Search

The ubiquitin proteasome pathway comprises a coordinated, dynamic cellular system critical to cellular metabolism, signaling and proliferation. The expanding clinical utility of the peptide boronate, bortezomib, in the treatment of patients with multiple myeloma and other hematologic malignancies has established the human 26S proteasome as a validated target in cancer. Still, only one FDA-approved proteasome inhibitor presently exists. Restricted activity against one enzymatic function of the proteasome and dose-limiting toxicities associated with bortezomib warrant further discovery efforts aimed at the identification of structurally and functionally distinct protein degradation inhibitors (PDIs). Here, we report a novel family of natural product proteasome inhibitors discovered by high-throughput, high-content screening at the National Cancer Institute Initiative for Chemical Genetics. A primary screen of 14,000 small molecules was performed in 384-well plate format using a cell line stably transfected with a destabilized fluorescent protein chimera. Assay positives were retested in the primary screen in dose-response format. Thiostrepton was selected for further characterization due to its unique macrocyclic chemical structure, the recent publication of its total synthesis, reports of anticancer properties and the lack of prior annotation as a PDI. First, thiostrepton was linked to previously characterized molecules acting on the protein degradation pathways by transcriptional small molecule connectivity mapping (CMAP). Subsequent cell-state analyses confirmed strong induction of functional and annotated gene sets associated with misfolded protein stress and proteasome inhibition. Mechanism of action was confirmed by biochemical profiling of human 20S proteasome active site inhibition and specificity using homogeneous assays and selective substrates for each of three catalytic active sites. Importantly, inhibitory activity of thiostrepton differs from bortezomib by blocking both the chymotryptic-like and PGPH active sites with sub-micromolar potencies. Dose-dependent inhibition of multiple myeloma cell growth was observed, with a concomitant increase in polyubiquitinated protein stress and induction of apoptosis. Inhibition of conferred proliferation by bone marrow stroma was confirmed using a novel miniaturized high-content assay modeling the bone marrow stroma-multiple myeloma microenvironment. Structurally related compounds to thiostrepton, nosiheptide and siomycin, were confirmed also as proteasome inhibitors as above. Our discovery of this class of natural products as proteasome inhibitors and a recent report of siomycin inhibition of Sonic Hedgehog (Shh) signaling begged the question whether established proteasome inhibitors would inhibit Shh signaling in human cancer. This hypothesis was confirmed in a set of reporter-gene assays. In sum, these studies identify thiopeptide macrocycles as a class of naturally-occurring proteasome inhibitors poised for clinical development in hematologic malignancies, establish novel high-throughput assays for modeling MM-stroma microenvironment interactions and pave the way for the development of proteasome inhibitors in disease states where Shh signaling is central to pathogenesis.

Multiple myeloma (MM) is selectively sensitive to perturbations of protein catabolism. Genetic evidence has identified a link between plasma cell differentiation and upregulation of the unfolded protein response. The effective, selective, cytotoxic effect of proteasome inhibition observed in MM therapy further validates pathways of protein degradation as therapeutic targets in this incurable disease. Upon proteasome inhibition, cancer cells utilize compensatory pathways such as aggresome formation and macroautophagy. Despite a remarkable increase in the granularity of knowledge surrounding autophagy in mammalian cell biology, the role of autophagy in MM cell survival has not been studied. Lysosome-dependent macroautophagy (autophagy) pathway is known to be up-regulated by chemical inhibition of the proteasome. Therefore we theorized that autophagy is a pathway of proteasome inhibitor resistance in MM, and that simultaneous inhibition of both the ubiquitin-proteasome system (UPS) and the autophagy pathway would result in synergistic toxicity in MM cells. We chose to explore this hypothesis utilizing a chemical genetic approach, with the goal of repurposing off-patent autophagy inhibitors for immediate translation in the context of human clinical investigation. We selected for study the combination of the only FDA-approved proteasome inhibitor, bortezomib, and the 4-aminoquinoline autophagy inhibitor, hydroxychloroquine (HCQ). Indeed, HCQ exhibits a dose-dependent cytotoxic effect on cultured human MM cells in vitro. Using high-throughput, high-content epifluorescence microscopy, we have demonstrated a dose-dependent increase in autophagosome number and average vesicle area in MM cells treated with HCQ consistent with inhibition of late autophagy. These findings were confirmed by electron microscopy. Autophagy inhibition by HCQ was verified by LC3 immunoblot with and without bafilomycin clamp. Gold-standard pulse-chase studies of HCQ verified inhibition of long-lived protein degradation in MM cells. Of note, in vivo studies of HCQ in a bioluminescent orthotopic xenograft model of MM failed to demonstrate a statistically evident disease response or prolongation of survival. Drawing on the aforementioned biological and clinical insights, we predicted that combination autophagy and proteasome inhibition would result in synergistic cytotoxicity. To explore multiple combinations of these classes of therapeutic agents, we devised a high-throughput translational platform for assessing synergy between therapeutic agents. In this model, library plates of compounds are established which may be “crossed” with each other by alternating plate orientation and delivering compounds to experimental assay microtiter plates by robotic pin transfer. Experimental measurements (i.e. cytotoxicity, protein content, post-translational modifications) are then processed in a semi-automated manner employing the median effect principle and illustrative graphical outputs. This platform was validated by modeling known, effective combination therapies in current clinical use or study in MM. We observe potent, synergistic toxicity from the combination of bortezomib and HCQ in both MM1.S and RPMI-8826 MM cell lines. Of note, the combination index observed is

