METHODS TO IDENTIFY INHIBITORS OF THE UNFOLDED PROTEIN RESPONSE

Abstract

Methods for identifying compounds that are inhibitors of the unfolded protein response are provided. In particular, the methods identify compounds that inhibit the activity of IRE1.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of U.S. Provisional Application No. 60/777,458, filed Feb. 27, 2006, the disclosure of which is incorporated herein by reference in its entirety.

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

This invention was made in part with government support under PHS Grant No. 1R01CA112108-01A1, awarded by the National Institutes of Health. The government may have certain rights in the invention.

FIELD OF THE INVENTION

The present invention relates generally to methods to identify inhibitors of the unfolded protein response. Inhibitors identified by the instant methods are of use, for example, in the treatment of disorders characterized by cell growth in hypoxic conditions, such as cancers, in particular solid tumors. The present invention includes methods to monitor the activity of IRE1 in cells under stress, in particular hypoxic stress.

BACKGROUND OF THE INVENTION

A defining feature of solid tumors is their capacity to divide aggressively and disseminate metastases under conditions of nutrient deprivation and limited oxygen availability. These severe stresses arise from inadequate perfusion as the primary tumor rapidly outgrows its initial blood supply, and from dramatic structural abnormalities of tumor vessels that can lead to disturbed microcirculation (Hockel and Vaupel, Semin. Oncol. 28(2 Suppl 8):36-41, 2001; Vaupel, et al. Med. Oncol. 18:243-59, 2001). As a result, regions of low O₂ tension, or hypoxia, are heterogeneously distributed within the tumor mass. While tumor hypoxia is a physiological barrier to cell survival, it paradoxically drives malignant progression by imposing a powerful selective pressure for cells that can best adapt to this stress and subsequently resume cell division.

Tumor hypoxia also correlates with a more aggressive disease course and increased failure following radiation and chemotherapy. The presence of hypoxia has been demonstrated in a wide variety of human cancers, including cervix, breast, lung, brain, pancreas, head and neck, and prostate (Evans S., & Koch C. Cancer Lett. 195:1-16, 2003). Many of these tumors contained regions of severe hypoxia (<5 mmHg oxygen). Clinically, the duration of disease- and progression-free survival correlates inversely with the degree of tumor hypoxia. For example, in patients with squamous carcinoma of the head and neck, the one year disease-free survival was 78% for patients with median tumor pO2>10 mm Hg but only 22% for median pO2<10 mm (Brizel, et al., Int. J. Radiat. Oncol. Biol. Phys. 38:285-9, 1997). Hypoxic cells also exhibit increased resistance to standard radiation and chemotherapy treatment programs, as these cells are relatively isolated from the blood supply and because radiation and chemotherapy preferentially kill rapidly dividing cell populations. Collectively, these findings provide strong evidence that hypoxia has a profound impact on tumor growth and clinical outcome.

Hypoxia dramatically reshapes cellular physiology, causing cell cycle arrest, a shift in energy production to glycolysis, elevated secretion of survival and pro-angiogenic factors, expression of genes involved in drug resistance, and increased cell motility and invasion. A watershed discovery linking these profound changes to the control of gene expression was made with the identification of hypoxia-inducible factor (HIF), a heterodimeric transcription factor that exerts control over a broad range of cellular pathways including glycolysis, angiogenesis and erythropoiesis (Semenza, Trends Mol. Med. 2002 8(4 Suppl):S62-7, 2002; Semenza, Nat. Rev. Cancer 3:721-32, 2003).

While HIF controls the expression of more than 60 genes and constitutes a key node in cellular stress signaling, HIF activation alone cannot account for the full repertoire of changes that occur intracellularly as oxygen becomes limiting. The hypoxic cell also elicits additional, HIF-1-independent, adaptive responses that contribute to increased survival under low oxygen conditions. For example, an immediate reaction to hypoxia is a reduction in the rates of global protein synthesis, which reduces energy demands when oxygen and ATP levels are low (Hochachka et al., Proc. Natl. Acad. Sci. USA, 93:9493-8, 1996). Further, hypoxia causes a sharp increase in the expression of molecular chaperones, which assist in protein refolding and in the degradation of terminally misfolded conformers. Underlying these changes is a coordinated cellular program called the unfolded protein response (UPR) that serves as a master regulator of cellular homeostasis and which plays a fundamental cytoprotective role during cellular stresses such as hypoxia.

The endoplasmic reticulum (ER) is an extensive intracellular membrane network that extends throughout the cytoplasm and functions primarily to process newly synthesized secretory and transmembrane proteins. Accumulation of unfolded proteins in this compartment causes ER stress, with prolonged ER stress resulting in cell death. The cellular response to ER stress consists of at least two coordinated pathways: 1) rapid translational arrest mediated by PERK (pancreatic ER kinase or PKR-like ER kinase); and 2) transcriptional activation of unfolded protein response (UPR) target genes (Ron D. J. Clin. Invest. 110:1383-1388, 2002; Harding H., et al. Annu. Rev. Cell. Dev. Biol. 18:575-599, 2002; Feldman D. E., et al. Mol. Cancer. Res. 3:597-605, 2005). In addition to solid tumors, the UPR has been implicated in diseases such as conformational diseases, diabetes, cardiovascular disease, atherosclerosis, viral infection, and cerebrovascular disease (Schroder M., et al. Mutat. Res. 569:29-63, 2005; Kaufman R. J. Clin. Invest. 110:1389-1398, 2002).

During normal embryonic development, activation of the UPR is essential for the maturation of secretory cells in the liver and pancreas, and drives an expansion of the ER in antibody-secreting B lymphocytes to accommodate increased secretory load. Iwakoshi et al., Immunological Reviews 194: 29-38 (2003); Harding et al., Molecular Cell 5: 897-904 (2000); Shaffer et al., Immunity 21: 81-93 (2004); Reimold et al., Genes Dev 14: 152-157 (2000). Several lines of evidence have also implicated the UPR in various disease processes, such as diabetes and cardiovascular disease, and as a survival mechanism underlying tumor growth and the adaptation of malignant cells to hypoxic stress. Ma and Hendershot, Nat Rev Cancer 4: 966-977 (2004); Feldman et al., Mol Cancer Res 3: 597-605 (2005); Koumenis, Curr Mol Med 6: 55-69 (2006).

A critical feature of malignant tumors is their capacity to survive and seed distant metastases under conditions of nutrient deprivation and limited oxygen availability. Hockel and Vaupel, Seminars in Oncology 28: 36-41 (2001); Vaupel et al., Methods in Enzymology 381: 335-354 (2004); Subarsky and Hill, Clin Exp Metastasis 20: 237-250 (2003). Intratumoral hypoxia arises solid tumors through severe structural abnormalities of tumor vasculature and disturbed microcirculation, resulting in tissue regions of extremely low O₂ partial pressures distributed heterogeneously within the tumor mass. Vaupel et al., Methods in Enzymology 381: 335-354 (2004); Hockel and Vaupel, Journal of the National Cancer Institute 93: 266-276 (2001); Vaupel et al., Medical Oncology 18: 243-259 (2001). Since the delivery of oxygen and nutrients to the tumor is determined by fluctuating blood flow, different regions of the tumor must constantly adjust to varying degrees of nutrient deprivation. The tumor microenvironment thus imposes a strong selective pressure for cells best adapted for survival under these stresses. Adaptation to hypoxia contributes to the diminished apoptotic potential of tumor cells and accounts for many of the clinical consequences of malignant progression, including locoregional tumor recurrence and distant metastases. Evans and Koch, Cancer Letters 195: 1-16 (2003); Le et al., Cancer Metastasis Rev 23: 293-310 (2004). Hypoxia-mediated clonal expansion of cells with diminished apoptotic potential has been demonstrated in vitro, and hypoxic cells exhibit increased metastatic potential. Erler et al., Nature 440: 1222-1226 (2006); Graeber et al., Nature 379: 88-91 (1996). Importantly, depletion of molecular oxygen or glucose impairs the posttranslational modification and oxidative folding of secretory proteins, providing a direct biochemical link between nutrient deprivation in tumors and activation of the UPR. Tu et al., Science 290: 1571-1574 (2000); Koumenis et al., Molecular & Cellular Biology 22: 7405-7416 (2002).

PERK, an ER transmembrane protein, was first identified as regulating translational attenuation during ER stress through the phosphorylation of translation initiation factor eIF2α. While most mRNA translation is repressed following phosphorylation of eIF2α, activating transcription factor 4 (ATF4) is selectively translated during ER stress leading to increased expression of chaperones, foldases, and downstream targets such as CHOP/GADD153, a pro-apoptotic gene. Koumenis et al demonstrated that translational control of protein synthesis during hypoxia also occurs through the activation of PERK. These investigators showed that PERK −/− MEFs where unable to phosphorylate eIF2α and had decreased survival after exposure to hypoxia compared to the wild-type MEFs. They concluded that PERK plays an important role in hypoxia-induced translation attenuation, further supporting a role for hypoxia in the development of ER stress (Koumenis et al., Mol. Cell. Biol. 22:7405-7416 (2002)). A rapid decrease in de novo protein synthesis upon exposure to hypoxia has also been observed (Chen et al., Cancer Res. 64:7302-7310 (2004)). Downstream of PERK, ATF4 is also activated by hypoxia in a HIF-1 independent manner. One consequence of ATF4 activation is induction of a GADD34 which feeds back to desphosphorylate eIF2α and release cells from translational inhibition.

In coordination with the inhibition of protein synthesis, the UPR is also responsible for the transcriptional activation of a discrete set of genes. These genes function to increase the cellular folding capacity through the induction of ER chaperone proteins and folding enzymes. The UPR is a conserved stress response and many of its downstream target genes have been characterized in yeast and mammalian cells. In mammalian cells, activating transcription factor 6 (ATF6) and X-box binding protein (XBP1) are critical regulators of the transcriptional response to ER stress.

The ER resident transmembrane protein IRE1 is conserved in throughout eukaryotic phylogeny and functions as both a proximal sensor of ER stress and as a critical UPR signal transducer via its dual cytoplasmic kinase and endoribonuclease domains. Tirasophon et al., Genes Dev 12: 1812-1824 (1998). Mammalian IREL1α, the major functional homolog of yeast IREL1α, excises a 26-nucleotide intron from the mRNA encoding the bZIP transcription factor XBP-1. This introduces a translational frame shift downstream of the splice site to generate XBP-1s, a potent transcription factor. Yoshida et al., Cell 107: 881-891 (2001); Calfon et al., Nature 415: 92-96 (2002); Lee et al., Genes & Development 16: 452-466 (2002). XBP-1s drives an expansion of ER capacity through the increased expression of molecular chaperones and components of the ER-associated protein degradation (ERAD) machinery that is required for the clearance of terminally misfolded proteins. Schroder and Kaufman, Mutation Research 569: 29-63 (2005); Lee et al., Molecular & Cellular Biology 23: 7448-7459 (2003). IRE1α is extensively activated in hypoxic regions of human tumor xenografts throughout tumorigenesis (Feldman et al., Mol Cancer Res 3: 597-605 (2005)), and transformed mouse fibroblasts genetically deleted for XBP-1 exhibit increased sensitivity to hypoxia and fail to grow as tumors when implanted into immune-deficient mice (Romero-Ramirez et al., Cancer Research 64: 5943-5947 (2004)). Activation of IRE1α by ER stress triggers multiple signaling outputs that extend beyond the splice-activation of XBP-1, including IRE1α endonuclease-mediated cleavage of a subset of mRNAs encoding secretory proteins (Hollien and Weissman, Science 313: 104-107 (2006)), and activation of autophagy and apoptosis pathways through the IRE1α kinase domain and its downstream effectors caspase-12, ASK1, and JNK1 (Ogata et al., Mol Cell Biol (2006); Urano et al., Science 287: 664-666 (2000)). Thus IRE1α may participate in both cytoprotective and pro-apoptotic pathways.

A schematic of the UPR pathway is shown in FIG. 1. In this model, GRP78 regulates each of the major branches of the UPR by direct association with ATF6, IRE1 and PERK. Given its importance in regulating the UPR, GRP78 levels can be increased by downstream signaling from each of these pathways, indicating that significant overlap occurs in activation of the UPR.

