Frequency of trisomy 20 in nonmalignant bronchial epithelium from lung cancer patients and cancer-free former uranium miners and smokers.

Lung cancer is the leading cause of cancer-related deaths. The development of sensitive screening methods to identify at-risk individuals before emergence of clinical disease would permit early intervention that could decrease this mortality. Our previous studies have shown that cells with trisomy 7 can be detected in bronchial epithelium from cancer-free smokers and former uranium miners. However, the use of more than one molecular marker could increase the chance of identifying at-risk individuals. Trisomy 20, which is found in 43-57% of non-small cell lung cancers, is a candidate marker. The purpose of the current investigation was to determine the percentage of cells with trisomy 20 in persons with a high risk for lung cancer. Bronchial epithelial cells that had been assayed for trisomy 7 were assayed for trisomy 20 by fluorescence in situ hybridization. Trisomy 20 was detected in bronchial epithelial cells from lung cancer patients and from smokers and ex-uranium miners without lung cancer. In some cases, patients who were negative for trisomy 7 exhibited trisomy 20. Consequently, more people with field cancerization were identified using both markers. However, the two markers combined did not appear to stratify the risk for lung cancer.

Individuals at high risk for lung cancer have airway epithelial cells with chromosome aberrations frequently found in lung tumor cells.

Identification of individuals at greatest risk of developing lung cancer could enhance the efficacy of intervention modalities, thereby greatly reducing mortality from this disease. One strategy for identifying these people is to establish molecular markers which reflect the severity of their cancerization field. Thus, investigations were initiated to determine of cells with chromosome aberrations frequently exhibited by lung tumor cells, i.e., trisomy 7, trisomy 20, and deletion of 9p23, are prevalent within the uninvolved airways of cancer patients. As a result, cells containing these aberrations were consistently found at significantly elevated levels by using fluorescence in situ hybridization (FISH). In contrast, cells collected from non-smokers who had never smoked were normal by this assay. The next objective was to determine of cells exhibiting these alterations are also present in upper airways of exposed, but cancer-free smokers and ex-uranium miners. The results showed that, although only a subset of these people will develop lung cancer in their lifetimes, they universally harbor increased numbers of abnormal cells within their airway epithelium. However, the number of sites with multiple verities of abnormal cells tended to be fewer compared with the cancer patients. Thus, quantifying cells with molecular alterations within the cancerization field of a smoker may delineate those with a lesser grade of damage, and these individuals may be at a lesser risk of developing disease. However, differences in the extent of genetic damage afforded by these assays may not clearly define individuals with pending disease, and additional molecular assays must be devised.

Concurrent fluorescence in situ hybridization and immunocytochemistry for the detection of chromosome aberrations in exfoliated bronchial epithelial cells.

OBJECTIVE: A procedure was developed to allow concurrent detection of chromosome aberrations and identification of bronchial epithelial cells. STUDY DESIGN: Fluorescence in situ hybridization for chromosome 7 and immunocytochemistry for cytokeratin were performed on exfoliated bronchial epithelial cells in a sputum sample from a cancer patient. RESULTS: The Spectrum Orange-labeled alpha satellite probe for chromosome 7 produced red fluorescence, nuclei were counterstained with 4,6-diamidino-2-phenylindole (blue), and cytokeratin was visualized using a fluorescein isothiocyanate (FITC)-conjugated secondary antibody (green). CONCLUSION: This procedure allowed the rapid identification of airway epithelial cells with numerical chromosome aberrations in this sample. Ultimately, this procedure could increase the sensitivity and specificity of sputum cytology as a laboratory diagnostic tool for the early detection of lung cancer.

Molecular identification of individuals at high risk for lung cancer.

The objective of the work reviewed herein was to evaluate whether a cancerization field-consisting of cells with genetic alterations can be detected within normal-appearing bronchial epithelium. By using fluorescence in situ hybridization (FISH) for trisomy 7, cancerization fields were detected in the majority of cancer patients and also in significant percentages of cancer-free tobacco smokers and former uranium miners. These results suggest that molecular analyses may enhance the power of detecting premalignant changes in bronchial epithelium and may ultimately lead to identifying persons at greatest risk for developing lung cancer.

Detection of trisomy 7 in nonmalignant bronchial epithelium from lung cancer patients and individuals at risk for lung cancer.

