| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Research Articles: Therapeutics, Targets, and Development
Comparison of biochemical and biological effects of ML858 (salinosporamide A) and bortezomib
Millennium Pharmaceuticals, Inc., Cambridge, Massachusetts
Requests for reprints: Mark J. Williamson, Millennium Pharmaceuticals, Inc., 40 Landsdowne Street, Cambridge, MA 02139. Phone: 617-551-7847; Fax: 617-551-8906. E-mail: mark.williamson{at}mpi.com
Abstract
Strains within the genus Salinospora have been shown to produce complex natural products having antibiotic and antiproliferative activities. The biochemical basis for the cytotoxic effects of salinosporamide A has been linked to its ability to inhibit the proteasome. Synthetically accessible salinosporamide A (ML858) was used to determine its biochemical and biological activities and to compare its effects with those of bortezomib. ML858 and bortezomib show time- and concentration-dependent inhibition of the proteasome in vitro. However, unlike bortezomib, which is a reversible inhibitor, ML858 covalently binds to the proteasome, resulting in the irreversible inhibition of 20S proteasome activity. ML858 was equipotent to bortezomib in cell-based reporter stabilization assays, but due to intramolecular instability is less potent in long-term assays. ML858 failed to maintain levels of proteasome inhibition necessary to achieve efficacy in tumor models responsive to bortezomib. Our results show that ML858 and bortezomib exhibit different kinetic and pharmacologic profiles and suggest that additional characterization of ML858 is warranted before its therapeutic potential can be fully appreciated. [Mol Cancer Ther 2006;5(12):3052–61]
Introduction
The modulation of intracellular protein homeostasis is an appealing therapeutic strategy because cancerous cells are dependent on the synthesis, ubiquitination, and subsequent degradation of proteins (1–4). The 26S proteasome is a multicatalytic protease that is responsible for the majority of intracellular protein turnover in eukaryotic cells (5, 6). The catalytic core of this complex is the 20S proteasome, a multisubunit complex of 
700 kDa molecular mass (7). Some of the cellular processes that use proteolysis as a regulatory mechanism are the cell cycle, stress signaling, apoptosis, and the inflammatory response (8–10). Many of these processes are deregulated in human cancers and are therefore considered to be of importance in the pathogenesis of human malignancies.
The therapeutic potential of targeting the ubiquitin proteasome pathway as a means of cancer treatment has been shown by the preclinical and clinical antitumor activity of bortezomib (VELCADE, Millennium Pharmaceuticals, Inc., Cambridge, MA), a first in class proteasome inhibitor that is approved by the Food and Drug Administration for the treatment of multiple myeloma patients who have received at least one prior therapy (11–13). Success in targeting this pathway has encouraged the evaluation of new proteasome inhibitors for clinical activity. Salinosporamide A is a bioactive metabolite derived from a group of obligate marine actinomycete bacteria that exhibits potent cancer cell cytotoxicity through inhibition of the 20S proteasome (14, 15). Structurally, salinosporamide A (Fig. 1 ) resembles the ß-lactone derived from the natural product lactacystin and the related compound, PS-519 (16–18). Recently, Corey and coworkers (19, 20) reported total syntheses of salinosporamide A and analogues.
|
Materials and Methods
Proteasome Inactivation Assays
Proteasome inactivation measurements were obtained by monitoring progress curves of hydrolysis for the site-specific 7-amido-4-methylcoumarin (AMC)–labeled peptide substrates (ß5, Suc-LLVY-AMC; ß2, Z-VLR-AMC, and ß1, Z-LLE-AMC). Assays were done in cuvettes with 2 mL of 50 mmol/L HEPES (pH 7.5), 0.5 mmol/L EDTA, at 37 ± 0.2°C, and with continuous stirring. The concentrations of substrates varied from 10 to 25 µmol/L (<1/2 KM). The concentration of human 20S proteasome was 1 nmol/L and was activated with 4 nmol/L recombinant PA28
(Boston Biochem, Cambridge, MA). The kobs values were estimated by nonlinear least-squares regression of fluorescence on time using the equation for time-dependent or slow-binding inhibition (26): F = vst + (vi – vs / kobs) [1 – exp(–kobst)], where F is fluorescence, vi and vs are the initial and final velocities of the reaction in the presence of inhibitor, and t is time. kobs/[I] values were estimated by taking the slope of a linear regression of kobs values on [I].