The functional link between the UPR and hypoxia was found through studies on GRP78, a critical regulator of the UPR. Expression of the glucose regulated family of proteins (GRPs) within solid tumors was recognized more than a decade ago. These experiments indicate that glucose starvation and hypoxia were physiologically relevant stresses occurring during the growth of solid tumors (Cai J., et al., J. Cell. Physiol. 154:229-237, 1993). Furthermore, cells in which GRP78 expression was inhibited through an antisense strategy exhibited increased sensitivity to hypoxia compared to the parental wild-type cell line (Koong A., et al., Int. J. Radiat. Oncol. Biol. Phys. 28:661-666, 1994).

Other UPR regulated genes such as GRP94 and protein disulfide isomerase (PDI) have also been implicated in mediating neuronal survival after ischemia/reperfusion injury (Sullivan D., et al., J. Biol. Chem. 278:47079-47088, 2003; Bando Y., et al., Eur. J. Neurosci. 18, 2003.). Similarly, oxygen regulated protein 150 kDal (ORP150, also known as GRP170), another ER chaperone protein, protected neurons from ischemic stress in a cell culture model and reduced the cerebral infarct area after middle cerebral artery occlusion in a transgenic mouse model (Tamatani M., et al., Nat. Med. 7:317-323, 2001).

These studies indicate that the UPR has a broad range of functions during hypoxia including promotion of cell survival and regulation of angiogenesis. Given its role in regulating survival under hypoxia and its requirement for tumor growth, targeting XBP-1 may be an effective therapeutic strategy. However, there are currently few examples of anti-cancer drugs that can effectively inhibit transcription factor activation. There thus remains a need for compositions that may be employed to inhibit the activity of XBP-1 and thereby prevent or inhibit tumor growth.

Identification of compounds capable of inhibiting the activity of XBP-1 and thereby capable of preventing or inhibiting tumor growth would be facilitated by assays suitable for use in high throughput screens. Direct measurement of XBP-1 levels in cells is not easily automated. Convenient and easily detectable substrates for the endonuclease or kinase activities of IRE1 are currently unavailable. US Patent Application No. 2003/0224428 reports methods purportedly useful in screening inhibitors of IRE1-mediated processing of untranslatable XBP-1 mRNA. The reported methods are limited to the screening of plasma cells or virus-infected cells, however, and are therefore unsuitable for identifying compounds useful in the treatment or prevention of disorders in more general cell types and tissues. The methods also fail to account for the effects of tumor microenvironment, such as, for example, hypoxia, on the activity of potential therapeutic compounds. The methods also lack steps to counterscreen for compounds causing non-specific effects on the detectable marker and for compounds that are toxic to cells even in the absence of ER stress. The methods would therefore falsely identify compounds that have nothing to do with the UPR and that would be unsuitable for therapeutic use. Furthermore, the methods have not been shown to be suitable for use in high throughput screening assays.

Due to the importance of the unfolded protein response in cellular metabolism, and, in particular, in pathological processes, there is great interest in developing inhibitors with defined specificities against this process. Such inhibitors can help to identify target enzymes in cells, particularly where the cells are associated with particular indications, and can provide new drug candidates. There is thus a need for inhibitors of the unfolded protein response and novel methods of inhibiting this pathway, as well as methods of treating or preventing disorders of the unfolded protein response and methods of identifying novel inhibitors of the pathway.

SUMMARY OF THE INVENTION

The present invention addresses these problems by providing novel methods to identify inhibitors of the unfolded protein response.

In one aspect, the invention provides methods comprising the steps of:

providing a first array of cells that stably express an mRNA fusion sequence, wherein the mRNA fusion sequence comprises a first mRNA segment comprising an unprocessed XBP-1 transcription factor gene sequence and a second mRNA segment comprising a reporter gene sequence, and wherein the first mRNA segment is processed by IRE1 to form a frameshifted mRNA fusion sequence that is translatable by a cell to produce a detectable protein;

contacting the first array of cells with a library of compounds; and

identifying a compound that inhibits the activity of IRE1.

In some embodiments, the library of compounds comprises at least 50, at least 100, at least 500, at least 1000, or at least 5000 different compounds.

In some embodiments, the first array of cells comprises a microtiter plate.

In some embodiments, the detectable protein is an enzyme.

In some specific embodiments, the enzyme is luciferase.

In some embodiments, the detectable protein is a fluorescent protein.

In some embodiments, the detectable protein is detected using an antibody.

In other embodiments, the method further comprises the step of counterscreening the library of compounds to identify a compound that is not toxic to cells grown in air.

In another aspect of the invention, the method further comprises the step of stimulating the unfolded protein response prior to contacting the first array of cells with the library of compounds.

In certain embodiments, the unfolded protein response is stimulated by treatment of the cells with tunicamycin and thapsigargin.

In other embodiments, the unfolded protein response is stimulated by treatment of the cells with hypoxic conditions.

In some embodiments, the library of compounds comprises at least 50, at least 100, at least 500, at least 1000, or at least 5000 different compounds.

In some embodiments, the first array of cells comprises a microtiter plate.

In some embodiments, the detectable protein is an enzyme.

In some specific embodiments, the enzyme is luciferase.

In some embodiments, the detectable protein is a fluorescent protein.

In some embodiments, the detectable protein is detected using an antibody.

In some specific embodiments, the method further comprises the step of counterscreening the library of compounds to identify a compound that is not toxic to cells grown in air.

In another aspect of the invention, the method further comprises the step of counterscreening the library of compounds to identify a compound that inhibits detection of the detectable protein.

In some embodiments, the counterscreening step comprises the use of a second array of cells that constituitively express the detectable protein.

In some embodiments, the library of compounds comprises at least 50, at least 100, at least 500, at least 1000, or at least 5000 different compounds.

In some embodiments, the first array and second array each comprise a microtiter plate.

In some embodiments, the detectable protein is an enzyme.

In specific embodiment, the enzyme is luciferase.

In some embodiments, the detectable protein is a fluorescent protein.

In some embodiments, the detectable protein is detected using an antibody.

In some specific embodiments, the method further comprises the step of counterscreening the library of compounds to identify a compound that is not toxic to cells grown in air.

In another aspect of the invention, the method further comprises the steps of stimulating the unfolded protein response prior to contacting the first array of cells with the library of compounds; and counterscreening the library of compounds to identify a compound that inhibits detection of the detectable protein.

In some embodiments, the unfolded protein response is stimulated by treatment of the cells with tunicamycin and thapsigargin.

In some embodiments, the unfolded protein response is stimulated by treatment of the cells with hypoxic conditions.

In some embodiments, the counterscreening step comprises the use of a second array of cells that constituitively express the detectable protein.

In some embodiments, the library of compounds comprises at least 50, at least 100, at least 500, at least 1000, or at least 5000 different compounds.

In some embodiments, the first array and second array each comprise a microtiter plate.

In certain embodiments, the detectable protein is an enzyme.

In more specific embodiments, the enzyme is luciferase.

In other embodiments, the detectable protein is a fluorescent protein.

In yet other embodiments, the detectable protein is detected using an antibody.

In some embodiments, the method further comprises the step of counterscreening the library of compounds to identify a compound that is not toxic to cells grown in air.

In another aspect, the invention provides methods to identify inhibitors of IRE1, wherein the processing by IRE1 is an RNA splicing reaction.

In yet another aspect, the invention provides methods to identify inhibitors of IRE1, wherein the compound inhibiting the activity of IRE1 inhibits the endonuclease activity of IRE1.

In another aspect of the invention, the method further comprises the step of counterscreening the library of compounds to identify a compound that is not toxic to cells grown in air.

In another aspect, the invention provides methods wherein the identifying step comprises comparing the amount of detectable protein in cells treated with the compound to the amount of detectable protein in untreated cells.

In yet another aspect of the invention, the cells in the first array of cells are cancer cells.

In specific embodiments, the cancer cells are solid tumor cells.

In more specific embodiments, the solid tumor cells are selected from the group consisting of sarcoma cells, carcinoma cells, and lymphoma cells.

In even more specific embodiments, the cells are fibrosarcoma cells.

In other embodiments, the cells are adenocarcinoma cells.

In some embodiments, the cancer cells are selected from the group consisting of: multiple myeloma, cervical cancer, brain cancer, pancreatic cancer, head and neck cancers, prostate cancer, breast cancer, soft tissue sarcomas, primary and metastatic liver cancer, primary and metastatic lung cancer, esophageal cancer, colorectal cancer, lymphoma, and leukemia.

LISTING OF DRAWINGS

FIG. 1 is a schematic of the unfolded protein response (UPR) signaling pathway.

FIG. 2A is a schematic of a fusion protein in which unspliced XBP-1 is fused in frame with luciferase. Under hypoxia or ER stress, IRE1 splices a 26 nt sequence in XBP-1 causing a translational frameshift that allows read through of a stop codon, resulting in the production of an XBP-1-luciferase fusion protein. FIG. 2B shows the fold change in luciferase activity (RLU), detected after 24 hours of exposure to hypoxia, when HT1080 cells stably expressing the IRE1 reporter are allowed to reoxygenate.

FIG. 3 is a schematic of an initial screen of a 66,000 small molecule library for specific inhibitors of XBP-1.

FIG. 4 shows a “heat map” view of a single plate from the primary screen for inhibitors of XBP-1.

FIG. 5A shows examples of individual compounds tested at 1 uM, 2 uM and 6 uM for inhibition of tunicamycin-(Tm) induced transactivation of a 5 repeat XBP-1 promoter element (5X-UPRE)-luciferase reporter construct transiently transfected into HT1080 cells. FIG. 5B shows individual compounds tested for inhibition of hypoxia (48 hours) induced transactivation of the same UPRE-luciferase report construct transiently transfected into HT1080 cells.

FIG. 6A shows XBP-1 expression as determined by RT-PCR in HT1080 cells treated with hypoxia in the presence of various candidate inhibitors compounds. FIG. 6B shows the inhibition of XBP-luciferase reporter activity in hypoxia by the inventive irestatins. HT1080 fibrosarcoma cells stably expressing the Xbp-luciferase reporter were treated with 1 μM of each Irestatin or left untreated, and incubated in hypoxia (0.01% of oxygen) for 48 hours at 37° C. Cells were harvested, lysed in reporter lysis buffer, and assayed for luminescence using a luminometer.

FIGS. 7A and B show the hypoxia-specific cytotoxicity of candidate IRE1 inhibitors on HT1080 sarcoma cells and MiaPACA-2 cells, respectively, as determined in a clonogenic survival assay. FIG. 7C shows the inhibition of hypoxia survival of human tumor cells by candidate IRE1 inhibitors.

FIG. 8 shows the inhibition of IRE1-mediated XBP-1 splicing in hypoxia by the inventive irestatins.

FIGS. 9A-D illustrate the effects of administration of two different potential irestatins to nude mice implanted with HT1080 cells stably expressing XBP-1s-luciferase. FIG. 9A shows bioluminescent activity prior to injection, FIG. 9B shows activity 8 hours after injection, FIG. 9C shows activity 24 hours after injection, and FIG. 9D shows activity 8 hours after a second injection of the potential irestatins.

FIG. 10 shows the ability of the inventive irestatins to inhibit tumor growth in vivo in a mouse model. Dose: 60 mg/kg ip bolus injection every 48 hours. 5 total doses. 5-7 tumors per group. PANC1 pancreatic adenocarcinoma cell line.

FIG. 11 shows the inhibitory effects of Irestatin 9389 on the IRE1α/XBP-1 pathway.

FIG. 12 shows the inhibitory effects of Irestatin 9389 on the endonuclease function of IRE1α.

FIG. 13 shows that exposure to irestatin 9389 induces apoptosis and impairs cell survival under hypoxia and ER stress.

FIG. 14 shows the in vivo antitumor activity of irestatin 9389.

FIG. 15 shows expression of XBP-1s in human pancreas tissue specimens.

FIG. 16 shows histopathological analysis of mouse pancreas and liver tissues.

