.
Statistics:
Data were analysed by pair-wise comparison of each dose group with the control group using the non-parametric one-tailed Wilcoxon rank test. A statistically significant difference was set at p < 0.05.
Sex:
male
Genotoxicity:
other: No increase was noted and even after an oral dose of cyclophosphamide the rate of micronuclei was not increased when the animals were studied 1-28 days after application.
Sex:
male
Genotoxicity:
other: Nasal histopathology and cell proliferation showed effects as to be expected at exposures of 6, 10, and 15 ppm

MNT: For the negative control group, a mean frequency of 4.75 cells with MN per 1000 cells was measured. Inhalation of FA for 28 days did not lead to a statistically significant induction of MN in nasal cells. However, there was also no induction of MN in nasal cells of rats exposed to a single dose of cyclophosphamide (CP, 20 mg/kg) by gavage and analysed 3, 7, 14 or 28 days after the treatment.

Histopathology: Qualitative analysis of HE-stained sections did not reveal any epithelial lesions in any of the regions after exposure to FA up to 6 ppm. Treatment with 10 ppm resulted in minimal or slight squamous metaplasia in nasal cavity level I. Exposure to 15 ppm showed in a continuum of (up to severe) site-specific epithelial degeneration and inflammation. The overall data demonstrate a clear concentration-related response with a no-observed effect level of 6 ppm. The cumulative incidence for metaplasia across the four levels shows a clear decreasing gradient from level I to level IV (anterior–posterior).

Cell proliferation: Inhaled FA induced cell proliferation in a site specific and concentration-related manner in the nasal cavity: dose-dependent higher mean values of the unit-length labelling indices (ULLI), were observed with values up to a 10-fold increase in the anterior parts of the nasal cavity of rats exposed to concentrations of 10 and 15 ppm for 28 days. Treatment with 6ppm still showed some borderline effects in the anterior part of the nasal cavity but the ULLI only reached a 1.8-fold increase. Site-specific, pathological lesions of the respiratory and transitional epithelium in the nasal cavity were only found after exposure to 10ppm and 15 ppm. Concentrations of 6ppm and lower did not result in histopathological changes.

Endpoint:
in vivo mammalian cell study: DNA damage and/or repair
Type of information:
experimental study
Adequacy of study:
key study
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
study well documented, meets generally accepted scientific principles, acceptable for assessment
Principles of method if other than guideline:
Monkey were exposed to 1.9 to 6.1 ppm isotope-labeled formaldehyde for 2 days (6 h/day) and exogenous dG adducts in nasal DNA were determined..
GLP compliance:
not specified
Type of assay:
other: formation of DNA adducts
Specific details on test material used for the study:
Name of test material (as cited in study report): formaldehyde
Species:
other: rats and monkeys
Route of administration:
inhalation: gas
Duration of treatment / exposure:
Rats: 1 day
Monkeys: 2 days (6 h/day)
Dose / conc.:
2 ppm
Remarks:
Rats
Dose / conc.:
5.8 ppm
Remarks:
Rats
Dose / conc.:
1.9 ppm
Remarks:
Monkeys
Dose / conc.:
6.1 ppm
Remarks:
Monkeys
Tissues and cell types examined:
nasal tissue
Genotoxicity:
other: Exogenous adducts formed in the nasal turbinates are lower for nonhuman primates that for rats. There are indications that endogenous dG adducts are 2-3-fold higher in monkeys than in rats.

Exogenous dG adducts in nasal DNA in monkeys after two days of formaldehyde exposure was similar to that in rats after one day of exposure to 2 ppm formaldehyde. Exogenous dG adducts in nasal DNA in monkeys after two days of exposure to 6.1 ppm formaldehyde was ~2.5-fold lower than in rats exposed to 5.8 ppm formaldehyde for 1 day. Relative to rats, primates were found to have two- to threefold higher endogenous dG adducts in the analysed tissues, further reducing the ratio exogenous/endogenous formaldehyde adducts in primates exposed to low concentrations.

Endpoint:
in vivo mammalian cell study: DNA damage and/or repair
Type of information:
experimental study
Adequacy of study:
key study
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
study well documented, meets generally accepted scientific principles, acceptable for assessment
Reason / purpose:
reference to same study
GLP compliance:
not specified
Type of assay:
other: DNA adduct formation
Specific details on test material used for the study:
- Name of test material (as cited in study report): formaldehyde
- Test material form: gas
Species:
other: Rat and monkeys
Strain:
other: Rat: Fischer 344; Monkey: Cynomolgus macaques
Sex:
male
Route of administration:
inhalation: gas
Details on exposure:
Air supply was maintained at approximately 23 L/min
Rat: nose only; Monkey: whole body
Duration of treatment / exposure:
Rats: 7, 14, 21, or 28 consecutive days (6 h/day).
Monkeys: 2 consecutive days (6 h/day).
Post exposure period:
Rats: post exposure at 6, 24, 72, and 168 h.
Dose / conc.:
2 ppm
Remarks:
Rats
Dose / conc.:
6 ppm
Remarks:
Monkeys
No. of animals per sex per dose:
3
Control animals:
yes
Tissues and cell types examined:
Rats and monkeys: nasal respiratory epithelium (from right and left sides of the nose and from the septum), blood, white blood cells, spleen, thymus, tracheal bronchial lymph nodes (TBLN), mediastinal lymph nodes, trachea, lung, kidney, liver, and brain. Bone marrow (from both femurs).
DNA was isolated from the tissues using a NucleoBond DNA isolation kit, reduced and digested, and quantified with ultrasensitive nano ultra performance LC-MS-MS.
Statistics:
The statistical significance of adduct concentration differences were assessed using statistical tests appropriate to the different experimental designs. Two-sided unpaired student’s t-tests were used to compare tissue-specific endogenous adduct concentrations in exposed and control animals. The 2-sided Dunnett’s test (Dunnett, 1964) was utilized in assessing differences by treatment duration in relation to the single control group employed in the 28-day exposure study. Concentration differences were considered to be statistically significant if P<.05 for all statistical tests.
Genotoxicity:
other: At an exposure level of 2 ppm FA-dG adducts accumulate to reach a steady state after 28 days (rats).
Genotoxicity:
other: Endogenous and exogenous N2-HOMe-dG adducts were found in all regions of nasal passages studied. The amount of exogenous adducts in different sections of the exposed monkey nasal epithelium was 5- to 11-fold lower than that of endogenous adducts (monkeys)

Rats: In nasal epithelial DNA, both endogenous and exogenous DNA adducts were identified and quantified after 28 days of exposure. The amount of exogenous [13CD2]-N2-Me-dG in nasal DNA increased with extended exposure time, indicating that DNA adducts in the form of N2-HOMe-dG accumulated during the 28-day exposure period in vivo. In control rats, equivalent amounts of endogenous adducts and no exogenous adducts were observed. In all other tissues, only endogenous formaldehyde-induced N2-HOMe-dG could be observed (except for 1 bone marrow sample from a 28-day exposure rat).

The t1/2 for the formation and repair/loss of [13CD2]-N2-HOMe-dG adducts in nasal respiratory epithelium was estimated to be 7.1 days (90% confidence interval (CI) = [6.0, 8.7] days using all of the data). The exogenous DNA adducts, [13CD2]-N2-HOMe-dG, were hence shown to approach a steady-state concentration during the 28-day exposure phase.

Monkeys: Both endogenous and exogenous N2-HOMe-dG adducts were found in all regions of nasal passages studied. The amount of exogenous adducts in different sections of the exposed monkey nasal epithelium was 5- to 11-fold lower than that of endogenous adducts. In monkey bone marrow, white blood cells, and trachea, only endogenous adducts were found; however, the amount of endogenous N2-HOMe-dG adducts present in direct scraped bone marrow was at least 2 times higher than in collection using saline extrusion, indicating distinct differences between these 2 bone marrow collection procedures.

Additional information

Genotoxicity in vitro

Numerous studies are available on genotoxicity of formaldehyde in vitro and a comprehensive summary is presented in IARC (1995 & 2006; see review Gentox in vitro; evaluation in IARC 2006,). A detailed presentation of selected studies (robust study summaries) on gene mutation in bacteria and mammalian cells, clastogenic activity in mammalian and human cells as well as data on the mechanisms of genotoxicity are given in IUCLID Section 7.6.1.

In the Salmonella microsome assays gene mutations were induced with and without metabolic activation. In mammalian cells, chromosome aberrations were detected independent of the metabolic activation system. There is also some indication for gene mutation in mammalian cells without a metabolic activation system. The DNA-protein cross-linking (DPC) activity which occurred in vivo at the site of first contact has also been demonstrated in in vitro experiments; repair and threshold concentrations have been reported.

In further studies (IARC, 1995/2006) formaldehyde induced DNA damage in bacteria, forward and reverse mutation in bacteria (S. typhimurium and E. coli), DNA damage in fungi including DNA-protein cross-links (DPC), gene conversion and gene mutation in fungi, and in mammalian cells DNA damage including DPC, sister chromatid exchange and cell transformation. The available data on DPC formation and repair in in vitro studies suggested that formaldehyde concentrations of >= 200 μM (6 μg/mL) which produce enough DPC to persist until replication, cannot be expected in blood of humans occupationally exposed to formaldehyde via the inhalation route. Therefore, the reported systemic effects in the SCE test and the micronucleus tests are most likely not related to the formaldehyde exposure. In conclusion, systemic cytogenetic effects in blood cells of humans exposed to formaldehyde are unlikely to occur because these conditions are not met.

Gene mutation in bacterial cells:

Marnett et al. (1985) have shown gene mutagenic activity in the Ames test without metabolic activation. S. typhimurium TA97, TA98, TA100, TA102, TA104 were exposed to formaldehyde at the following concentrations: 0, 0.3, 0.7, 1.4, 2 μmoles/plate (0, 9, 21, 30, 42, 60 μg/plate). The test substance was tested to its toxic limit (documented only for TA104). The test substance induced gene mutation in TA104 at dose levels below the cytotoxicity threshold. Mutagenic activity is also detected in TA102. Weak effects were observed in TA97 and TA98. TA102 is known to detect cross-linking mutagens.

S. typhimurium TA1535, TA1537, TA98 and TA100 were incubated at concentrations of 3.3-300 μg formaldehyde/plate with and without metabolic activation (MA). Negative and positive controls were valid. Weak positive effects were found in TA100 and TA98 without MA but more pronounced effects with MA (even at non-cytotoxic concentrations) but again with a weak mutagenic potency. Under the experimental conditions described in this study the test substance has mutagenic activity in TA98 and TA100 but only weak potency (Haworth et al., 1983).

It is important to note that point mutations in bacteria cannot be directly translated to mammalian systems because bacteria are lacking histones and therefore the predominant genotoxic effect in mammalian cells, i. e. DPC formation, is not possible in bacteria or only to a much lesser extent.

Gene mutation in mammalian cells:

Formaldehyde has mutagenic activity in the mouse lymphoma assay (Blackburn et al., 1991). Mouse lymphoma L5178Y cells were exposed to 0, 8, 12, 16, 20 nL/mL formalin (containing 37% formaldehyde) without metabolic activation (MA) and to 0, 40, 45, 50, 55, 60, 65 nL/mL with MA (related to 37% formalin). Cytotoxicity (% total growth) and mutant frequency were scored. The test substance induced dose dependent increases in the mutant frequencies with and without MA. Cytotoxicity was obvious at the same concentrations. There was no differentiation given between gene and chromosome mutation (large and small colonies). Addition of formaldehyde dehydrogenase inhibited the mutagenic effects suggesting no mutagenic activity of any impurity (e. g. methanol). The test substance was considered to be mutagenic under the condition of this study.

V79 Chinese hamster lung fibroblast were incubated for 4 h with 0, 3.75, 7.5, 15 μg formaldehyde/mL medium without MA and processed for detection of gene mutations in the HPRT assay (Merk & Speit, 1998). In contrast to the positive control the test substance did not induce gene mutation at the HPRT locus even at cytotoxic dose levels. But the same concentrations induced dose dependent increases in the number of micronuclei in parallel experiments. At the same concentrations formaldehyde induced DNA-protein crosslinks parallel to the induction of cytotoxicity. These results indicated that chromosome mutation (clastogenicity) and cytotoxicity but not gene mutation might be related to DNA-protein crosslinks. Conclusion: Formaldehyde has no gene mutagenic activity in the HPRT assay.

In addition, Speit and Merk (2002) evaluated the mutagenic effects of formaldehyde in the mouse lymphoma assay (L5178Y cells). There was a clear and concentration related mutagenic effect. But as this effect was mainly due to a strong increase in small colony mutants the authors conclude that formaldehyde mainly causes mutations by induction of chromosomal aberrations.

In a further HPRT assay in V79 cells formaldehyde has gene mutagenic activity without metabolic activation at cytotoxic dose levels (Grafström et al., 1993). V79 cells were exposed for 1 h without a metabolic activation system to formaldehyde at concentration levels of 0, 3, 9, 18, 30 μg/mL. The positive control was valid. The test substance induced dose dependently gene mutation at the HPRT locus at concentrations which were cytotoxic. Positive results were found at >= 300 µM (9 μg/mL). According to Merk and Speit (1998) this effect is mainly due to reduced cell survival and not to increased occurrence of mutant colonies.

