James Watters

Kif1C, a kinesin-like motor protein, mediates mouse macrophage resistance to anthrax lethal factor.

BACKGROUND Inbred mouse strains exhibit striking differences in the susceptibility of their macrophages to the effects of anthrax lethal toxin (LeTx). Previous data has shown that this difference in susceptibility lies downstream of toxin entry into macrophages. A locus controlling this phenotype, called Ltxs1, has been mapped to chromosome 11, but the responsible gene has not been identified. RESULTS Here, we report the identification of the Ltxs1 gene as Kif1C, which encodes a kinesin-like motor protein of the UNC104 subfamily. Kif1C is the only gene in the Ltxs1 interval exhibiting polymorphisms between susceptible and resistant strains. Multiple alleles of Kif1C determine the susceptibility or resistance of cultured mouse macrophages to LeTx. Treatment of resistant macrophages with brefeldin-A (which alters the cellular localization of Kif1C) induces susceptibility to LeTx, while ectopic expression of a resistance allele of Kif1C in susceptible macrophages causes a 4-fold increase in the number of cells surviving LeTx treatment. We also show that cleavage of map kinase kinase 3, a target of LeTx proteolysis, occurs in resistant cells. CONCLUSIONS We conclude that mutations in Kif1C are responsible for the differences in the susceptibility of inbred mouse macrophages to LeTx and that proper Kif1C function is required for LeTx resistance. Since the LeTx-mediated proteolysis of map kinase kinase 3 occurs even in resistant cells, Kif1C does not affect cellular entry or processing of LeTx and likely influences events occurring later in the intoxication pathway.

Ltx1, a mouse locus that influences the susceptibility of macrophages to cytolysis caused by intoxication with Bacillus anthracis lethal factor, maps to chromosome 11

The lethal factor (LF) toxin that is produced by Bacillus anthracis plays an important role in the pathogenesis of anthrax. LF has mononuclear phagocyte‐specific intoxicating effects that are not well understood. We have identified genetic differences in inbred mouse strains that determine whether their cultured macrophages are susceptible to the cytolytic effect of LF intoxication. Our identification of resistant and susceptible mouse strains enabled us to analyse crosses between these strains and to map a single responsible gene (called Ltx1 ) to chromosome 11. Ltx1 probably influences intoxication events that occur after the delivery of LF to the cytosol, as all mouse macrophages are killed by polypeptides containing the catalytic domain of Diphtheria toxin fused to the domain of LF required for cytosolic transport. Furthermore, the susceptibility phenotype is dominant to resistance, suggesting that resistance is caused by an absence of or polymorphism in a molecule that acts jointly with, or downstream of, the activity of LF. Our mapping of Ltx1 is a crucial first step in its positional cloning, which will provide more information about the mechanism of LF intoxication.

De novo Discovery of a γ-Secretase Inhibitor Response Signature Using a Novel In vivo Breast Tumor Model

Notch pathway signaling plays a fundamental role in normal biological processes and is frequently deregulated in many cancers. Although several hypotheses regarding cancer subpopulations most likely to respond to therapies targeting the Notch pathway have been proposed, clinical utility of these predictive markers has not been shown. To understand the molecular basis of gamma-secretase inhibitor (GSI) sensitivity in breast cancer, we undertook an unbiased, de novo responder identification study using a novel genetically engineered in vivo breast cancer model. We show that tumors arising from this model are heterogeneous on the levels of gene expression, histopathology, growth rate, expression of Notch pathway markers, and response to GSI treatment. In addition, GSI treatment of this model was associated with inhibition of Hes1 and proliferation markers, indicating that GSI treatment inhibits Notch signaling. We then identified a pretreatment gene expression signature comprising 768 genes that is significantly associated with in vivo GSI efficacy across 99 tumor lines. Pathway analysis showed that the GSI responder signature is enriched for Notch pathway components and inflammation/immune-related genes. These data show the power of this novel in vivo model system for the discovery of biomarkers predictive of response to targeted therapies, and provide a basis for the identification of human breast cancers most likely to be sensitive to GSI treatment.

Abstract 3436: Integrated profiling of p53 wild-type cell lines identifies differentially responsive populations and a gene expression signature that predicts sensitivity to SAR405838, a potent and selective disruptor of the p53-MDM2 interaction.

