Inhibiting the palmitoylation/depalmitoylation cycle selectively reduces the growth of hematopoietic cells expressing oncogenic Nras

Ras proteins regulate cell fate by cycling between active GTP-bound and inactive GDP-bound conformations (Ras-GTP and Ras-GDP). RAS genes encode 4 proteins (N-Ras, H-Ras, K-Ras4A, and K-Ras4B) that have identical guanine nucleotide and effector binding domains but diverge substantially within the hypervariable region (HVR).^1,2  Prenylation of the C-terminal cysteine and palmitoylation of other cysteines within the HVR of H-Ras and N-Ras induce a dynamic cycle of depalmitoylation and repalmitoylation that regulates subcellular trafficking. By contrast, K-Ras4B localizes to the plasma membrane (PM) by a mechanism that does not involve palmitoylation.¹  Perturbation of palmitate turnover leads to a nonspecific distribution of H- and N-Ras to endomembranes and decreases signaling from the PM.³  This observation suggests that interfering with depalmitoylation might selectively reduce the growth of cancer cells with NRAS mutations, as normal K-Ras4B function would be preserved. Inhibiting oncogenic N-Ras signaling is particularly relevant in hematologic malignancies where NRAS is mutated more frequently than KRAS.^4-9 

Acyl protein thioesterase 1 (APT1) catalyzes Ras depalmitoylation.¹⁰  Dekker et al¹⁰  recently developed palmostatin B, a small molecule inhibitor with activity against APT1 that disrupted normal H-Ras subcellular localization and attenuated signaling in fibroblasts transformed with oncogenic Hras. Here we present genetic, cell biologic, and biochemical data with in vitro studies of palmostatin B that support targeting the depalmitoylation/repalmitoylation cycle in hematologic cancers characterized by oncogenic NRAS mutations.

Nras and Kras alleles containing an N terminal green fluorescent protein (GFP) marker were cloned into the murine stem cell virus (MSCV) vector with expression driven by the internal ribosomal entry site.^11,12  Retrovirally transduced E14.5 fetal liver cells from inbred C57Bl/6 mice were plated in methylcellulose medium to assess CFU-GM growth as described previously.^1,11,12  Biochemical analyses were performed on cultured macrophages that were differentiated from transduced GFP⁺ fetal liver cells in 50 ng/mL M-CSF.^11,12  Cells were analyzed by confocal microscopy 7 days after differentiation using the Zeiss LSM 510 NLO Meta. Mx1-Cre; LSL-Nras^G12D and Mx1-Cre; LSL-Kras^G12D mice were injected with a single dose of polyinosinic-polycytidylic acid to induce Kras/Nras expression as previously described.^13,14  Bone marrow cells from Mx1-Cre, LSL-Nras^G12D and Mx1-Cre, LSL-Kras^G12D mice with myeloproliferative disease (MPD) and acute myeloid leukemia (AML) were harvested, and colony assays were performed as described.^14-16  Mice with MPD were generated in the C57Bl/6 strain background, and the AML cells were generated by insertional mutagenesis in which C57Bl/6 × 129/Sv F1 mice were infected with the MOL4070LTR retrovirus.^14,15  Palmostatin B was synthesized and suspended as described.¹⁰ 

We generated MSCV vectors encoding N-terminal GFP fusions of wild-type (WT) N-Ras, N-Ras^G12D, N-Ras^{G12D, C181S}, and N-Ras^{G12D, SSDD}, which lacks a conserved SSDD (serine-serine-aspartic acid-aspartic acid) motif in the HVR domain that promotes stable palmitoylation.¹⁷  After transduction and sorting, CFU-GM growth from GFP⁺ mouse fetal liver cells was assessed over a range of GM-CSF concentrations. In all experiments, the growth of cells infected with an “empty” MSCV-internal ribosomal entry site–GFP vector was also assessed and was consistently similar to cells expressing WT N-Ras (supplemental Figure 1, available on the Blood Web site; see the Supplemental Materials link at the top of the online article). N-Ras^G12D expression induced cytokine-independent CFU-GM growth and conferred pronounced GM-CSF hypersensitivity (Figure 1A). Fetal liver cells expressing N-Ras^{G12D, SSDD} formed fewer colonies in the absence of GM-CSF and were less hypersensitive at low cytokine doses (Figure 1A). The C181S substitution abrogates the palmitoylation site in the N-Ras HVR. Cells expressing this mutant protein unexpectedly formed fewer CFU-GM colonies than control cells, suggesting dominant negative effects on hematopoietic growth (Figure 1A).