Cell cultures of human lymphoblasts (Liber et al., 1989) were treated with 4.5 μg/mL test substance for 8 x 2 h without MA for induction of gene mutations at the HPRT gene. Each treatment resulted in approx. 50% survival. These extreme conditions were necessary to obtain sufficient mutants for subsequent analysis leading finally to a mutation frequency 12-fold higher than the background. The authors presented an examination of the mutational specificity of formaldehyde in comparison to spontaneous mutations using Southern blot and Northern blot analysis as well as DNA-sequence analysis. The mutations are caused by complete or partial deletions but also by point mutations (comparable with spontaneous mutations). The limited number of mutants investigated indicated to a shift of the mutational spectrum from point mutations to deletions under formaldehyde treatment. Spontaneous mutations in untreated cells gave 67% point mutations, 27% small and 6% complete deletions while formaldehyde treatment resulted in 53%, 27% and 20%, respectively. DNA sequence analysis revealed an AT → GC transition in the spontaneous mutant. 6 out of the 7 formaldehyde-induced mutants were base substitutions at the AT base pairs, 4 of them were AT → CG transversions at one specific site. Conclusion: Formaldehyde has mutagenic activity in the HPRT assay on human lymphoblasts and DNA sequence analysis indicated to some differences to spontaneous mutations. In comparison with spontaneous mutations formaldehyde led to a shift from point mutations in favor of complete deletions.

In conclusion, point mutations are only of minor importance for in vitro mutagenicity of formaldehyde.

Chromosome aberrations:

In the mouse lymphoma assay (Speit & Merk, 2002; see above) the mutagenic activity was mainly caused by induction of chromosomal aberrations as indicated by the predominant formation of small colony mutants.

In CHO cells formaldehyde has chromosome mutagenic activity (Galloway et al., 1985). The cytogenetic studies were performed in 2 independent laboratories. CHO cells were exposed with and without MA to formaldehyde at concentrations ranging between 1.1 and 16 μg/mL (without MA) or 1.1 and 50 μg/mL (with MA). The cytotoxicity threshold was reached (but not clearly documented in the result part). Questionable positive results were obtained without MA but dose dependent clastogenic effects with MA.

V79 cells were incubated for 4 h with 0, 0.45, 0.93, 1.86, 3.75, 7.5 μg formaldehyde/mL medium without MA and processed for detection of clastogenic effects in the micronucleus assay. The test substance induces dose dependent increases in the number of micronuclei at dose levels which also showed cytotoxic effects (3.75 μg/mL). At the same concentrations formaldehyde induced DNA-protein crosslinks in parallel experiments. These results indicated that clastogenicity and cytotoxicity might be related to DNA-protein crosslinks. No relationship was seen with gene mutation in the parallel HPRT assay (Merck & Speit, 1998). Conclusion: Formaldehyde has clastogenic activity in the micronucleus test.

In a recent study by Speit et al. (2007; supporting data) V79 cells were exposed for 2 h to formaldehyde at dose levels of 0, 5, 10, 25, 50, 75, 100, 200, 300, 400 μM (0, 0.15, 0.3, 0.75, 1.5, 2.25, 3, 6, 9, 12 μg/mL) without MA. Significantly increased MN frequency was seen at >= 75 μM (2.25 μg/mL), a doubling of the MN frequency (compared with control) at >= 100 μM (3 μg/mL). The concentration-effect for micronuclei induction was characterized by fitting different curves to the data. A two-phase regression model fitted best in comparison with a linear or quadratic model and indicated a practical threshold for micronuclei induction.

Schmid et al. (1986) reported chromosome mutagenic activity in human lymphocytes exposed in vitro. Human lymphocytes (primary culture, one healthy donor) were exposed for 1 h to 0, 1, 1.9, 3.8, 7.5, 15, 30 μg formaldehyde/mL medium with and without MA and processed for analysis of chromosomal aberrations. Dose-dependent clastogenic activity was found at >= 7.5 μg/mL (with MA) and >= 3.8 μg/mL (without MA). Cytotoxic effects were detected at >= 7.5 μg/mL independent of metabolic activation.

Similar results were presented in a study report (Boots Company, 1986). Clastogenic effects were detected in primary cultures of human lymphocytes at dose levels inducing also a reduced mitotic index. In this cytogenetic assay mainly chromatid deletions and exchanges were found. A decrease in mitotic index was found at >= 6 μg/mL and significant clastogenic effects at 8 μg/mL.

In a recent study (Schmid & Speit, 2007) primary cultures of blood cells collected from healthy, young volunteers were used for determination of micronuclei (MN) after exposure to 0, 100, 200, 250, 300, 400μM (corresponding to 0, 3.0, 6.0, 7.5, 9.0, 12 μg/mL, no MA); 2000 cells per dose level were analyzed for MN and 500 cells for cytotoxicity (nuclear division index; NDI). Three trials per experiments were performed with blood from different donors. Protocol I (addition of formaldehyde at the start of the culture; total culture time of 68 h in protocol I-III) did not lead to increased MN frequencies; concentrations up to 250 μM (7.5 μg/mL) could be tested but caused strong cytotoxic effects. At Protocol II (blood cultures treated with formaldehyde 24 h after the start of the cultures) concentrations up to 400 μM (12 μg/mL) could be tested but no significant induction of MN was measured, however, data on cytotoxicity indicated that the NDI is strongly reduced under these experimental conditions. Protocol III (i. e. addition of formaldehyde 44 h after the start of the cultures) a clear and concentration-related induction of MN is observed, statistically significant at of 300 μM (9 μg/mL) and higher. Reduction in the NDI as an indicator of cytotoxicity parallels the induction of MN. Protocol III is expected to be most sensitive because proliferating lymphocytes are treated during the last cell cycle before preparation. Blood cells with increased amounts of DPC at the start of the cell culture are either able to repair induced DPC or are too heavily damaged to be detected as binucleated cells with induced micronuclei. FISH analysis to differentiate between a clastogenic effect (centromere-negative MN) and an aneugenic effect (centromere positive MN) in the induction of MN at 350 μM revealed that 81% of the analyzed MN in binucleated cells (119 out of 147) were centromere-negative and 19% were centromere-positive (more clastogenic effects). Conclusion: In vitro formaldehyde induced MN formation in human blood cells at concentrations >= 300 μM (9 μg/mL) if formaldehyde was added 44 h after the start of the cultures; the clastogenic effects were paralleled by cytotoxic effects. Addition of formaldehyde at earlier time point of the cell culture did not induce micronuclei but cytotoxicity. Taking into account the results obtained with the different protocols to test for micronuclei the authors concluded that cytogenetic effects in blood cultures of humans exposed to formaldehyde are unlikely to occur because the conditions of protocol III are not met.

DNA-protein cross-links (DPC) and DNA damage:

DNA strand breaks:

By the standard alkaline comet assay no DNA migration was induced in V79 Chinese hamster cells. Instead the values for the tail moment were even lower than in untreated controls. Even after extended electrophoresis time or proteinase K treatment, induction of DNA strand breaks was not observed. Again there was a trend for reduced migration under standard conditions that became significant by prolongation of electrophoresis time. The reduced migration is an indication for DNA-protein cross links (Speit et al., 2007).

In former years the induction of single strand breaks (SSB) in the alkaline elution assay was reported by several authors, often in conjunction with formation of DPC (e. g. Cosma and Marchok, 1988). These authors showed a very rapid and complete repair of SSB within 2 h. Under consideration of this rapid repair the mutational consequences of SSB are minor. Further the test method may lead to variable results regarding SSB. Therefore the data obtained by the comet assay should have preference showing that DPC formation is of major importance as compared to the formation of SSB.

SCE:

Primary cultures of blood cells collected from healthy, young volunteers were used for determination of SCE without MA after exposure to 0, 25, 50, 100, 200μM (corresponding to 0, 0.75, 1.5, 3.0, 6.0 μg/mL). No SCE induction was found at dose levels up to 100 μM (3 μg/mL) but significant effects (p<0.01) at 200 μM (6 μg/mL). Cytotoxicity: proliferation index was decreased (non-significant) at 100 μM and significantly (p< 0.01) at 200 μM. Conclusion: In vitro formaldehyde induced SCE formation in human blood cells at concentrations >= 200 μM (6 μg/mL); cytotoxic effects occur in parallel or even precede the genotoxic effect in the SCE test (Schmid & Speit, 2007). Based on the relationship of cytotoxicity and SCE formation and under consideration of the formaldehyde concentrations necessary to induce SCE in human blood cells, the authors conclude that positive SCE findings reported for humans are most likely not related to formaldehyde exposure.

In V79 cells SCEs were induced at concentrations >= 100 μM (3 μg/mL); the genotoxic effects were accompanied by cytotoxicity (Speit et al., 2007; supporting study). In this experiment the number of concentrations studied allowed to establish a concentration-effect relationship. The concentration-effect for SCE induction was characterized by fitting different curves to the data. A two-phase regression model fitted best in comparison with a linear or quadratic model and indicated a practical threshold for SCE induction.

Formaldehyde induced SCE in V79 Chinese hamster cells and A549 human lung cells in a concentration related manner (and Speit, 2008). After treatment with formaldehyde for 1 h, fresh BrdU containing medium was added. With the same exposure duration of 1 h, but addition of BrdU medium after 4 h the induction of SCE was strongly diminished. This result clearly shows that a relevant amount of SCE inducing lesions had been repaired within the time span of 4 h in fresh medium and that repair of the primary DPC is important for the magnitude of further “downstream” genotoxic effects.

DPC:

Cosma & Marchok (1988) incubated cells of the rat tracheal epithelial cell line C18 (nontumorigenic) without MA at concentration levels of 0, 3, 6, 12 μg formaldehyde/mL medium for 1.5 h and processed for alkaline elution assay for quantifying DPC and DNA single-strand breaks (SSB) and their repair. The test substance induced DPC and SSB. The DPC effects were also seen at non-cytotoxic dose levels. DPC and SSB were immediately repaired within 4 and 2 h, respectively. Conclusion: Formaldehyde has DNA-protein crosslinking activity and induces single strand breaks; both effects were rapidly repaired.

Merk & Speit (1998) have shown the DNA-protein cross-linking activity in V79 cells, which were incubated for 4 h with 0, 0.45, 0.93, 1.86, 3.75, 7.5, 15 μg formaldehyde/mL medium without MA. The K-SDS assay was used for quantifying DPC plus a parallel modified comet assay to detect also the cross-linking activity (leading to reduced tail moment after gamma-irradiation, compare with following DPC studies). A dose dependent increase in % DPC and reduced tail moment in the modified comet assay were found even at a concentration resulting only in slight cytotoxic effects (relative cloning efficiency 73% at 3.75 μg/mL). In parallel experiments formaldehyde was also clastogenic in the micronucleus test but did not induce gene mutation at the HPRT locus at concentrations which were clearly cytotoxic. This indicates a possible relationship between DPC and cytotoxicity/clastogenicity but not gene mutation.

In a recent study of Schmid & Speit (2007) primary cultures of blood cells collected from healthy, young volunteers were used for determination of DPC by the alkaline comet assay; DNA damage was measured without MA via the parameter tail moment. For the detection of DPC, formaldehyde-treated blood cultures (0, 25, 50, 75, 100, 200μM; corresponding to 0, 0.75, 1.5, 2.25, 3.0, 6.0 μg/mL) and controls were exposed to 2 Gy gamma-rays. In the presence of DPC, gamma-ray-induced DNA migration is reduced (indirect measure of DPC; reduction of the irradiation-induced tail moment due to formaldehyde-induced DPC formation). Removal of DPC was determined by the reduction of the inhibition of gamma-ray-induced DNA migration when cells are irradiated at different time points after treatment with formaldehyde. DPC: Formaldehyde induced a concentration-related reduction in gamma radiation induced DNA migration due to DPC formation (significant at >= 25 μM or 0.75 μg/mL). Repair: DPC induced in blood samples at the start of the culture are removed in time; 8 h after the start of the culture, DPC induced by < =100 μM (3 μg/mL) formaldehyde are completely removed while a portion of DPC induced by higher concentrations (200 and 300 μM) still persists after 24 h. DPC induced by concentrations up to 100 μM (3 μg/mL) are completely removed before lymphocytes start to replicate. Conclusion: In vitro formaldehyde induced DPC formation in human blood cells at concentrations >= 25 μM (0.75 μg/mL); DPC repair occurred, DPC induced by concentrations up to 100 μM (3 μg/mL) are completely removed.

A similar threshold for DPC formation in vitro was reported in V79 cells by Speit et al. (2007; supporting study). DPC was found at >10 μM (0.3 μg/mL), significant effects were measured at >=25 μM (0.75 μg/mL).

Further experiments of this working group (Speit et al., 2008) were performed with the human permanent lung cell line A549. DPC induced by formaldehyde (0, 100, 200, 300 μM or 0, 3, 6, 9 μg/mL) in these cells without MA were also determined indirectly by the alkaline comet assay after irradiation with gamma-rays. The abolition of formaldehyde-induced reduction in DNA migration is taken as a measure for time-dependent DPC repair. DPC: Treatment of cells with formaldehyde for 1 h resulted in a concentration-related induction of DPC; the effect was statistically different in comparison with the irradiated control at a concentration of >= 200 μM (6 μg/mL), slight (non-significant) effects were seen at 100 μM (3 μg/mL) in confluent cultures; however, significant effects were detected at >= 100 μM in proliferating cultures. Repair: Formaldehyde induced DPC were removed in a time-dependent manner: when cells were treated for 1 h with 300 µM formaldehyde and incubated in fresh medium for another 4 or 8 h a numerical but statistically non-significant reduction of the cross-linking effect was still seen after 4 hr. Eight hour after the end of the treatment DPC were completely removed. In contrast, permanent formaldehyde exposure (up to 24 h) has shown, that DPC are continuously induced over longer periods of time (significant at initial concentrations >= 200 μM) and that formaldehyde is not rapidly inactivated in the cell culture medium. Cytotoxicity: Using a more sensitive method on cytotoxic effects (clonal growth of A549 cells) significant cytotoxic effects were found at >= 20 μM (0.6 μg/mL); using cell counts after 48 h exposure to formaldehyde slight cytotoxicity was found at 100 μM (3 μg/mL), statistically significant at 200 μM (6 μg/mL). No significant reduction in cell counts up to a concentration of 500 μM (15 μg/mL) after 1 h exposure and further cultivation in fresh medium. Conclusion: Formaldehyde induced DPC formation in a human lung cell line at concentrations >= 100 μM (3 μg/mL, 1 h exposure, significant at 200 μM); DPC repair occurred, DPC induced by concentrations up to 300 μM (9 μg/mL) are completely removed within 8 h. Depending on the method to determine cytotoxicity (colony formation or inhibition of cell growth) DPC formation either occurred at higher (colony formation) or lower concentrations (inhibition of cell growth) than cytotoxicity.

