Small molecule inhibitors of Hsp90 potently affect inflammatory disease pathways and exhibit activity in models of rheumatoid arthritis.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 58, No. 12, December 2008, pp 3765–3775 DOI 10.1002/art.24047 © 2008, American College of Rheumatology Small Molecule Inhibitors of Hsp90 Potently Affect Inflammatory Disease Pathways and Exhibit Activity in Models of Rheumatoid Arthritis John W. Rice, James M. Veal, R. Patrick Fadden, Amy F. Barabasz, Jeffrey M. Partridge, Thomas E. Barta, Laura G. Dubois, Kenneth H. Huang, Sarah R. Mabbett, Melanie A. Silinski, Paul M. Steed, and Steven E. Hall Objective. To evaluate the ability of SNX-7081, a novel small molecule inhibitor of Hsp90, to block components of inflammation, including cytokine production, protein kinase activity, and angiogenic signaling. A close analog was evaluated in preclinical in vivo models of rheumatoid arthritis (RA). Methods. SNX-7081 binding to Hsp90 was characterized in Jurkat cells and RA synovial fibroblasts (RASFs). Inhibition of NF-B nuclear translocation was evaluated in cellular systems, using lipopolysaccharide (LPS), tumor necrosis factor ␣, or interleukin-1␤ stimulation. Suppression of cytokine production in THP-1 cells, human umbilical vein endothelial cells, and RASFs was studied. Disruption of MAPK signaling cascades by SNX-7081 following growth factor stimulation was assessed. SNX-7081 was tested in 2 relevant angiogenesis assays: platelet-derived growth factor activation of fibroblasts and LPS-induced nitric oxide (NO) release in J774 macrophages. A close analog, SNX-4414, was evaluated in rat collagen-induced arthritis and adjuvant-induced arthritis, following oral treatment. Results. SNX-7081 showed strong binding affinity to Hsp90 and expected induction of Hsp70. NF-B nuclear translocation was blocked by SNX-7081 at nanomolar concentrations, and cytokine production was potently inhibited. Growth factor activation of ERK and JNK signaling was significantly reduced by SNX-7081. NO production was also sharply inhibited. In animal models, SNX-4414 fully inhibited paw swelling and improved body weight. Scores for inflammation, pannus formation, cartilage damage, and bone resorption returned to normal. Conclusion. The present results demonstrate that a small molecule Hsp90 inhibitor can impact inflammatory disease processes. The strong in vivo efficacy observed with SNX-4414 provides preclinical validation for consideration of Hsp90 inhibitors in the treatment of RA. Rheumatoid arthritis (RA) is a prevalent autoimmune inflammatory disease characterized by abnormal infiltration of a range of cell types, including macrophages, T helper cells, B cells, mast cells, and plasma cells, into the joint synovial tissue (1–5). Accompanying the cellular invasion is synovial hyperplasia, in which synovial fibroblasts increase in number and exhibit cancer-like properties, including anchorage-independent growth and proliferation. Central to disease development and progression is the role of cytokine production and release from the localized immune cells. This progression involves synovial fibroblast invasion of cartilage and bone destruction and is in part osteoclast mediated, utilizing matrix metalloproteases and other matrix-degrading enzymes coupled with an ongoing angiogenic component. Disease progression ultimately leads to disability, and in its most severe forms, RA is associated with increased mortality. Current therapy for RA includes a range of options (1). Methotrexate and other disease-modifying antirheumatic drugs exert their effects through general immunosuppressive properties. More recently, recogni- John W. Rice, James M. Veal, PhD, R. Patrick Fadden, PhD, Amy F. Barabasz, Jeffrey M. Partridge, Thomas E. Barta, PhD, Laura G. Dubois, Kenneth H. Huang, PhD, Sarah R. Mabbett, Melanie A. Silinski, PhD, Paul M. Steed, PhD, Steven E. Hall, PhD: Serenex Inc., Durham, North Carolina. All authors own stock or stock options in Serenex Inc. Address correspondence and reprint requests to James M. Veal, PhD, Serenex Inc., 323 Foster Street, Durham, NC 27701. E-mail: email@example.com. Submitted for publication April 21, 2008; accepted in revised form August 15, 2008. 3765 3766 tion of the importance of cytokine signaling in RA has led to the discovery of cytokine-targeting agents, and such agents are now widely prescribed (3). Drug discovery strategies targeting T cells and B cells have also proved successful (5). However, despite this range of options, there remains substantial unmet medical need in the treatment of RA. For example, in a significant percentage of patients, the disease fails to respond to anti–tumor necrosis factor (anti-TNF) therapies, and disease progression frequently occurs in patients whose RA initially responded to therapy. Hsp90 is a member of the chaperone family and is responsible for guiding late-stage tertiary folding for multiple protein clients (6,7). Beyond initial folding, Hsp90 function continues to be needed for maintaining the conformational integrity of many of its clients. The chaperoning process is complex, involves multiple cochaperone proteins, and requires an ATP binding and hydrolysis cycle on the part of Hsp90 for proper functioning. If ATPase activity is disrupted, e.g., by inhibitors of ATP binding, client proteins lose their proper folding, which in turn activates degradation-related cellular processes. In particular, client proteins are ubiquitinated via E3 ligases, leading to proteasome-mediated degradation. Inhibition of Hsp90 as a treatment strategy has been primarily evaluated in the context of cancer treatment (7,8). However, there are substantive data indicating that an Hsp90-targeted agent could be useful in the treatment of inflammatory diseases, including RA. Aberrant cytokine and receptor signaling, angiogenesis, and cellular invasion are common to both cancer and inflammation. Natural product inhibitors of Hsp90 block activation of the NF-B pathway, leading to loss of cytokine