Proteasome inhibitors in multiple myeloma: 10 years later

The ubiquitin-proteasome pathway is responsible for degradation of the majority of regulatory proteins in eukaryotic cells, including proteins that control cell-cycle progression, apoptosis, and DNA repair and therefore plays an essential role in maintaining normal cellular homeostasis.^1-3  The 26S proteasome consists of a barrel-shaped 20S proteolytic core, composed of 2 identical α-subunit rings and 2 identical β-subunit rings, plus 2 19S regulatory complexes that cap the 20S barrel.⁴  Proteins destined for degradation are first polyubiquitinated; the 19S cap recognizes and binds ubiquitinated proteins and directs them to the 20S core, where proteolytic cleavage is mediated by 3 β-subunits: β1 (caspase-like activity), β2 (trypsin-like activity), and β5 (chymotrypsin-like activity).^1-3  Disruption of proteasome activity results in growth arrest and cell death because of induction of an apoptotic cascade as a result of the rapid accumulation of incompatible regulatory proteins within the cell.⁵ 

Cancer cells generally have higher levels of proteasome activity compared with normal cells and, moreover, are more sensitive to the proapoptotic effects of proteasome inhibition than normal cells,⁶  making the proteasome a rational therapeutic target in oncology.⁷  Based on promising preclinical results, proteasome inhibition has been extensively explored as a therapeutic strategy in multiple myeloma (MM), and proteasome inhibitors (PIs) now form a cornerstone of antimyeloma therapy.

Bortezomib was the first-in-class PI to be introduced into the clinic. Recently, a number of second-generation PIs have been developed and are undergoing intense examination in clinical trials. The different PIs are distinct regarding their specificities and affinities for the different catalytic sites within the proteasome core (Table 1).^8-10  Based on chemical structure and active moiety, PIs can be classified into 3 groups: boronates, epoxyketones, and salinosporamides.

Bortezomib is a dipeptidyl boronic acid-based specific, reversible PI that targets the chymotrypsin- and caspase-like active sites, with minimal effect on trypsin-like activity.⁸  By inhibiting the proteasome, bortezomib acts through multiple mechanisms to suppress tumor survival pathways and to arrest tumor growth, tumor spread, and angiogenesis.¹¹  The mechanisms of its antitumor activity in MM have been elucidated through in vitro and in vivo experiments. Bortezomib directly induces apoptosis of tumor cells, inhibits the activation of NF-κB in cells and in the tumor microenvironment, reduces adherence of myeloma cells to bone marrow stromal cells, blocks production and intracellular signaling of IL-6 in myeloma cells, stops the production and expression of proangiogenic mediators, and overcomes defects in apoptotic regulators, such as Bcl-2 overexpression and alterations in tumor suppressor p53.^11,12  In addition, proteasome inhibition with bortezomib has been shown to induce the endoplasmic reticulum stress response, associated with disruption of the unfolded protein response, an important aspect of the mechanism of action of proteasome inhibition in MM because of the high production of immunoglobulins by MM cells.¹³  These mechanisms of action of proteasome inhibition have provided the rationale for the combination of bortezomib and other PIs with numerous chemotherapeutic and targeted agents^12-14 ; for example, the disruption of protein quality control and the consequent activation of aggresomal degradation of misfolded proteins have provided a strong rationale for the combination of PIs and histone deacetylase inhibitors.¹³  As a result of its mechanisms of action, bortezomib has also been associated with increased bone formation and osteoblastic activity, and decreased bone resorption and osteoclastic activity.¹⁵ 

MLN9708 is another boronate PI that is a reversible inhibitor of primarily the chymotrypsin-like activity of the 20S proteasome. However, in contrast to bortezomib, MLN9708 has a shorter dissociation half-life (Table 1) and has demonstrated greater tissue penetration compared with bortezomib in preclinical studies.¹⁰  Furthermore, MLN9708 is orally available and is the first oral PI to enter clinical trials in MM. Like bortezomib, MLN9708 has been shown to inhibit NF-κB activation and demonstrated antitumor activity in MM and other hematologic malignancies in in vitro and in vivo studies.¹⁰  A third boronate PI, CE P-18770,¹⁶  was also in development, but data from clinical trials have not been reported.