Sample Preparation for Proteasome Enzymatic Assays
To prepare tissue culture cell samples, cells were harvested and resuspended in 5 mmol/L EDTA (pH 8.0), frozen at –80°C, then thawed and centrifuged at 10,000 x g for 10 min at 4°C. The supernatant served as the source of 20S. For preparation of blood samples, 10 µL of whole blood were added to 300 µL cold lysis buffer (5 mmol/L EDTA), mixed, and incubated on ice for 15 to 20 min. Samples were centrifuged at 6,600 x g, 4°C, for 20 min. The resulting supernatant (200 µL) was added to 200 µL buffer [40 mmol/L HEPES (pH 8.0), 1 mmol/L EDTA, and 20% glycerol], mixed, and assayed immediately or stored at –80°C. Tumor and brain samples were prepared as follows: Samples (100–300 mg) were frozen and mechanically disrupted in a Covaris E–series instrument (Covaris, Inc., Woburn, MA) in 50 mmol/L HEPES (pH 8.0) and 1 mmol/L DTT. Samples were centrifuged for 10 min at 14,000 rpm at 4°C. Supernatants were mixed with equal volumes of stabilizing buffer and assayed immediately or stored at –80°C.
20S Proteasome Assays
For 20S ß5 proteasome assay, lysates prepared from cells, blood, or tissues were thawed at room temperature and diluted 1:5 with proteasome assay buffer [20 mmol/L HEPES, 0.5 mmol/L EDTA (pH 7.4)]. Five microliters of lysate were added to wells of a 96-well plate followed by 100 µL of proteasome assay buffer at 37°C containing 40 µmol/L Suc-LLVY-AMC and 12 nmol/L recombinant human PA28. The progress curves were monitored in a plate reader for 1 h at 37°C. The slopes of the progress curves were converted to substrate turnover by use of an AMC calibration standard. 20S ß1 and ß2 proteasome assays were identical to the ß5 proteasome assay except that substrates Z-LLE-AMC and Z-VLR-AMC were used, respectively.
Irreversibility Experiments
Rat brain 20S proteasome (200 nmol/L; Boston Biochem) was activated with 1.2 µmol/L recombinant PA28 in 50 mmol/L Tris (pH 7.5), 20 mmol/L NaCl, 0.5 mmol/L EDTA, and 5 mmol/L ß-mercaptoethanol. Aliquots were treated with 2 µmol/L ML858, bortezomib, or DMSO, respectively, and incubated for 24 h at room temperature to ensure complete reaction and also total hydrolysis of residual ML858. Excess inhibitor was removed by G-25 size-exclusion spin chromatography pre-equilibrated with the above buffer. Immediately, proteasome activity was assessed by mixing 2 µL of sample with 200 µL of 20 µmol/L Suc-LLVY-AMC, 19 nmol/L PA28 in 20 mmol/L HEPES (pH 7.5), and 0.5 mmol/L EDTA; reactions were read at 37°C in a plate reader. Rates of reaction and koff values were derived as above.
Equilibrium Dialysis
Human whole blood (heparinized) was either fortified with [14C]bortezomib (200 nmol/L, 5 min at 37°C) or pretreated with 10 mmol/L ML858 (30 min at 37°C) and then fortified with [14C]bortezomib. RBCs were obtained by centrifugation at 1,500 x g; fractions were washed once with PBS. Washed RBC fractions were diluted with PBS (1:3) and subjected to three freeze/thaw cycles. The resulting lysates were dialyzed against naïve RBC lysate (1-mL volumes, 37°C, and 10-kDa membranes). At 0.5, 1, and 4 h, samples from the donor and acceptor reservoirs were analyzed with flame oxidation/liquid scintillation counting.
Plasmids
The Ub-FL plasmid was licensed from Washington University (St. Louis, MO). The plasmid is based on the EGFP-N1 vector (Clontech, Mountain View, CA). Briefly, enhanced green fluorescent protein was removed from EGFP-N1 by the HindIII and NotI sites and replaced with firefly luciferase, which was digested from the pGL-3 plasmid (Promega, Madison, WI) by HindIII and XbaI sites. Four tandem copies of ubiquitin G76V were then fused in frame with the NH2 terminus of firefly luciferase, using NheI and HindIII sites. The resultant ubiquitin-luciferase fusion construct was designated Ub-FL and subsequently renamed 4xUb-Luc at Millennium Pharmaceuticals. The NFkB-TA-Luc/UB-Bsd vector was made by inserting the blasticidin selection marker from pUB/Bsd (Invitrogen, Grand Island, NY) into pNFkB-TA-Luc (Clontech).