production in macrophages and other cell types (9–13). Receptor-interacting protein (RIP) and IKK, members of the NF-B signaling pathway, have been shown to be Hsp90 clients that are degraded following Hsp90 inhibition (14–17). Similarly, activation of MAP kinases (ERK-1/2, p38, JNK) can be blocked by Hsp90 inhibition (11,13,18). Interleukin-1 receptor–associated kinase is also a client of Hsp90, and consequently, inhibition of Hsp90 also has the ability to diminish innate immunity responses via Toll-like receptor signaling (18). CD14 cell surface expression and activity are regulated by Hsp90 (19,20). SGT1, a component of the inflammasome, has also recently been reported to be degraded following Hsp90 inhibition (21). The receptor tyrosine kinase KIT, which regulates mast cell actions, appears to be a sensitive client for Hsp90 with relevance to multiple diseases (22). In vivo efficacy of Hsp90 inhibitors in RICE ET AL preclinical models of arthritis, sepsis, and uveitis supports the relevance of Hsp90 inhibition observed in vitro (23,24). Hsp90 chaperone functions are also central to angiogenesis and invasion. When Hsp90 is inhibited, hypoxia-inducible factor 1␣ (HIF-1␣) is degraded under both normoxic and hypoxic conditions, and vascular endothelial growth factor (VEGF) signaling and production are reduced (25–29). VEGF receptor 1 (VEGFR-1) and VEGFR-2 are observed to be sensitive clients (30). Additionally, Hsp90 both regulates Akt activation (31,32) and forms a critical complex with endothelial nitric oxide synthase (eNOS) and activated Akt, and Hsp90 inhibition leads to loss of NO production (33–38). Hsp90 inhibition is effective in angiogenesis models such as the Matrigel assay (39). Finally, the transforming activities of platelet-derived growth factor receptor ␣ (PDGFR␣) are sensitive to Hsp90 inhibition (40). We have recently reported on a novel series of small molecule inhibitors of Hsp90 (41). The chemical scaffold binds competitively to the ATP-binding site of Hsp90 and is composed of a benzamide moiety coupled to a tetrahydroindolone. X-ray data demonstrate that the benzamide functionality acts as a structural isostere to the adenine ring of ATP. The compounds are of low molecular weight and demonstrate good pharmacokinetic properties, including bioavailability on oral dosing. In the present study we evaluated one of these inhibitors, SNX-7081, and confirmed that it is a potent and specific Hsp90 inhibitor that is able to affect multiple inflammation-related signaling processes in vitro. Additionally, we found that a close analog of SNX-7081 was highly active in preclinical in vivo models of RA. MATERIALS AND METHODS Reagents. The NIH3T3, THP-1, and J774 mouse macrophage cell lines were obtained from American Type Culture Collection (Manassas, VA) and grown in Dulbecco’s modified Eagle’s medium (DMEM)/Ham’s F-12 (Hyclone, Logan, UT) or RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 20 mM HEPES buffer. Primary cultures of human umbilical vein endothelial cells (HUVECs) and EGM-2 medium (insulin-like growth factor 1 [IGF-1], epidermal growth factor [EGF], fibroblast growth factor [FGF], VEGF, and 2% FBS) were from Lonza (Rockland, ME). Cultures of synovial fibroblasts from RA patients (RASFs) were from Asterand (Detroit, MI) and were also maintained in DMEM/Ham’s F-12 supplemented with 10% FBS. Primary antibodies used for high-content analysis of Hsp70 were from Assay Design (Ann Arbor MI), phospho–ERK-1/2, phospho– c-Jun, and phospho–tyrosine 100 were from Cell Signaling (Boston, MA), and NF-B was from Santa Cruz Biotechnology SMALL MOLECULE INHIBITORS OF Hsp90 IN INFLAMMATORY DISEASE (Santa Cruz, CA). Conjugated secondary antibodies were from Invitrogen (Carlsbad, CA). ATP displacement assay and mass spectrometry identification of Hsp90. Ground porcine spleen tissue and Jurkat cell lysates were prepared by flash freezing followed by homogenization in saline solution at 4°C. These lysates were then incubated with a Sepharose affinity resin column in order to reversibly bind a range of proteins containing purine-binding sites, as described previously (41). SNX-7081 at various concentrations (0.8–500 M) was then incubated with the resin for 5 minutes (spleen lysates) or 90 minutes (Jurkat cell lysates). Column elution, gel electrophoresis, trypsin digestion of isolated proteins, and mass spectrometry (MS)–based protein identification were all performed as previously described (41). High-content analysis assays. A Cellomics (Pittsburgh, PA) Arrayscan 4.5 was used to perform all high-content analysis assays and to obtain all images, except for the use of a BD 435 imager to collect the 20⫻ images of HUVEC NF-B translocation. Cells were seeded onto 96-well plates (Greiner, Frickenhausen, Germany) and allowed to adhere overnight. The following day, test compound was added. For PDGF and growth factor activation assays, cells were treated with medium containing 0.1% FBS. For NF-B and phospho–c-Jun assays in HUVECs, cells were pretreated for 24 hours before activation for 30 minutes with 1 ng/ml (final concentration) cytokine. J774 cells were plated onto collagen and pretreated for 6 hours before activation for 30 minutes with 100 ng/ml (final concentration) lipopolysaccharide (LPS). NIH3T3 cells were pretreated for 24 hours and then activated for 15 minutes with a mixture of recombinant human PDGF-BB and PDGF-DD (100 ng/ml final concentration; R&D Systems, Minneapolis, MN). HUVECs were pretreated for 24 hours in 0.1% FBS medium, followed by removal of the treatment medium, which was replaced for 15 minutes with EGM-2 medium. Following treatment and activation, cells were fixed and permeabilized (15 minutes each step) with 4% phosphate buffered formalin and phosphate buffered saline (PBS) plus 0.1% Triton X-100, respectively. This was followed by 1 hour of incubation with primary antibody (diluted 1:100 in PBS–1% bovine serum albumin) and finally, 1 hour of incubation with fluorescence-conjugated secondary antibody (diluted 1:1,000) and Hoechst