Investigating DNA methylation as a mediator of genetic risk in childhood acute lymphoblastic leukemia.

Genome-wide association studies have identified a growing number of single nucleotide polymorphisms (SNPs) associated with childhood acute lymphoblastic leukemia (ALL), yet the functional roles of most SNPs are unclear. Multiple lines of evidence suggest that epigenetic mechanisms may mediate the impact of heritable genetic variation on phenotypes. Here, we investigated whether DNA methylation mediates the effect of genetic risk loci for childhood ALL. We performed an epigenome-wide association study (EWAS) including 808 childhood ALL cases and 919 controls from California-based studies using neonatal blood DNA. For differentially methylated CpG positions (DMPs), we next conducted association analysis with 23 known ALL risk SNPs followed by causal mediation analyses addressing the significant SNP-DMP pairs. DNA methylation at CpG cg01139861, in the promoter region of IKZF1, mediated the effects of the intronic IKZF1 risk SNP rs78396808, with the average causal mediation effect (ACME) explaining ~30% of the total effect (ACME P = 0.0031). In analyses stratified by self-reported race/ethnicity, the mediation effect was only significant in Latinos, explaining ~41% of the total effect of rs78396808 on ALL risk (ACME P = 0.0037). Conditional analyses confirmed the presence of at least three independent genetic risk loci for childhood ALL at IKZF1, with rs78396808 unique to non-European populations. We also demonstrated that the most significant DMP in the EWAS, CpG cg13344587 at gene ARID5B (P = 8.61 × 10-10), was entirely confounded by the ARID5B ALL risk SNP rs7090445. Our findings provide new insights into the functional pathways of ALL risk SNPs and the DNA methylation differences associated with risk of childhood ALL.

Association between birth characteristics and incidence of pituitary adenoma and craniopharyngioma: a registry-based study in California, 2001-2015.

PURPOSE: To evaluate the association between birth characteristics, including parental sociodemographic characteristics, and early-onset pituitary adenoma (PA) and craniopharyngioma. METHODS: Leveraging the population-based California Linkage Study of Early-onset Cancers, we identified the birth characteristics of incident cases with PA (n = 1,749) or craniopharyngioma (n = 227) who were born from 1978 to 2015 and diagnosed 1988-2015, as well as controls in a 50:1 ratio matched on birth year. Adjusted odds ratios (OR) and 95% confidence interval (CI) estimates were computed using unconditional multivariable logistic regression. RESULTS: Males had a lower risk of PA than females (OR = 0.37, 95%CI: 0.34-0.41), and Black (OR = 1.55, 95%CI: 1.30-1.84) or Hispanic (OR = 1.53, 95%CI: 1.34-1.74) individuals had a higher risk compared to non-Hispanic Whites. Older maternal age was positively associated with PA (OR = 1.09, 95%CI: 1.04-1.15 per 5 years, p < 0.01), as was higher maternal education (OR = 1.12, 95%CI: 1.04-1.20 per year, p < 0.01). There were no statistically significant associations between birthweight (OR = 1.04, 95%CI: 0.99-1.09 per 500 g, p = 0.12), birth plurality, or birth order and PA. When stratified by race and ethnicity, the significant association with maternal education was identified only for non-Hispanic White individuals. On multivariable logistic regression, no statistically significant associations were identified between birth characteristics and incidence of craniopharyngioma, except that risk was higher among Hispanic (OR = 1.45, 95%CI: 1.01-2.08) compared to non-Hispanic White individuals. CONCLUSION: In this large, population-based study, female sex, older maternal age, higher maternal education, and Hispanic ethnicity and Black race compared to non-Hispanic White race, were associated with an increased risk of PA in children and young adults.

Birth characteristics and risk of Ewing sarcoma.