Carfilzomib (PR-171) is an irreversible PI that belongs to the epoxyketone class and is structurally and mechanistically distinct from bortezomib.¹⁷  It demonstrates potent and sustained inhibition of chymotrypsin-like activity and appears to have a greater selectivity for the chymotrypsin-like protease compared with bortezomib, with lower affinity for the trypsin- and caspase-like proteases.⁹  Unlike the boronate PIs, carfilzomib has minimal activity against off-target enzymes, including serine proteases.¹⁸  Like bortezomib, carfilzomib inhibits both the constitutive and immunoproteasome, and carfilzomib has equivalent potency against the β5 and LMP7 subunits (Table 1).¹⁹ 

Carfilzomib has been shown to trigger cell cycle arrest, induce apoptosis, and activate stress response pathways in human tumor cell lines, including MM and other hematologic malignancies and solid tumors.¹⁷  Importantly, carfilzomib has demonstrated activity against bortezomib-resistant cell lines and primary MM cells.¹⁹  The carfilzomib analog ONX 0912 is another peptide epoxyketone PI that is in development²⁰ ; however, no data from clinical studies of this orally available agent have yet been reported. In addition to antimyeloma effects, these epoxyketone PIs have also been shown to inhibit bone resorption in preclinical models.²¹ 

Marizomib (NPI-0052) is a natural lactone compound derived from the marine bacterium Salinospora tropica. This agent belongs to a unique class of PIs, the salinosporamides. Marizomib is an irreversible PI that, unlike bortezomib and carfilzomib, inhibits both the chymotrypsin-like and trypsin-like protease activities, but only minimally affects the caspase-like activity within the proteasome. As a result, marizomib has a unique efficacy and safety profile and does not exhibit cross-resistance with other PIs. Marizomib-induced apoptosis is predominantly the result of activation of caspase-8–mediated signaling pathways.²²  Marizomib has demonstrated antitumor activity in preclinical models of MM, other hematologic malignancies, and solid tumors.²³ 

The efficacy and safety of bortezomib were established through a number of important phase 1 to 3 studies. The earliest phase 1 studies, which were reported 10 years ago, demonstrated manageable toxicity with promising antineoplastic activity, particularly in patients with relapsed/refractory MM.^24-26  When administered intravenously (IV) on a twice-weekly schedule, the maximum tolerated dose (MTD) was determined to be 1.3 mg/m², based on dose-limiting adverse events (AEs) of diarrhea and peripheral neuropathy (PN).²⁴  Notable in these phase 1 studies was the relative lack of significant myelosuppression; in addition, cardiac, hepatic, or renal toxicities were infrequent.

After these early studies, the efficacy of bortezomib in MM was confirmed in the large phase 2 SUMMIT study (Table 2), which investigated bortezomib in relapsed/refractory MM patients.²⁷  Overall, 27% of patients achieved partial response (PR) or better, including 10% complete/near-complete responses (CR/nCR). Median time to progression (TTP) was 7 months, compared with 3 months with patients' previous therapy. Another phase 2 study (CREST) prospectively compared 2 doses of bortezomib in patients with relapsed or refractory MM.²⁸  Of note, the survival, response, and TTP data suggested that a starting dose of 1.3 mg/m² was preferable; however, the trial also showed that the dose of bortezomib could be reduced to 1.0 mg/m² if required while still offering patients a substantial survival benefit, thereby providing valuable information regarding the use of this agent in clinical practice.²⁹  On the basis of these phase 2 results, bortezomib was approved for the treatment of relapsed and refractory MM and moved into phase 3 investigation. The international phase 3 APEX trial demonstrated the superiority of bortezomib over high-dose dexamethasone,³⁰  which at that time could be considered a standard of care in the relapsed setting. Bortezomib demonstrated superior response rates, TTP (the primary endpoint), and overall survival (OS; Table 2),³⁰  and, in an updated analysis with a median follow-up of 22 months, median OS was 29.8 months versus 23.7 months, despite more than 62% of dexamethasone patients crossing over to receive bortezomib.³¹  Based on these results, bortezomib soon became one of the treatments of choice for relapsed MM.