DETAILED DESCRIPTION OF THE INVENTION

In some embodiments, the present invention provides methods for identifying compounds that are capable of inhibiting the unfolded protein response, in particular the activity of IRE1. Methods for identifying compounds that modulate activity of IRE1 comprise providing a construct in which a reporter gene (e.g., a gene encoding a detectable protein such as, for example, firefly luciferase, Renilla luciferase, beta-galactosidase, green fluorescent protein, red fluorescent protein, yellow fluorescent protein, thymidine kinase, or a protein detectable by the binding of a further detector molecule, such as an antibody) is fused downstream and in frame with unspliced XBP-1, and stably transfecting this construct into a cell line, for example a tumor cell line such as fibrosarcoma or pancreatic adenocarcinoma cell lines (e.g. Panc1, MiaPaca and HT1080), or commonly used cell lines such as MDCK or HEK293. The transformed cells are then subjected to ER stress, for example, by adding drugs known to cause ER stress, such as tunicamycin and/or thapsigargin, or subjected to hypoxia, thereby activating the unfolded protein response. The cells are then incubated with a candidate inhibitor of the unfolded protein response, and the activity of the reporter gene in the treated cells is compared with that control, untreated, cells. If a reduction in reporter gene activity is observed, the candidate modulator is an inhibitor of the unfolded protein response, and in particular, of IRE1 activity.

In one aspect, the instant invention provides a method for identifying a compound as an inhibitor of IRE-1 activity, comprising:

(a) providing a construct in which a reporter gene is fused downstream and in-frame with unspliced XBP-1;

(b) stably transfecting cells with the construct to provide transformed cells;

(c) inducing endoplasmic reticulum (ER) stress in the transformed cells;

(d) contacting the transformed cells with a test compound to provide treated cells; and

(e) comparing activity of the reporter gene in the treated cells with activity of the reporter gene in control, untreated, cells,

whereby a difference in the activity of the reporter gene in the treated cells compared to the activity of the reporter gene in control, untreated, cells, indicates that the test compound is an inhibitor of IRE1 activity.

In specific embodiments, methods are provided wherein a reduction in the activity of the reporter gene in the treated cells compared to the activity of the reporter gene in control, untreated cells, indicates that the test compound is an inhibitor of IRE1 activity, wherein the reporter gene is firefly luciferase, wherein the cells are tumor cells, wherein the cells are human fibrosarcoma cells, wherein ER stress is induced by contacting the transformed cells with a compound known to induce ER stress, wherein the compound is selected from the group consisting of: tunicamycin and thapsigargin, and wherein the ER stress is induced by subjecting the transformed cells to hypoxia.

The inhibitors of IRE1 activity identified according to the instant methods may be referred to herein as irestatins. Compositions that contain one or more inhibitors of IRE1 activity may be effectively employed in the treatment of cancers, particularly those cancers characterized by the presence of moderate to severe hypoxia. Examples of such cancers include solid tumors and secretory cell malignancies, including multiple myeloma. Examples of solid tumor cells include sarcomas, carcinomas, and lymphomas. Cancers that may be effectively treated employing the inventive compositions include, for example, cervix, brain, pancreas, breast, head and neck, and prostate cancers, and soft tissue sarcomas. Accordingly, the methods for identifying inhibitors of IRE1 may make use of cells derived from any of the above cancers or tissues.

The methods of the instant invention take advantage of high throughput screening techniques. The methods of the invention can be performed, for example, using common disposable laboratory assay platforms such as microtiter plates and microarray slides. Microtiter plates and microarray slides suitable for use in the methods of the instant invention may conventionally contain, for example, 24, 96, 384, 768, or 1536 separate spots or wells. Alternative formats and sizes are, however, considered within the scope of the invention. Each separate microarray spot or microtiter plate well may contain cells expressing, for example, a fusion of unprocessed XBP-1 and a detectable protein. Alternatively, a microarray spot or microtiter plate well may contain control cells expressing the detectable protein alone or control cells lacking any expression construct. The microarray spots or microtiter plate wells will further contain an appropriate culture medium for proper maintenance and/or growth of the cultured cells, as would be understood by one of skill in the art.

As noted above, the detectable protein of the instant invention may be, for example, firefly luciferase, Renilla luciferase, β-galactosidase, green fluorescent protein, red fluorescent protein, yellow fluorescent protein, thymidine kinase, or a protein detectable by the binding of a further detctor molecule, such as an antibody. In some embodiments, the detectable protein may be α-galactosidase, alkaline phosphatase, horseradish peroxidase, exoglucanase, Bar1, Pho5 acid phosphatase, chitinase, or chloramphenicol acetyl transferase. In some embodiments, the detectable protein is an antigen that is specifically recognized by an antibody or fragment of an antibody that is itself detectable. In preferred embodiments, the detectable protein is luciferase.

The chemical libaries of use in the methods of the instant invention are readily available from commercial sources, for example, from Specs (Wakefield, R.I.), Chembridge (San Diego, Calif.), Maybridge (Maybridge, Cornwall, UK), MicroSource Discovery Systems, Inc. (Gaylordsville, Conn.), Prestwick Chemical Inc. (Washington, D.C.), BIOMOL International L.P. (Plymouth Meeting, Pa.), Sigma-Aldrich (St. Louis, Mo.), ChemRX, or others.

The high throughput screening techniques disclosed herein permit the rapid analysis of large chemical libraries to identify inhibitors of the unfolded protein response, and in particular of the activity of IRE1. The compound libaries screened according to the instant techniques may comprise, for example, at least 50, at least 100, at least 500, at least 1000, at least 5000 different compounds, or even more distinct compounds.

ER stress and the unfolded protein response may in some cases be stimulated in the cells of the instant methods prior to contacting the cells with the compounds of the compound library. As described above, ER stress and the unfolded protein response may be stimulated in a variety of ways, any of which may be usefully employed in this aspect of the invention. In specific embodiments, ER stress and the unfolded protein response is stimulated by treatment of the cells with tunicamycin and thapsigargin, either separately or in combination. In other embodiments, ER stress and the unfolded protein response is stimulated by treatment of the cells with hypoxic conditions. Other methods to cause ER stress and the unfolded protein response are likewise considered within the scope of this aspect of the invention.

The methods of the instant invention may in some embodiments include the step of counterscreening the library of compounds to identify those compounds that are not toxic to cells grown in the absence of ER stress or the unfolded protein response. Such compounds would be expected to display an increased specificity of activity toward cells associated with a disease in which the UPR has been implicated, such as, for example, cancer, conformational diseases, diabetes, cardiovascular disease, atherosclerosis, viral infection, and cerebrovascular disease. Compounds may be screened on cells grown in the absence of ER stress or the unfolded protein response by, for example, omitting tunicamycin and/or thapsigargin from the culture media. Alternatively, or in combination, compounds may be screened, for example, on cells grown in air. Compounds displaying stronger inhibitory activity toward cells subjected to ER stress and the unfolded protein response and low toxicity toward cells in the absence of ER stress and the unfolded protein response would generally be of most interest for use as therapeutics.

The methods of the instant invention may likewise in some embodiments include the step of counterscreening the library of compounds to identify those compounds that inhibit detection of the detectable protein. Such compounds could, for example, inhibit the activity of an enzyme or quench the fluorescence of a fluorescent protein used as the detectable protein. Alternatively, such compounds could, for example, inhibit the binding of an antibody to the detectable protein and thus inhibit the detection of the protein. Such counterscreens may, in some cases, make use of a second array of cells that constituitively express the detectable protein. Any effects of a compound on the detection of the detectable protein could therefore be identified in these cells. Such effects could be considered in assessing the therapeutic potential of the compound.

As described above, inhibitor compounds identified using the methods of the instant invention may be usefully employed in the treatment of disorders in which the UPR has been implicated, such as, for example, cancers characterized by the presence of moderate to severe hypoxia. Cells used in the instant methods are therefore preferably cultured cells from tissues affected by these disorders, such as, for example, cultured cancer cells. In particular, the cells of the instant methods may include cells from solid tumors and secretory cell malignancies, including multiple myeloma. Examples of solid tumor cells useful in the methods of the instant invention include sarcomas, carcinomas, and lymphomas. Other cultured cancer cells that may be usefully employed in the methods of the instant invention include, for example, cervix, brain, pancreas, breast, head and neck, prostate cancers, soft tissue sarcomas, primary and metastatic liver cancer, primary and metastatic lung cancer, esophageal cancer, colorectal cancer, lymphoma, and leukemia. In specific embodiments of the invention, the cells are fibrosarcoma cells or adenocarcinoma cells.

It will be readily apparent to one of ordinary skill in the relevant arts that other suitable modifications and adaptations to the methods and applications described herein may be made without departing from the scope of the invention or any embodiment thereof. Having now described the present invention in detail, the same will be more clearly understood by reference to the following Examples, which are included herewith for purposes of illustration only and are not intended to be limiting of the invention.

EXAMPLES

Example 1

Involvement of XBP-1 in Hypoxia and Tumor Growth

We have demonstrated that UPR related genes represent a major class of genes that are transcriptionally induced under hypoxia, that XBP-1 is activated during hypoxia in a HIF-1 independent manner, and that cell survival and apoptosis under hypoxia was mediated by XBP-1 (Romero L., et al. Cancer Res. 64:5943-5947, 2004). We have demonstrated that XBP-1 is essential for tumor growth. We implanted spontaneously transformed XBP-1 wild-type and knockout mouse embryonic fibroblasts (MEFs) as tumor xenografts into SCID mice and found that XBP-1 knockout MEFs were completely unable to grow as tumors. Furthermore, tumor growth was dependent upon the spliced form of XBP-1. We transfected spliced XBP-1 (XBP1s) into XBP-1 knockout MEFs and were able to restore the growth rate of these tumors back to that of the wild-type cells. We also transfected a mutant form of unspliced XBP-1 (XBP1u) in which the splice site was deleted. Transfection of this construct resulted in expression of an “unspliceable” form of XBP-1. Reintroduction of XBP1u into an XBP-1 null background was not able to restore tumor growth. These studies indicate that the spliced (activated) form of XBP-1 is a critical component of tumor growth. We obtained similar results using HT1080 cells overexpressing mutants of IRE1 in which either the kinase domain was deleted (IRE1ΔC) or both the kinase and endonuclease domain were deleted (IRE1ΔEn). Both of these deletion mutants were found to be defective in XBP-1 splicing and transactivation of a UPRE reporter.

Furthermore, we observed that tumor growth was impaired in tumor cells expressing IRE1 deletion mutants or an XBP-1 dominant negative (overexpression of mutant XBP-1 in which the transactivation domain was deleted). Conversely, hypoxia survival was increased and tumor growth was accelerated when the spliced form of XBP-1 was overexpressed. Taken together, these data strongly indicate that XBP-1 is an important regulator of tumor growth.

To further investigate the role of XBP-1 on tumor growth, we have developed an HT1080 cell line in which XBP-1 expression was regulated using a tetracycline inducible XBP-1 shRNA expression vector. In these cells, XBP-1 expression was inhibited in the presence of doxycycline, allowing us to determine the effect of inhibiting XBP-1 on an established tumor. In these experiments, doxycycline was added into the drinking water of tumor bearing mice when the tumors reached a size of 50-100 mm³. In the presence of doxycycline, there was a significant delay in the growth of these tumors as compared to the controls. We observed even greater tumor growth delay with constitutive inhibition of XBP-1 by shRNA. We also obtained similar results when XBP-1 was inhibited in a dominant negative manner in both an inducible and constitutively expressed manner. From these experiments, we concluded that XBP-1 plays a critical role in tumor growth and inhibition of XBP-1 is a may therefore be an effective therapeutic strategy.

To validate the clinical significance of XBP-1 as a potential therapeutic target in pancreatic tumors, we performed immunohistochemical analysis on 30 pancreatic tumor specimens taken from consecutive surgical specimens, 30 surrounding stroma samples, 29 chronic pancreatitis samples, and twenty normal pancreas samples. We have previously reported on the oxygenation status of a subset of these pancreatic tumors and found that they were extremely hypoxic while the normal adjacent pancreas was well-oxygenated (Koong A., et al. Int. J. Radiat. Oncol. Biol. Phys. 48:919-922, 2000). Because they are so profoundly hypoxic, pancreatic tumors are ideal tumors for the development of hypoxia targeted therapies. For these studies, we generated an affinity purified peptide antibody that was specific for the spliced form of human XBP-1. The strongest XBP1s expression was observed in the pancreatic tumor with minimal expression in the surrounding stroma or normal pancreas.