DNA binding dye (diluted 1:50,000). Average staining intensity for each end point was determined using either the target activation or compartmental analysis algorithms (Cellomics). For each well of the plate, a minimum of 250 (compartmental analysis) or 500 (target activation) individual cells per well were identified by nuclear staining intensity. Corresponding antibody staining intensity for each individual cell was then determined. Translocation inhibition was measured as the ratio of staining intensity in activated cells to that in nonactivated cells. Cytokine release assays. Cells were pretreated for either 6 hours (THP-1) or 24 hours (HUVECs and RASFs), followed by activation for 18–24 hours with interleukin-1␤ (IL-1␤) or LPS. After activation, the supernatants were removed and analyzed using commercially available Luminex reagents (Bio-Rad, Los Angeles, CA). Pharmacokinetics. In vivo studies were conducted at Washington Biotech (Columbia, MD). SNX-4414 was administered by oral gavage at 5 mg/kg to each of 3 female Sprague-Dawley rats weighing 160–170 gm. Blood samples 3767 were collected into prechilled EDTA-treated tubes by retroorbital sinus bleeding at 5, 15, and 30 minutes and 1, 2, 4, 8, 12, 24, and 48 hours posttreatment and converted to plasma by centrifuging at 4°C. The samples were stored frozen until liquid chromatography–tandem MS (LC-MS/MS) analysis. The concentration of SNX-4414 active species in rat plasma was determined by LC-MS/MS following protein precipitation with acetonitrile. Samples were thawed on ice and were not allowed to reach room temperature. To 50 l of sample, 10 l of 160 mM ammonium acetate (pH 7.4) was added, followed by 180 l of acetonitrile that contained internal standard (1 g/ml). The LC-MS/MS system consisted of a Prominence high-performance liquid chromatograph (Shimadzu, Columbia, MD) coupled to a hybrid triple quadrupole/ linear ion trap mass spectrometer with a Turbospray source (Q-Trap; Applied Biosystems, Foster City, CA). Ten microliters of sample was injected onto an Eclipse XDB-C18 column (2.1 ⫻ 30 mm, 3.5 mm; Agilent, Santa Clara, CA), and elution of analytes and internal standard was achieved using a water– methanol gradient in the presence of 0.08–0.1% formic acid and 4 mM ammonium formate. The electrospray ionization mass spectrometer was operated in positive ion mode, using multiple reaction monitoring. Fragmentation was achieved by collision-activated dissociation, with collision energy of 31 eV. Quantitation was performed using analyte-to-standard peak area ratios, with calibration curves of 2–4,000 ng/ml. In vivo rat arthritis models. In vivo studies and followup histopathologic analysis were conducted at Bolder BioPath (Boulder, CO) (42). Animal care including room, cage, and equipment sanitation conformed to the guidelines cited in the Guide for the Care and Use of Laboratory Animals and the applicable standard operating procedures of the University of Colorado vivarium. Collagen-induced arthritis (CIA). Female Lewis rats (140–170 gm; 10 per group with arthritis, 4 per normal control group) were anesthetized with isoflurane and injected at the base of the tail and 2 sites on the back with 300 l Freund’s incomplete adjuvant (Difco, Detroit, MI) containing 2 mg/ml bovine type II collagen (Elastin Products, Owensville, MO), on days 0 and 6. Treatment by oral gavage of a 5 ml/kg solution of SNX-4414 (at 48- or 72-hour intervals) was initiated on day 9 of the study and continued through day 16, with animals killed on day 17. Rats were weighed on day 0 and days 9–17 of the study, and caliper measurements of the ankles were obtained daily beginning on day 9. Histopathologic features (inflammation, pannus formation, cartilage damage, and bone resorption) in ankle and knee joint specimens were scored on a 0–5 scale (0 ⫽ normal; 5 ⫽ severe disease). Adjuvant-induced arthritis (AIA). Male Lewis rats (165– 185 gm; 8 per group with arthritis, 8 per normal control group) were anesthetized with isoflurane and injected at the base of the tail with 100 l Freund’s complete adjuvant and lipoidal amine, on day 0. Dosing was the same as in the CIA study and was initiated on day 7 (at the time of semi-established disease). Animals were weighed on days 0, 4, 7, and daily from day 8 through day 14, at which time, dose volumes were adjusted. Starting on day 7 (prior to onset of swelling but after establishment of systemic disease), caliper measurements of the ankles were obtained daily through day 14. Final body weight was recorded on day 14. Animals were killed on day 14. Histopathologic analysis was as described for the CIA study. 3768 RICE ET AL Statistical analysis. The significance of differences in mean paw weights and histopathologic scores between treatment groups was analyzed by Student’s t-test or other tests as appropriate. P values less than or equal to 0.05 were considered significant. RESULTS SNX-7081 binding to Hsp90 and induction of Hsp70. SNX-7081 is a member of a novel class of small molecule binders to Hsp90 (Figure 1A). SNX-7081 binding to Hsp90 was evaluated using a previously described affinity chromatography method (41), that allows nonenzymatic evaluation of binding of ligands to proteins that contain ATP or other purine ligand– binding sites. SNX-7081 bound strongly to Hsp90 from both tissue and cell lysates (Figure 1B). A 10-point concentration evaluation in human Jurkat cells yielded an apparent Ki of 26 nM for binding of SNX-7081 to Hsp90 (data not shown). No apparent binding to other proteins captured on the resin was detected. Hsp70 induction was evaluated to assess the ability of SNX7081 to inhibit Hsp90 in RASF cells. Consistent with inhibition of Hsp90 as a mechanism of action, SNX-7081 potently induced Hsp70, with a 50% inhibition concentration (IC50) of 44 ⫾ 9 nM (mean ⫾ SD) (Figures 1C and D). Inhibition of NF-B nuclear translocation by SNX-7081. Inflammatory response due to bacteria (e.g., LPS), cytokine signaling (e.g., IL-1␤ or TNF␣), or other agents involves activation