The tumor suppressor protein p53 functions at the center of a complex biological network that translates diverse stress signals into cell cycle arrest or apoptosis. MDM2 inhibits p53 activity by binding to the transactivation domain of p53 and by acting as a p53-specific E3 ubiquitin ligase. The disruption of the interaction between p53 and MDM2 with small molecular mass molecules has demonstrated encouraging single agent anti-tumor activity in preclinical in vitro and in vivo settings. While p53 wild-type status is necessary for response to antagonists of the p53-MDM2 protein-protein interaction (PPI), there is diversity in response to such therapeutic agents across p53 wild-type tumor models. As such, the ability to predict p53 wild-type tumor responsiveness to antagonists of the p53-MDM2 PPI is an unmet need that could significantly impact clinical development of modulators of the p53-MDM2 PPI. Here, we explored the effect of SAR405838, a selective p53-MDM2 antagonist, on a panel of 65 p53 wild-type cell lines derived from different tumors of origin. Over 80% of these cell lines showed some degree of sensitivity to SAR405838 when used as a single agent. Three broad response classes were observed: non-responders, responders involving apoptosis, and responders not involving apoptosis. A small number of tissue types were consistently enriched among responders (melanoma and leukemia) and non-responders (cervical and uterine tumors). By comparing pre-treatment gene expression profiling data with the single agent sensitivity to SAR405838, we identified a signature of 39 genes that clearly predicts non-response vs response. Within these 39 genes, CDKN2A was upregulated in non-responders, and MDM2 was upregulated in responders. Gene set enrichment analysis and GO category enrichments were consistent with predicted sensitivity in melanoma and leukemia, and indicate that a gene expression signature of loss of p53 pathway function predicts lack of single agent activity of SAR405838. 3-fold cross-validation of this classifier showed an accuracy of 94%, and principle component analyses was used to down-select the signature to the smallest number of genes carrying the maximum information and derive the optimal cut-off for making prospective response predictions. These data indicate that the development of p53-MDM2 PPI modulators could be targeted to a subset of p53 wild-type cancer patients and eliminate more predicted non-responders than those harbouring somatic mutations in p53. Citation Format: James Watters, Steve Rowley, Laurent Debussche. Integrated profiling of p53 wild-type cell lines identifies differentially responsive populations and a gene expression signature that predicts sensitivity to SAR405838, a potent and selective disruptor of the p53-MDM2 interaction. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3436. doi:10.1158/1538-7445.AM2013-3436

Abstract 3350: Plucked hair as a biomarker platform for monitoring transcriptional consequences of clinical exposure to antagonism of the HDM2/P53 interaction in tumors.

Proceedings: AACR 104th Annual Meeting 2013; Apr 6-10, 2013; Washington, DC The ability to assess and monitor target engagement is crucial for informing early drug development decisions. High vascularisation of the hair follicle, frequent epithelial origin of tumors, and high degree of congruence of expression in hair of pathways dysregulated in cancers, makes the cellular bulb on plucked human scalp hair an excellent surrogate tissue for non-invasive monitoring of PD effects in clinical trials. Disruption of the p53-HDM2 interaction with small molecules has demonstrated single agent anti-tumor activity in preclinical models and represents an attractive treatment strategy in oncology. Development of a peripheral tissue based gene expression signature of inhibition of the p53-HDM2 interaction could facilitate the early development of these compounds. To develop a peripheral PD biomarker of antagonism of this interaction, we used two selective p53-HDM2 antagonists, Nutlin-3 and Sanofis SAR405838 and applied our plucked hair biomarker platform to develop a gene expression signature indicative of compound exposure. SAR405838 displays potent activity in vitro and in vivo against p53 WT cell lines / xenograft models, but not in the p53 mutant context. Three hairs from each of four healthy donors were exposed to either SAR405838 or Nutlin-3 over a range of compound concentrations for 6hr or 24hr in our proprietary ex vivo cultures. We extracted RNA from the cellular bulb of individual hairs and assessed the transcriptome by microarray analysis. Biological enrichment analysis of genes differentially expressed revealed strong correlation to activated P53 signaling pathways. Further analysis revealed a core set of congruent genes as candidate PD biomarkers of p53-HDM2 antagonism, one of which was MIC-1, a secreted plasma protein that demonstrates a strong PK/PD relationship in patients treated with p53-MDM2 antagonists in Phase 1 trials. In ex vivo plucked hair, we have demonstrated biologically relevant differential expression of a panel of transcriptional markers exhibiting common response to Nutlin-3 and SAR405838. These reflect compound mechanism of action, and can provide further evidence of target engagement in addition to plasma MIC-1 levels. While the temporal and kinetic relationship between gene expression changes, toxicity, and clinical efficacy remains to be determined, the genes identified in this study may be used to provide further MOA information in clinical settings to monitor PD responses in plucked scalp hair obtained from patients exposed to SAR405838. Citation Format: Gino Miele, Elliot Harrison, Tim Mefo, Jo Read, Lydia Meyer Turkson, Alan Murdoch, Michael Teufel, Laurent Debussche, Donald Bergstrom, James Watters. Plucked hair as a biomarker platform for monitoring transcriptional consequences of clinical exposure to antagonism of the HDM2/P53 interaction in tumors. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3350. doi:10.1158/1538-7445.AM2013-3350

