Introduction
Genetic mutations have been recognized as one of the primary etiological factors in the development of cancer. Pediatric cancers present a paradoxical situation wherein cancer is present, yet seemingly insufficient time has elapsed for the sufficient genetic "hits" to accrue for cancer to develop. Recently, heritable genetic mutations have been investigated as potential cancer predisposing factors. Specifically, mutations in the function of TP53, RB1, and VHL genes have all shown functional roles in the development of sarcomas, retinoblastoma, and hemangioblastomas in children. However, research investigating predisposition factors in the etiology of the most common pediatric cancer, leukemia, still remains in its infancy. Preliminary research has suggested the role of mutations in Phase I and Phase II xenobiotic metabolizing enzymes in the development of leukemia. Nevertheless, much more work is needed in order to better elucidate these mutations' role in the etiology of leukemia.

To date, no study to date has specifically investigated differences in mutation prevalence in different populations of children with leukemia. Specifically, little has been done in minority populations such as the Hispanic and African-American populations, which constitute a substantial portion of the leukemia caseloads in our and many other pediatric hospitals. Such variations in mutation prevalence could explain the variable assortment of cancer predispositions in these different populations, and better elucidate why certain children are more/less predisposed to different forms of leukemia. It has also been speculated that the differing prevalences of these mutations in different ethnic groups might account for the variable responses to treatment that children experience when placed on a common chemotherapeutic regimen. Furthermore, current advances in cancer treatment suggest that unique "patient-specific" chemotherapeutic regimens might soon become the model of treatment, but this will only be possible with further elucidation of specific enzymes variations between patients which would effect treatment success and toxicity.

Therefore we have developed this case control study of Hispanic, African-American and Caucasian children, wherein we will seek to determine if different genetic factors might contribute to leukemia development. DNA samples (n ~360) were collected from leukemia patients in the hematology/oncology ward at Children's Medical Center of Dallas (CMC), and these samples will be matched by age, ethnicity, and sex to samples obtained from the out-patient clinics at CMC. All samples will be subjected to PCR-ASO testing for specific, pre-established polymorphisms in phase I and phase II enzymes known to have etiological roles in other cancers:

" CYP3A4, CYP1A1, CYP2D4, CYP1E1
" N-acetyltransferase 1 and 2
" Glutathione-S-transferase mu1, pi1, theta1
" Myeloperoxidase

Upon completion of data collection, these samples will be analyzed to determine:

" If significant differences in prevalence exist between the different ethnic groups
" If any polymorphisms are significantly associated with leukemia when cases are compared to their matched controls
" If different combinations of mutations have synergistic/interactive effects on cancer risk

It is our hypothesis that each ethnic group should have its own novel mutation profile, in a fashion suggested by prior epidemiological research. We anticipate that this study will better elaborate on these prior findings, and more succinctly compare the mutation variations between these communities. In addition, we expect that our study may find novel differences between cases and controls of the same ethnicity, which should elucidate how different mutation profiles may function in the etiology of leukemia. Finally, we anticipate that the effects of combinations of mutations in several xenobiotic metabolizing enzymes should have synergistic effects on leukemia predisposition.

Methods:
Samples: DNA samples were obtained via blood draw from roughly 360 children diagnosed with B-lineage leukemia over the past 5 years. All samples where amplified via PCR and stored as part of a separate investigation protocol, and will be accessed for this study. Matched controls from non-leukemic children will be obtained from residual blood left over from blood draws in the Childrens Medical Center laboratory. Cases and controls will be matched based on sex, age, and ethnicity information that were retained from the patient's medical charts. All other identifying information has withheld from this study's investigators.

The controls and cases will be amplified via standard polymerase chain-reaction, and subsequently sequenced for novel polymorphisms in the prestated enzymes via allele-specific oligonucleotide analysis utilizing primers obtained from prior research studies. Each patient will be identified as either homozygous or heterozygous for the wild-type and mutations alleles. All protocols have been approved by the Institutional Review Board.

Upon completion of data analysis, all data will be analyzed by an independent biostatistician at UT Southwestern Medical Center, who will construct odds ratios, significance values, and confidence intervals for each mutation between the different ethnic groups and the case/control dyads. The alpha value will be set at the standard .05 value, and any value found to have a p-value below this will be considered significant. Currently, power studies are being completed to determine if our sample sizes are sufficient and what the smallest detectable difference in values should be. (need to update after power calculations done)