Background: Waldenström’s Macroglobulinemia (WM) is an incurable lymphoplasmacytic lymphoma with limited options of therapy. Histone deacetylase (HDAC) inhibitors represent promising new treatment strategy in B cell malignancies. We therefore investigated the in vitro effect of the novel hydroxamic acid derivative HDAC inhibitor LBH589 in WM. Methods: WM cell lines (BCWM1 and WSU-WM) and IgM secreting low-grade lymphoma cell lines (MEC1, RL) were used. Bone marrow primary CD19+ cells and bone marrow stromal cells (BMSC) were obtained from patients with WM after informed consent. Cytotoxicity and DNA synthesis were measured by MTS assay and thymidine uptake assay. Cell signaling and apoptotic pathways were determined by Western Blot and immunofluorescence. Results: LBH589 induced a significant decrease of proliferation and triggered cytotoxicity in all cell lines tested and primary CD19+ WM cells (IC50 of 20–40nM), even in the presence of BMSC, IL-6 and IGF-1, which induce resistance to conventional therapies. Importantly, LBH589 did not induce cytotoxicity in healthy donor peripheral blood mononuclear cells. LBH589 induced both intrinsic and extrinsic apoptotic pathways, with caspase-9, caspase-8, caspase-3, and PARP cleavage in a dose-dependent manner. We also demonstrated significant upregulation of the proapoptotic transcription factor p53 and down-regulation of the anti-apoptotic proteins BclxL, Mcl-1 and c-myc. We then demonstrated that LBH589 induced apoptosis in WM cells in a caspase-independent manner through induction of autophagy, as shown by upregulation of LC3B and Rab7 expression. We further determined the mechanism of action of LBH589 in WM, investigating the effect of LBH589 on histone acetylation and NF-kB pathways. We found that LBH589 induced a dose-dependent increase in histone H3-H4 acetylation; and inhibited both canonical and non-canonical pathways of NF-κB, as shown by western blot and immunofluorescence. In addition, LBH589 augmented rituximab, fludarabine, bortezomib, and perifosine-induced cyotoxicity in WM cells. Conclusion: LBH589 has significant antitumor activity in WM in vitro, providing the framework for clinical trials evaluating LBH589 as a new therapeutic agent in patients with WM.

Autophagy is an evolutionarily conserved process by which cellular structures may be degraded to support ongoing biogenesis. Recent evidence has identified the induction of autophagy following growth factor pathway inhibition, rapid tumor expansion and proteasome inhibition. Thus, autophagy is a highly desirable pathway for the development of targeted therapeutics in cancer, in particular in tumors characterized by misfolded protein stress such as multiple myeloma (MM). Unfortunately, there are only several known classes of autophagy inhibitors and the emerging biology of autophagy has identified few validated targets. We therefore performed a forward chemical genetic study of autophagy using novel techniques in high-content imaging, high-throughput screening and multidimensional data analysis. Using LN229 glioblastoma cells stably transfected with an LC3-EGFP construct, we created a robust screening platform capable of quantifying the number, size and fluorescence intensity of autophagosomes. Assay validation was performed with the known autophagy inhibitor chloroquine, which induces a characteristic shift from diffuse to punctate fluorescence. Initially, we profiled a library of approximately 3,500 structurally diverse bioactive small molecules to leverage possible downstream mechanistic insights. Assay positives and controls were re-tested in the adherent H1299 adenocarcinoma cell line likewise stably expressing LC3-EGFP. Known autophagy inhibitors were enriched with this approach: nocodazole, bafilomycin A1, colchicine and monensin. Structure-activity relationship analysis of the primary screening data and 34 novel autophagy modulators revealed a compelling activity of bis-indolyl maleimides. We confirmed sub-micromolar inhibition of autophagy by three such compounds using a quantitative, radiolabeled protein degradation assay (tritiated tyrosine pulse-chase), LC3-II immunoblotting and quantitative measurement of autophagosomes across three cell lines. From these studies we selected a lead compound, K252a, for further characterization and tested twenty-two closely related analogues. Detailed SAR analysis has identified features of the molecule required for inhibition of autophagy and sites permissive for further functionalization in ongoing target identification studies. Using MM as a translational model system of misfolded protein stress, we found dose-dependent, selective cytotoxicity in MM of K252a compared to analogues lacking autophagy inhibitory activity. We report here the results of a forward chemical genetic study of autophagy which identifies and characterizes novel small molecule inhibitors of autophagy and establishes a rationale for clinical studies of autophagy inhibitors in the treatment of MM.

