PROJECT SUMMARY
Tumors of the central nervous system (CNS) are the most fatal malignancies. Determining the optimal treatment
for patients relies on accurate diagnoses, which at present is only achievable by histopathological analyses of
the tissue obtained by invasive brain surgery. Operations in the brain instill tremendous anxiety in patients and
pose a significant risk for neurological morbidity and even mortality. In addition, accurate longitudinal monitoring
of response to treatment, and distinguishing recurrence from treatment-related effects (pseudoprogression),
relies on limited information obtained on radiologic imaging as the option to invasively obtain tissue
longitudinally over the course of disease is prohibitive due to the morbidity of repeat procedures. Given these
limitations, there has been significant interest in identifying non-invasive or “liquid” biomarkers for the diagnosis
of CNS tumors. Most groups have looked at mutations in circulating cell-free tumor DNA in the plasma or
cerebrospinal fluid. Unfortunately, CNS tumors usually do not release enough DNA into the systemic circulation
for such an approach to be reliable on a prospective clinical basis. Furthermore, searching for single mutations
does not allow for the discrimination of the many different types of brain tumors included in a differential
diagnosis. Instead of focusing on mutations, we focused on DNA methylation alterations that are highly
abundant and cancer specific. We developed, optimized, and validated a novel approach termed cell-free
methylated DNA immuno-precipitation and sequencing (cfMeDIP-seq) that involves isolating and enriching for
methylated fragments of tumor cell-free DNA using a methylation-specific antibody and then sequencing.
Published data from our group provides proof-of-principle that plasma cfMeDIP-seq can reliably and accurately
diagnose tumors throughout the body, including those in the CNS. Based on these data, we hypothesize that
methylation profiling of circulating DNA, via cfMeDIP-seq, identifies reliable biomarkers for the
diagnosis, prognosis, and monitoring of CNS tumors non-invasively. We will build upon our existing data
to use cfMeDIP-seq as a novel technique to establish and validate a comprehensive, non-invasive CNS tumor
classifier (Aim 1); build an accurate predictive model for the non-invasive prognostication of meningiomas (Aim
2); identify plasma biomarkers of response to treatment, recurrence, and malignant transformation in gliomas
(Aim 3); and identify methylation and plasma biomarkers of brain metastases development from systemic
cancers (Aim 4). Success in this proposal will genuinely help shape a much-needed paradigm-shift in the field
of neuro-oncology by establishing cfMeDIP-seq as a reliable liquid biomarker for the non-invasive diagnosis,
prognostication, and monitoring of patients with primary and metastatic CNS tumors.