PURPOSE: The incidence of Ewing sarcoma varies according to race and ethnicity, and genetic susceptibility is known to affect disease risk. Apart from these factors, the etiology of Ewing sarcoma is largely unknown. METHODS: We compared the birth characteristics of a population-based series of 556 Ewing sarcoma cases born in California in 1978-2015 and diagnosed in 1988-2015 with those of 27,800 controls selected from statewide birth records and frequency-matched to cases on the year of birth, using multivariable logistic regression models. We also assessed whether Ewing sarcoma clustered within families. RESULTS: Compared to non-Hispanic White subjects, Black (odds ratio [OR] = 0.07, 95% confidence interval [CI] 0.03-0.18), Asian (OR = 0.57, 95% CI 0.41-0.80), and Hispanic (OR = 0.73, 95% CI 0.62-0.88) individuals had a significantly lower risk of Ewing sarcoma. Race and ethnicity differences were more profound for metastatic Ewing sarcoma. Birthweight was also identified as a significant risk factor (OR = 1.09, 95% CI 1.00-1.18 for each 500 g increase in birthweight). A separate family-based cancer clustering analysis did not suggest any strong role for familial predisposition alleles. CONCLUSIONS: This population-based study with minimal selection bias provides support for a role of accelerated fetal growth in the etiology of Ewing sarcoma in addition to more precise estimates of racial and ethnic variations in disease risk. This comparatively large analysis of birth characteristics and Ewing sarcoma in a multiethnic population should stimulate further investigations into genetic and environmental causes.

History of Early Childhood Infections and Acute Lymphoblastic Leukemia Risk Among Children in a US Integrated Health-Care System.

Surrogate measures of infectious exposures have been consistently associated with lower childhood acute lymphoblastic leukemia (ALL) risk. However, recent reports have suggested that physician-diagnosed early-life infections increase ALL risk, thereby raising the possibility that stronger responses to infections might promote risk. We examined whether medically diagnosed infections were related to childhood ALL risk in an integrated health-care system in the United States. Cases of ALL (n = 435) diagnosed between 1994-2014 among children aged 0-14 years, along with matched controls (n = 2,170), were identified at Kaiser Permanente Northern California. Conditional logistic regression was used to estimate risk of ALL associated with history of infections during first year of life and across the lifetime (up to diagnosis). History of infection during first year of life was not associated with ALL risk (odds ratio (OR) = 0.85, 95% confidence interval (CI): 0.60, 1.21). However, infections with at least 1 medication prescribed (i.e., more "severe" infections) were inversely associated with risk (OR = 0.42, 95% CI: 0.20, 0.88). Similar associations were observed when the exposure window was expanded to include medication-prescribed infections throughout the subjects' lifetime (OR = 0.52, 95% CI: 0.32, 0.85).

Genetic determinants of childhood and adult height associated with osteosarcoma risk.

BACKGROUND: Although increased height has been associated with osteosarcoma risk in previous epidemiologic studies, to the authors' knowledge the relative contribution of stature during different developmental timepoints remains unclear. Furthermore, the question of how genetic determinants of height impact osteosarcoma etiology remains unexplored. Genetic variants associated with stature in previous genome-wide association studies may be biomarkers of osteosarcoma risk. METHODS: The authors tested the associations between osteosarcoma risk and polygenic scores for adult height (416 variants), childhood height (6 variants), and birth length (5 variants) in 864 osteosarcoma cases and 1879 controls of European ancestry. RESULTS: Each standard deviation increase in the polygenic score for adult height, corresponding to a 1.7-cm increase in stature, was found to be associated with a 1.10-fold increase in the risk of osteosarcoma (95% confidence interval [95% CI], 1.01-1.19; P =.027). Each standard deviation increase in the polygenic score for childhood height, corresponding to a 0.5-cm increase in stature, was associated with a 1.10-fold increase in the risk of osteosarcoma (95% CI, 1.01-1.20; P =.023). The polygenic score for birth length was not found to be associated with osteosarcoma risk (P =.11). When adult and childhood height scores were modeled together, they were found to be independently associated with osteosarcoma risk (P =.037 and P = .043, respectively). An expression quantitative trait locus for cartilage intermediate layer protein 2 (CILP2), rs8103992, was significantly associated with osteosarcoma risk after adjustment for multiple comparisons (odds ratio, 1.35; 95% CI, 1.16-1.56 [P = 7.93×10-5 and Padjusted =.034]). CONCLUSIONS: A genetic propensity for taller adult and childhood height attainments contributed independently to osteosarcoma risk in the current study data. These results suggest that the biological pathways affecting normal bone growth may be involved in osteosarcoma etiology.

Perinatal factors associated with clinical presentation of osteosarcoma in children and adolescents.