Collectively, these data demonstrate that the spliced form of XBP-1 (XBP1s) is essential for tumor growth, important for survival during hypoxia, and overexpressed in human pancreatic tumors. These observations strongly indicate that inhibition of XBP-1 is a promising therapeutic strategy.

Example 2

Identification of Inhibitors of XBP-1 Splicing

A high throughput screen for small molecule inhibitors of IRE1 activity was developed as detailed below. The sequence for XBP-1 is described in, for example, Liou, H-C. et al. Science 247:1581-1584, 1990; and Yoshimura, T. et al. EMBO J. 9:2537-2542, 1990. The amino acid sequence for unspliced XBP-1 protein is provided in SEQ ID NO: 1, with corresponding cDNA sequence being provided in SEQ ID NO: 3. The amino acid sequence for the spliced form is provided in SEQ ID NO: 2.

As shown in FIG. 2A, we developed a reporter construct in which luciferase was fused downstream and in frame with the unspliced form of XBP-1, containing the IRE-1 splice site. In the unspliced form, no luciferase is translated because of an endogenous stop codon. However, during hypoxia and ER stress, a 26 nt sequence is spliced out by IRE1 resulting in a frame-shift and read-through of the stop codon (Iwawaki et al., Nat. Med. 10:98-102, 2004). This results in production of an XBP1-luciferase fusion protein in which luciferase activity is detected only when XBP-1 is spliced by IRE1. This construct was stably transfected into HT1080 cells (human fibrosarcoma cell line). As shown in FIG. 2B, luciferase activity, detected after 24 hours of exposure to hypoxia, rapidly decreases when the HT1080 cells are allowed to reoxygenate, demonstrating that XBP-1 splicing is tightly controlled and largely restricted to hypoxic/ER stress conditions.

These tumor cells were used to screen a 66,000 chemically diverse small molecule library for inhibitors of XBP-1 splicing (Stanford High Throughput Facility compound library, which contains compounds from: SPECS & BioSPECS (Wakefield R.I.), Chembridge (San Diego, Calif.), and ChemRx libraries (Disclovery Partners International, San Diego, Calif.)). In this screen, we used two drugs, tunicamycin (“Tm”) (which blocks protein glycosylation) and thapsigargin (“Tg”) (an inhibitor of ER Ca-ATPase) that cause ER stress to activate the IRE1 reporter.

Specifically, HT1080 fibrosarcoma cells stably transfected with the unspliced XBP-1-luciferase reporter construct (3000/well) were plated onto a solid white 384 well microplate with a multidrop dispenser (40 μL per well). The plates were then placed into an automated incubator. After 24 hours of growth, a mixture of tunicamycin (1 μg/ml) and thapsigargin (100 nM) inducers were added, and candidate compounds were then added to the plates. After 24 hours, luciferase reagent (10 μl) was added to each well and the plates were read in a Molecular Devices Analyst GT (0.2 second read per well). Compounds that blocked IRE1 activation showed reduced levels of luciferase activity compared to control wells.

Compounds were selected for further investigation on the basis of their ability to block IRE1 reporter activation. In order to be selected, a compound must have demonstrated >95% inhibition of the reporter. Using this selection criteria, we selected the top 400 compounds for further testing. In this group, we performed a secondary screen comparing the ability of these compounds to inhibit IRE1-regulated luciferase activity without having an effect on CMV-regulated luciferase activity. From this analysis, we selected 58 compounds and repeated the IRE1 reporter screen on each compound individually.

This resulted in 38 compounds that were then tested individually in five separate cell based assays including the following: 1)>95% inhibition of hypoxia-activated XBP1-luciferase reporter; 2)>95% inhibition of tunicamycin activated XBP1-luciferase reporter; 3)>95% inhibition of hypoxia induced UPRE-luciferase reporter (multimer of unfolded protein response element which XBP-1 can transactivate); 4)>95% inhibition of tunicamycin induced UPRE-luciferase reporter; and 5) inhibition of XBP-1 splicing by RT-PCR. To qualify for further testing, each compound must have satisfied 4/5 of the conditions described above. A total of 18 compounds, referred to as candidate irestatins, met these criteria and were identified for further testing as described below. The structure of each of these compounds is shown in Table 1, above. A schematic of this screen is shown in FIG. 3.

A “heat map” view of a single plate from the primary screen is shown in FIG. 4. HT1080 cells stably expressing the XBP1-luciferase construct described above were plated in 384 well format (4,000 cells/well) and a different compound was added robotically into each individual well. Compounds were selected for further testing based upon demonstrating >95% inhibition of luciferase activity. The two lanes on the far left of FIG. 4 were negative controls (tunicamycin/thapsigargin alone) and the two lanes on the far right were positive controls (media alone).

FIG. 5A shows examples of compounds that were tested individually at 1 uM, 2 uM and 6 uM for inhibition of a UPRE-luciferase reporter following exposure to tunicamycin (Tm). In these studies, the luciferase reporter was under the control of 5 repeats of the XBP-1 promoter element (5X-UPRE). FIG. 5B shows compounds that were tested for inhibition of hypoxia (48 hours) induced transactivation of the same UPRE-luciferase report construct transiently transfected into HT1080 cells. More specifically, HT1080 fibrosarcoma cells transiently transfected with a luciferase reporter under the control of 5 repeats of the XBP-1 promoter element (5X-UPRE) were treated with 1 μM of each irestatin or left untreated, and incubated in normoxia or hypoxia (0.1% oxygen) for 48 hrs at 37° C. Cells were harvested, lysed in reporter lysis buffer, and assayed for luminescence using a luminometer. Fold induction is calculated as the luminesence in hypoxia divided by the normoxic luminescence value. The irestatin used is identified by a four-digit number below each bar.

Individual testing of the most promising compounds for inhibition of endogenous XBP-1 splicing (FIG. 6A) was also performed. In this assay, HT1080 cells were treated with hypoxia in the presence of various compounds and XBP-1 was amplified by RT-PCR. Not every compound inhibited XBP-1 splicing in this assay. Under aerobic conditions, only the unspliced form of XBP-1 XBP-1u) was detectable (lane 1). The spliced form of XBP-1 (XBP-1s) was detectable under hypoxia (lane 2). The ability of each individual compound to inhibit XBP-1 splicing was variable. In this set of compounds, only two were effective inhibitors of XBP-1 splicing (lanes 5 and 7). Interestingly, two compounds (lanes 3 and 4) resulted in inhibition of both the spliced and unspliced forms of XBP-1.

FIG. 6B shows the results of studies in which HT1080 fibrosarcoma cells stably expressing the XBP-luciferase reporter were treated with 1 uM of each irestatin or left untreated, and incubated in hypoxia (0.01% oxygen) for 48 hrs at 37° C. Cells were harvested, lysed in reporter lysis buffer, and assayed for luminescence using a luminometer.

Several of the candidate irestatins were tested in a hypoxia clonogenic survival assay. FIG. 7A is an example of some of the candidate irestatins that demonstrated selective sensitization of HT1080 cells to hypoxia. HT1080 fibrosarcoma cells stably were treated with 1 uM of the indicated irestatin or left untreated, and incubated in hypoxia (0.01% oxygen) for 48 hrs at 37° C. Cells were harvested and counted, and allowed to form colonies under normal oxygen tension. Survival rate is expressed as the fraction of colonies formed divided by the total number of cells seeded for each condition. For all experiments, cells were plated in triplicate, and all experiments were repeated at least three times. These experiments were repeated using MiaPaCa2 cells in place of the HT 1080 fibrosarcoma cells. As shown in FIG. 7B, the three compounds shown in FIG. 7A also sensitized MiaPaca2 cells to hypoxia, indicating that even though the screen was performed in HT1080 cells, the results may be generalized to other cell types.

FIG. 7C shows results of experiments demonstrating that candidate irestatins inhibit survival of human tumor cells in hypoxia. PANC1 pancreatic adenocarcinoma cells were treated with 1 uM of the indicated irestatin or left untreated, and incubated in hypoxia (0.01% oxygen) for 48 hrs at 37° C. Cells were harvested and counted, and allowed to form colonies under normal oxygen tension. After 10-11 days, colony formation was analyzed by staining with crystal violet.

FIG. 8 shows the results of studies in which HT1080 fibrosarcoma cells were treated with 1 uM of each Irestatin or left untreated, and incubated in hypoxia (0.01% oxygen) for 24 hrs at 37° C. Cells were harvested, lysed, and analyzed by Western blot using anti-XBP-1 antisera (lower panel) or anti-HIF-1 antisera (top panel) to confirm hypoxia exposure. The results confirm that the tested irestatins inhibit IRE1 signaling and XBP-1 splicing during hypoxia.

Example 3

Inhibition of XBP-1 Splicing in Tumors by Inhibitors of IRE1 Activity

Several nude mice were implanted with HT1080 cells stably expressing a XBP-1s-luciferase construct and XBP-1 activation was examined using bioluminescence imaging. Imaging was performed using the In Vivo Imaging System (IVIS, Xenogen Corporation, Alameda, Calif.) in the Stanford Center for Innovation in In Vivo Imaging (SCI3). This device consists of a cooled CCD camera mounted on a light-tight specimen chamber. In these experiments, two different potential irestatins (3281 & 5500) were injected IP into nude mice implanted with HT1080 stably expressing XBP1s-luciferase (described in FIG. 2A). We estimated that injecting mice at a concentration of 50 mg/kg (no apparent toxicity) was within a 10-fold range of the in vitro drug concentrations used (assuming uniform distribution and ignoring excretion/metabolism) for the above described cell culture assays.

As shown in FIGS. 9A-D, XBP-1 splicing activity was undetectable 8 hrs after irestatin 3281 injection and became detectable within 16 hrs later. Following a second injection, XBP-1 splicing was again inhibited after 8 hrs. These data strongly indicate that this compound had a direct effect on the inhibition of XBP-1 splicing, and may be effectively employed in the treatment of solid tumors. A second candidate irestatin (5500) was tested in the same manner and did not have any affect on XBP-1 splicing, at least at the time points assayed.

Example 4

Inhibition of Tumor Growth In Vivo by Inhibitors of IRE1 Activity

The ability of inhibitors of the inventive inhibitors of IRE1 activity to inhibit tumor growth in vivo was examined in a mouse model as follows.

PANC1 pancreatic adenocarcinoma cells were implanted subcutaneously into nude mice. Mice were then given a bolus injection of one of the inventive irestatins (1401, 9337, 3611 or 9389) at a dose of 60 mg/kg every 48 hours for a total of 5 doses, with 5-7 tumors being treated per group. As shown in FIG. 10, significant tumor growth was observed in untreated mice, but not in mice treated with the irestatins. These results indicate that the inventive irestatins may be effectively employed to inhibit tumor growth in vivo.

Example 5

Identification and Characterization of Potent Inhibitors of the IREE1α/XBP-1 Pathway

To date, the contribution of IRE1α to hypoxia tolerance and tumorigenesis has not been directly addressed and remains poorly understood. In this study, we employed a reverse chemical genetics approach to investigate the role of IRE1α in tumor growth. The use of small molecules to study protein function allows for the rapid and selective targeting of individual functions of multifunctional proteins, and serves as a powerful complement to conventional genetic strategies. Soderholm et al., Nat Chem Biol 2: 55-58 (2006). Indeed, genetic deletion in mice of IREla or XBP-1 causes embryonic lethality (Reimold et al., Genes Dev 14: 152-157 (2000); Harding et al., Mol Cell 7: 1153-1163 (2001)), and PERK and XBP-1 are required for the correct development of secretory organs such as the liver, pancreas and salivary gland (Lee et al., Embo J 24: 4368-4380 (2005); Zhang et al., Cell Metab 4: 491-497 (2006)). Thus, the UPR is necessary for the survival of tissues exposed to physiological levels of ER stress during fetal and postnatal development. The identification of small-molecule inhibitors provides an alternate strategy to inactivate IRE1α, enabling a functional analysis of this core UPR component in diverse cell types, including transformed cells cultured under hypoxia. This approach can also yield potential drug leads that may be utilized to address whether inactivation of a core UPR component can be tolerated in animals and applied as an antitumor strategy.