These initial studies of single-agent bortezomib also established the common bortezomib-associated toxicities, the most frequent being gastrointestinal symptoms, anemia, thrombocytopenia, fatigue, and PN. As treatment-related PN was noted in these key studies, substantial efforts were made to characterize this toxicity. It is thought that bortezomib mainly causes direct dorsal root ganglion toxicity, based on studies in mice, which showed that ubiquitinated aggregates accumulated in the cytoplasm of the dorsal root ganglion.^32-34  There is now substantial experience regarding bortezomib-induced PN, and a number of excellent reviews provide a comprehensive overview of its etiology, clinical characteristics, and management.^35-38  Bortezomib-induced PN can be effectively managed with dose modification and is generally reversible in more than 50% of cases, with weekly dosing^39-41  and subcutaneous administration of bortezomib providing approaches to limit treatment-emergent PN, as described in more detail later.⁴²  The most common hematologic toxicity associated with bortezomib is transient thrombocytopenia,⁴³  which shows a cyclical pattern of platelet decrease and recovery, without evidence of cumulative thrombocytopenia.^43,44  The transient, predictable decrease in platelet counts generally results in a low requirement for platelet support, and so bortezomib can usually be administered to thrombocytopenic patients.⁴⁴ 

As an alternative to the previous standard of IV administration, bortezomib has been infused subcutaneously, and this route of administration has recently been approved by the United States Food and Drug Administration. A randomized phase 1 trial of subcutaneous (SC) versus IV bortezomib in relapsed/refractory MM showed similar systemic bortezomib exposure, 20S proteasome inhibition, response rates, and safety with SC and IV bortezomib.⁴⁵  A large phase 3 study has confirmed these preliminary results (Table 2),⁴²  demonstrating the noninferiority of SC versus IV bortezomib in terms of overall response rate (ORR) after 4 cycles. After median follow-up of 12 months, there were no significant differences in TTP and 1-year OS with SC versus IV bortezomib. PN of any grade (38% vs 53%; P = .04), grade more than or equal to 2 (24% vs 41%; P = .01), and grade more than or equal to 3 (6% vs 16%; P = 0·03) was significantly less common with SC than IV administration. These results have validated a more convenient and less toxic route of administration that is likely to become standard in the near future.

Given the finite treatment course of bortezomib, as used in the phase 2 and 3 studies described earlier, patients may remain sensitive to bortezomib after relapse after initial therapy. Studies have specifically addressed the issue of bortezomib retreatment in relapsed MM, confirming that it is feasible, without evidence of cumulative toxicity. In a retrospective multicenter survey of 94 relapsed MM patients who had responded to initial bortezomib treatment, no uncommon toxicity with bortezomib retreatment was identified, and efficacy data showed that 63% of patients responded to retreatment, with a median TTP of 9.3 months, consistent with sustained sensitivity to bortezomib.⁴⁶  A similar prospective phase 2 trial demonstrated an ORR of 32% and 42% in patients who received single-agent bortezomib and bortezomib-dexamethasone, respectively, as retreatment after initial response to bortezomib.⁴⁷ 