Similar results were presented in parallel studies (Speit et al., 2008) on human nasal epithelial cells (HNEC). DPC induced by formaldehyde (0, 50, 100, 200, 300 μM or 0, 1.5, 3, 6, 9 μg/mL) in HNEC were also determined using the same experimental design. A reduction of gamma-ray induced increase in the tail moment was detected also in HNEC, the effect was statistically different in comparison with the irradiated control at a concentration of >= 200 μM (6 μg/mL), slight (non-significant) effects were seen at 100 μM (3 μg/mL). DPC Repair was also time-dependent. Compared with DPC formation (reduced tail moment) immediately after the 1 h formaldehyde exposure reduction of the cross-linking effect was measured after an incubation period in fresh medium of further 4 hr, but a statistically significant effect was still seen for the concentration of 200 μM (6 μg/mL, highest concentration tested). Eight hour after the end of the treatment, DNA migration was still slightly (but non-significantly) reduced; after incubation in fresh medium for 24 h cultures treated with 100 or 200 μM did not reveal any reduction of the gamma-ray-induced DNA migration indicating that DPC are completely removed. Conclusion: Formaldehyde induced DPC formation in primary cultures of human nasal epithelial cells at concentrations >= 200 μM (6 μg/mL), slight (non-significant) effects were seen at 100 μM (3 μg/mL). DPC repair occurred, DPC induced by concentrations up to 200 μM (6 μg/mL) are completely removed within 8-24 h.

In a recent publication by Neuss et al. (2010) in vitro co-cultivation experiments with primary human nasal epithelial cells (HNEC) and isolated lymphocytes were performed to investigate whether reactive formaldehyde (FA) can be passed on from nasal epithelial cells (site of first contact) to lymphocytes located in close proximity and induce DNA damage in these cells. A modified comet assay was used as a sensitive method for the detection of FA induced (dose levels 0, 100, 200, or 300 μM) DNA–protein cross links (DPC) in gammy ray exposed cells. Co-cultivation of lymphocytes with HNEC exposed to FA for 1 h causes a concentration related induction of DPC in lymphocytes when co-cultivation takes place in the exposure medium, however, when the exposure medium is changed after FA treatment of HNEC and before lymphocytes are added, no induction of DPC was measured in lymphocytes even after exposure of HNEC to high FA concentrations (300 mM) and extended co-cultivation (4 h). Measurement of FA concentration in the cell culture medium indicated that FA is actually not released even from highly exposed cells into the cell culture medium. In conclusion, FA that has entered nasal epithelial cells is not released and does not damage other cells in close proximity to the epithelial cells. These results are in contrast to recently proposed hypothetic mechanism for FA-induced leukaemia by damaging circulating haematopoietic stem cells or haematopoietic progenitor cells in nasal passages, which then travel to the bone marrow and become initiated leukemic stem cells (see Summary and discussion Carcinogenicity). This study reinforced the findings of Neuss and Speit (2008) that formaldehyde does not migrate from human A549 lung cells to V79 cells after washing and change of the culture medium. In this investigation the endpoint was SCE induction.

DNA adducts

Lu et al. (2009) have shown in vitro formation of S-[1-(N-2-deoxyguanosinyl) methyl] glutathione between glutathione and DNA in the presence of formaldehyde. This adduct is unique because of the involvement of S-hydroxymethylglutathione which is a key player during the detoxication of formaldehyde. Further studies might be valuable for answering the question of whether formaldehyde exhibits systemic toxicity.

A literature search after the last IUCLID update was carried out up to April 20, 2015 and provided the following new information:

After Lu et al. (2009) had shown that FA readily reacts with the thiol group of GSH to form a crosslink with N2-dG via a methylene group, Lu et al. (2010a, supporting) systematically studied crosslinking reactions of FA with different amino acids and nucleosides. The highest yields of crosslinked products were obtained with FA + lysine + dG followed by the reaction with cysteine and dG. Yields from the other reaction partners were lower by a factor of 10 or more. While the lysine adduct was unstable at ambient temperature, that derived from cysteine was stable.

Speit et al. (2014, supporting) did not observe co-mutagenic effects when A549 cells were treated simultaneously with formaldehyde and various mutagens. By the comet assay to study induction and removal of DPC and by the CBMN to measure micronuclei, formaldehyde did not enhance the genotoxicity of the other mutagens. In addition, no effect was noted on the expression (mRNA level) of the gene of O6-methylguanosine-DNA methyltransferase.

Luch et al. (2014, supporting) studied in Simian Virus 40-transformed fibroblasts the effect of low concentrations of formaldehyde (<100 µM) on a specific repair complex (UV-DBB) and on DNA excision repair of DNA lesions. They found that formaldehyde slows down the relocation of DNA damage sensors to DNA damage sites and thereby delays excision repair. A threshold concentration of 25 µM was established.

Grogan and Jinks-Robertson (2012, supporting) studied formaldehyde in a saccharomyces cerevisiae frameshift detection system for its effect on nucleotide excision repair (NER) and translesion DNA synthesis (TLS). By using strains defective for different repair pathways the author concluded that formaldehyde generated lesions (DPC) triggered error prone TLS and are substrates for NER.

Formaldehyde treatment (100 µM) of A549 cells led to an increase of DPC and malondialdehyde while superoxide dismutase and glutathione peroxidase activities were reduced. In addition, NF-kB and AP-1 were induced (Zhang et al., 2013, supporting). These effects were counteracted by pre-treatment of cells with the antioxidant curcumin. The authors suggest that the genotoxicity is mediated by oxidative stress.

Ji et al. (2014, supporting) used an in vitro erythroid expansion system to study the effect of formaldehyde on cultured mouse and human hematopoetic stem/progenitor cells. Induction of micronuclei in polychromatic erythrocytes differentiated from mouse bone marrow was significantly increased and the expansion of human erythroid progenitor cells was significantly suppressed in a dose related manner. In addition, in the human cells at the highest concentration (100 µMol), the proportion of cells in G2/M was slightly increased and aneuploidy of chromosomes 7 and 8 was significantly increased at 50 but not at 100 µMol. The authors interpret these findings as an indication for a potential mode of action for induction of leukaemia by formaldehyde.

She et al. (2013, supporting) studied genotoxicity of formaldehyde on mouse bone marrow mesenchymal stem cells in vitro at concentrations between 0.75 and 200 µM. At ≥75 µM cell survival was inhibited and DNA strand breaks were increased in the proteinase K modified comet assay. DPC were induced at ≥125 µM concordant with a decrease of strand breaks in the standard comet assay. The incidence of SCE was increased at ≥125 µM and that of micronuclei at ≥150 µM. Thus, DNA damage generally was observed at concentrations when cytotoxicity started and the findings were well in line with those of Speit and co-workers using other cell lines.

The relevance of these in vitro findings remains unclear in relation to specific sensitivities of hematopoetic stem/progenitor cells. Thus, the group of Speit and co-workers have shown micronuclei induction in many different cell lines. In addition, Kuehner et al. (2012, supporting) did not find leukaemia specific aneuploidies in vitro in cultured human progenitor cells and there was no indication that these cells are more sensitive to the cytotoxic effect of formaldehyde than other cell lines (human A549 lung cells or V79 cells). Furthermore, Speit et al. (2011c, supporting) demonstrated that in vitro formaldehyde does not induce aneuploidy using the micronucleus CBMN test with in situ hybridisation with a pan-centromeric probe of A 549 cells or when studying aneuploidy in V79 cells in comparison with standard clastogens or aneugens. Finally also by gene expression profiling in TK6 human lymphoblastoid cells, Kuehner et al. (2013, supporting) demonstrated that formaldehyde rather resembles clastogens and not aneugens.

Yoshida and Ibuki (2014, supporting) studies histone H3 phosphorylation after treatment of A549 cells in vitro with relatively high formaldehyde concentrations of 0.3-3 mM. By analysing the histone modifications they suggested that formaldehyde may act both as an initiator and promoter for carcinogenesis.

Recent studies related to Fanconi anemia (Fan) shed important light on the mechanism involved in cyto- and genotoxicity of (endogenous) formaldehyde. Fan is a rare disease associated with early development of cancer, including acute myeloid leukaemia. Cells from such patients are highly sensitive to DNA-DNA crosslinking agents associated with increased chromosomal damage. Ridpath et al. (2007, key) used chicken DT40 cells with targeted mutations defective in particular DNA repair and cell cycle checkpoint response pathways. Formaldehyde caused a significant reduction of survival of mutants deficient in homologous recombination repair, especially to FANCD2 and FANCD1 (BRCA2). This increased sensitivity was not dependent on intracellular GSH levels indicating that an impairment of the major pathway of formaldehyde detoxification is not responsible for this effect. Nucleotide (NER), base pair excision repair (BER) and some other repair mechanisms were not heavily involved in the elimination of DNA lesions. Hypersensitivity against formaldehyde was found at concentrations between 10-15 µM. These concentrations are in the very low range or below of those of endogenous formaldehyde. In addition, cells of human origin showed similar hypersensitivity to formaldehyde. It was concluded that endogenous formaldehyde plays a critical role in the predisposition to malignant tumors in Fan patients.

Rosado et al. (2011, key) reported that viability of human FANCB knockout cells is reduced by formaldehyde associated with increased chromosomal aberrations. They further investigated upstream and downstream components of the Fan repair pathway. Cells deficient for these components showed high sensitivity to formaldehyde, while other pathways of DNA repair were largely resistant to formaldehyde with exception of XAP showing mild sensitivity. But in contrast to Ridpath et al. (2007) they did not observed that interstrand crosslinks are involved in formaldehyde induced damage. Formaldehyde is generated within the nucleus by histone demethylation and is detoxified by FDH (ADH5). Cells deficient in FANCD2 and ADH5 gene stopped growing after 3 days and cell survival improved by removing endogenous formaldehyde with ß-mercaptoethanol. It is concluded that inactivation of formaldehyde catabolism by disrupting ADH5 results in cell lethality in cells deficient in upstream (FANCL) or downstream components (FANCD2) of the Fan pathway.

Ren et al. (2013, supporting) compared human lymphoblast cell lines sufficient and deficient for FANCD2, the latter derived from a Fan patient. After treatment with formaldehyde (0-150 µM), cytotoxicity, DPC, micronuclei, chromosomal aberrations, and apoptosis showed larger increases in the deficient as compared to the sufficient cells. These findings support the conclusion that FANCD2 protein is essential to protect human lymphoblastoid cells against formaldehyde induced toxicity.

Neuss et al. (2010b, supporting) studied gene expression in primary human nasal epithelial cells exposed in vitro to formaldehyde. Repeated treatments with low formaldehyde concentrations (20 and 50 µM) did not lead to a significant induction of DPC, but at 50 µM the expression of more than 100 genes was changed. The altered genes were not involved in the main pathways for formaldehyde detoxification or repair of DPC. There was a substantial overlap of the gene categories affected in this study with those of the in vivo study of Andersen et al. (2008). As these repeated treatments with time intervals of 24 h did not lead to an increase of DPC it was concluded that DPC are continuously removed without accumulation in vitro.

Cheah et al. (2013, supporting) studied gene expression profiles of aldehydes known to occur in tobacco smoke in human lung A549 cells. Exposure concentration leading to 20-25% cell death were selected for an exposure duration of 2 h; formaldehyde was tested at a somewhat higher concentration as compared to Neuss et al. (2010b), i.e. 83.2 µM. In comparison to the other aldehydes tested, formaldehyde led to the highest number of affected genes, 42 genes being up- and 25 down-regulated. Upregulated genes were mainly associated with apoptosis, DNA damage response, stress response and transcriptional regulation and the the down regulated with transcriptional regulation.

Genotoxicity in vivo:

Systemic effects in laboratory animals:

Standard cytogenetic and micronucleus assays on systemic effects in samples of bone marrow or peripheral blood after oral or inhalation exposure in experimental animals revealed negative results.

In the cytogenetic assay of Kligerman et al. (1984) 6 F344 rats per group were exposed 6 h per day for 5 days to 0, 0.5, 6, or 15 ppm but chromosome aberrations were only determined at 0 and 15 ppm. One h after the last exposure blood samples were taken by cardiac puncture and lymphocytes extracted, cultivated for 54 h and stained for chromosome analysis. Under the experimental conditions of this study, formaldehyde did not induce increases in chromosomal aberrations in lymphocytes of rats at a dose level of 15 ppm, indicating no systemic activity. The mitotic index did not reveal cytotoxicity. The parallel SCE assay gave also negative results.

In the standard micronucleus assay (Morita et al., 1997) 5 male CD-1 mice per group were gavaged with 0, 100, 200 mg/kg bw (for studying micronucleated polychromatic erythrocytes in bone marrow, two application, interval 24 h) or with 25, 50, 100, 200 mg/kg bw (for studying micronucleated reticulocytes in peripheral blood; single application). 0, 24, 48, 72 h after application (peripheral blood) or 24 and 48 h after the last application (bone marrow) mice were killed and tissues processed for scoring micronuclei incidences. No substance-related increase in micronuclei in any treatment group in bone marrow and peripheral blood at dose levels up to 200 mg/kg bw (MTD) was detected. In additional experiments (2 x i. v. 0, 10, 20, 30 mg/kg bw, 60% of lethal dose) also no induction of micronuclei in reticulocytes of the peripheral blood was observed after i. v. injection.