Cutaneous T-Cell Lymphoma (CTCL) comprises a group of related lymphoproliferative disorders characterized by the presence of malignant lymphocytes in the skin. Most patients with CTCL enjoy a normal life expectancy, though all experience chronic morbidity due to the symptomatic and cosmetic manifestations of epidermotropism. There is no curative therapeutic option for patients with CTCL and consequently topical approaches have become a core component of standard care. Recently, activity against advanced CTCL has been demonstrated by a new class of targeted agents: histone deacetylase inhibitors (HDACi). One HDACi has been approved by the FDA for use in advanced disease (vorinostat; SAHA; Merck Research Laboratories, Rahway, NJ). Dose-limiting toxicities with parenteral HDAC inhibitors include fatigue, diarrhea, nausea and myelosuppression. The poor tolerability of the currently available, nonselective small molecules render most patients with this indolent disease unlikely to benefit from this pharmacologic class. We therefore conceived of reverse pro-drug HDAC inhibitors to allow high dose-intensity at the site of disease and presystemic metabolism to mitigate side effects and reduce exposure to genotoxic agents. Suberohydroxamic acid phenyl ester (SHAPE) is a first-in-class soft-drug HDACi. SHAPE is a soluble, scalable analog of SAHA possessing an isosteric, isoelectronic ester bond at a site predicted by molecular modeling not to impact target recognition features. Serum metabolism was confirmed using a real-time absorbance assay capable of monitoring SHAPE degradation to suberic acid hydroxamate and phenol. We have determined that the rapid hydrolysis of SHAPE in serum is due to the catalytic activities of butyrylcholinesterase and paraoxonase. Cellular inhibition of Class I and II HDACs was confirmed using high content imaging; effects on histone and tubulin acetylation comparable to SAHA are witnessed at concentrations above 20 micromolar. Active site binding to HDAC isoforms was determined using purified protein and a novel, miniaturized fluorescence polarization assay we designed to permit instantaneous kinetic measurements of labile compounds. Tolerability studies of topical SHAPE were performed on shaved immunocompetent mice and on immunodeficient animals bearing human skin xenografts. Histologic evidence of inflammation or damage was not visualized. We next performed experiments testing efficacy against CTCL in vivo. Using cell lines established from neoplastic, skin-homing lymphocytes arising from an interleukin-7 transgenic murine model of CTCL, we developed a transplantation model in syngeneic animals with rapid, homogeneous disease onset compatible with studies of investigational agents. Daily administration of SHAPE (50 mg/mL) for two weeks resulted in clearance of skin-homing lymphocytes, compared to vehicle treated controls. For use in murine model studies and anticipating human clinical trials, we identified acetylation response biomarkers using mass spectrometry and developed immunohistochemical protocols with acetylation state-specific antibodies. In sum, we report the design, characterization and translational studies supporting the clinical development of a novel, soft-drug HDAC inhibitor in CTCL.

Acute lymphoblastic leukemia (ALL) is the most common form of childhood cancer. Despite effective chemotherapy, 20% of patients will relapse. Therefore, it is critical that we identify novel therapies to treat ALL. We are studying a new small molecule compound known as tubacin (tubulin acetylation inducer) that selectively inhibits histone deacetylase 6 (HDAC6). HDAC6 binds to polyubiquitinated misfolded proteins and to dynein motor proteins, including alpha-tubulin, thereby recruiting misfolded or unwanted proteins to aggresomes and subsequent degradation by the lysosome. Tubacin was discovered through a chemical genetic screen of a 7,392 small molecule library and was found to induce acetylation of alpha-tubulin by inhibiting one of the two catalytic domains of HDAC6. This inhibition disrupts the interaction of HDAC6 with dynein resulting in marked accumulation of ubiquitinated proteins. Previous work demonstrated that treatment of multiple myeloma cells with tubacin inhibited growth at low micromolar concentrations. Tubacin does not appear to affect global histone acetylation, gene expression, or cell cycle regulation. To determine the effects of tubacin in human ALL cells, we treated both B- and T-cell ALL cell lines with varying concentrations of the drug and performed MTT assays. In T-ALL cells (Jurkat, Loucy), the IC50 of tubacin was found to be 1 to 3 uM, while in B-cell ALL cells (REH, Nalm-6), the IC50 was 2 to 5uM. Tubacin induced apoptosis of Jurkat and Loucy cells stained with Annexin V and propidium iodide. Within 12 hours, we observed increased protein polyubiquitination and PARP cleavage, but no difference in Rb phosphorylation in cells treated with tubacin. Furthermore, tubacin treatment increased acetylation of alpha-tubulin in Loucy and Jurkat cells within 3 hours. To study whether tubacin was toxic to normal hematopoietic cells, we treated human bone marrow cells cultured in methylcellulose containing IL-3, IL-6, and Stem Cell Factor with varying concentrations of tubacin. The IC50 of 20uM was determined from numbers of colonies plated in triplicate. Similarly, treatment of normal human lymphocytes cultured in IL-2, demonstrated an IC50 of 16uM. Finally, we examined the effects of tubacin on growth of primary ALL cells. Bone marrow cells from three patients with B-cell ALL at diagnosis were cultured in tubacin at varying concentrations and MTT assays were performed. In two of the three ALL samples, the IC50 was less than 5uM. Experiments to study the effects of tubacin in mouse models of leukemia are in progress. Our results suggest that inhibition of HDAC6 and the aggresome pathway provides a novel approach to treat ALL.

