US20150018406A1

US20150018406A1 - Modulation of breast cancer growth by modulation of xbp1 activity

Info

Publication number: US20150018406A1
Application number: US14/383,687
Authority: US
Inventors: Laurie H. Glimcher; Xi Chen
Original assignee: Cornell University
Current assignee: Cornell University
Priority date: 2012-03-09
Filing date: 2013-03-11
Publication date: 2015-01-15
Also published as: WO2013134774A1; US10655130B2; US20170240903A1

Abstract

Described herein is a previously unknown function of XBP1 in triple-negative breast cancer (TNBC). It is shown that XBP1 is preferentially spliced and activated in TNBC, and that deletion of XBP1 significantly blocks triple negative breast tumor growth. Strikingly, XBP1 is required for the self-renewal of breast tumor initiating cells (TICs). Genome-wide mapping of the XBP1 transcriptional regulatory network identified a fundamental role for XBP1 in regulating the response to hypoxia via the transcription factor hypoxia-inducible factor 1a (HIF1a). Importantly, activation of this pathway appears to carry prognostic implications, as expression of the XBP1-dependent signature is associated with shorter survival times in human TNBC.

Description

GOVERNMENT SUPPORT

This invention was made with government support under CA112663 and AI032412 awarded by National Institutes of Health. The government has certain rights in the invention.

BACKGROUND OF THE INVENTION

During tumor development and progression, cancer cells encounter cytotoxic conditions such as hypoxia, nutrient deprivation, and low pH due to inadequate vascularization (Hanahan, D., et al. 2011. Cell 144, 646-674). To maintain survival and growth in the face of these physiologic stressors, a set of adaptive response pathways are induced. One adaptive pathway well studied in other contexts is the unfolded protein response (UPR), which is induced by factors affecting the endoplasmic reticulum (ER) such as changes in glycosylation, redox status, glucose availability, calcium homeostasis or the accumulation of unfolded or misfolded proteins (Hetz, C., et al. 2001. Physiol Rev 91, 1219-1243). Notably, features of the tumor microenvironment, such as hypoxia and nutrient deprivation, can disrupt ER homeostasis by the perturbation of aerobic processes such as oligosaccharide modification, disulphide bond formation, isomerization, and protein quality control and export (Wouters, B. G., et al. 2008. Nat Rev Cancer 8, 851-864).
In mammalian cells, the UPR is mediated by three ER-localized transmembrane protein sensors: Inositol-requiring transmembrane kinase/endonuclease-1 (IRE1), PKR-like ER kinase (PERK) and activating transcription factor 6 (ATF6) (Walter, P., et al. 2011. Science 334, 1081-1086). Of these, IRE1 is the most evolutionarily conserved branch. An increase in the load of folding proteins in the ER activates IRE1, an ER-resident kinase and endoribonuclease that acts as an ER-stress sensor (Walter, P., et al. 2011. Science 334, 1081-1086). Activated IRE1 removes a 26 bp intron from XBP1 mRNA and results in a frame shift in the coding sequence, with the spliced form encoding a 226 amino acid transcriptional activation domain (Calfon, M., et al. 2002. Nature 415, 92-96; Yoshida, H., et al. 2001. Cell 107, 881-891). In contrast to the unspliced XBP1 (XBP1u), which is unstable and quickly degraded, spliced XBP1 (XBP1s) is stable and is a potent inducer of target genes that orchestrate the cellular response to ER stress (Hetz, C., et al. 2011. Physiol Rev 91, 1219-1243). XBP1 deficient mice display severe abnormalities in differentiation of several lineages of specialized secretory cells, including plasma cells (Reimold, A. M., et al. 2001. Nature 412, 300-307), exocrine pancreas cells (Lee, A. H., et al. 2005. EMBO J. 24, 4368-4380) and intestinal epithelial cells (Kaser, A., et al. 2008. Cell 134, 743-756). As the mammary gland is a secretory tissue that undergoes extensive secretory compartment expansion during the transition from pregnancy to lactation, the function of XBP1 in the normal mammary gland and in breast cancer is of special interest. XBP1 expression was reported to be regulated by estrogen receptor and induced in primary human breast cancer (Fujimoto, T., et al. 2003. Breast Cancer 10, 301-306), however, the functional role of the UPR and XBP1 in the normal and malignant mammary gland is largely unknown.

SUMMARY OF THE INVENTION

The unfolded protein response (UPR) is essential for tumor cells to survive the pathologic stresses intrinsic to the tumor microenvironment. The instant invention is based, at least in part, on the new finding of an unexpected function of XBP1 (X box binding protein 1), a key component of the UPR, in human triple negative breast cancer (TNBC). In particular, the instant inventors have discovered that XBP1 promotes TNBC and does so by controlling the hypoxia response. Triple negative breast cancer (TNBC) is a highly aggressive malignancy with limited treatment options and TNBC-targeted therapies do not yet exist. Here, it is reported that XBP1, a key component of the Unfolded Protein Response (UPR), is activated in TNBC and plays a pivotal role in the tumorigenicity and progression of this human breast cancer subtype. The instant inventors show that XBP1 is required for the transformation of immortalized mammary epithelial cells. Silencing of XBP1 significantly suppressed the growth and invasiveness of TNBCs. Activation of the XBP1 pathway is associated with poor prognosis in human TNBC patients. Intriguingly, XBP1 is preferentially activated in tumor initiating cells (TICs) and is essential for sustaining TIC self-renewal. Moreover, overexpression of the active form of XBP1 (XBP1s) in non-TICs is sufficient to confer stem-like or tumor-initiating properties on them, while depletion of XBP1 inhibited tumor relapse due to a preferential depletion of TICs (by reducing the population of chemotherapy-resistant TICs).
Genome-wide mapping of the XBP1 transcriptional regulatory network revealed that XBP1 regulates the hypoxia response through controlling HIF1α transcriptional activity and the expression of HIF1α targets. The instant inventors have identified a genetic fingerprint (gene expression signature) indicative of XBP1 pathway activation that is associated with poor prognosis in human TNBC patients. These findings, for the first time, reveal a key function for this branch of the UPR in TNBC (linking the UPR pathway with TNBC and TIC), opening new avenues for therapeutics for TNBC patients.

DESCRIPTION OF THE FIGURES

FIG. 1. The UPR is activated in human breast cancer.

(A) A TMA containing normal breast tissue or breast cancer tissue sections was subjected to IHC for phospho-PERK (Thr980) (DAB staining, brown). Representative pictures are shown from normal and human breast cancer tissues.

(B). Comparison of PERK phosphorylation in normal breast tissue samples and breast cancer samples. 66 normal human breast tissues and 40 human breast cancer tissues were evaluated.

(C) The TMA were subjected to IHC for phospho-EIF2α (Ser51) (DAB staining, brown). Representative pictures are shown from normal and human breast cancer tissues.

(D) Comparison of EIF2α phosphorylation in normal breast tissue samples and breast cancer samples. 59 normal human breast tissues and 41 human breast cancer tissues were evaluated.

FIG. 2. XBP1 is required for transformation of immortalized mammary epithelial cells

(A) XBP1 silencing blocks the phenotypic transformation of MCF10A ER-Src cells. MCF10A ER-Src cells were infected with lentivirus encoding XBP1 shRNA (shXBP1) or control shRNA (shCtrl), and treated with tamoxifen (TAM) for 36 hr. Phase-contrast images are shown.

(B) Quantification of invasive cells in untreated and TAM-treated MCF10A ER-Src cells in the presence or absence of control or XBP1 shRNA.

(C) Quantification of soft agar colony formation in untreated and TAM-treated MCF10A ER-Src cells in the presence or absence of control or XBP1 shRNA. Experiments were performed in triplicate and data are shown as mean±SD.

(D) Tumor growth (mean±SD) of untreated, control shRNA, and XBP1 shRNA treated MCF10A ER-Src (TAM treated) cells. TX: treatment with shRNA.

(E) MCF10A ER-Src cells were infected with retrovirus encoding XBP1s or empty vector. Phase-contrast images are shown.

(F) Quantification of soft agar colonies in MCF10A ER-Src cells infected with empty vector or spliced XBP1 (XBP1s) expressing retroviruses. Phase-contrast images are shown in the lower panel.

All experiments were performed in triplicate and data are shown as mean±SD.

FIG. 3. XBP1 inhibition blocks breast cancer cell growth and invasiveness in vitro and in vivo.

(A) RT-PCR analysis of XBP1 splicing in different luminal and basal-like cell lines. XBP1u: unspliced XBP1, XBP1s: spliced XBP1.

(B) Quantification of soft agar colony formation in untreated and control shRNA or XBP1 shRNA infected breast cancer cells.

(C) Quantification of invasive cells in untreated and control shRNA or XBP1 shRNA infected breast cancer cells. **p,0.01

(D) Quantitative RT-PCR analysis of XBP1 expression in MDA-MB-231 cells infected with doxycycline (DOX) inducible lentiviruses encoding shRNAs against XBP1 or scrambled LACZ control, in the presence or absence of doxycycline for 48 h. Data are presented relative to β-actin. Experiments were performed in triplicate and data are shown as mean±SD.

(E) Representative bioluminescent images of orthotopic tumors formed by MDA-MB-231 cells as in (D) that were then superinfected with a retrovirus encoding firefly luciferase. A total of 1.5×10⁶cells were injected into the fourth mammary glands of NOD/SCID/IL2Rγ−/− mice. Bioluminescent images were obtained 5 days later and serially after mice were begun on chow containing doxycycline (day 19). Pictures shown are the day 19 image (Before Dox) and day 64 image (After Dox).

(F) Quantitation of imaging studies as in (E). *p<0.05, **p<0.01.

(G) Tumor incidence of TNBC patient-derived BCM-2147 tumor treated with scrambled siRNA (n=11) or XBP1 siRNA (n=9). Tumor incidence is reported at 10 weeks post-transplantation. Statistical significance was determined by Barnard's test. (Barnard, G. A., 1945. Nature 156, 177; Barnard, G. A., 1947. Biometrika 34, 123-138).

(H) Tumor growth (mean±SD) of BCM-2147 tumors as in (G). *p<0.05, **p<0.01.

(I) Knockdown efficiency of XBP1 in MDA-MB-231 derived xenograft tumor (as in FIG. 3E). Quantitative RT-PCR analysis of XBP1 expression in shCtrl or shXBP1 xenograft tumor. Data are presented relative to β-actin. There are 5 mice in each group and data are shown as mean±SD.

(J) Knockdown efficiency of XBP1 in MDA-MB-231 cells with two shRNA constructs targeting different regions of XBP1.

(K) Bioluminescent images of orthotopic tumors formed by luciferase-expressing MDA-MB-231 cells infected with different lentiviruses. A total of 1.5×10 cells were injected into the fourth mammary glands of nude mice. Bioluminescent images were obtained 1 week later and serially after mice were begun on chow containing doxycycline (Dox) (day 10). Pictures shown are the day 10 images (Before Dox) and day 45 images (After Dox).

(L) Tumor growth (mean±SD) of untreated or control shRNA, and XBP1 shRNA treated MDA-MB-436 cells. **p<0.01.

(M) Tumor growth (mean±SD) of untreated or control shRNA, and XBP1 shRNA treated HBL-100 cells. **p<0.01. TX: treatment with shRNA.

FIG. 4. XBP1 is required to sustain cancer stem cell self-renewal

(A) RT-PCR analysis of XBP1 splicing in untreated and TAM treated NTICs (CD44^low/CD24^high) and TICs (CD44^high/CD24^low). XBP1u: unspliced XBP1, XBP1s: spliced XBP1.

(B) Flow cytometry analyzing CD44 and CD24 expression of untreated and TAM treated (36 h) MCF10A ER-Src cells infected with control GFP shRNA or XBP1 shRNA encoding lentivirus.

(C) Number of mammospheres per 1,000 cells generated by TAM treated MCF10A ER-Src cells uninfected, or infected with control shRNA or XBP1 shRNA encoding lentivirus.

(D) The indicated number of TAM-treated MCF10A-ER-Src cells infected with control shRNA or XBP1 shRNA were injected into NOD/SCID/IL2Rã−/− mice and the tumor incidence was reported at 12 weeks post-transplantation.

(E) RT-PCR analysis of XBP1 splicing in NTICs and TICs purified from TNBC patient. XBP1u: unspliced XBP1, XBP1s: spliced XBP1.

(F) Number of mammospheres per 1,000 cells generated from untreated and control shRNA or XBP1 shRNA encoding lentivirus infected primary tissue samples from five patients with TNBC

(G) 10 NTIC sorted from two human TNBC patients or NTIC overexpressing XBP1s were injected into NOD/SCID/IL2Rγ−/− mice and the incidence of tumors was monitored.

(H) Knockdown efficiency of XBP1 in MCF10A-ER-Src cells.

(I) Percentage of TICs (CD44high/CD24low) in TAM treated MCF10A-ER-Src cells infected with control shRNA or XBP1 shRNA encoding lentivirus.

(J) Cell viability assay (Cell-titer Glo) on TICs (CD44high/CD24low) isolated from transformed MCF10A-ER-Src cells infected with control shRNA or XBP1 shRNA encoding lentivirus (72 h after infection). Data were normalized to the control (cell infected with shCtrl). Experiments were performed in triplicate and data are shown as mean±SD.

(K) Cell viability assay (Cell-titer Glo) on NTICs (CD44low/CD24high). Data analysis is the same as (J).

(L) Tumor growth (mean±SD) of MDA-MB-231 cells untreated or treated with doxorubicin, or doxorubicin (dox)+control shRNA, or doxorubicin+XBP1 shRNA. TX: treatment with Dox or Dox+shRNA.

(M) Number of mammospheres per 1,000 cells generated from day 20 xenograft tumors under different treatments as indicated. Data are shown as mean±SD.

FIG. 5. XBP1 interacts with HIF1α and co-occupies promoters of HIF1α target genes.

(A) Motif enrichment analysis in the XBP1 binding sites. The average HIF1α motif enrichment signal is shown for the 1 kb region surrounding the center of the XBP1 binding site.

(B) FLAG-tagged HIF1α and XBP1s were co-expressed in 293T cells and the cells were treated in 0.1% O2 for 16 h. Co-IP was performed with M2 anti-FLAG antibody. Western blot was carried out with anti-XBP1s antibody or anti-FLAG antibody. Empty vector was used as negative control.

(C) Nuclear extracts from MDA-MB-231 cells treated with TM (1 ug/ml, 6 h) in 0.1% O2 (16 h) were subjected to co-IP with anti-HIF1α antibody or rabbit IgG. Western blot was carried out with anti-XBP1s antibody or anti-HIF1α antibody.

(D-F) Schematic diagram of the primer locations across the JMJD1A promoter (D). XBP1 and HIF1α cobind to JMJD1A, DDIT4, VEGFA, and PDK1 promoters under hypoxic conditions. A ChIP assay was performed using anti-XBP1 polyclonal antibody (D-E) or anti-HIF1α polyclonal antibody (D, F) to detect enriched fragments. Fold enrichment is the relative abundance of DNA fragments at the amplified region over a control amplified region. GST antibody was used as mock ChIP control (D-F). Primer locations correspond to (D).

(G) Schematic of the luciferase reporter constructs containing three copies of HRE (3×HRE)

(H) 3×HRE reporter was co-transfected with XBP1s expression plasmid or empty vector into MDA-MB-231 cells and luciferase activity measured.

(I) 3×HRE reporter was co-transfected with doxycycline (DOX) inducible constructs encoding two shRNAs targeting different regions of XBP1 or scrambled LACZ control into MDA-MB-231 cells. Cells were treated in 0.1% O₂for 24 h in the presence or absence of doxycycline, and luciferase activity assayed. All luciferase activity was measured relative to the renilla luciferase internal control. Experiments were performed in triplicate and data are shown as mean±SD. *p<0.05, **p<0.01.

(J) Western blotting analysis of XBP1s expression in nuclear extract of MDA-MB-231 cells cultured under unstressed or stressed condition (0.1% O2 and glucose deprivation) for 16 h. Lamin B was used as loading control.

(K) Distribution of XBP1 binding sites. Locations of XBP1 binding sites relative to the nearest tran transcription units. The percentages of binding sites at the respective locations are shown

(L) Identification of XBP1 motif in ChIP-seq. Matrices predicted by the de novo motif-discovery algorithm Seqpos. p=1×10⁻³⁰.

(M) Nuclear extracts from Hs578T cells treated with TM (1 ug/ml, 6 h) in 0.1% O2 (16 h) were subjected to co-IP with anti-HIF1α antibody or rabbit IgG. Western blot was carried out with anti-XBP1s antibody or anti-HIF1α antibody.

(N) XBP1 and HIF1α co-bind to the JMJD2C promoter under hypoxic conditions.

FIG. 6. XBP1 regulates the hypoxia response.

(A) Plot from GSEA showing enrichment of the HIF1α mediated hypoxia response pathway in XBP1-upreuglated genes.

(B) Gene expression microarray heatmap showing that genes involved in the HIF1α mediated hypoxia responses were differentially expressed after XBP1 knockdown.

(C-D) Quantitative RT-PCR analysis of VEGFA, PDK1, GLUT1, JMJD1A and DDIT4 expression after knockdown of XBP1 in MDA-MB-231 under hypoxic conditions (C) or MDA-MB-231 derived xenograft tumors (d, n=5). Results are presented relative to β-actin expression. Experiments were performed in triplicate and data are shown as mean±SD. *p<0.05, **p<0.01.

(E) Plot showing the genome-wide association between the strength of the XBP1 binding and the occurrence of the HIF1α motif. The signal of XBP1 ChIP-seq peaks was shown as a heatmap using red (the strongest signal) and white (the weakest signal) color scheme. Each row shows ±300 bp centered on the XBP1 ChIP-seq peak summits. Rows are ranked by XBP1 occupancy. The horizontal blue lines denote the presence of the HIF1α motif.