Negative results were also reported by Natarajan et al. (1983) investigating micronuclei and chromosomal aberrations in bone marrow cells and chromosomal aberrations in spleen cells. Mice were treated intraperitoneally with doses up to 25 mg/kg bw.

Similarly, after exposure of rats by inhalation up to 15 ppm over 1 and 8 weeks (6 h/d; 5 d/week) no significant increase in chromosomal abnormalities was observed in bone marrow cells (Dallas et al., 1992).

In the most valid micronucleus (MN) study (Speit et al., 2009) 6 male F344 rats per dose level were exposed 6 h/day, 5 days/week for 4 weeks to 0, 0.5, 1, 2, 6, 10, or 15 ppm formaldehyde. Positive controls (n=6) received twice orally, with an interval of 24 h, 10 mg/kg bw/day cyclophosphamide. After the exposure period peripheral blood samples were collected and the percentage of reticulocytes (RET) among all erythrocytes (RET plus normochromatic erythrocytes, NCE) and the percentage of reticulocytes with MN (MN-RET) were determined by flow cytometric analysis. Toxicity: No clinical signs were detected but the body weight gain was reduced at >= 10 ppm. MN: The positive control was valid. Biometric analysis showed no statistically significant difference between the mean micronucleus frequencies in the treated groups and the negative control group; also no significant difference were seen in the trend test.

In the parallel comet assay (normal alkaline procedure and after gamma-irradiation) (Speit et al., 2009) the results excluded a DNA strand breaking effect, the induction of alkali-labile sites (ALS) as well as a cross-linking effect of formaldehyde in the blood of rats exposed to formaldehyde by inhalation in a 28-day study at concentrations up to 15 ppm (no systemic genotoxic effects). Furthermore, this treatment of rats resulted also in no statistically significant difference between the mean SCE frequencies of peripheral blood cells in the treated groups and the negative control group (Speit et al., 2009).

No systemic gene mutagenic activity was detected in the mouse spot test (Jensen et al., 1982) when dams received dynamic exposure for 6 h/day at gestation day 8, 9 and 10 at concentration levels of 0, 6, or 18 mg/m³ (0, 5, or 15 ppm).

An oral rodent dominant lethal assay of limited validity was also negative, while i. p. studies (limited validity) gave positive results (see supporting data gentox in IULID Section 7.6.2). Studies in Drosophila on different endpoints gave positive results after feeding of formaldehyde; however, the relevance for mammals is questionable.

In a recent study by Lu and co-workers (2010) on DNA adduct formation male F344 rats were exposed once (6 h) or 6 h/day for 5 days to 10 ppm [13CD2]-formaldehyde. The applied method allowed differentiation of DNA adducts and DNA-DNA cross-links originating from endogenous and inhalation-derived formaldehyde exposure. Tissue samples for DNA analysis were nasal respiratory epithelium (local effects presented below) from the right and left sides of the nose and from the septum; entire tissues of spleen, thymus, lung, liver, and bone marrow., No mono-adducts (-dG or –dA) or DNA-DNA crosslinks were formed by exogenous formaldehyde in sites remote to the portal-of entry, even when five times more DNA was analyzed. Endogenous dG-CH2-dG cross-links were present in all tissues as well as hydroxymethyl-dG and –dA mono-adducts. The study does not support the biological plausibility that inhaled formaldehyde causes leukaemia.

A literature search after the last IUCLID update was carried out up to April 20, 2015 and provided the following new information:

Lu et al. (2011, key) increased the sensitivity of the approach of Lu et al. (2011) and did not find exogenous DNA adducts in the bone marrow of rats exposed once over 6 h to 15.2 ppm. Taking into consideration the sensitivity of the method and the amounts of endogenous adducts they concluded that less than 1 exogenous adduct would be present in 1500 endogenous adducts at 15.2 ppm. It was considered highly implausible that this one additional exogenous adduct could induce malignant transformation in the bone marrow if 1500 endogenous adducts do not. These considerations prompted a letter to the editor reinforcing the implausibility of leukaemia induction by formaldehyde inhalation (Lu et al., 2011a, supporting).

By the same method, no exogenous DNA adducts were detected in the bone marrow of cynomolgus macaques exposed to [13CD2]-formaldehyde at 1.9 and 6.1 ppm for 6 h/d over 2 days (Moeller et al., 2011, supporting). Similarly, Yu et al. (2015, key) did not detect exogenous dG-adducts in the bone marrow, white blood cells or 9 further organs (including tracheal bronchial lymph nodes and the trachea) of rats exposed to 2 ppm over 28 days (6 h/d). Also, in the bone marrow, white blood cells, and the proximal trachea including the carina no exogenous adducts were detected in monkeys exposed to 6 ppm over 2 days (6 h/d).

Ye et al. (2013, supporting) exposed BALB/c mice to 0, 0.5, 1, and 3 mg/m3 (8 h/d over 7 consecutive days). DPC levels were significantly increased as measured by the unspecific KCl-SDS method. In addition, parameters of oxidative stress were significantly increased like reactive oxygen species and malonaldehyde while GSH was significantly decreased. All tissues investigated (bone marrow, peripheral blood mononuclear cells, lung, liver, spleen, testes) were affected starting at different doses except for DPC in the lung. Bone marrow was among the organs with the strongest effects. While according to the authors these findings would strengthen the biological plausibility of formaldehyde induced leukaemia, the question how formaldehyde may lead to such effects if it does not enter systemic circulation remains unanswered.

Yu et al. (2014, supporting) studied bone marrow toxicity in mice exposed to 0, 20, 40, and 80 mg/m³. Histopathological alterations of the bone marrow were observed at ≥40 mg/m³. Other effects, some of them already starting at 20 mg/m³, were increases of micronuclei, DNA damage in the comet assay, myeloperoxidase and a decrease of glutathion peroxidase activities. Gene expression of peroxiredoxin 2 was reduced and that of myeloperoxidase increased. The authors suggest that peroxiredoxin 2 expression is involved in bone marrow toxicity.

Katsnelson et al. (2013, supporting) exposed female rats to 10-15 mg/m³ formaldehyde (mean 12.8 mg/m³) (4 h/d, 5 d/week over 10 weeks) and described an increased incidence of micronuclei in polychromatic erythrocytes in the bone marrow.

Indications for oxidative stress were also described by Matsuoka et al. (2010, supporting) in male ICR mice. After a single exposure to 0.1 ppm over 24 h the ratio of 8OHdG/dG was significantly increased in plasma, but was decreased in urine, lung and liver, and remained unchanged in brain. The NO3- levels mirrored the 8OHdG/dG ratios. After 3 ppm over 24 h, NO3- levels in liver and plasma were significantly decreased while the SOD activities in blood and urine were increased without a change in the liver. It is suggested that that formaldehyde inhalation at low concentrations influences oxidative stress response in a tissue specific manner, but the somehow different findings at 0.1 and 3 ppm are difficult to explain.

Zhao et al. (2009, supporting) studied DPC formation in the liver of mice exposed to 0, 0.5, 1, and 3 ppm formaldehyde continuously over 72 h by the unspecific KCl-SDS method. DPC formation was significantly increased at 1 and 3 ppm. At 3 ppm DPC were completely repaired within 12 h. In vitro DPC repair in a liver cell line was clearly slower and took about 18-24h. The authors offer no explanation how formaldehyde after inhalation may reach the liver.

Systemic genotoxic effects in humans:

In a weight of evidence approach it is concluded that there is not sufficient evidence for systemic genotoxic effects in humans.

He et al. (1998) reported weak effects in the micronucleus assay, for chromosomal aberrations and SCE in lymphocytes of 13 students exposed during an anatomy class. Results of further studies are contradictory.

Slightly increased sister chromatid exchange rates (SCE) were observed in 8 students exposed to an average of 1.2 ppm formaldehyde during an anatomy course (Yager et al. 1986, in Greim 2000). No increases in SCE but increased numbers of chromosome aberrations were observed in 20 workers of a paper factory (Bauchinger & Schmid 1985, in Greim 2000).

Six persons exposed to concentrations up to 9.8-11 ppm in a pathological lab showed no clastogenic effects or increased SCEs (Thomson et al., 1984; in Greim, 2000).

Negative results were also reported in 3 other studies on human occupational exposed to formaldehyde (OECD, 2004; IARC, 1995). Evaluation of these studies is hampered by low number of subjects, co-exposure and lack of details (OECD, 2004).

In the study of Shaham et al. (2002; see “review Gentox human” ) SCE was measured in peripheral lymphocytes of 90 formaldehyde exposed workers (hospital pathology) versus 52 unexposed workers (administrative section). The adjusted mean of SCEs was significantly higher among the exposed compared with that of the unexposed group (P <0.01). However, in this study exposure (and co-exposure) was not clearly documented.

In a further study (Shaham et al., 2003; see “review Gentox human” in IUCLID Section 7.10.2) 186 exposed workers from pathology departments were compared with 213 unexposed controls. No details on exposure are given. Exposure estimates were related to ambient air measurements over 15 min at various periods of a typical working day. The exposed group was divided into those with low exposure (mean 0.4 ppm) and those with high exposure (mean 2.24 ppm). The mean exposure period was 15.9 years. The level of DPC was significantly higher in the exposed after adjustment for age, sex, years of education, smoking and origin. In addition pantropic (wild type + mutant) and mutant p53 protein was measured in serum. There was a significant correlation between pantropic and mutant p53. Formaldehyde increased the risk (non-significantly) of having a level of pantropic p53 above a predefined cut-off of 150 pg/ml (OR 1.6; 95% CI 0.8-3.1). The risk of having a high level of pantropic p53 was mainly determined by the levels of DPC: for workers with DPC levels above the median the risk of having pantropic p53 of > 150 pg/ml was significantly higher (OR 2.5; 95% CI 1.2-5.4). The authors conclude that DPC and mutations in p53 may represent steps in formaldehyde carcinogenesis.

Heck and Casanova (2004) have critically assessed the methodological approach of Shaham for determination of DPC by the K-SDS method. The method was validated by reacting human lymphocytes in vitro with concentrations of formaldehyde ranging from 0.005-2 mM. Based on the calibration curve Shaham and co-workers claimed that DPC could be detected at “concentrations as low as 0.001 mM or even lower”. By analyzing the calibration curve given by Shaham et al., Heck and Casanova (2004) concluded that an unequivocal concentration response was only evident at concentrations above 0.3 mM. This higher quantification limit is also supported by other investigators referenced in the Heck and Casanova paper.

An overview on results obtained with studies on systemic genotoxicity in humans is presented by BfR (see Table 5 in BfR, 2006). It is concluded that “there is no sufficient evidence to reject the plausible assumption that formaldehyde does not induce systemic genotoxicity in man. ”

In a study by Orsiere et al. (2006; see “review Gentox human” in IUCLID Section 7.10.2) the DNA damage and the incidence of micronuclei in peripheral lymphocytes of pathology and anatomy laboratory workers (n=59) was analysed (37 controls). Exposure levels were determined by passive air monitoring in the breathing zone of the workers. Short-term sampling (high levels expected) over a period of 15 minutes revealed a mean concentration of 2 ppm (range <0.1 – 20.4 ppm) and sampling during a typical working day (8 h) resulted in a mean of 0.1 ppm (range <0.1 – 0.7 ppm). No difference was found in DNA damage determined by a chemiluminescence microplate assay in exposed workers in the morning and at the end of the shift. However, in the micronucleus test a significant increase in micronuclei was detected in exposed workers. In a subgroup of 18 exposed and 18 control workers the micronucleus test was combined with FISH with a pan-centromeric DNA probe. The incidence of micronuclei containing one centromere was significantly higher in exposed than in controls while there was no difference for micronuclei containing more than one centromere or no centromere. No explanation was given to the (implausible) observation that in controls more micronuclei with more than one centromere were found as compared to the number of micronuclei with only one centromere. While no clastogenicity as a consequence for interaction with DNA was found in this study the authors suggest that formaldehyde leads to aneugenicity by disturbing the spindle apparatus. But such an effect, i. e. aneugenicity not accompanied by clastogenicity, is very unlikely. Finally, the authors did not discuss a possible co-exposure (especially in pathology and anatomy laboratories).

Costa et al. (2008) evaluated the genetic effects of occupational exposure to formaldehyde in a group of 30 Pathological Anatomy laboratory workers and controls (similar group characteristics) for the endpoints cytogenetics (micronuclei, MN; sister chromatid exchange, SCE) and DNA damage (comet assay). The level of exposure to FA was evaluated near the breathing zone of workers (individual TWA determined). The mean level of formaldehyde exposure of the 30 individuals studied was 0.44 ppm (range 0.04–1.58 ppm). No further data on exposure were given (e. g. exposure duration and frequency). The MN frequency was significantly higher in the peripheral blood cells of exposed subjects as well as SCE values. In the comet assay data showed a significant increase of tail length. The validity of this study is restricted by 1) limited data on exposure, 2) no data on possible co-exposure of laboratory workers, 3) methodological shortcomings in the comet assay(only tail length measured; increased tail length difficult to interpret).

In a recent study (Zhang et al., 2010) evidence was given for leukaemia-specific chromosome changes in cultured myeloid progenitor cells of workers which might be related to formaldehyde exposure. In this study 43 workers exposed to formaldehyde and 51 frequency-matched controls were compared; effects on haematopoiesis were examined by measuring complete blood counts and peripheral stem/progenitor cell colony formation; myeloid progenitor cells, the target for leukemogenesis, were cultured to determine colony formation and in addition to quantify monosomy 7 and trisomy 8 in a subset of the 10 most highly exposed workers and 12 controls. The 8 h TWA formaldehyde concentration at the work place was 1.28 ppm (10th and 90th percentile: 0.63-2.51 ppm) for exposed workers and 0.026 ppm in controls. Among exposed workers, peripheral blood cell counts were significantly lowered (total WBC count; counts for myeloid cell types like granulocytes, platelets, and RBC as well as lymphocytes) and this effect was interpreted as being consistent with toxic effects on the bone marrow. The leukaemia-specific chromosome changes (monosomy 7 and trisomy 8) that were claimed to be leukaemia specific were significantly elevated in progenitor cells. Colony formation of progenitor cells was decreased (but not to a statistically significant extent) in exposed workers. Finally colony formation of progenitor cells taken from an unexposed subject were reported to be sensitive to very low formaldehyde concentrations in vitro.

