Kaiko Kunii

FGFR2-Amplified Gastric Cancer Cell Lines Require FGFR2 and Erbb3 Signaling for Growth and Survival

We have identified a critical role for amplified FGFR2 in gastric cancer cell proliferation and survival. In a panel of gastric cancer cell lines, fibroblast growth factor receptor 2 (FGFR2) was overexpressed and tyrosine phosphorylated selectively in FGFR2-amplified cell lines KatoIII, Snu16, and OCUM-2M. FGFR2 kinase inhibition by a specific small-molecule inhibitor resulted in selective and potent growth inhibition in FGFR2-amplified cell lines, resulting in growth arrest in KatoIII cells and prominent induction of apoptosis in both Snu16 and OCUM-2M cells. FGFR2-amplified cell lines also contained elevated phosphotyrosine in EGFR, Her2, and Erbb3, but the elevated phosphorylation in EGFR could not be inhibited by gefitinib or erlotinib. We show that the elevated EGFR, Her2, and Erbb3 phosphotyrosine is dependent on FGFR2, revealing EGFR family kinases to be downstream targets of amplified FGFR2. Moreover, shRNA to Erbb3 resulted in a loss of proliferation, confirming a functional role for the activated EGFR signaling pathway. These results reveal that both the FGFR2 and EGFR family signaling pathways are activated in FGFR2-amplified gastric cancer cell lines to drive cell proliferation and survival. Inhibitors of FGFR2 or Erbb3 signaling may have therapeutic efficacy in the subset of gastric cancers containing FGFR2 amplification.

MK-2461, a Novel Multitargeted Kinase Inhibitor, Preferentially Inhibits the Activated c-Met Receptor

The receptor tyrosine kinase c-Met is an attractive target for therapeutic blockade in cancer. Here, we describe MK-2461, a novel ATP-competitive multitargeted inhibitor of activated c-Met. MK-2461 inhibited in vitro phosphorylation of a peptide substrate recognized by wild-type or oncogenic c-Met kinases (N1100Y, Y1230C, Y1230H, Y1235D, and M1250T) with IC(50) values of 0.4 to 2.5 nmol/L. In contrast, MK-2461 was several hundredfold less potent as an inhibitor of c-Met autophosphorylation at the kinase activation loop. In tumor cells, MK-2461 effectively suppressed constitutive or ligand-induced phosphorylation of the juxtamembrane domain and COOH-terminal docking site of c-Met, and its downstream signaling to the phosphoinositide 3-kinase-AKT and Ras-extracellular signal-regulated kinase pathways, without inhibiting autophosphorylation of the c-Met activation loop. BIAcore studies indicated 6-fold tighter binding to c-Met when it was phosphorylated, suggesting that MK-2461 binds preferentially to activated c-Met. MK-2461 displayed significant inhibitory activities against fibroblast growth factor receptor (FGFR), platelet-derived growth factor receptor, and other receptor tyrosine kinases. In cell culture, MK-2461 inhibited hepatocyte growth factor/c-Met-dependent mitogenesis, migration, cell scatter, and tubulogenesis. Seven of 10 MK-2461-sensitive tumor cell lines identified from a large panel harbored genomic amplification of MET or FGFR2. In a murine xenograft model of c-Met-dependent gastric cancer, a well-tolerated oral regimen of MK-2461 administered at 100 mg/kg twice daily effectively suppressed c-Met signaling and tumor growth. Similarly, MK-2461 inhibited the growth of tumors formed by s.c. injection of mouse NIH-3T3 cells expressing oncogenic c-Met mutants. Taken together, our findings support further preclinical development of MK-2461 for cancer therapy.