of NF-B transcription factor and its translocation to the nucleus, where it activates the transcription of multiple genes (43). Stimulation of HUVECs with 1 ng/ml IL-1␤ or TNF␣ for 30 minutes caused the expected migration of NF-B into the nucleus (Figure 2A). Pretreatment with SNX-7081 for 24 hours fully blocked IL-1␤–induced NF-B nuclear translocation, with a mean ⫾ SD IC50 of 114 ⫾ 29 nM. For comparison, the natural product Hsp90 inhibitor, 17AAG, was also evaluated. Both potency and maximal inhibition levels were similar to those observed with SNX-7081 (data not shown). NF-B translocation in J774 macrophage cells was also examined (Figure 2B), and SNX-7081 inhibited translocation, with an IC50 of 241 ⫾ 57 nM. Additionally, in RASF cells, both IL-1␤– initiated and TNF␣-initiated NF-B nuclear translocation were inhibited by SNX-7081 (Figure 2C), with IC50 values of 44 ⫾ 19 nM and 88 ⫾ 11 nM, respectively. We also noted that in RASF cells, the level of overall NF-B translocation was reduced only ⬃50–75% as compared Figure 1. SNX-7081 binding to Hsp90 and induction of Hsp70. A, Chemical structure of SNX-7081. B, Elution of Hsp90 from a porcine spleen lysate bound reversibly to a Sepharose capture resin. The resin was developed to allow reversible binding of a wide range of proteins that utilize ATP or other purines as ligands. Compound concentrations increase from left to right (20, 100, and 500 M). C, Representative images of vehicle control– and SNX-7081–treated rheumatoid arthritis synovial fibroblast (RASF) cells (magnification ⫻ 10). D, Induction of Hsp70 in RASFs by SNX-7081. Values are the mean ⫾ SD from triplicate wells. IC50 ⫽ 50% inhibition concentration. with levels in control cells that were not activated with cytokine (Figure 2D). Down-regulation of cytokine signaling. The ability of SNX-7081 to reduce cytokine levels was evaluated in several cell types following application of distinct stimuli. Production of cytokines, including TNF␣, IL-6, SMALL MOLECULE INHIBITORS OF Hsp90 IN INFLAMMATORY DISEASE 3769 Figure 2. Effects of SNX-7081 (SNX) on NF-B translocation. A, Images of human umbilical vein endothelial cells (HUVECs), showing NF-B nuclear localization after stimulation with 1 ng/ml interleukin-1␤ (IL-1␤) and reduced levels of translocation with 24-hour pretreatment with 10 M SNX-7081. B, Images of J774 macrophage cells, showing NF-B nuclear localization after stimulation with 100 ng/ml lipopolysaccharide (LPS) and reduced levels of translocation with 6-hour pretreatment with 10 M SNX-7081. C, Images of rheumatoid arthritis synovial fibroblast (RASF) cells, showing NF-B nuclear localization after stimulation with 1 ng/ml tumor necrosis factor ␣ (TNF␣) and reduced levels of translocation with 24-hour pretreatment with 10 M SNX-7081. (Magnification ⫻ 10.) D, Levels of NF-B translocation inhibition achieved with each cell line and agonist. The degree of translocation inhibition by SNX-7081 was measured by scaling to the ratio of translocation in activated versus nonactivated cells (mean ⫾ SD of quadruplicate determinations). IL-1␤, granulocyte–macrophage colony-stimulating factor, and RANTES, was inhibited by SNX-7081, with IC50 values of ⬍100 nM observed in several cases (Table 1). In LPS-stimulated THP-1 monocytes, potent inhibition of IL-1␤ and TNF␣ production (IC50 33 nM and 61 nM, respectively) occurred after only 6 hours of SNX-7081 pretreatment. IL-8 inhibition was relatively weaker. IL-6 production in both HUVECs and RASFs was markedly inhibited. In contrast to the partial abrogation of NF-B nuclear translocation, if a particular cytokine appeared sensitive to Hsp90 inhibition, SNX-7081 treatment at higher concentrations led to complete abrogation of its production. In addition, overall proliferation of activated HUVECs was potently inhibited by SNX-7081 (IC50 3 nM), but RASF proliferation showed minimal sensitivity (IC50 7.8 M). Effects of SNX-7081 on MAP kinase and angiogenic signaling. Activation of MAP kinase signaling pathways (ERK-1/2, p38, JNK) is implicated in aberrant inflammatory responses (44). The ability of SNX-7081 to inhibit two of these pathways, i.e., ERK-1/2 signaling as measured by phospho-ERK levels after growth factor activation and JNK signaling as measured by phospho– c-Jun levels after IL-1␤ activation, was evaluated in stimulated HUVECs (Figure 3). Growth factor–starved HUVECs showed increased levels of ERK-1/2 signaling upon activation with a mixture of EGF, IGF-1, VEGF, FGF, and a low percentage of FBS (2% final in medium); pretreatment with SNX-7081 fully blocked this effect, with a mean ⫾ SD IC50 of 11 ⫾ 6 nM. In contrast to ERK-1/2 signaling, a higher baseline level of JNK signaling was observed as measured by phospho–c-Jun levels, which increased ⬃2.5-fold upon IL-1␤ stimulation. Nonetheless, SNX-7081 also demonstrated the ability to markedly reduce phospho–c-Jun levels, with an IC50 of 94 ⫾ 19 nM. At maximum inhibition, levels of phospho–c-Jun were ⬃25% below levels measured in growth factor–starved HUVECs (Figure 3C). SNX-7081 was further investigated for its ability to block PDGF-initiated signaling. Using a mixture of 100 ng/ml each of PDGF-BB and PDGF-DD to activate NIH3T3 fibroblasts, global tyrosine signaling was assessed by measuring phospho–tyrosine 100 staining (Figures 3B and C). This activation was fully reversible by 3770 RICE ET AL Table 1. Inhibition of cytokine release following treatment with SNX-7081* Cells, treatment Cytokine THP-1 cells, 18-hour activation with LPS HUVECs, 24-hour activation with IL-1␤ RASFs, 24-hour activation with IL-1␤ IL-1␤ IL-6 IL-8 IL-15 G-CSF GM-CSF MCP-1 MIP-1␣ MIP-1␤ RANTES TNF␣ IFN␥ IP-10 33 ⫾ 13 NE 791 ⫾ 300 NE NE NE 76 ⫾ 5 132 ⫾ 66 NE 250 ⫾ 59 61 ⫾ 24 NE NE ND 50 ⫾ 19 161 ⫾ 81 88 ⫾ 10 57 ⫾ 10 43 ⫾ 22 94 ⫾ 9 NE 60 ⫾ 12 53 ⫾ 22 937 ⫾ 37 241 ⫾ 91 56 ⫾ 16 ND 27 ⫾ 16 44 ⫾ 9 56 ⫾ 15 13 ⫾ 4 95 ⫾ 41 NE NE NE 31 ⫾ 23 NE 189 ⫾ 129 NE * Values are the mean ⫾ SEM 50% inhibition concentration (IC50; nM). THP-1 cells were pretreated with SNX-7081 for 6 hours; human