(F-G) Chromatin extracts from control MDA-MB-231 cells or XBP1 knockdown MDA-MB-231 cells (treated with 0.1% O₂for 24 h) were subjected to ChIP using anti-HIF1α antibody (F), and anti-RNA polymerase II antibody (G). The primers used to detect ChIP-enriched DNA in (F-G) were the peak pair of primers in JMJD1A, DDIT4, NDRG1, PDK1 and VEGFA promoters (Table 2). Primers in the β-actin region/promoter were used as control. Data are presented as the mean±SD.

(H) Quantitative RT-PCR analysis of VEGFA, PDK1, GLUT1, MCT4, JMJD1A and XBP1 expression after knockdown of XBP1 in Hs578T cells treated with 0.1% O2 for 24 h. Results are presented relative to β-actin expression. Experiments were performed in triplicate and data are shown as mean±SD. *p<0.05, **p<0.01.

(I) Chromatin extracts from control MDA-MB-231 cells or XBP1 knockdown MDA-MB-231 cells (treated with 0.1% O2 for 24 h) were subjected to ChIP using anti-XBP1s antibody. Data are presented as the mean±SD.

(J) Immunoblotting analysis of control MDA-MB-231 cell lysates and XBP1 knockdown lysates (treated with 0.1% O2 for 24 h) were performed using anti-HIF1α or anti-HSP90 antibody.

FIG. 7. XBP1 genetic signature is associated with human breast cancer prognosis.

(A) Heatmap showing the expression profile of genes bound by XBP1 and differentially expressed after XBP1 knockdown

(B-C) Kaplan-Meier graphs demonstrating a significant association elevated expression of the XBP1 signature with shorter relapse-free survival in two cohorts of triple negative breast cancer patients (B and C). The log-rank test P values are shown.

(D). Kaplan-Meier graphs showing the significant association of expression of HIF1α gene signature with shorter relapse-free survival in a cohort of 383 TNBC patients. The log-rank test P values are shown.