Discovery of a 5H-benzo[4,5]cyclohepta[1,2-b]pyridin-5-one (MK-2461) inhibitor of c-Met kinase for the treatment of cancer.

c-Met is a transmembrane tyrosine kinase that mediates activation of several signaling pathways implicated in aggressive cancer phenotypes. In recent years, research into this area has highlighted c-Met as an attractive cancer drug target, triggering a number of approaches to disrupt aberrant c-Met signaling. Screening efforts identified a unique class of 5H-benzo[4,5]cyclohepta[1,2-b]pyridin-5-one kinase inhibitors, exemplified by 1. Subsequent SAR studies led to the development of 81 (MK-2461), a potent inhibitor of c-Met that was efficacious in preclinical animal models of tumor suppression. In addition, biochemical studies and X-ray analysis have revealed that this unique class of kinase inhibitors binds preferentially to the activated (phosphorylated) form of the kinase. This report details the development of 81 and provides a description of its unique biochemical properties.

PDK1 Attenuation Fails to Prevent Tumor Formation in PTEN-Deficient Transgenic Mouse Models

PDK1 activates AKT suggesting that PDK1 inhibition might suppress tumor development. However, while PDK1 has been investigated intensively as an oncology target, selective inhibitors suitable for in vivo studies have remained elusive. In this study we present the results of in vivo PDK1 inhibition through a universally applicable RNAi approach for functional drug target validation in oncogenic pathway contexts. This approach, which relies on doxycycline-inducible shRNA expression from the Rosa26 locus, is ideal for functional studies of genes like PDK1 where constitutive mouse models lead to strong developmental phenotypes or embryonic lethality. We achieved more than 90% PDK1 knockdown in vivo, a level sufficient to impact physiological functions resulting in hyperinsulinemia and hyperglycemia. This phenotype was reversible on PDK1 reexpression. Unexpectedly, long-term PDK1 knockdown revealed a lack of potent antitumor efficacy in 3 different mouse models of PTEN-deficient cancer. Thus, despite efficient PDK1 knockdown, inhibition of the PI3K pathway was marginal suggesting that PDK1 was not a rate limiting factor. Ex vivo analysis of pharmacological inhibitors revealed that AKT and mTOR inhibitors undergoing clinical development are more effective than PDK1 inhibitors at blocking activated PI3K pathway signaling. Taken together our findings weaken the widely held expectation that PDK1 represents an appealing oncology target.

Abstract 1185: H3B-8800, a novel orally available SF3b modulator, shows preclinical efficacy across spliceosome mutant cancers