Cell Culture
Human cell lines RPMI-8226, HT29, MDA-MB-231, HEK293, and HeLa were obtained from the American Type Culture Collection (Manassas, VA). Cell line AP6 was adapted from RPMI 8226 cells at Millennium Pharmaceuticals by selecting adherent RPMI 8226 cells. All cells were maintained in DMEM (Invitrogen) containing high glucose, L-glutamine (Invitrogen), 110 mg/L sodium pyruvate and pyridoxine hydrochloride (Invitrogen), and supplemented with 10% fetal bovine serum (Hyclone, Logan, UT). Cells were cultured at 37°C, 5% CO2 atmosphere, 97% relative humidity, and were routinely passaged by trypsin-EDTA (Life Technologies, Grand Island NY) treatment.
Cell Culture Transfections
Cells were seeded at 1 x 106/10-mm plate and were transfected with plasmid DNA using lipofectamine 2000 (Invitrogen). Clonal cell lines expressing 4xUb-Luc or NFkB-TA-Luc/UB-Bsd were selected by limiting dilution or ring cloning of transfection pools maintained in parental growth medium containing 750 µg/mL geneticin (Invitrogen) or 5 µg/mL blasticidin (Invitrogen), respectively. Luciferase activity in cell lysates was determined with a luciferase assay kit (Promega) and normalized to total cell protein as measured by BCA assay kit (Pierce, Rockford, IL).
Viability Assays
One day before the intended start of the assay, AP6 cells were plated in 96-well microtiter plates in a final volume of 75 µL at 1,500 per well. Cells were placed in a tissue culture incubator overnight (37°C, 5% CO2, and 21% O2, 97% relative humidity) to allow the cells to equilibrate. Test compounds (25 µL) were then systematically added to the microtiter plates containing cells to a final volume of 100 µL and replaced into the tissue culture incubator. For ML858, the medium also contained a final concentration of 0.1% DMSO. Cell viability, based on mitochondrial function, was measured 48 h later using 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium assays as indicated by the manufacturer (Roche, Indianapolis, IN). Fifty-percent maximal effective concentration (EC50) values were calculated using Prism software (GraphPad Software, Inc., San Diego, CA).
Washout Experiments
HeLa cells were incubated with either 500 nmol/L ML858 or 500 nmol/L bortezomib for 1 h, washed in PBS, and returned to medium either lacking or containing 500 nmol/L ML858 or 500 nmol/L bortezomib for the times indicated. Cell extracts were prepared and tested for peptidase activity as indicated above.
In vitro Reporter Assays
NFkB-Luc and 4xUb-Luc activities were measured using clonally derived stable HEK293 and MDA-MB-231 cell lines, respectively. Cells were seeded at 1 x 104 per well in 384-well plates 16 to 24 h before compound treatment. For NFkB-Luc assays, cells were stimulated with 10 ng/mL tumor necrosis factor- in the presence of the indicated concentration of proteasome inhibitor for 3 h. For the 4xUb-Luc assays, cells were incubated with compound alone for 8 h. Luciferase activity was measured using Bright-Glo assay reagents (Promega) in a LEADseeker instrument (GE Healthcare Bio-Sciences Corp, Piscataway, NJ). Inhibition of NFkB-Luc activity was calculated relative to a vehicle (DMSO) control, whereas 4xUb-Luc reporter stabilization was expressed as fold luciferase activity over vehicle (DMSO).
In vivo 4xUb-Luc Reporter Assays
Tumor xenografts were formed by s.c. implanting 4 x 106 HT-29 cells stably transfected with the 4xUb-Luc reporter into the right flank of female NCR nude mice (Taconic, Hudson, NY). Tumors were allowed to grow to an average of 1,300 mm3 before bortezomib (1 mg/kg) and ML858 (0.4 mg/kg) were administered by tail-vein injection (n = 5 mice per group). Bioluminescence imaging at 0, 3, and 6 h was done with a Xenogen IVIS 100 Imaging System (Xenogen, Alameda, CA) as follows: Each mouse was dosed with 60 mg/kg luciferin (Xenogen) by i.p. injection; 15 min later, a dorsal image was taken, which was immediately followed by a ventral-image capture. Signal intensities from regions of interest were defined manually, and data were expressed as average photon flux (photons per hour).