DETAILED DESCRIPTION OF THE INVENTION

The unfolded protein response (UPR) is essential for tumor cells to survive the pathologic stresses intrinsic to the tumor microenvironment. Here, it is reported an unexpected function of XBP1 (X box binding protein1), a key component of the UPR, in human triple negative breast cancer (TNBC). It is shown that XBP1 is required for the transformation of immortalized mammary epithelial cells. Silencing of XBP1 significantly suppressed the growth and invasiveness of TNBCs. Activation of the XBP1 pathway is associated with poor prognosis in human TNBC patients. Intriguingly, XBP1 is preferentially activated in tumor initiating cells (TICs) and is essential for sustaining TIC self-renewal. Moreover, overexpression of the active form of XBP1 in non-TICs is sufficient to confer stem-like properties on them, while depletion of XBP1 inhibited tumor relapse due to a preferential depletion of TICs. Genome-wide mapping of the XBP1 transcriptional regulatory network revealed that XBP1 regulates the hypoxia response through controlling HIF1α transcriptional activity and the expression of HIF1α targets. The instant inventors have identified a genetic fingerprint indicative of XBP1 pathway activation that is associated with poor prognosis in human TNBC patients. These findings, for the first time, link the UPR pathway with TNBC and TIC, opening new avenues for therapeutics for TNBC patients.
Accordingly, in one aspect, the invention pertains to a method of inhibiting growth of triple negative breast cancer (TNBC) in a subject, the method comprising administering to the subject a direct or indirect inhibitor of XBP1 such that growth of the TNBC in the subject is inhibited. Non-limiting examples of direct inhibitors of XBP1 include a nucleic acid molecule that is antisense to an XBP1-encoding nucleic acid molecule, an XBP1 shRNA, an XBP siRNA, a microRNA that targets XBP1, a dominant negative XBP1 molecule and small molecule inhibitors of XBP1. Non-limiting examples of indirect inhibitors of XBP1 include agents that target IRE1, an endonuclease essential for proper splicing and activation of XBP1, such that inhibition of IRE1 leads to inhibition of the production of the spliced, active form of XBP1. Non-limiting examples of IRE1 inhibitors include a nucleic acid molecule that is antisense to an IRE1-encoding nucleic acid molecule, an IRE1 shRNA, an IRE1 siRNA, a microRNA that targets IRE1, a dominant negative IRE1 molecule and small molecule inhibitors of IRE1.
In another aspect, the invention pertains to a method of identifying a compound useful in inhibiting the growth of triple negative breast cancer (TNBC) cells, the method comprising:
a) providing an indicator composition comprising XBP1 and HIF1α, or biologically active portions thereof;
b) contacting the indicator composition with each member of a library of test compounds;
c) selecting from the library of test compounds a compound of interest that decreases the interaction of XBP1 and HIF1α, or biologically active portions thereof, wherein the ability of a compound to inhibit growth of TNBC cells is indicated by a decrease in the interaction as compared to the amount of interaction in the absence of the compound.
The indicator composition can be, for example, a cell-free preparation comprising XBP1 and HIF1 cc, or biologically active portions thereof (e.g., isolated recombinant proteins), or a cell comprising XBP1 and HIF1α, or biologically active portions thereof (e.g., a recombinant cell transfected to express XBP1 and HIF1α proteins). The read-out for the method to determine the amount of interaction between XBP1 and HIF1α can be, for example, a direct read-out that measures the amount of binding between XBP1 and HIF1α (e.g., one or both proteins can be labeled or tagged), such as co-immunnoprecipitation, or an indirect read-out that measures the amount of transcriptional activity of the XBP1/HIF1α complex, such as use of a reporter gene responsive to the XBP1/HIF1α complex and measurement of the level of the reporter.
In yet another aspect, the invention pertains to a method for determining a prognosis status for a subject with triple negative breast cancer (TNBC), the method comprising:
a) determining an XBP1 gene signature for the TNBC of the subject; and
b) correlating the XBP1 gene signature with a prognosis status for the subject, wherein higher expression of the XBP1 gene signature, relative to a control, correlates with shorter relapse-free survival of the subject and lower expression of the XBP1 gene signature, relative to a control, correlates with longer relapse-free survival of the subject.
The XBP1 gene signature can comprise, for example, a plurality of genes regulated by XBP1 in TNBC, such as a plurality of genes selected from the 133 genes shown in Table 1.
The contents of all references, patents, and published patent applications cited throughout this application, as well as the figures and the sequence listing, are hereby incorporated by reference.
Various aspects of the invention are described in further detail in the following subsections:
I. XBP1 and Triple Negative Breast Cancer
During tumor development and progression, cancer cells encounter cytotoxic conditions such as hypoxia, nutrient deprivation, and low pH due to inadequate vascularization (Hanahan, D., et al. 2011. Cell 144, 646-74). To maintain survival and growth in the face of these physiologic stressors, a set of adaptive response pathways are induced. One adaptive pathway well studied in other contexts is the unfolded protein response (UPR), which is induced by factors affecting the endoplasmic reticulum (ER) such as changes in glycosylation, redox status, glucose availability, calcium homeostasis or the accumulation of unfolded or misfolded proteins (Hetz, C., et al. 2011. Physiol Rev 91, 1219-43). Notably, features of the tumor microenvironment, such as hypoxia and nutrient deprivation, can disrupt ER homeostasis by the perturbation of aerobic processes such as oligosaccharide modification, disulphide bond formation, isomerization, and protein quality control and export (Wouters, B. G., et al. 2008. Nat Rev Cancer 8, 851-64). In mammalian cells, the UPR is mediated by three ER-localized transmembrane protein sensors: Inositol-requiring transmembrane kinase/endonuclease-1 (IRE1), PKR-like ER kinase (PERK) and activating transcription factor 6 (ATF6) (Walter, P., et al. 2011. Science 334, 1081-6). Of these, IRE1 is the most evolutionarily conserved branch. An increase in the load of folding proteins in the ER activates IRE1, an ER-resident kinase and endoribonuclease that acts as an ER-stress sensor4. Activated IRE1 removes a 26 bp intron from XBP1 mRNA and results in a frame shift in the coding sequence, with the spliced form encoding a 226 amino acid transcriptional activation domain (Calfon, M., et al. 2002. Nature 415, 92-6; Yoshida, H., et al. 2001. Cell 107, 881-91). In contrast to the unspliced XBP1 (XBP1u), which is unstable and quickly degraded, spliced XBP1 (XBP1s) is stable and is a potent inducer of target genes that orchestrate the cellular response to ER stress (Hetz, C., et al. 2011. Physiol Rev 91, 1219-43). Several studies have reported on the activation of the UPR in various human tumors and its relevance to combinatorial therapy (Ma, Y., et al. 2004. Nat Rev Cancer 4, 966-77; De Raedt, T., et al. 2011. Cancer Cell 20, 400-13; Mahoney, D. J., et al. 2011. Cancer Cell 20, 443-56; Healy, S. J., et al. 2009. Eur J Pharmacol 625, 234-46; Carrasco, D. R., et al. 2007. Cancer Cell 11, 349-60). However, the role of the UPR and XBP1 in the malignant mammary cell is largely unknown.
As described in detail above, the UPR is a major cellular stress response pathway activated in tumors that allows them to adapt to the stresses of the tumor microenvironment. Several studies have reported on the activation of the UPR in various human tumors and its relevance to combinatorial therapy (Carrasco, D. R., et al. 2007. Cancer Cell 11, 349-36; De Raedt, T., et al. 2011. Cancer Cell 20, 400-413; Healy, S. J., et al. 2009. Eur J Pharmacol 625, 234-246; Ma, Y., et al. 2004. Nat Rev Cancer 4, 966-977; Mahoney, D. J., et al. 2011. Cancer Cell 20, 443-456). However, the role of the UPR in breast cancer pathogenesis remains elusive. Here, the instant inventors have identified a previously unknown function of XBP1 in triple-negative breast cancer (TNBC). It is demonstrated that XBP1 is spliced and activated in TNBC, and that deletion of XBP1 significantly blocks triple negative breast tumor growth. Here, it is demonstrates that XBP1, a key component of the most evolutionarily conserved branch of the UPR, is essential for the transformation of mammary epithelial cells and is preferentially activated in tumor initiating cells (TICs) where it is essential for sustaining TIC self-renewal. Furthermore, XBP1 silencing suppressed tumor relapse along with depleting the breast tumor initiating cells (TICs). Genome-wide mapping of the XBP1 transcriptional regulatory network identified its key downstream target to be the hypoxia response via the transcription factor hypoxia-inducible factor 1α (HIF1α). XBP1 regulates HIF1α transcriptional activity by controlling HIF1α binding to promoter DNA and by the recruitment of RNA polymerase II. We generated a genetic fingerprint indicative of XBP1 pathway activation that we found to be associated with poor prognosis in human TNBC patients. Moreover, activation of the hypoxia response pathway appears to carry prognostic implications, as expression of the XBP1-dependent signature is associated with shorter survival times in patients with TNBC.
XBP1 was reported to be highly expressed in ER+ breast tumors and to activate ERα in a ligand-independent manner (Ding, L., et al. 2003. Nucleic Acids Res 31, 5266-5274; Fujimoto, T., et al. 2003. Breast Cancer 10, 301-306). Splicing of XBP1 confers estrogen independence and anti-estrogen resistance to breast cancer cell lines (Gomez, B. P., et al. 2007. Faseb J 21, 4013-4027). Here, by manipulating the expression of XBP1 in a panel of breast cancer cell lines and in a human xenograft model, we discovered a key function for XBP1 in TNBC. TNBC is a subtype of breast tumors characterized by a of the absence of expression of ER, PR and HER2, signaling receptors known to fuel most breast cancers. TNBC is extremely aggressive and more likely to recur and metastasize than the other subtypes (Foulkes, W. D., et al. 2010. N Engl J Med 363, 1938-1948). While ER+, PR+ or Her2 tumors respond well to ER antagonist, aromatase inhibitor, or Her2-targeted therapies, TNBC is unresponsive to most receptor targeted treatments. TNBC is a highly heterogeneous group of cancers, the genes linked to TNBC are not well understood and thus, targeted therapies do not yet exist. We found that XBP1 was preferentially activated in TNBC cells, and that silencing of XBP1 was very effective in suppressing the tumorigenicity and progression of TNBCs.
A TNBC
Triple-negative breast cancer (TNBC) refers to any breast cancer that does not express the genes for estrogen receptor (ER), progesterone receptor (PR) or Her2/neu. Triple negative is sometimes used as a surrogate term for basal-like; however, more detailed classification may provide better guidance for treatment and better estimates for prognosis. (Hudis, C. A., et al. 2011. The Oncologist 16, 1-11). Triple-negative breast cancer (TNBC) is breast cancer characterized by malignant tumors. As used herein, the term “malignant” refers to a non-benign tumor or a cancer. In one embodiment a malignancy expands to other parts of the body as well (metastasizes). A malignant tumor is usually life-threatening, causing death if it remains untreated. If treated, the spread of a malignant tumor can be slowed or even arrested. Depending on the amount of tissue damage prior to treatment, tissue or organ function can be compromised.
Triple negative breast cancers have a relapse pattern that is very different from hormone-positive breast cancers: the risk of relapse is much higher for the first 3-5 years but drops sharply and substantially below that of hormone-positive breast cancers after that. This relapse pattern has been recognized for all types of triple negative cancers for which sufficient data exists although the absolute relapse and survival rates differ across subtypes. (Hudis, C. A., et al. 2011. The Oncologist 16, 1-11; Cheang, M. C. U., et al. 2008. Clinical Cancer Research 14 (5), 1368-1376).
Triple-negative breast cancers are sometimes classified into “basal-type” and other cancers; however, there is no standard classification scheme. Basal type cancers are frequently defined by cytokeratin 5/6 and EGFR staining. However no clear criteria or cutoff values have been standardized yet. (Hudis, C. A., et al. (2011). The Oncologist 16, 1-11). About 75% of basal-type breast cancers are triple negative. Some TNBC overexpresses epidermal growth factor receptor (EGFR). Some TNBC over expresses transmembrane glycoprotein NMB (GPNMB). On histologic examination triple negative breast tumors mostly fall into the categories secretory carcinoma or adenoid cystic types (both considered less aggressive), medullary cancers and grade 3 invasive ductal carcinomas with no specific subtype, and highly aggressive metastatic cancers. (Hudis, C. A. et al. 2011. The Oncologist 16, 1-11). Medullary TNBC in younger women are frequently BRCA1-related. Rare forms of triple negative breast cancer are apocrine and squamous carcinoma. Inflammatory breast cancer is also frequently triple negative.
B. UPR
The term “Unfolded Protein Response” (UPR) or the “Unfolded Protein Response pathway” refers to an adaptive response to the accumulation of unfolded proteins in the ER and includes the transcriptional activation of genes encoding chaperones and folding catalysts and protein degrading complexes as well as translational attenuation to limit further accumulation of unfolded proteins. Both surface and secreted proteins are synthesized in the endoplasmic reticulum (ER) where they need to fold and assemble prior to being transported.
Since the ER and the nucleus are located in separate compartments of the cell, the unfolded protein signal must be sensed in the lumen of the ER and transferred across the ER membrane and be received by the transcription machinery in the nucleus. The unfolded protein response (UPR) performs this function for the cell. Activation of the UPR can be caused by treatment of cells with reducing agents like DTT, by inhibitors of core glycosylation like tunicamycin or by Ca-ionophores that deplete the ER calcium stores. First discovered in yeast, the UPR has now been described in C. elegans as well as in mammalian cells. In mammals, the UPR signal cascade is mediated by three types of ER transmembrane proteins: the protein-kinase and site-specific endoribonuclease IRE-1; the eukaryotic translation initiation factor 2 kinase, PERK/PEK; and the transcriptional activator ATF6. If the UPR cannot adapt to the presence of unfolded proteins in the ER, an apoptotic response is initiated leading to the activation of JNK protein kinase and caspases 7, 12, and 3. The most proximal signal from the lumen of the ER is received by a transmembrane endoribonuclease and kinase called IRE-1. Following ER stress, IRE-1 is essential for survival because it initiates splicing of the XBP-1 mRNA the spliced version of which, as shown herein, activates the UPR.
C. XBP1
The term “XBP-1” refers to a X-box binding human protein that is a DNA binding protein and has an amino acid sequence as described in, for example, Liou, H. C., et. al. 1990. Science 247, 1581-1584 and Yoshimura, T., et al. 1990. EMBO J. 9, 2537-2542, and other mammalian homologs thereof, such as described in Kishimoto T., et al. 1996. Biochem. Biophys. Res. Commun. 223, 746-751 (rat homologue). Exemplary proteins intended to be encompassed by the term “XBP-1” include those having amino acid sequences disclosed in GenBank with accession numbers A36299 [gi:105867], NP_—005071 [gi:4827058], P17861 [gi:139787], CAA39149 [gi:287645], and BAA82600 [gi:5596360] or e.g., encoded by nucleic acid molecules such as those disclosed in GenBank with accession numbers AF027963 [gi: 13752783]; NM_—013842 [gi:13775155]; or M31627 [gi:184485]. XBP-1 is also referred to in the art as TREB5 or HTF (Yoshimura, T., et al. 1990. EMBO Journal. 9, 2537; Matsuzaki, Y., et al. 1995. J. Biochem. 117, 303). Like other members of the b-zip family, XBP-1 has a basic region that mediates DNA-binding and an adjacent leucine zipper structure that mediates protein dimerization.
As described above, there are two forms of XBP-1 protein, unspliced and spliced, which differ markedly in their sequence and activity. Unless the form is referred to explicitly herein, the term “XBP-1” as used herein includes both the spliced and unspliced forms. Spliced XBP-1 (“XBP1s”) directly controls the activation of the UPR, while unspliced XBP-1 functions due to its ability to negatively regulate spliced XBP-1.
As used herein, the term “spliced XBP-1” (“XBP1s”) refers to the spliced, processed form of the mammalian XBP-1 mRNA or the corresponding protein. Human and murine XBP-1 mRNA contain an open reading frame (ORF1) encoding bZIP proteins of 261 and 267 amino acids, respectively. Both mRNA's also contain another ORF, ORF2, partially overlapping but not in frame with ORF1. ORF2 encodes 222 amino acids in both human and murine cells. Human and murine ORF1 and ORF2 in the XBP-1 mRNA share 75% and 89% identity respectively.
As used herein, the term “unspliced XBP-1” refers to the unprocessed XBP-1 mRNA or the corresponding protein. As set forth above, unspliced murineXBP-1 is 267 amino acids in length and spliced murine XBP-1 is 371 amino acids in length. The sequence of unspliced XBP-1 is known in the art and can be found, e.g., Liou, H. C., et. al. 1990. Science 247, 1581-1584 and Yoshimura, T., et al. 1990. EMBO J. 9, 2537-2542, or at GenBank accession numbers NM_—005080 [gi:14110394] or NM_—013842 [gi:13775155].
II. XBP1 and Tumor Initiating Cells
TNBC typically contain a higher proportion of “stem-like” breast cancer cells, also known as tumor initiating cells (TICs), characterized by a CD44⁺CD24^−/lowsurface phenotype of and the expression of aldehyde dehydrogenase 1 (Al-Hajj, M., et al. 2003. Proc Natl Acad Sci USA 100, 3983-3988; Ginestier, C., et al. 2007. Cell Stem Cell 1, 555-567). TICs resemble stem cells, as they are capable of both indefinite self-renewal and differentiation. Relative to NTICs, TICs contribute to a significantly higher incidence of recurrence and distant metastasis, and are responsible for tumor initiation and maintenance (Smalley, M., et al. 2003. Nat Rev Cancer 3, 832-844; Stingl, J., et al. 2007. Nat Rev Cancer 7, 791-799). Although conventional therapies have shown great promise in killing the bulk of differentiated tumor cells, TICs are resistant to chemotherapy (Stingl, J., et al. 2007. Nat Rev Cancer 7, 791-799). The development of effective therapies targeting the TIC is urgently needed to treat breast cancer metastasis and relapse. Although several self-renewal regulatory pathways including the Notch, Wnt and Hedgehog pathways (Visvader, J. E., et al. 2008. Nat Rev Cancer 8, 755-768), as well as microenvironmental stress, such as hypoxia (Keith, B., et al. 2007. Cell 129, 465-472; Schwab, L. P., et al. 2012. Breast Cancer Res 14, R6), are known to be essential in promoting a stem-like phenotype, progress in targeting TICs with novel therapeutics is still hindered by our incomplete knowledge of the molecular pathways contributing to TIC identity.
Here we have demonstrated that XBP1 is essential for the self-renewal of breast TICs. In support of this claim, we showed that XBP1 was selectively activated in TICs, XBP1 inhibition blocked the formation of TICs, and depletion of XBP1 greatly suppressed the growth of mammospheres derived from human TNBC patients and various breast cancer cell lines, a key measure of TIC function. Overexpression of XBP1s in non-TICs conferred stem-like traits and tumorigenic potential at very low dilutions (10 cells). Finally, XBP1 depletion in combination with chemotherapy blocked xenograft tumor growth and relapse, which was attributed to the decreased TIC population after combinatorial treatment. Ours is the first study to demonstrate that compromising the ER stress response significantly impairs TIC growth and self-renewal. We speculate that the rapid proliferation of TICs requires robust ER protein folding, assembly, and transport, functions which rely on XBP1 activation and which are compromised in its absence. XBP1 serves as one of the major cellular adaptive mechanisms activated to protect TICs in a non-dividing dormant state, and XBP1 confers on TICs growth and survival advantages over non-TICs. The specific acquisition of XBP1 activation in TICs is intriguing and provides new insights into pathways that may be used to target this subpopulation of cancer cells.
III. XBP1 Regulates the Hypoxia Response Through HIF1α
Hypoxia is known to promote aggressive tumor phenotypes. A growing body of evidence indicates that hypoxia is required for TIC survival and tumor propagation in glioma, lymphoma and acute myeloid leukemia (Heddleston, J. M., et al. 2009. Br J Cancer 102, 789-795; Jogi, A., et al. 2002. Proc Natl Acad Sci USA 99, 7021-7026; Li., Z., et al. 2009. Cancer Cell 15, 501-513; Wang, Y., et al. 2011. Cell Stem Cell 8, 399-411). HIF transcription factors are crucial to the maintenance of the undifferentiated state of stem cells residing in hypoxic niches. TNBCs also display increased levels of hypoxia (Rakha, E. A., et al. 2009. Clin Cancer Res 15, 2302-2310; Tan, E. Y., et al. 2009. Br J Cancer 100, 405-411) and HIF1α was recently demonstrated to be essential for their maintenance of breast TICs. HIF1α promotes expansion of breast TICs in vivo, and deletion of HIF 1α results in reduced mammosphere formation, primary breast tumor growth and pulmonary metastases in the MMTV-PyVT breast cancer mouse model (Schwab, L. P., et al. 2012. Breast Cancer Res 14, R6). Increased HIF1α levels are also associated with increased metastasis and decreased survival in patients with breast cancer (Bos, R., et al. 2003. Cancer 97, 1573-1581; Semenza, G. L., 2010. Cell 107, 1-3).
Our data reveal that XBP1 acts in breast TICs and TNBC through regulating the response to hypoxia. HIF1α requires XBP1 to sustain downstream target expression under hypoxic conditions. XBP1 interacts with HIF1α to co-occupy a set of, if not all, HIF1α targets. Depletion of XBP1 leads to reduction in classic HIF1α targets expression and HRE activity by blocking HIF1α binding to its target genes, which subsequently affects the recruitment of RNA polymerase II to target promoters. Hypoxia is a physiological inducer of the UPR in cancer (Wouters, B. G., et al. 2008. Nat Rev Cancer 8, 851-864). In this study, we found that XBP1 functions in a positive feedback loop to sustain the hypoxia response via regulating HIF1α transcriptional activity. This feed-forward circuit ensures maximum HIF activity and an efficient adaptive response to the cytotoxic microenvironment of solid tumors. HIF activity is tightly controlled during tumor progression, through translational and post-translational regulation of HIF1α but relatively little is known about how HIF1α transcriptional activity is controlled (Kaelin, W. G., Jr., et al. 2008. Mol Cell 30, 393-402). Our study reveals an unexpected function for XBP1 as a HIF1α transcriptional cofactor. We propose a model in which these two critical pathways, the UPR and the hypoxia response, are physically interconnected and act together to mount an appropriate adaptive response that promotes the survival of TICs in the hostile tumor microenvironment
IV. Therapeutic Targeting of the UPR in TNBC
We have highlighted the importance of the IRE1/XBP1 pathway in TNBC growth and metastasis, in part through regulating TICs. XBP1s expression is directly correlated with poor patient survival in human TNBC patients. Strikingly, while XBP1 is selectively activated in rapidly growing TICs, UPR pathways remain in a quiescent state in most normal unstressed cells. Hence inhibition of the UPR may offer a means to exclusively target tumor cells.
XBP1 is a transcription factor, and traditionally transcription factors other than hormone receptors have been difficult to target with small molecules. However, the upstream kinase and endoribonuclease IRE1, which drives the splicing of XBP1 mRNA, is a viable drug target. Recently, two groups have identified specific IRE1 endoribonuclease inhibitors (Papandreou, I., et al. 2011. Blood 117, 1311-1314; Volkmann, K., et al. 2011. J Biol Chem 286, 12743-12755). Intriguingly, these compounds efficiently inhibit XBP1 splicing in vivo and dramatically impair tumor growth in a xenograft model (Mahoney, D. J., et al. 2011. Cancer Cell 20, 443-456; Papandreou, I., et al. 2011. Blood 117, 1311-1314; Volkmann, K., et al. 2011. J Biol Chem 286, 12743-12755). While large-scale small molecule screens have provided potentially promising candidates that target the IRE1/XBP1 pathway, attention needs to be paid to the specificity and cytotoxity of these compounds in vivo. Recent advances in solving the crystal structure of IRE1 (Korennykh, A. V., et al. 2009. Nature 457, 687-693; Lee, K. P., et al. 2008. Cell 132, 89-100; Zhou, J., et al. 2006. Proc Natl Acad Sci USA 103, 14343-14348) should accelerate the design of more potent and specific IRE1 inhibitors. The use of UPR inhibitors in combination with standard chemotherapy may greatly enhance the effectiveness of anti-tumor therapies.
The methods of the invention using inhibitory compounds which inhibit the expression, processing, post-translational modification, or activity of spliced XBP-1 or a molecule in a biological pathway involving XBP-1 can be used in the treatment of TNBC. In one embodiment of the invention, an inhibitory compound can be used to inhibit (e.g., specifically inhibit) the expression, processing, post-translational modification, or activity of spliced XBP-1. In another embodiment, an inhibitory compound can be used to inhibit (e.g., specifically inhibit) the expression, processing, post-translational modification, or activity of unspliced XBP-1.
Inhibitory compounds of the invention can be, for example, intracellular binding molecules that act to specifically or directly inhibit the expression, processing, post-translational modification, or activity e.g., of XBP-1 or a molecule in a biological pathway involving XBP-1 (e.g., HIF1α). As used herein, the term “intracellular binding molecule” is intended to include molecules that act intracellularly to inhibit the processing expression or activity of a protein by binding to the protein or to a nucleic acid (e.g., an mRNA molecule) that encodes the protein. Examples of intracellular binding molecules, described in further detail below, include antisense nucleic acids, intracellular antibodies, peptidic compounds that inhibit the interaction of XBP-1 or a molecule in a biological pathway involving XBP-1 and a target molecule (e.g., HIF1α), and chemical agents that specifically or directly inhibit XBP-1 activity or the activity of a molecule in a biological pathway involving XBP-1 (e.g., HIF1α).
In one embodiment, an inhibitory compound of the invention is an antisense nucleic acid molecule that is complementary to a gene encoding XBP-1 or a molecule in a signal transduction pathway involving XBP-1, e.g., a molecule with which XBP-1 interacts), or to a portion of said gene, or a recombinant expression vector encoding said antisense nucleic acid molecule. The use of antisense nucleic acids to downregulate the expression of a particular protein in a cell is well known in the art (see e.g., Weintraub, H., et al 1986. Reviews—Trends in Genetics, Vol. 1(1); Askari, F. K., et al 1996. N. Eng. J. Med. 334, 316-318; Bennett, M. R., et al. 1995. Circulation 92, 1981-1993; Mercola, D., et al. 1995. Cancer Gene Ther. 2, 47-59; Rossi, J. J., 1995. Br. Med. Bull. 51, 217-225; Wagner, R. W., 1994. Nature 372, 333-335). An antisense nucleic acid molecule comprises a nucleotide sequence that is complementary to the coding strand of another nucleic acid molecule (e.g., an mRNA sequence) and accordingly is capable of hydrogen bonding to the coding strand of the other nucleic acid molecule. Antisense sequences complementary to a sequence of an mRNA can be complementary to a sequence found in the coding region of the mRNA, the 5′ or 3′ untranslated region of the mRNA or a region bridging the coding region and an untranslated region (e.g., at the junction of the 5′ untranslated region and the coding region). Furthermore, an antisense nucleic acid can be complementary in sequence to a regulatory region of the gene encoding the mRNA, for instance a transcription initiation sequence or regulatory element. Preferably, an antisense nucleic acid is designed so as to be complementary to a region preceding or spanning the initiation codon on the coding strand or in the 3′ untranslated region of an mRNA. Given the known nucleotide sequence for the coding strand of the XBP-1 gene and thus the known sequence of the XBP-1 mRNA, antisense nucleic acids of the invention can be designed according to the rules of Watson and Crick base pairing. For example, the antisense oligonucleotide can be complementary to the region surrounding the translation start site of an XBP-1 An antisense oligonucleotide can be, for example, about 5, 10, 15, 20, 25, 30, 35, 40, 45 or 50 nucleotides in length. An antisense nucleic acid of the invention can be constructed using chemical synthesis and enzymatic ligation reactions using procedures known in the art. To inhibit expression in cells, one or more antisense oligonucleotides can be used.
Alternatively, an antisense nucleic acid can be produced biologically using an expression vector into which all or a portion of a cDNA has been subcloned in an antisense orientation (i.e., nucleic acid transcribed from the inserted nucleic acid will be of an antisense orientation to a target nucleic acid of interest). The antisense expression vector can be in the form of, for example, a recombinant plasmid, phagemid or attenuated virus. The antisense expression vector can be introduced into cells using a standard transfection technique.
The antisense nucleic acid molecules of the invention are typically administered to a subject or generated in situ such that they hybridize with or bind to cellular mRNA and/or genomic DNA encoding a protein to thereby inhibit expression of the protein, e.g., by inhibiting transcription and/or translation. An example of a route of administration of an antisense nucleic acid molecule of the invention includes direct injection at a tissue site. Alternatively, an antisense nucleic acid molecule can be modified to target selected cells and then administered systemically. For example, for systemic administration, an antisense molecule can be modified such that it specifically binds to a receptor or an antigen expressed on a selected cell surface, e.g., by linking the antisense nucleic acid molecule to a peptide or an antibody which binds to a cell surface receptor or antigen. The antisense nucleic acid molecule can also be delivered to cells using the vectors described herein.
In yet another embodiment, an antisense nucleic acid molecule of the invention is an α-anomeric nucleic acid molecule. An α-anomeric nucleic acid molecule forms specific double-stranded hybrids with complementary RNA in which, contrary to the usual β-units, the strands run parallel to each other (Gautier, C., et al. 1987. Nucleic Acids. Res. 15, 6625-6641). The antisense nucleic acid molecule can also comprise a 2′-o-methylribonucleotide (Inoue, H., et al. 1987. Nucleic Acids Res. 15, 6131-6148) or a chimeric RNA-DNA analogue (Inoue, H., et al. 1987. FEBS Lett. 215, 327-330).