umbilical vein endothelial cells (HUVECs) and rheumatoid arthritis synovial fibroblasts (RASFs) were pretreated with SNX-7081 for 24 hours. LPS ⫽ lipopolysaccharide; IL-1␤ ⫽ interleukin-1␤; ND ⫽ not determined (since the cytokine was used for stimulation); NE ⫽ no observed effect (IC50 ⬎ 10 M); G-CSF ⫽ granulocyte colony-stimulating factor; GM-CSF ⫽ granulocyte–macrophage colony-stimulating factor; MCP-1 ⫽ monocyte chemotactic protein 1; MIP-1␣ ⫽ macrophage inflammatory protein 1␣; TNF␣ ⫽ tumor necrosis factor ␣; IFN␥ ⫽ interferon-␥; IP-10 ⫽ IFN␥-inducible 10-kd protein. Figure 3. Effects of SNX-7081 on cellular signaling events. A, Images depicting levels of phospho-ERK in control growth factor–starved HUVECs, growth factor–stimulated HUVECs, and HUVECs pretreated for 24 hours with SNX-7081 prior to stimulation with growth factor. SNX-7081 prevented both activation of and increased levels of phospho-ERK. B, Images depicting levels of phospho–tyrosine 100 (p-tyr) and phospho-ERK in control serum-starved NIH3T3 cells, serum-starved NIH3T3 cells activated with platelet-derived growth factor (PDGF), and serum-starved NIH3T3 cells pretreated for 24 hours with SNX-7081 prior to activation with PDGF. Levels of phospho-tyrosine and phospho-ERK were quantified by high-content analysis methods. SNX-7081 prevented both activation of and increased levels of phospho–tyrosine and phospho-ERK. (Magnification ⫻ 10.) C, Signal transduction inhibition by SNX-7081. The degree of inhibition was measured by scaling to the ratio of signal transduction in activated versus nonactivated cells (mean ⫾ SD of triplicate determinations). D, Levels of nitric oxide (NO) production in J774 macrophage cells. Activation of J774 cells with 100 ng/ml LPS caused an increase in NO production, which was blocked by 6-hour pretreatment with SNX-7081. The degree of inhibition of NO release by SNX-7081 was measured by scaling to the ratio of NO release in activated versus nonactivated cells (mean ⫾ SD of quadruplicate determinations). See Figure 2 for other definitions. SMALL MOLECULE INHIBITORS OF Hsp90 IN INFLAMMATORY DISEASE treatment with SNX-7081, and concentration response evaluation demonstrated an IC50 of 56 ⫾ 7 nM. Additionally, staining for phospho–ERK-1/2 following PDGF activation demonstrated that phospho-ERK signaling was inhibited by SNX-7081, with an IC50 of 14 ⫾ 4 nM. Nitric oxide production. Stimulation of angiogenesis occurs in part through increased production of NO, and the increase is accomplished by activation of a complex involving Akt, Hsp90, and eNOS (33–38). SNX7081 was evaluated to determine whether its inhibition of Hsp90 would disrupt NO production. J774 macrophages were activated with LPS for 18 hours, either in the absence of SNX-7081or following 6-hour pretreatment with SNX-7081. SNX-7081 pretreatment caused potent disruption of NO production (Figure 3D), with an IC50 of 6 ⫾ 2 nM. In vivo evaluation of an orally active analog of SNX-7081 in preclinical models of arthritis. Pharmacokinetic evaluation of SNX-7081 indicated that its exposure levels following oral dosing were inadequate to support evaluation of the compound in vivo. However, SNX-4414, a closely related analog containing a simple solubilizing prodrug functionality, was identified and found to have favorable pharmacokinetic properties following oral dosing (Figure 4A). SNX-4414 is rapidly converted from prodrug to active species in vivo, and the parent compound showed a well-behaved absorption profile with a half-life of 8.7 hours and a plasma exposure of 23 g/ml/hour following administration of a 5 mg/kg dose. SNX-4414 was evaluated in both the rat CIA model and the rat AIA model (42). Consistent with the in vitro profile observed with this class of compounds, significant inhibition of inflammation was demonstrated in vivo (Figures 4B and C). A clear dose-response and full suppression of ankle swelling were attained with both every-2-day and every-3-day treatment schedules. Measurement of body weights showed improved weight relative to both untreated and dexamethasone-treated arthritic rats (Figure 5B), and splenomegaly was reversed. Histopathologic evaluation of ankle and knee joint specimens for inflammation, pannus formation, cartilage damage, and bone erosion showed marked inhibition of these parameters (Figure 5A). Joint sections from both the ankle and the lateral knee showed extensive pannus formation and severe synovitis in vehicle-treated rats (Figure 5C). In contrast, joint sections from rats treated with 8 mg/kg oral SNX-4414 did not exhibit histopathologic abnormalities and were characterized by normal synovium and absence of cartilage damage or bone erosion. 3771 Figure 4. In vivo evaluation of the effects of SNX-4414 treatment. A, Pharmacokinetics following oral (PO) dosing. Values are the mean from 3 rats treated with a single 5 mg/kg dose of SNX-4414. B, Ankle diameter in rats with collagen-induced arthritis treated with SNX-4414 administered once every 2 days (Q2D) or once every 3 days (Q3D). Values are the mean ⫾ SEM in 8 rats per group; values in rats treated with 8 mg/kg SNX-4414 every 2 days and in positive control rats treated with 0.6 mg/kg dexamethasone (Dex) per day (QD) overlie the values in normal control rats. C, Ankle diameter in rats with adjuvantinduced arthritis treated with SNX-4414 administered once every 2 days. Values are the mean ⫾ SEM in 8 rats per group; values in rats treated with 8 mg/kg, 10 mg/kg, and 12 mg/kg SNX-4414 overlie the values in normal control rats. DISCUSSION Previous clinical evaluations of the therapeutic potential of Hsp90 inhibition have focused on oncology. 