Early identification and subsequent intervention are needed to decrease the high mortality rate associated with lung cancer. The examination of bronchial epithelium for genetic changes could be a valuable approach to identify individuals at greatest risk. The purpose of this investigation was to assay cells recovered from nonmalignant bronchial epithelium by fluorescence in situ hybridization for trisomy of chromosome 7, an alteration common in non-small cell lung cancer. Bronchial epithelium was collected during bronchoscopy from 16 cigarette smokers undergoing clinical evaluation for possible lung cancer and from seven individuals with a prior history of underground uranium mining. Normal bronchial epithelium was obtained from individuals without a prior history of smoking (never smokers). Bronchial cells were collected from a segmental bronchus in up to four different lung lobes for cytology and tissue culture. Twelve of 16 smokers were diagnosed with lung cancer. Cytological changes found in bronchial epithelium included squamous metaplasia, hyperplasia, and atypical glandular cells. These changes were present in 33, 12, and 47% of sites from lung cancer patients, smokers, and former uranium miners, respectively. Less than 10% of cells recovered from the diagnostic brush had cytological changes, and in several cases, these changes were present within different lobes from the same patient. Background frequencies for trisomy 7 were 1.4 +/- 0.3% in bronchial epithelial cells from never smokers. Eighteen of 42 bronchial sites from lung cancer patients showed significantly elevated frequencies of trisomy 7 compared to never smoker controls. Six of the sites positive for trisomy 7 also contained cytological abnormalities. Trisomy 7 was found in six of seven patients diagnosed with squamous cell carcinoma, one of one patient with adenosquamous cell carcinoma, but in only one of four patients with adenocarcinoma. A significant increase in trisomy 7 frequency was detected in cytologically normal bronchial epithelium collected from four sites in one cancer-free smoker, whereas epithelium from the other smokers did not contain this chromosome abnormality. Finally, trisomy 7 was observed in almost half of the former uranium miners; three of seven sites positive for trisomy 7 also exhibited hyperplasia. Two of the former uranium miners who were positive for trisomy 7 developed squamous cell carcinoma 2 years after collection of bronchial cells. To determine whether the increased frequency of trisomy 7 reflects generalized aneuploidy or specific chromosomal duplication, a subgroup of samples was evaluated for trisomy of chromosome 2; the frequency was not elevated in any of the cases as compared with controls. The studies described in this report are the first to detect and quantify the presence of trisomy 7 in subjects at risk for lung cancer. These results also demonstrate the ability to detect genetic changes in cytologically normal cells, suggesting that molecular analyses may enhance the power for detecting premalignant changes in bronchial epithelium in high-risk individuals.

Induction of genomic instability in normal human bronchial epithelial cells by 238Pu alpha-particles.

Pulmonary deposition of alpha-particle-emitting radon daughters is estimated to account for 10% of all lung cancer deaths in the USA. However, the nature and timing of early (preneoplastic) genetic alterations in radon-associated lung cancer are still relatively uncertain. The purpose of this investigation was to determine whether genomic instability occurs after exposure of cultured normal human bronchial epithelial cells to six equal, fractionated doses of alpha-particles (total doses 2-4 Gy). Two weeks after the final exposure, foci of phenotypically altered cells (PACs) were detected in 0, 63 and 77% of control, low and high dose cultures respectively. Of these, 18% exhibited extended life spans relative to unexposed controls. Elevated frequencies of binucleated cells (BNCs), a marker of genomic instability, were observed in 60 and 38% of the PAC cultures from the low and high dose groups respectively. The micronucleus assay also showed evidence of genomic instability in 40 and 38% of PAC cultures from the low dose and high dose groups respectively. No changes in microsatellite length, another marker of genomic instability, were detected in any of the PAC samples with the 28 markers used for this assay. However, one PAC (L2) showed a hemizygous deletion at 9p13.3. Another PAC (H9), which exhibited the highest frequency of cells containing micronuclei (MN), exhibited a hemizygous deletion at 7q31.3. Each loss may represent a stable mutation that resulted either directly from irradiation or later in progeny of exposed cells because of alpha-particle-induced genomic instability. The fact that elevated levels of BNCs and MN were present in the progeny many generations after irradiation indicates that the genetic alterations detected with these two markers were not a direct consequence of radiation exposure, but of resulting genomic instability, which may be an early change after exposure to alpha-particles.