In still another embodiment, an antisense nucleic acid molecule of the invention is a ribozyme. Ribozymes are catalytic RNA molecules with ribonuclease activity which are capable of cleaving a single-stranded nucleic acid, such as an mRNA, to which they have a complementary region. Thus, ribozymes (e.g., hammerhead ribozymes (described in Haselhoff, J., et al. 1988. Nature 334, 585-591)) can be used to catalytically cleave mRNA transcripts to thereby inhibit translation mRNAs. Alternatively, gene expression can be inhibited by targeting nucleotide sequences complementary to the regulatory region of a gene (e.g., an XBP-1 promoter and/or enhancer) to form triple helical structures that prevent transcription of a gene in target cells. See generally, Helene, C., 1991. Anticancer Drug Des. 6(6), 569-84; Helene, C., et al. 1992. Ann. N.Y. Acad. Sci. 660, 27-36; and Maher, L. J., 1992. Bioassays 14(12), 807-15.
In another embodiment, a compound that promotes RNAi can be used to inhibit expression of XBP-1 or a molecule in a biological pathway involving XBP-1. The term “RNA interference” or “RNAi”, as used herein, refers generally to a sequence-specific or selective process by which a target molecule (e.g., a target gene, protein or RNA) is downregulated. In specific embodiments, the process of “RNA interference” or “RNAi” features degradation of RNA molecules, e.g., RNA molecules within a cell, said degradation being triggered by an RNA agent. Degradation is catalyzed by an enzymatic, RNA-induced silencing complex (RISC). RNAi occurs in cells naturally to remove foreign RNAs (e.g., viral RNAs). Natural RNAi proceeds via fragments cleaved from free dsRNA which direct the degradative mechanism to other similar RNA sequences. Alternatively, RNAi can be initiated by the hand of man, for example, to silence the expression of target genes. RNA interference (RNAi is a post-transcriptional, targeted gene-silencing technique that uses double-stranded RNA (dsRNA) to degrade messenger RNA (mRNA) containing the same sequence as the dsRNA (Sharp, P. A., et al. 2000. Science 287, 5462:2431-3.; Zamore, P. D., et al. 2000. Cell 101, 25-33. Tuschl, T., et al. 1999. Genes Dev. 13, 3191-3197; Cottrell T. R., et al. 2003. Trends Microbiol. 11, 37-43; Bushman F., 2003. Mol Therapy 7, 9-10; McManus M. T., et al. 2002. Nat Rev Genet. 3, 737-47). The process occurs when an endogenous ribonuclease cleaves the longer dsRNA into shorter, e.g., 21-23-nucleotide-long RNAs, termed small interfering RNAs or siRNAs. As used herein, the term “small interfering RNA” (“siRNA”) (also referred to in the art as “short interfering RNAs”) refers to an RNA agent, preferably a double-stranded agent, of about 10-50 nucleotides in length (the term “nucleotides” including nucleotide analogs), preferably between about 15-25 nucleotides in length, more preferably about 17, 18, 19, 20, 21, 22, 23, 24, or 25 nucleotides in length, the strands optionally having overhanging ends comprising, for example 1, 2 or 3 overhanging nucleotides (or nucleotide analogs), which is capable of directing or mediating RNA interference. Naturally-occurring siRNAs are generated from longer dsRNA molecules (e.g., >25 nucleotides in length) by a cell's RNAi machinery (e.g., Dicer or a homolog thereof). The smaller RNA segments then mediate the degradation of the target mRNA. Kits for synthesis of RNAi are commercially available from, e.g. New England Biolabsor Ambion. In one embodiment one or more of the chemistries described above for use in antisense RNA can be employed in molecules that mediate RNAi.
Alternatively, compound that promotes RNAi can be expressed in a cell, e.g., a cell in a subject, to inhibit expression of XBP-1 or a molecule in a biological pathway involving XBP-1. In contrast to siRNAs, shRNAs mimic the natural precursors of micro RNAs (miRNAs) and enter at the top of the gene silencing pathway. For this reason, shRNAs are believed to mediate gene silencing more efficiently by being fed through the entire natural gene silencing pathway. The term “shRNA”, as used herein, refers to an RNA agent having a stem-loop structure, comprising a first and second region of complementary sequence, the degree of complementarity and orientation of the regions being sufficient such that base pairing occurs between the regions, the first and second regions being joined by a loop region, the loop resulting from a lack of base pairing between nucleotides (or nucleotide analogs) within the loop region. shRNAs may be substrates for the enzyme Dicer, and the products of Dicer cleavage may participate in RNAi. shRNAs may be derived from transcription of an endogenous gene encoding a shRNA, or may be derived from transcription of an exogenous gene introduced into a cell or organism on a vector, e.g., a plasmid vector or a viral vector. An exogenous gene encoding an shRNA can additionally be introduced into a cell or organism using other methods known in the art, e.g., lipofection, nucleofection, etc.
The requisite elements of a shRNA molecule include a first portion and a second portion, having sufficient complementarity to anneal or hybridize to form a duplex or double-stranded stem portion. The two portions need not be fully or perfectly complementary. The first and second “stem” portions are connected by a portion having a sequence that has insufficient sequence complementarity to anneal or hybridize to other portions of the shRNA. This latter portion is referred to as a “loop” portion in the shRNA molecule. The shRNA molecules are processed to generate siRNAs. shRNAs can also include one or more bulges, i.e., extra nucleotides that create a small nucleotide “loop” in a portion of the stem, for example a one-, two- or three-nucleotide loop. The stem portions can be the same length, or one portion can include an overhang of, for example, 1-5 nucleotides.
In certain embodiments, shRNAs of the invention include the sequences of a desired siRNA molecule described supra. In such embodiments, shRNA precursors include in the duplex stem the 21-23 or so nucleotide sequences of the siRNA, desired to be produced in vivo.
Another type of inhibitory compound that can be used to inhibit the expression and/or activity of XBP-1 or a molecule in a biological pathway involving XBP-1 (e.g., HIFα1 is an intracellular antibody specific for said protein. The use of intracellular antibodies to inhibit protein function in a cell is known in the art (see e.g., Carlson, J. R., 1988. Mol. Cell. Biol. 8, 2638-2646; Biocca, S., et al. 1990. EMBO J. 9, 101-108; Werge, T. M., et al. 1990. FEBS Letters 274, 193-198; Carlson, J. R., 1993. Proc. Natl. Acad. Sci. USA 90, 7427-7428; Marasco, W. A., et al. 1993. Proc. Natl. Acad. Sci. USA 90, 7889-7893; Biocca, S., et al. 1994. Bio/Technology 12, 396-399; Chen, S. Y., et al. 1994. Human Gene Therapy 5, 595-601; Duan, L., et al. 1994. Proc. Natl. Acad. Sci. USA 91, 5075-5079; Chen, S. Y., et al. 1994. Proc. Natl. Acad. Sci. USA 91, 5932-5936; Beerli, R. R., et al. 1994. J. Biol. Chem. 269, 23931-23936; Beerli, R. R., et al. 1994. Biochem. Biophys. Res. Commun. 204, 666-672; Mhashilkar, A. M., et al. 1995. EMBO J. 14, 1542-1551; Richardson, J. H., et al. 1995. Proc. Natl. Acad. Sci. USA 92, 3137-3141; PCT Publication No. WO 94/02610 by Marasco et al.; and PCT Publication No. WO 95/03832 by Duan et al.).
To inhibit protein activity using an intracellular antibody, a recombinant expression vector is prepared which encodes the antibody chains in a form such that, upon introduction of the vector into a cell, the antibody chains are expressed as a functional antibody in an intracellular compartment of the cell. For inhibition of transcription factor activity according to the methods of the invention (e.g., inhibition of HIFα1, preferably an intracellular antibody that specifically binds the protein is expressed within the nucleus of the cell. Nuclear expression of an intracellular antibody can be accomplished by removing from the antibody light and heavy chain genes those nucleotide sequences that encode the N-terminal hydrophobic leader sequences and adding nucleotide sequences encoding a nuclear localization signal at either the N- or C-terminus of the light and heavy chain genes (see e.g., Biocca, S., et al. 1990. EMBO J. 9, 101-108; Mhashilkar, A. M., et al. 1995. EMBO J. 14, 1542-1551). A preferred nuclear localization signal to be used for nuclear targeting of the intracellular antibody chains is the nuclear localization signal of SV40 Large T antigen (see Biocca, S., et al. 1990. EMBO J. 9, 101-108; Mhashilkar, A. M., et al. 1995. EMBO J. 14, 1542-1551).
In another embodiment, an inhibitory compound of the invention is a peptidic compound derived from the XBP-1 amino acid sequence or the amino acid sequence of a molecule in a biologicalon pathway involving XBP-1 (e.g., HIFα1). For example, in one embodiment, the inhibitory compound comprises a portion of, e.g., XBP-1 or HIFα1 (or a mimetic thereof) that mediates interaction of XBP-1, for example, with HIF1α such that contact of XBP-1 or HIF1α with this peptidic compound competitively inhibits the interaction of XBP-1 and HIF1α.
The peptidic compounds of the invention can be made intracellularly in cells by introducing into the cells an expression vector encoding the peptide. Such expression vectors can be made by standard techniques using oligonucleotides that encode the amino acid sequence of the peptidic compound. The peptide can be expressed in intracellularly as a fusion with another protein or peptide (e.g., a GST fusion). Alternative to recombinant synthesis of the peptides in the cells, the peptides can be made by chemical synthesis using standard peptide synthesis techniques. Synthesized peptides can then be introduced into cells by a variety of means known in the art for introducing peptides into cells (e.g., liposome and the like).
In addition, dominant negative proteins (e.g., of XBP-1 or HIF1α) can be made which include XBP-1 or HIF1α molecules (e.g., portions or variants thereof) that compete with native (i.e., wild-type) molecules, but which do not have the same biological activity. Such molecules effectively decrease, e.g., XBP-1 or HIF1α activity in a cell.
Other inhibitory agents that can be used to specifically inhibit the activity of an XBP-1 or a molecule in a biological pathway involving XBP-1 are chemical compounds that directly inhibit expression, processing, post-translational modification, and/or activity of, e.g., an XBP-1 (or HIF1α) or inhibit the interaction between, e.g., XBP-1 and HIF1α. Such compounds can be identified using screening assays that select for such compounds, as described in detail above as well as using other art recognized techniques.
In exemplary embodiments, one or more of the above-described inhibitory compounds is formulated according to standard pharmaceutical protocols to produce a pharmaceutical composition for therapeutic use. A pharmaceutical composition of the invention is formulated to be compatible with its intended route of administration.
V. Prognostic Uses
Triple negative breast cancers comprise a very heterogeneous group of cancers. There is conflicting information over prognosis for the various subtypes but it is believed that, at least for more aggressive subtypes, present method of prognosis are poor. It is characterized by distinct molecular, histological and clinical features including a particularly unfavorable prognosis despite increased sensitivity to standard cytotoxic chemotherapy regimens.
The present invention is based, at least in part on the discovery of a gene expression signature indicative of XBP pathway activation that is associated with poor prognosis in patients with TNBC. As used herein, the term “gene expression signature” refers to a specific pattern of detectable signals indicative of gene expression in a sample. In one embodiment, the detectable signals are nucleic acid hybridization signals, for example, signals generated by hybridization of mRNAs in the sample to mRNA nucleic acid probes, e.g. probes having sequence complementarity to the mRNAs. Exemplary detectable labels include, but are not limited to, radioactive labels, fluorescent labels probes, colorometric labels, biotin labels, etc. Probes and/or mRNAs can be immobilized, for example, on a chip, membrane, slide, film, etc. In other embodiments, hybridization can be accomplished with one or more components in solution. In exemplary aspects of the invention, a “gene expression signature” consists of a plurality of signals of varied intensity, the pattern of which is reproducible when detected in replicate samples. In preferred aspects of the invention, a “gene expression signature” consists of a plurality of signals of increased intensity, for example, genes exhibiting increased expression in a TNBC sample or cell. In other aspects of the invention, a “gene expression signature” consists of a plurality of signals of decreased intensity, for example, genes exhibiting decreased expression in a TNBC sample or cell. In still other aspects of the invention, a “gene expression signature” consists of a plurality of signals of increased and decreased intensity, for example, genes exhibiting increased and decreased expression in a TNBC sample or cell.
In exemplary embodiments of the invention, a “gene expression signature” is detected in a test sample (e.g., a biological sample from a patient suspected of having or at risk for developing TNBC, and compared to an appropriate control gene expression signature profile (e.g., a signature from a known TNBC sample or cell). In preferred embodiments, the “test sample” is a sample isolated, obtained or derived from a subject, e.g., a human subject. The term “subject” is intended to include living organisms but preferred subjects are mammals, and in particular, humans. In particularly preferred embodiments, the “test sample” is a sample isolated, obtained or derived from a female subject, e.g., a female human.
In some embodiments, the gene expression signature is associated with a specific stage of TNBC. In some embodiments, the gene expression signature features or consists essentially of mRNAs that are coordinately regulated. These mRNAs may be coordinately regulated, for example, by HIF1α transcriptional activity and can comprise or consist of specific HIF1α targets, i.e., genes expressed as a result of HIF1α transcriptional activity.
In preferred embodiments, a gene expression profiling test is used to analyze the patterns of a plurality of genes, e.g., those set forth in Table 1 within a sample from a TNBC subject, e.g., within a sample of cells from a breast tissue tumor in said subject or from another sample of cancer cells from said subject to help predict how likely it is that breast cancer, e.g., an early-stage breast cancer will recur after initial treatment.
In exemplary embodiments, the invention features diagnostic tests that quantify the likelihood of disease recurrence in subjects, e.g., women subjects with triple-negative breast cancer (TNBC). Such likelihood of disease recurrence is referred to herein as “prognostic significance”. In referred embodiments, the diagnostic tests of the invention further assess the likely benefit from certain types of cancer therapeutics, e.g., chemotherapy. Such assessment is referred to herein as “predictive significance”.
In exemplary aspects of the invention, the diagnostic tests are designed or formatted to analyzes a panel genes within a sample from a TNBC subject, e.g., cells or a tissue sample from a tumor of said subject. From such an analysis, a practitioner or other health professional (e.g., pathologist) can determine, for example, prognostic significance and/or predictive significance. In exemplary embodiments, the test provides for determination of a “recurrence score”. in exemplary embodiments, a recurrence score is a numerical value, e.g., a number between 0 and 100, that corresponds to a specific likelihood of breast cancer recurrence within a certain time period after an initial diagnosis or treatment. In some embodiments, the score corresponds to a likelihood of recurrence within 5 years of the initial diagnosis or treatment. In some embodiments, the score corresponds to a likelihood of recurrence within 10 years of the initial diagnosis or treatment. Based on such a score, a subject (e.g., a TNBC patient) may be classified as low, intermediate or high risk for recurrence. Such a classification may assume that said subject follows a course of treatment including, for example, treatment with anti-hormonal therapy, such as tamoxifen or aromatase inhibitors (e.g., anastrozole), over the period of time following diagnosis or treatment. Depending on the subject risk for recurrence, treatment protocols may include anti-cancer drugs, chemotherapy, treatment with anti-hormonal therapy, such as tamoxifen or aromatase inhibitors, neoadjuvant hormonal therapy (oncology) and the like.
In exemplary embodiments of the invention, the diagnostic test is a noninvasive test that is performed on a small amount of the tissue removed during the original surgery lumpectomy, mastectomy, or core biopsy. In preferred embodiments, the tissue sample (after the surgical procedure) is fixed (e.g., formalin-fixed) and embedded (e.g., paraffin-embedded) so as to be preserved for further diagnostic testing. In other preferred embodiments, the sample (specimen) is fresh tissue sample/specimen. If using a fresh sample, the sample (from an unfixed tumor specimen) can be placed in a preservative solution within a short period of time, e.g., within an hour of surgery. Exemplary preservatives include, but are not limited to, solutions containing RNAse inhibitors.
In exemplary embodiments, a practitioner or other health professional (e.g., pathologist) prepares the samples for testing, (e.g., fixing, embedding, thin-sectioning) samples are analyzed, e.g., in a laboratory or at a testing facility, for example, via RT-PCR to determine expression of a plurality of genes, e.g., 10-20, 20-30, 30-40 or more, from a gene signature of the invention. In preferred embodiments, a panel of genes strongly correlated with recurrence-free survival is features in a diagnostic assay or kit of the invention. In exemplary embodiments of the invention, the results of the featured diagnostic tests can be integrated with other standard laboratory test results to help practitioners and/or health care professionals formulate a treatment plan based on the unique characteristics of the tumor or cell sample.
Pluralities or panels of genes featured in the diagnostic assays and/or kits of the invention can include cancer genes (those correlated with recurrence) and can include, for example, reference or control genes used to normalize the expression of the cancer genes.
Various methodologies of the instant invention include step that involves comparing a value, level, feature, characteristic, property, etc. to a “suitable control”, referred to interchangeably herein as an “appropriate control”. A “suitable control” or “appropriate control” is any control or standard familiar to one of ordinary skill in the art useful for comparison purposes. In one embodiment, a “suitable control” or “appropriate control” is a value or level, of one or more genes (or mRNAs of said genes, or proteins expressed therefrom) as determined in a cell or sample positive for TNBC, as described herein. In another embodiment, a “suitable control” or “appropriate control” is a value or level, of one or more genes (or mRNAs of said genes, or proteins expressed therefrom) as determined in a cell or sample negative for TNBC, e.g., that determined in a cell or organism, e.g., a control or normal cell or organism, exhibiting, for example, normal traits. In yet another embodiment, a “suitable control” or “appropriate control” is a predefined value or level of one or more genes (or mRNAs of said genes, or proteins expressed therefrom).
VI. Screening Assays
In certain aspects, the invention features methods for identifying compounds useful in inhibiting the growth of TNBC cells, such compounds having potential therapeutic use in the treatment of TNBC. As described herein, the instant invention is based, at least in part, on the discovery of a previously unknown role for XPB1 is TNBC, such a role being linked to transcriptional activity of HIF1α. Genome-wide mapping of the XBP1 transcriptional regulatory network revealed that XBP1 regulates the hypoxia response through controlling HIF1α transcriptional activity and the expression of HIF1α targets. Accordingly, in exemplary aspects the invention features methods of identifying for identifying compounds useful in inhibiting the growth of TNBC cells, the methods featuring screening or assaying for compounds that modulate, e.g., activate or increase, or inhibit or decrease, the interaction of XBP1 and HIF 1α, or biologically active portions thereof. In exemplary aspects, the methods comprise: providing an indicator composition comprising XBP1 and HIF1α, or biologically active portions thereof; contacting the indicator composition with each member of a library of test compounds; and selecting from the library of test compounds a compound of interest that decreases the interaction of XBP1 and HIF1α, or biologically active portions thereof, wherein the ability of a compound to inhibit growth of TNBC cells is indicated by a decrease in the interaction as compared to the amount of interaction in the absence of the compound
As used herein, the term “contacting” (i.e., contacting a cell e.g. a cell, with a compound) includes incubating the compound and the cell together in vitro (e.g., adding the compound to cells in culture) as well as administering the compound to a subject such that the compound and cells of the subject are contacted in vivo. The term “contacting” does not include exposure of cells to an XBP-1 modulator that may occur naturally in a subject (i.e., exposure that may occur as a result of a natural physiological process).
As used herein, the term “test compound” refers to a compound that has not previously been identified as, or recognized to be, a modulator of the activity being tested. The term “library of test compounds” refers to a panel comprising a multiplicity of test compounds.
As used herein, the term “indicator composition” refers to a composition that includes a protein of interest (e.g., XBP-1 or a molecule in a biological pathway involving XBP-1, e.g., HIF1α), for example, a cell that naturally expresses the protein, a cell that has been engineered to express the protein by introducing one or more of expression vectors encoding the protein(s) into the cell, or a cell free composition that contains the protein(s) (e.g., purified naturally-occurring protein or recombinantly-engineered protein(s)).
As used herein, the term “cell” includes prokaryotic and eukaryotic cells. In one embodiment, a cell of the invention is a bacterial cell. In another embodiment, a cell of the invention is a fungal cell, such as a yeast cell. In another embodiment, a cell of the invention is a vertebrate cell, e.g., an avian or mammalian cell. In a preferred embodiment, a cell of the invention is a murine or human cell. As used herein, the term “engineered” (as in an engineered cell) refers to a cell into which a nucleic acid molecule e.g., encoding an XBP-1 protein (e.g., a spliced and/or unspliced form of XBP-1) has been introduced.
As used herein, the term “cell free composition” refers to an isolated composition, which does not contain intact cells. Examples of cell free compositions include cell extracts and compositions containing isolated proteins.
The ability of the test compound to modulate XBP-1 binding to HIF1α can also be determined. Determining the ability of the test compound to modulate XBP-binding to HIF1α can be accomplished, for example, by coupling the HIF1α with a radioisotope or enzymatic label such that binding of HIF1α to XBP-1 can be determined by detecting the labeled HIF1α in a complex. Alternatively, XBP-1 could be coupled with a radioisotope or enzymatic label to monitor the ability of a test compound to modulate XBP-1 binding to HIF1α in a complex. Determining the ability of the test compound to bind to XBP-1 (or HIF1α) can be accomplished, for example, by coupling the compound with a radioisotope or enzymatic label such that binding of the compound to XBP-1 (or HIF1α) can be determined by detecting the labeled compound in a complex. For example, targets can be labeled with ¹²⁵I, ³⁵S, ¹⁴C, or ³H, either directly or indirectly, and the radioisotope detected by direct counting of radioemmission or by scintillation counting. Alternatively, compounds can be labeled, e.g., with, for example, horseradish peroxidase, alkaline phosphatase, or luciferase, and the enzymatic label detected by determination of conversion of an appropriate substrate to product.
It is also within the scope of this invention to determine the ability of a compound to interact with XBP-1 or HIF1α without the labeling of any of the interactants. For example, a microphysiometer can be used to detect the interaction of a compound with XBP-1 or HIF1α without the labeling of either the compound or the XBP-1 or HIF1α (McConnell, H. M., et al. 1992. Science 257, 1906-1912). As used herein, a “microphysiometer” (e.g., Cytosensor) is an analytical instrument that measures the rate at which a cell acidifies its environment using a light-addressable potentiometric sensor (LAPS). Changes in this acidification rate can be used as an indicator of the interaction between a compound and XBP-1 or HIF1α.
The cells used in the instant assays can be eukaryotic or prokaryotic in origin. For example, in one embodiment, the cell is a bacterial cell. In another embodiment, the cell is a fungal cell, e.g., a yeast cell. In another embodiment, the cell is a vertebrate cell, e.g., an avian or a mammalian cell. In a preferred embodiment, the cell is a human cell. The cells of the invention can express endogenous XBP-1 or HIF1α or can be engineered to do so. For example, a cell that has been engineered to express the XBP-1 protein and/or HIF1α can be produced by introducing into the cell an expression vector encoding the protein. Recombinant expression vectors that can be used for expression of XBP-1 or a HIF1α.
In another embodiment, the indicator composition is a cell free composition. XBP-1 or HIF1α expressed by recombinant methods in a host cells or culture medium can be isolated from the host cells, or cell culture medium using standard methods for protein purification. For example, ion-exchange chromatography, gel filtration chromatography, ultrafiltration, electrophoresis, and immunoaffinity purification with antibodies can be used to produce a purified or semi-purified protein that can be used in a cell free composition. Alternatively, a lysate or an extract of cells expressing the protein of interest can be prepared for use as cell-free composition.
In one embodiment, the amount of binding of XBP-1 to HIF1α in the presence of the test compound is greater than the amount of binding of XBP-1 binding to HIF1α in the absence of the test compound, in which case the test compound is identified as a compound that enhances binding of XBP-1 to HIF1α. In another embodiment, the amount of binding of the XBP-1 to HIF1α in the presence of the test compound is less than the amount of binding of the XBP-1 to HIF1α in the absence of the test compound, in which case the test compound is identified as a compound that inhibits binding of XBP-1 to HIF 1α.
Binding of the test compound to XBP-1 or HIFα1 can be determined either directly or indirectly as described above. Determining the ability of XBP-1 (or HIF1α) protein to bind to a test compound can also be accomplished using a technology such as real-time Biomolecular Interaction Analysis (BIA) (Sjolander, S., et al. 1991. Anal. Chem. 63, 2338-2345; Szabo, A., et al. 1995. Curr. Opin. Struct. Biol. 5, 99-705). As used herein, “BIA” is a technology for studying biospecific interactions in real time, without labeling any of the interactants (e.g., BIAcore). Changes in the optical phenomenon of surface plasmon resonance (SPR) can be used as an indication of real-time reactions between biological molecules.
In the methods of the invention for identifying test compounds that modulate an interaction between XBP-1 protein and HIF1α, the complete XBP-1 (or e.g HIF1α) protein can be used in the method, or, alternatively, only portions of the protein can be used. In one embodiment of the above assay methods of the present invention, it may be desirable to immobilize either XBP-1 (or HIF 1α) for example, to facilitate separation of complexed from uncomplexed forms of one or both of the proteins, or to accommodate automation of the assay. Binding of a test compound to a XBP-1 with HIF1α in the presence and absence of a test compound, can be accomplished in any vessel suitable for containing the reactants. Examples of such vessels include microtitre plates, test tubes, and micro-centrifuge tubes. In one embodiment, a fusion protein can be provided in which a domain that allows one or both of the proteins to be bound to a matrix is added to one or more of the molecules. For example, glutathione-S-transferase fusion proteins or glutathione-S-transferase/target fusion proteins can be adsorbed onto glutathione sepharose beads (Sigma Chemical, St. Louis, Mo.) or glutathione derivatized microtitre plates, which are then combined with the test compound or the test compound and either the non-adsorbed target protein or XBP-1 (or HIF1α) protein, and the mixture incubated under conditions conducive to complex formation (e.g., at physiological conditions for salt and pH). Following incubation, the beads or microtitre plate wells are washed to remove any unbound components, the matrix is immobilized in the case of beads, and complex formation is determined either directly or indirectly, for example, as described above. Alternatively, the complexes can be dissociated from the matrix, and the level of binding or activity determined using standard techniques.
Other techniques for immobilizing proteins on matrices can also be used in the screening assays of the invention. For example, either an XBP-1 protein or HIF1α can be immobilized utilizing conjugation of biotin and streptavidin. Biotinylated protein or target molecules can be prepared from biotin-NHS (N-hydroxy-succinimide) using techniques known in the art (e.g., biotinylation kit, Pierce Chemicals, Rockford, Ill.), and immobilized in the wells of streptavidin-coated 96 well plates (Pierce Chemical). Alternatively, antibodies which are reactive with protein but which do not interfere with binding of the proteins can be derivatized to the wells of the plate, and unbound XBP-1 or HIF1α protein is trapped in the wells by antibody conjugation. Methods for detecting such complexes, in addition to those described above for the GST-immobilized complexes, include immunodetection of complexes using antibodies reactive with XBP-1 or HIF1α, as well as enzyme-linked assays which rely on detecting an enzymatic activity associated with the XBP-1 or HIF1α.
Another aspect of the invention pertains to kits for carrying out the screening assays, modulatory methods or diagnostic assays of the invention.
This invention is further illustrated by the following examples which should not be construed as limiting. The contents of all references, patents, and published patent applications cited throughout this application, as well as the figures and the sequence listing, are hereby incorporated by reference.