There are several aspects that raise doubt on the validity of this investigation and the conclusions:

The study population was insufficiently characterized 1) the numbers of workers at the plants were not given which means that essential information is missing that is necessary to understand the set-up of the study. 2) it remains unclear how the exposed study group was selected from the population at the plant. Only a rather vague description "about 1 to 2 ppm on most of the days" was given by Zhang et al 2010 as the inclusion criterion. However, such a fuzzy criterion cannot be operationalized (programmed). Thus, the selection of the exposed cannot be reproduced. 3) the basis of the participation rates is unclear: no information is given on the number and on the representativeness of workers monitored for formaldehyde exposure. Because this was a condition for recruitment of the exposed group a relevant selection bias - masked by reporting a high but conditional “participation rate” - cannot be ruled out.4) the jobs the controls held were not described. The only information on the type of work of this group was that they “were engaged primarily in manufacturing”. There was no information on which products they produced / were exposed to, on their type of work (level of physical activity), whether shift work was involved / working times were similar (circadian variation of biological parameters) etc. 5) no definition and no data were given for the “comparable demographic and socioeconomic characteristics”. Differences in these covariates between the groups may have confounded the comparison. 6) Zhang et al 2010 reported to have requested information on medical history and current medications but no specific information is presented. Medication may well have affected the results: there was a clearly higher percentage in exposed vs. unexposed regarding recent respiratory infections (40% vs. 29%). 7) there was also no information given on differences between the five different plants of the study participants and how information varied across plants, e. g., it is unclear whether subsets of exposed and unexposed used for FISH analyses were from two, three, four or five different plants. 8) Formaldehyde exposure levels were only measured under rather recent conditions; no attempt was made to estimate historical exposures. Thus, it is unclear whether the effects described can really be related to the level of Formaldehyde exposure reported. 9) only unadjusted summary measures are reported but although multiple regression analyses were performed taking age, gender and smoking into account as described in the Methods section of Zhang et al 2010. These three variables were used in the adjustment because they changed the estimates relevantly according to the criterion given. Therefore, the unadjusted (crude) estimates presented cannot be taken for granted.

- All blood cell counts are in the normal range, but information on physical activity and time of day for blood sampling are missing albeit both factors influence blood cell counts.

- The origin of the aneuploidies observed remains unclear, i. e. were they already formed in the bone marrow or did they occur only in vitro.

- The incidence of aneuploidy in controls is very high suggesting that mitotic mal-segregation is unusually frequent in these cells or that artefacts occurred.

- Monosomy 7 and trisomy 8 were only scored in pooled cultures from colonies and it is unclear how many colonies were isolated from each blood sample and whether the effects may have been driven by single colonies

- The sensitivity of progenitor cells against formaldehyde is not specific for this cell type and has been shown to be in the same range by Speit et al. (2008) and Schmid and Speit (2007) also for other cell types; thus this finding is nothing more than unspecific in vitro cytotoxicity.

In the literature search after the last IUCLID update some further studies were identified up to April 20, 2015, some of them discussing the question whether the investigation of Zhang et al. (2010) provided mechanistic evidence that formaldehyde might be a human leukemogen.

Hosgood et al. (2013, supporting) investigated the same workers of the Zhang et al. (2010) study to identify the subsets of lymphocytes affected in the previous study. Total natural killer cells and T cells were significantly decreased in the exposed workers and among the T cells, CD8+ T cells, CD8+ effector memory T cells and regulatory T cells showed a significant decrease. The authors caution that these findings need to be independently confirmed due to the small sample size.

Goldstein (2010, supporting) discussed the difference between findings in mechanistic animal and epidemiological studies regarding leukaemia. As possible explanations he offered species differences or that myeloid precursor cells within the nasal mucosa might be the site for leukemogenesis. But cloromas, which are local collections of myeloid tumor cells, are rarely if ever found in the nose. He discussed potential modes of action (including those proposed by Zhang et al., 2010) how formaldehyde could lead to leukaemia and the evidence for these available. In total, he concluded that it is highly probable that formaldehyde causes human leukaemia although the evidence is not sufficiently strong to warrant its classification as human leukemogen. Key for additional substantial proof would be a replication of the findings of Zhang et al. (2010).

Gentry et al. (2013, supporting) provided a detailed analysis of the methods used by Zhang et al. (2010). They found that the protocol for scoring of metaphases was not followed and that in fact monosomy 7 and trisomy 8 were analysed in less samples that required by their protocol. Furthermore, the assays used did not actually measure the events in the stem cells that might be involved in the development of acute myeloid leukaemia. The aneuploidies reported could not have arisen in vivo but rather during the in vitro cell culture up to 14 days. Due to the nature of the assay and the doubling of cells that would occur over the 14-day culturing period, each original stem cell with aneuploidy should give rise to a colony of aneuploidy metaphases. In contrast to the expected development of totally aneuploid colonies, Zhang et al. (2010) reported small numbers of aneuploid cells not from individual colonies, but rather from pooled colonies. Because of the low frequency of aneuploidy and the lack of whole colonies of aneuploid cells, it is unlikely that the aneuploidy arose from a cell that was affected in vivo. Rather, it strongly suggests that the aneuploidy in these cells arose late during the period of in vitro culture and was not caused by formaldehyde exposure in vivo. Cultures from both exposed and unexposed workers had aneuploid cells, consistent with development of aneuploidy in vitro. The authors suggested that factors other than from formaldehyde exposure might have contributed to the effects reported and they concluded that the data of Zhang et al. (2010) do not support a mechanism for a causal association between formaldehyde and myeloid or lymphoid malignancies.

In the table below, the findings on blood cell counts in different studies are listed in comparison to the observations of Zhang et al. (2010). As can be seen, the findings of Zhang et al. (2010) do not correspond to those of several other investigations even under similar exposure conditions (Lyapina et al., 2004). (All these studies are supporting, Klimisch score 2)

Author

Population

(Expo/contr

Level (ppm)

Blood effects

Red cells

Blood effects

White cells

Blood effects

Neutrophils

Lymphocytes

total

Lymphocytes

B

Lymphocytes

T

Lymphocytes

PLT

Thrasher, 1987

Mobile homes (8/8)

0.07-0.117

 

 

 

 

= (a)

- (a)

 

Madison, 1991

3 years after accident (42/29)

2-5 estimate

 

=

 

=

 

 

 

Srivastava, 1992

Workers (6) (b)

No data

- (4/6)

 

 

+(3/6)

 

 

 

Kuo, 1997

Nurses (50/71)

<0.3-0.89

=

=/-

=

=

 

 

=

Ying, 1999

Anatomy students (23) (c)

0.51+/-0.3

 

 

 

 

+(a)

-(a)

 

Lyapina, 2004

Workers (29/21)

0.64-1.92 mg/m³

=

=

 

 

 

 

=

Ye, 2005

Workers (18/23), waiters (12)

0.985+/-0.286 mg/m³

 

 

 

 

+(a)

-(a)

 

Aydin, 2013

Workers (46/46)

0.2

=

=

=

=

=

+

 

Zhang, 2010

Hosgood, 2013

Workers (43/51)

1.28

(0.63-2.51)

-

-

-

-

 

 

=

 

 

-

-

Costa, 2013

Pathology/anatomy (35/35)

0.36+/-0.3 ppm

 

 

 

 

-(a)

=(a)

 

Sancini 2014

Clinical laboratories

Men 44/44

Women 42/42

No data

 

 

+

=

 

 

+

=

 

 

 

+,=,-: increased, unchanged, decreased: (a): only %-change given (no absolute counts); (b): comparison with normal values; (c): pre-exposure served as control;

Some comments on studies in the table

Thrasher, 1987: investigation of immunological parameters in inhabitants in mobile homes with symptoms; absolute cell counts not given, only the percentage of B- and T-lymphocytes

Madison, 1991: investigation of general population with acute exposure 3 years after an accident; increase of CD26 cell count; no effect on cell counts of other subsets

Kuo, 1997: exposed nurses from hemodialysis department compared to ward nurses. Eosinophils, basophils and monocytes no difference to controls. Decrease in white blood cells only at 2nd investigation after 1 year. There was a difference between exposed and controls in the number affected by common cold and allergic rhinitis

Ying, 1999: absolute cell counts not given, only the percentage of B- and T-lymphocytes. Subsets of T lymphocytes also decreased

Lyapina, 2004: there was a negative correlation between exposure duration and red blood cell count, but was a possible influence of age tested as covariate?

Ye, 2005: absolute cell counts not given, only the percentage of B- and T-lymphocytes

Aydin, 2013: no difference in monocytes, increase of NK cells

Hosgood, 2013: same study population as Zhang et al. (2010); CD8+ and NK cells decreased

Costa, 2013: absolute cell counts not given, only the percentage of B- and T-lymphocytes

Tang, 2009: referenced several papers in Chinese language reporting changes in blood cell counts that are not available on PubMed.

Sancini, 2014: in addition, monocytes and eosinophils were also significantly higher in the exposed male workers as compared to their controls, but not in the female workers.

The genotoxic effects in peripheral blood cells including positive results in MN frequency, SCE induction and increased DNA damage in the comet assay are clearly in contrast to the results in experimental animal studies under high exposure conditions (see Speit et al., 2009). In the discussion of the paper published by Speit and colleagues (2009) it is argued that systemic genotoxic effects of formaldehyde reported in human studies lack plausibility because positive test results cannot be expected in peripheral lymphocytes after inhalation exposure. "The reported effects are either test artefacts, chance findings or due to another kind of exposure.” Furthermore, the authors discussed the mechanisms of formaldehyde genotoxicity in vitro and in vivo and gave several arguments against these systemic effects in human studies. For example, an increase in DNA migration in the comet assay (e. g. Costa et al., 2008; increased tail length) is difficult to interpret and to explain by biological mechanisms given the abundance of evidence that formaldehyde predominantly induces DNA-protein-crosslinks (DPC) and DPC lead to a reduction of DNA migration (Speit et al., 2009; see also Merk & Speit, 1998).

From in vitro studies on genotoxic effects induced by formaldehyde in human blood cells came evidence that the above mentioned systemic cytogenetic effects in humans are not related to formaldehyde exposure because the corresponding conditions are not met (Schmid & Speit, 2007; 3 entries).

In conclusion, there is not sufficient evidence for systemic genotoxic effects of formaldehyde.

A literature search after the last IUCLID update was carried out up to April 20, 2015.

As a substantial number of new studies were identified, all these investigations, including the former ones, are summarised in a table and are briefly discussed here.

Systemic genotoxicity in exposed workers (the new studies in this table are supporting studies)

Only the first author is mentioned and the year of publication

Author

Study group

N: exp./contr.

Exp. Level

Endpoints

Result

Thomson, 1984

Pathology

6/5

1.14-6.93 during tasks lasting over 2-4 h/d; peaks up to >11mg/m³

CA

SCE

-

-

Bauchinger, 1985

Paper factory

20/20

No data

CA

SCE

+

-

Yager,1986

Anatomy students

10(after)/10

(before course)

1.2 ppm

SCE

+

Suruda, 1993

Anatomy course over 85 days

29(after)/29

(before course)

14.8 ppmxh (cumulative), peaks up to 4.33 ppm

MN

SCE

+

-

Shaham, 1996

Pathology, anatomy

12/8

No data

DPX

+(e)

Shaham, 1997

Pathology, anatomy

12/8 (for DPX)

13/20 (for SCE)

Measurements over 15 min: mean: 1.46 ppm (peaks up to 3.1)

DPX

SCE

+(e)

+

Shaham, 2002

Pathology

90/52

Measurements over 15 min: 0.04-5.6 ppm

SCE

+(f)

Shaham, 2003

Pathology

186/213

Same as in 2002 study

DPX

+

Ying, 1997

Anatomy students

23(after)/23

(before course)

0.508 ±0.299 mg/m³

MN

-

Ying, 1999

Anatomy students

23(after)/23

(before course)

0.508 ±0.299 mg/m³

SCE

-

He, 1998

Anatomy students

13/10

2.37 ppm (wean)

CA

SCE

CBMN

+

+

+

Ye, 2005

FA factory

18 workers /

16 waiters /

23 students

0.985±0.296 mg/m³

0.107±0.067 mg/m³

SCE

+ (against students)

Orsiere, 2006

Pathology, anatomy

59/37

18/18

TWA: 0.1 (<0.1-0.7)

Peaks up to 20.4 ppm

CBMN

CBMN+FISH

+***

+(d↑)

Pala 2008

Workers in different cancer research laboratories

7/25

5/15

2/17

Workers divided into low (0.005-0.026 µg/m³) and high (0.026-0.269 µg/m³) exposure groups

CBMN

CA

SCE

-

-

-

Costa, 2008(h)

Pathology

30/30

Mean 0.44, range 0.04-1.58 ppm

CBMN

SCE

Comet

+

+

+

Jiang, 2010

Plywood industry

151/112

0.08-6.30 ppm

CBMN

Comet

+

+

Jakab, 2010

Pathology (women)

37/37

Mean 0.9 (range 0.23-1.21) mg/m³

CA

SCE

HPRT mutat.