Genetic toxicity of 2-acetylaminofluorene, 2-aminofluorene and some of their metabolites and model metabolites.

2-Acetylaminofluorene and 2-aminofluorene are among the most intensively studied of all chemical mutagens and carcinogens. Fundamental research findings concerning the metabolism of 2-acetylaminofluorene to electrophilic derivatives, the interaction of these derivatives with DNA, and the carcinogenic and mutagenic responses that are associated with the resulting DNA damage have formed the foundation upon which much of genetic toxicity testing is based. The parent compounds and their proximate and ultimate mutagenic and carcinogenic derivatives have been evaluated in a variety of prokaryotic and eukaryotic assays for mutagenesis and DNA damage. The reactive derivatives are active in virtually all systems, while 2-acetylaminofluorene and 2-aminofluorene are active in most systems that provide adequate metabolic activation. Knowledge of the structures of the DNA adducts formed by 2-acetylaminofluorene and 2-aminofluorene, the effects of the adducts on DNA conformation and synthesis, adduct distribution in tissues, cells and DNA, and adduct repair have been used to develop hypotheses to understand the genotoxic and carcinogenic effects of these compounds. Molecular analysis of mutations produced in cell-free, bacterial, in vitro mammalian, and intact animal systems have recently been used to extend these hypotheses.

The Agency for Toxic Substances and Disease Registry's role in development and application of biomarkers in public health practice.

An overview of the Agency for Toxic Substances and Disease Registry's (ATSDR) biomarker program is presented in the context of the paradigm for biomarkers developed by the National Research Council (NRC, 1987, 1991). The status and projected utility of four biomarker studies conducted by NRC and sponsored by ATSDR, the Environmental Protection Agency (EPA), and the National Institute of Environmental Health Sciences (NIEHS) are discussed. These studies include a review of relevant research on biomarkers for specific toxicologic end points, including reproductive toxicology, pulmonary toxicology, neurotoxicology, and immunotoxicology. Also, the scope of related research on exposure characterization being conducted by the ATSDR-sponsored research program at Rutgers University is reviewed. The potential impact of biomarkers on public health assessments and on the range of ATSDR programs is described. Specifically, the role of biomarkers in dose reconstruction, in ATSDR's health studies program, and in the emerging field of molecular epidemiology is reviewed. In addition, future directions and research needs are addressed.

Dinitropyrene metabolism, DNA adduct formation, and DNA amplification in an SV40-transformed Chinese hamster embryo cell line.

The environmental pollutants 1,6-dinitropyrene (1,6-DNP) and 1,8-dinitropyrene (1,8-DNP) are strongly carcinogenic in a number of animal models. These DNPs are metabolized by nitroreduction to N-hydroxy arylamine derivatives that either directly or after acetylation bind to cellular DNA. In the experiments reported here, we examined whether DNA adduct formation by 1,6-DNP and 1,8-DNP was associated with amplification of specific DNA sequences, a process that may be causally related to tumorigenesis. CO60 cells, an SV40-transformed Chinese hamster embryo cell line, were incubated with 2.5 or 50 ng/mL [4,5,9,10(-3)H]1,6-DNP for 5 h. High-pressure liquid chromatographic analysis of organic extracts of the medium indicated the presence of 1-acetylamino-6-nitropyrene, suggesting that these cells are capable of nitroreduction and acetylation. 32P-Postlabeling analysis of DNA isolated from cells exposed to 1.0 or 2.5 ng/mL 1,6-DNP revealed dose-related formation of N-(deoxyguanosin-8-yl)-1-amino-6-nitropyrene. A similar adduct, presumably N-(deoxyguanosin-8-yl)-1-amino-8-nitropyrene, was detected after incubations with 1,8-DNP. DNA isolated from analogous experiments was slot-blotted onto nylon membranes and hybridized with 32P-labeled SV40, c-fos, or beta-actin DNA probes. beta-Actin was not amplified and c-fos was amplified only a small amount; however, there was dose-related amplification of SV40 sequences, whose levels were in some instances approximately 20 times that observed in solvent-treated controls. These data indicate that DNA adduct formation by 1,6-DNP and 1,8-DNP is associated with the amplification of certain DNA sequences, a response that may be related to the tumorigenic potential of these compounds.