EXAMPLES

Example 1

The UPR is Activated in Human Breast Cancer Patients

To determine whether the UPR is activated in breast cancer, we used immunohistochemistry (IHC) to examine the phosphorylation of PERK, a marker of UPR activation, in human primary breast tumor samples. By staining breast cancer tissue microarrays (TMA) containing 66 normal breast tissue samples and 40 tumor tissue samples, we found that PERK was preferentially phosphorylated in breast tumors, but not in normal breast tissue (FIG. 1A, 1B), suggesting that activation of the UPR occurs specifically in tumors. Next, the same TMA were stained with antibodies specifically recognizing phosphorylation of eukaryotic translational initiation factor 2α (eIF2α), another marker of UPR activation. Similarly, eIF2α was phosphorylated in malignant breast tumors but not normal breast tissue (FIG. 1C, 1D). Thus, the UPR is preferentially activated in breast tumors.

Example 2

XBP1 is Required for Transformation of Immortalized Mammary Epithelial Cells

The IRE1-XBP1 axis of the UPR shows robust conservation from yeast to metazoans, including humans. To investigate the role of XBP1 in cellular transformation, we used MCF10A immortalized mammary epithelial cells that express ER-Src, a fusion of the Src kinase oncoprotein (v-Src) and the ligand binding domain of the estrogen receptor. Treatment of these cells with tamoxifen (TAM) for 36 hr results in neoplastic transformation, including the ability to form colonies in soft agar, increased motility and invasive ability, and tumor formation upon injection into nude mice (Iliopoulos, D., et al. 2009. Cell 139, 693-706). Knockdown of XBP1 expression with a highly effective shRNA (Figure S1) blocked the neoplastic transformation of MCF10A ER-Src cells (FIG. 2A). Furthermore, XBP1 silencing reduced the invasiveness and the ability of MCF10A ER-Src cells to form colonies in soft agar and tumors in immunodeficient mice (FIG. 2B-D). We tested the ability of enforced XBP1 expression to transform MCF10A cells by overexpression of the XBP1 spliced form (XBP1s) in MCF10A ER-Src cells in the absence of tamoxifen. XBP1 overexpression was sufficient to induce transformation in the absence of tamoxifen (FIG. 2E). Furthermore, XBP1s overexpression increased colony formation in a soft agar assay (FIG. 2F). Collectively, these results demonstrate that XBP1 is both necessary and sufficient for the transformation of mammary epithelial cells.

Example 3

XBP1 Inhibition Blocks Breast Cancer Cell Growth and Invasiveness Both Ex Vivo and In Vivo; XBP1 Silencing Blocks Triple Negative Breast Cancer Progression

To further characterize the function of XBP1 in breast cancer, we first determined the activation status of XBP1 in different breast cancer cell lines. Breast cancers can be classified as luminal or basal-like, depending on their expression of different cytokeratins (Perou, C. M., et al. 2000. Nature 406, 747-752; Vargo-Gogola, T., et al. 2007. Cancer 7, 659-672). Unexpectedly, XBP1 was preferentially spliced and activated in basal-like breast cancer cells (FIG. 3A), which harbor a transcriptome similar to that of triple negative breast cancer (TNBC), a subtype of breast cancer that is extremely aggressive and difficult to target due to the lack of expression of the estrogen (ER), progesterone (PR) and human epidermal growth factor 2 (HER2) receptors (Foulkes, W. D., et al. 2010. N Engl J Med 363, 1938-1948). In particular, while XBP1 expression was readily detected in both luminal and basal-like breast cancer cells, the level of its spliced (activated) form was higher in the latter cell type (FIG. 1A), which comprises primarily TNBC (Perou, C. M., et al. 2000. Nature 406, 747-52; Vargo-Gogola, T., et al. 2007. Nat Rev Cancer 7, 659-72; Herschkowitz, J. I., et al. 2007. Genome Biol 8, R76). Furthermore, silencing XBP1 expression decreased the ability of different breast cancer cell lines to form colonies in soft agar (FIG. 3B).
TNBC is a highly aggressive subtype of breast cancer characterized by the absence of estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor 2 (HER2) expression (Foulkes, W. D., et al. 2010. N Engl J Med 363, 1938-48). We next demonstrated that silencing of XBP1 significantly impaired soft agar colony formation (FIG. 1B) and invasiveness (FIG. 1C) of multiple TNBC cell lines (MDA-MB-231, MBD-MB-468, HBL-100, MDA-MB-436, MDA-MB-157), suggesting a potential role of XBP1 in the regulation of anchorage-independent growth and invasiveness of TNBC.
Interestingly, knockdown of XBP1 was more effective in suppressing the proliferation of basal-like (MDA-MB-231, MBD-MB-468, HBL-100, MDA-MB-157, MDA-MB-435, MDA-MB-436, SUM-159) than luminal (MCF7, BT-474, SKBR3, T47D, MDA-MB-361) breast cancer cell lines, consistent with the preferential splicing of XBP1 in basal-like cells. Similarly, knockdown of XBP1 decreased the invasiveness of breast cancer cell lines, a phenotype that was more dramatic in basal-like lines (FIG. 3C). These data suggest that XBP1 regulates the growth and invasiveness of breast cancer cells, especially basal-like breast cancer cells.
To assess the function of XBP1 in vivo, we established an orthotopic xenograft mouse model with inducible expression of shRNA against XBP1. In particular, we infected MDA-MB-231 cells, a TNBC cell line, with lentiviruses encoding XBP1 shRNAs under the control of a doxycycline-inducible promoter. Cells infected with a lentivirus encoding a scrambled LacZ shRNA served as a control. Doxycycline treatment of cells infected with the XBP1 shRNA lentivirus led to an 85% reduction in XBP1 mRNA levels compared to cells grown in the absence of doxycycline (FIG. 3D).
Next, these MDA-MB-231 cells infected with the shRNA lentiviruses were further infected with a retrovirus encoding luciferase. After injection with retroviruses, these cells were implanted (injected) orthotopically in the mammary glands of NOD/SCID/IL2Rγ−/− mice. The kinetics of tumor growth were monitored with bioluminescent imaging. At > two weeks after implantation (19 days), prior to induction of the XBP1 shRNA, XBP1 shRNA and control tumors exhibited similar luciferase signals (FIG. 3E). These mice were then fed chow containing doxycycline to induce the XBP1 shRNA and serially monitored using bioluminescence. After 4 weeks of XBP1 depletion a significant inhibition of tumor growth was observed (FIGS. 3E and 3F). XBP1 was efficiently silenced in the tumor (FIG. 3I). While tumors expressing control shRNA (n=8) began to metastasize to the lungs 9 weeks after transplantation, no metastasis was observed in the XBP1 shRNA xenograft tumors (n=8) (FIG. 3E). To rule out off-target effects of the XBP1 shRNA, the same assays were conducted with another inducible XBP1 shRNA construct targeting a different region of XBP1 (FIG. 3J), which yielded similar results (FIG. 3K). To exclude the possibility of cell line specific effects, subcutaneous xenograft experiments were performed using two other TNBC cell lines: MDA-MB-436 and HBL-100 cells. As expected, XBP1 silencing significantly repressed the formation of MDA-MB-436 and HBL-100 TNBC-derived tumors (FIG. 3L). Importantly, we examined the functional relevance of XBP1 in primary human breast tumor cells. We inhibited XBP1 by siRNA in a patient-derived TNBC xenograft model (BCM-2147). Silencing of XBP1 in this model significantly decreased tumor incidence (FIG. 3G) and suppressed tumor growth (FIG. 3H), further supporting the role of XBP1 in TNBC. Collectively, these results demonstrate that loss of XBP1 suppresses the growth and metastasis (tumorigenicity and progression) of human triple negative breast tumors.

Example 4

XBP1 is Required to Sustain Tumor Initiating Cell (TIC) Self-Renewal; XBP1 is Required for Tumor Initiating Cells

Previous studies have shown that basal-like breast cancer cells are more aggressive than luminal cells due to increased numbers of a stem cell-like CD44^high/CD24^lowsubpopulation, termed tumor initiating cells (TICs) (Al-Hajj, M., et al. 2003. Proc Natl Acad Sci USA 100, 3983-3988; Mani, S. A., et al. 2008. Cell 133, 704-715). To interrogate the effect of XBP1 on TICs, we used a model of breast epithelial cells (MCF10A) carrying an inducible Src oncogene (ER-Src), in which the Src kinase oncoprotein (v-Src) was fused with the ligand binding domain of the estrogen receptor (Iliopoulos, D., et al. 2009. Cell 139, 693-706). Recently, it has been shown that during transformation of MCF10A ER-Src cells, there is formation of a CD44^high/CD24^lowpopulation with TIC characteristics (Iliopoulos et al., 2011). In particular, Treatment of these cells with tamoxifen (TAM) for 24-36 hr results in neoplastic transformation and the gain of a CD44high/CD24low population with tumor-initiating property (Iliopoulos, D., et al. 2011. Proc Natl Acad Sci USA 108, 1397-402). In transformed MCF10A ER-Src cells, knockdown of XBP1 blocked the formation of the CD44^high/CD24^lowER-Src TIC population (reducing the CD44^high/CD24^lowTIC fraction)(FIG. 4B and Figures H-I). In this system, XBP1 was more highly spliced in TICs (CD44^high/CD24^low) relative to non-TICs (NTICs) (FIG. 4A). XBP1 silencing also suppressed the ability of transformed MCF10A ER-Src cells to form mammospheres (FIG. 4C), an assay used to assess the self-renewal of breast TICs (Dontu, G., et al. 2003. Genes Dev 17, 1253-1270). These phenotypes were not due to a direct effect of XBP1 on cell viability (FIGS. 4J-K).
To test if expression of XBP1 was sufficient to induce TIC properties in NTICs, XBP1s was overexpressed in CD44^low/CD24^highNTICs derived from MCF10A ER-Src cells. This induced the formation of a population with a TIC-like CD44^high/CD24^lowsurface phenotype and enhanced mammosphere forming ability.
Tumorigenicity in a murine host is the gold standard for evaluating the stem cell-like properties of TICs (Clarke, M. F., et al. (2006). Cancer Res 66, 9339-9344). To further investigate if the TICs induced by XBP1s expression in NTICs also display TIC properties in a murine tumor formation assay, NTICs or NTICs with enforced expression of XBP1s (XBP1s-NTIC) were injected into NOD/SCID mice at a range of dilutions. Remarkably, as few as 100 XBP1s-NTICs cells were able to generate a tumor, whereas control NTICs failed to form tumors at any dilution. Thus, under these conditions, XBP1s is sufficient for the induction of functional breast TICs.
In a limiting dilution experiment, TAM-treated MCF10A-ER-Src cells bearing control shRNA were able to initiate tumors when as few as 1×10⁴or 1×10⁵cells were implanted. However, XBP1-depleted cells showed complete loss of tumor-seeding ability even when 1×10⁶cells were xenografted (FIG. 4D).
In addition to MCF10A ER-Src cells, we examined the effects of XBP1 inhibition in TICs derived from breast cancer cell lines. XBP1 inhibition suppressed the growth of mammospheres derived from MDA-MB-231, MDA-MB-468 and MDA-MB-436 cells (FIG. 4F).
To evaluate the functional relevance of XBP1 in human cancer patients, we sorted the CD44^high/CD24^lowsubpopulation directly from human TNBC patient samples, and confirmed XBP1 splicing to be elevated in this fraction compared to the CD44low/CD24high cells (FIG. 4E). Infection of the CD44^high/CD24^lowcells with lentivirus expressing XBP1 shRNA, inhibited the formation of mammospheres derived from a number of patient-derived TNBC tissues (FIG. 4F). Conversely, overexpression of XBP1s in NTICs (CD44^high/CD24^low) sorted from primary human TNBC (or derived from breast cancer cell lines) transformed them into TICs as based on surface phenotype. Remarkably, these XBP1s-induced TICs are able to form tumors in immunodeficient mice at very low dilutions (as low as 10 xenografted cells) whereas none of the control parental NTICs were tumorigenic (FIG. 4G).
Collectively, these data establish a critical and unexpected role of XBP1 in TICs, likely contributing to its function in promoting triple-negative breast cancer