Genomic characterization of hematologic and solid cancers has revealed recurrent somatic mutations affecting genes encoding the RNA splicing factors SF3B1, U2AF1, SRSF2 and ZRSR2. Recent data reveal that these mutations confer an alteration of function inducing aberrant splicing and rendering spliceosome mutant cells preferentially sensitive to splicing modulation compared with wildtype (WT) cells. Here we describe a novel orally bioavailable small molecule SF3B1 modulator identified through a medicinal chemistry effort aimed at optimizing compounds for preferential lethality in spliceosome mutant cells. H3B-8800 potently binds to WT or mutant SF3b complexes and modulates splicing in in vitro biochemical splicing assays and cellular pharmacodynamic assays. The selectivity of H3B-8800 was confirmed by observing lack of activity in cells expressing SF3B1R1074H, the SF3B1 mutation previously shown to confer resistance to other splicing modulators. Although H3B-8800 binds both WT and mutant SF3B1, it results in preferential lethality of cancer cells expressing SF3B1K700E, SRSF2P95H, or U2AF1S34F mutations compared to WT cells. In animals xenografted with SF3B1K700E knock-in leukemia K562 cells or mice transplanted with Srsf2P95H/MLL-AF9 mouse AML cells, oral H3B-8800 treatment demonstrated splicing modulation and inhibited tumor growth, while no therapeutic impact was seen in WT controls. These data were also evident in patient-derived xenografts (PDX) from patients with CMML where H3B-8800 resulted in a substantial reduction of leukemic burden only in SRSF2-mutant but not in WT CMML PDX models. Additionally, due to the high frequency of U2AF1 mutations in non-small cell lung cancer, H3B-8800 was tested in U2AF1S34F-mutant H441 lung cancer cells. Similar to the results from leukemia models, H3B-8800 demonstrated preferential lethality of U2AF1-mutant cells in vitro and in in vivo orthotopic xenografts at well tolerated doses. RNA-seq of isogenic K562 cells treated with H3B-8800 revealed dose-dependent inhibition of splicing. Although global inhibition of RNA splicing was not observed; H3B-8800 treatment led to preferential intron retention of transcripts with shorter and more GC-rich regions compared to those unaffected by drug. Interestingly, H3B-8800-retained introns commonly disrupted the expression of spliceosomal genes, suggesting that the preferential effect of H3B-8800 on spliceosome mutant cells is due to the dependency of these cells on expression of WT spliceosomal genes. These data identify a novel therapeutic approach with selective lethality in leukemias and lung cancers bearing a spliceosome mutation. Despite the essential nature of splicing, cancer cells without a spliceosome mutation were less sensitive to H3B-8800 compared with potent eradication of mutant counterparts. H3B-8800 is currently undergoing clinical evaluation in patients with MDS, AML, and CMML. Citation Format: Silvia Buonamici, Akihide Yoshimi, Michael Thomas, Michael Seiler, Betty Chan, Benjamin Caleb, Fred Csibi, Rachel Darman, Peter Fekkes, Craig Karr, Gregg Keaney, Amy Kim, Virginia Klimek, Pavan Kumar, Kaiko Kunii, Stanley Chun-Wei Lee, Xiang Liu, Crystal MacKenzie, Carol Meeske, Yoshiharu Mizui, Eric Padron, Eunice Park, Ermira Pazolli, Sudeep Prajapati, Nathalie Rioux, Justin Taylor, John Wang, Markus Warmuth, Huilan Yao, Lihua Yu, Ping Zhu, Omar Abdel-Wahab, Peter Smith. H3B-8800, a novel orally available SF3b modulator, shows preclinical efficacy across spliceosome mutant cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1185. doi:10.1158/1538-7445.AM2017-1185

Abstract B125: Mutant SF3B1 downregulates proteins involved in differentiation, including ABCB7