Materials and Methods

IRE1α Inhibitor Screen

As described above in Example 2, HT1080 fibrosarcoma cells stably expressing the XBP-luciferase reporter were plated in a 384 well microplate (4000 cells/well). After 24 hours, cells were treated with a mixture of tunicamycin (4 μg/ml) and thapsigargin (0.4 μM), followed by the addition of one compound per well (10 μM). We screened a total of 66,000 diverse molecules obtained from Specs, Chembridge and ChemRX. Twenty-four hours post-induction, BriteGlo luciferase substrate (10 μl) was added to each well and the signal intensity determined in a plate reader (0.2 s read per well). Hits were determined as compounds that significantly (>75%) inhibited activation of the XBP-luciferase signal by ER stress. We retested 431 compounds from the initial screen, and selected 58 compounds for additional analysis, including calculation of IC50 values and inhibition of a CMV-luciferase reporter. A total of 12 molecules, including irestatin 9389, exhibited potent and specific inhibition of IRE1α and were further characterized.

Plasmids, Cell Lines, and Antibodies

The human fibrosarcoma cell line HT1080 and myeloma cell line RPMI-8226 were obtained from American Type Culture Collection (ATCC, Manassas, Va.). Cells were maintained at 37° C. with 5% CO₂ in DMEM (HT1080) or RPMI 1640 media (RPMI-8226 cells) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin antibiotics. Rabbit polyclonal antisera raised against human XBP-1 and phospho-IRE1α were a gift from Dr. Fumihiko Urano (University of Massachusetts, Worcester, Mass.). Additional antibodies were obtained from the following commercial sources: Grp78 (Stressgen); IRE1α, ATF6, and CHOP/GADD153 (Santa Cruz Biotechnology, Santa Cruz, Calif.); Flag M2 monoclonal (Sigma, St. Louis, Mo.); cleaved caspase 3, JNK1 and phospho-JNK1 (Cell Signaling Technologies, Danvers, Mass.); HIF-1α (Novus Biologicals, Littleton, Colo.); (hypoxyprobe and anti-pimonidazole antibody kits (Chemicon, Temecula, Calif.).

To generate the XBP-luciferase reporter, N-terminally Flag-tagged, unspliced human XBP-1 (amino acids 1-208) was amplified by PCR using Pfx polymerase (Invitrogen, San Diego, Calif.). The PCR product was digested with EcoRI and BamHI, and subcloned into pEGFP-N1 (Clontech, Mountain View, Calif.) to generate pFlag-XBP1(1-208)-EGFP. This plasmid was subsequently digested with BamHI and Not I to remove EGFP. Firefly luciferase containing BamHI and Not I sites was amplified by PCR and subcloned downstream of XBP-1 such that luciferase is translated only in the ‘spliced’ reading frame. All constructs were verified by sequencing.

Immunoblotting

Cells (2×10⁶) were cultured in 10-cm dishes, collected using a cell scraper at 4° C., and lysed by addition of 150 μl cell lysis buffer [50 mM Tris pH 7.4, 150 mM NaCl, 10% glycerol, 0.5% Triton X-100. 0.5% NP-40, 2 mM Na₃VO₄, 20 mM beta-glycerophosphate, 10 mM NaF, 1 mM DTT, 1 mM PMSF). Lysates were centrifuged for 5 min at 10,000×g, and proteins (˜40 μg) were resolved by SDS-PAGE followed by semi-dry transfer to nitrocellulose membranes. Membranes were blocked in TBS-5% milk supplemented with 0.1% Tween-20. The blots were then probed overnight with relevant antibodies, washed, and incubated for 2 hours with species-specific secondary antibodies conjugated to horseradish peroxidase. After washing in block solution, immunoreactive material was detected by enhanced chemiluminescence (SuperSignal West Dura Extended, Pierce, Inc., Rockville, Ill.).

Reporter Assays

HT1080 cells stably expressing the XBP-luciferase construct were grown in 60 mm dishes to 60-70% confluency. Following hypoxia treatment, cells were washed twice with PBS, lysed in 400 μl 1× reporter lysis buffer (RLB) (Promega, Madison Wis.) for 30 min at 24° C. Lysates (100 μl) were mixed with an equal volume of luciferase substrate (Promega), and assayed using a luminometer. For 5×-UPRE-luciferase reporter assays, cells were co-transfected with the appropriate reporter plasmid and a control plasmid (pSV40-beta-gal) using Lipofectamine 2000 (Invitrogen, San Diego, Calif.). Twenty-four hours after transfection, fresh media was added, and cells were treated with Tm or shifted to hypoxia. After treatment, cells were lysed in 1× RLB and analyzed for luciferase activity as described above. Beta-galactosidase activity was determined using the beta-galactosidase enzyme assay system (Promega).

Northern Blots

Cells were cultured in 10 cm plates, harvested, and total RNA recovered with Trizol (Invitrogen, San Diego, Calif.). Total RNA (10 μg) was resolved on a 1% agarose-formaldehyde gel. ³²P-labeled probes were prepared using the Rediprime II random-prime labeling kit (GE-Amersham, Buckinghamshire, UK). The primers used to PCR amplify probes are as follows. P58^IPK: 5′GTGGCCCCCGGCTCCGTGACCAGCCGGCTGGGCTCGGTA 3′ (SEQ ID NO: 4); 5′ ACGCTTCAGTATTATCATTCTTCAACTTTGACGCAGCTTT 3′ (SEQ ID NO: 5). DER-1: 5′ GTCGGACATCGGAGACTGGTTCAGGAGCATCCCGGCGAT 3′ (SEQ ID NO: 6); 5′TCCTACTGGGCAGCCAGCGGTACAAAAACTGAGGGTGTGG 3′ (SEQ ID NO: 7). Blots were incubated with probe overnight, washed three times in 2×SSC/0.2% SDS, dried, and exposed to a phosphorimager screen overnight. Images were analyzed using ImageQuant software (Molecular Dynamics).

Ribonuclease Assay

The in vitro ribonuclease assays were carried out using purified IRE1α-cyto essentially as described. Gonzalez and Walter, Methods Mol Biol 160: 25-36 (2001); Gonzalez et al., Embo J 18: 3119-3132 (1999). For each reaction, 5 μg purified IRE1α-cyto was incubated with 300 ng of fluorescein-labeled RNA stem-loop substrate at 37° C. in a total volume of 300 μl. Aliquots (50 μl) were withdrawn at the indicated times and mixed with an equal volume of stop solution. Id. Reactions were analyzed by SDS-PAGE using 10-20% acrylamide gradient gels. The sequence for the hXBP-1 3′ RNA stem-loop substrate is as follows: 5′CAGCACUCAGACUACGUGCACCUCUGCAGCAGGUGCAGGCCCAGUU G 3′ (SEQ ID NO: 8). For the RNAse A cleavage assay, 300 ng of labeled XBP-1 RNA substrate were incubated with 1 ng bovine RNAse A (Sigma) in the presence of RNAsin (40 units), irestatin 9389 (2 μM) or DMSO vehicle control at 30° C. for the indicated times.

Mouse Immunohistochemistry and Histopathology

Tumor-bearing mice were injected i.p. with hypoxyprobe (50 mg/kg) 1 hour prior to sacrifice. Mice were euthanised under anesthesia by cervical dislocation, and tumors were surgically resected, embedded in OCT compound (Sakura Tissue Tek), and frozen at −80° C. Tumors were sectioned at 8 μm, fixed in 4% paraformaldehyde, and blocked in PBS-4% BSA. Tissue sections were incubated overnight in block solution containing antisera specific for hypoxyprobe (1:250) and cleaved caspase-3 (1:400). Slides were washed three times with block solution and incubated for 2 hours at room temperature with anti-mouse Alexa 488 or anti-rabbit Alexa 594 (Invitrogen, San Diego, Calif.). Slides were washed five times in block solution, and coverslips mounted with Permount supplemented with DAPI.

Complete blood counts (CBC's) and clinical chemistry panels were performed on blood obtained by cardiac puncture after euthanasia with CO₂. Gross necropsies were performed, all major viscera were harvested, fixed in 10% buffered neutral formalin, routinely processed for paraffin embedding, and stained with hematoxylin and eosin (H&E). Sections were analyzed by a board-certified veterinary pathologist (DMB).

Clonogenic Survival Assays

For hypoxia survival assays, cells were grown in 60 mm dishes until reaching at 50-70% confluence and shifted to hypoxia (0.1% O₂) for 48 hrs. Cells were trypsinized, counted using a hemocytometer, and replated in triplicate at 1,000-20,000 cells per plate in normal culture medium. After 10-12 days of growth under normal oxygen conditions, colonies were stained with 0.2% crystal violet in ethanol and counted. Survival values are expressed as the number of colonies divided by the total number of cells seeded for each condition, normalized to the plating efficiency under normal oxygen conditions. At least three independent experiments were performed.

Tumor Xenografts

Female 4-6 week-old SCID (B6.CB17) mice supplied by Stanford University Animal Facility were housed in the same facility (American Association of Laboratory Animal Care-approved) with 12 hour light cycles. Food and water were provided ad libitum. All experiments were approved by the institutional care and use committee. The potential toxicities of irestatin 9389 were examined in SCID mice injected i.p. once daily over 4 consecutive days with increasing doses of irestatin 9389 or vehicle control. A dosing regimen of 50-60 mg/kg, equal to 75% of the LD50 value, resulted in robust inhibition of IRE1α function without apparent toxicity. For xenografts, 2×10⁶ HT1080 fibrosarcoma cells were resuspended in 50-75 μl PBS and injected s.c. in the dorsal flanks of host mice. When the implanted tumors reached a mean volume of ˜150 mm³, mice were randomly assigned into different treatment groups. Mice were dosed by i.p. bolus injection with either vehicle (50% DMSO, 20% cremophor EL, 30% ethanol) or irestatin 9389 (50 mg/kg). Tumors (6-8 per group) were measured every 2-4 days with calipers. Tumor volume was calculated using the formula [(W²×L) 0.52] where W=width and L=length.

In Vivo Bioluminescence Imaging

HT1080 fibrosarcoma cells (2×10⁶) stably expressing the XBP-luciferase reporter were implanted s.c. into severe combined immune deficient (SCID) mice. Ten minutes prior to imaging, mice were injected i.p. with D-luciferin (150 mg/kg) solubilized in PBS. Optical bioluminescence imaging was performed using the IVIS charged-coupled device camera system (Caliper Life Sciences, Hopkinton, Mass.). Mice were imaged for 1-4 minutes per acquisition scan. Signal intensities were analyzed using Living Image software (Caliper).