Subsequent to the demonstration of single-agent activity, studies showed that the efficacy of bortezomib was enhanced through combination with agents with different modes of action. Based on observations from in vitro studies of synergistic activity between bortezomib and anthracyclines,⁴⁸  and on phase 1 data,⁴⁹  a phase 3 study compared single-agent bortezomib with bortezomib plus pegylated liposomal doxorubicin (PLD) in patients with relapsed/refractory MM (Table 2).⁵⁰  The combination was superior, with median TTP increased from 6.5 to 9.3 months with bortezomib-PLD versus bortezomib, albeit at the cost of increased grade 3 or 4 toxicity. These results led to the approval of the combination for relapsed MM in 2007. In vitro studies also showed that dexamethasone and bortezomib act synergistically,⁵¹  and clinical observations showed that the addition of dexamethasone improved responses in up to 34% of patients after suboptimal response to single-agent bortezomib.^42,52,53  Although no randomized trial comparing bortezomib versus bortezomib-dexamethasone has been performed, it is now common practice to add dexamethasone, with or without other agents, to bortezomib when treating relapsed MM.

Apart from the bortezomib-PLD trial, few data on bortezomib-based combinations in the relapsed setting are available from randomized phase 3 trials. Recently, the Intergroupe Francophone du Myélome (IFM) and European Group for Blood and Marrow Transplantation conducted a prospective comparison of bortezomib-thalidomide-dexamethasone (VTD) versus thalidomide-dexamethasone (TD) in patients relapsing after high-dose therapy plus autologous stem cell transplantation (HDT-ASCT) and showed that median TTP was significantly longer with VTD than TD, with a trend for a survival benefit (Table 2).⁵⁴  This study showed, for the first time, that a triplet bortezomib-based combination was superior to a 2-drug thalidomide-based regimen for relapsed MM, in terms of significant greater response rates and longer TTP. Less striking, the VANTAGE 088 phase 3 trial showed only a weak clinical benefit with bortezomib plus the histone deacetylase-inhibitor vorinostat versus single-agent bortezomib⁵⁵ ; median progression-free survival (PFS) benefit was just 0.8 months, indicating that alternative combinations should be tested. Currently, other novel agents, such as perifosine or panobinostat, are being evaluated in combination with bortezomib in phase 3 trials.

Multiple other bortezomib-based combinations have been studied in phase 1 or 2 trials, incorporating steroids plus alkylators, immunomodulatory drugs (IMiDs), monoclonal antibodies, and heat shock protein-90, histone deacetylase, or AKT inhibitors.^14,56  Widely used regimens include bortezomib plus cyclophosphamide and dexamethasone (VCD) or prednisone (VCP), or bortezomib plus lenalidomide and dexamethasone (VRD). Response rates range from 60% to 90%, with median PFS of approximately 7 to 15 months and median OS of 16 to 37 months, and bortezomib can be delivered at or near the standard dose and schedule with few overlapping toxicities.

For many years, the standard of care in elderly patients was melphalan-prednisone (MP). Preclinical studies demonstrated in vitro synergy with bortezomib combined with melphalan,⁴⁸  and consequently bortezomib was studied in combination with MP (VMP). A phase 1 or 2 trial of VMP demonstrated an ORR of 89%, including 32% CR,⁵⁷  median TTP of 27 months, and OS at 38 months of 85%.⁵⁸  These impressive results were the basis for the randomized phase 3 VISTA trial that compared MP versus VMP in previously untreated MM patients ineligible for HDT-ASCT (Table 3).⁵⁹  At the first report, median TTP (primary endpoint) was 24.0 versus 16.6 months with VMP versus MP (P < .001),⁵⁹  the ORR was 71% versus 35%, and CR rates were 30% and 4%, respectively (P < .001). Importantly, the hazard ratio for OS was 0.61 in favor of VMP, with a median follow-up of 16.3 months.⁵⁹  These results led to the approval of VMP as induction therapy in patients ineligible for ASCT in the United States and European Union, and this regimen is now considered one of the standards of care in this population. Updated results confirming a survival advantage for VMP have been reported,^60,61  with 5-year follow-up data demonstrating a continued statistically significant OS benefit with VMP versus MP despite a higher proportion of MP patients receiving subsequent bortezomib (43% vs 22%).⁶¹ 