Refractory Anemia with Ringed Sideroblasts (RARS), a subtype of Myelodysplatic Syndrome (MDS), occurs with a high frequency of hotspot mutations in HEAT (Huntingtin, Elongation factor 3, protein phosphatase 2A, Targets of rapamycin 1 domains) domains of SF3B1. This protein component of the U2 snRNP complex of the spliceosome is essential in the proper selection and usage of 39 splice sites. RNAseq analysis of MDS and other tumor types in which SF3B1 hotspot mutations have been found show that alternative 39 splice site usage is the predominant cause of RNA transcript aberration. These modifications can result in mRNAs encoding novel peptides, or they can introduce premature termination codons into the pre-mRNA, most likely directing it to the Nonsense Mediated Decay (NMD) pathway for degradation. Using a predictive tool to determine the likelihood of a given aberrant transcript to be targeted for NMD, we determined that nearly 50% of the SF3B1-mutant-associated aberrant transcripts were candidates for degradation. We confirmed this experimentally by treating isogenic Nalm-6 cells (engineered by AAV homology to express SF3B1 K700E or K700K) with or without cycloheximide, an agent known to inhibit translation and RNA degradation by NMD. Investigation of the resulting RNAseq data showed significant rescue of gene expression only for the transcripts predicted to be NMD targets. Ingenuity Pathway Analysis indicated that many of the downregulated genes in SF3B1 mutant samples were involved in differentiation, which has been shown to be dysregulated in MDS. We tested the idea that such modifications in the transcriptome confer selective advantage or impair differentiation in SF3B1 mutant cells. We began by manipulating the expression of ABCB7, one of the genes identified in our RNAseq analysis to be downregulated by aberrant splicing and subsequent NMD. ABCB7 is a mitochondrial transporter important in cellular iron metabolism and, indirectly, in heme production. Additionally, loss of function of ABCB7 is causal in X-linked sideroblastic anemia and has been implicated in RARS MDS. We discovered in our SILAC proteomic analysis that ABCB7 protein was dramatically decreased in K700E SF3B1 Nalm-6 cells relative to K700K Nalm-6, in agreement with our RNAseq analysis. Using doxycycline-inducible shRNA expression, we knocked down ABCB7 mRNA and protein expression in TF-1 erythroblasts. These cells show significant decreases in erythropoeitin (EPO)-induced differentiation when expressing exogenous K700E SF3B1, but not K700R (a very conservative mutation) or WT SF3B1. With direct knock down of ABCB7, we observed a similar phenotype - impairment of EPO-induced differentiation in ABCB7 shRNA-induced cells by Day 7, with no overall decline in cell viability. Interestingly, knock down of SF3B1 expression with shRNA also reduces ABCB7 mRNA. However, it also promotes cell death. This is consistent with the heterozygous nature of SF3B1 hotspot mutations; severe loss of SF3B1 function is deleterious. We propose that hotspot SF3B1 mutants promote aberrant splicing of multiple genes, inducing a general “spliceosomal sickness” in addition to downregulating key genes (e.g. ABCB7) responsible for erythroid differentiation impairment, such as that observed in RARS. Citation Format: Rachel B. Darman, Samantha A. Perino, Michael Seiler, Shouyong Peng, Jacob Feala, Peter Fekkes, Gregg F. Keaney, Kaiko Kunii, Linda Lee, Kian Huat Lim, Yoshiya Oda, Khin Myint, Esther A. Obeng, Ermira Pazolli, Eun Sun Park, John Yuan Wang, Markus Warmuth, Lihua Yu, Ping Zhu, Yoshiharu Mizui, Benjamin L. Ebert, Peter G. Smith, Silvia Buonamici. Mutant SF3B1 downregulates proteins involved in differentiation, including ABCB7. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr B125.

Abstract C8: Targeting MCL1-dependent cancers with SF3B splicing modulators

Myeloid cell leukemia 1 (MCL1) is a member of the BCL2 family of proteins governing the apoptosis pathway and is one of the most frequently amplified genes in cancer. MCL1 overexpression often results in dependence on MCL1 for survival and is linked to resistance to anticancer therapies. However, the development of direct MCL1 inhibitors has proven challenging and new modalities for targeting MCL1 are required. Alternative splicing of MCL1 converts the anti-apoptotic MCL1 long (MCL1L) isoform to the BH3-only MCL1 short (MCL1S) isoform, which has been reported to be pro-apoptotic. Thus, changing MCL1 isoform levels through modulation of RNA splicing may represent an attractive approach to targeting MCL1-amplified cancers. To this end, we tested a collection of small molecule SF3B modulators that impact RNA splicing on MCL1-dependent and MCL1-independent NSCLC cell lines. SF3B modulators induced rapid downregulation of the long form and upregulation of the short- and intron-containing form of MCL1 across models; however, apoptosis was only observed in MCL1-dependent cells. Importantly, SF3B modulators preferentially killed MCL1-dependent cell lines and sensitivity correlated with MCL1 amplification. To dissect the mechanism of SF3B modulator-induced cytotoxicity, we overexpressed either the cDNA for the BH3-only short isoform or the full length isoform of MCL1. Surprisingly, overexpression of MCL1S cDNA had no significant effect on cells by itself and did not sensitize cells to SF3B modulator cytotoxicity. Conversely, MCL1L-specific shRNA knockdown was sufficient to kill MCL1-dependent cells and SF3B modulator cytotoxicity was rescued by expression of MCL1L cDNA. Together, these results argue that MCL1L modulation and not MCL1S upregulation is the effector of SF3B modulator cytotoxicity. In immunocompromised mice bearing MCL1-dependent xenograft models, SF3B1 modulator treatment resulted in significant downregulation of MCL1 levels accompanied by induction of apoptosis and robust efficacy at well-tolerated doses. Moreover, MCL1L cDNA expression in MCL1-dependent models rescued apoptosis induced by SF3B1 modulator treatment. These studies provide proof-of-concept that splicing modulation is an effective strategy for targeting cancers dependent on MCL1. Citation Format: Daniel Aird, Ermira Pazolli, Craig Furman, Linda Lee, Kaiko Kunii, Eun Sun Park, Craig Karr, Betty Chan, Michelle Aicher, Silvia Buonamici, John Yuan Wang, Jacob Feala, Lihua Yu, Markus Warmuth, Peter Smith, Peter Fekkes, Ping Zhu, Baudouin Gerard, Yoshiharu Mizui, Laura Corson. Targeting MCL1-dependent cancers with SF3B splicing modulators. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr C8.