We next performed confocal microscopy on macrophages grown from GFP⁺ fetal liver cells. WT N-Ras and N-Ras^G12D showed similar distributions to the PM and perinuclear Golgi (Figure 1B). Loss of the SSDD motif reduced PM targeting and increased the intensity of perinuclear Golgi staining, whereas the nonpalmitoylated N-Ras^{G12D, C181S} protein showed a complete lack of PM localization (Figure 1B).

To analyze Ras signaling, cultured macrophages were deprived of M-CSF and serum for 16 hours followed by stimulation with GM-CSF. Cells expressing WT N-Ras displayed low basal levels of Ras-GTP and phosphorylated ERK (pERK), which increased on GM-CSF stimulation (Figure 1C). By contrast, N-Ras^G12D and N-Ras^{G12D, SSDD} expression resulted in elevated basal Ras-GTP and pERK levels, which increased further in response to GM-CSF. Although the N-Ras^{G12D, C181S} mutant protein also accumulated in the GTP-bound conformation, basal and cytokine-induced pERK levels were greatly attenuated. Moreover, total Ras levels were markedly lower in macrophages expressing N-Ras^{G12D, C181S} relative to the other 3 proteins (Figure 1C). These biochemical data are consistent with the adverse effects of N-Ras^{G12D, C181S} on CFU-GM growth (Figure 1A). To further address how N-Ras^{G12D, C181S} expression affects normal Ras signaling and progenitor growth, we generated the C181S mutation in WT N-Ras. This mutant did not alter CFU-GM growth (Figure 1D), which supports the hypothesis that GTP-bound N-Ras^{G12D, C181S} functions in a dominant negative manner by sequestering effectors.

To investigate the therapeutic potential of interfering with subcelluar N-Ras^G12D localization, we assessed the effects of palmostatin B¹⁰  on the growth of primary hematopoietic cells. We first assayed the effects of palmostatin B on CFU-GM growth from fetal liver cells transduced with GFP-N-Ras^G12D and K-Ras4B^G12D. Importantly, palmostatin B promoted a dose-dependent reduction in cytokine-independent CFU-GM growth that was selective for cells expressing GFP-N-Ras^G12D (Figure 2A). Similar effects were observed in the presence of a sub-saturating concentration of GM-CSF (0.1 ng/mL; Figure 2A). We next generated vectors in which the HVR domains of N-Ras^G12D and K-Ras4B^G12D were exchanged (GFP-N-Ras^{G12D, KHVR} and GFP-K-Ras4B^{G12D, NHVR}). Studies of progenitors expressing these fusion proteins confirmed the importance of the N-Ras HVR in conferring sensitivity to palmostatin B as cells expressing K-Ras4B^{G12D, NHVR} were highly responsive to the drug, whereas cells transduced with the GFP-N-Ras^{G12D, KHVR} were not (Figure 2B). Confocal microscopy showed that treatment with palmostatin B increased endomembrane localization of N-Ras^G12D and K-Ras4B^{G12D, NHVR}, whereas PM targeting of K-Ras4B^G12D and N-Ras^{G12D, KHVR} was unaffected (Figure 2C).

