Parental Exposure to Medications and Hydrocarbons and ras Mutations in Children with Acute Lymphoblastic Leukemia: A Report from the Children's Oncology Group

The ras proto-oncogene consists of three genes: H-ras, K-ras, and N-ras. The small protein products (p21) of ras genes are highly homologous and are in the inner plasma membrane surface; they transmit proliferative signals from cytokines and growth factors to the nucleus (1). Five domains of the ras p21 protein are highly conserved throughout the G-protein superfamily and are essential for GTP binding and hydrolysis. Ras proto-oncogene mutations have been implicated in the pathogenesis of many malignancies, with frequency of mutation varying from 95% in pancreatic cancer to 5% in breast cancer (2). The frequency of ras mutations is reported to be between 5% and 20% of patients with acute lymphoblastic leukemia (ALL; refs. 3-11). These mutations result in a constitutively active ras protein, which disrupts the normally tightly regulated ras-dependent signal transduction pathways (2). The most common ras mutations are found in codons 12, 13, and 61 and result in a continuously activated ras protein that can autonomously stimulate cell growth or differentiation (2).

A variety of physical and chemical carcinogens have been shown to induce ras mutations in both human and animal tumor models (e.g., K-ras mutations are associated with methylnitrosourea, cigarette smoke, and organochlorines; refs. 12, 13). We conducted a case-case study of 837 childhood (age <15 years) ALL cases who participated in a large case-control study of ALL to examine whether parental occupational exposures and use of specific medications, factors that we previously found to be associated with an increased risk of childhood leukemia, are related to ras gene mutation (14, 15).

Cases were identified from Children's Cancer Group institutions and have been described in detail elsewhere (14-16). Briefly, the Children's Cancer Group was one of two cooperative clinical trial groups in the 1990s that cared for about 93% of pediatric cancer patients in the United States. Nearly 50% of childhood leukemias in the United States were treated by a Children's Cancer Group hospital or institution (17). During the study recruitment period, there were 120 Children's Cancer Group institutions; 108 participated in the current study. Eligible cases were children ages ≤14 years who had new ALL diagnoses between January 1, 1989 and June 15, 1993. Additional eligibility criteria included a telephone in the patient's residence and availability of the biological mother for a telephone interview. A total of 2,081 eligible cases were identified, and 1,914 mothers were interviewed (92%). Among the 167 nonrespondents were 41 (2.0%) physician refusals, 70 (3.4%) parental refusals, and 18 (0.9%) lost to follow-up after first contact; 38 (1.8%) nonrespondents fell into other miscellaneous categories.

Exposure information was collected by independent telephone interviews using structured questionnaires with mothers and, whenever available, fathers of cases. The mother's questionnaire included questions about demographics, maternal disease history, medication use, occupation, personal habits, household exposure before and during the index pregnancy and birth, reproductive and family medical history as well as the child's disease history, medication use, and exposure to environmental hazards (e.g., pesticides and insecticides). The father's questionnaire focused on medication use, personal habits, household exposures, occupational history, and family medical history. The father's questionnaire was completed for 1,801 of the 2,081 eligible cases (86.5%): 83.4% directly by fathers and 26.6% by mothers as the surrogate. Only information obtained directly from the father was included in the paternal exposure analysis.

The current ras study was initiated in 1996, 3 years after the survey interview for the parent study was completed. Diagnostic bone marrow slides or cryopreserved marrow samples for 837 ALL cases who had participated in the case-control study were available from the original diagnostic institutions or from the Children's Cancer Group reference laboratory. Of the 1,914 interviewed cases, we obtained specimens from 837 subjects.

DNA samples from the 837 ALL cases were extracted from archival bone marrow slides or cryopreserved marrow samples as described previously (18). DNA cells were scraped from the slide with a scalpel using a sterile technique to prevent contamination. Cells were suspended in PCR buffer [50 mmol/L KCl, 10 mmol/L Tris-HCl (pH 9.0), and 1% Triton X-100], boiled for 10 minutes, extracted twice with phenol, and precipitated with ethanol. DNA was washed once with 70% ethanol, resuspended in Tris-EDTA buffer, and amplified by PCR.