BACKGROUND: Osteosarcoma typically develops during puberty with tumors arising at sites of rapid bone growth, suggesting a role for growth-regulating pathways in tumor etiology. Birthweight is one measure of perinatal growth that has been investigated as an osteosarcoma risk factor. Whether birthweight affects clinical features of osteosarcoma remains unexplored. METHOD: Six hundred seventy patients with osteosarcoma, aged 0-19 years, were identified through the California Cancer Registry. We analyzed birth certificate data from the California Department of Public Health vital statistics unit for these patients and 2,860 controls, matched by sex, birth-year, and race/ethnicity. We examined the impact of birthweight on the risk, timing, and clinical presentation of pediatric osteosarcoma including tumor location, size, extension, differentiation, presence of metastasis, and age at onset. Regression models were adjusted for race, sex, gestational age, socioeconomic status, and tumor site. RESULTS: Higher birthweight was associated with more advanced tumor stage (P = 0.017), a trend toward greater tumor extension into surrounding tissues (P = 0.083), and with occurrence of tumors in sites other than the long bones of the arms/legs (P = 9.7 × 10-3 ). Higher birthweight was also associated with an increased likelihood of metastases present at diagnosis (P = 0.047), with each 200 g increase in birthweight associated with a 1.11-fold increase in the odds of having metastatic disease (95% confidence interval: 1.01-1.22). CONCLUSIONS: The association between higher birthweight and more aggressive osteosarcoma, frequently occurring at sites other than the long bones, suggests that growth pathways active during gestation may play an important role in future osteosarcoma progression, especially at anatomic sites with diminished rates of osteoblastic proliferation.

Pre- and Postnatal Exposures to Tobacco Smoking and Survival of Childhood Acute Lymphoblastic and Myeloid Leukemias in California, United States.

BACKGROUND: Tobacco smoke adversely affects the prognosis of adult cancers including myeloid leukemia, but less is known in children. METHODS: We evaluated whether pre- and postnatal exposures to tobacco smoke decrease 5-year survival of 1,235 childhood acute lymphoblastic leukemia (ALL) and 188 childhood acute myeloid leukemia (AML) cases derived from a population-based case-control study in California. Cases were diagnosed between 1995 and 2015 (median follow-up time of 13.2 years overall). We obtained data on tobacco smoking (before conception, during pregnancy, after birth), parental education and income, clinical features, and vital status through 2020. Cox proportional hazards regression was used to calculate hazard ratios (HR) and 95% confidence intervals (CI) for mortality associated with smoking, adjusting for sociodemographic characteristics and risk group (ALL only). RESULTS: About 23% of mothers and 39% of fathers reported smoking and 130 children with ALL and 52 with AML died within 5 years. For AML, increased risks of death were observed among children whose fathers smoked before conception compared with nonsmoking fathers [HR = 1.41; 95% confidence interval (CI), 0.95-3.44 and 3.47; 95% CI, 1.37-8.81, respectively for <20 vs. ≥20 cigarettes per day; Ptrend = 0.01]. HR for child's passive smoking was 1.74, 95% CI, 0.81-3.73. Paternal preconception smoking may also reduce 5-year survival among ALL with favorable prognostic molecular subtypes (high hyperdiploidy and absence of IKZF1 gene deletion), although the associations did not reach statistical significance (Pheterogeneity = 0.07). CONCLUSIONS: Paternal preconception smoking decreased 5-year survival of childhood AML. IMPACT: Knowledge of exposure to tobacco smoking should be integrated in the treatment plan of childhood leukemias.

Genetic variation near GRB10 associated with bone growth and osteosarcoma risk in canine and human populations.

BACKGROUND: Canine and human osteosarcoma are similar in clinical presentation and tumor genomics. Giant breed dogs experience elevated osteosarcoma incidence, and taller stature remains a consistent risk factor for human osteosarcoma. Whether evolutionarily conserved genes contribute to both human and canine osteosarcoma predisposition merits evaluation. METHODS: A multi-center sample of childhood osteosarcoma patients and controls underwent genome-wide genotyping and imputation. Ancestry-adjusted SNP associations were calculated within each dataset using logistic regression, then meta-analyzed across the three datasets, totaling 1091 patients and 3026 controls. Ten regions previously associated with canine osteosarcoma risk were mapped to the human genome, spanning ∼6 Mb. We prioritized association testing of 5985 human SNPs mapping to candidate osteosarcoma risk regions detected in Irish wolfhounds, the largest dog breed studied. Secondary analyses explored 6289 additional human SNPs mapping to candidate osteosarcoma risk regions identified in Rottweilers and greyhounds. RESULTS: Fourteen SNPs were associated with human osteosarcoma risk after adjustment for multiple comparisons, all within a 42 kb region of human Chromosome 7p12.1. The lead variant was rs17454681 (OR=1.25, 95 %CI: 1.12-1.39; P=4.1×10-5), and independent risk variants were not observed in conditional analyses. While the associated region spanned 2.1 Mb and contained eight genes in Irish wolfhounds, associations were localized to a 50-fold smaller region of the human genome and strongly implicate GRB10 (growth factor receptor-bound protein 10) in canine and human osteosarcoma predisposition. PheWAS analysis in UK Biobank data identified noteworthy associations of the rs17454681 risk allele with varied measures of height and pubertal timing. CONCLUSIONS: Our comparative oncology analysis identified a novel human osteosarcoma risk allele near GRB10, a growth inhibitor that suppresses activated receptor tyrosine kinases including IGF1R, PDGFRB, and EGFR. Epidemiologists may benefit from leveraging cross-species comparisons to identify haplotypes in highly susceptible but genetically homogenous populations of domesticated animals, then fine-mapping these associations in diverse human populations.