Abstract 2941: Targeting MCL1-dependent cancers through RNA splicing modulation

Proceedings: AACR 106th Annual Meeting 2015; April 18-22, 2015; Philadelphia, PA Myeloid cell leukemia 1 (MCL1) is a member of the BCL2-family of proteins governing the apoptosis pathway and is one of the most frequently amplified genes in cancer. MCL1 overexpression often results in dependence on MCL1 for survival and is linked to resistance to anticancer therapies. However, the development of direct MCL1 inhibitors has proven challenging and thus far has been unsuccessful. Alternative splicing of MCL1 converts the anti-apoptotic MCL1 long (MCL1-L) isoform to the BH3-only containing MCL1 short (MCL1-S) isoform. As a potential approach for targeting MCL1-dependent cancers, we explored the use of MCL1 splicing modulators. We screened a unique chemical library of compounds that span a range of splicing activities on various substrates in an in vitro assay. Interestingly, we found a subset of general splicing modulators, as well as a subset of SF3B1 inhibitors, that are capable of driving the distinctive alterations in MCL1 splicing that in turn can trigger preferential killing of MCL1-dependent cell lines. The best modulators induce a prominent down-regulation of MCL1-L, up-regulation of MCL1-S, and accumulation of intron-retained MCL1 transcript. Somewhat surprisingly, several additional avenues of investigation pointed to MCL1-L down-regulation rather than MCL1-S up-regulation as the driver of preferential killing of MCL1-dependent cells. This includes the fact that compound-induced cytotoxicity can be rescued by expression of a MCL1-L cDNA and MCL1-L specific shRNA knockdown is sufficient to kill MCL1-dependent cells. On the other hand, overexpression of MCL1-S cDNA had no significant effect on cells and splicing modulators that induced very high levels of MCL1-S mRNA in the absence potent MCL1-L down-regulation exhibit minimal cytotoxicity. Biochemical characterization and understanding of these MCL1 splicing modulators has enabled further optimization of compounds that can induce potent and preferential killing of MCL1-dependent cancer cell lines in vitro. Preliminary studies in mice bearing MCL1-dependent NSCLC xenografts confirmed current lead compounds can indeed induce rapid down-regulation of MCL1-L, induction of apoptosis, and antitumor activity. Collectively these data yield insight into mechanisms of MCL1 splicing modulation that can trigger acute apoptosis in MCL1-dependent cancers and provides support for the idea of using splicing modulators to target difficult-to-drug oncogenic drivers such as MCL1. Citation Format: Eun Sun Park, Michelle Aicher, Daniel Aird, Silvia Buonamici, Betty Chan, Cheryl Eifert, Peter Fekkes, Craig Furman, Baudouin Gerard, Craig Karr, Gregg Keaney, Kaiko Kunii, Linda Lee, Ermira Pazolli, Sudeep Prajapati, Takashi Satoh, Peter Smith, John Yuan Wang, Karen Wang, Markus Warmuth, Lihua Yu, Ping Zhu, Yoshiharu Mizui, Laura B. Corson. Targeting MCL1-dependent cancers through RNA splicing modulation. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2941. doi:10.1158/1538-7445.AM2015-2941

