Grantee Research Project Results

Childhood leukemia is a multifactorial disease; consequently, our group established a complex research “nexus” that includes complementary approaches and disciplines (i.e., tumor biology, molecular studies, epidemiologic population-based studies, international collaborations, and exposure assessment). We seek to understand the timing of the initiation and progression of childhood leukemia and the causal biological mechanisms; to identify biomarkers of relevant chemical exposures and their early effects; and to discover environmental, genetic, and epigenetic risk factors, as well as identify differences in childhood leukemia risk by ethnicity. CIRCLE’'s overarching theme during the first cycle was to examine the effects of in utero and early life exposure to potentially carcinogenic chemicals (i.e., pesticides, tobacco-related contaminants, polychlorinated biphenyls (PCBs), and polybrominated diphenyl ethers (PBDEs)) present in homes and the interplay of these chemicals with genetic and epigenetic factors that can lead to the development of childhood acute lymphoblastic leukemia (ALL), as well as acute myeloid leukemia (AML), a rare subtype.

 

Our approach was to first develop and then increasingly refine exposure estimates, beginning with a “conventional” epidemiologic study using self-reports and then incorporating state-of-the-art exposure assessments, using home sampling and biomarkers. We then investigated epigenetic mechanisms underlying the risks associated with these exposures using data from the California Childhood Leukemia Study (CCLS) and other studies conducted by members of the Childhood Leukemia International Consortium (CLIC). Below we summarize our key achievements and findings from the scientific projects and outreach/translation activities, which received strong administrative support from the CIRCLE leadership during the first cycle (2009-2015).

 

To address the barrier posed by childhood leukemia being a rare disease, Project 1 supported the expansion of the CLIC consortium to 33 case-control studies conducted in North and South America, Europe, Africa, and Oceania. We used information on up to 9,000 childhood leukemia cases and 15,000 controls to evaluate the effect of tobacco and pesticide exposures on childhood leukemia risk overall, by subtype, alone, and in combination with genetic factors. Complementary CLIC pooled analyses on other environmental factors, birth characteristics, lifestyle, and genetic factors indirectly benefited from our Center’s work.

 

In addition, we overcame the formidable challenges of sharing data among CLIC members from different countries and cultures by creating an atmosphere of trust and obtaining agreements on procedures, including rules for pooling data, reviewing analyses and preparing manuscripts. We also strengthened our partnership with the International Agency for Research on Cancer (IARC) who agreed to host the CLIC data center. 

 

Most studies describing environmental risk factors for childhood leukemia rely on interviews and questionnaires to obtain exposure data; an approach which is limited by a lack of specificity and a susceptibility to recall bias. We characterized levels of persistent chemicals in dust samples collected from participating households as an alternative, and potentially more objective, measure of children’s environmental exposures to targeted chemicals. Independently funded epidemiological analyses of these dust measurements have revealed novel environmental risk factors for childhood leukemia, including PCBs, PBDEs, and PAHs. We evaluated the validity of this approach of using residential-dust samples to characterize children’s environmental exposures to indoor contaminants for epidemiological studies. Specifically, we estimated correlations between chemical levels in children’s blood samples and in dust samples from their homes. We estimated the temporal variability of dust-chemical levels and identified determinants of dust-chemical levels using mixed-effects models of dust measurements from CIRCLE homes.

 

We also developed methods for omics analyses of chemicals extracted from archived neonatal blood spots (ANBS), which had been obtained by CIRCLE from the State of California. Because these chemicals represented in utero exposures of childhood leukemia cases and controls, they offered a unique opportunity to investigate effects of early life exposures on childhood leukemia. Specifically, we developed an untargeted method for measuring adducts of human serum albumin (HSA) that had been modified at Cys34, which is the most nucleophilic hot spot in HSA and serum.

 

We examined epigenetic mechanisms in the development of childhood leukemia, as well as links of these epigenetic changes to putative risk factors for childhood leukemia. We also explored the timing of methylation profiles that may be independently associated with the risk of childhood leukemia. To account for confounding introduced by differences in cell distribution, we significantly improved our analytical procedures to quantify associations between DNA methylation at CpG sites and the environmental exposures studied in Projects 1 and 2 (i.e., self-reported smoking and exposure to pesticides, folate intake during pregnancy, and house-dust levels of PCBs).

 

We established the normal developmental pattern of DNA methylation in pre-cancerous pre-B cells, and identified several precise DNA methylation changes that contribute to the development of childhood ALL. We have also discovered that folic acid levels around the time of conception affect the birth methylation levels of neural crest development (with possible impact on birth defects) and cancer development genes. In further work, we have constructed a DNA methylation biomarker for maternal smoking, and discovered a significant link between this biomarker and leukemia risk, providing a link between maternal smoking and leukemia that was previously not discovered in epidemiology studies on this topic. 

 

CIRCLE investigators worked together to enhance our understanding of the continuum from exposure to disease initiation. For example, Project 2 demonstrated that PAH levels measured in dust samples collected after birth provide an acceptable surrogate for estimating in utero PAH exposure, which were previously associated with an increased risk of ALL. Following these findings, Project 3 conducted genome-wide DNA methylation analyses in archived neonatal blood spots available in a subset of CCLS households with PAH dust measurement. Consequently, we found that exposure to PAHs was associated with changes in DNA methylation in genes that may be implicated in ALL risk.

