¹The Jerome Lipper Multiple Myeloma Center, Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115

²Nereus Pharmaceuticals, San Diego, California 92121

³Harvard Medical School, Boston, Massachusetts 02115

Summary

Bortezomib therapy has proven successful for the treatment of relapsed and/or refractory multiple myeloma (MM); however, prolonged treatment is associated with toxicity and development of drug resistance. Here, we show that the novel proteasome inhibitor NPI-0052 induces apoptosis in MM cells resistant to conventional and Bortezomib therapies. NPI-0052 is distinct from Bortezomib in its chemical structure, effects on proteasome activities, mechanisms of action, and toxicity profile against normal cells. Moreover, NPI-0052 is orally bioactive. In animal tumor model studies, NPI-0052 is well tolerated and prolongs survival, with significantly reduced tumor recurrence. Combining NPI-0052 and Bortezomib induces synergistic anti-MM activity. Our study therefore provides the rationale for clinical protocols evaluating NPI-0052, alone and together with Bortezomib, to improve patient outcome in MM.

Article Outline

SIGNIFICANCE

Introduction

Results and discussion

NPI-0052 is structurally distinct from Bortezomib and inhibits all three protease activities within the proteasome both in vitro and in vivo

NPI-0052 is orally bioactive

NPI-0052 targets nuclear factor-κB

NPI-0052 blocks proteasome activity in MM cells

NPI-0052 inhibits growth and triggers apoptosis in MM cells

NPI-0052 does not affect viability of patient MM-derived bone marrow stromal cells

NPI-0052 overcomes recombinant human interleukin-6- and recombinant human insulin-like growth factor-I-mediated antiapoptotic effects

NPI-0052 blocks vascular endothelial growth factor-induced migration of MM cells

NPI-0052 inhibits human MM cell growth in vivo and prolongs survival in a murine model

Comparative analysis of in vivo antitumor activity of NPI-0052 and Bortezomib

Mechanisms mediating anti-MM activity of NPI-0052

Differential requirement of caspases and mitochondrial signaling during NPI-0052- and Bortezomib-induced MM cell apoptosis

Differential effects of NPI-0052 and Bortezomib on Bcl-2-overexpressing MM cells

NPI-0052 and Bortezomib have differential effects on normal lymphocytes

Combined Bortezomib and NPI-0052 treatment triggers synergistic apoptosis in MM cells

Experimental procedures

In vitro and in vivo proteasome activity assays

Assaying proteasome activity by immunoblotting

Human plasmacytoma xenograft models

Expression vectors and transfections

Statistical analysis

Isobologram analysis

Acknowledgements

Supplemental data

References

SIGNIFICANCE

Introduction

The successful development of Bortezomib/PS-341 Velcade therapy for treatment of relapsed/refractory multiple myeloma (MM) has established proteasome inhibition as an effective therapeutic strategy. The dipeptide boronic acid analog Bortezomib is a potent, highly selective, and reversible proteasome inhibitor that targets the 26S proteasome complex and inhibits its function (Adams et al., 1999). Besides inhibiting NF-κB, Bortezomib has pleiotropic effects on MM biology by targeting (1) cell cycle regulatory proteins; (2) the unfolded protein response (UPR) pathway; (3) p53-mediated apoptosis; and (4) DNA repair mechanisms; as well as (5) classical stress response pathways via both intrinsic (caspase-9-mediated) and extrinsic (caspase-8-mediated) cell death cascades. Specifically, Bortezomib activates c-Jun amino-terminal kinase (JNK) (Chauhan et al., 2003), which triggers mitochondrial apoptotic signaling: release of cytochrome-c (cyto-c) and second mitochondrial activator of caspases (Smac) from mitochondria to cytosol, followed by activation of caspase-9 and caspase-3 (Chauhan and Anderson, 2003). Despite the potent anti-MM activity of Bortezomib, both intrinsic and acquired resistance has already been observed in MM patients (Anderson, 2004). The mechanisms conferring Bortezomib resistance are now being defined (Chauhan et al., 2005). Nonetheless, the combination of Bortezomib with other novel and conventional agents can overcome Bortezomib resistance (Chauhan et al., 2005).