3772 RICE ET AL Figure 5. Effects of SNX-4414 in rat arthritis models. A, Histopathologic scores in the collagen-induced arthritis (CIA) model. Histopathologic parameters in rats treated with vehicle or with 8 mg/kg SNX-4414 once every 2 days were scored on a scale of 0–5. B, Weight effects in the adjuvant-induced arthritis model. Rats were treated with 10 mg/kg SNX-4414 once every 2 days (Q2D), with vehicle, or with 0.1 mg/kg dexamethasone (Dex) per day (QD). Values in A and B are the mean ⫾ SEM in 8 rats per group. C, Photomicrographs of the lateral knee and ankle in the CIA model. Specimens from a normal rat and a rat treated with 8 mg/kg SNX-4414 once every 2 days exhibit normal synovium, no cartilage damage (large arrows), and no marginal zone pannus or bone resorption (small arrows). The specimen from a vehicle-treated arthritic rat exhibits severe synovitis, marked cartilage damage (large arrow), and moderate pannus and bone resorption (small arrows). F ⫽ fibula; T ⫽tibia (original magnification ⫻ 25). However, research has indicated that Hsp90 is also critically involved in multiple processes related to inflammation, and that inhibition of Hsp90 might therefore also be of benefit in autoimmunity-related diseases. The commonalities of RA and cancer have been frequently noted (1,2,4). At least 2 compounds, methotrexate and rituximab, are approved for both cancer and RA. Recently, there was an anecdotal report of RA responding to the cancer drug imatinib in patients who had both diseases (45). In the present study, we evaluated the ability of a novel small molecule chemical scaffold, focusing on the compound SNX-7081, to impact inflammation generally and preclinical models of RA in particular. The objective was to assess the ability of SNX7081 to affect specific cellular processes such as NF-B and PDGF pathway signaling, and also to consider downstream effects relevant to inflammation, such as cytokine and NO production. SNX-7081 was observed to inhibit NF-B nuclear translocation, consistent with reports that proteins involved in NF-B pathway signaling, IKK and RIP, are Hsp90 clients (14–17). The induction of Hsp70, as a result of Hsp90 inhibition (46), may also contribute to the observed effects. Hsp70 induction has antiapoptotic effects (7,47) and, unlike in cancer cells, in which Hsp70 levels are typically strongly up-regulated, this effect may be beneficial in the setting of inflammatory disease. We also observed that the maximal level of inhibition attainable with SNX-7081 did not completely abolish NF-B translocation. However, full blockade of cytokine production was observed with multiple cytokines following SNX-7081 treatment. One possible explanation for these observations is that, even if some limited transcriptional activation remains due to residual NF-B translocation, posttranscriptional processes are also affected by Hsp90 inhibition. Supporting this rationale is the observation by others that activation of p38 can be blocked by inhibiting Hsp90 (11,13,18). Activated p38 is implicated in maintaining messenger RNA (mRNA) stability through its actions on 3⬘ AU-rich elements involved in transcriptional stability (9,11). The effect of JNK signaling knockdown, as shown here, on mRNA stability is less clear. Another interesting consideration is whether heat-shock SMALL MOLECULE INHIBITORS OF Hsp90 IN INFLAMMATORY DISEASE factor 1 (HSF-1) may participate in the observed effects of SNX-7081 on cytokine production. HSF-1 is known to block transcription of certain cytokine genes (48). This function could be relevant to the results obtained with SNX-7081, since HSF-1 is normally maintained in a repressed monomeric state through binding to Hsp90, but is released when Hsp90 is inhibited (49). Hsp90 inhibition by a compound related to SNX7081 has been shown previously to impact p-ERK signaling in a cancer cell line (41). We have demonstrated in this study that SNX-7081 is capable of abrogating ERK-1/2 and JNK signaling in HUVECs and 3T3 fibroblasts following growth factor stimulation. It is interesting to note that, while p38 has been a well-established therapeutic target for inflammation and RA, the RAF/ MEK/ERK pathway has been traditionally thought to be more relevant to oncology. However, MEK inhibition has also recently shown promise as a strategy for RA treatment (50). An Hsp90 inhibitor may be attractive in that different activated cell types in the RA phenotype may either preferentially use certain MAP kinase signaling pathways or make redundant use of them (2,44), so a general blockade of activated signaling would be preferred. Angiogenesis is viewed as a key element of the invasive and destructive phenotype of RA (1,4), and the importance of Hsp90 for VEGF and HIF-1␣–related signaling has been established (25–30). The findings described herein focus on additional aspects of angiogenesis and the role of Hsp90. First, we evaluated the effects of Hsp90 inhibition on PDGF signaling, given the increasing understanding of its importance in stromal angiogenesis and the recent observation that PDGFR␣ is also a client of Hsp90 (40). Treatment with SNX-7081 potently disrupted PDGF-BB activation of protein kinase signaling in HUVECs, RASF cells, and NIH3T3 fibroblasts. SNX-7081 also abrogated NO production in stimulated macrophages. Multiple mechanisms may account for this strong inhibition. More specifically, maximal eNOS activity requires a complex between eNOS and Hsp90 with functional ATPase activity, as well as recruitment of activated Akt (32–38). Disruption of either the Hsp90 ATPase activity or Akt phosphorylation down-regulates eNOS activity. SNX-7081 potently inhibited Hsp90 ATPase activity. Akt phosphorylation was not assessed in this study, but an SNX-7081–related compound has been shown to block Akt activation in proliferating cells (41). A key feature of the chemical series described herein is that many of its members are potent Hsp90 inhibitors that can be administered orally with good 3773 pharmacokinetics. The in vitro effects shown for SNX7081 were validated in vivo by the efficacy of the analog SNX-4414 in preclinical models of RA. Ankle and knee swelling were eliminated in both CIA and AIA. Efficacy at well-tolerated doses was equal to or superior to that of dexamethasone, and in contrast to arthritic rats treated with dexamethasone, continued weight loss was not observed in those treated with