Establishment and Treatment of the CWR22 Model
Donor mice bearing CWR22 tumors 500 to 2,000 mm3 were euthanized by CO2 asphyxiation, and tumors were dissected using sterile technique. Viable tumor fragments measuring 2 x 1.5 x 1.5 mm were implanted into the right flank of male 6-week-old CB-17 severe combined immunodeficient mice (Charles River Laboratories, Wilmington, MA) using a 13-gauge trocar 10 mm into the subcutis. Tumor volume was monitored twice weekly with caliper measurements. The mean tumor volume was calculated using the formula volume = width2 x length / 2. When the mean tumor volume reached 200 mm3, mice were randomized into treatment groups containing 12 animals per group. Animals were dosed as described in text on a twice per week schedule for a maximum period of 4 weeks, or until the humane end point was reached in >50% of the mice in that group (when tumors in vehicle arm achieved 10% of animal body weight). Tumor growth inhibition was calculated at day 21 as follows: [(control average tumor volume – treated average tumor volume) x 100 / control average tumor volume]. Statistical comparisons between study groups were made with a Student's t test.
Establishment and Treatment of the OCI-LY3/7D1-Luciferase Model
Severe combined immunodeficient-nonobese diabetic mice (Taconic) were inoculated with 1 x 106 OCI-LY3/7D1-luciferase cells in the tail vein on day 0. Thirty minutes postinjection, mice were imaged in real time with Xenogen IVIS 100 Imaging System as follows: Each mouse was dosed with 120 mg/kg luciferin by i.p. injection; 10 min later, a dorsal image was taken, which was immediately followed by a ventral-image capture. Both total signal photon flux and distribution was analyzed to ensure good grafting of tumor cells. A total of 50 mice were grafted, and five groups of eight mice were randomized into treatment groups. Beginning at day 1 or 8 after tumor cell grafting and continuing for 3 weeks thereafter, mice were dosed with compounds as indicated in text. Mice were monitored weekly to assess progression of tumor burden and tolerance to the treatment protocol. Total body mean flux was calculated on day 21 following tumor cell grafting. Tumor growth inhibition was calculated as follows: [(control average photon flux – treated average photon flux) x 100 / control average photon flux]. Statistical comparisons between study groups were made with a Student's t test.
Results
ML858 Is a Potent, Irreversible Proteasome Inactivator
First, we compared the rate constants of proteasome peptidase inactivation by ML858 and the related ß-lactones omuralide and PS-519 by monitoring progress curves for the hydrolysis of site-specific AMC-labeled peptide substrates (Table 1
). The selectivity of ML858 for the three different catalytic centers of the proteasome was similar to PS-519 and omuralide in that all three preferentially inactivated the ß5 (chymotrypsin-like) activity. Inactivation of ß5 by ML858 was five times faster than inactivation by PS-519 and 11 times faster than inactivation by omuralide under the conditions of our assay.
|
In contrast, the mechanism of proteasome inhibition by bortezomib is reversible and thus differs from the other three molecules in Table 1 (24, 25). However, bortezomib acts as a slow-binding (time-dependent) inhibitor; thus, its behavior closely resembles that of ML858 on the short time scale of these enzyme assays and therefore can be compared with the other molecules (26). The second-order rate constant for bortezomib association to the ß5 site was estimated to be 200 mmol/L–1 s–1, virtually equivalent to the rate constant for reaction at this site by ML858, 210 mmol/L–1 s–1. Bortezomib and ML858 differed in relative selectivity for the ß1 (caspase-like) and ß2 (trypsin-like) active sites as has been previously shown (16).
Irrespective of any minor differences in subunit selectivity, both bortezomib and ML858 were potent inhibitors of proteasome-catalyzed I
B degradation in cells as evidenced by their ability to inhibit tumor necrosis factor-–stimulated induction of a nuclear factor-B reporter gene (Table 1). Importantly, this assay was done by using simultaneous exposure of the cells to tumor necrosis factor- and test compound (see Materials and Methods). In independent experiments, we found that IB degradation in these cells occurs within 10 to 20 min after exposure to tumor necrosis factor- (data not shown).
Thus, we interpreted potent activity in this assay to reflect the fact that ML858, like bortezomib, rapidly entered the cells and blocked proteasome function.
