NRAS palmitoylation and oncogenic fitness

In this issue of Blood, Zambetti et al validate that palmitoylation of mutant NRAS is essential for its oncogenic function by using a genetic mouse model. Their study highlights that the palmitoylation/depalmitoylation cycle is an attractive candidate for therapeutic intervention in hematologic malignancies with NRAS mutations.¹ 

A refined understanding of the mechanistic underpinnings driving neuroblastoma-RAS (NRAS) oncogenic function is critical for the development of effective cancer therapies. NRAS is frequently mutated in hematologic malignancies, primarily in multiple myeloma, myelodysplastic syndrome/myeloproliferative neoplasia (MDS/MPN) overlap syndromes, and acute myeloid leukemia.²  These mutations lock the enzyme in an active conformation and drive ligand-independent anti-apoptotic signals.³ 

Like other mutant RAS proteins, oncogenic NRAS has been considered undruggable because it lacks deep pockets suitable for stable binding of small molecule inhibitors. Alternatively, the disruption of posttranslational modifications that activate NRAS function may define molecular targets of clinical value. Efforts to inhibit farnesylation, a lipid modification at cysteine-186 (C186) that is necessary for its biologic and oncogenic function, have been largely unsuccessful because RAS-mutant cancer cells become resistant by using geranyl-geranylation. This alternative modification, which is not normally used at this site in hematopoietic progenitors, reactivates oncogenic NRAS (and KRAS) function in the neoplastic cells.⁴ 

Farnesylated NRAS is palmitoylated at C181 in the Golgi apparatus and is then transported to the plasma membrane by the VPS35 chaperone protein. In contrast to farnesylation, palmitoylation is reversible, and NRAS undergoes rapid palmitoylation and depalmitoylation that shuttles the protein between an active and palmitoylated form in the plasma membrane and an inactive and depalmitoylated form in the Golgi apparatus.⁵  Despite the well-documented implication of palmitoyl modifications in NRAS function, previous studies have not been able to establish the essentiality of palmoitylation in hematologic malignancies. Studies by Cuiffo and Ren⁶  and Xu et al⁷  used a retroviral transduction approach with overexpressed palmitoylation-defective oncogenic Nras (Nras^G12D,C181S) protein to demonstrate that C181 is important for its oncogenic activity and for its proper subcellular localization.

Notably, Nras^G12D,C181S expression had a strong negative impact on signaling activation in hematopoietic cells. However, data showing that expression levels of Ras oncoprotein affect specificity and function have shed doubts on studies in which nonphysiologic levels of Nras expression were used. Indeed, inducible and physiologic expression levels of the oncogenic Nras^G12D from its endogenous locus (Nras^LSL−G12D) resulted in the development of a variety of hematologic malignancies in mice in a dose-dependent manner.^8,9 

To clearly define the relevance of palmitoylation in hematopoietic transformation by Nras^G12D, Zambetti et al developed an Nras^{LSL−G12D,C181S}-inducible double knockin mouse strain, in which physiologic levels of the palmitoylation-defective oncogenic Nras protein (Nras^G12D,C181S) are expressed from the Nras locus. The authors’ compelling results demonstrate that the suppression of C181 palmitoyl chains is essential for Nras^G12D oncogenic activity.

An important point that emerges from the study is that genetic suppression of palmitoylation completely abrogates Nras^G12D-driven neoplasia. Because NRAS mutations are heterozygous in practically all cases of human hematologic malignancies, these results suggest that efficient inhibition of palmitoylation could hinder oncogenic fitness and reverse neoplastic expansion. Notably, this suppression was effective when either 1 or both alleles express Nras^G12D,C181S for lymphoid and myeloid neoplasms, suggesting that palmitoylation may be an attractive target for different hematologic malignancies.

Previous studies have reported that increased gene dosage drives Nras^G12D-dependent aggressive MPNs in vivo.^7,9  Zambetti et al evaluated the impact of reduced palmitoylation on the latency of Nras-driven MPNs by using mice with either 2 copies of G12D or 1 copy of G12D and 1 copy of G12D,C181S mutations at the Nras locus. They determined that the C181S mutation rendered the encoded protein non-oncogenic, because mice with C181S showed an intermediate disease latency. Further analysis unexpectedly revealed a clonal outgrowth of cells that had reverted the C181S mutation in G12D,G12D background in mice that developed aggressive MPN and other hematologic diseases. These data indicate that suppression of palmitoylation may result in a fitness disadvantage.

Hence, the essentiality of the palmitoylation/depalmitoylation modifications in hematopoietic neoplasias with NRAS driver mutations provides proof-of-principle evidence for therapeutic intervention. However, further mechanistic and functional studies are necessary before embarking on clinical studies. For example, there is a limited understanding of which (and how many) enzymes are involved in the palmitoylation/depalmitoylation cycle, and whether there is a redundancy in different hematopoietic compartments. Targeting the ABHD17 family of serine hydrolases, which depalmitoylate NRAS,¹⁰  is an alternative potential therapeutic strategy. Finally, the identification of revertant aggressive clones with growth advantage may indicate that C181 is under high selective pressure, and that pharmacologic interference of palmitoylation may promote the emergence of resistant clones that escape this block and reactivate neoplastic outgrowth. This could be envisioned by a mutation in the targeted palmitoyl-transferase that excludes the inhibitor and reestablishes palmitoylation, a mutation in an alternative pathway that reactivates MAPK signaling, or by the use of alternative enzymes and/or lipid moieties, as was found when using farnesylation inhibitors.