UC Berkeley

The etiology of childhood acute lymphoblastic leukemia (ALL), the most common childhood cancer, is unknown. A role for infection is supported by considerable research over the last twenty years. Despite this evidence and many infection related hypotheses, no definitive infectious agent has been identified. Until recently broad surveys of the microbiome were not possible. Modern developments in next generation sequencing and bioinformatics however now allow unbiased investigation into the astounding diversity of organisms that inhabit the human body.

This dissertation describes the first comprehensive analysis of tissue from children with childhood leukemia using metagenomic (the study of non-human nucleic acids) and epidemiologic methods. Metagenomes in pre treatment bone marrow from the Fallon, NV ALL cluster and other ALL non-clustered cases from California were interrogated for known and novel infections. A putative and potentially novel human virus was found in the Fallon cases. To investigate viruses involved more broadly in childhood ALL, viromes of pre-treatment bone marrow from children with acute myloid leukemia (AML) and from children with ALL (who we assumed to have similar background infections) were compared. In children with ALL, a greater intensity of viral infection as well as a significant difference in the expression of viral packets originating from double stranded DNA viruses was observed. Using 16s based bacterial metagenomics, no difference was observed in bacterial diversity between the groups showing the absence of differential contamination and suggesting that bacteria do not play a role in ALL. These results suggest a possible putative causal agent in the Fallon Nevada leukemia cluster and increased viral infection driven by double stranded DNA viruses in non-clustered children with ALL.

These data contribute to an etiologic hypothesis of ALL where a combination of pre-natal and post-natal events lead to leukemia. First, an underlying genetic predisposition, combined with an in utero infectious exposure, cause chromosomal anomalies and immune dysregulation. This followed by postnatal events, including toxic and infectious exposures that further modify adaptive immunity lead to a greater intensity of infections. This feeds the loop of enhanced b-cell production that facilitates virally induced self- mutation caused by foreign nucleic acids invading pre b-cells. Resulting in an ABOBEC arbitrated hypermutation. Eventually the loop of enhanced yet less competent b-cell production, RAG mediated over mutation, increased infection leading to pre b-cell viral nucleic acid loading, and APOBEC caused self mutation remove the proliferative and apoptotic cellular checks. The result is childhood ALL.