In both these VMP studies,^57,59  important toxic effects were recorded, particularly PN and gastrointestinal symptoms. Consequently, the Spanish group designed a novel, less intensive bortezomib-based regimen to maintain efficacy while reducing toxic effects (Table 3).⁴⁰  Patients received induction therapy based mainly on weekly rather than twice-weekly dosing of bortezomib, with prednisone plus melphalan (VMP) or thalidomide (VTP). This was followed by a second randomization to maintenance with bortezomib-thalidomide (VT) or bortezomib-prednisone (VP). There was no difference in response rates after induction between VMP and VTP, but the grade 3 or 4 PN rate was only 7% with VMP arm (half that in VISTA). After maintenance, the CR rate was approximately 40% in both arms, and only 3% of VP and 9% of VT patients developed grade 3 PN.⁶²  Overall median PFS and TTP were 31 and 35 months, respectively, and the 3-year OS was 70%, which compared favorably with VISTA results. Therefore, the goal of reducing toxicity without impairing outcomes was achieved.

Another randomized phase 3 trial has provided data on weekly administration of bortezomib and the potential of bortezomib-based maintenance. The Italian group examined VMP-thalidomide (VMPT) followed by VT maintenance (VMPT-VT) compared with VMP alone in transplant-ineligible patients (Table 3).^39,41  comparing the 2 arms of the study, the efficacy results (including response rates and PFS) were in favor of VMPT-VT. After inclusion of the first 139 patients, the protocol was amended with the aim of reducing the incidence of PN, and both VMPT-VT and VMP induction schedules were changed from the planned VISTA-like twice-weekly schedule to compose nine 5-week cycles using weekly bortezomib dosing. Remarkably, there was a substantial reduction in the incidence of nonhematologic severe AEs in the weekly versus twice-weekly group, whereas long-term outcomes appeared similar (Table 3).³⁹  Notably, the incidence of any grade 3 or 4 PN was 8% versus 28% with weekly versus twice-weekly bortezomib (P < .001). The authors concluded that the weekly schedule resulted in a substantial improvement in safety and did not impact efficacy.

These 2 phase 3 trials are of major importance because they validate the weekly VMP schedule in elderly patients, demonstrating that this modified schedule is able to significantly decrease toxicity, especially PN, without adversely affecting the overall outcome. The studies, together with the phase 3b community-based UPFRONT study (Table 3), have also introduced the concept of bortezomib-based maintenance in elderly patients. To date, data regarding a role for bortezomib maintenance in this setting remain inconclusive because the Spanish trial proposed maintenance in both arms and the Italian trial used different induction therapy in the 2 arms. Data from a “pure” maintenance trial comparing bortezomib versus no bortezomib after a common induction therapy are currently not available.

The activity of bortezomib in the relapsed setting prompted its evaluation upfront as part of induction before HDT-ASCT. Several phase 2 trials investigated bortezomib-dexamethasone induction, with response rates of 66% to 90%, including 15% to 21% CR and 31% to 70% very good PR (VGPR) or better.^63-65  This translated into high CR and VGPR rates after transplantation. Toxicities were generally mild to moderate and proved manageable, without treatment-related mortality, and stem cell collection was adequate. Therefore, this promising combination was compared prospectively with vincristine-doxorubicin-dexamethasone (VAD), which was considered the standard induction regimen.

The IFM2005-01 phase 3 study compared bortezomib-dexamethasone with VAD as pre-ASCT induction in previously untreated patients (Table 3).⁶⁶  Postinduction response rates were significantly higher with bortezomib-dexamethasone, regardless of disease stage or adverse cytogenetic abnormalities. After first transplantation, CR/nCR and more than or equal to VGPR rates remained significantly higher with bortezomib-dexamethasone (Table 3). There was a trend for improved PFS with bortezomib-dexamethasone versus VAD (P = .06), but 3-year survival rates were not different. It should be noted that more than half the patients in each group subsequently received lenalidomide consolidation and were randomized to lenalidomide or placebo maintenance as part of another IFM study. The incidence of severe AEs appeared similar between groups, but hematologic toxicity and treatment-related mortality were more frequent with VAD. Conversely, rates of grade 3 or 4 PN during induction through first transplantation were significantly higher with bortezomib-dexamethasone than VAD (9.2% vs 2.5%). The level of response achieved with bortezomib-dexamethasone after induction in IFM2005-01 is now considered the goal for current therapies; as a result, bortezomib-dexamethasone has become the backbone of pre-ASCT induction therapy to which other, more complex regimens should be compared.