Recent studies have focused on developing other proteasome inhibitors as therapeutics in cancer. NPI-0052 is one such molecule derived from fermentation of Salinospora, a new marine gram-positive actinomycete (Macherla et al., 2005). In the present study, we show that both NPI-0052 and Bortezomib can be distinguished by chemical structure, their effects on proteasomal activities, and differential toxicity profiles against normal cells. NPI-0052 induces apoptosis in MM cell lines and patient cells. Both biochemical and genetic studies show that NPI-0052-induced cell death, in contrast to that triggered by Bortezomib, relies primarily on caspase-8-mediated signaling pathways. In vivo studies show that NPI-0052 significantly inhibits tumor growth in mice, is well tolerated, and prolongs survival. Based on their structural and mechanistic differences, NPI-0052 and Bortezomib show synergistic anti-MM activity. Finally, current Bortezomib therapy requires intravenous (i.v.) administration, whereas NPI-0052 is orally bioactive. Overall, our study provides the rationale for clinical protocols evaluating NPI-0052 to inhibit tumor cell growth, overcome conventional and Bortezomib resistance, limit toxicity profiles, and improve patient outcome in MM.

Results and discussion

NPI-0052 is structurally distinct from Bortezomib and inhibits all three protease activities within the proteasome both in vitro and in vivo

The naturally occurring and synthetic inhibitors of the ubiquitin-proteasome pathway include peptide aldehydes, peptide boronates, nonpeptide inhibitors, peptide vinyl sulfones, and peptide epoxyketones ([Adams, 2004], [Adams et al., 1999] and [Chauhan et al., 2005]). Bortezomib/PS-341/Velcade is a boronic acid dipeptide derivative and inhibits proteasome function via interaction of boronic acid at the C terminus of Bortezomib with an active threonine site in the proteasome (Adams et al., 1999) (Figure 1A). NPI-0052 is a nonpeptide proteasome inhibitor and shares structural features with another proteasome inhibitor, Omuralide (Figure 1A), which inhibits proteasome activity by covalently modifying the active site threonine residues of the proteasome (Corey and Li, 1999). Omuralide is clasto-lactacystin β-lactone, the active form of the well-known proteasome inhibitor lactacystin (Corey and Li, 1999). Despite the structural similarity with Omuralide, NPI-0052 can be distinguished by the presence of a uniquely methylated C3 ring juncture, chlorinated alkyl group at C2, and cyclohexene ring at C5 (Figure 1A). Evaluation of the effects of NPI-0052, Omuralide, and Bortezomib on cathepsin A activity show that Omuralide is a more specific inhibitor of cathepsin A (IC₅₀ = 65 ± 6 nM). Both NPI-0052 and Bortezomib also inhibited cathepsin A with different IC₅₀ (NPI-0052, 1.4 ± 0.1 μM; Bortezomib, 14 ± 5 μM).

Omuralide is known to inhibit all three protease activities in the proteasome: the chymotrypsin-like (CT-L), trypsin-like (T-L), and caspase-like (C-L) activities (Corey and Li, 1999); however, the effect of NPI-0052 on these proteasome activities is undefined. We therefore examined whether NPI-0052 affects these proteasome activities using human erythrocyte 20S proteasomes and fluorogenic peptide substrates. We also simultaneously compared the effect of Bortezomib on proteasome activities. Both NPI-0052 and Bortezomib inhibit all three proteasome activities, albeit at different concentrations (Figures 1B–1D). Results show that (1) NPI-0052 inhibits CT-L and T-L activities at lower concentrations than Bortezomib, and (2) higher concentrations of NPI-0052 than Bortezomib are required to inhibit C-L activity. Our in vitro data suggest that NPI-0052, like Bortezomib, targets proteasomes.

We next compared the effects of NPI-0052 and Bortezomib on all three proteasome activities in vivo in mice. For these studies we selected the MTD dose for each agent. Based on our previous study (LeBlanc et al., 2002), the MTD of Bortezomib is 1.0 mg/kg (i.v.) given twice weekly. Additional experimentation in beige-nude-xid (BNX) mice established the MTD of NPI-0052 at 0.15 mg/kg (i.v.) twice weekly (data not shown). Mice were treated with a single dose of NPI-0052 (0.15 mg/kg i.v.) or Bortezomib (1 mg/kg i.v.); blood samples were collected at 90 min, 24 hr, 48 hr, 72 hr, or 168 hr; and whole blood cells were then analyzed for proteasome activity (Figure 2). NPI-0052 completely inhibited CT-L activity by 90 min, which was recoverable by 168 hr (Figure 2A), whereas Bortezomib-inhibited CT-L activity was markedly restored at 24 hr (Figure 2B). T-L activity is significantly inhibited (50% inhibition) by NPI-0052 at 90 min, 24 hr, 48 hr, and 72 hr and is restored by 168 hr (Figure 2C); in contrast, Bortezomib enhances T-L activity, which remains elevated even at 168 hr (Figure 2D). Finally, NPI-0052 inhibits C-L activity at 90 min, 24 hr, 48 hr, and 72 hr, and this activity recovered at 168 hr (Figure 2E), whereas Bortezomib significantly inhibits C-L activity at 90 min, 24 hr, 48 hr, and 72 hr and is similarly recoverable at 168 hr (Figure 2F). Therefore, after NPI-0052 treatment, all three proteasome activities remain inhibited at 72 hr and were restored to significant levels by 168 hr. These data suggest that NPI-0052 and Bortezomib differentially affect all three proteasome activities.