 

CLIC pooled studies in Project 1 confirmed an increased risk of childhood leukemia with paternal (but not maternal) tobacco smoking. The “paradoxical” lack of association with maternal smoking during pregnancy may be due to recall bias, and in utero biomarkers of exposure in Projects 2 and 3 provided useful insight. Adductomics analyses in neonatal blood spots of infants without leukemia revealed that levels of two Cys34 adducts differed between infants of smoking and nonsmoking mothers. Specific methylation changes associated with self-reported parental tobacco smoking were identified in two independent sets of children from the CCLS. Other CLIC pooled studies confirmed that prenatal maternal folate supplementation reduced risk of childhood ALL. Following these findings, we identified folate-induced methylation changes in two independent sets of children from the CCLS.

 

The overarching goals of Core B were to develop narratives that communicated the key questions, findings, and trajectory of future research aimed at finding the causes of leukemia in children. The narrative was used as the basis for the research translation and community outreach program; and to engage various audiences including: non-governmental organizations that serve parents, health/child care professionals, community leaders, policy leaders, and other advocates interested in children’s environmental health; targeted earned media; state and national policy and stakeholder audiences such as the California Department of Public Health and the California Environmental Protection Agency.

 

Core B investigators and the Pediatric Health Specialists successfully increased the awareness of childhood leukemia and its risk factors. The “visibility” of the research conducted in CIRCLE was heightened by an invitation to join a Congressional Briefing on October 28, 2015 where Dr. Metayer, Center PI, presented key findings on pre- and post-natal chemical exposures and childhood leukemia. CIRCLE also initiated collaborations with other Children’s Centers to better assess the cumulative impacts of chemical exposures on several childhood diseases such as cancer, asthma, neurocognitive disorders.

 

We conducted a survey of nearly 200 pediatric oncology clinicians indicating that clinicians feel ill-prepared to answer questions about the role of environmental factors in childhood illness. This observation strengthens the need for increasing environmental health education among health professionals. With support from CIRCLE investigators, Dr. Miller (Pediatric Health Specialist) developed an interactive E-book chapter on leukemia and the environment (http://coeh.berkeley.edu/ucpehsu/) that highlights CIRCLE research. The E-book has been an important resource to the public and health professional and is available for continuing education credits through the Centers for Disease Control and Prevention.

We identified—with enhanced precision and robust statistical methods—associations of childhood leukemia with multiple chemicals (pesticides, paints, and tobacco), immune factors, prenatal folic acid, vitamin supplementation, and fetal growth. We reported that, the risk of leukemia varies by histologic and cytogenetic subtype for selected exposures. In addition, we significantly strengthened CLIC, a key infrastructure and resource for future international research on childhood leukemia, as well as other childhood cancers.

 

The funded research validated the use of home dust samples to characterize children’s environmental exposures to indoor contaminants such as PBDEs, PCBs, PAHs, and nicotine. Our findings have formed the basis for two subsequent epidemiological analyses, which, for the first time, identified PBDEs and PAHs as risk factors for childhood leukemia. The omics methods that we have developed for characterizing children’s in utero exposures from analysis of neonatal blood spots will allow us to investigate unknown causes of childhood leukemia. Thus, our findings will likely lead to future public health interventions designed to reduce children’s exposure to harmful chemicals and thus reduce risks of childhood leukemia. 

 

In our first 5 years of funding, we exhaustively characterized the DNA methylation biology of our disease of interest, and developed an analytical pipeline to enable further discovery of the direct links between pregnancy exposures and subsequent leukemia in the offspring. We utilized environmental measures from epidemiology (Project 1) and modern exposure science (Project 2) to characterize key DNA methylation changes that impact disease risk. The DNA methylation changes associated with both environmental exposures and leukemia status in neonatal blood spots provided a template for the studies proposed in next funding cycle, which will also incorporate genetics and non-chemical factors.

 

Research translation initiated during the first cycle of CIRCLE laid the foundation for future initiatives to improve environmental health literacy among health care professionals, concerned families (including minorities such as Latinos who are disproportionally affected by childhood leukemia), and patient advocacy groups, as well as provided a scientific basis for developing prevention programs for childhood leukemia.

 

 

Leukemia is the most common type of childhood cancer. About 2,400 cases of childhood leukemia (ages 0-14 years) are diagnosed annually in the United States. Leukemia incidence rates have increased over the past 50 years, and part of this increase is likely due to changes in patterns of exposure to chemicals introduced into a child’s environment, alone or in combination with genetic susceptibility. The etiology of childhood leukemia is complex, and research to identify risk factors and the mechanisms by which they affect leukemia risk requires multidisciplinary and multi-institutional efforts. Our center, CIRCLE, was designed to foster the collaborative research network to make this work possible. We collaborated with a large international group of investigators to examine ubiquitous environmental exposures (i.e., tobacco smoking and residential pesticides) that may be causally associated with the most frequent cancer in children. We provided important information regarding the contribution of various environmental, genetic and epigenetic factors to the risk of childhood leukemia, with a focus on children's exposures to common chemicals with strong biologic plausibility (i.e., benzene, nicotine, PAHs, PCBs, and PBDEs). We quantified PCBs and PBDEs —the most common brominated flame retardants in the United States— in both house dust and serum from childhood leukemia cases and their mothers, strengthening our understanding of exposure pathways possibly leading to childhood leukemia. Novel measurements of adducts and small molecules in archived blood collected at birth provided valuable information about fetal exposures during the final month of gestation. Our comprehensive examination and integrated statistical analysis revealed new relationships and pathways between environmental exposures and key epigenetic factors that impact the genesis of ALL in children. We increased awarness to various audiences about the risk factors associated with the development of childhood lekeumia and possible ways to reduce risk.