ML858 Is Hydrolytically Unstable
Bortezomib is relatively stable in aqueous solutions near neutral pH, but omuralide is not; it hydrolyzes to the inactive dihydroxy acid at a considerable rate (27, 28). PS-519 had a similar liability, and Macherla et al. (15) found that salinosporamide A also hydrolyzes. We measured rates of hydrolysis for PS-519 and ML858 by two methods. The half-life in aqueous solution was estimated by directly monitoring the loss of parent compound and the appearance of the hydrolysis product by liquid chromatography-mass spectrometry as a function of time. Also, the half-life was estimated indirectly by following the loss of inhibitory activity (i.e., following the decrease in kobs/[I]) as a function of time. Both methods yielded similar results within experimental error; because the kinetics of ML858 disappearance paralleled the loss of inhibitory activity, the data suggested that its hydrolysis product is virtually devoid of inhibitory activity, analogous to the hydrolysis product of omuralide. Whereas PS-519 was unstable, with a t1/2 
32 min, the stability of ML858 was only modestly improved with t1/2 of 56 min. Our results were entirely consistent with those reported by Macherla et al. (15), who found that, unlike the hydrolysis product of omuralide, the salinosporamide A hydrolysis product underwent an additional chemical transformation, precluding its isolation in sufficient quantity to assay directly.
ML858 Competes with Bortezomib for Proteasome Binding
Bortezomib is a selective inhibitor of the proteasome with only modest activity toward serine and cysteine proteases and as shown above can inhibit the chymotrypsin-like and caspase-like activities (ß5 and ß1, respectively) of the 20S proteasome (29, 30). To assess how ML858 and bortezomib may differentially affect proteasome activity, an in vitro experiment was done with human whole blood to determine if ML858 could compete with bortezomib for proteasome binding. Human whole blood was prepared with either [14C]bortezomib or pretreated with ML858 before addition of [14C]bortezomib before subjecting the RBC lysate to equilibrium dialysis against naïve RBC lysate. [14C]bortezomib associated with RBC lysate (no ML858 pretreatment) traversed the dialysis membrane at an exceptionally slow rate (<20% after 24 h), whereas [14C]bortezomib associated with the ML858 pretreated lysate reached equilibrium within 2 h, indicating that ML858 is likely blocking binding of bortezomib to the proteasome in RBCs (see Supplementary Fig. S2).1 These data suggest that combination treatments with bortezomib and ML858 may force the distribution of either or both of these compounds to tissues not typically accessible to them, potentially changing their relative toxicity profiles.
ML858 Instability Affects Cellular Potency
Bortezomib strongly inhibits cell viability in many human tumor cell lines, demonstrating an average 50% inhibition of cell growth (GI50) of 3.8 nmol/L across the entire National Cancer Institute cell panel (11, 31–33). Salinosporamide A is also reported to affect cell viability in tissue culture (15, 34). A comparison was made between the in vitro activities of these two compounds and EC50 values were established using a dose-range treatment of bortezomib and ML858. Studies in AP6 cells, a derivative of RPMI 8226, showed that bortezomib was 5.7-fold more cytotoxic than ML858 in a 48-h viability assay (Fig. 2A
). Six additional cell lines (SKOV3, HT29, MDA-MB-231, H460, H358, and PC3) were tested in similar assays and the shift in EC50 values ranged from 2.3- to 11.9-fold for bortezomib versus ML858 after 48 h of treatment (data not shown).
|
80 and 40 nmol/L for bortezomib and ML858, respectively. Taken together, these data suggest that in vitro ML858 is as potent as bortezomib when run in short-term (3 and 8 h) reporter assays (NFkB-Luc and 4xUb-Luc, respectively), but is less potent in long-term (48 h) viability assays. This is likely a direct consequence of ML858 instability. To test this hypothesis, HeLa cells were exposed to 500 nmol/L ML858 or bortezomib for either 1 or 48 h, and proteasome activity was monitored over time. Both bortezomib and ML858 potently inhibited proteasome activity after transient exposure to compound and, as expected, proteasome activity recovered to baseline over the next 24 h (Fig. 2C and D). However, when cells were exposed to ML858 for 48 h, proteasome activity also recovered to untreated levels over time, whereas exposure to bortezomib for 48 h resulted in continuous inhibition and ultimately led to cell death (Fig. 2C and D).