UDS

Apoptosis

+ (d↓)

-

-

-

+

Zhang, 2010

FA-resin workers

10/12

Median 1.28, 90 percentile 2.51 ppm (g)

Aneuploidy

+

Viegas, 2010; 2012 (i,k)

FA-resin production

Pathology, anatomy

30

 

50

85 total controls

TWA: 0.21; ceiling up to 1.04 ppm

TWA: 0.28; ceiling up to 5.02 ppm

CBMN

-

 

+

Ladeira, 2011(i)

Histopathology

56/85

TWA mean 0.16 (range 0.04-0.51), peaks up to 2.93 ppm

CBMN

+

Santovito, 2011

Pathology

20/16

Mean 0.0727; SE 0.0128 mg/m³

CA

+

Zeller, 2011a

Volunteers

41*

Up to 0.7 ppm; 0.4 + 4 peaks of 0.8 ppm;

4x15 min cycling at 89 W

CBMN

SCE

Comet

Expression of FDH gene

-

-

-

-

Costa, 2011(h)

Pathology/anatomy

48/50

Mean 0.43, range 0.04-1.58 ppm

CBMN

Comet

+

+

Bouraoui, 2013

Pathology/anatomy

31/31

Between 0.2 and 3.4 ppm

CBMN with FISH

+**

Aydin, 2013

MDF production

46/46

0.10-0.33 ppm

Comet

-

Costa 2013 (h)

Pathology/anatomy

35/35

0.36 ± 0.03 ppm

(range 0.23–0.69 ppm)

CBMN

SCE

TCR mutation

+

+

-

Costa 2013a (h)

Pathology/anatomy

38/42

0.35 ppm (calculated TWA) (range 0.18-0.69)

CBMN

+

Ladeira 2013 (i)

Histopathology

54/82

0.16 ppm (min–max: 0.04–0.51

ppm)

CBMN

+

Lin, 2013

Plywood industry

82

58

38

 

62

0.13 mg/m³

0.68 mg/m³

1.48 mg/m³

(range 0.02-2.04)

0.27 mg/m³ (measurement before / after work)

Comet

CBMN

DPC

 

Comet

DPC

+ (a)

- (b)

-

 

+

+

Musak, 2013 (l)

pathology

105/250

0.32 mg/m³

CA

+

Costa, 2015 (h)

Pathology

84/87

0.38 (range 0.08-1.39) ppm

Comet

CA

+ (c)

+ (c)(d↑)

Lan 2015 (j)

Exposed workers

29/32

1.38 ppm (0.78, 2.61 10th, 90th percentile)

Aneuploidy

CA

+

+

CBMN=cytokinesis-block micronucleus assay; CA=chromosomal aberrations; SCE=sister chromatid exchange; UDS=UV induced UDS; DPC: DPC; MN: Micronuclei without cytokinesis-block

* blood sampling before (internal control) and after last exposure

** significant increase only of centromere positive micronuclei (aneugenicity)

*** increased micronuclei predominantly explained by aneugenicity

(a) Group comparison positive

(b) only trend test positive by comparison for number of work years

(c) No association with time of exposure

(d) Aneuploidy increased/decreased

(e) Same data for DPX in Shaham, 1986 and 1987

(f) Large overlap of participants in 2002 and 2003 study as judged by exposure data

(g) Subgroup of the most highly exposed workers of a group of 43 exposed workers; aneuploidy measured in CFU-GM colonies.

(h) Costa et al. (2015) comprises a group of 35 individuals already studied in the pilot study of Costa et al. (2013); whether there also is an overlap with the study population of Costa et al. (2008) or Costa et al. (2011) cannot be ascertained, but there are obvious similarities between the groups. Similarly it is unclear whether and to what extent there is an overlap with the study of Costa et al. (2013a).

(i) A comparison of the control populations or the exposure levels of Viegas et al. (2010) and Ladeira et al. (2011 and 2013) indicate a substantial overlap. Therefore these studies may not be considered completely independent.

(j) This study investigated circulating myeloid progenitor cells from exposed workers. The authors point specifically to the fact that increased levels of monosomy 7 and chromosomal aberrations of chromosome 5 are frequently observed in acute myeloid leukaemia. They suppose that his may be taken as an indication for a causal link between formaldehyde exposure and induction of myeloid leukaemia.

(k) The same study has been published by Viegas et al. (2013) giving more detailed analytical data, especially for peak exposures (called “ceilings”): peaks ranged for the different exposure groups from 0.34 to 5.02 ppm. Also the same data are reported by Viegas et al. (2012).

(l) The total study comprised 247 subjects in anaesthesia, 249 in oncology departments, and 105 in pathology departments. The odds ratios (95% CI) for chromatid + chromosome type aberrations were for anaesthetics 3.9 (2.7-5.8), cytostatics 2.7 (1.9-3-9), and formaldehyde 1.7 (1.1-2.7). Separate analyses showed increased odds ratios for both types of aberrations for anaesthetics and cytostatics, but for formaldehyde only a borderline significance for chromosome type aberrations (OR 1.6; 1.0-2.5) was observed.

For the study of Santovito et al. (2011) the low exposure concentrations of formaldehyde and the small difference between exposed and control subjects (72.2 vs. 36.4 µg/m³) is noted. The authors also investigated a possible influence of GSTM1 and GSTT1 gene polymorphism. These polymorphisms did not affect the levels of cytogenetic damage,

Some further studies are mentioned with a Klimisch score of 4. In a study on a small number (20) of nurses exposed to cytostatic drugs, anaesthetics, FA and other sterilising gases, elevated sister chromatid exchange counts in blood cells were observed vs. a control group. Quantitative exposure data were not given (Santovito et al 2014). Therefore, no conclusions may be drawn regarding a specific effect of FA. Musak et al (2013) studied chromosomal damage among 601 medical professionals occupationally exposed to volatile anesthetics, antineoplastic agents, and formaldehyde. An increased frequency of chromosomal aberrations was associated with exposure to anesthetics, cytostatics, and formaldehyde, but exposure levels for these agents are not given. Peteffi et al. (2015) compared 46 workers exposed to low levels of formaldehyde (0.03-0.09 ppm, 8 h TWA, measured in 7 sectors of the working area) to 45 referents (mean exposure 0.012 ppm). DNA damage was statistically significantly increased in the exposure group. Also urinary formic acid, serving as biomonitoring index, was significantly increased in the exposed. This latter finding is highly improbable taking into account the small difference of the exposure levels and the report of Gottschling et al. (1984) who did not detect significant changes in formate excretion over a 3-week period of exposure to FA at a concentration in air of less than 0.4 ppm. Souza and Devi (2013) studied micronuclei frequency by the cytokinesis-block micronucleus assay (CBMN) in 30 formaldehyde exposed workers in forensic medicine compared to 30 controls. Micronuclei were significantly elevated in the exposed group and correlated with duration of exposure. No information is given on extent of exposure. Gomaa et al. (2012) compared 30 formaldehyde exposed workers to 15 unexposed controls. Airborne formaldehyde concentrations are not given. Chromosomal aberrations and DNA migration in the comet assay were significantly increased in peripheral lymphocytes of exposed workers.

As can be seen in table 1 many of the older studies (<2000) only comprised small study populations (apart from Shaham et al., 2002, 2003) and positive as well as negative results were obtained. More recent investigations often report on larger groups and positive genotoxic findings predominate. The cytokinesis-block micronucleus test (CBMN) and the comet assay are the methods most frequently applied.

An assessment of these findings predominantly has to take into consideration that after inhalation in experimental animals FA does not reach systemic circulation as recently confirmed by Kleinnijenhuis et al. (2013) nor does it lead to DNA adducts (Lu et al., 2010, 2011: Moeller et al., 2011) or DPC, SCE or micronuclei (Speit et al., 2009) in organs distant from the site of first contact or in the blood. Therefore these findings lack biological plausibility and “were not considered (by RAC, 2012) for inclusion in the discussion on classification of FA.” This mechanistic argument is still valid for the interpretation of the new studies. In addition further points have to be taken into consideration:

1)The reliability of the scoring of micronuclei in the CBMN appears questionable. For instance, Ladeira et al (2011) claimed a moderately positive correlation between micronucleus frequency in peripheral blood lymphocytes and the duration of FA exposure. However, a blinded re-evaluation showed that repeated measurements of the same slides were highly variable not only between two scorers, but also when slides were evaluated by the same scorer (Speit et al 2012) at different times.

2)The applicability of the CBMN to human biomonitoring has been severely challenged by Speit et al. (2012a) and Speit (2013, 2013a) based on mechanistic grounds. While the CBMN is well suited for in vitro testing of mutagenicity, the in vivo method should be rather insensitive for the detection of mutagens/clastogens (Speit et al., 2012a). Thus the reliability of positive results obtained with the CBMN in human biomonitoring is questioned because “it is highly unlikely that DNA damage induced by exposures toward environmental and occupational chemicals in vivo leads to increased micronuclei frequencies” in the CBMN (Speit, 2013).

3) Only the investigations of Orsiere et al. (2006) and Bouraoui et al. (2013) differentiated by FISH staining whether the micronuclei scored were derived from clastogenicity or aneugenicity. In both studies the micronuclei predominantly contained the centromere indicating to the latter mechanism. But in vitro data of Speit et al. (2011b) and Kuehner et al. (2012, 2013) clearly demonstrated that FA predominantly leads to clastogenicity and not to aneugenicity. Therefore the CBMN results obtained by FISH staining again lack biological plausibility.

4) The induction of increased DNA migration as described in human biomonitoring studies also lacks plausibility. Speit et al. (2007) have shown in vivo, that FA only leads to DPC (with decreased migration) and no increases have been observed down to concentrations by a factor of 10,000 below those at which crosslinking begins.

5) And finally, the relevance of positive SCE and micronuclei findings in biomonitoring studies have been questioned by Speit et al. (2009) mainly because DPC present at the start of lymphocyte culture are removed during cell culture before lymphocytes start to replicate (Schmid and Speit, 2007).

In conclusion, in spite of the new publications the previous assessment of RAC (2012) is still valid. These biomonitoring studies, based primarily on mechanistic considerations, cannot be taken as proof that FA leads to systemic genotoxicity in exposed workers. This assessment is supported by the negative results obtained with human volunteers in the study of Mueller et al. (2013). Under these conditions, Zeller et al. (2011a) did not observe genotoxic effects in the CBMN, the comet assay, and the SCE test in blood samples taken after the last exposure.

Apart from studying genotoxicity in the volunteers of the Mueller et al. (2013, supporting) study, Zeller et al. (2011a) also investigated the expression of the GSH-dependent FDH (ADH5) and performed DNA microarray analyses in blood samples using a full-genome human microarray. No exposure related effects on the expression of the FDH gene or alterations by the microarray analyses were noted.

Apart from the increased incidence of micronuclei in lymphocytes as shown in the table, Ladeira et al. (2013, supporting) also investigated the potential influence of polymorphism. They found that single nucleotide polymorphisms (SNP) of XRCC3 that participates in DNA double-strand break/repair were associated with significantly increased incidences of nuclear buds in lymphocytes. But polymorphism of ADH5 (FDH), the enzyme most important for formaldehyde detoxification, had no effect.

Apart from studying genotoxicity, Jiang et al. (2010, supporting) also genotyped glutathione-S-tranferases by multiplex PCR. They reported that the positive findings in the comet assay were slightly higher in workers with the GSTM1 null genotype (p=0.07). Also, workers with the GSTP1 codon 105 Val allele had a slightly higher micronuclei rate than those with the wild type allele. In control subjects GST polymorphism had no correlation with the genotoxicity markers investigated. The authors concluded that polymorphism of GST genes may modulate genotoxicity of formaldehyde.

Local genotoxic effects in laboratory animals:

In animals formaldehyde showed local genotoxic effects only at the site of first contact after oral exposure, no valid data are available on the inhalation route. However, concerning premutagenic lesions like DNA-protein cross-links there are indication that the dose-response is nonlinear possibly due to the saturation of molecular defence mechanisms. In several studies on rats and monkeys the formation of DNA-protein cross-links (DPC) at the site of first contact has been demonstrated after inhalation exposure. DPC is discussed as premutagenic lesion. DPC formation was detected already at low concentrations (0.3 ppm in rats). The dose response was nonlinear with a steep rise at higher dose levels suggesting saturation of defence mechanisms and showed coincidence with damage of epithelium. Rapid removal of DPC has been reported in corresponding experiments (half-lives 2-4 h) and no accumulation was detected in in vivo studies.

In the study of Casanova et al. (1989) four male Sprague-Dawley rats per group were exposed to 14C-labelled formaldehyde (0.3, 0.7, 2, 6, or 10 ppm; single 6 h exposure period). Rats were killed immediately after exposure and nasal respiratory mucosal tissue removed and tissues of 4 rats per dose combined, homogenized and DNA isolated and prepared for analysis of DPC. Shape of dose response curve of DPC was nonlinear indicating possible saturable metabolic defence mechanisms. The detoxification pathway (oxidation of inhaled formaldehyde) is half saturated at an airborne formaldehyde concentration of 2.6 ppm. Covalent binding to DNA was noted even at the lowest concentration of 0.3 ppm (1.4+-0.6 pmole/mg DNA). In rats DPC were induced in vivo at the predilection sites for tumor development of the nasal mucosa after inhalation exposure (Casanova et al., 1994). When rats were exposed to 0.7, 2, 6, or 15 ppm either without or with preexposure to the same concentrations over 11 weeks /6 h/d; 5 d/week) no accumulation of DPC was found due to the rapid removal of DPC (Casanova et al., 1994). The authors concluded that DPC produced in any single day’s exposure must be completely or almost completely removed by the time of the next day’s exposure, i. e. within at most 18 h.

In further studies of this working group (Casanova et al., 1991) it was reported that DPC are formed in the nasal mucosa of monkeys at concentrations >= 0.7 ppm (single exposure, 6 h) and in other tissues of the respiratory tract at dose levels >= 2 ppm. There is no study available demonstrating mutagenic effects in the nasal cavity of experimental animals after inhalation. No local clastogenic effects but histopathological effects and cell proliferation in the nasal cavity of 3 rats has been shown after repeated inhalation exposure to 20 ppm (only one dose level tested) for 6 h on each of 5 consecutive days (BASF AG, 2001).