Example 5

XBP1 Silencing Increases Sensitivity and Reduces Resistance to Chemotherapy; Inhibition of XBP1 Suppresses Tumor Relapse

Chemotherapy is the only systemic therapy currently used clinically to treat TNBC. However, patients with TNBC have the highest rate of relapse within 1-3 years despite the use of adjuvant chemotherapy (Lehmann, B. D., et al. 2011. J Clin Invest 121, 2750-67). Moreover, TICs are resistant to chemotherapy and are believed to be responsible for tumor relapse after chemotherapy (Dean, M., et al., 2005. Nat Rev Cancer 5, 275-284). Given that XBP1 appears to induce TIC differentiation, the role of XBP1 in mediating the relapse of the MDA-MB-231 xenograft tumor after chemotherapy was evaluated. It was believed that this approach would yield further insights into the function of XBP1 in TNBC. Treatment of MDA-MB-231 xenograft tumors with doxorubicin (i.p.) every 5 days, from day 15 until day 30, suppressed tumor growth (FIG. 4M). Relapse from treatment occurred on day 60, i.e., Tumor relapse after treatment was detected from day 60 onwards. Strikingly, combinatorial treatment with doxorubicin and XBP1 shRNA not only blocked tumor growth but also inhibited tumor relapse (FIG. 4N).
The presence of tumor initiating cells (TICs), characterized by the cell surface phenotype CD44high/CD24low and the expression of ALDH1 (Ginestier, C., et al. 2007. Cell Stem Cell 1, 555-67), are thought to play a role in chemotherapy resistance and tumor relapse after systemic adjuvant therapy (Dean, M., et al. 2005. Nat Rev Cancer 5, 275-84; Al-Hajj, M., et al. 2003. Proc Natl Acad Sci USA 100, 3983-8; Creighton, C. J., et al. 2009. Proc Natl Acad Sci USA 106, 13820-5; Li, X., et al. 2008. J Natl Cancer Inst 100, 672-9). In order to test whether suppression of tumor relapse (this increased sensitivity to chemotherapy) is due to an effect of XBP1 on TICs, we examined mammosphere-forming ability of cells (the number of mammospheres) derived from the treated tumors (day 20). Mammosphere assays are used to assess the activity of breast TICs in vitro (Dontu, G., et al. 2003. Genes Dev 17, 1253-70). Consistent with the previously observed enrichment of TIC following chemotherapy (Creighton, C. J., et al. 2009. Proc Natl Acad Sci USA 106, 13820-5), mammosphere formation was increased in cells derived from doxorubicin treated tumors (FIG. 4L). Intriguingly, tumors treated with doxorubicin in combination with XBP1 knockdown demonstrated substantially suppressed mammosphere growth (FIG. 4M), suggesting that XBP1 silencing blunted chemotherapy-induced expansion of the TIC pool. Thus, the combination of chemotherapy and XBP1 knockdown suppresses breast tumor growth and prolongs remission in breast xenografts.
Collectively, these data demonstrate that XBP1 is required to sustain TIC self-renewal in breast cancer.

Example 6

XBP1 Interacts with HIF1α and Co-Occupies the Promoters of HIF1α Targets; HIF1α is a Co-Regulator of XBP1 in TNBC

Given the importance of XBP1 in the breast cancer models above and to further understand how XBP1 contributes to TNBC, we sought to identify transcriptional networks regulated by XBP1 and to dissect the underlying mechanism by mapping the physiological targets of XBP1s using ChIP coupled with high-throughput sequencing (ChIP-seq). Tumor cells are exposed to hypoxia and glucose deprivation, and these factors are appreciated to have a large impact on tumor pathophysiology (Semenza, G. L. 2003. Nat Rev Cancer 3, 721-32). XBP1s was highly expressed in MDA-MB-231 cells by exposure to the physiological stressors (FIG. 5J) To examine if these stressors of cellular physiology might induce XBP1 activation via splicing, MDA-MB-231 cells were grown in hypoxic and glucose deprivation conditions for 24 h. Exposure to hypoxia and glucose deprivation induced splicing of XBP1, and this resulted in a corresponding increase in the signal intensity detected in ChIP-seq experiments. Using a ChIP-seq approach (using a polyclonal antibody specifically recognizing the XBP1s protein), we identified a total of 6317 high-confidence XBP1 binding sites in MDA-MB-231 cells. 13.9% of the binding sites mapped to promoters, and 73.6% were found at distal intergenic and intronic regions (FIG. 5K). Notably, the overlap of the genes bound by XBP1 in MDA-MB-231 cells versus those bound in plasma cells or pancreatic beta cells was small (Acosta-Alvear, D., et al. 2007. Mol Cell 27, 53-66). Therefore, our study revealed a unique repertoire of XBP1 binding sites specific for TNBCs. As expected, XBP1 extensively bound to genes involved in the UPR pathway, such as DNAJB9, HSPA5, and EDEM. By performing microarray and gene set enrichment analysis (GSEA) of genes differentially expressed upon XBP1 depletion in MDA-MB-231 cells, we found that the UPR pathway was among the most enriched categories, with significant enrichment of genes involved in ER stress and UPR pathways indicating that XBP1 directly regulates the UPR in TNBC cells.
To determine the in vivo sequence specificity of XBP1, we derived the consensus sequence motifs by using a motif-discovery algorithm MDScan (Liu, X. S., et al. 2002. Nat Biotechnol 20, 835-839). Notably, the predominant motif found was a perfect match to the XBP1 consensus site GC/ACACGT (FIG. 5L), confirming the validity of the ChIP-seq dataset. Remarkably, a HIF1α binding motif showed statistically significant enrichment in our dataset (enrichment of the HIF1α binding motif in the XBP1 sites (p=1.0×10⁻³⁰)) (FIG. 5A), suggesting potential cooperation between HIF1α and XBP1, e.g., that HIF1α frequently co-localizes to the same transcriptional regulatory elements as XBP1. HIF1α is a ubiquitously expressed, O₂dependent subunit of Hypoxia Induced Factor (HIF1), known to play essential roles in TNBC and in breast TICs self-renewal (Schwab, L. P., et al. 2012. Breast Cancer Res 14, R6). The enrichment of the HIF1α motif in the XBP1 ChIP-seq dataset raised the possibility that XBP1 and HIF1α might interact in the same transcriptional complex.
To assess this possibility, Flag-tagged HIF1α was co-expressed with XBP1s in 293T cells cultured under hypoxia. Treatment of cells with the proteasome inhibitor MG132 for 16 hours was necessary to inhibit the basal turnover of HIF1α. Extracts were harvested and immunoprecipitated with M2 FLAG antibody, and HIF1α was found to co-precipitate with XBP1s (FIG. 5B). This interaction could also be observed with endogenous proteins in the context of two TNBC cell lines. MDA-MB-231 and Hs578T cells were treated with tunicamycin (TM), a potent pharmacologic ER-stress inducer that triggers robust XBP1 splicing. Nuclear extracts were harvested, and immunoprecipitation using an anti-HIF1α antibody demonstrated the co-precipitation of XBP1 (FIGS. 5C and 5M). Thus, endogenous XBP1 interacts with HIF1α in the nucleus.
To extend these results, we next asked whether XBP1 binds together with HIF1α specifically at the site of HIF1α target genes. Direct ChIP-qPCR was performed to examine the co-occupancy XBP1 and HIF α at several well known HIF1α direct targets including VEGFA, PDK1, DDIT4, JMJD1A and JMJD2C (Xia, X, et al. 2009. Proc Natl Acad Sci USA 106, 4260-4265). As shown in FIGS. 5D-F, and FIG. 5N, both XBP1 and HIF1α bind to the promoters of VEGFA, PDK1, DDIT4, JMJD1A and JMJD2C under hypoxic conditions, whereas control GST ChIP did not show any enrichment. Next, we ascertained the functional contribution of XBP1 to the regulation of HIF1α targets. As the physiologic response to tissue hypoxia is initiated by the binding of the HIF-1 transcription factor to the hypoxia response element (HRE) (Semenza, G. L., 2001. Cell 107, 1-3), a luciferase construct containing three copies of HRE (FIG. 5G) was co-transfected together with a construct encoding XBP1s into MDA-MB-231 cells. XBP1s was able to transactivate the HRE reporter in a dose dependent manner, whereas the empty vector had no effect (FIG. 5H). Conversely, depletion of XBP1 by two independent shRNA constructs dramatically reduced HRE activity under hypoxic conditions (FIG. 5I). Taken together, these data demonstrate that XBP1s interacts with HIF1α and in turn the two collaborate to regulate the promoters of HIF1α targets.

Example 7

XBP1 Regulates the Response to Hypoxia (the Hypoxia Response Pathway)

Next, we profiled the differential transcriptome regulated by XBP1 silencing in MDA-MB-231 cells using gene expression microarray analysis. In particular, to identify the transcriptional programs regulated by XBP1, we perturbed XBP1 expression in MDA-MB-231 cells by shRNA and examined the effects on gene expression by microarray analysis under the same conditions as the above ChIP-seq assay. Gene set enrichment analysis (GSEA) identified significant enrichment of genes in the hypoxia response pathway (FIG. 6A, B). To verify the regulation of the hypoxia response by XBP1, we exposed cells to hypoxia, and demonstrated that depletion of XBP1 resulted in downregulation of HIF1α targets VEGFA, PDK1, GLUT1 and DDIT4 expression (FIG. 6C, Figure S4). This result indicates that XBP1 regulates the expression of HIF1α targets under hypoxic conditions. Performing the same experiment in another TNBC cell line, HS578T, yielded similar results (FIG. 6D, FIG. 6H). Thus, XBP1 is an essential mediator of the hypoxic response via its key function in regulating the expression of HIF1α target genes.
To further understand the mechanism by which XBP1 regulates HIF1α transcriptional pathways, we first examined the correlation between XBP1 and HIF1α at genome-wide level. As shown in FIG. 6E, a high level of XBP1 occupancy was associated with increased occurrence of the HIF1α motif in TNBC (p<1×10-5), suggesting a requirement of XBP1 for HIF1α occupancy. Next, we depleted XBP1 and examined the occupancy of HIF1α at HIF1α-XBP1 co-bound sites near well-established HIF1α targets. MDA-MB-231 cells infected with control shRNA or XBP1 shRNA were treated for 24 h under hypoxic conditions, and the extracts were subjected to ChIP. As expected, XBP1 knockdown reduced the occupancy of XBP1 on co-bound sites (FIG. 6I). HIF1α levels were not altered by XBP1 depletion (FIG. 6J). XBP1 depletion substantially attenuated HIF1α occupancy at the targets (FIG. 6F), suggesting that the recruitment of HIF 1α is dependent on XBP1
To further understand the relationship between XBP1, HIF1α and the basal transcription machinery, we examined the recruitment of RNA polymerase II at the promoters of HIF1α target genes. In particular, we carried out ChIP against RNA polymerase II. Consistent with the reduction in HIF1α target transcripts after XBP1 depletion, the binding of RNA polymerase II to the XBP1-HIF1α co-bound sites was also significantly reduced in the absence of XBP1 (FIG. 6G). As a control, RNA polymerase II binding to β-actin, which is not occupied by XBP1, was not altered (FIG. 6G). Collectively, these data suggest that XBP1 regulates HIF1α transcriptional activity by controlling the binding of HIF1α to its targets and by the recruitment of RNA polymerase II.

Example 8

XBP1Activation is Associated with Human Breast Cancer Prognosis

Through integrated analysis of XBP1 ChIP-seq data and gene expression profiles, we identified a plurality of genes that are directly bound and up-regulated by XBP1. This gene set was defined as the XBP1 signature (FIG. 7A). The gene signature is also defines by the genes set forth in Table 1.

TABLE 1

XBP1 gene signature

Refseq	Gene Symbol	RP value	FDR

NM_005080	XBP1	0.000308166	0.046
NM_001079539	XBP1	0.000616333	0.033
NM_173354	SIK1	0.007660895	0.026
NM_001177	ARL1	0.007856733	0.0225
NM_015021	ZNF292	0.010608268	0.0192
NM_001113182	BRD2	0.018319709	0.017
NM_005104	BRD2	0.020101806	0.016444444
NM_024116	TAF1D	0.021276877	0.014727273
NM_005321	HIST1H1E	0.026452097	0.013733333
NM_134470	IL1RAP	0.026969403	0.01425
NM_177444	PPFIBP1	0.031355838	0.014380952
NM_144949	SOCSS	0.031921637	0.014818182
NM_014011	SOCSS	0.032025593	0.014956522
NM_014840	NUAK1	0.032058195	0.015166667
NM_003410	ZFX	0.032864181	0.015851852
NM_012421	RLF	0.035372081	0.017483871
NM_002610	PDK1	0.036732609	0.018571429
NM_001259	CDK6	0.037469791	0.018666667
NM_001134368	SLC6A6	0.037723647	0.018918919
NM_003670	BHLHE40	0.038232511	0.018894737
NM_006265	RAD21	0.039985705	0.0195
NM_012330	MYST4	0.041773318	0.020095238
NM_004792	PPIG	0.041827844	0.020232558
NM_006699	MAN1A2	0.042287349	0.020347826
NM_006427	SIVA1	0.043459113	0.020857143
NM_001145306	CDK6	0.046056189	0.022346154
NM_021709	SIVA1	0.046086078	0.022566038
NR_027856	CLK1	0.047788652	0.023758621
NR_027855	CLK1	0.048191354	0.024305085
NM_004071	CLK1	0.048592674	0.024833333
NM_001162407	CLK1	0.04931818	0.025419355
NM_001135581	SLC1A4	0.050712807	0.026338462
NM_003286	TOP1	0.051189956	0.026848485
NM_018463	ITFG2	0.05599817	0.028027778
NM_020791	TAOK1	0.056306819	0.028273973
NM_004642	CDK2AP1	0.058411965	0.028973684
NM_004354	CCNG2	0.059493678	0.029777778
NM_006810	PDIA5	0.059980932	0.030292683
NM_003038	SLC1A4	0.060388037	0.030952381
NM_033026	PCLO	0.060740842	0.031035294
NM_001031723	DNAJB14	0.063887884	0.032593407
NM_022044	SDF2L1	0.068126387	0.034
NM_012328	DNAJB9	0.06931484	0.034632653
NM_018386	PCID2	0.070132067	0.035030303
NM_001127203	PCID2	0.07045563	0.03532
NM_052834	WDR7	0.07101882	0.035960784
NM_015285	WDR7	0.071327044	0.036368932
NM_003432	ZNF131	0.072904644	0.037364486
NM_018725	IL17RB	0.073397321	0.038558559
NM_014629	ARHGEF10	0.076846594	0.040537815
NM_005834	TIMM17B	0.078254854	0.041289256
NM_001127202	PCID2	0.078373692	0.04157377
NM_178812	MTDH	0.078864716	0.042080645
NM_015565	RNF160	0.079551116	0.042384
NM_173214	NFAT5	0.079829972	0.042692913
NM_138714	NFAT5	0.080330418	0.043410853
NM_138713	NFAT5	0.080828942	0.044333333
NM_020182	PMEPA1	0.080955732	0.044820896
NM_006599	NFAT5	0.081325577	0.045066667
NM_001113178	NFAT5	0.081820358	0.046246377
NM_001307	CLDN7	0.08389827	0.046628571
NM_206866	BACH1	0.085201186	0.046822695
NM_001006622	WDR33	0.085351526	0.047070423
NM_021913	AXL	0.085668045	0.047496503
NM_001080512	BICC1	0.086429554	0.047708333
NM_014607	UBXN4	0.086642454	0.048246575
NM_001699	AXL	0.086705455	0.048489796
NM_001186	BACH1	0.087932789	0.050313725
NM_001706	BCL6	0.089419251	0.051261146
NM_001042370	TROVE2	0.089715599	0.051594937
NM_005734	HIPK3	0.09027961	0.0521875
NM_001048200	HIPK3	0.091696467	0.054060606
NM_004641	MLLT10	0.095742597	0.057737143
NM_020354	ENTPD7	0.096224931	0.058034091
NM_001009569	MLLT10	0.097206843	0.058905028
NM_004600	TROVE2	0.097239566	0.059233333
NM_001042369	TROVE2	0.097349374	0.059436464
NM_015659	RSL1D1	0.097841743	0.059747253
NM_032991	CASP3	0.098056014	0.060174863
NM_004346	CASP3	0.098873885	0.0605
NM_002360	MAFK	0.100285797	0.061659574
NM_013409	FST	0.100786907	0.061978836
NM_033300	LRP8	0.102030543	0.062492147
NM_003376	VEGFA	0.102146358	0.06307772
NM_022066	UBE2O	0.103051603	0.063897959
NM_017522	LRP8	0.103179158	0.064304569
NM_004083	DDIT3	0.104217214	0.06504
NM_004631	LRP8	0.104316037	0.065792079
NM_001001925	MTUS1	0.105039098	0.066868293
NM_199170	PMEPA1	0.105246837	0.06763285
NM_032711	MAFG	0.10539111	0.068210526
NM_001018054	LRP8	0.105441559	0.068580952
NM_199169	PMEPA1	0.105647608	0.069549296
NM_001001924	MTUS1	0.106466031	0.070608295
NM_033668	ITGB1	0.107856482	0.071909502
NM_001025368	VEGFA	0.108110645	0.072198198
NM_001025367	VEGFA	0.109049116	0.07275
NM_005067	SIAH2	0.109127465	0.072915556
NM_199171	PMEPA1	0.109425982	0.073274336
NM_001025366	VEGFA	0.109980443	0.073929825
NM_006287	TFPI	0.112000579	0.07525
NM_018433	KDM3A	0.112719	0.075476395
NM_001455	FOXO3	0.113000887	0.075794872
NM_001146688	KDM3A	0.113091614	0.076153191
NM_025090	USP36	0.113105469	0.076559322
NM_012224	NEK1	0.113246019	0.077268908
NM_002359	MAFG	0.113434126	0.077548117
NM_001033756	VEGFA	0.114262021	0.078248963
NM_201559	FOXO3	0.115000173	0.079853659
NM_004850	ROCK2	0.11677596	0.08116
NM_177951	PPM1A	0.117567808	0.081698413
NM_015640	SERBP1	0.117780863	0.082086957
NM_001018069	SERBP1	0.118092488	0.082433071
NM_001018068	SERBP1	0.118404096	0.083276265
NM_001018067	SERBP1	0.118715685	0.08355814
NM_015497	TMEM87A	0.118930277	0.084030769
NM_001025369	VEGFA	0.119400069	0.084557252
NM_001973	ELK4	0.120484396	0.085222642
NM_022828	YTHDC2	0.121842701	0.087516484
NM_016578	RSF1	0.121898417	0.087744526
NM_206909	PSD3	0.122170784	0.08792
NM_006466	POLR3F	0.123368602	0.088527076
NM_012334	MYO1O	0.123689567	0.088834532
NM_014945	ABLIM3	0.123956467	0.089039427
NM_015046	SETX	0.127055781	0.091531469
NM_174907	PPP4R2	0.127746035	0.092090278
NM_006350	FST	0.128346778	0.092914089
NM_005135	SLC12A6	0.128533103	0.093130137
NM_005649	ZNF354A	0.128561915	0.093372014
NM_024949	WWC2	0.129706945	0.09427027
NM_031899	GORASP1	0.130596765	0.095006711
NM_138927	SON	0.132138364	0.097980456
NM_001143886	PPP1R12A	0.133036549	0.099647436