In vivo 4xUb-Luc Reporter Activity of Bortezomib and ML858
Luker et al. (35) have reported the ability to assess 26S proteasome function in living animals using an ubiquitin-luciferase reporter for bioluminescence imaging. We wanted to determine whether ML858 would stabilize the 4xUb-Luc reporter in vivo. Mice bearing s.c. HT-29 tumors stably expressing the 4xUb-Luc reporter were injected with either bortezomib or ML858 (1 and 0.4 mg/kg, respectively) and luminescence was monitored 0, 3, and 6 h after injection (Fig. 3A
). We found robust time-dependent increases in luciferase activity in bortezomib-treated animals, but only a modest signal over background in the ML858-treated group (6-h time point; Fig. 3B). We also assessed exposure by measuring proteasome inhibition in whole blood, tumors, and brains taken after the 6-h time point. Both compounds exhibited significant (P < 0.001 in both cases) proteasome inhibition in whole blood (Fig. 3C); however, only bortezomib caused considerable (P < 0.05) proteasome inhibition in tumors (Fig. 3D), whereas only ML858 showed significant (P < 0.001) proteasome inhibition in the brain (Fig. 3E).
|
A comparative efficacy study was done by treating mice bearing CWR22 tumors with the maximum tolerated dose (MTD) of bortezomib (0.8 mg/kg i.v.) and ML858 (0.3 mg/kg i.v. and 0.8 mg/kg p.o.) twice weekly, corresponding to the optimal dose and schedule determined for bortezomib. Dosing was continued for 3 weeks, and 20S proteasome activity of blood and tumors was assessed at both the midpoint and the end of the study. Efficacy was determined based on treated/control (T/C) tumor weight ratio guidelines defined by National Cancer Institute studies. [According to National Cancer Institute studies, moderate efficacy is a T/C of <0.42; total growth inhibition of 58% or more; ref. 42]. Figure 4 shows that better than moderate tumor growth inhibition (T/C, 0.3) was achieved in the bortezomib treatment group (P < 0.001), but not observed in the ML858 i.v. treatment group (T/C, 0.58) or the ML858 p.o. treatment group (T/C, 0.6).
|
In vivo Antitumor Activity of Bortezomib and ML858 in the OCI-LY3/7D1-Luciferase Model
In all of our s.c. xenograft studies, treatment with ML858 consistently resulted in prolonged 20S proteasome inhibition in peripheral blood. For that reason, assessing ML858 activity in a disseminated model where tumors have the potential to exist in many tissues throughout the body was of significant interest. The OCI-LY3/7D1-luciferase model has been developed for this purpose and is sensitive to both bortezomib and Rituximab.2 Derived from the activated B cell–like, diffuse large B-cell lymphoma cell line OCI-LY3, when grafted via tail vein injection into nonobese diabetic/severe combined immunodeficient mice, OCI-LY3/7D1-luciferase cells distribute into the lower spine and limbs, causing paralysis after 3 to 4 weeks (43).
A comparative efficacy study was done in this model where bortezomib was dosed once weekly at 1.4 mg/kg i.v. and ML858 was dosed twice per week at 0.6 mg/kg via p.o. gavage. Dosing was initiated on either day 1 or day 8 (data not shown) following tail vein grafting and continued until day 21. Tumor burden was assessed using Xenogen IVIS Imaging System. The response to bortezomib and ML858 dosed from days 1 to 21 is shown in Fig. 5 . Significant tumor growth inhibition was again shown in the bortezomib treatment group (P < 0.001) but not in the ML858 groups.
|
Discussion
The present studies provide new insights into the biochemical and pharmacologic differences between the proteasome inhibitors bortezomib and ML858. Although bortezomib is a reversible proteasome inhibitor, we have shown that ML858 is a potent, irreversible proteasome inactivator. It remains to be determined if irreversibility will be a desirable feature or a liability, given that potential toxicities may only be manifested in longer-term clinical trials.
We report here for the first time that ML858, like other members of the ß-lactone family, is hydrolytically unstable. Our data suggests that this intramolecular instability affects the potency of ML858 over time. We provide evidence that ML858 exhibits similar potency to bortezomib only in short-term ubiquitin proteasome pathway reporter cell-based assays, but is less potent in long-term cell viability assays. To further explore this observation, we did washout experiments and monitored proteasome activity. These data show that some cells overcome the initial effect of proteasome inhibition caused by ML858 and bortezomib and synthesize new proteasomes. However, because ML858 has been inactivated by hydrolysis at this time, these de novo proteasomes are no longer susceptible to inhibition by ML858 and cells are able to recover and divide. Conversely, bortezomib remains intact when it is released from the proteasome and is therefore able to provide continuous proteasome inhibition, which led to cell death.
In vivo, both bortezomib and ML858 are selectively sequestered by RBC due to the micromolar intracellular proteasome concentration. Recovery of proteasome activity in RBCs, which are devoid of a nucleus, is dependent on erythrocyte turnover. Thus, to sustain higher levels of proteasome inhibition in cells other than RBCs in vivo with an irreversible, unstable compound like ML858, one may need to dose at levels that would permit preferential distribution to distant tissues, and this may not be possible given potential toxicities.