Folate Metabolism and Risk of Childhood Acute Lymphoblastic Leukemia: A Genetic Pathway Analysis from the Childhood Cancer and Leukemia International Consortium.

BACKGROUND: Prenatal folate supplementation has been consistently associated with a reduced risk of childhood acute lymphoblastic leukemia (ALL). Previous germline genetic studies examining the one carbon (folate) metabolism pathway were limited in sample size, scope, and population diversity and led to inconclusive results. METHODS: We evaluated whether ∼2,900 single-nucleotide polymorphisms (SNP) within 46 candidate genes involved in the folate metabolism pathway influence the risk of childhood ALL, using genome-wide data from nine case-control studies in the Childhood Cancer and Leukemia International Consortium (n = 9,058 cases including 4,510 children of European ancestry, 3,018 Latinx, and 1,406 Asians, and 92,364 controls). Each study followed a standardized protocol for quality control and imputation of genome-wide data and summary statistics were meta-analyzed for all children combined and by major ancestry group using METAL software. RESULTS: None of the selected SNPs reached statistical significance, overall and for major ancestry groups (using adjusted Bonferroni P-value of 5 × 10-6 and less-stringent P-value of 3.5 × 10-5 accounting for the number of "independent" SNPs). None of the 10 top (nonsignificant) SNPs and corresponding genes overlapped across ancestry groups. CONCLUSIONS: This large meta-analysis of original data does not reveal associations between many common genetic variants in the folate metabolism pathway and childhood ALL in various ancestry groups. IMPACT: Genetic variants in the folate pathway alone do not appear to substantially influence childhood acute lymphoblastic leukemia risk. Other mechanisms such as gene-folate interaction, DNA methylation, or maternal genetic effects may explain the observed associations with self-reported prenatal folate intake.

Evaluating genomic polygenic risk scores for childhood acute lymphoblastic leukemia in Latinos.

The utility of polygenic risk score (PRS) models has not been comprehensively evaluated for childhood acute lymphoblastic leukemia (ALL), the most common type of cancer in children. Previous PRS models for ALL were based on significant loci observed in genome-wide association studies (GWASs), even though genomic PRS models have been shown to improve prediction performance for a number of complex diseases. In the United States, Latino (LAT) children have the highest risk of ALL, but the transferability of PRS models to LAT children has not been studied. In this study, we constructed and evaluated genomic PRS models based on either non-Latino White (NLW) GWAS or a multi-ancestry GWAS. We found that the best PRS models performed similarly between held-out NLW and LAT samples (PseudoR2 = 0.086 ± 0.023 in NLW vs. 0.060 ± 0.020 in LAT), and can be improved for LAT if we performed GWAS in LAT-only (PseudoR2 = 0.116 ± 0.026) or multi-ancestry samples (PseudoR2 = 0.131 ± 0.025). However, the best genomic models currently do not have better prediction accuracy than a conventional model using all known ALL-associated loci in the literature (PseudoR2 = 0.166 ± 0.025), which includes loci from GWAS populations that we could not access to train genomic PRS models. Our results suggest that larger and more inclusive GWASs may be needed for genomic PRS to be useful for ALL. Moreover, the comparable performance between populations may suggest a more oligogenic architecture for ALL, where some large effect loci may be shared between populations. Future PRS models that move away from the infinite causal loci assumption may further improve PRS for ALL.