We examined mutations in K-ras and N-ras genes at codons 12, 13, and 61 by PCR and allele-specific oligonucleotide hybridization and confirmed them by DNA sequencing. PCR of N-ras and K-ras exon gene fragments was done in a thermocycler (Perkin-Elmer Corp., Norwalk, CT) with ∼200 ng DNA employing standard conditions in 100 μL total reaction volume. The PCR reaction buffer contained 50 mmol/L KCl, 10 mmol/L Tris-HCl (pH 8.3), 1.5 mmol/L MgCl₂, 0.001% (w/v) gelatin (Sigma Chemical Co., St. Louis, MO), 200 μmol/L deoxynucleotide triphosphates (Boehringer Mannheim, Mannheim, Germany), and 2.5 units AmpliTaq (Perkin-Elmer). We amplified sequences surrounding N-ras and K-ras codons 12, 13, and 61 using the following protocol: cycle 1, 94°C for 5 minutes, 55°C for 1 minute, and 72°C for 1 minute; cycles 2 to 35, 94°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute. The oligonucleotide primers were used at a concentration of 0.5 μmol/L. Each PCR amplification of the ras genes exon gene fragments from patient samples included positive and negative controls, and the efficiency and specificity of each amplification were assessed by analyzing the products after electrophoresis on a 3% NuSieve (FMC, Rockland, ME)-1% agarose gel and ethidium bromide staining. Studies in our laboratory had demonstrated that direct DNA sequencing of amplified ras exon fragments was sensitive enough to detect mutation in samples with >50% blasts (data not shown). Samples with <50% blasts were screened by single-strand conformational polymorphism at 4°C and 22°C. Bands with abnormal migration on single-strand conformational polymorphism were excised and subjected to PCR amplification with nested primers and analyzed with subsequent direct DNA sequencing. The DNA of ras mutation-positive samples had an independent second PCR amplification and repeat DNA sequencing to confirm the ras gene mutation sequence.

Case-case comparisons were applied in the statistical analyses. Parental occupational exposure and medication use among cases with any ras mutation, cases with any K-ras mutation, or cases with any N-ras mutation were compared with cases without any ras mutation. Odds ratios (ORs), as approximations of relative risk, were used to measure the association between exposures and ras mutations. Unconditional logistic regression was employed in the data analyses to obtain ORs and 95% confidence intervals (CIs) with adjustment for family income, parental race, education, and age and the child's age and sex.

A total of 127 (15.2%) ALL cases were found to have a K-ras (4.7%) or N-ras (10.6%) mutation in their diagnostic bone marrow specimens; two cases had both mutations. The most frequent ras mutations were found in codon 12 or 13, exon 1 of K-ras (95%) and N-ras genes (90%), most of which involved a G:C to A:T transition (78%). Cases with ras mutations, especially those with a K-ras mutation, were more likely to be younger at diagnosis than ras mutation-negative cases. Cases with K-ras mutations (30%) were more likely to be <2 years old than were ras-negative cases (10%). There were more early B-cell ALL and fewer T-cell ALL among cases with any ras mutation. N-ras mutation cases appeared less likely to have pre–B-cell ALL. ALL cases with a ras mutation did not differ significantly from those without ras mutation by sex, family income, and parental race and age. Subsequent analyses, however, adjusted for age and the above-mentioned demographic variables because they were confounders in our earlier publications on parental occupational exposure and medication use with childhood ALL (Table 1; refs. 14, 15).

Table 2 presents the association between maternal occupational exposure and ras mutation. K-ras mutations were associated with maternal exposure to solvents, degreasers, or cleaning agents (OR 3.1, 95% CI 1.0-9.7) and plastic materials (OR 6.9, 95% CI 1.2-39.7) during pregnancy and plastic materials after pregnancy (OR 8.3, 95% CI 1.4-48.8). Maternal ever exposure to oil or coal products before case diagnosis or during the postnatal period was associated with a significantly elevated risk of N-ras mutation (OR 2.3), although risk estimates of N-ras mutation were also not significantly elevated among ALL cases whose mothers reported exposure to these chemicals before and during pregnancy with the index child.

Associations between paternal exposure to hydrocarbons and ras mutation are summarized in Table 3. Paternal exposure to plastic materials before pregnancy (OR 2.4, 95% CI 1.1-5.1) was related to N-ras mutations. Paternal exposure to miscellaneous hydrocarbons (including epoxy resins, formaldehyde, glues, exhaust, fuels, cooking oils, thermal decomposition products, glycols, ethylene oxide, polycyclic aromatic hydrocarbons, and alcohol) during the child's postnatal period was related to a significantly increased risk of N-ras mutations (OR 1.8, 95% CI 1.0-3.1).

Table 4 presents the association of selected parental medication use and ras mutation among ALL cases. Reported maternal use of mind-altering drugs including marijuana, lysergic acid diethylamide, and cocaine was associated with N-ras mutations (OR 3.2, 95% CI 1.7-6.1) and K-ras mutations (OR 2.4, 95% CI 0.9-6.7). Paternal use of amphetamines or diet pills was associated with N-ras mutations (OR 4.1, 95% CI 1.1-15.0) and not significantly with K-ras mutations (OR 2.4, 95% CI 0.3-22.7). Paternal use of mind-altering drugs was associated with an elevated risk of N-ras mutations (OR 2.0, 95% CI 1.1-3.7) and not significantly related to K-ras mutations (OR 1.6, 95% CI 0.6-4.7).