NPI-0052 is orally bioactive

Previous studies showed that Bortezomib is orally active ([Palombella et al., 1998] and [Teicher et al., 1999]); however, the current Bortezomib therapy in MM is administered i.v. The rationale for using the i.v. route instead of oral administration may include bioavailability, rapid distribution (Adams et al., 1999), solubility, and reversibility/half-life issues. We therefore next examined whether NPI-0052 is also orally bioactive. Mice were treated with various oral and i.v. concentrations of NPI-0052; at 90 min whole-blood lysates (WBL) were analyzed for proteasome activity. Oral administration of NPI-0052 significantly inhibits CT-L activity of 20S proteasomes in a dose-dependent manner between 0.025 mg/kg and 0.50 mg/kg (Figure 3A). Both T-L and C-L activities were also inhibited, albeit to lesser extent (data not shown). These findings show that NPI-0052 given orally inhibits proteasome activity in vivo.

NPI-0052 targets nuclear factor-κB

A major rationale for using Bortezomib therapeutically is its ability to inhibit nuclear factor-κB (NF-κB) activation ([Adams, 2002], [Hideshima et al., 2002] and [Russo et al., 2001]). We therefore next asked whether NPI-0052 similarly affects NF-κB activation. To address this issue, a stable HEK-293 clone was generated carrying a luciferase reporter gene under the regulation of 5× NF-κB binding sites. Stimulation of these cells with human TNF-α leads to increased luciferase activity due to NF-κB activation. Pretreatment of NF-κB/luc HEK-293 cells with NPI-0052 results in a significant (p < 0.001) dose-dependent decrease of luciferase activity after TNF-α stimulation (Figure 3B), indicating inhibition of NF-κB. Bortezomib also downregulates NF-κB activity, but at higher concentrations than NPI-0052 (Figure 3B). These findings indicate that NPI-0052 is a more potent inhibitor of NF-κB than Bortezomib.

Activation of NF-κB triggers transcription and secretion of various proinflammatory cytokines, such as TNF-α, interleukin-1-β (IL-1-β), and interleukin-6 (IL-6), which mediate the growth and survival of tumor cells ([Adams, 2002], [Hideshima et al., 2002] and [Russo et al., 2001]). Inhibition of cytokine synthesis and function by means of proteasome inhibition thereby has clinical benefit. Since both NPI-0052 and Bortezomib inhibit NF-κB, we examined the effect of these agents on LPS-induced cytokine production in human peripheral blood mononuclear cells (PBMNCs). NPI-0052 inhibited LPS-triggered secretion of all three cytokines at lower concentrations than Bortezomib (Figure 3C). These findings show that NPI-0052, like Bortezomib, targets NF-κB and related cytokine secretion.

NPI-0052 blocks proteasome activity in MM cells

Our preclinical and clinical studies have already shown Bortezomib to be an effective therapy in MM ([Hideshima and Anderson, 2002] and [Richardson et al., 2005]). Our recent study showed the composition of active proteasome in MM.1S cell line and the proteasome targets of Bortezomib in these cells (Berkers et al., 2005). In that study, we utilized a novel methodology to measure proteasome activity by immunoblotting using DansylAhx₃L₃VS as a probe (Berkers et al., 2005). The results from this method correlated with those obtained using fluorogenic substrates and, in addition, allow for determining subunit specificity of a given proteasome inhibitor (details provided in the Experimental Procedures). In the present study, we examined the effects of NPI-0052 and Bortezomib on the catalytic activities of proteasome subunits. Cells were cultured in the presence or absence of various concentrations of either NPI-0052 (2 nM, low toxic dose; 7 nM, IC₅₀; and 20 nM, highly toxic to MM.1S cells), and compared these effects with those triggered by Bortezomib (2 nM, low toxic dose; 5 nM, IC₅₀; and 20 nM, highly toxic to MM.1S cells). Competition experiments between either NPI-0052 or Bortezomib and DansylAhx₃L₃VS revealed that NPI-0052 (7 nM) markedly inhibits the CT-L activity as represented by β-5 (β-5) subunit of the proteasome (Figure 3D). Furthermore, NPI-0052 also decreased the DansylAhx₃L₃VS labeling of the β-1 (C-L activity) and β-2 (T-L activity) in a dose-dependent manner. Slightly higher concentrations of Bortezomib are necessary to markedly inhibit β-5 and β-1 subunits, whereas β-2 subunits are not inhibited. Together, these findings demonstrate the ability of NPI-0052 to inhibit all three proteasome activities in MM cells and are consistent with in vitro results using fluorogenic substrates shown in Figure 1 and Figure 2.