SNX-4414. Histopathologic evaluation of joint sections revealed normal synovium, bone, and cartilage. Given the broad spectrum of cellular roles played by Hsp90, the central question arises as to whether an Hsp90 inhibitor would have a sufficient therapeutic index to be efficacious in the treatment of RA or other inflammatory diseases. Obviously, although the animal data presented here are encouraging, this question can ultimately be addressed only by clinical evaluation. However, several observations may suggest that Hsp90 inhibition would be effective in treating inflammation, possibly even more than has been found in cancer. In cancer, one treatment goal is to disrupt oncogenic signaling via Hsp90 client degradation, but the more difficult and stringent goal, perhaps achieved in combination with other agents, is to cause cancer cell apoptosis. For inflammation, simple disruption of Hsp90 client–based signaling may well be sufficient to have meaningful impact on disease. Encouragingly, 6-houronly treatments with SNX-7081 were able to cause substantial cellular response. Additionally, the rebound of elements such as cytokine and phosphorylation signaling, following Hsp90 client recovery, may be more muted in RA or other inflammatory diseases relative to proliferation in cancer, such that less frequent dosing with an Hsp90 inhibitor would be needed for achievement of a prolonged effect. RA is a multifaceted disease for which additional treatments are still needed. The pleiotropic abilities of the Hsp90 inhibitors described herein to affect cytokine production, signaling cascades, and angiogenesis support their further evaluation as therapies for RA and other inflammatory diseases. AUTHOR CONTRIBUTIONS Dr. Veal had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study design. Rice, Veal, Fadden, Barta, Steed, Hall. Acquisition of data. Rice, Barabasz, Partridge, Dubois, Mabbett, Silinski, Steed. Analysis and interpretation of data. Rice, Veal, Barabasz, Partridge, Dubois, Mabbett, Silinski, Steed, Hall. Manuscript preparation. Rice, Veal, Hall. 3774 RICE ET AL Statistical analysis. Rice. Design and synthesis of compounds. Barta, Huang. REFERENCES 1. Lee DM, Weinblatt ME. Rheumatoid arthritis. Lancet 2001;358: 903–11. 2. Firestein GS. Evolving concepts of rheumatoid arthritis. Nature 2003;423:356–61. 3. McInnes IB, Schett G. Cytokines in the pathogenesis of rheumatoid arthritis. Nat Rev Immunol 2007;7:429–42. 4. Bainbridge J, Sivakumar B, Paleolog E. Angiogenesis as a therapeutic target in arthritis: lessons from oncology. Curr Pharm Des 2006;12:2631–44. 5. Smolen JS, Aletaha D, Koeller M, Weisman MH, Emery P. New therapies for treatment of rheumatoid arthritis. Lancet 2007;370: 1861–74. 6. Pratt WB, Toft DO. Regulation of signaling protein function and trafficking by the hsp90/hsp70-based chaperone machinery. Exp Biol Med (Maywood) 2003;228:111–33. 7. Whitesell L, Lindquist SL. HSP90 and the chaperoning of cancer. Nat Rev Cancer 2005;5:761–72. 8. Neckers L. Heat shock protein 90: the cancer chaperone. J Biosci 2007;32:517–30. 9. Kastelic T, Schnyder J, Leutwiler A, Traber R, Streit B, Niggli H, et al. Induction of rapid IL-1␤ mRNA degradation in THP-1 cells mediated through the AU-rich region in the 3⬘UTR by a radicicol analogue. Cytokine 1996;8:751–61. 10. Zhu FG, Pisetsky DS. Role of the heat shock protein 90 in immune response stimulation by bacterial DNA and synthetic oligonucleotides. Infect Immun 2001;69:5546–52. 11. Wax S, Piecyk M, Maritim B, Anderson P. Geldanamycin inhibits the production of inflammatory cytokines in activated macrophages by reducing the stability and translation of cytokine transcripts. Arthritis Rheum 2003;48:541–50. 12. Tago K, Tsukahara F, Naruse M, Yoshioka T, Takano K. Hsp90 inhibitors attenuate effect of dexamethasone on activated NF-B and AP-1. Life Sci 2004;74:1981–92. 13. Hsu HY, Wu HL, Tan SK, Li VP, Wang WT, Hsu J, et al. Geldanamycin interferes with the 90-kDa heat shock protein, affecting lipopolysaccharide-mediated interleukin-1 expression and apoptosis within macrophages. Mol Pharmacol 2007;71: 344–56. 14. Lewis J, Devin A, Miller A, Lin Y, Rodriguez Y, Neckers L, et al. Disruption of hsp90 function results in degradation of the death domain kinase, receptor-interacting protein (RIP), and blockage of tumor necrosis factor-induced nuclear factor-B activation. J Biol Chem 2000;275:10519–26. 15. Broemer M, Krappmann D, Scheidereit C. Requirement of Hsp90 activity for IB kinase (IKK) biosynthesis and for constitutive and inducible IKK and NF-B activation. Oncogene 2004;23:5378–86. 16. Pittet JF, Lee H, Pespeni M, O’Mahony A, Roux J, Welch WJ. Stress-induced inhibition of the NF-B signaling pathway results from the insolubilization of the IB kinase complex following its dissociation from heat shock protein 90. J Immunol 2005;174: 384–94. 17. Wang X, Ju W, Renouard J, Aden J, Belinsky SA, Lin Y. 17-allylamino-17-demethoxygeldanamycin synergistically potentiates tumor necrosis factor-induced lung cancer cell death by blocking the nuclear factor-B pathway. Cancer Res 2006;66: 1089–95. 18. De Nardo D, Masendycz P, Ho S, Cross M, Fleetwood AJ, Reynolds EC, et al. A central role for the Hsp90.Cdc37 molecular chaperone module in interleukin-1 receptor-associated-kinasedependent signaling by Toll-like receptors. J Biol Chem 2005;280: 9813–22. 19. Vega VL, De Maio A. Geldanamycin treatment ameliorates the response to LPS in murine macrophages by decreasing CD14 surface expression. Mol Biol Cell 2003;14:764–73. 20. Triantafilou K, Triantafilou M, Dedrick RL. A CD14-independent LPS receptor cluster [published erratum appears in Nat Immunol 2001;2:658]. Nat Immunol 2001;2:338–45. 21. Mayor A, Martinon F, De Smedt T, Petrilli V, Tschopp J. A crucial function of SGT1 and HSP90 in inflammasome activity links mammalian and plant innate immune responses. Nat Immunol 2007;8:497–503. 22. Fumo G, Akin C, Metcalfe DD, Neckers L. 17-allylamino-17demethoxygeldanamycin (17-AAG) is effective in down-regulating mutated, constitutively activated KIT protein in human mast cells. Blood 2004;103:1078–84. 