(P 21) LSD1 INHIBITOR SP-2509 DISRUPTS IRON-SULFUR CLUSTER BIOGENESIS

Objective: We sought to determine the mechanism of action by which the LSD1 inhibitor SP-2509 leads to the reversal of EWSR1::FLI1 transcriptional activity and cell death in Ewing sarcoma cells.

Methods: LSD1 was depleted with a dox inducible dCas9-KRAB system in A673 cells. Cellular Thermal Shift Assay (CETSA) was used to assess shifts in thermal stability of proteins in response to compounds of interest. Cell-titer glo was used to assess cell viability. Mass spectrometry coupled with cysteine reactivity was used to assess Iron-Sulfur (Fe-S) cluster protein abundance and prosthetic group occupancy. RNA-seq was performed to assess gene expression.

Results: We first evaluated SP-2509 binding to hits from our published CRISPR screen. After 3 hours of treatment with 1 uM SP-2509 UQCRFS1 and its chaperone LYRM7 show reduced thermal stability. Cells treated for 24 or 48 hours show reductions in protein abundance of both UQCRFS1 and NDUFS1, but not UQCRFS1 interactor CYC1.
UQCRFS1 and NDUFS1 are both Fe-S cluster containing proteins. To assay the whether other Fe-S proteins were affected by SP-2509, cells were treated with 1 uM SP-2509 for 16 hours and assayed for protein abundance and cysteine reactivity. Fe-S proteins showed reduced protein abundance as well as increased cysteine reactivity indicating decreased binding to the Fe-S cofactor.
Depletion of Fe-S cluster proteins, such as UQCRFS1 and NDUFS1 was previously noted in cuproptosis, but, the copper chelator tetrathiomolybdate shows minimal effects on the potency of SP-2509. The addition of copper chloride and the iron transport mediator transferrin both induced resistance. These data suggest that SP-2509 is not a copper or iron ionophore.
We then tested whether Fe-S protein disruption depends upon LSD1 inhibition using genetic and chemical approaches. Inducible CRISPRi KD of LSD1 does not change SP-2509 IC50. We then tested 3 compounds structurally similar to SP-2509 (cpds 15, 19, and 20), but lacking LSD1 inhibitor activity. These compounds show comparable cytotoxicity and UQCRFS1 destabilization, indicating Fe-S protein disruption is independent from LSD1 inhibition.
We then used SP-2509, the catalytic LSD1 inhibitor OG-L002, and cpds 15, 19, and 20 to ask whether Fe-S disruption correlated with the reversal of EWSR1::FLI1 gene regulation. Principle component analysis shows cpd 15, 19, and 20 samples cluster with SP-2509, while OG-L002 clusters with control samples. Pathway analysis shows several pathways that are upregulated following treatment such as Protein Secretion/uptake, Hypoxia, Heme Metabolism, Unfolded Protein Response and Apoptosis.

Conclusion: Inhibition of LSD1, by CRISPRi or OG-L002 shows no impact on cell viability over 72 hours. Changes in gene regulation and cell viability caused by SP-2509 appears independent of LSD1 inhibition. Our data indicate that SP-2509 and similar compounds disrupt the incorporation of Fe-S clusters into nascent Fe-S proteins. The resulting instability and subsequent unfolding leads to stress response and cell death. SP-2509 does not appear to function as a copper or iron ionophore and how SP-2509 disrupts Fe-S biogenesis is an area of ongoing study. There remain several unexplored aspects of this activity, such as the mechanism by which Fe-S cluster disruption affects fusion oncoprotein activity, the reason why fusion oncoprotein driven sarcomas are sensitive to SP-2509 treatment and why this type of Fe-S cluster disruption is generally well tolerated in clinical studies.