In this study, we found that 15.2% of childhood ALL cases had a ras mutation in their leukemia cells; N-ras (10.6%) was twice as likely to be affected. Ras mutations occurred more often among ALL cases who were <2 years old at diagnosis and among those with B-lineage leukemia. N-ras mutations were associated with maternal ever exposure to oil and coal products before case diagnosis or during pregnancy, paternal exposure to plastic materials before index pregnancy or other hydrocarbons during the postnatal period, and parental use of mind-altering drugs. K-ras mutations were associated with maternal exposure to solvents and plastic materials during pregnancy and plastic materials after pregnancy and paternal use of amphetamines or diet pills.

It has been recognized that human carcinogenesis results from mutations in a series of critical genes that control cell proliferation, differentiation, and programmed cell death. Ras mutations are one of the most common genetic alterations found in human malignancies (2, 19). Ras genes mutation result in constitutively active ras protein, which disrupts the ras-dependent signal transduction pathways, stimulating tumor growth and/or modulating the immune response against the tumor cells (2). Although the specific mechanisms of ras mutations remain unknown, epidemiologic and animal studies have suggested that specific point mutations in ras genes can be induced by certain chemical carcinogens. In a murine model for lymphohematopoietic neoplasia, exposure to N-methyl-N-nitrosurea has been shown to result in a predominance of K-ras activating mutations at codon 12 with a consistent G:C to A:T transition (20). Studies of human adenocarcinoma of the lung have observed K-ras mutations in 30% of patients with a history of smoking, while <5% of nonsmokers carry the mutation (21). In a study of 62 adult acute myeloid leukemia cases and 630 healthy control subjects, Taylor et al. (22) found that ras-positive cases were about twice as likely to have been employed in occupations associated with an increased leukemia risk compared with ras-negative acute myeloid leukemia cases or healthy controls. Another study found aspartic acid mutations in codon 13 of K-ras gene in vinyl chloride–induced liver cancer and an elevation of serum ras protein p21 among liver cancer patients and workers exposed to vinyl chloride but not among healthy nonexposed controls (23). Serum concentrations of organochlorine compounds have been found to be related to glycine to valine substitution at codon 12 of K-ras gene in pancreatic cancer patients (12). In the current study, we found that certain parental exposures were related to N-ras or K-ras mutations at exon 1, mainly involving G:C to A:T transition.

Reports from previous studies have shown ras mutations, mainly N-ras, in 5% to 20% of ALL patients (3-11). However, these earlier studies have several limitations: most included only patients from a single institution, had small sample sizes (14 to 150 subjects) to characterize ras-positive patients accurately, and lacked environmental exposure information. The current study contains the largest series of childhood ALL cases who have been evaluated for the ras mutation and confirmed by DNA sequencing. Therefore, our data represent the most accurate assessment of ras mutation among ALL patients thus far. The extensive exposure information collected in the current study allowed us to evaluate whether ras mutation is related to particular environmental exposures.

Parental occupational exposure to hydrocarbons, such as chlorinated solvents, benzene, or paints, has been linked previously to elevated childhood leukemia risk (24, 25). In earlier reports from the parent study (14, 15), we found that maternal exposure to solvents, paints, or thinners (preconception and during pregnancy) and plastic materials (postnatal) were related to an increased risk of childhood ALL. Paternal exposure to plastic materials before conception was also associated with ALL. Parental use of amphetamines, diet pills, or mind-altering drugs before and during the index pregnancy was also found to be related to an increased childhood ALL risk, particularly among children <1 year old at diagnosis. The current study found some of these previously identified high-risk chemical exposures to be associated with ras mutations, thus providing additional evidence that parental exposure to these agents affected ALL risk in offspring. We found that ALL cases with ras mutations were more likely to be younger at diagnosis, and the association with parental use of amphetamines, diet pills, or mind-altering drugs was confined to young children. This may suggest that ras mutations occurred during pregnancy or preconceptionally in germ cells.

This analysis employed a case-case comparison study design. The advantage of this study approach over a case-control design is that it reduces recall bias, where case parents may recall higher exposure levels than control parents. In this study, it is unlikely that a case parent of a child with a ras mutation would recall exposures differently than a case parent of a child without ras mutations. However, ras mutations were only evaluated in 837 of 1,914 (44%) ALL cases included in the parent study, which substantially reduced the statistical power. This case-case study is limited by its inability to identify risk factors shared by ras mutation positive and negative cases (i.e., risk factors that can induce both ras mutation and other gene mutations could have been missed with this type of study design). Therefore, the association between ras mutations and parental chemical exposure could be underestimated in this study.

In summary, we found that parental occupational exposure to hydrocarbons and mind-altering drugs, chemicals that have been previously suggested to increase the risk of childhood leukemia, were related to specific ras mutations in childhood ALL.