To characterize the effects of ML858 and bortezomib in vivo, we dosed mice bearing HT29 4xUb-Luc tumors with the MTD of each compound and monitored accumulation of the luciferase reporter. We chose the determined MTD dose for efficacy studies for these experiments because the physiochemical properties of the two compounds are different and an equal concentration comparison would not be a fair assessment of their activities. We observed a robust time-dependent accumulation of the luciferase reporter only in the bortezomib-treated group. Subsequent assessment of proteasome inhibition levels in whole blood and tumors 6 h posttreatment revealed that although both compounds achieved profound levels of proteasome inhibition in blood, only bortezomib achieved significant levels of proteasome inhibition in tumors at this time point.
The best method to reconcile the almost complete proteasome inhibition observed in blood with the variable levels of proteasome inhibition attained in tissues is unclear. Due to ML858 instability, we do not have a pharmacokinetic assay that allows us to assess exposure in tumors and permeability studies in cells. A couple of hypotheses can be proposed to explain the poor tissue proteasome inhibition effects. It may be that all the available drug binding sites within the blood compartment may need to become saturated before the compound is able to reach peripheral sites. Indeed, preliminary experiments with repeated dosing (comparing qd with qd·3) show titration of whole blood proteasome activity, but not tumor proteasome activity (data not shown). Alternatively, it may be that owing to the instability of ML858, very little of it reaches outside of the systemic circulation intact and the opportunity for inhibition of proteasomes other than those found in RBCs is minimal.
Chauhan et al. (21) reported that salinosporamide A shows efficacy comparable with bortezomib in a multiple myeloma xenograft tumor model. We report here that ML858 did not show significant efficacy in either a prostate s.c. or a lymphoma-disseminated xenograft tumor model. There are several possibilities as to why this discrepancy of in vivo activity may exist. First, to the best of our knowledge, it is unlikely that the source and preparation of ML858 and NPI-0052 used in these experiments are alike. ML858 is a synthetic compound, whereas NPI-0052 is a bioactive metabolite. However, through collaboration, we have analyzed ML858 and NPI-0052 by nuclear magnetic resonance and have concluded that they are identical by this analytic method. Second, our pharmacodynamic data suggested that the lack of efficacy observed in our models is likely owing to the absence of significant proteasome inhibition in the tumors (Supplementary Table S1). It is possible that ML858 was able to achieve higher levels of proteasome inhibition in the multiple myeloma model used by Chauhan et al., but this is difficult to assess as no assessment of proteasome inhibition (tumor or blood) was reported for their efficacy study. The in vivo proteasome data shown in Chauhan et al.'s Fig. 3A depicts proteasome inhibition in a separate study with whole blood after 1 dose of 10 mL/kg NPI-0052. It is clear from our internal studies with bortezomib that blood proteasome inhibition does not correlate with efficacy. Additionally, it is likely that the level of proteasome inhibition necessary to achieve efficacy will vary between cell types and models.
It is worthwhile to point out that the dosing differs significantly in the reported efficacy studies with salinosporamide A and ML858. In our studies, the MTD of ML858 was determined by dose-escalation toxicity studies where the animals were monitored for dose-limiting affects (i.e., weight loss, posture, feeding, etc.) over 7 days. As to why the MTDs may vary, it may be that the MM1.S model reported by Chauhan et al. is a more sensitive model and requires less drug, or it may be that different mouse strains used for the models may have differing MTDs. Finally, it is also possible that the isolation and purification protocols for ML858 and salinosporamide A may provide means for different potencies.
Whole-body autoradiography of rats administered [14C]bortezomib reveal that whereas most organs receive drug, there is minimal exposure in the central nervous system, testes, and eyes (44). Although we have not carried out similar studies with radiolabeled ML858, our mouse studies here suggested that ML858 distributed to the brain in amounts sufficient to cause substantial proteasome inhibition in this organ. To determine if this finding was species-dependent, we investigated ML858 in Wistar rats. Preliminary data suggest significant levels (>40%) of proteasome inhibition in the brain after an acute dose (data not shown). Surprisingly, when delivered i.v., the level of brain proteasome inhibition caused by ML858 was greater than the inhibition in the tumors of the same animals. Although brain proteasome inhibition was much reduced when ML858 was administered p.o., it is yet to be determined if this molecule could be safely and sufficiently dosed to maximize efficacy while minimizing brain exposure. The relevance of these findings is not yet known, but defects in the ubiquitin proteasome pathway have been linked with numerous neurodegenerative diseases such as Alzheimer's, Parkinson's, and polyglutamine diseases (45–48). Even the slightest risk of such a compound getting into the central nervous system coupled with covalent, irreversible proteasome inhibition has the potential to produce unacceptable toxicities that may only manifest in larger-scale, longer-term clinical trials.