Nitrobenzo[a]pyrene-induced DNA amplification in SV40-transformed Chinese hamster embryo cells.

Nitrobenzo[a]pyrenes (NBaPs) are ubiquitous environmental pollutants that produce mutations in Salmonella typhimurium and Chinese hamster ovary cells. In this study, 1-, 3-, and 6-NBaP induced amplification of SV40 DNA sequences in an SV40-transformed Chinese hamster embryo cell line which is sensitive to DNA amplification by various known carcinogens. Of the three isomers, 3-NBaP produced the highest level of gene amplification, which was 4.8 relative to untreated controls at a dose of 5 micrograms/ml. Considering the relationship between gene amplification and tumorigenesis, it seems prudent to carry out a more exhaustive analysis of the carcinogenic potential of these agents.

Dot-blot hybridization: quantitative analysis with direct beta counting.

The suitability of using direct beta counting (DBC) for quantitating radioactivity of the probe:target complex in dot-blot hybridization was evaluated using a Packard Matrix 96. A comparison of blots analyzed using autoradiography followed by densitometry scanning (film/densitometry) with those analyzed using direct beta counting revealed similar data trends with the two methods. However, direct beta counting quantitated the amount of radioactivity in the dot blots directly (without film exposure or additional sample preparation), which significantly reduced the time required to obtain results. Blots analyzed first with direct beta counting and then liquid scintillation counting exhibited similar data trends with both methods. Despite a decreased counting efficiency, analysis with direct beta counting has the following advantages compared with liquid scintillation counting: 1) no additional sample preparation is required (no vials or cocktail are used), 2) no sample destruction occurs due to analysis and 3) quantitative results are obtained more rapidly (since the radioactivity for all 96 samples in a dot blot is simultaneously determined in real time). Analysis with direct beta counting was also shown not to interfere with the successful reprobing of stripped dot blots with either unique sequence or total genomic probes. Overall, direct beta counting provides quick, quantitative results for dot blots while saving considerable time and effort.

The induction of aneuploidy by 3-nitrobenzo[a]pyrene in Chinese hamster ovary cells.

Nitrobenzo[a]pyrenes are found in urban air particulates and particulates from diesel exhaust, gasoline engines and wood burning stoves. Following exposure of Chinese hamster ovary cells (CHO-K1-BH4) to 3-nitrobenzo[a]pyrene (3-NB[a]P), cells with multiple nuclei and/or nuclei with multiple lobes were observed. When CHO cells were treated with 5 micrograms/ml 3-NB[a]P for 5 or 20 h, aneuploidy was noted in these cells at 24-96 h post-exposure. The addition of N6, O2'-dibutyryl adenosine 3':5'-cyclic monophosphate to 3-NB[a]P-exposed CHO cell cultures appeared to reduce the amount of aneuploidy in treated cultures. Structure--activity studies showed that 1-NB[a]P was a much less effective inducer of aneuploidy than 3-NB[a]P and 6-NB[a]P was ineffective. 1-, 3- and 6-nitrosobenzo[a]pyrenes were not effective inducers of aneuploidy in CHO cells, and aneuploidy was not observed in cultures treated with 3-NB[a]P in the presence of S9 activation. It appears that the parent 3-NB[a]P is responsible for producing aneuploidy in CHO cells.

Evaluation of the genotoxicity of gentian violet in bacterial and mammalian cell systems.

Previous studies indicate that gentian violet (GV), a triphenylmethane dye used in agriculture and human medicine, is a clastogen in vitro and a carcinogen in chronically exposed mice and rats. Data on its genotoxic activity, however, have been incomplete and partly contradictory. Mutagenesis and DNA damage experiments were conducted to re-evaluate the genotoxic potential of GV in both bacterial and mammalian cell systems. GV was mutagenic in Salmonella typhimurium tester strains TA97 and TA104, but there was little mutagenic activity detected in strains TA98 and TA100. A rat liver homogenate fraction (S9) tended to increase mutagenicity. The major microsomal metabolites of GV, pentamethylpararosaniline and N,N,N',N'-tetramethylpararosaniline were less mutagenic in TA97 and TA104, while N,N,N',N"-tetramethylpararosaniline was a weak mutagen in Salmonella. GV was not mutagenic in Chinese hamster ovary (CHO) cell strain CHO-K1-BH4, and was a questionable mutagen in CHO-AS52 cells. While GV produced DNA damage as measured by sedimentation of nucleotids derived from B6C3F1 mouse lymphocytes treated in vitro, no damage was found in lymphocytes isolated from mice dosed with GV. GV was also a weak producer of gene amplification in an SV40-transformed Chinese hamster cell line. The results indicate that GV is a point mutagen in bacteria; however, since similar exposure conditions produced weak mutagenic activity in mammalian cells, GV may be carcinogenic by virtue of its clastogenic activity.