Six male F344 rats per dose level were exposed 6 h/day, 5 days/week for 4 weeks to 0, 0.5, 1, 2, 6, 10, or 15 ppm formaldehyde. Genotoxic effects in broncho-alveolar lavage (BAL) cells were investigated (Neuss et al., 2010). No increase was observed for micronuclei or for DNA strand breaks and alkali-labile sites (standard alkaline comet assay) or DPC formation (modified comet assay with gamma-irradiation). MMS (once orally) as positive control substance for DNA strand breaks and alkali-labile sites led to an increased DNA migration in the standard comet assay as expected. This study could not reproduce the clastogenic effect on rat lung-lavage cells reported by Dallas et al. (1992; limited validity).

Recio et al. (1992) investigated point mutations of the p53 gene in nasal squamous cell carcinomas induced in the rat bioassay of Kerns et al. (1983). Point mutations in the p53 complementary DNA sequence were found in 5 of 11 tumors analyzed. One of these occurred at a mutational hot spot (rat codon 271; CGT-CAT). But as these mutations were identified in tumor tissues it is unclear whether they were directly induced by formaldehyde or developed secondarily in the course of prolonged cell division.

Recently, (Meng et al., 2010) in a sub-chronic inhalation study at exposure concentrations of 0, 0.7, 2, 6, 10 and 15 ppm (6 h/d; 5 d/week over 13 weeks) DNA was isolated from sites of the nasal mucosa where the incidence of squamous cell carcinomas was greatest in the carcinogenicity bioassay. The p53 codon 271 CGT to CAT mutant fractions were determined by ACB-PCR measurements. The geometric mean mutant frequencies were very similar at all exposure levels and there was no dose response relationship. It was concluded that 13 weeks of exposure up to 15 ppm did not cause a significant p53 codon 271 CGT to CAT mutation. In addition, no GGT to GAT mutations at K-Ras codon 12 were found at any exposure concentration, including the 0-ppm control group. In contrast to the mutational endpoints, the percentage of proliferating cells did increase with FA exposure. Statistically significantly increased cell proliferation was observed at 10 and 15 ppm. Therefore the findings of Recio et al. (1992) are to be regarded as secondary to increased cell proliferation. These results are consistent with a mode of action for tumor formation driven by cytotoxicity and compensatory cell proliferation and not by mutations.

When rat nasal tissue was analyzed in the study of Lu et al. (2010) (see above, systemic effects in laboratory animals) exogenous formaldehyde induced N2-hydroxymethyl-dG mono-adducts and dG-dG cross-links in DNA from rat respiratory nasal mucosa in similar amounts as produced by endogenous formaldehyde. While the amount of endogenous hydroxymethl-dA mono-adduct roughly corresponded to that of the endogenous –dG mono-adduct no exogenous –dA mono-adduct was found. This study supports the local carcinogenic activity of formaldehyde but the relationship to the mutational activity of the adducts measured at 10 ppm is still unclear.

After gavage formaldehyde induced local clastogenic effects at a dose level inducing also local irritation (Migliore et al., 1989). In this study 5 male Sprague-Dawley rats per group were gavaged with 0 or 200 mg/kg bw formaldehyde. The rats were sacrificed 16 h, 24 or 30 h after the treatment and tissues of forestomach, duodenum, ileum, colon processed for scoring micronucleated cells and cells with other nuclear anomalies. The mitotic index was not significantly altered in any tissue by the treatment. The test substance did induce an increased number of micronuclei more pronounced in the forestomach than in the intestine and more pronounced 30 h after treatment than after 16 h. The same results were obtained concerning total nuclear anomalies. Local irritation in the gastrointestinal tract was found 30 h after treatment.

A literature search after the last IUCLID update was carried out up to April 20, 2015 and provided the following new information:

Speit et al. (2011a, key) exposed male Fischer 344 rats (6/group) to concentrations of 0.5-15 ppm formaldehyde (6 h/d, 5 d/week over 4 weeks) and investigated micronuclei formation in nasal epithelial cells. No increase was noted and even after an oral dose of cyclophosphamide the rate of micronuclei was not increased when the animals were studied 1-28 days after application. Nasal histopathology and cell proliferation showed effects as to be expected at exposures of 6, 10, and 15 ppm.

Lu et al (2011, key) measured endogenous and exogenous N2-hydroxymethyl-dG adducts in nasal DNA of rats exposed to 0.7, 2, 5.8, 9.1 or 15.2 ppm 13CD2 for 6 hours by a highly sensitive ultra-performance liquid chromatography tandem mass spectrometry method. Exogenous FA DNA adducts were formed in a highly non-linear fashion, with a 21.7-fold increase in exposure causing a 286-fold increase in exogenous adducts. Endogenous DNA adducts dominated at low exposures, comprising more than 99 % of total adduct levels. In this context, it was demonstrated that N2-hydroxymethyl-dG was the primary DNA adduct formed in nasal cells following FA exposure while endogenous FA also led to the corresponding dA-adducts in amounts comparable to endogenous dG-adducts. Also in monkeys exposed to 2 or 6 ppm, 6 hours/day for 2 days, the external FA-dG adduct was only detected in the nose and not in the bone marrow. At 6 ppm, the FA-dG adduct level was lower in the nasal tissue in the monkeys than in rats with a single 6-hour exposure, suggesting a lower risk in primates than in rats (Moeller et al. 2011).

Swenberg et al. (2011, key) compared endogenous and exogenous FA induced DNA-dG adducts in the nasal tissue of primates and rats. Exogenous adducts in monkeys after 2 days of exposure were similar to those of rats exposed for 1 day at 2 ppm and were ~2.5 times lower in monkeys at 6.1 ppm for 2 days compared to rats at 5.8 ppm for 1 day (6 h/d). These data demonstrate that exogenous adducts formed in the nasal turbinates are lower for nonhuman primates that for rats. In addition, there are indications that endogenous dG adducts are 2-3-fold higher in monkeys than in rats. This reduces the ratio of exogenous/endogenous adducts in primates exposed to low FA concentrations by a factor of ~5.

Yu et al. (2015, key) determined formation, accumulation, and hydrolysis of endogenous and exogenous FA DNA adducts in rats after exposure to 2 ppm over 28 consecutive days (6 h/d) followed by a 7 day post-exposure period. Monkeys were exposed to 6 ppm on 2 consecutive days (6 h/d) and DNA dG adducts were measured in different parts of the respiratory tract. Again exogenous DNA adducts were only found in nasal tissue of rats and monkeys. In the lower respiratory tract no exogenous adducts could be measured in the trachea or carina (monkeys). The exogenous dG FA adducts in rats approached a steady state concentration during the 28 d exposure period with a rapid loss of nearly 20% during the first 6 h post exposure followed by a much slower decrease thereafter. The half-life for formation and loss of the exogenous adducts was estimated to be 7.1 days. Combining the data for monkeys in the present study with those of Moeller et al. (2011) showed that exogenous adducts in different sections of the nasal epithelium were always 5-11-fold lower than endogenous adducts.

Yu et al. (2015) also studied the relationship between the formation of FA DNA adducts and DPC. Based on the investigations of Lu et al. (2009, 2010a) (see above) Yu et al. (2015) showed that the N2-dG-methylene adducts with cysteine and GSH were unstable at physiological pH and room temperature with a half-life of 11.6 min and 79.6 min, respectively. Cleavage occurred at the methylene-S-bond but not at the N2-dG-methylene bond leading to the N2-hydroxymethyl-dG adduct identified in former investigations (Lu et al., 2010, 2011; Moeller at al., 2011). These results suggested that DPCs may be important sources of FA induced DNA mono adducts.

In the light of the instability of FA induced DPCs the authors questioned the reported increase of DPCs after FA exposure in circulating lymphocytes in workers (Shaham et al., 1996, 2003) or in several tissues of mice (Ye et al., 2013). They proposed that these unexpected findings may be due to the use of non-specific DPC assays that cannot differentiate between exogenous and endogenous FA induced DPCs.

At an exposure level of 2 ppm Yu et al. (2015) have shown that FA-dG adducts accumulate to reach a steady state after 28 days. By combining the data of Lu et al. (2011) for a single exposure to 0.7 and 2 ppm with those of Yu et al. (2015) at 2 ppm over 28 days, the exogenous steady state DNA adduct levels at 0.7 ppm may be approximated. Exogenous adducts at 2 ppm, single exposure, were 0.19 adducts/107 dG and after 28 days 1.05 (factor of 5.5). At 0.7 ppm, single exposure, 0.039 exogenous adducts/107 dG were found and therefore at steady state after 28 days of 0.21 adducts/107 dG might be expected. A direct comparison with endogenous adducts is somehow hampered because there was a difference between both of the studies: mean endogenous adducts 4.57 adducts/107 dG for Lu et al. (2011) and 2.91 for Yu et al. (2015). But the steady state exogenous adducts of about 0.2 adducts/107 dG were by a factor of 14 or 22 lower than the endogenous adducts. In addition, these exogenous steady state adducts were always within the standard deviations of both studies (Lu et al., 2011; Yu et al., 2015). Taking into account the low dose non-linearity of the response curve for exposures below 0.7 ppm (for example at 0.3 ppm) a more than proportional decrease of exogenous adducts is to be expected.

By in vitro experiments with TK6 cells Edrissi et al. (2013, supporting) had shown that endogenous formaldehyde is a major source of N6-formyllysine (FA-Lys), an abundant protein modification in cells. This adduct is evenly distributed among different classes of histone proteins. In rats, exposures to isotope labelled FA (13C2H2O) at 0.7, 2, 6 and 9 ppm for 6 hours were used in differentiating between adducts from exogenous and endogenous FA-Lys adducts in the total, the cytoplasmic, the membrane and the nuclear proteins. After proteolysis and analysis of FA-Lys, the ratio between exogenous and endogenous adducts was shown to increase with increasing exposure; for example for the total nasal epithelial proteins, the ratio was 0.035, 0.14, 0.15 and 0.40, respectively. At each of these FA exposures, the ratios were in the order cytoplasmic ≈ membrane > soluble nuclear > chromatin protein bound formaldehyde, indicating a decrease in the exogenous FA-Lys concentration from the cytoplasmic to the nuclear proteins. Opposite, the endogenous FA-Lys adducts were similar at all exposure concentrations in all cellular compartments. Also, this indicated that the external FA exposure did not influence the endogenous FA production. No external FA-Lys adducts were detected in the lungs, liver and bone marrow and thus, the results paralleled studies on FA-dG adducts, confirming that direct external FA adducts are limited to the nasal epithelium (Edrissi et al. 2013a, key).

Andersen et al (2010, key) combined studies with different FA exposure levels and exposure duration with toxicokinetic modelling for tissue FA acetal and glutathione levels and with histopathology and gene expression in nasal epithelium from rats exposed to 0, 0.7, 2, 6, 10 or 15 ppm FA 6 hours/day for 1, 4 or 13 weeks. At 0.7 and 2 ppm FA, the cellular levels of FA acetal showed a very minor increase with exposures and GSH a very minor decrease; several ppm FA would be required to achieve significant changes. Treatment-related nasal lesions were found in the respiratory epithelium at 2 ppm FA and higher. Patterns of gene expression varied with concentration and duration. At 2 ppm, sensitive response genes associated with cellular stress, thiol transport/reduction, inflammation and cell proliferation were up-regulated at all exposure durations. At 6 ppm and higher, gene expression changes showed enrichment of pathways involved in cell cycle, DNA repair, and apoptosis. ERBB, EGFR, WNT, TGF-β, Hedgehog, and Notch signalling were also enriched. Benchmark doses for significantly enriched pathways were lowest at 13 weeks. Seven genes were combined in a grouping referred to as the “Sensitive Response Genes”, showing Benchmark Dose around 1 ppm for all three exposure periods. Transcriptional and histological changes at 6 ppm and greater corresponded to dose ranges in which the toxicokinetic model predicted significant reductions in free glutathione levels and increases in FA acetal levels. Genomic changes at 0.7–2 ppm likely represent changes in extracellular FA acetal and glutathione levels. DNA replication stress, enhanced proliferation, squamous metaplasia, and stem cell niche activation appear to be associated with FA carcinogenesis. It was concluded that dose dependencies, high background levels of FA acetal, and nonlinear FA acetal/glutathione tissue kinetics indicated that FA concentrations below 1 or 2 ppm would not increase the risk of cancer in the nose or any other tissue, or affect FA homeostasis within epithelial cells. Overall, this conclusion is in agreement with a histologic NOAEC of 1 ppm for a 2-year inhalation in rats (Woutersen et al 1989, Gelbke et al 2014).

Although not directly related to genotoxicity, in this section recent studies on microRNAs (miRNA) are described. Such effects are not mutations but represent epigenetic changes by post-transcriptional regulation of gene expression including altered DNA methylation, histone methylation, or regulation of gene expression by binding to mRNA. If miRNA is increased the targeted mRNA and protein can be reduced and vice versa (Swenberg et al., 2013).

Rager et al. (2011, supporting) studied miRNA patterns in vitro by microarray analysis after exposure of human lung A549 cells to formaldehyde in air at 1 ppm over 4 h. miRNA expression was verified by quantitative real-time polymerase chain reaction and the enriched biological functions were determined. Cytotoxicity and IL-8 also were measured. 89 miRNAs were significantly down-regulated and it was concluded that formaldehyde potentially alters signalling pathways associated with cancer, inflammation, and the endocrine system. IL-8 release was also increased and lethality of the cells was minimally increased.