Abstract 741: Combined inhibition of MDM2 and MEK for KRAS mutant non-small cell lung cancer

There are currently no effective targeted therapies for KRAS mutant NSCLC. Combination drug strategies that target downstream pathways activated by mutant KRAS are currently being investigated, however efficacy may be limited by inability to induce a sufficient apoptotic response. MDM-2 (murine double minute 2) is a ubiquitin ligase that negatively regulates p53 function by promoting its degradation. Inhibitors of the MDM2-p53 interaction lead to p53 activation and transcription of genes promoting cell cycle arrest and apoptosis. We investigated the efficacy of combined MEK and MDM2 inhibition for KRAS mutant NSCLC. In KRAS mutant NSCLC cell lines with wild-type p53, pimasertib (MEKi) and SAR405838 (MDM-2i) led to upregulation of BIM and PUMA respectively. Induction of apoptosis and cell cycle arrest correlated with in vitro efficacy, and the combination exhibited marked in vivo activity against KRAS mutant p53 wild-type NSCLC xenograft tumors. These studies provide rationale for future clinical investigation of combined MEK and MDM-2 inhibition for KRAS mutant NSCLC. Citation Format: Aaron N. Hata, Hannah L. Archibald, Maria D. Gomez-Caraballo, Laurent R. DeBussche, Sukhvinder Sidhu, James Watters, Jeffrey A. Engelman. Combined inhibition of MDM2 and MEK for KRAS mutant non-small cell lung cancer. [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 741. doi:10.1158/1538-7445.AM2014-741

Abstract 1799: Targeted murine double minute 2 (MDM2) inhibition results in dramatic tumor regression in an MDM2-amplified glioblastoma multiforme (GBM) xenograft model

Proceedings: AACR Annual Meeting 2014; April 5-9, 2014; San Diego, CA Disruption of the MDM2/p53 pathway is an oncogenic driver alteration observed in the vast majority of GBM. This study evaluated the efficacy of the small molecule, human-specific, MDM2 inhibitor SAR405838 across a panel of GBM xenografts with distinct alterations within the MDM2/p53 pathway. Through screening a panel of 56 GBM xenografts with a combination of array CGH and MDM2-specific PCR amplification, five xenograft models were identified with high-level MDM2 amplification (GBM 5, 102, 108, 122, and 143). MDM2 transcript expression was elevated in all of these lines except GBM102, which had expression levels similar to non-amplified lines. The effects of SAR405838 were studied in 2 of the MDM2 amplified lines (GBM102, GBM108), and non-amplified lines with wild-type p53 (GBM10) or mutant p53 (GBM12). Using a combination of neurosphere and Cyquant assays to assess in vitro sensitivity, 100 nM SAR405838 completely inhibited growth of GBM108 (6% +/- 1% absorbance relative to control), partially inhibited GBM102 neurosphere formation (34 +/- 17% relative to control), and had nominal impact on relative neurosphere formation in GBM10 (51% +/- 4% relative to control) or GBM12 (96% +/- 4% relative to control). Twenty-four hours after treatment with 100 nM SAR405838, transcript expression of p53 target genes p21 and PUMA were elevated by 14 and 12 fold relative to control in GBM108, while more limited changes in expression were observed in GBM102 (4 and 2.5 fold increase, respectively). Similarly, analysis of apoptosis after 72h by Annexin-V staining demonstrated a 5-fold increase in apoptotic fraction from control in GBM108 compared to no increase in GBM102. A more striking difference in efficacy was observed in flank tumor xenograft studies where the time to reach a median tumor volume of 1000 mm3 was marginally extended in GBM102 (36 days for placebo vs. 44 days for SAR405838). In contrast, GBM108 placebo-treated mice grew to a median volume of 1000 mm3 in 53 days while treatment with SAR405838 resulted in complete tumor regression at 70 days. SAR405838 is a known substrate for the P-glycoprotein efflux pump that functions in the blood brain barrier to exclude xenobiotics from the central nervous system. Consistent with limited penetration of SAR405838 into the brain, a parallel study in an intracranial xenograft model for GBM108 demonstrated no significant survival benefit with SAR405838 treatment compared to placebo. Thus, although SAR405838 provides marked efficacy in vitro and in a flank tumor model of an MDM2 amplified tumor line, the lack of efficacy in an orthotopic model of this study suggests that more prudent investigation is needed before advancing the development of this specific drug in human tumors growing in the brain. Citation Format: Ann C. Mladek, Isabelle Meaux, Katrina Bakken, Pascal Pannier, Cedric Barriere, James W. Watters, Laurent R.J. Debussche, Jann N. Sarkaria. Targeted murine double minute 2 (MDM2) inhibition results in dramatic tumor regression in an MDM2-amplified glioblastoma multiforme (GBM) xenograft model. [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 1799. doi:10.1158/1538-7445.AM2014-1799