Results and Discussion

FIG. 11 shows the identification of Irestatin 9389 as a potent inhibitor of the IRE1α/XBP-1 pathway. A. XBP-luciferase reporter construct. Firefly luciferase was inserted downstream of the IRE1α splice site in human XBP-1 to enable the conditional translation of luciferase under ER stress in an IRE1α-dependent manner. B. Selective inhibition of the XBP-luciferase reporter by irestatin 9389. HT1080 human fibrosarcoma cells stably expressing the XBP-luciferase reporter or CMV-luciferase were cultured in the presence of Tm (4 μg/ml) and Tg (0.4 μM) and irestatin 9389 at the indicated concentrations. After 24 hours, luciferase activity was analyzed in an automated plate reader. For each cell line, values are expressed as the percent inhibition of the median for Tm/Tg-treated wells, corrected for background. C. Structure of irestatin 9389. D. XBP-luciferase reporter assay. HT1080 cells stably expressing the XBP-luciferase reporter were exposed to Tm (4 μg/ml) for 24 hours or hypoxia (0.1% oxygen) for 24 or 48 hours, in the presence of DMSO or irestatin 9389 (1 μM) as indicated. Values are expressed as the fold increases over uninduced levels. E. 5X-UPRE reporter assay. HT1080 cells were co-transfected with 5X-UPRE luciferase and SV40-beta-gal reporter plasmids, followed by exposure to Tm or hypoxia as in D. For each condition, luciferase activity is normalized to beta-galactosidase expression levels as an internal control for transfection efficiency. F. Western immunoblot analysis of XBP-1s. HT1080 cells were left untreated (lane 2) or exposed to Tm (4 μg/ml) for 20 hours in the presence of DMSO vehicle (lane 1) or the indicated irestatins (2 μM; lanes 3-6). Cell lysates were resolved by SDS-PAGE and immunoblotting using antisera specific for XBP-1s (top panel) or actin and GAPDH (bottom panel) as loading controls. G. Irestatin 9389 blocks the accumulation of XBP-1s under hypoxic conditions. HT1080 cells were treated with DMSO or exposed to irestatin 9389 (2 μM; lane 3) in normoxia (N) or under hypoxia for 24 hours (H 24; lanes 2,3). Cells were harvested, lysed, and analyzed by immunoblotting with antisera specific for HIF-1α (top), XBP-1s (middle) or actin (bottom). H. Northern blot analysis of XBP-1s transcription targets. Cells were exposed to Tm (4 μg/ml) or hypoxia for 24 hours (H 24) in the absence or presence of irestatin 9389 (2 μM). Total RNA was analyzed by Northern blotting using radiolabeled probes specific for P58^IPK or DER-1. Total rRNA is shown as loading control.

FIG. 12 shows that irestatin 9389 inhibits the endonuclease function of IRE1α. A. Irestatin 9389 does not modulate the expression of Grp78. HT1080 cells were exposed to DMSO vehicle (lane 1), irestatin 9389 (2.5 μM; lane 2) for 16 hours or Tm (5 μg/ml; lane 3) for 8 hours. Following treatments, cells were harvested, lysed, and analyzed by immunoblotting using anti-Grp78 antibody (top) or anti-actin (bottom) as a loading control. B. Effect of irestatin 9389 on IRE1α expression and kinase function. HT1080 cells were preincubated for 16 hours with either vehicle or irestatin 9389 (2.5 μM), followed by addition of Tm (5 μg/ml) for the indicated times. Cell lysates were analyzed by Western immunoblotting using anti-IRE1α (bottom) or anti-phospho-IRE1α antibodies (top). C. Effect of irestatins on JNK1 activation under ER stress. HT1080 cells were untreated (lane 1), exposed to TNF-α (10 ng/ml, 10 min), or Tm (4 μg/ml, 1.5 hrs) (lanes 3-8) following a 2 hour preincubation in the presence of vehicle (lane 3) or the indicated irestatins (2.5 μM; lanes 4-8). Lysates were analyzed by Western blot using antisera specific for phospho-JNK1 (top) or total JNK1 (bottom). D. Purification of IRE1α-cytosolic. 6x-His-IRE1α-cyto containing the IRE1α kinase and endonuclease was expressed in bacteria (lane 1) and isolated by Nickel resin affinity chromatography to >95% purity (lane 2). E. IRE1α endonuclease assay. Fluorescein end-labeled RNA minisubstrate (300 ng) corresponding to the downstream (3′) human XBP-1 intron-exon cleavage site was incubated in the absence (lanes 1-3) or presence (lanes 4-9) of purified His6-IRE1α-cyto (5 μg), and exposed to either vehicle or irestatin 9389 (2.5 μM). The reactions were stopped at the indicated times and reaction aliquots were resolved by SDS-PAGE and visualized by UV illumination. F. Quantification of RNA cleavage kinetics. Results represent the mean from 3 independent experiments +/−SEM. G. RNAse A activity assay. Labeled XBP-1 RNA minisubstrate (300 ng) was exposed for the indicated times to RNAse A (1 ng) in the presence of either RNAse inhibitor (40 units), irestatin 9389 (2.5 μM), or vehicle only for the indicated times. Samples were analyzed as in (E).

FIG. 13 shows that exposure to irestatin 9389 induces apoptosis and impairs cell survival under hypoxia and ER stress. A. Effect of irestatin 9389 on PERK and ATF6 pathways. HT1080 cells were treated with vehicle alone (lanes 1-4) or 2.5 μM irestatin 9389 (lanes 5-8) and cultured under aerobic conditions for 18 hours (N) or shifted to hypoxia for the indicated times. Protein lysates were analyzed by Western blot analysis using antisera specific for ATF6 (top), CHOP/GADD153 (middle) or actin (bottom). Arrow indicates the cleaved, transcriptionally active form of ATF6. B. Cleavage of caspase-3 in irestatin-treated cells under hypoxia. HT1080 cells were cultured in normoxia (N) or under hypoxia for 36 hours (H 36) in absence or presence of irestatin 9389 (2.5 μM). Arrows indicate proteolytically cleaved caspase-3. C. Colony formation assay. HT1080 cells were treated as in B under normoxia (N) or hypoxia for 48 hours (H 48). Cells were harvested, counted, and allowed to grow under normal culture conditions for 11-12 days. Colonies were visualized with crystal violet staining. D. Quantification of clonogenic survival assay. Values represent the mean+/−SEM from at least 4 independent experiments. E. TUNEL staining of cells treated as in C. F. Quantification of TUNEL-positive cells. Values represent the mean+/−SEM from at least 3 experiments. G. HT1080 tet-off Flag-XBP-1s cells were cultured in the presence or absence of dox (1 μg/ml), followed by lysis and anti-Flag immunoblot. H. Rescue of irestatin-mediated cell death by enforced expression of XBP-1s. Tet-off XBP-1s cells were cultured with or without irestatin 9389 (2.5 μM) in the absence or presence of dox, under hypoxia for 48 hours (H 48). Cells were processed as in C, and colonies were visualized with crystal violet staining. I. Cell proliferation assays. Equal numbers (1×10⁵) of HT1080 fibrosarcoma (left) or RPMI 8226 myeloma cells (right) were seeded on day 0, and cultured in the presence of vehicle control or irestatin 9389 at the indicated concentrations. Cells were harvested at the indicated times and counted by hemocytometer. Values represent the mean calculated from triplicate experiments +/−SEM.

FIG. 14 shows the in vivo antitumor activity of irestatin 9389. A. Irestatin 9389 impairs IREla activity in implanted tumor xenografts. Equal numbers (2×10⁶) of XBP-luciferase or CMV-luciferase reporter cells were implanted s.c. into SCID mice. After one week, mice were treated with irestatin 9389 (50 mg/kg), followed by optical bioluminescence imaging. B. Inhibition of tumor growth by irestatin 9389. HT1080 s.c. tumor xenografts were established in SCID mice and allowed to reach a mean volume of 150 mm³ before treatment. Irestatin 9389 (50 mg/kg) or vehicle control was administered q 2d by i.p. injection and continued for 2 weeks, for a total of 6 doses. Tumor volumes were calculated based on caliper measurements taken every 3-5 days. Data points represent the mean volume calculated from at least 7 tumors per group, with SEM shown in one direction. Mean mouse weights +/−SEM are shown in bottom graph. C.

Immunohistochemical analysis of tumor xenografts. Tissue sections prepared from cryo-preserved tumors following 3 doses with either vehicle control or irestatin 9389 were immunostained using hypoxyprobe (pimonidazole) or antisera specific for cleaved caspase-3. D. Quantification of tumor immunohistochemistry. At least 8 randomly chosen fields (>300 cells/field) per tumor were scored for pimonidazole and cleaved caspase-3 staining. A minimum of 3 tumors (250-300 mm³ at harvest) were analyzed per treatment group. Values represent mean+/−SEM.

FIG. 15 shows the expression of XBP-1s in human pancreas tissue specimens. Tissues surgically recovered from normal pancreas, chronic pancreatitis, or pancreatic tumors were sectioned and stained using antisera specific for XBP-1s (400× magnification). Images were scored on the basis of staining intensity and quantified as shown in the table.

FIG. 16 shows the histopathological analysis of mouse pancreas and liver tissues. Pancreas and liver specimens recovered from mice treated with three doses of either vehicle (top) or irestatin 9389 (50 mg/kg; bottom) were sectioned and stained with hematoxylin and eosin.

As described above, a HT1080 fibrosarcoma cell line stably expressing a fusion of unprocessed XBP-1 inserted upstream of firefly luciferase has been developed to identify small molecule inhibitors of the IRE1α/XBP-1 signaling module. Under ER stress conditions, IRE1α catalyzes the removal of a 26-nt intronic sequence from the XBP-1 mRNA, introducing a shift in reading frame that allows for the translation of luciferase (FIG. 11A). We screened a chemical library of 66,000 small molecules for inhibitors of XBP-luciferase activity stimulated by incubation of the reporter cell line with a mixture of tunicamycin and thapsigargin, two mechanistically distinct chemical inducers of ER stress. We also utilized a counterscreen consisting of HT1080 cells stably expressing a constitutively-expressed, CMV promoter-driven luciferase construct to exclude agents that caused non-specific inhibition of luciferase activity. We identified 12 molecules, termed irestatins, which consistently inhibited the IRE1α/XBP-1 signaling module without significantly affecting the activity of CMV-luciferase. We pursued several of the most potent irestatins, and describe here in detail our analysis of irestatin 9389, which inhibited XBP-luciferase activity with mean inhibitory concentration (IC50) of ˜25 nM (FIG. 11B). The structure of this molecule is shown in FIG. 11C.

To determine if irestatin 9389 impairs IRE1α/XBP-1 signaling triggered by oxygen deprivation, we cultured XBP-luciferase reporter cells for 24 or 48 hours under hypoxia (<0.1% oxygen) in the absence or presence of irestatin 9389 (1 μM), and then assayed for luciferase activity. As a separate control, cells were also treated with Tm for 24 hours, which increased luciferase activity by 60-fold (FIG. 11D). As expected, exposure to irestatin 9389 inhibited Tm-mediated activation of the reporter by more than 90%. Exposure to irestatin 9389 also diminished activation of the XBP-luciferase reporter under hypoxia for 24 or 48 hours. Whereas control (DMSO-treated) cells increased XBP-luciferase activity by 95-fold after 48 hours of hypoxia, the addition of irestatin 9389 robustly inhibited this response (FIG. 11D, right panel).

Since these assays employed a chimeric XBP-luciferase substrate, we next determined whether irestatin 9389 could inhibit the function of endogenous XBP-1s. HT1080 cells were transfected with a firefly luciferase reporter under the transcriptional control of 5 tandem repeats of the unfolded protein response element (5X-UPRE), a canonical DNA binding site for XBP-1s identified in the promoter regions of XBP-1 target genes. Yoshida et al., Molecular & Cellular Biology 20: 6755-6767 (2000); Yamamoto et al., Journal of Biochemistry 136: 343-350 (2004). Following exposure to Tm, luciferase activity increased by ˜12-fold over untreated cells, while cells exposed to both Tm and irestatin 9389 exhibited less than a 4-fold induction (FIG. 11E). Irestatin 9389 also robustly inhibited UPRE promoter activity under hypoxic conditions. After 48 hours of hypoxia, vehicle-treated cells increased luciferase activity by 170-fold, while the addition of irestatin 9389 diminished this response by more than 90% (FIG. 11E, right panel). In support of these findings, western immunoblot analysis demonstrated that irestatin 9389 blocked the accumulation of XBP-1s following treatment with Tm, while structurally unrelated irestatin candidates exhibited little or no effect (FIG. 11F, lanes 3-5). Similarly, irestatin 9389 decreased levels of XBP-1s following 24 hours of hypoxia (FIG. 11G), while the expression of HIF-1α, a hypoxia-inducible transcription factor that functions independently of the UPR (Romero-Ramirez et al., Cancer Research 64: 5943-5947 (2004)), was not affected by irestatin 9389 (FIG. 11G, top panel).

Gene expression profiling studies have identified several XBP-1-dependent target genes that are transcriptionally induced during ER stress. Lee et al., Molecular & Cellular Biology 23: 7448-7459 (2003). These include the DnaJ/Hsp40-like gene P58^IPK and DER-1, a component of the ERAD pathway. Oda et al., J Cell Biol 172: 383-393 (2006). To analyze the effect of irestatin 9389 on the expression of these genes, HT1080 cells were treated with Tm or cultured under hypoxia for 24 hours, followed by isolation of total RNA and Northern blot analysis. Expression of these key UPR genes increased significantly (>5-fold) under hypoxia or following treatment with Tm, while the addition of irestatin 9389 fully inhibited this response (FIG. 11H). We conclude that irestatin 9389 specifically blocks the production or accumulation of XBP-1s following ER stress and diminishes the expression of its downstream effectors.