Rager et al. (2013, supporting) studies miRNA profiles in the nose of cynomolgus macaques after exposure to formaldehyde at 0, 2, and 6 ppm over 2 consecutive days (6 h/d). >500 miRNAs were analysed. 3 and 13 miRNAs were dysregulated at 2 and 6 ppm, respectively. The greatest increase was found for miRNA-125b involved in apoptosis signalling and the greatest decrease for miRNA-142-3p that shows altered expression in nasopharyngeal cancer. It is concluded that formaldehyde exposure likely influences apoptosis.

A similar analysis was carried out by Rager et al. (2014, supporting) in rats exposed to 0 or 2 ppm for 7, 28 or 28 d followed by a 7 d recovery period. miRNA profiles were assessed in respiratory epithelium, circulating white blood cells and bone marrow. In the nose 84 miRNAs showed increased or decreased response after 7 d of exposure, 59 after 28 d, and 0 after recovery. The corresponding figures for white blood cells were 31, 8, and 3, while no effects were noted in the bone marrow. In the nose miRNA-10b and members of the let-7 family, known nasopharyngeal cancer players, showed decreased expression. To integrate miRNA responses with transcriptional changes, genome wide mRNA profiles were analysed in the nose and white blood cells. Pathway analyses revealed enrichment of immune system/inflammation signalling. The authors suggest that inflammation in the nose may drive effects in white blood.

Li et al. (2015, supporting) described markedly changed miRNA expression in the olfactory bulb of mice exposed to 3 ppm formaldehyde over 1 or 7 d (6 h/d). After 1 d 25 miRNAs were differentially expressed, and after 7 d 9. Functional annotation analysis showed that the main predicted targets were associated with cancer, transcriptional regulation and axonal growth.

Local cytogenetic effects in humans

From the data available to day it is difficult to come to final conclusion regarding local genotoxic effects in humans. However, there is an indication that formaldehyde may react at the site of first contact.

A human cohort study on workers of a plywood factory revealed also evidence for slight local clastogenic effects in the nasal epithelium in coincidence with increased inflammation (Ballarian et al., 1992).

Further, slight clastogenic effects but not aneugenic effects were shown in students exposed during an embalming course. These effects were more obvious in buccal cells than in cells of the nasal epithelium (Titenko-Holland et al., 1996).

Beside these 2 studies described in detail , further data are available on local effects in nasal or buccal mucosa cells after inhalation exposure. Positive findings were presented in these studies. However, these studies are not fully reliable or not sufficiently documented to derive relevant information on dose-effect relationship (BfR, 2006; see “review Gentox human”). Furthermore, in a review presented by Speit & Schmid (2006) the local chromosome mutagenic effects of formaldehyde in humans after inhalation exposure in the micronucleus test with exfoliated nasal or buccal epithelial cells, the actual targets of formaldehyde were evaluated. The evaluation is based on 8 studies including the two studies described in detail (Titenko-Holland et al.1996; Ballarian et al.1992). The data suggest an increase in micronuclei frequencies in nasal and/or buccal cells after inhalation exposure. However, methodological shortcomings and limited documentation were found and it is yet not possible to assess the local genotoxicity of formaldehyde in humans. But the data have to be taken as an indication that formaldehyde can express its genotoxicity at the site of first contact.

In a recent micronucleus study on local effects by Speit et al. (2007) it was found that formaldehyde inhalation exposure does not induce micronuclei in buccal mucosa of human volunteers. The exposure concentration varied between 0.15-0.5 ppm at constant levels, up to 0.5 ppm with 4 peaks of 1.0 ppm for 15 min each (4 h exposure per day over a period of 10 working days).

An evaluation and summary of the endpoint genetic toxicity is also presented in a recent review by IARC (2006), details are given in Section Mutagenicity - Non human information - in vivo data.

A literature search after the last IUCLID update was carried out up to April 20, 2015 and provided the following new information:

Viegas et al. (2010, supporting) studied 30 formaldehyde resin production workers and 50 workers in pathology/anatomy in comparison to 85 non-exposed controls. Exposure levels were 0.21 ppm (mean TWA) with peaks up to 1.04 ppm for production and 0.28 ppm (TWA) with peaks up to 5.02 ppm for pathology/anatomy. A significantly increased incidence of micronuclei in buccal cells was observed in both exposure groups as compared to controls. The same study has been published by Viegas et al. (2013) giving more detailed analytical data, especially for peak exposures (called “ceilings”): peaks ranged for the different exposure groups from 0.34 to 5.02 ppm. Also the same data are reported by Viegas et al. (2012).

A similar study in 56 histopathology workers again in comparison to 85 controls also showed an increase in micronuclei in buccal cells (Ladeira et al., 2011, supporting). Exposure levels were 0.16 ppm (mean TWA) with peaks up to 2.93 ppm. A comparison of the control population of this study with that of Viegas et al. (2010) indicates a substantial overlap of both studies that was not mentioned in the study. Whether a partial overlap may also relate to the exposed workers remains unclear. However, these 2 studies may not be considered as completely independent.

Ladeira et al. (2013, supporting) found an increased incidence of micronuclei in buccal cells of 54 exposed histopathology workers compared to 82 unexposed controls. Exposure levels were 0.15 ppm (8h TWA) (range 0.04-0.51). A comparison of the control populations or the exposure levels of Viegas et al. (2010) and Ladeira et al. (2011) indicate a substantial overlap. Therefore these studies may not be considered completely independent.

Costa et al. (2013a, supporting) reported an increase in micronuclei in buccal cells in 38 pathology/anatomy workers compared to 42 controls. Mean 8 h TWA concentrations (no details on the sampling procedure) were 0.35 ppm (range 0.18-0.69).

Peteffi et al. (2015, klimisch 4) compared 46 workers exposed to low levels of formaldehyde (0.03-0.09 ppm, 8 h TWA, measured in 7 sectors of the working area) to 45 referents (mean exposure 0.012 ppm. In the micronucleus test with buccal cells there was a significant group difference only for binucleated cells. Also urinary formic acid, serving as biomonitoring index, was significantly increased in the exposed. This latter finding is highly improbable taking into account the small difference of the exposure levels and the report of Gottschling et al. (1984) who did not detect significant changes in formate excretion over a 3-week period of exposure to FA at a concentration in air of less than 0.4 ppm.

Zeller et al. (2011a, supporting) did not find an increased incidence of micronuclei in the nasal cells of 41 volunteers exposed for 4 h over 5 days at concentrations ranging from 0-0.7 ppm (4 h TWA) and at 0.3 and 0.4 ppm each with 4 15 min-peaks of 0.6 and 0.8 ppm, respectively. In addition no formaldehyde related alterations of gene expression were observed by microarray analysis of nasal biopsies. Pre-exposure values served as negative control. The strength of this study and that of Speit et al. (2007) is that the exposure conditions were clearly defined.

Similar to the review of Speit and Schmid (2006), Knasmueller et al. (2011, supporting) concluded that genotoxicity tests in exfoliated human nasal cells need further standardisation of applied methods and/or that sufficient information on the role of confounding factors was lacking for most protocols.


Short description of key information:
Gentoxicity in vitro:
Chromosome mutagenic activity of formaldehyde (and to a much lesser extent gene mutations) is well documented from in vitro studies and numerous studies on other endpoints suggested further evidence for genotoxicity of formaldehyde in vitro. DNA-protein cross-links (DPC) as pre-mutagenic lesion have been sufficiently investigated including threshold and repair. The threshold for DPC formation in cultured human lymphocytes is >10 μM (0.3 μg/mL), significant effects were reported at >=25 μM (0.75 μg/mL); DPC induced by concentrations up to 100 μM (3 μg/mL) are completely removed before lymphocytes start to replicate. There is some evidence that clastogenic effects are related to DPC formation.

Genotoxicity in vivo:
The available data in experimental animals demonstrated the genotoxic activity of formaldehyde at the site of first contact after oral exposure. Studies on local mutagenic effects in humans suggested increased micronucleus frequencies in nasal and buccal cells, however, a final conclusion is not yet possible. The mechanism of clastogenicity might be related to DNA-protein cross-links and their repair. DNA-protein cross-links at the site of first contact have been demonstrated after inhalation exposure in rats and monkeys. There is no clear evidence for systemic genotoxicity in experimental animals or in humans.

Endpoint Conclusion: Adverse effect observed (positive)

Justification for classification or non-classification

Genotoxicity in vitro

Chromosome mutagenic activity of formaldehyde (and to a much lesser extent gene mutations) is well documented from in vitro studies and numerous studies on other endpoints suggested further evidence for genotoxicity of formaldehyde in vitro. DNA-protein cross-links (DPC) as pre-mutagenic lesion have been sufficiently investigated including threshold and repair. The threshold for DPC formation in cultured human lymphocytes is >10 μM (0.3 μg/mL), significant effects were reported at >=25 μM (0.75 μg/mL); DPC induced by concentrations up to 100 μM (3 μg/mL) are completely removed before lymphocytes start to replicate. There is some evidence that clastogenic effects are related to DPC formation.

No co-mutagenicity was observed under simultaneous exposure to formaldehyde and other established mutagens.

Some authors claimed that hematopoetic stem cells are especially sensitive to the cytotoxicity of formaldehyde and that formaldehyde leads to aneugenicity in these cells. This could not be verified by independent studies: toxicity of formaldehyde in such stem cells was similar to that to other cell lines and aneuugenicity could not be verified in stem cells or in other cell lines or by gene expression profiling. It has been shown thin studies zsunf an analytical method with very high sensitivity to differentiate between exogenous and endogenous at the effects of formaldehyde depend heavily on the FANCD1 and FANCD2 repair genes that are deficient in Fanconi Anemia patients.

Genotoxicity in vivo

The available data in experimental animals demonstrated the genotoxic activity of formaldehyde at the site of first contact after oral exposure. Studies on local mutagenic effects in humans suggested increased micronucleus frequencies in nasal and buccal cells, however, a final conclusion is not yet possible. The mechanism of clastogenicity might be related to DNA-protein cross-links and their repair. DNA-protein cross-links at the site of first contact have been demonstrated after inhalation exposure in rats and monkeys. There is no clear evidence for systemic genotoxicity in experimental animals or in humans. There are a large number of new studies on genotoxicity in vivo as summarised below.

Systemic genotoxicity in animals: in studies using an analytical method with very high sensitivity to differentiate between DNA adducts caused by endogenous and exogenous formaldehyde, DNA adducts could only be detected in the nasal epithelium but not in tissues far off the portal of entry in rats and non-human primates. These findings are in contrast to observations mainly in mice that formaldehyde led to genotoxicity and indications for genotoxicity and oxidative stress in the bone marrow and other organs. The different findings in rats and monkeys on the one side and in mice on the other are not resolved yet.

Systemic genotoxicity in humans: a study claimed to have identified a mechanism by which formaldehyde could induce leukaemia. The authors reported in exposed workers aneugenic effects in myeloid stem cells and reduced peripheral blood cell counts similar to findings associated with exposure to benzene. The study was criticised by several authors concerning various methodological aspects and the implausibility of formaldehyde induced aneuploidy. When comparing different publications measuring blood cell counts no consistent pattern could be identified.

A number of studies reported genotoxic effects in the blood of exposed workers measuring micronuclei mainly by the CBMN method, SCE, DPX or chromosomal aberrations. But there are also studies not supporting these findings. The relevance of these positive observations remains unclear for several reasons as there are: formaldehyde does not reach the peripheral blood, similar effects were not observed in guideline animal studies, the reliability of scoring of micronuclei was found to be questionable, by differentiation of micronuclei with and without centromeres they were found to arise mainly from aneugenicity which does not correspond to the primary mode of action of formaldehyde, and no effects were found in a volunteer study under controlled exposure conditions. In addition, the suitability of the in vivo CBMN procedure has been questioned on basic methodological grounds. Also it is unclear how formaldehyde as a typical cross linker may lead to increased DNA migration in the comet assay.

Local genotoxicity in animals: the predominant DNA lesion caused by inhalation of formaldehyde is the N2-hydroxymethyl-dG adduct with a highly non-linear dose response curve in relation to the exposure concentrations. When comparing the amount of this adducts caused by endogenous or exogenous formaldehyde, the exogenous adduct only amounted to about 1% of the endogenous adduct after a single exposure to 0.7 ppm. Exogenous adducts in the nose of non-human primates are lower than those in rats. In rats the adduct has a half-life of 7.1 days and steady state concentrations will be attained after 28 days of exposure. Therefore after prolonged exposures, exogenous adducts at 0.7 ppm will be by a factor of ~5.5 higher as was extrapolated from inhalation to 2 ppm over 28 days. DPCs are mainly formed by binding between DNA and lysine in proteins and are most probably the precursors of the N2-hydroxymethyl-dG adduct. The ratios of endogenous/exogenous DNA-lysine adducts were clearly higher in the cytoplasma compartment than in the nucleus compartment, while the concentrations of endogenous adducts were similar in both compartments.

Some studies were carried out on modifications of miRNA that are to be considered as epigenetic effects modifying post-transcriptional regulation of gene expression. These studies, in vitro and in vivo, indicated that formaldehyde may alter signalling pathways associated with cancer, inflammation, apoptosis, and the immune system.

Local genotoxicity in humans: local genotoxicity in formaldehyde exposed humans have been described in the form of increased micronuclei rates in nasal or buccal cells. These findings are difficult to interpret because similar effects were not found in volunteers under controlled exposure conditions, the methods employed lack sufficient standardisation, and such effects were not observed in highly exposed experimental animals.

Under consideration of the RAC (2012) proposal for classification and labelling, formaldehyde is classified as Mutagen Cat. 3; R68, according to Directive 67/548/EEC1999/45/ EC, and according to EU Classification, Labelling and Packaging of Substances and Mixtures (CLP) Regulation (EC) No. 1272/2008, Annex VI ,the classification isMutagen Cat 2 (H341). The proposal of RAC (2012) for mutagenicity classification was not based on germ cell or systemic mutagenicity, but on the local genotoxicity in the nose of exposed rats under consideration of the classification guideline of ECHA.