Abstract 3135: In vitro and in vivo synthetic lethal screens to identify novel targets in the context of PTEN deficiency.

Proceedings: AACR 104th Annual Meeting 2013; Apr 6-10, 2013; Washington, DC Functional genomic screens have been employed by many groups to identify novel targets for cancer therapeutics. However, translation of these data sets into new drug discovery programs has proved challenging, in part due to difficulties in obtaining strong on-target knockdown, coupled with misleading off-target effects. Most commercially available pooled shRNA screening libraries provide 5-6 shRNAs per gene, with little knowledge of knock-down efficiency. In an effort to overcome these issues, we designed a custom shRNA library targeting 580 human genes, including key nodes in signal transduction pathways and genes of interest for oncology drug development. This pathway-centric design allows us to determine critical pathways for cancer cell growth in addition to individual gene phenotypes. For each gene, we have total of 11 shRNAs, a subset of which have been previously shown high efficiency knock-down. In particular, we emphasized genes where tool or lead compounds were available for rapid follow-up. Both in size and scope, this library was designed to quickly identify important signaling pathways that are essential on their own or when perturbed in the presence of small molecule drugs in our pipeline. With this focused library, we performed functional genomic shRNA screens in PTEN mutant or null lines across cancer types in vitro, including four cell lines that are PTEN deficient (UACC62, MDA-MB-468, U87-MG and PC3), and one cell line, (HCT-116) that is PTEN wild type. Meanwhile, we performed in vivo screens with PC3 cells. To identify hits from our screens, we used an internally developed algorithm, c-FOLD, to calculate p-value and fold changes for each shRNA. As confirmation that our library and methods can identify essential genes in a context specific manner, we observe BRAF as the top essential gene hit in the cell line which contains an activated BRAF allele, UACC62, but not in other lines tested. We then examined the concordance of hits in our in vitro and in vivo experiments with PC3 cells. In general, individual shRNAs were highly correlated between in vitro and in vivo assays, but a subset is specific to one setting, and may present particularly interesting targets. To discover synthetic lethal partners to PTEN deficiency, we required that a gene is essential in all four PTEN deficient lines in vitro, as well as in PC3 in vivo, but not essential in HCT-116. Among the top hits are multiple shRNAs against ENTPD5, an ER enzyme hydrolyzing UDP to UMP. In addition, using our pathway approach, we identified components in JNK pathways as synthetic lethal partners to PTEN deficiency. The results obtained from these studies established a strong platform for in vivo and in vitro identification of novel targets and novel combination partners. These approaches can also be integrated with other Omics data, such as mutation and expression, to ultimately find promising targets to develop novel cancer therapeutics. Citation Format: Jing-xin Zhang, Joshua Murtie, Oleg Iartchouk, Hui Cao, Gary Shapiro, Zhifang Li, Hongyun Wang, Zhihu Ding, Yu-an Zhang, Madelyn Light, Dietmar Hoffman, May Cindhuchao, Saurin Jani, Richard Newcombe, Eva Bric-Furlong, Bin Wu, Angela Virone-Oddos, Stephan Reiling, Joachim Theihaber, Christoph Lengauer, Jack Pollard, James Watters, Serena Silver, Venkat Reddy. In vitro and in vivo synthetic lethal screens to identify novel targets in the context of PTEN deficiency. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3135. doi:10.1158/1538-7445.AM2013-3135

Abstract 4648: In vitro characterization of spiro-oxindole-based modulators of the MDM2-p53 interaction and their interspecies selectivity.