We next sought to determine the mechanism by which irestatin 9389 inhibits IRE1α/XBP-1 function. We first examined if irestatin 9389 deregulates the expression of Grp78, thereby increasing the fraction of Grp78-bound IRE1α and raising the activation threshold for IRE1α. Liu et al., Journal of Biological Chemistry 277: 18346-18356 (2002); Zhou et al., Proc Natl Acad Sci USA 103: 14343-14348 (2006); Bertolotti et al., Nat Cell Biol 2: 326-332 (2000). HT1080 cells were incubated with vehicle or irestatin 9389 (2.5 μM) for 16 hours, followed by western immunoblot analysis using Grp78 antisera. As a positive control, cells were treated with Tm for 8 hours, which robustly induced Grp78 (FIG. 12A, lane 3). In contrast, irestatin 9389 had no effect on Grp78 levels (FIG. 12A) at 16 hours or following longer treatments of 24 or 36 hours (data not shown). Similarly, cells incubated in the presence of irestatin 9389 for 16 hours exhibited no significant changes in the total level of IRE1α, as judged by Western immunoblotting (FIG. 12B, lower panel).

Activation of IRE1α is preceded by ATP binding and autophosphorylation, and the IRE1α kinase is required for endonuclease activity. Tirasophon et al., Genes & Development 14:2725-2736 (2000). To determine if irestatin 9389 inhibits the IRE1α kinase, HT1080 cells were preincubated for 16 hours with irestatin or vehicle followed by addition of Tm to induce ER stress. Cells were then harvested at regular intervals, and activation of the IRE1α kinase was assessed by immunoblotting using anti-phospho-IRE1α antisera. In both control and irestatin-treated cells, the addition of Tm triggered a rapid increase in levels of phospho-IRE1α (FIG. 12B). Preincubation with irestatin 9389 also failed to block the phosphorylation of JNK1, a downstream effector of IRE1α kinase signaling (Urano et al., Science 287: 664-666 (2000)), during Tm-induced ER stress (FIG. 12C). Interestingly, several structurally unrelated irestatins strongly inhibited the IRE1α-dependent phosphorylation of JNK1 under ER stress (FIG. 12C, lanes 7-8), indicating that mechanistically distinct classes of irestatins were identified by the initial screen.

Next we determined whether irestatin 9389 inhibited the endonuclease function of IRE1α. To monitor endonuclease activity, we devised an in vitro ribonuclease assay in which a fluorescein labeled RNA hairpin corresponding to the 3′ intron-exon boundary of human XBP-1 serves as a cleavage substrate for the IRE1α nuclease. Because the isolated IRE1α endonuclease lacks significant catalytic activity (Dong et al., RNA 7: 361-373 (2001); Nock et al., Methods Enzymol 342: 3-10 (2001); D. F. and A. K., unpublished data), we expressed in bacteria and purified the full cytosolic portion of IRE1α (His6-IRE1α-cyto) containing both kinase and endonuclease domains (FIG. 12D). In the presence of ATP and purified His6-IRE1α-cyto, the XBP-1 target RNA sequence was efficiently cleaved, with a mean half-life of ˜25 minutes (FIG. 12E). Addition of irestatin 9389 (2.5 μM) to the reaction strongly inhibited cleavage (FIG. 12E). However, irestatin is not a general ribonuclease inhibitor, as a >100-fold molar excess of irestatin 9389 failed to inhibit degradation of the XBP-1 3′ intronic loop by RNAse A (FIG. 11G). Thus, irestatin 9389 functions as a selective inhibitor of the IRE1α endoribonuclease without impairing IRE1α kinase function.

Activation of IRE1α alleviates ER stress through the splice-activation of XBP-1 and by the co-translational cleavage of mRNAs encoding secreted proteins. Hollien and Weissman, Science 313: 104-107 (2006). To assess the impact of inhibiting IRE1α signaling on the cellular response to ER stress, we performed a kinetic analysis of the two other major UPR pathways, ATF6 and PERK, in hypoxic cells exposed to irestatin 9389. Treatment of hypoxic cells with irestatin 9389 significantly increased the proteolytic cleavage of ATF6 into its transcriptionally active 50 kDa form (FIG. 13A, top). Likewise, the expression of CHOP/GADD153, a downstream target of the PERK-ATF4 signaling module, was increased in irestatin-treated cells following exposure to hypoxia for 6-12 hours (FIG. 13A, middle panel). As persistent activation of the PERK-ATF4-CHOP signaling module triggers apoptotic cell death (McCullough et al., Molecular & Cellular Biology 21: 1249-1259 (2001); Yamaguchi and Wang, Journal of Biological Chemistry 279: 45495-45502 (2004); Marciniak et al., Genes & Development 18: 3066-3077 (2004); Boyce et al., Science 307: 935-939 (2005)), we also examined the activation of caspase-3, the major apoptotic effector caspase, in irestatin-treated cells. Whereas vehicle-treated cells exhibited minimal activation of caspase-3 after 36 hours of hypoxia, exposure to irestatin 9389 stimulated cleavage of caspase-3 (FIG. 13B, lanes 3-4). This effect was specific to hypoxia-stressed cells, as irestatin 9389 had no effect on caspase-3 processing in cells cultured under normal oxygen conditions (FIG. 13B, lanes 1-2). Taken together, these findings indicate that irestatin 9389 overwhelms the adaptive capacity of the UPR, leading to initiation of programmed cell death.

We corroborated these biochemical findings using colony formation assays as an indicator of cell viability. Addition of irestatin 9389 (2.5 μM) to the culture medium had a negligible effect on the survival of HT1080 cells cultured under normal oxygen conditions (FIG. 13C). However, in cells cultured under hypoxia for 48 hours, irestatin 9389 strongly inhibited colony formation (FIG. 13D). Exposure of hypoxic cells to irestatin 9389 for a shorter duration (hours 40-48 of hypoxia) also resulted in a 8-fold decrease in the rate of colony formation (data not shown). Consistent with the increased activation of caspase-3, treatment with irestatin 9389 significantly increased the proportion of hypoxic cells undergoing programmed cell death, as indicated by TUNEL-positive cells under hypoxia (FIG. 13E). After 48 hours of hypoxia, only 6% of vehicle-treated cells were TUNEL-positive, as compared with 35% of irestatin-treated cells (FIG. 13F).

To determine if the irestatin-mediated inhibition of IRE1α/XBP-1s pathway accounts for decreased viability under hypoxia, we generated a cell line in which Flag-tagged XBP-1s is expressed under the control of a tetracycline-regulated promoter. Cells cultured in the presence of doxicycline (dox, 1 μg/ml) do not express Flag-XBP-1s, while removal of dox restores robust expression of Flag-XBP-1s (FIG. 13G). In the presence of both dox and irestatin 9389 (2.5 μM), we again observed a significant (˜60 fold) decrease in viability following exposure to hypoxia for 48 hours. In contrast, the same concentration of irestatin 9389 had a minimal effect on the survival of hypoxic cells expressing Flag-XBP-1s (FIG. 13H). Thus, inhibition of the IRE1α/XBP-1s signaling module, and not an off-pathway effect of the irestatin, is primarily responsible for the poor survival of irestatin-treated tumor cells under hypoxia. Importantly, exposure to irestatin 9389 also strongly inhibited the growth of the myeloma cell line RPMI 8226, a secretory plasmacytoma, in a dose-dependent manner (FIG. 13I, right panel). In contrast, exposure to the same concentrations of irestatin 9389 had a negligible effect on the growth rate of HT1080 cells cultured under normal conditions (FIG. 13I, left panel). We conclude that irestatin 9389 selectively impairs the growth and survival of a variety of transformed cell types subjected to mechanistically distinct forms of ER stress.

The increased sensitivity of irestatin-treated cells to hypoxic stress in vitro indicate that selective inhibition of IRE1α signaling could impact tumor growth. In support of an active role for IRE1α in tumor growth, we found that >50% (16/30) of surgically resected human pancreatic adenocarcinoma specimens exhibited moderate or strong immunoreactivity for XBP-1s. In contrast, XBP-1s was not detected in normal pancreas specimens (0/20), and infrequently observed in chronic pancreatitis (1/29) (FIG. 15). To explore the effects to irestatin 9389 in vivo, we first established animal dosing parameters using real-time bioluminescence imaging of SCID mice that had been implanted subcutaneously (s.c.) with tumor cells stably expressing the XBP-luciferase reporter. Irestatin 9389 administered in single doses of 50-60 mg/kg robustly inhibited the XBP-luciferase reporter for 6-8 hours after the injection (FIG. 14A). The XBP-luciferase signal returned to basal levels by 24 hours after treatment. A complete blood count and analysis of blood chemistry indicated that 3-4 doses of irestatin 9389 (50 mg/kg), administered every other day, were well tolerated and did not result in significant impairment of kidney, liver, or bone marrow function (Table 3). Although IRE1α has been implicated in glucose tolerance (Lipson et al., Cell Metab 4: 245-254 (2006); Ozcan et al., Science 306: 457-461 (2004)), we found no significant difference in fasting blood glucose levels between irestatin- and vehicle-treated animals (Table 3). These findings are further supported by histopathological analysis of all major organs, which revealed no significant differences between the vehicle and irestatin treatment groups. (FIG. 16).

TABLE 3
Analysis of blood chemistry and cell composition.
	Vehicle	Irestatin 9389
	mean	SEM	mean	SEM
Chemistry Panel
Glucose mg/dL	112.5	20.56696	124.5	7.14
AST IU/L	107.6	22.92408	117.775	14.25
ALT IU/L	30	10.15513	29.4	6.68
Total Bilirubin mg/dL	0.525	0.287228	0.3	0
Cholesterol mg/dL	102.25	8.261356	102	8.8
Electrolyte Panel
Sodium mM	151.5	2.12132	152.25	1.89
Potassium mM	7.875	0.388909	7.5175	0.49
Chloride mM	116	1.414214	116.75	2.22
Carbon Dioxide mM	22.55	0.777817	25.075	0.71
Na/K Ratio mM	19.25	1.202082	20.325	1.36
Anion Gap mM	20.9	0.565685	17.975	0.71
Complete Blood Count
WBC K/uL	5.55	1.340398	5.19	1.23
RBC M/uL	9.8	0.583095	10.375	0.3
HGB gm/dL	13.75	0.818535	14.625	0.59
HCT %	43.9	2.946184	47	1.39
Platelets K/uL	574.5	159.9281	805.5	124.9
Vehicle-treated or irestatin-treated nude mice were euthanized with carbon dioxide, and a terminal cardiac blood draw performed. Blood was collected using a heparinzed syringe for CBC and clinical chemistries. Based on comparisons with the vehicle control mice, the only lesion that may be related to treatment is a mild leukopenia noted in both treated mice. The degree is mild and histologically, the bone marrow was not impacted.

Next, we tested if treatment with irestatin 9389 could have a direct impact on tumor growth. Equal numbers (2×10⁶) of HT1080 cells were injected in the flanks of nude mice and allowed to grow for 2 weeks until tumors reached a mean volume of ˜150 mm³. Mice were then randomly assigned into vehicle control or irestatin groups, and dosed by intraperitoneal (i.p.) injection of vehicle or irestatin 9389 (50 mg/kg) every other day for a total of 6 doses. Although this dosing regimen resulted in a transient inhibition of IRE1α, significant cytostatic antitumor effects were soon evident (FIG. 14B). The inhibition of tumor growth continued even after the final injection of irestatin 9389. One week after the last treatment, the mean volume of irestatin-treated tumors was significantly less than vehicle-treated tumors (1790+/−380 mm³ versus 480+/−210 mm³; P<0.01) (FIG. 14B). Irestatin-treated mice did not exhibit significant long-term weight loss compared to vehicle-treated mice (FIG. 14B, top).