Notch signaling represents a central pathway regulating hematopoiesis, stem cell differentiation, and malignant transformation in human cancer. Activation of highly conserved Notch1 receptors results in cleavage and release of an intracellular domain (ICN1). Following translocation to the nucleus, ICN1 forms a ternary complex with the transcriptional repressor CSL (CBF-1, Suppressor of Hairless and Lag-1) bound to cognate DNA. This event triggers a repressor-to-activator switch, as an interfacial groove is formed which recruits the Mastermind-Like (MAML1) co-activator protein. Activating mutations in NOTCH1 are found in more than 50% of patients with T-Cell Acute Lymphoblastic Leukemia (T-ALL), promoting protein stability and establishing a direct link to disease pathogenesis. Pharmacologic efforts to target the Notch pathway in T-ALL have been directed at gamma secretase, a regulatory enzyme in Notch activation. Recently, the observed clinical resistance to gamma secretase inhibitors has been explained, in part, by additional mutations in the Notch-targeting ubiquitin ligase, Fbxw7, which further increases oncoprotein stability. Therefore, direct inhibitors of ICN1 function are highly desirable. Drawing upon insights afforded by the resolved crystal structure of the DNA-bound ICN1:MAML1:CSL complex, we synthesized a series of hydrocarbon stapled alpha-helical peptides targeting Notch (SAHNs) based on minimal motifs of the MAML protein predicted to engage the composite ICN1:CSL interface. Direct, high-affinity binding to purified components of the Notch complex was confirmed using surface plasmon resonance (SPR). Nuclear access of SAHN1 was confirmed using quantitative epifluorescent and confocal microscopy. Intracellular association with ICN1 and CSL was established using bidirectional affinity chromatography. Using a novel CSL-responsive reporter construct, we observed inhibition of endogenous Notch transactivation by SAHN1 in T-ALL cell lines. Furthermore, SAHN1 induces a dose-dependent knockdown of endogenous Notch1 target genes including HES1, HEY1 and cMYC in T-ALL cell lines. Remarkably, inhibition of Notch signaling by SAHN1 confers selective cytotoxicity at 48 hours in a panel of T-ALL cell lines with known mutations in NOTCH, including those resistant to gamma secretase inhibitors. Supporting an on-target mechanism of action, we have prepared a damaged analogue of SAHN1 containing a two-residue rearrangement (SAHN1D). SAHN1D possesses reduced binding affinity for the Notch complex and despite comparable intracellular access, SAHN1D lacks both transcriptional and cytotoxic effects on cultured T-ALL cell lines in vitro. Efficacy studies have also been performed in vivo using a novel murine model of T-ALL. In summary, we report here the design, biochemical characterization and translational rationale supporting the first direct inhibitor of the Notch transactivation complex in T-ALL.

Graft-versus-host disease (GVHD) is a serious and common complication of allogeneic hematopoetic stem cell transplantation (HSCT). GVHD results from genetic incompatibilities between donor and host. Although matching of donor and recipient human leukocyte antigens (HLA) is critical to transplant outcomes, GVHD is common even in transplants involving HLA-identical sibling donors. The additional determinants of graft compatibility are poorly understood, though a small number of minor histocompatibility antigens have been reported. The human genome has recently been found to show large-scale structural variation, including large gene deletions sufficiently common to appear as homozygous gene deletions in many patients. We investigated whether donor-recipient disparity for common gene deletion polymorphisms (homozygous gene deletion in donor but not in recipient) was associated with GVHD following allogeneic HSCT in 821 HLA-identical sibling donor-recipient pairs. Patient samples and clinical outcomes data were assembled collaboratively from institutions in Finland and North America, in strict compliance with ethical standards and IRB-approved protocols. Donors and patients were typed for the presence of six common gene deletions using real-time PCR. Donor-recipient disparities were then assessed for association with acute and chronic GVHD. Donor-recipient disparities in the expected direction (-/- in donor, +/- or +/+ in recipient) for two common gene deletions, UGT2B28 and UGT2B17, were associated with chronic GVHD (OR > 4, P=0.001; and OR=2.0, P=0.05 respectively). Donor-recipient disparity for UGT2B17 was also associated with acute GVHD (OR=2.0, P=0.05). In the cohorts assembled for this study, we also observed an expected robust association for sex mismatch with chronic GVHD (OR=1.6, p=0.001) and a non-significant trend toward association with acute GVHD (OR=1.1, p=0.30). No significant association with acute or chronic GVHD was observed for the other four common gene deletion polymorphisms (GSTT1, GSTM1, OR51A2, and LCE3C). Consistent with prior studies of minor histocompatibility antigens on the Y-chromosome, we observed evidence of serologic immunoreactivity to a nine residue peptide on UGT2B17 in an HSCT recipient with mismatch at this locus. Because disparities for UGT2B17 and UGT2B28 occur in fewer than 6% of sibling-donor transplants in populations with European ancestry, they can explain only a fraction of the genetically attributable risk of GVHD. Of note, deletion of UGT2B28 appears to be more common in individuals with African and Latin American ancestry (allele frequency of 25–35%), and therefore merits investigation as a potential contributor to transplant outcome in those populations. We conclude that variation in genome structure associates with graft-versus-host disease and merits further investigation as a cause of minor histocompatibility barriers to hematopoietic stem cell transplantation. For the Transplant Structural Genetics Consortium: David Altshuler, Shannon Chilewski, Steven McCarroll (Broad Institute of Harvard and MIT, Cambridge, MA); Joseph Antin, James Bradner, Stephanie Lee, Jerome Ritz, Robert Soiffer (Dana-Farber Cancer Institute, Boston, MA); Aarno Palotie (Finnish Genome Center, Helsinki, Finland); Jukka Partanen, Hannu Turpeinen (Finnish Red Cross, Helsinki, Finland); Tapani Ruutu, Liisa Volin (Helsinki University Central Hospital, Helsinki, Finland); David Ginsburg, David Siemeniak (University of Michigan, Ann Arbor, MI).