23. Bucci M, Roviezzo F, Cicala C, Sessa WC, Cirino G. Geldanamycin, an inhibitor of heat shock protein 90 (Hsp90) mediated signal transduction has anti-inflammatory effects and interacts with glucocorticoid receptor in vivo. Br J Pharmacol 2000;131:13–6. 24. Poulaki V, Iliaki E, Mitsiades N, Mitsiades CS, Paulus YN, Bula DV, et al. Inhibition of Hsp90 attenuates inflammation in endotoxin-induced uveitis. FASEB J 2007;21:2113–23. 25. Gradin K, McGuire J, Wenger RH, Kvietikova I, Whitelaw ML, Toftgard R, et al. Functional interference between hypoxia and dioxin signal transduction pathways: competition for recruitment of the Arnt transcription factor. Mol Cell Biol 1996;16:5221–31. 26. Minet E, Mottet D, Michel G, Roland I, Raes M, Remacle J, et al. Hypoxia-induced activation of HIF-1: role of HIF-1␣-Hsp90 interaction. FEBS Lett 1999;460:251–6. 27. Mabjeesh NJ, Post DE, Willard MT, Kaur B, Van Meir EG, Simons JW, et al. Geldanamycin induces degradation of hypoxiainducible factor 1␣ protein via the proteosome pathway in prostate cancer cells. Cancer Res 2002;62:2478–82. 28. Isaacs JS, Jung YJ, Mimnaugh EG, Martinez A, Cuttitta F, Neckers LM. Hsp90 regulates a von Hippel Lindau-independent hypoxia-inducible factor-1␣-degradative pathway. J Biol Chem 2002;277:29936–44. 29. Liu YV, Baek JH, Zhang H, Diez R, Cole RN, Semenza GL. RACK1 competes with HSP90 for binding to HIF-1␣ and is required for O(2)-independent and HSP90 inhibitor-induced degradation of HIF-1␣. Mol Cell 2007;25:207–17. 30. Sanderson S, Valenti M, Gowan S, Patterson L, Ahmad Z, Workman P, et al. Benzoquinone ansamycin heat shock protein 90 inhibitors modulate multiple functions required for tumor angiogenesis. Mol Cancer Ther 2006;5:522–32. 31. Sato S, Fujita N, Tsuruo T. Modulation of Akt kinase activity by binding to Hsp90. Proc Natl Acad Sci U S A 2000;97:10832–7. 32. Wei Q, Xia Y. Roles of 3-phosphoinositide-dependent kinase 1 in the regulation of endothelial nitric-oxide synthase phosphorylation and function by heat shock protein 90. J Biol Chem 2005;280: 18081–6. 33. Garcia-Cardena G, Fan R, Shah V, Sorrentino R, Cirino G, Papapetropoulos A, et al. Dynamic activation of endothelial nitric oxide synthase by Hsp90. Nature 1998;392:821–4. 34. Song Y, Zweier JL, Xia Y. Heat-shock protein 90 augments neuronal nitric oxide synthase activity by enhancing Ca2⫹/calmodulin binding. Biochem J 2001;355(Pt 2):357–60. 35. Brouet A, Sonveaux P, Dessy C, Balligand JL, Feron O. Hsp90 ensures the transition from the early Ca2⫹-dependent to the late phosphorylation-dependent activation of the endothelial nitricoxide synthase in vascular endothelial growth factor-exposed endothelial cells. J Biol Chem 2001;276:32663–9. 36. Fontana J, Fulton D, Chen Y, Fairchild TA, McCabe TJ, Fujita N, et al. Domain mapping studies reveal that the M domain of hsp90 serves as a molecular scaffold to regulate Akt-dependent phosphorylation of endothelial nitric oxide synthase and NO release. Circ Res 2002;90:866–73. 37. Takahashi S, Mendelsohn ME. Synergistic activation of endothe- SMALL MOLECULE INHIBITORS OF Hsp90 IN INFLAMMATORY DISEASE 38. 39. 40. 41. 42. 43. 44. lial nitric-oxide synthase (eNOS) by HSP90 and Akt: calciumindependent eNOS activation involves formation of an HSP90Akt-CaM-bound eNOS complex. J Biol Chem 2003;278:30821–7. Sun J, Liao JK. Induction of angiogenesis by heat shock protein 90 mediated by protein kinase Akt and endothelial nitric oxide synthase. Arterioscler Thromb Vasc Biol 2004;24:2238–44. Kaur G, Belotti D, Burger AM, Fisher-Nielson K, Borsotti P, Riccardi E, et al. Antiangiogenic properties of 17-(dimethylaminoethylamino)-17-demethoxygeldanamycin: an orally bioavailable heat shock protein 90 modulator. Clin Cancer Res 2004;10: 4813–21. Matei D, Satpathy M, Cao L, Lai YC, Nakshatri H, Donner DB. The platelet-derived growth factor receptor ␣ is destabilized by geldanamycins in cancer cells. J Biol Chem 2007;282:445–53. Chandarlapaty S, Sawai A, Ye Q, Scott A, Silinski M, Huang K, et al. SNX2112, a synthetic heat shock protein 90 inhibitor, has potent antitumor activity against HER kinase-dependent cancers. Clin Cancer Res 2008;14:240–8. Bendele A. Animal models of rheumatoid arthritis. J Musculoskelet Neuronal Interact 2001;1:377–85. Li Q, Verma IM. NF-B regulation in the immune system. Nat Rev Immunol 2002;2:725–34. Schett G, Tohidast-Akrad M, Smolen JS, Schmid BJ, Steiner CW, Bitzan P, et al. Activation, differential localization, and regulation of the stress-activated protein kinases, extracellular signal–regulated kinase, c-Jun N-terminal kinase, and p38 mitogen-activated protein 45. 46. 47. 48. 49. 50. 3775 kinase, in synovial tissue and cells in rheumatoid arthritis. Arthritis Rheum 2000;43:2501–12. Paniagua RT, Sharpe O, Ho PP, Chan SM, Chang A, Higgins JP, et al. Selective tyrosine kinase inhibition by imatinib mesylate for the treatment of autoimmune arthritis. J Clin Invest 2006;116: 2633–42. Salminen A, Paimela T, Suuronen T, Kaarniranta K. Innate immunity meets with cellular stress at the IKK complex: regulation of the IKK complex by HSP70 and HSP90. Immunol Lett 2008; 117:9–15. Daugaard M, Rohde M, Jaattela M. The heat shock protein 70 family: highly homologous proteins with overlapping and distinct functions. FEBS Lett 2007;581:3702–10. Calderwood SK, Xiao X, Xie Y. The inside story: antiinflammatory roles of HSF1 and heat shock proteins. In: Asea AA, De Maio A, editors. Heat shock proteins: potent mediators of inflammation and immunity. Dordrecht (The Netherlands): Springer; 2007. p. 95–113. Zou J, Guo Y, Guettouche T, Smith DF, Voellmy R. Repression of heat shock transcription factor HSF1 activation by HSP90 (HSP90 complex) that forms a stress-sensitive complex with HSF1. Cell 1998;94:471–80. Thiel MJ, Schaefer CJ, Lesch ME, Mobley JL, Dudley DT, Tecle H, et al. Central role of the MEK/ERK MAP kinase pathway in a mouse model of rheumatoid arthritis: potential proinflammatory mechanisms. Arthritis Rheum 2007;56:3347–57.