We further examined tumors from control and irestatin-treated mice for differences in cell survival. In tumors treated with three doses of irestatin 9389 (50 mg/kg), we observed a significant increase in cleaved caspase-3, an indicator of apoptosis, relative to vehicle-treated controls (FIG. 14C). The increase in apoptosis was most pronounced in hypoxic tissue regions of tumors, as determined by co-immunoreactivity for pimonidazole adducts (FIG. 14C, bottom panel). Quantitative analysis of immunostained tumor sections indicated that, in vehicle-treated tumors, less than 15% of hypoxic cells were apoptotic, compared to nearly 45% in irestatin-treated tumors (FIG. 14D). Interestingly, some pimonidazole-negative areas also exhibited increased levels of apoptosis following treatment with irestatin 9389, indicating that ER stress or sensitivity to irestatin occurs in tissue regions that are not acutely hypoxic (FIG. 14D). Taken together, these observations indicate that transient intratumoral inhibition of the UPR can potentiate cell death and impair tumor growth.

Severe hypoxia triggers the accumulation of misfolded proteins in the ER (Koumenis et al., Molecular & Cellular Biology 22: 7405-7416 (2002)), a potentially lethal condition that is remedied through the action of the UPR. In this study, we sought to determine the function of the IRE1α branch of the UPR in cellular tolerance to hypoxia and tumor growth. We employed a chemical genetic strategy to identify inhibitors of this pathway, and obtained multiple, mechanistically distinct classes of irestatins, including molecules that selectively target either the IRE1α kinase or endonuclease. We found that selective inactivation of the IRE1α endonuclease critically incapacitates the adaptive capacity of the UPR, resulting in increased ER stress and cell death under hypoxia. Irestatins therefore define a novel category of ER stress-selective antitumor agents specifically targeted to the underlying physiological response of tumor cells to the tumor microenvironment.

Several reports have demonstrated an essential role for the UPR in embryonic development, raising the possibility that systemic application of UPR-targeting molecules could cause severe toxicity to normal tissues, particularly those with secretory function such as the pancreas and liver. Twakoshi et al., Immunological Reviews 194: 29-38 (2003); Reimold et al., Genes Dev 14: 152-157 (2000); Reimold et al., Nature 412: 300-307 (2001). However, we found that multiple bioactive doses of irestatin 9389 were well tolerated and did not result in acute injury to these organ systems, as indicated by analysis of blood chemistry and organ pathology. Without intending to be bound by theory, our observations are consistent with the finding that expression of XBP-1 in the liver rescues the embryonic lethality of XBP-1 deficient mice, indicating that most tissues can function adequately in the absence of this key UPR transcription factor. Lee et al., Embo J 24: 4368-4380 (2005). Likewise, deletion of PERK results in a multitude of developmental abnormalities, including hyperglycemia and atrophy of the exocrine pancreas. Harding et al., Mol Cell 7: 1153-1163 (2001). However, PERK is necessary for the development of insulin-secreting pancreatic beta cells specifically during the fetal and early neonatal period and is not required in adults to maintain beta cell functions or glucose homeostasis. Zhang et al., Cell Metab 4: 491-497 (2006). Without intending to be bound by theory, these findings indicate that the major UPR pathways are required in a subset of secretory tissues during temporally delimited developmental windows, and that inactivation of core UPR signaling modules using drug-like molecules can be well tolerated in mature animals.

Although individual UPR pathways are dispensable under most circumstances, we found that pharmacological inhibition of IRE1α significantly impaired the growth of implanted tumors. This finding reinforces the idea that tumors are subjected to significantly elevated levels of ER stress relative to the surrounding normal tissues, a condition that may arise through the distinct contrasts in oxygenation status between normal tissues and solid tumors. Hockel and Vaupel, Seminars in Oncology 28: 36-41 (2001); Vaupel et al., Methods in Enzymology 381: 335-354 (2004). Without intending to be bound by theory, the antitumor effects of irestatin 9389 are consistent with a report demonstrating that inhibition of UPR target gene expression during glucose-deprivation can impair tumor growth. Park et al., Journal of the National Cancer Institute 96: 1300-1310 (2004). Without intending to be bound by theory, the rate of tumor growth may be naturally constrained by the severity of ER stress and by the capacity of the UPR to restore cellular homeostasis. Inhibition of this response induces proteotoxicity in hypoxic tumor cells, as indicated by the increased output of parallel UPR pathways downstream of ATF6 and PERK following treatment with irestatin 9389. In support of this model, irestatin 9389 potently blocks the induction of the XBP-1 targets DER-1 and P58^IPK, essential components of the ERAD machinery that mediate clearance of misfolded proteins from the ER. Ye et al., Nature 429: 841-847 (2004); Oyadomari et al., Cell 126: 727-739 (2006).

The pharmacological induction of ER proteotoxicity represents an effective therapeutic strategy in the treatment of solid tumors or secretory cell malignancies such as multiple myeloma, in which the UPR sustains cell viability under conditions of elevated secretory output. Iwakoshi et al., Nat Immunol 4: 321-329 (2003). Without intending to be bound by theory, since activation of the UPR can confer drug resistance to cancer cells (Gray et al., Mol Pharmacol 68: 1699-1707 (2005); Li and Lee, Curr Mol Med 6: 45-54 (2006)), our findings indicate that coordinated treatment with UPR-targeting agents may potentiate the efficacy of conventional chemotherapies. Inhibition of the UPR may also sensitize tumors to vascular targeting agents or anti-angiogenic drugs, which increase the fraction of hypoxic or nutrient-deprived tumor tissues (El-Emir et al., Eur J Cancer 41: 799-806 (2005); Boyle and Travers, Anticancer Agents Med Chem 6: 281-286 (2006); Dong et al., Cancer Research 65: 5785-5791 (2005)), or to radiation therapy, which preferentially kills oxygenated cell populations (Vaupel et al., Medical Oncology 18: 243-259 (2001); Vaupel et al., Seminars in Oncology 28: 29-35 (2001)). Likewise, proteasome inhibitors such as bortezomib (Velcade) have been shown to cause ER stress, while also inhibiting the UPR. Lee et al., Proceedings of the National Academy of Sciences of the United States of America 100: 9946-9951 (2003); Nawrocki et al., Cancer Res 65: 11510-11519 (2005); Obeng et al., Blood 107: 4907-4916 (2006). A combination of an irestatin and one or more proteasome inhibitors may exhaust the protective capacity of the UPR, pushing tumor cells into a decompensated state and ultimately cell death.

Example 6

Activity of Irestatins with 9389-Like Structure

Compounds of the screening library with structural similarity to compound 9389 (see Table 1) have been identified and in some cases further assayed for inhibitory activity. See Table 4. Compounds listed with “IC50” values were assayed secondarily after initially being identified in the high throughput screen. Each value represents a separate calculation of reporter inhibition based upon the high throughput robotic screening platform. The actual IC50 values are calculated and represent an estimate of the potency of each compound. This assay is not considered to be accurate below a concentration of 10 nM. Compounds classified with “mild” activity inhibited the XBP1-luciferase reporter by 10-30%. Compounds classified with “moderate” activity inhibited the XBP1-luciferase reporter by 30-75%. Compounds classified with “potent” activity inhibited the XBP1-luciferase reporter by 75-100%. Compounds classified with “undetected” activity inhibited the XBP1-luciferase reporter by less than 10% under the defined conditions.

Compounds with activities classified as “undetected” in Table 4 were identified by manual review of the structures of compounds reportedly present in the chemical libraries. Compounds displaying at least some structural similarity to the compounds with demonstrated activity are shown. The presence of these compounds in the assays has not been independently confirmed, however, so a lack of detectable activity may not necessarily be due to a compound's lack of activity.

TABLE 4
Activities of compounds having structural similarity to Compound 9389.
			IC50	Conc	%	Activity
Compound	STRUCTURE	Assay	(uM)	(uM)	Inh	Class
1567		HTS		10	−41.2	undetected
2399		HTS		10	13.3	mild
3290		HTS		10	−30.3	undetected
1491		HTS HTS		10 10	11.0 63.4	mild
1740		HTS HTS		10 10	25.1 5.9	mild
2750		HTS HTS		10 10	11.7 16.6	mild
4335		IRE IC50 IRE IC50	0.09   6.30	20   12	67.4   70.4	moderate
5500		IRE IC50 IRE IC50	0.06     0.000048	20   20	100.4    104.4	potent
8878		IRE IC50 IRE IC50	0.023   5.14	20   20	72.4   50.0	moderate
2853		HTS		10	26.5	mild
3371		IRE IC50	13.90	20	72.6	moderate
3398		HTS		10	−56.2	undetected
4645		HTS		10	−8.3	undetected
4950		HTS		10	−6.2	undetected
6392		HTS		10	2.7	undetected
6451		HTS		10	−55.6	undetected
8233		HTS		10	−59.6	undetected
8920		HTS		10	25.7	mild
9165		HTS		10	−6.5	undetected
9388		HTS		10	−40.8	undetected
9389		IRE IC50 IRE IC50	0.0063    0.031	20   20	87.1   100.3	potent
9668		HTS		10	19.0	mild
9766		HTS		10	26.7	mild
9787		HTS		10	−122.3	undetected
0040		HTS		10	−4.6	undetected
0069		HTS		10	−5.4	undetected
6068		HTS		12.3	5.8	undetected

From the foregoing, it will be appreciated that, although specific embodiments of the invention have been described herein for the purpose of illustration, various modifications may be made without deviating from the spirit and scope of the invention.

All references disclosed herein, including patent references and non-patent references, are hereby incorporated by reference in their entirety as if each was incorporated individually.

Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, numerous equivalents to the specific method and reagents described herein. Such equivalents are considered to be within the scope of this invention and are covered by the following claims.

Claims

1. A method to identify an inhibitor of the unfolded protein response comprising the steps of:

providing a first array of cells that stably express an mRNA fusion sequence, wherein the mRNA fusion sequence comprises a first mRNA segment comprising an unprocessed XBP-1 transcription factor gene sequence and a second mRNA segment comprising a reporter gene sequence, and wherein the first mRNA segment is processed by IRE1 to form a frameshifted mRNA fusion sequence that is translatable by a cell to produce a detectable protein;

contacting the first array of cells with a library of compounds; and

identifying a compound that inhibits the activity of IRE1.

2. The method of claim 1, wherein the library of compounds comprises at least 50, at least 100, at least 500, at least 1000, or at least 5000 different compounds.

3. The method of claim 1, wherein the first array of cells comprises a microtiter plate.

4. The method of claim 1, wherein the detectable protein is an enzyme.

5. The method of claim 4, wherein the enzyme is luciferase.

6. The method of claim 1, wherein the detectable protein is a fluorescent protein.

7. The method of claim 1, wherein the detectable protein is detected using an antibody.

8. The method of claim 1, further comprising the step of:

counterscreening the library of compounds to identify a compound that is not toxic to cells grown in the absence of ER stress.

9. The method of claim 8, wherein the compound is not toxic to cells grown in air.

10. The method of claim 1, further comprising the step of:

stimulating the unfolded protein response prior to contacting the first array of cells with the library of compounds.

11-20. (canceled)

21. The method of claim 1, further comprising the step of:

counterscreening the library of compounds to identify a compound that inhibits detection of the detectable protein.

22-30. (canceled)

31. The method of claim 1, further comprising the steps of:

stimulating the unfolded protein response prior to contacting the first array of cells with the library of compounds; and

counterscreening the library of compounds to identify a compound that inhibits detection of the detectable protein.

32-42. (canceled)

43. The method of claim 1, wherein the processing by IRE1 is an RNA splicing reaction.

44. The method of claim 1, wherein the compound inhibiting the activity of IRE1 inhibits the endonuclease activity of IRE1.

45. The method of claim 1, further comprising the step of:

counterscreening the library of compounds to identify a compound that is not toxic to cells grown in the absence of ER stress.

46. (canceled)

47. The method of claim 1, wherein the identifying step comprises comparing the amount of detectable protein in cells treated with the compound to the amount of detectable protein in untreated cells.

48. The method of claim 1, wherein the cells in the first array of cells are cancer cells.

49-53. (canceled)