Histone deacetylase 6 (HDAC6) has an essential role to recruit ubiquitinated proteins to transport to aggresomes, which ultimately induces lysosomal protein degradation. We have shown that inhibition of proteasomes with bortezomib and of aggresomes with HDAC6 inhibitor Tubacin demonstrated significant cytotoxicity in MM cell lines and MM patient tumor cells in vitro (Hideshima T et al., PNAS2005, 102: 8597–8572). In this study, we further examined the biologic significance of HDAC6 inhibition by Tubacin in MM cells. We found that HDAC6 is constitutively associated with heat shock protein (Hsp) 90 in MM cell lines which is enhanced by Tubacin, as evidenced by co-immunoprecipitation. Since Akt and STAT3 have been shown to play important role in proliferation, anti-apoptosis, and drug resistance in MM cells; and all are client proteins of Hsp90, we next further examined whether inhibition of HDAC6 could modulate activities of these proteins via Hsp90. Importantly, Tubacin enhanced phosphorylation of Akt, associated with augmentation of Hsp90 acetylation. Hsp90 inhibitor 17-AAG downregulated Akt phosphorylation associated with enhanced interaction of Hsp90 with Akt, which was partially blocked by Tubacin. On the other hand, 17-AAG did not enhance acetylation of α-tubulin or ubiquitination of proteins, suggesting that Hsp90 does not affect HDAC6 function. Furthermore, we found that STAT3 is also constitutively associated with Hsp90. Importantly, both Tubacin and 17-AAG inhibit phosphorylation of STAT3 in a dose- and time-dependent fashion in MM cells. Taken together, our data indicate that HDAC6 has an important role not only in aggresomal protein degradation, but also in MM cell pathogenesis by modulating Akt and STAT3 signaling cascades via Hsp90 acetylation in MM cells.

Cutaneous T-Cell Lymphoma (CTCL) comprises a group of related lymphoproliferative disorders characterized by the presence of malignant lymphocytes in the skin. The most common variant is Mycosis Fungoides (MF), which affects approximately 500 patients per year in the United States. Though most patients with MF enjoy a normal life expectancy, they experience chronic morbidity due to the symptomatic and cosmetic manifestations of epidermotropism: erythematous patches, elevated plaques, alopecia, and cellulitis. Many patients will experience progressive disease with lymph node infiltration, tumor and a leukemic phase termed the Sezary Syndrome. The lack of a survival benefit offered by clinical trials of combination chemotherapy and radiation supports the current standard of initial topical therapy for most patients with CTCL. Topical or skin-directed approaches widely offered include: nitrogen mustards (meclorethamine or carmustine), corticosteroids, oral psoralen with UVA radiation, UVB phototherapy and electron beam irradiation. Recent, early phase clinical studies of histone deacetylase inhibitors (HDACi) have illustrated a remarkable activity among patients with advanced CTCL. The structural dissimilarity of these ligands supports an on-target antineoplastic effect. Unfortunately, the frequently severe side effects associated with the systemic delivery of these nonselective ligands may limit their downstream clinical utility in the majority of patients with this disease, even with topical administration. We therefore devised a chemical strategy to achieve high dose-intensity at the site of cutaneous disease following topical administration with limited systemic exposure. Suberohydroxamic acid phenyl ester (SHAPE) is a first-in-class reverse prodrug inhibitor of histone deacetylases. Between the aliphatic chain and aromatic capping element of this canonical HDACi, we positioned an ester bond so as to promote presystemic metabolism by serum esterases. Preclinical studies of this ligand demonstrate potent activity against nuclear and cytosolic HDAC proteins. Doses required to achieve maximal histone hyperacetylation in human cancer cell tissue culture are comparable to vorinostat (SAHA; Merck & Co., Rahway, NJ) and MS-275 (Berlex Pharmaceuticals, Montville, NJ). These studies support the stability of SHAPE in the intracellular environment amid cellular esterases. The rapid metabolism (t-½ less than 5 minutes) by serum esterases has been established biochemically using continuous spectrophotometry, and is mediated by both butyrylcholinesterase and paraoxonase. Tolerability and preliminary activity have recently been demonstrated in an IL-7 transgenic mouse model of CTCL supporting clinical development as a therapeutic strategy in humans. Application to premalignant, neoplastic and inflammatory conditions of the oropharynx is being explored. Further studies characterizing the impact of SHAPE on T-cell function are also ongoing, to assess utility in conditions such as psoriasis and cutaneous graft-versus-host disease.