Temporal SCE and cytotoxicity responses of murine cells following in vivo treatment with MNU or L-PAM.

Time-dependent SCE responses of bone marrow and cultured spleen lymphocytes of BDF1 mice to in vivo treatment with MNU or L-PAM were studied. L-PAM was generally more active than MNU in producing elevated SCEs. Increases of 4-5 times control levels were produced in lymphocytes cultured at 1 or 24 h after i.p. injection of 4.95 mumoles/kg of L-PAM whereas an approximately 3-fold increase was produced by an acute injection of 190 mumoles/kg of MNU. Temporal SCE responses of bone marrow cells were carried out with doses of L-PAM (4.95 mumoles/kg) and MNU (131 mumoles/kg) found to be noncytotoxic by analysis of relative percentages of first, second, and third generation cells. The SCE response of second generation bone marrow cells (greater than 7 time baseline) to MNU was maximum when treatment was at the first cycle and decreased rapidly with increasing time prior to, or after the start of BrdUrd infusion. By contrast nonreciprocal (NR) SCE responses of third generation progeny never exceeded a 2-fold increase over baseline. Dramatic inhibition of cell cycling by MNU was evident as the reciprocal (R) SCEs in third generation cells increased, and exceeded NR SCEs, with increased treatment time intervals after the start of BrdUrd infusion. Similar dramatic cytotoxicity of L-PAM was apparent in time-dependent SCE response studies. An increased BrdUrd infusion time (28 h rather than the usual 26.5 h) was necessary to achieve adequate numbers of third division cells. Maximum SCE responses of the latter cells to L-PAM did not exceed 3 times baseline levels, whereas maximum responses of greater than 9 times control levels were produced in second generation cells. Comparison of SCE responses of second and third generation progeny of similarly treated cell populations, appears to provide a more sensitive assessment of cytotoxicity than does the conventional method of cell cycle analysis.

Comparison of sister chromatid exchanges in spleen and thymus lymphocytes from AKR, C57BL/6N x DBA/2J F1, and CBA mice following in vivo exposure to N-methyl-N-nitrosourea.

The induction of sister chromatid exchange (SCE) by N-methyl-N-nitrosourea (MNU) was studied in spleen and thymus lymphocytes from AKR mice, which are highly susceptible to MNU-produced thymomas, CBA mice, which are much less sensitive to induction of thymomas by MNU, and C57BL/6N x DBA/2J F1 mice. MNU produced dose-related increases in SCE in concanavalin A (Con A)- and lipopolysaccharide-stimulated spleen lymphocytes and Con A-stimulated thymus lymphocytes from each mouse strain at 1 and 24 h posttreatment. MNU-induced SCE were generally higher in Con A-stimulated spleen lymphocytes compared to lipopolysaccharide-stimulated spleen lymphocytes and Con A-stimulated thymus lymphocytes from each mouse strain. On the whole, MNU-produced SCE were lower in AKR and CBA spleens than in C57BL/6N x DBA/2J F1 spleens. In addition, MNU-induced SCE levels in thymus lymphocytes from all three strains of mice were, for the most part, similar. These results indicate that if differences in MNU-induced genotoxicity in AKR, CBA, and C57BL/6N x DBA/2J F1 thymus lymphocytes exist, these differences cannot be ascertained by use of the in vivo/in vitro SCE assay.

Triethylene melamine-induced sister-chromatid exchange in murine lymphocytes exposed in vivo.

The induction of sister-chromatid exchange (SCE) by triethylene melamine (TEM), a known animal carcinogen, was investigated in an in vivo exposure/in vitro culture murine lymphocyte assay. Dose-related increases in SCE were observed in B6D2F1 mice following a single i.p. injection of 0.5, 1 or 2 mg/kg TEM. SCE frequencies remained elevated over baseline levels at 24 h post exposure. It is hoped that studies of this nature can determine whether the in vivo/in vitro murine lymphocyte SCE assay is useful for predicting the carcinogenic potential of an agent.