We have identified two spiro-oxindole-based modulators -compounds A and B- of the MDM2-p53 protein-protein interaction that display suitable pharmacological properties for human clinical testing (1) and a high affinity to the N-terminal domain of human MDM2. For instance, in fluorescence polarization (FP) displacement assay using N-terminal domain of human MDM2 and a p53-based peptide (2), compounds A and B display Ki values of 5 and 5,9 nM, respectively, while Nutlin 3 shows a Ki value of 80,3 nM. The superior binding activity of the spiro-oxindole derivatives was confirmed by isothermal titration calorimetry (ITC) and surface plasmon resonance (SPR). Having established their binding activities for the human target and in order to properly address potential on-target adverse effects in preclinical settings, we cloned MDM2 N-terminal domains of monkey, dog, rat and mouse. Sequencing analysis showed that, while the rat data from NCBI Reference Sequence database (3) is inaccurate, there is a high identity between human and monkey sequences. FP displacement assays were developed using MDM2 N-terminal domains of dog, rat and mouse which matched above-mentioned N-terminal domain of human MDM2. Ki value determinations revealed that the binding affinities of compounds A and B but not of Nutlin 3 significantly shifted to higher values by using both rodent MDM2 while, by using canine MDM2, they were similar to the ones determined with human MDM2. These findings were unexpected since MDM2 N-terminal domain sequences are conserved between rodent and non-rodent species. Cellular p53 activation data obtained with Nutlin 3, compounds A and B in a panel of monkey, dog, rat and mouse cell lines were consistent with the biochemical findings and confirmed that compounds A and B are modest p53 activators in rodent cells. In line with the results obtained in cellular settings, acute in vivo induction of p53-dependent transactivation by compounds A and B was significantly weaker in mouse spleen than in human xenografted tumor samples. In conclusion, these data demonstrate that compounds A and B both display high affinity against human MDM2 and that modulators of the MDM2-p53 protein-protein interaction can exhibit inter-species selectivity. These findings should be taken into consideration in the assessment of the tolerabily of MDM2-p53 antagonists in preclinical settings. 1. Wang S. et al., AACR, Orlando FL, 2011, Abstract LB-204 2. Ding K. et al, J. Am. Chem. Soc. 127, 10130-1, 2005. 3. http://www.ncbi.nlm.nih.gov/protein/NP_001101569.1 Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4648. doi:1538-7445.AM2012-4648

Abstract 4731: Cytotoxic versus cytostatic effects in tumor models by spiro-oxindole-based modulators of the MDM2-p53 interaction

Proceedings: AACR 103rd Annual Meeting 2012‐‐ Mar 31‐Apr 4, 2012; Chicago, IL Two spiro-oxindole-based modulators -compounds A and B - of the MDM2-p53 protein-protein interaction display suitable pharmacological properties and are currently under advanced preclinical development (1). In vitro, these compounds induce massive cell death in a subset of relevant tumor models such as MDM2-amplified SJSA-1, while they can significantly block cell proliferation without detectable cytotoxic effect in other human cell lines (e.g. HCT116). To better characterize in vivo the different tumor cell sensitivity of these novel MDM2 antagonists, quantitative and longitudinal pharmacodynamic effects have been determined in several preclinical settings: i) anti-tumor efficacy, ii) activation of p53-dependent molecular biomarkers such as p53 itself, p21, MDM2, PUMA, Noxa, Bax and MIC-1 and iii) downstream biological effects monitored by immuno-assays as well as by in vivo fluorescence tomography and PET imaging. The results obtained from these studies have revealed striking differences in the biological read-outs associated with the observed cytotoxic versus cytostatic effects and provide guidance for the use of certain biomarkers in the clinical development of this new generation of anti-cancer agents. 1. Wang S. et al., AACR, Orlando FL, 2011, Abstract LB-204 Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4731. doi:1538-7445.AM2012-4731