Autophagy is an evolutionarily conserved process by which cellular structures may be degraded for ongoing biogenesis. Following nutrient deprivation, autophagy functions critically to replenish metabolic precursors by recycling nonessential structures. Recent evidence has identified the induction of autophagy following growth factor withdrawal in preclinical models of cancer. As the autophagic response in these studies requires abrogation of apoptotic effectors, autophagy emerges as a relevant cell survival mechanism in human cancer. While growth factor depletion is an uncommon approach in cancer therapy, small molecule inhibition of dominant growth factor pathways is a central focus of targeted chemotherapeutic development. Therefore, we predicted that small molecule inhibition of growth factor pathways will prompt an autophagic response in human cancers, and that additional inhibition of autophagy will result in a synergistic and tumor-selective cytotoxicity. As there are few known inhibitors of autophagy, we developed a forward chemical genetic discovery approach utilizing high-throughput, high-content, epifluorescent microscopy. A human glioblastoma cell line stably expressing an EGFP-LC3 fusion protein was cultured at low confluency in 384-well plate format. Four hours following 100 nL pin-transfer of an arrayed small molecule library, cells were fixed and nuclei were counter-stained with Hoechst. Images were acquired and analyzed using automated microscopy and scripting software (MetaXPress; Molecular Devices, Sunnyvale, CA). The designed algorithm proved capable of quantitative cell-scoring for the presence of characteristic, fluorescent, LC3-positive autophagic punctae, as well as discrete annotation of vesicle size permitting the differentiation of likely autophagy inducers from inhibitors of autophagosome-lysosome fusion. Statistical rendering of four phenotypic measurements for the selection of assay positives was performed in the ChemBank analysis environment (NCI-Initiative for Chemical Genetics, Cambridge, MA). With an interest in discovering immediate candidates for clinical investigation, we profiled a 3000-member small molecule library of off-patent pharmaceutical products and bioactive compounds with known pharmacology. Six ligands were discovered which demonstrated a phenotype characteristic of late inhibition of autophagosomal maturation. Each is an established pharmaceutical product. At least three of these molecules have previous annotation as inhibitors of vesicle function in the medical literature. With an interest in profiling these ligands for synergy with growth factor pathway inhibitors in cancer, we established a novel platform for assessing synergistic cytotoxicity in high-throughput format. Data collected from these experiments confirm the synergistic antineoplastic activity of known and novel inhibitors of autophagy when combined with receptor tyrosine kinase inhibition in model systems of hematologic and epithelial tumors. These experiments serve as the scientific basis for clinical studies planned in lung cancer, breast cancer and multiple myeloma. Additionally, these data identify a panel of small molecule inducers of autophagy of plausible utility in protein aggregation disorders such as Huntington’s, Parkinson’s and Prion Diseases.

The response of refractory multiple myeloma to the proteasome inhibitor bortezomib reveals an intriguing sensitivity of this incurable malignancy to perturbations of protein catabolism. However, an overall clinical response rate of approximately 30% as a single agent suggests the importance of chemoresistance mediated by compensatory mechanisms of protein degradation. With proteasome inhibition, juxtanuclear inclusion bodies accumulate. These “aggresomes” are specific cellular structures comprised of chaperones, misfolded proteins, and proteasome components. The cytoplasmic histone deacetylase inhibitor (HDAC6) is essential for aggresome formation. Recently, we have demonstrated robust cytotoxic synergy in multiple myeloma cells between bortezomib and the carboxy-terminal domain-selective inhibitor of HDAC6, tubacin (1). Tubacin is a hydroxamic acid member of a diversity-oriented synthetic chemical library realized and validated previously by members of our laboratory (2). In our study of bortezomib and tubacin in multiple myeloma, we noted the dose-dependent, significant association between cytotoxicity and the marked accumulation of polyubiquitinated proteins in sensitized cells. With an interest in further interruption of the misfolded protein response, we have explored the cytosolic chaperone protein, hsp90, as an additional target using a chemical biologic approach. 17-AAG is an analog of the benzoquinone ansamycin antibiotic geldanamycin, known to bind to the ADP/ATP pocket of hsp90. 17-AAG stabilizes a conformation of the chaperone favoring targeted degradation of its client proteins via recruitment of the hsp70 co-chaperone complex (3). Further supporting this strategy, collaborators have recently identified that HDAC6 binds to and principally mediates the deacetylation of hsp90 (4). Inhibition with non-selective HDAC inhibitors was shown to augment hsp90 acetylation and inhibit ATP binding, resulting in the promotion of protein degradation by polyubiquitination. With the support of these data, we investigated whether the addition of 17-AAG to bortezomib and tubacin results in increased cytotoxicity in multiple myeloma cells. Indeed, we demonstrate potent cytotoxicity in cultured myeloma cells at low concentrations of each small molecule. Focused study of the MM.1S cell line demonstrates that the addition of 17-AAG to tubacin and bortezomib markedly increases the burden of ubiquitinated, cytosolic cellular protein by 24 hours, correlating with enhanced cell killing. These data further credential the protein degradation pathway in multiple myeloma, demonstrate the contribution of targeted, combined approaches with active small molecules, and provide a blueprint for a rational therapeutic strategy.