Induction of sister chromatid exchange in multiple murine tissues in vivo by various methylating agents.

In order to study the reliability of in vivo sister chromatid exchange (SCE) assays for predicting carcinogenicity, several known animal carcinogens were tested in a multicellular in vivo SCE assay and an in vivo/in vitro murine lymphocyte assay. The methylating agents 1,2-dimethylhydrazine.2 HCl (DMH), dimethylnitrosamine (DMN), methylnitrosourea (MNU), methyl methane-sulphonate (MMS), and methylazoxymethanol acetate (MAM) were tested for SCE induction in several murine tissues in vivo, including bone marrow, alveolar macrophages, regenerating and intact liver, and kidney from B6D2F1 mice. In all cell types, clear dose-responses were observed following exposure of mice to subcytotoxic fractions of the LD50 dose of DMH, MNU, or MMS. DMN (0.03-0.27 mmol/kg) produced small, although not dose-related, increases in SCE in all cell types. At the doses tested (0.06 and 0.08 mmol/kg), MAM did not induce elevated SCE in the various cell types. Following a series of multiple i.p. injections of low, non-toxic doses of DMH (0.15 mmol/kg, once a week, for 10 weeks), significant differences were observed in intact vs. regenerating liver and in single vs. multiple injections in regenerating liver. Following exposure of B6D2F1 mice to a single i.p. injection of 0.25 mmol/kg DMN, DMH, or MMS or 0.19 mmol/kg MNU, SCE responses were evaluated in Concanavalin A (Con A)- and LPS-stimulated blood and spleen lymphocytes. Considerable cytotoxicity was observed in blood lymphocytes. In Con A- and LPS-stimulated spleen lymphocytes, DMH-, and DMN- and MMS-induced SCE frequencies were approximately 1.5-2 x baseline levels and MNU-induced SCE were approximately three- to fourfold higher than baseline values in cultures initiated at 1 and 24 h postexposure. At 48 and 72 h after an i.p. injection of 0.131 mmol/kg MNU, SCE responses in lymphocytes were approximately 2 x baseline levels. At 24 h following one, two, or four injections (one/week) of 0.075 mmol/kg MNU dose-related increases in SCE were observed in spleen lymphocytes. These studies illustrate that carefully designed in vivo SCE assays may have the capacity to predict the tumorigenic potential of chemical agents.

Long-term persistence of ethyl carbamate-induced sister chromatid exchanges in murine lymphocytes.

Ethyl carbamate-induced sister chromatid exchange (SCE) was evaluated at 20 min and 1, 3, 4.5, 5.5, 7, and 9 h postexposure (acute dose, ethyl carbamate, 3.3 mmol/kg) in concanavalin A (Con A)- and lipopolysaccharide (LPS)-stimulated murine peripheral blood lymphocytes (PBL). In both Con A- and LPS-stimulated PBL, SCE responses peaked between 4.5 and 5.5 h postinjection, a time which corresponds to complete biotransformation of ethyl carbamate. Peak induced SCEs for Con A- and LPS-stimulated PBL were 6.43 and 7.44, respectively. SCE responses were also evaluated in Con A- and LPS-stimulated PBL at 3 and 24 h following the last of a series of two, four, or six i.p. injections of ethyl carbamate (3.3 mmol/kg) given every other day. Dose-related increases (presumably reflecting the accumulation of unrepaired SCE-inducing damage) in SCEs were observed at both times following two and four injections of ethyl carbamate. However, following six injections a decrease in SCE response and increased cytotoxicity were observed. Persistence of SCE-inducing DNA lesions was observed in blood, spleen, and parathymic node lymphocytes following the last of a series of 12 i.p. injections (three times weekly) of ethyl carbamate (2.2 mmol/kg). With the exception of LPS-stimulated blood lymphocytes, exposed blood and spleen Con A- and LPS-stimulated lymphocyte populations contained a significantly higher number of high-frequency cells than did their respective controls at 16 weeks postexposure. The gradual return of SCE levels to base-line values appears to be primarily a consequence of slow population turnover. Parathymic node lymphocytes exhibited elevated SCE responses (2 times base-line levels) for up to 4 weeks postexposure.