Metabolic Determinants of Ferroptosis in B-Cell Lymphoma

Background and Significance: Approaches for targeted elimination of B-cells as the root-cause of disease have been highly effective in the treatment of B-cell lymphoma (rituximab, ibrutinib). To expand the portfolio of B-cell-selective drugs, we developed an interactive computational tool (www.lymphoblasts.org) and identified ferroptosis previously unrecognized selective vulnerability in B-cell lymphomas. Ferroptosis, a non-apoptotic form of cell death driven by iron-dependent membrane lipid peroxidation, has shown therapeutic potential in highly therapy-resistant tumors. However, no potent ferroptosis inducers are available clinically, and genetic ablation of Gpx4, the main anti-ferroptotic defense, leads to lethal renal and neurological toxicities in mice. A better understanding of the factors regulating ferroptosis is therefore required to allow its therapeutic targeting and the development of clinical grade ferroptosis inducers.

Results: Compound screening data (CTD, GDSC), together with gene-dependency scores from CRISPR and RNAi screens, were re-analyzed and integrated for B-ALL/mature B-cell lymphomas compared to myeloid leukemia and solid tumors. Most notably, among 4,518 compounds screened, we identified 5 top-ranking compounds that converge on targeting the ferroptosis pathway: Based on our computational approach (www.lymphoblasts.org), we identified ferroptosis inducers as particularly lethal to B-cell lymphoma compared to other tumor types. This was the case for multiple classes of ferroptosis inducers, including GPX4 inhibitors (RSL3, ML162, ML210), the cysteine-glutamate antiporter (system x_c^-, SLC7A11) inhibitor Erastin, and the iron oxidizer and GPX4 inactivator FINO₂.

Analysis of gene dependency in 89 B-cell lymphoma compared to 779 myeloid leukemia and solid tumor cell lines based on CRISPR and RNAi data (DepMap) revealed a selective dependence of B-cell lymphomas on anti-ferroptotic molecules. These genes notably include components of the glutathione synthesis machinery (GCLC, GCLM, GSS) and selenoprotein synthesis (SEPH2, LRP8, SELENOI). Interestingly, B-cell malignancies also showed marked dependence on molecules that sensitize to ferroptosis, including the transferrin receptor (TFRC; iron import) and ACSL4 (polyunsaturated fatty acid, PUFA). This suggests that B-cells might be addicted to iron- and PUFA-metabolism that make them intrinsically vulnerable to ferroptosis. Accordingly, iron chelation treatment showed a much higher toxicity in B-cell lymphoma compared to solid tumors.

Based on transduction of mature splenic B-cells from Tfrc-fl/fl mice with MYC, BCL2 and dominant-negative p53 as B-cell lymphoma model, we confirmed that Cre-mediated ablation of Tfrc induced rapid cell death, selectively in mature B-lymphoma. In addition, analysis of clinical data from the DLBCL MMMLNP trial cohort revealed that greater than median expression of ACSL4 is associated with significantly worse survival (P=0.003). The opposite was seen with ACSL3 (P=0.0004), which counteracts ACSL4-dependent PUFA-metabolism. Combination of these two gene expression profiles strongly predicted prognosis, as individuals with the ACSL4_high/ACSL3_low gene expression signature had much higher mortality than those with the ACSL4_low/ACSL3_high signature (hazard ratio for mortality by log-rank test: 2.484, 95% confidence interval: 1.709 – 3.609).

Whole-genome CRISPR screens in B-cell malignancies under the selective pressure of RSL3, Erastin, and FINO₂ highlighted several known ferroptosis regulators and uncovered new ferroptosis-related genes and pathways including sphingolipid metabolism (KDSR), phosphatidylcholine synthesis (FLVCR1, PCYT1A), inositol metabolism (ITPK1, IPMK), and lipid membrane remodelling (ATP8B2).

Conclusions: Integrating CRISPR dependency and drug sensitivity data revealed that B-cell lymphomas are uniquely sensitive to ferroptosis and highly depend on anti-ferroptotic defense mechanisms. This likely reflects the intrinsic dependency of B-cell lymphomas on iron import and PUFA-metabolism, at the cost of increased vulnerability to ferroptosis. Furthermore, our CRISPR screens performed with the ferroptosis inducers RSL3, Erastin, and FINO₂ revealed phosphatidylcholine synthesis (FLVCR1, PCYT1A) and lipid membrane remodelling (ATP8B2) as central mechanistic targets in B-cell lymphomas.