Abstract 2932: SF3B1 mutations induce aberrant mRNA splicing in cancer and confer sensitivity to spliceosome inhibition

Proceedings: AACR Annual Meeting 2014; April 5-9, 2014; San Diego, CA Recurrent heterozygous mutations of the spliceosome protein SF3B1 have been identified in myelodysplastic syndromes, chronic lymphocytic leukemia (CLL), breast, pancreatic and skin cancers. SF3B1 is a component of the U2 snRNP complex which binds to the pre-mRNA branch point site and is involved in recognition and stabilization of the spliceosome at the 3′ splice site. To understand the impact of SF3B1 mutations, we compared RNAseq profiles from tumor samples with SF3B1 hotspot mutations (SF3B1-MUT) or wild-type SF3B1 (SF3B1-WT) in breast cancer, melanoma and CLL. This analysis revealed significant increases in the usage of novel alternative splice junctions in SF3B1-MUT samples including selection of alternative 3′ splice sites and less frequently exon skipping. These events induce expression of alternative mRNAs that are translated into novel proteins or aberrant mRNAs that are decayed by cells. A common alternative splicing profile was shared across different hotspot mutations and lineages (e.g. ZDHHC16 and COASY); however, unique alternative splicing profiles were also observed suggesting lineage specific effects. RNAseq analysis of several cell lines with endogenous SF3B1 hotspot mutations confirmed the presence of the same spliced isoforms as observed in tumor samples. To prove that SF3B1-MUT were inducing alternative splicing, transient transfection of several SF3B1 hotspot mutations in 293FT cells induced the expression of the common alternatively spliced genes suggesting functional similarity. Selective shRNA depletion of mutant SF3B1 allele in SF3B1-MUT cells resulted in downregulation of the same splice isoforms. Furthermore, isogenic B-cell lines (NALM-6) expressing the most frequent SF3B1 mutation (K700E) were generated and profiled by RNAseq. As expected, similar alternatively spliced genes were observed in NALM-6 SF3B1-K700E cells exclusively. To investigate the role of nonsense-mediated mRNA decay (NMD) in eliminating aberrant mRNAs induced by SF3B1-MUT, we treated NALM-6 SF3B1-K700E cells with cycloheximide, a translation inhibitor known to inhibit NMD. In the treated samples, expression of several aberrant mRNAs was revealed and some of these transcripts were shown to be downregulated in patient samples. Taken together, these results confirm the association between different SF3B1 hotspot mutations and the presence of novel splice isoforms. We demonstrated that E7107, a potent and selective inhibitor of wild-type SF3B1, also binds and inhibits SF3B1-MUT protein. In addition, E7107 represses the expression of several common aberrant splice mRNA products in SF3B1-MUT cells in vitro and in vivo. When tested in a NALM-6 mouse model, E7107 induced tumor regression and increased the overall survival of animals implanted with NALM-6 SF3B1-K700E cells. These data suggest splicing inhibitors as a promising therapeutic approach for cancer patients carrying SF3B1 mutations. Citation Format: Silvia Buonamici, Kian Huat Lim, Jacob Feala, Eunice Park, Laura Corson, Michelle Aicher, Daniel Aird, Betty Chan, Erik Corcoran, Rachel Darman, Peter Fekkes, Gregg Keaney, Pavan Kumar, Kaiko Kunii, Linda Lee, Xiaoling Puyang, Jose Rodrigues, Anand Selvaraj, Michael Thomas, John Wang, Markus Warmuth, Lihua Yu, Ping Zhu, Peter Smith, Yoshiharu Mizui. SF3B1 mutations induce aberrant mRNA splicing in cancer and confer sensitivity to spliceosome inhibition. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 2932. doi:10.1158/1538-7445.AM2014-2932