In exemplary embodiments, subset of the genes listed in Table 1 can be selected to constitute a more simple gene signature. For example, a subset of genes, e.g., 10-20, 20-30 or more genes from Table 1 (or, for example, 5%, 10%, 15% 20% or more of the genes in Table 1) can be selected having a high degree of expression or representation in the gene signature. Alternatively, a subset of genes, e.g., 10-20, 20-30 or more genes from Table 1 (or, for example, 5%, 10%, 15% 20% or more of the genes in Table 1) can be selected having a low degree of expression or representation in the gene signature.
Differentially expressed genes (DEGs) can be selected based on low false discovery rate (FDR) (e.g., FDR for p-values from t-test.) For example, genes with a RP value of <0.1, <0.09, <0.08, <0.07, <0.06, <0.05, <0.04 or <0.02 can be selected as DEGs. Alternatively, genes with a FDR<0.05, <0.04, <0.3, or <0.2 can be selected as DEGs. Alternatively, or in combination, DEGs can be selected based on rank product (RP) value A lower absolute value for RP indicates a higher degree of differential expression. The genes in Table 1 were ranked in descending order of the absolute RP value.
RP ranking can characterize up-regulated genes and down-regulated genes under one class. To obtain one RP value per gene for comparison within results (or for comparison with ranking according to other methods), a lower value can be defined as a net value for a gene. A small net value for RP is therefore evidence of differential expression. (See e.g., Kadota K et al. (2009). Algorithm Mol. Biol. 4:7.)
To investigate the correlation of the XBP1 gene signature with patient relapse-free survival, we performed survival analysis using an aggregate breast cancer dataset that contains the gene expression profile and the survival information for 109 TNBC patient samples from 21 datasets (Lehmann, B. D., et al. 2011. J Clin Invest 121, 2750-2767). Of the plurality of genes in the XBP1 signature, a subset of genes were represented on the TNBC microarray datasets (FIG. 7A, Table 1).
As shown in FIG. 7B, the activation of the XBP1 pathway, as represented by the higher expression of the XBP1 signature, correlates with shorter relapse-free survival (Log-rank test, p=0.00768). These findings were confirmed in an independent validation cohort of 193 TNBC patients (FIG. 7C, Log-rank test, p=6.3×10⁻⁶).
We have identified both the UPR and the hypoxia response as XBP1 dependent pathways in TNBC. Interestingly, growing evidence indicates that increased expression of HIF1α and HIF1α targets, such as CA9 and GLUT1, are associated with worse clinical outcome in basal-like human breast tumors (Bos, R., et al. 2003. Cancer 97, 1573-1581; Hussein, Y. R., et al. 2011. Transl Oncol 4, 321-327; Semenza, G. L., 2010. Oncogene 29, 625-634; Tan, E. Y., et al. 2009. Br J Cancer 100, 405-411), consistent with the association of XBP1 with TNBC. To understand the clinical relevance of these two XBP1-regulated pathways in TNBC. we examined mRNA expression levels of multiple UPR markers in TICs and NTICs derived from five human TNBC patients. This analysis (survival analysis) revealed up-regulation of these marker genes in TICs relative to NTICs, indicative of an association of the UPR pathway with TICs and TNBC. Intriguingly, we also found that an elevated expression of the UPR gene signature in TNBC was associated with decreased relapse free survival (Log-rank test, p=0.00911) (FIG. 7D).
Collectively these data demonstrate that activation of XBP1 in TNBC patients is associated with poor clinical outcome.

Discussion

Patients with TNBC have a relatively poorer prognosis and are more likely to recur and develop metastatic disease than other breast cancer subtypes (Foulkes, W. D., et al. 2010. N Engl J Med 363, 1938-48; Lehmann, B. D., et al. 2011. J Clin Invest 121, 2750-67). The genes linked to TNBC are not well understood and thus, unlike other breast cancer subtypes, effective targeted therapies have not yet been identified for TNBC (Foulkes, W. D., et al. 2010. N Engl J Med 363, 1938-48). Here, by manipulating the expression of XBP1, the key component of the most evolutionarily conserved branch of the UPR, in a panel of breast cancer cell lines and in the patient-derived xenograft model, a key function for XBP1 in TNBC was discovered. XBP1 was activated in TNBC cells, and silencing of XBP1 was very effective in suppressing the tumorigenicity and progression of TNBCs. In addition to its essential role in TNBC, it is expected that XBP1 may also affect other subtypes of human breast cancer. TNBC typically contains a higher proportion of tumor-initiating cells (TICs) (Blick, T., et al. 2010. J Mammary Gland Biol Neoplasia 15, 235-52; Ricardo, S., et al. 2011. J Clin Pathol 64, 937-46). Relative to NTICs, TICs are resistant to chemotherapy, and contribute to a significantly higher incidence of recurrence and distant metastasis (Smalley, M., et al. 2003. Nat Rev Cancer 3, 832-44; Stingl, J., et al. 2007. Nat Rev Cancer 7, 791-9). Progress in targeting this subpopulation with novel therapeutics continues to be hampered by our incomplete knowledge of the molecular pathways contributing to TIC identity. It is thus demonstrated herein that XBP1 is a novel regulator for breast TICs.
These studies are the first to demonstrate that compromising the ER stress response significantly impairs the TIC population. It is speculated that TICs residing in the stem cell niche require robust UPR activation to cope with external stress. Hence TICs rely on XBP1 activation and their function is compromised in its absence. The increased activation of XBP1 in TICs is intriguing and provides potentially novel strategies to target this subpopulation of cancer cells. Hypoxia is known to promote aggressive tumor phenotypes and HIF1α was recently demonstrated to be essential for TNBC and breast TICs (Schwab, L. P., et al. 2012. Breast Cancer Res 14, R6; Conley, S. J., et al. 2012. Proc Natl Acad Sci USA 109, 2784-9; Montagner, M., et al. 2012. Nature 487, 380-4). Increased HIF1αá levels are also associated with increased metastasis and decreased survival in patients with TNBC (Semenza, G. L., 2010. Oncogene 29, 625-34; Bos, R., et al. 2003. Cancer 97, 1573-81). The data presented herein reveal that XBP1 acts in TNBC through regulating the HIF1α transcriptional program. HIF1α requires XBP1 to sustain downstream target expression. Hypoxia is a physiological inducer of the UPR in cancer (Wouters, B. G., et al. 2008. Nat Rev Cancer 8, 851-64). In the studies, it was found that XBP1 functions in a positive feedback loop to sustain the hypoxia response via regulating HIF1α transcriptional activity. This feed-forward circuit ensures maximum HIF activity and an efficient adaptive response to the cytotoxic microenvironment of solid tumors. HIF activity is tightly controlled during tumor progression, through translational and post-translational regulation of HIF1α, but relatively less is known about how HIF1α transcriptional activity is controlled (Kaelin, W. G., Jr., et al. 2008. Mol Cell 30, 393-402). These studies reveal a novel function for XBP1 as a HIF1α transcriptional cofactor. Herein is proposed a model in which these two critical pathways, the UPR and the hypoxia response, are physically interconnected and act together to mount an appropriate adaptive response to perpetuate cancer cells in the hostile tumor microenvironment. These data highlight the importance of XBP1 in TNBC progression and recurrence. Activation of the XBP1 pathway is correlated with poor patient survival in human TNBC patients, hence inhibition of this pathway may offer novel treatment strategies for this aggressive subtype of breast cancer. The use of UPR inhibitors in combination with standard chemotherapy may greatly enhance the effectiveness of anti-tumor therapies.

Experimental Procedures

Detailed protocols for all experimental procedures are provided below.

Cell Culture and Treatments

The non-transformed breast cell line MCF10A cells contains ER-Src, an integrated fusion of the v-Src oncoprotein, and the ligand-binding domain of estrogen receptor (ER) (Iliopoulos, D., et al., 2009. Cell 139, 693-706). These cells were grown in DMEM/F12 medium supplemented with 5% donor horse serum (Invitrogen), 20 ng/ml epidermal growth factor (EGF) (R&D systems), 10 ug/ml insulin (Sigma), 100 ug/ml hydrocortisone (Sigma), 100 ng/ml cholera toxin (Sigma), 50 units/ml pen/step (Gibco), with the addition of puromycin (Sigma). Src induction and cellular transformation was achieved by treatment with 1 uM 4-OH tamoxifen (TAM), typically for 36 h as described previously (Iliopoulos, D., et al. 2009. Cell 139, 693-706; Iliopoulos, D., et al. 2010. Mol Cell 39, 761-72).
All breast cancer cells were cultured according to Neve, R. M., et al. 2006. Cancer Cell 10, 515-27. Following retroviral or lentiviral infection, cells were maintained in the presence of puromycin (2 ug/ml) (Sigma). For all hypoxia experiments, cells were maintained in an anaerobic chamber (Coy laboratory) with 0.1% O₂. For glucose deprivation experiments, cells were maintained in DMEM without glucose medium (Gibco) with 10% FBS (Gibco) and 50 units/ml of penicillin/streptomycin.

Orthotopic Tumor Growth Assays

Six week old female NOD/SCID/IL2Rγ−/− mice (Taconic) were used for xenograft studies. Approximately 1.5×10⁶viable tumor cells were resuspended in 40 ul growth factor reduced Matrigel (BD Biosciences) and injected orthotopically into mammary gland four as previously described (Zhang, Q., et al. 2009. Cancer Cell 16, 413-424). Mice were supplied with chow containing 6 g doxycycline/kg (Bioserv) for treatment. For bioluminescent detection and quantification of cancer cells, mice were given a single i.p. injection of a mixture of luciferin (50 mg/kg), ketamine (150 mg/kg), and xylazine (12 mg/kg) in sterile water. Five minutes later, mice were placed in a light tight chamber equipped with a charge coupled device IVIS imaging camera (Xenogen). Photons were collected for a period of 1-60 s, and images were obtained by using LIVING IMAGE 2.60.1 software (Xenogen) and quantified using IGOR Pro 4.09 A image analysis software (WaveMatrics). The imaging intensity was normalized to the luminescence signal of each individual mouse taken before the Doxycycline chow treatment. The average luminescence ratio of treatment group (LacZ or XBP1 shRNA) was plotted over the course of doxycycline chow treatment. Results are presented as mean±standard error of the mean (SEM).

Sorting of TICs and NTICs (General)

To separate TICs from NTICs, flow cytometric cell sorting was performed on single-cell suspensions that were stained with CD44 antibody (FITC-conjugated) and with CD24 antibody (PE-conjugated) (BD Biosciences) for 30 min. As used throughout, TICs are defined by the minority CD44^high/CD24^lowpopulation, whereas NTICs are defined by the majority CD44^low/CD24^high.
Purification of TICs and NTICs from Patients with TNBC (Detailed)
Five human invasive triple negative ductal carcinoma tissues (stage III) were used in our TIC experiments (Iliopoulos, D., et al. 2011. Proc Natl Acad Sci USA 108, 1397-402). Immunomagnetic purification of TICs and NTICs was performed according to Shipitsin, M., et al. 2007. Cancer Cell 11, 259-73. Briefly, the breast tissues were minced into small pieces (1 mm) using a sterile razor blade. The tissues were digested with 2 mg/ml collagenase I (C0130, Sigma) and 2 mg/ml hyalurimidase (H3506, Sigma) in 370C for 3 h. Cells were filtered, washed with PBS and followed by Percoll gradient centrifugation. The first purification step was to remove the immune cells by immunomagnetic purification using an equal mix of CD45 (leukocytes), CD15 (granulocytes), CD14 (monocytes) and CD19 (B cells) Dynabeads (Invitrogen). The second purification step was to isolate fibroblasts from the cell population by using CD10 beads for magnetic purification. The third step was to isolate the endothelial cells by using an “endothelial cocktail” of beads (CD31 BD Pharmingen cat no. 555444, CD146 P1H12 MCAM BD Pharmingen cat no. 550314, CD105 Abcam cat no. Ab2529, Cadherin 5 Immunotech cat no. 1597, and CD34 BD Pharmingen cat no. 555820). In the final step the CD44high cells were purified from the remaining cell population using CD44 beads.
These cells were sorted for CD44high/CD24low (TIC) cells, CD24high cells were also purified using CD24 beads. These cells were sorted for CD44low/CD24high (NTICs) cells. These TIC and NTIC populations were sorted again with CD44 antibody (FITC-conjugated) (555478, BD Biosciences) and CD24 antibody (PE-conjugated) (555428, BD Biosciences) in order to increase their purity (>99.2% in all cases).

Mammosphere Formation Assay

Mammospheres were generated by placing cell lines in suspension (1,000 cells/ml) in serum-free DMEM/F12 media, supplemented with B27 (1:50, Invitrogen), 0.4% BSA, 20 ng/mL EGF, and 4 μg/ml insulin. After 6 days of incubation, mammospheres were typically >75 mM in size with 97% bearing the CD44^high/CD24^lowphenotype. For serial passaging, 6-day old mammospheres were harvested using a 70 um cell strainer, whereupon they were dissociated to single cells with trypsin and then re-grown in suspension for 6 days.

ChIP and ChIP-seq

ChIP assays were carried out as described previously (Chen, X., et al. 2008. Cell 133, 1106-1117). Briefly, cells were crosslinked with 1% formaldehyde for 10 min at room temperature, and formaldehyde was then inactivated by the addition of 125 mM glycine. Chromatin extracts containing DNA fragments with an average size of 500 bp were immunoprecipitated by using the antibodies described below. All ChIP experiments were repeated at least three times.
ChIP was performed with XBP1 antibody (Biolegend, 619502); HIF1α antibody (Abcam, ab2185), RNA Polymerase II antibody (Millipore, 05-623) or GST antibody (Santa Cruz, sc-33613). The primers used in FIG. 6 are listed in Table 2.