Bortezomib (Velcade®) triggers significant anti-tumor activity in multiple myeloma (MM) both in preclinical models and in patients with relapsed refractory disease. However, 60–65% of patients did not respond to bortezomib, and mechanisms of resistance to bortezomib are not fully understood. Recent studies have revealed an alternative system to the proteasome for degradation of polyubiquitinated misfolded/unfolded proteins, termed the aggresome. Aggresome formation ultimately induces autophagic clearance, which terminates in lysosomal degradation. The aggresome pathway therefore likely provides a novel system for delivery of aggregated proteins from cytoplasm to lysosomes for degradation. In aggresomal protein degradation, histone deacetylase6 (HDAC6) has an essential role. We therefore hypothesized that inhibition of both mechanisms of protein catabolism could induce accumulation of ubiquitinated proteins, followed by significant cell stress and cytotoxicity in MM cells. We demonstrate that HDAC6 specific inhibitor Tubacin specifically triggers acetylation of α-tubulin, as a result of HDAC6 inhibition, in a dose- and time-dependent fashion. It induces cytotoxicity in MM cells, which is mediated via caspase-dependent apoptosis; no toxicity is observed in normal peripheral blood mononuclear cells. Tubacin inhibits the interaction of HDAC6 with dynein and induces marked accumulation of ubiquitinated proteins. It synergistically augments bortezomib-induced cytotoxicity via c-Jun NH2-terminal kinase (JNK)/caspase activation. Importantly, this combination also induces significant cytotoxicity in plasma cells isolated from MM patient bone marrow (BM). Adherence of MM cells to BM stromal cells confers growth and resistance to conventional treatments; in contrast, the combination of tubacin and bortezomib triggers toxicity even in adherent MM cells (Hideshima et al. PNAS102:8567–8572, 2005). In the present study, we further examined the biologic sequelae of inhibiting both proteasome and aggresome pathways in MM. Bortezomib induces aggresome formation, evidenced by immunohistochemical analyses using DAPI, ubiquitin and γ-tubulin Abs. Tubacin enhances 20S proteasome activity in a time-dependent fashion; conversely bortezomib significantly augments HDAC6 protein expression. Taken together, these results suggest that proteasome function could compensate for aggresome function and vise versa. Importantly, treatment with the combination of Tubacin and bortezomib induces caspase-12 cleavage and upregulation of Grp78 in MM cells suggesting endoplasmic reticulum (ER) stress, associated with increased apoptosis. Finally, heat shock protein (Hsp)90 is a chaperone for transport of ubiquitinated proteins to the proteasome. HDAC6 is also constitutively associated with Hsp90, and the HDAC6 inhibitor Tubacin augments Hsp90 acetylation. Ongoing studies will delineate the biologic role of HDAC6 in the chaperone function both in proteasome and aggresome protein degradation cascades. Our studies therefore demonstrate that Tubacin combined with bortezomib mediates significant anti-MM activity by blocking both the aggresome and proteasome, respectively, thereby providing the framework for clinical evaluation of combined therapy to improve patient outcome in MM.

Background The replication (G2/M) checkpoint is principally mediated by the serine/threonine protein kinase ATR (ataxia telangiectasia mutated and Rad3-related). ATR is a large (350 kD) member of the phosphatidylinositol kinase related kinase family. After exposure to genotoxic or replication stress, ATR halts cell cycle progression, allowing DNA repair complexes time enough to restore the fidelity of the genome prior to cell division. Previous experiments have demonstrated that cancer cells with p53 mutation are critically dependent on ATR-mediated arrest of the cell cycle. Industrial approaches to identify ATR inhibitors have failed likely as a result of protein insolubility. Methods We have undertaken a novel chemical genetic approach employing small molecule microarrays (SMMs) to identify molecules with high binding specificity for ATR. Three diversity-oriented combinatorial chemical libraries of more than 15,000 entities were generated by split-pool synthesis in solid phase on polystyrene macrobead supports. Compounds were robotically printed in microarray format on glass slides. Four analogs of FK506 were printed as positive controls. Extracts were prepared from mammalian cells transfected with over-expression constructs of FLAG-tagged ATR, FKBP12 and GFP. A protocol was developed and optimized for screening employing a primary anti-FLAG mouse monoclonal antibody and Cy5-fluorophore labeled anti-mouse antibody. Data analysis for small molecule binders was performed with GenePix software on an Axon Scanner. Biological activity of these molecules was analyzed in the context of mitotic spread and chromosomal fragility assays. Results Protein expression and antibody fidelity was verified by Western blot. The lysate-based SMM screening approach was optimized and validated by recognition of an interaction between over-expressed, epitope-tagged FKBP12 and analogs of FK506. Six small molecule hits suggesting ATR binding were identified and verified by triplicate microarray assays. Positive compounds were structurally similar members of a dihydropyrancarboxamide library suggesting recognition of a common target. Mitotic spread analysis of cells treated with two of these molecules and hydroxyurea demonstrated the premature chromatin condensation phenotype characteristic of replication checkpoint inhibition. Chromosomal fragility was notably augmented by these molecules as well. Chemosensitivity following replication stress was witnessed in p53-negative cells relative to an otherwise identical wild-type cell line. Conclusions Classical approaches to drug discovery are often limited by challenges in protein biochemistry such as protein size, solubility, activity and yield. We present compelling data that the small molecule microarray format can effectively be tailored for use with cellular lysates over-expressing a protein target of biological interest. Furthermore, we have used an optimized protocol to identify two novel, active small molecule inhibitors of the replication checkpoint (SMIRC-1 and SMIRC-2). The enhanced chemosensitivity in p53-negative cell lines supports a plausible role for ATR inhibitors as potentially useful chemotherapeutic agents.