To evaluate the effects of palmostatin B on cells expressing oncogenic N- and K-Ras from their endogenous loci, we grew CFU-GM colonies from Mx1-Cre, LSL-Nras^G12D and Mx1-Cre, LSL-Kras^G12D mice. In contrast to Mx1-Cre, LSL-Kras^G12D/+ mice,¹³  heterozygous Nras mutant (Mx1-Cre, LSL-Nras^G12D/+) mice do not develop MPD during the first 6 months of life on an inbred C57Bl/6 strain background, and the bone marrow contains rare cytokine-independent CFU-GM.^14,18  However, consistent with a recent report,¹⁹  mice that are homozygous for the Nras^G12D allele developed overt MPD by 3 months of age (supplemental Figure 3), and bone marrow cells from these animals and from Mx1-Cre, LSL- Kras^G12D/+ mice formed large numbers of CFU-GM colonies in the absence of GM-CSF. Palmostatin B reduced cytokine-independent CFU-GM growth from heterozygous and homozygous Nras mutant bone marrow but had no effect on Kras mutant cells (Figure 2D). We also cultured cells from Mx1-Cre, LSL-Nras^G12D/+ and Mx1-Cre, LSL-Kras^G12D/+ mice that developed AML after retroviral mutagenesis.^14,15  The growth of blast colonies from 2 independent Nras leukemias was reduced by approximately 50% in response to 30μM palmostatin B. Importantly, this drug concentration has no effect on colony growth from a Kras mutant AML or from WT marrow (Figure 2E).

Cuiffo and Ren recently reported that mutating the palmitoylation site at cysteine 181 in N-Ras^G12D inhibited PM localization in fibroblasts and prevented the development of MPD in a transduction/transplantation model.²⁰  However, our observation that exogenous expression of the C181S substitution has dominant negative effects in primary progenitors precludes drawing therapeutic inferences from these data. We therefore expressed N-Ras^{G12D, SSDD} and demonstrated that this protein does not interfere with either normal Ras signaling or CFU-GM growth. Importantly, the SSDD mutation reduced N-Ras trafficking to the PM and attenuated the effects of oncogenic N-Ras^G12D on progenitor growth. Treatment with palmostatin B had similar effects, and studies in which we created N-Ras and K-Ras4B fusion proteins localized the inhibitory effects to the N-Ras HVR. Palmostatin B also showed selective efficacy in bone marrow cells from mice in which Nras^G12D is expressed from the endogenous locus.

Palmostatin B locks N-Ras and H-Ras in the palmitoylated form, which results in entropy-driven diffusion throughout the cell via membrane exchange processes.²¹  Developing palmitoyl transferase inhibitors is an alternative strategy for selectively inhibiting oncogenic N-Ras signaling; however, this is challenging because of the existence of more than 20 known family members.²²  Together, our data support investigating inhibitors of depalmitoylation in hematologic cancers characterized by oncogenic NRAS mutations.

This work was supported by the National Institutes of Health (grant R37 CA72614, and an American Recovery and Reinvestment Act supplement to this award; grant K08 CA134649), the Leukemia & Lymphoma Society (Specialized Center of Research award), the Frank A. Campini Foundation, and the American Cancer Research Society (Post-Doctoral Fellowship award, J.X.). K.S. is an American Cancer Society Research Professor.

National Institutes of Health

Contribution: J.X. designed experiments, performed research studies, analyzed data, and wrote the manuscript; F.J.D. and C.H. synthesized and characterized palmostatin B and assisted in writing and editing the manuscript; Q.L. generated and characterized Mx1-Cre, Nras^G12D knockin mice, generated the Kras and Nras mutant AMLs used in these studies, and assisted in writing and editing the manuscript; K.M.H. developed Nras^G12D knockin mice and assisted in writing and editing the manuscript; E.H. generated reagents and assisted J.X. with some of the research studies; H.W. designed and supervised the work leading to the synthesis and characterization of palmostatin B and assisted in writing and editing the manuscript; and K.S. designed experiments, reviewed the data, and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Kevin Shannon, Helen Diller Family Cancer Research Bldg, University of California–San Francisco, 1450 3rd St, Rm 240, San Francisco, CA 94158-9001; e-mail: shannonk@peds.ucsf.edu.