SUPPLEMENTARY TABLE 2

ChIP primer sequence

Gene	Forward

JMJD1A
1	TGTTCCTTCAGGTTCAATAGAATTTTTCCC (SEQ ID NO:)
JMJD1A 2	CATCATTCATTATGGCCTTCAACTACTTTA (SEQ ID NO:)
JMJD1A 3	CTTTCCTGTGAGATTCTTCCGCCA (SEQ ID NO:)
JMJD1A 4	GGGTCCGGGAGGCTGTGCGTGTCTTGTGAG (SEQ ID NO:)
JMJD1A 5	TCCCACACCGACGTTACCAAGAAGGATCTG (SEQ ID NO:)
JMJD2C 1	AACTTCAAGGGGAATCTATGTATTGTTCAT (SEQ ID NO:)
JMJD2C 2	TCCCGTTAGCCTTAGCTCAATTAATCACAT (SEQ ID NO:)
JMJD2C 3	TCCTTCTACGCGAGTATCTTTCCC (SEQ ID NO:)
JMJD2C 4	GATTATCGCTTGCTTTCTTACCTTGCTGGC (SEQ ID NO:)
VEGFA	TCTTCGAGAGTGAGGACGTGTGT (SEQ ID NO:)
PDK1	CGCCCTGTCCTTGAGCC (SEQ ID NO:)
DDIT4	CTAGAGCTCGCGGTCTGGTCTGGTCT (SEQ ID NO:)
NDRG1	AACACGTGAGCTAAGCTGTCCGA (SEQ ID NO:)
BETA-ACTIN	GGGACTATTTGGGGGTGTCT (SEQ ID NO:)
Control	TGAGGGTTCATCAAGCTGGTGTCT (SEQ ID NO:)

		Reference

JMJD1A
1	Reverse
JMJD1A
2	TGGCCTATCCTAAGGTGACGCTATGA (SEQ ID NO:)
JMJD1A 3	GAAGAAAGGCGTGGAGTTACTGGATA (SEQ ID NO:)	Xia et al., 2009
JMJD1A 4	CCGCGAAATCGGTTATCAACTTTGGG (SEQ ID NO:)
JMJD1A 5	CGGCGCTTTCACCTTTCTCTCCCCTCT (SEQ ID NO:)
JMJD2C 1	ACTCGGCTCTATACAACCATTCCAAA (SEQ ID NO:)
JMJD2C 2	CTACTAGAAAATCAACTGGACTCATGGCAC (SEQ ID NO:)
JMJD2C 3	CTGGGTCCCTTGTGGCGTTTTCTCTA (SEQ ID NO:)	Xia et al., 2009
JMJD2C 4	GTCACGTGGGCTTACAAACAGCTT (SEQ ID NO:)
VEGFA	ACTGTATTACCAAGTTTGCGGGATACTGTA (SEQ ID NO:)	Lee et al., 2009
PDK1	AAGGCGGAGAGCCGGAC (SEQ ID NO:)	Lee et al., 2009
DDIT4	CGGTATGGAGCGTCCCCT (SEQ ID NO:)
NDRG1	GGCGAAGAGGAGGTGGACGACGACGAG (SEQ ID NO:)AAG	Xia et al., 2009
BETA-ACTIN	ATGGAGGCAGAAGGAACATGTGAG (SEQ ID NO:)	Gromak et al., 2006
Control	TCCCATAGGTGAAGGCAAAG (SEQ ID NO:)	Xia et al., 2009

The ChIP-seq library was prepared using ChIP-Seq DNA Sample Prep Kit (Illumina) according to the manufacturer's instructions. XBP1 ChIP-seq peaks were identified using MACS package (Zhang, Y., et al. 2008. Genome Biol 9, R137) with a p-value cutoff of 1×10⁻⁷.

Tumor Initiation Assay Using Patient-Derived Tumors

Tumorgraft line BCM-2147 was derived by transplantation of a fresh patient breast tumor biopsy (ER-PR-HER2−) into the cleared mammary gland fat pad of immune-compromised SCID/Beige mice and retained the patient biomarker status and morphology across multiple transplant generations in mice. To overcome the challenge of limited cell viability by dissociation of solid tumors, 10 mg tumor pieces containing 1.3×10⁵cells were transplanted with basal membrane extract (Trevigen, Gaithersburg, Md.). The cell number was calculated as average cell yield 1.3×10⁷cells/gram×0.01 gram=1.3×10⁵cells. For sustained siRNA release in the first two weeks following transplantation, porous silicon particles loaded with siRNA (scrambled control or XBP1 siRNA) packaged in nanoliposomes were injected into the tumor tissue with basal membrane extract at the time of transplantation. Scrambled sequence [5′ CGAAGUGUGUGUGUGUGGCdTdT 3′]; XBP1 siRNA sequence [5′ CACCCUGAAUUCAUUGUCUdTdT 3′]. Two weeks post-transplantation, nanoliposomes containing siRNA (15 mg per mouse) were injected I.V. twice weekly for 8 weeks. Mice were monitored thrice weekly for tumor development, and tumors were calipered and recorded using LABCAT Tumor Analysis and Tracking System v6.4 (Innovative Programming Associates, Inc., Princeton, N.J.). Tumor incidence is reported at 10 weeks post-transplantation.

Invasion Assay

We performed invasion assays according to 49. Invasion of the matrigel was conducted by using standardized conditions with BD BioCoat growth factor reduced MATRIGEL invasion chambers (PharMingen). Assays were conducted according to manufacturer's protocol, by using 5% horse serum (GIBCO) and 20 ng/ml EGF (R&D Systems) as chemoattractants.

Colony Formation Assay

1×105 breast cancer cells were mixed 4:1 (v/v) with 2.0% agarose in growth medium for a final concentration of 0.4% agarose. The cell mixture was plated on top of a solidified layer of 0.8% agarose in growth medium. Cells were fed every 6 to 7 days with growth medium containing 0.4% agarose. The number of colonies was counted after 20 days. The experiment was repeated three times and the statistical significance was calculated using Student's t test.

Subcutaneous Xenograft Experiments

MCF10A ER-Src TAM-treated (36 h) cells or MDA-MB-436 or HBL-100 breast cancer cells were injected subcutaneously in the right flank of athymic nude mice (Charles River Laboratories). Tumor growth was monitored every five days and tumor volumes were calculated by the equation V (mm³)=a×b²/2, where a is the largest diameter and b is the perpendicular diameter. When the tumors reached a size of ˜100 mm³(15 days) mice were randomly distributed into 3 groups (5 mice/group). The first group was used as control (non-treated), the second group was intratumorally treated with shCtrl and the third group was intratumorally treated with shXBP1. For each injection 10 ug of shRNA was mixed with 2 ul of vivo-jetPEI (polyethylenimine) reagent (cat. no 201-50G, PolyPlus Transfection SA) in a final volume of 100 ul. These treatments were repeated every five days for 4 cycles ( days 15, 20, 25, 30). In addition, in vivo dilution xenotransplantation assays were performed in NOD/SCID/IL2Rγ−/− mice. Mice were evaluated on a weekly basis for tumor formation. All mice were maintained in accordance with Dana-Farber Cancer Institute Animal Care and Use Committee procedures and guidelines.

Gene Expression Microarray Analysis

MDA-MB-231 cells infected with control shRNA or XBP1 shRNA lentiviruses grown in glucose free medium were treated in 0.1% O₂in a hypoxia chamber for 24 h. Total RNA was extracted by using RNeasy mini kit with on column DNase digestion (QIAGEN). Biotin labeled cRNA was prepared from 1 ug of total RNA, fragmented, and hybridized to Affymetrix human U133 plus 2.0 expression array. All Gene expression microarray data were normalized and summarized using RMA (Irizarry, R. A., et al. 2003. Nucleic Acids Res 31, e15). The differentially expressed genes were identified using Limma (Smyth, G. K., et al. 2003. Methods Mol Biol 224, 111-36) (q≦10%, fold change ≧1.5).

Motif Analysis

Flanking sequences around the summits (±300 bp) of the top 1,000 XBP1 binding sites were extracted and the repetitive regions in these flanking sequences were masked. The consensus sequence motifs were derived using Seqpos (Lupien, M., et al. 2008. Cell 132, 958-70).

XBP1 Signature Generation

The XBP1 signature was generated by integrative analysis of ChIP-seq and differential expression data using the method as previously described (Tang, Q., et al. 2011. Cancer Res 71, 6940-7). Briefly, we first calculated the regulatory potential for a given gene, Sg, as the sum of the nearby binding sites weighted by the distance from each site to the TSS of the gene:
S _g=Σ_i=1 ^k e ^−(0.5+4Δ ⁱ ⁾
where k is the number of binding sites within 100 kb of gene g and Δi is the distance between site i and the TSS of gene g normalized to 100 kb (e.g., 0.5 for a 50 kb distance). We then applied the Breitling's rank product method (Breitling, R., et al. 2004. FEBS Lett 573, 83-92; Klisch, T. J., et al. 2011. Proc Natl Acad Sci USA 108, 3288-93) to combine regulatory potentials with differential expression t-values to rank all genes based on the probability that they were XBP1 targets. Only genes with at least one binding site within 100 kb from its TSS and a differential expression t-value above the 75th percentile were considered (Tang, Q., et al. 2011. Cancer Res 71, 6940-7). The FDR of XBP1 target prediction was estimated by permutation (Breitling, R., et al. 2004. FEBS Lett 573, 83-92). At a FDR cutoff of 10% and differential expression fold-change cutoff of 1.5, we obtained 119 up-regulated genes (HUGO gene symbol) as direct targets of XBP1.

Survival Analysis (General)

Principle component analysis (PCA) was applied to patient expression profiles of genes of interest and separated the samples into 2 groups based on the median value of the first component. Kaplan-Meier survival analysis was used to assess the significance of survival difference. In cases where XBP1 signature genes were the relevant gene set, a correlation value was calculated between the relevant gene expression indexes of each patient and those of the MDA-MB-231 cell line, and the correlations of the 2 groups were compared and the significance of difference was assessed by t-test.

Survival Analysis (Detailed)

We performed survival analysis using an aggregated compendium of gene expression profiles of 383 TNBC samples from 21 breast cancer datasets (Rody, A., et al. 2011. Breast Cancer Res 13, R97). Of the 119 XBP1 signature genes, 91 genes had corresponding probes in this dataset. To avoid potential confounding factors such as heterogeneity among the samples, we randomly split all 383 TNBC samples into two datasets with similar size (190 and 193 cases) and evaluated the correlation of the XBP1 gene signature with relapse free survival using these two datasets respectively. We separated patients into two subgroups: one with higher and the other with lower expression of XBP1 signature. The subgroup classification was performed as described previously (Marotta, L. L., et al. 2011. J Clin Invest 121, 2723-35). Patients were considered to have higher XBP1 signature if they had average expression values of all the genes in the XBP1 signature above the 60th percentile (Marotta, L. L., et al. 2011. J Clin Invest 121, 2723-35). Kaplan-Meier survival analysis was performed and log-rank test was used to assess the statistical significance of survival difference between these 2 groups. A similar analysis was performed for the HIF pathway signature (VEGFA, PDK1, DDIT4, SLC2A1, KDM3A, NDRG1, PFKFB3, PIK3CA, RORB, CREBBP, PIK3CB and EGLN1).
Virus Production and Infection
The Phoenix packaging cell line was used for the generation of ecotropic retroviruses and all retroviral infections were carried out as described previously (Martinon, F., et al. 2010. Nat Immunol 11, 411-8). The 293T packaging cell line was used for lentiviral amplification and all lentiviral infections were carried out as previously described (Martinon, F., et al. 2010. Nat Immunol 11, 411-8). In brief, viruses were collected 48 and 72 hr after transfection, filtered, and used for infecting cells in the presence of 8 mg/ml polybrene prior to drug selection with puromycin (2 μg/ml). shRNA constructs were generated by The Broad Institute. Targeting of GFP mRNA with shRNA served as a control. Optimal targeting sequences identified for human XBP1 were 5′-GACCCAGTCATGTTCTTCAAA-3′, and 5′-GAACAGCAAGTGGTAGATTTA-3′, respectively. Knockdown efficiency was assessed by real-time PCR for XBP1.
Luciferase Assay
For FIG. 5H, MDA-MB-231 cells were co-transfected with 3×HRE luciferase (3×HRE-Luc) plasmid (Yan, Q., et al. 2007. Mol Cell Biol 27, 2092-102) and XBP1s overexpression construct (Kaser, A., et al. 2008. Cell 134, 743-56) or control vector by using Lipofectamine 2000 (Invitrogen). A Renilla luciferase plasmid (pRL-CMV from Promega) was co-transfected as an internal control. Cells were harvested 36 hr after transfection, and the luciferase activities of the cell lysates were measured by using the Dual-luciferase Reporter Assay System (Promega). For FIG. 5I, MDA-MB-231 cells were co-transfected with 3×HRE-Luc and two inducible XBP1 shRNA construct (in pLKO-Tet-On vector) or control shRNA construct by using Lipofectamine 2000 (Invitrogen). Cells were treated with doxycycline for 48 h and hypoxia for 24 h before the luciferase activities of the cell lysates were measured.
Statistical Analysis
The significance of differences between treatment groups were identified with a Student's t-test. P values of less than 0.05 were considered statistically significant.
Coimmunoprecipitation
Transfected cells were lysed in cell lysis buffer (50 mM Tris HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, and 10% glycerol with protease inhibitor cocktail) for 1 hour. M2 beads (Sigma) were incubated with the whole cell extracts at 4° C. for overnight. The beads were washed with cell lysis buffer four times. Finally, the beads were boiled in 2× sample buffer for 10 minutes. The eluents were analyzed by Western blot. Nuclear extracts were used to perform the endogenous co-IP as described previously (Xu, J., et al. 2010. Genes Dev 24, 783-98). Briefly, 5 mg of nuclear extracts were incubated with 5 ug of anti-HIF1α antibody (Novus Biologicals, NB 100-479) at 4° C. for overnight. The protein complexes were precipitated by addition of protein A agarose beads (Roche) with incubation for 4 hr at 4° C. The beads were washed four times and boiled for 5 min in 2× sample buffer.
Real-Time PCR Analysis
1 ug of RNA sample was reverse-transcribed to form cDNA, which was subjected to SYBR Green based real-time PCR analysis. Primers used for β-actin forward: 5′-CCTGTACGCCAACACAGTGC-3′ and reverse 5′-ATACTCCTGCTT GCTGATCC-3′; for VEGFA forward 5′-CACACAGGATGGCTTGAAGA-3′ and reverse 5′-AGGGCAGAATCATCACGAAG-3′; for PDK1 forward 5′-GGAGGTCTCAACACGAGGTC-3′ and reverse 5′-GTTCATGTCACGCTGGGTAA-3′; for GLUT1 forward 5′-TGGACCCATGTCTGGTTGTA-3′ and reverse 5′-ATGGAGCCCAGCAGCAA-3′; for JMJD1A forward 5′-TCAGGTGACTTTCGTTCAGC-3′ and reverse 5′-CACCGACGTTACCAAGAAGG-3′; for DDIT4 forward 5′-CATCAGGTTGGCACACAAGT-3′ and reverse 5′-CCTGGAGAGCTCGGACTG-3′; for MCT4 forward 5′-TACATGTAGACGTGGGTCGC-3′ and reverse 5′ CTGCAGTTCGAGGTGCTCAT-3′; for XBP1 splicing forward 5′-CCTGGTTGCTGAAGAGGAGG-3′ and reverse 5′-CCATGGGGAGATGTTCTGGAG-3′; for XBP1 total forward 5′-AGGAGTTAAGACAGCGCTTGGGGATGGAT-3′ and reverse 5′-CTGAATCTGAAGAGTCAATACCGCCAGAAT-3′.
Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, suitable methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.
This invention is further illustrated by the following examples which should not be construed as limiting.

EQUIVALENTS

Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed by the following claims.

Claims

1. A method of treating triple negative breast cancer (TNBC) in a subject, the method comprising administering to the subject a direct or indirect inhibitor of XBP1 in an amount effective to inhibit growth of cancer cells in said subject, such that TNBC in the subject is treated.

2. The method of claim 1, wherein the subject has an advanced stage of breast cancer.

3. The method of claim 1, wherein the inhibitor of XBP1 is a direct inhibitor.

4. The method of claim 3, wherein the inhibitor of XBP1 is selected from the group consisting of a nucleic acid molecule that is antisense to an XBP1-encoding nucleic acid molecule, an XBP1 shRNA, an XBP siRNA, a microRNA that targets XBP1, a dominant negative XBP1 molecule and a small molecule inhibitor of XBP1.

5. The method of claim 4, wherein the inhibitor of XBP1 is an XBP1 shRNA or an XBP siRNA.

6. The method of claim 1, wherein the inhibitor of XBP1 is an indirect inhibitor.

7. The method of claim 6, wherein the inhibitor of XBP1 is an agent that inhibits IRE1 or an agent that inhibits an endonuclease that produces XBP1s.

8. The method of claim 1, wherein the inhibitor of XBP1 is administered to breast tissue of said subject.

9. The method of claim 8, wherein the inhibitor of XBP1 is administered directly to a tumor in said tissue.

10. The method of claim 1, wherein the inhibitor of XBP1 is administered in combination with a second cancer therapeutic agent.

11. The method of claim 10, wherein the inhibitor of XBP1 is administered in combination with a chemotherapeutic agent.

12. The method of claim 1, wherein the treatment promotes longer relapse-free survival of the subject.

13. A method for determining a prognosis status for a subject with triple negative breast cancer (TNBC), the method comprising:

a) determining an XBP1 gene signature for the TNBC of the subject; and

b) correlating the XBP1 gene signature with a prognosis status for the subject, wherein higher expression of the XBP1 gene signature, relative to a control, correlates with shorter relapse-free survival of the subject and lower expression of the XBP1 gene signature, relative to a control, correlates with longer relapse-free survival of the subject.

14. The method of claim 13, wherein higher expression of the XBP1 gene signature, relative to a control, correlates with shorter relapse-free survival of the subject.

15. The method of claim 14, wherein the XBP1 gene signature comprises a plurality of genes set forth in Table 1.

16. The method of claim 15, wherein at least 5% of the genes set forth in Table 1 are expressed at a higher level relative to control.

17. The method of claim 15, wherein at least 20% of the genes set forth in Table 1 are expressed at a higher level relative to control.

18. The method of claim 13, wherein the gene signature is determined for a population of cells isolated from a breast tissue tumor breast of said subject.

19. The method of claim 13, wherein the gene signature is determined for a population of cancer cells isolated from said subject.

20. The method of claim 13, wherein the subject is negative for one or more of estrogen receptor (ER), progesterone receptor (PR) or Her2/neu.

21. The method of claim 13, wherein the subject is positive for BRCA1.

22. The method of claim 21, wherein the subject is further subjected to histopathological analysis of tumors.

23. A method of identifying a compound useful in inhibiting the growth of triple negative breast cancer (TNBC) cells, the method comprising:

a) providing an indicator composition comprising XBP1 and HIF1α, or biologically active portions thereof;

b) contacting the indicator composition with each member of a library of test compounds;

c) selecting from the library of test compounds a compound of interest that decreases the interaction of XBP1 and HIF1α, or biologically active portions thereof, wherein the ability of a compound to inhibit growth of TNBC cells is indicated by a decrease in the interaction as compared to the amount of interaction in the absence of the compound.

24. The method of claim 20, wherein the subject is positive for BRCA1.