In summary, our results have established that ML858 is a potent proteasome inactivator and represents an improvement in potency and stability in this class of compounds. Importantly, we report that the preclinical profile of ML858 has revealed limited efficacy in our tumor models and central nervous system penetrance not seen with bortezomib. These observations are significant in that novel proteasome inhibitors with potentially similar properties are entering into the clinic. Further characterization of ML858 is necessary to fully understand how these different properties could influence clinical response. Results of these future studies with ML858 and similar analogues should help determine the therapeutic potential for this class of proteasome inhibitors.
Acknowledgments
We thank Prof. E.J. Corey (Harvard University, Cambridge, MA) for synthetic samples of salinosporamide A and helpful discussions; Dr. G. Luker (Washington University, St. Louis, MO) for the Ub-FL plasmid; Millennium Biological Assay group for sample processing and 20S proteasome assays; Timothy Stefl for help with figures; and members of the 2GPI Project Team for useful discussions.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/). 
2 C.L. Reimer and A.M. Mazzola, unpublished data.
Received 4/ 4/06; revised 8/17/06; accepted 10/26/06.
References
B activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J Exp Med 2001;194:1861–74.This article has been cited by other articles:
|
|
 |
|
  C. M. Sloss, F. Wang, R. Liu, L. Xia, M. Houston, D. Ljungman, M. A. Palladino, and J. C. Cusack Jr. Proteasome Inhibition Activates Epidermal Growth Factor Receptor (EGFR) and EGFR-Independent Mitogenic Kinase Signaling Pathways in Pancreatic Cancer Cells Clin. Cancer Res., August 15, 2008; 14(16): 5116 - 5123. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Baritaki, E. Suzuki, K. Umezawa, D. A. Spandidos, J. Berenson, T. R. Daniels, M. L. Penichet, A. R. Jazirehi, M. Palladino, and B. Bonavida Inhibition of Yin Yang 1-Dependent Repressor Activity of DR5 Transcription and Expression by the Novel Proteasome Inhibitor NPI-0052 Contributes to its TRAIL-Enhanced Apoptosis in Cancer Cells J. Immunol., May 1, 2008; 180(9): 6199 - 6210. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Piva, B. Ruggeri, M. Williams, G. Costa, I. Tamagno, D. Ferrero, V. Giai, M. Coscia, S. Peola, M. Massaia, et al. CEP-18770: A novel, orally active proteasome inhibitor with a tumor-selective pharmacologic profile competitive with bortezomib Blood, March 1, 2008; 111(5): 2765 - 2775. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. Demo, C. J. Kirk, M. A. Aujay, T. J. Buchholz, M. Dajee, M. N. Ho, J. Jiang, G. J. Laidig, E. R. Lewis, F. Parlati, et al. Antitumor Activity of PR-171, a Novel Irreversible Inhibitor of the Proteasome Cancer Res., July 1, 2007; 67(13): 6383 - 6391. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. P. Miller, K. Ban, M. E. Dujka, D. J. McConkey, M. Munsell, M. Palladino, and J. Chandra NPI-0052, a novel proteasome inhibitor, induces caspase-8 and ROS-dependent apoptosis alone and in combination with HDAC inhibitors in leukemia cells Blood, July 1, 2007; 110(1): 267 - 277. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
) and bortezomib (
) for 48 h. Points, mean from three independent experiments; bars, SE. B, proteasome-dependent stabilization of the 4xUb-Luc reporter. MDA-MB-231 cells stably transfected with the 4xUb-Luc reporter were treated with increasing concentrations of ML858 (
) and ML858 (
). Points, ratio of photon flux from tumors at time 0 relative to tumors at 3 or 6 h. B, regions of interest analysis of HT-29 4xUb-Luc tumors after the 6-h time point by Xenogen imaging. Inhibition of ß5 activity of the proteasome in lysates of whole blood (C), tumor (D), and brain (E) from animals sacrificed after the 6-h time point. Columns, mean; bars, SE.
) via i.v., ML858 (