Uncovering Distinct Metabolic Properties of Relapse and Relapse-Fated Clones in B-Cell Acute Lymphoblastic Leukemia

B-cell ALL (B-ALL) accounts for 80% of pediatric and 75% of adult cases of ALL. Although primary treatment is often effective, 15-20% of pediatric, and 40-50% of adult patients will relapse. Upon relapse, disease is often resistant to treatment and survival is poor. Thus, there is an unmet need to identify patients at risk of relapse earlier and develop treatments to target relapse-fated cells before they fully evolve and progress to relapse.

Through xenografting of primary patient samples at clonal doses and sequencing of both xenograft and bulk patient samples, our lab previously discovered a set of minor genetic B-ALL subclones which are present at diagnosis and fated to initiate relapse (Dobson, Cancer Discovery, 2020). These subclones, termed diagnosis relapse initiating clones (dRIs), have a phenotype intermediate to diagnosis and relapse clones, with dRIs sharing many features of relapse clones. Relapse and dRI clones possess chemotherapy drug tolerance and a unique metabolism, including enrichment of genes involved in oxidative phosphorylation (OXPHOS).

Altered metabolism (increased OXPHOS) causes chemotherapy resistance in acute myeloid leukemia (AML; Farge, Cancer Discovery, 2017). Thus, we have focused our study on the unique metabolism of dRI and relapse clones as compared to the dominant diagnosis clone to determine if metabolism and chemotherapy resistance are linked in dRI and relapse B-ALL clones, which can be leveraged to target these clones therapeutically. We discovered metabolic phenotypes are highly consistent across the individual clonal populations (diagnosis, dRI, and relapse) belonging to the DUX-4 subtype of B-ALL.

To functionally validate increased OXPHOS in dRI, and relapse clones compared to diagnosis clones, as predicted by enrichment analysis, we performed the SeaHorse Mito Stress Test on diagnosis, dRI, and relapse clones from patient derived xenografts. ATP production, a measure of relative ATP demand, was significantly increased in relapse clones compared to diagnosis clones of the DUX-4 subtype. dRI clones had a trend for ATP production that was intermediate to diagnosis and relapse clones.

As another approach to evaluate the reliance of dRI and relapse clones on OXPHOS we developed an in vitro drug assay in which we cultured diagnosis, dRI, and relapse clones with the drug venetoclax, which has been shown to target OXPHOS independent of BCL-2 expression and is used clinically to treat other leukemias, including chronic lymphocytic leukemia and AML. Diagnosis, dRI, and relapse clones all showed a decrease in viability upon treatment with venetoclax. However, there was a greater proportional decrease in viability relative to the vehicle control in relapse clones than in diagnosis and dRI clones with the dRI clones being intermediate.

To examine global differences in the metabolism of diagnosis, dRI, and relapse clones we performed mass spectrometry-based metabolomics. DUX-4 relapse clones had depletion of TCA cycle intermediates compared to diagnosis clones. The abundance of TCA cycle intermediates in dRI clones was between that of diagnosis and relapse clones. The observed depletion of TCA cycle intermediates in relapse and dRI clones may indicate increased TCA cycle activity to fuel the increased OXPHOS observed in relapse and dRI clones compared to diagnosis clones.

Through our global metabolomics analysis, we also observed significant decrease in the abundance of long chain fatty acids in DUX-4 relapse clones compared to diagnosis clones, with a trend for dRIs intermediate to diagnosis and relapse clones. These findings suggest long chain fatty acids may undergo fatty acid oxidation (FAO) where fatty acids are converted to acetyl-CoA, which enters the TCA cycle to generate NADH and FADH₂ to fuel the increased OXPHOS observed in relapse and to a lesser extent dRI clones; a hypothesis we are testing.

Together, our data shows that dRI and relapse clones possess a unique metabolism with increased reliance on OXPHOS. Global metabolomics analysis suggest that this increased OXPHOS may be fuelled by FAO via the TCA cycle. Additional metabolomics experiments are ongoing to evaluate the reliance of dRI and relapse clones on FAO to fuel increased OXPHOS and to further characterize the unique metabolic phenotype of dRI and relapse clones. Overall, our data suggest that the unique metabolic requirements of dRI and relapse clones might reveal novel therapeutic targets.

Dick:Celgene/BMS: Research Funding; Trillium Therapeutics/Pfizer: Patents & Royalties: patent licencing; Graphite Bio: Membership on an entity's Board of Directors or advisory committees.