SummaryChromatin regulators have become attractive targets for cancer therapy, but it is unclear why inhibition of these ubiquitous regulators should have gene-specific effects in tumor cells. Here, we investigate how inhibition of the widely expressed transcriptional coactivator BRD4 leads to selective inhibition of the MYC oncogene in multiple myeloma (MM). BRD4 and Mediator were found to co-occupy thousands of enhancers associated with active genes. They also co-occupied a small set of exceptionally large super-enhancers associated with genes that feature prominently in MM biology, including the MYC oncogene. Treatment of MM tumor cells with the BET-bromodomain inhibitor JQ1 led to preferential loss of BRD4 at super-enhancers and consequent transcription elongation defects that preferentially impacted genes with super-enhancers, including MYC. Super-enhancers were found at key oncogenic drivers in many other tumor cells. These observations have implications for the discovery of cancer therapeutics directed at components of super-enhancers in diverse tumor types.

Summary Elevated expression of the c-Myc transcription factor occurs frequently in human cancers and is associated with tumor aggression and poor clinical outcome. The effect of high levels of c-Myc on global gene regulation is poorly understood but is widely thought to involve newly activated or repressed "Myc target genes." We report here that in tumor cells expressing high levels of c-Myc the transcription factor accumulates in the promoter regions of active genes and causes transcriptional amplification, producing increased levels of transcripts within the cell's gene expression program. Thus, rather than binding and regulating a new set of genes, c-Myc amplifies the output of the existing gene expression program. These results provide an explanation for the diverse effects of oncogenic c-Myc on gene expression in different tumor cells and suggest that transcriptional amplification reduces rate-limiting constraints for tumor cell growth and proliferation.

Abstract 3484 Acute myeloid leukemia (AML) is a stem cell-derived malignancy characterized by uncontrolled proliferation and accumulation of myeloblasts in hematopoietic tissues. The clinical course and prognosis in AML vary depending on deregulated genes, cell type(s) involved, and the biological properties of the clone. In most variants of AML, the complexity and heterogeneity of oncogenomes pose a challenge for the development of effective targeted therapeutics. However, diverse genetic aberrations in AML typically converge functionally to dysregulate the same cellular core processes. One key event is the corruption of myeloid cell-fate programs resulting in the generation of aberrantly self-renewing leukemia stem cells (LSC), which maintain and propagate the disease and are often resistant to conventional chemotherapy. Hence, strategies aimed at terminating aberrant self-renewal and eradicating LSC are considered as key for the development of more effective AML therapies. In an effort to systematically probe genes involved in chromatin regulation as potential therapeutic targets, we recently employed an unbiased screening approach combining AML mouse models and new in-vivo RNAi technologies, through which we identified the epigenetic ‘reader' BRD4 as new candidate drug target in AML (Zuber et al., Nature, in press). Inhibition of BRD4 using RNAi or a new small-molecule inhibitor (JQ1) blocking BRD4 binding to acetylated histones, showed profound antileukemic effects in AML mouse models, in all human AML cell lines tested (n=8) as well as in primary AML cells. In all models tested, BRD4 suppression was found to trigger apoptosis as well as terminal myeloid differentiation, and potently suppressed expression programs previously associated with LSC. As one key target, we observed a dramatic transcriptional repression of MYC, which recently has been discussed as core component of an LSC associated transcriptional module. To further evaluate suppression of BRD4 as a potential therapeutic approach to eradicate LSC in human AML, we analyzed the effects of JQ1 in primary AML cells obtained from 17 patients with freshly diagnosed or relapsed/refractory AML (females, n=5, males, n=12, median age: 54 years; range: 21–80 years). In unfractionated primary AML cells, submicromolar doses of JQ1 were found to induce major growth-inhibitory effects (IC50 between 0.05 and 0.5 μM) in a broad spectrum of AML subtypes. No differences in IC50 values were seen when comparing drug effects in AML cells kept in the presence or absence of growth-stimulating cytokines (G-CSF, IL-3, SCF). In addition, JQ1 treatment effectively triggered apoptosis in all patients tested, with similar anti-leukemic activities observed in newly diagnosed pts and refractory/relapsed AML. To further evaluate the clinical value of BRD4 as a clinically relevant target in AML, we analyzed the effect of JQ1 on AML LSC. In these experiments, JQ1 effectively induced apoptosis in CD34+/CD38+ progenitor cells as well as in CD34+/CD38− AML stem cells in all donors examined as evidenced by combined surface/Annexin-V staining. Furthermore, JQ1 was found to induce morphologic signs of maturation in 6 of 7 patients examined, thereby confirming our previous data obtained in mouse AML cells. Finally, we were able to show that JQ1 synergizes with Ara-C in inducing growth inhibition in HL60 cells and KG-1 cells. In summary, our data show that small-molecule inhibition of BRD4 has strong anti-leukemic effects in a broad range of AML subtypes. Furthermore, our results support the notion that JQ1's ability to suppress LSC specific transcriptional modules may translate into a therapeutic entry point for eradicating LSC in primary AML. While a more extensive in vivo evaluation of these effects, as well as the development of pharmacologically improved compounds will be required, all existing data unambiguously highlight small-molecule inhibition of BRD4 as a new promising concept in AML therapy. Disclosures: No relevant conflicts of interest to declare.