NPI-0052 inhibits growth and triggers apoptosis in MM cells

We next asked whether NPI-0052-induced proteasome inhibition correlates with cytotoxicity in MM cells. Treatment for 24 hr of MM cell lines (MM.1S, RPMI-8226, OPM2, U266), including those that are resistant to the conventional anti-MM agents Dexamethasone (Dex) (MM.1R) and Doxorubicin (Dox-40), with NPI-0052 induces a dose-dependent significant (p < 0.005) decrease in viability of all cell lines (IC₅₀ range 7–24 nM) (Figure 4A). To determine whether NPI-0052 similarly affects purified patient cells, tumor cells from nine MM patients relapsing after multiple prior therapies including Dex, melphalan, Bortezomib, and Thalidomide were treated for 24 hr with NPI-0052 (10 nM) and then analyzed for apoptosis. NPI-0052 induced significant apoptosis in these cells, as measured by DNA fragmentation assays (p < 0.005) (Figure 4B). Importantly, of nine patients examined four were refractory to Bortezomib therapy, and five were resistant to Thalidomide and Dex therapies. Patients were deemed to be refractory to Bortezomib when they had progressive disease while on Bortezomib therapy. In addition, data from the phase II clinical studies in MM show that only 35% of patients with relapsed/refractory MM respond to the treatment with Bortezomib; and in this context, the 65% of patients who were nonresponders were considered to be a Bortezomib-resistant patient population. We next examined whether cells from such patient populations are affected by Bortezomib treatment in vitro and whether NPI-0052 exerts a cytotoxic effect in these cells. CD138-positive cells from seven MM patients were treated with Bortezomib (10 nM) and NPI-0052 (10 nM) in vitro and then analyzed for viability (Figure 4C). Results demonstrate a varying sensitivity to Bortezomib in vitro, with a 15%–50% decrease in viability; however, all patient cells were significantly more sensitive to NPI-0052 (IC₅₀ ≤ 10 nM in 6 of 7 patients) (Figure 4C).

The finding that cells from Bortezomib-refractory patients are sensitive to the treatment with Bortezomib in vitro suggests that extracellular factors, such as the BM microenvironment, play a role in conferring drug resistance in vivo. Absence of the BM milieu under the ex vivo culture conditions may restore sensitivity of MM cells to Bortezomib; however, the activation of intrinsic cellular drug mechanisms cannot be excluded. Additionally, as noted above, NPI-0052 and Bortezomib differentially affect proteasome activities, which may explain their variable in vitro cytotoxicity. Furthermore, it is unclear at present whether alteration in any proteasome subunit confers Bortezomib resistance versus sensitivity. Our data obtained from viability studies suggest that NPI-0052 is a more potent inducer of MM cell apoptosis than Bortezomib in tumor cells obtained from Bortezomib-refractory MM patients.

NPI-0052 does not affect viability of patient MM-derived bone marrow stromal cells

Interaction of MM cells with bone marrow stromal cells (BMSCs) induces cytokine secretion, which mediates paracrine growth of MM cells, as well as protects against drug-induced apoptosis (Hideshima and Anderson, 2002). We examined whether NPI-0052 affects viability of BMSCs. Treatment of BMSCs (patients 1–5) for 24 hr with NPI-0052 (20 nM) does not induce apoptosis in these cells, whereas NPI-0052 triggered a significant (10- to 12-fold) increase in apoptosis of purified patient MM cells (Figure 4D). Importantly, NPI-0052 significantly decreased the secretion of IL-6 triggered by adhesion of MM cells to BMSCs (IC₅₀, 80–100 nM; p < 0.05), as is observed using Bortezomib (Anderson, 2004). Together, these results suggest that NPI-0052 does not directly affect BMSC viability, but blocks the secretion of BMSC-derived MM cell growth factor IL-6 within the BM milieu.

NPI-0052 overcomes recombinant human interleukin-6- and recombinant human insulin-like growth factor-I-mediated antiapoptotic effects

Both IL-6 and IGF-I trigger growth and protect against chemotherapy-induced apoptosis in MM cells ([Chauhan and Anderson, 2003] and