The addition of a third agent to bortezomib-dexamethasone, including thalidomide (VTD),⁶⁷  (liposomal) doxorubicin (VDD/PAD),^68,69  lenalidomide (VRD),⁷⁰  and cyclophosphamide (VCD),⁷¹  has been tested in several phase 2 studies, and outcomes appear even better than with the doublet. Three prospective studies have already shown that VTD is superior to TD or bortezomib-dexamethasone in terms of response rates (Table 3).^72-74  The Italian group compared TD versus VTD as induction before, and consolidation after, tandem ASCT and found that VTD resulted in higher CR and more than or equal to VGPR rates after induction and after ASCT, which translated into better PFS (Table 3).⁷²  The Spanish group also compared TD versus VTD versus a more complex chemotherapy regimen, including bortezomib, before ASCT, and confirmed that VTD achieved the best pre- and post-ASCT CR rates.⁷⁴  In the IFM2007-02 trial, 4 cycles of “standard” bortezomib-dexamethasone induction were compared with 4 cycles of VTD, using lower bortezomib and thalidomide doses to reduce the neuropathy rate.⁷³  VTD was again found to result in superior response rates before and after ASCT, and the reduced bortezomib and thalidomide doses were associated with reduced neurotoxicity. The IFM2007-02 study therefore provided further evidence for the superiority of a 3-drug over a 2-drug combination for tumor burden reduction as pre-ASCT induction. Further, the HOVON group reported a phase 3 randomized trial comparing VAD versus PAD as pre-ASCT induction, followed by thalidomide and bortezomib maintenance, respectively.⁷⁵  This study confirmed the superiority of the bortezomib-based triplet over VAD in terms of postinduction and post-ASCT response rates; and, notably, OS was also superior after PAD plus bortezomib maintenance.

No data are available to draw conclusions regarding the superiority of one bortezomib-based combination over another. Although response rates are clearly improved with novel agent-based regimens, demonstration of a significant OS advantage will often be difficult given the large numbers of patients and the long duration of follow-up required, and the availability of effective salvage therapies. Thus, based on response rates and depth of response as surrogate markers for outcome, 3-drug combinations are, in 2012, the standard of care before ASCT.⁷⁶ 

As well as being studied in induction regimens, bortezomib has been investigated for use elsewhere within the transplantation treatment paradigm, broadening the utility of PIs in MM. For example, based on the observed synergy with melphalan,⁴⁸  bortezomib has been examined as part of conditioning regimens. In a phase 2 IFM study of bortezomib plus melphalan 200 mg/m² (MEL200; the standard of care) as conditioning (Bor-HDM),⁷⁷  70% of patients achieved at least VGPR, including 32% CR after ASCT. Bortezomib did not increase hematologic toxicity, and only 1 case of grade 3 or 4 PN was reported. A matched control analysis with patients from the IFM2005-01 trial (MEL200 alone) showed the CR rate to be higher in the Bor-HDM group, regardless of induction therapy.⁷⁷  The results suggest that Bor-HDM is a safe and promising conditioning regimen. These findings were confirmed in a similar United States phase 1 or 2 trial involving 39 patients.⁷⁸  Nevertheless, randomized studies are needed to assess whether bortezomib-containing conditioning is superior to MEL200 alone.

Currently, novel agents, including bortezomib, are also being tested after ASCT, with the objective of further improving response rate and quality. Single-agent bortezomib consolidation after ASCT has been investigated by the Nordic group,⁷⁹  in a phase 3 trial in which 370 patients were randomized to receive no treatment or bortezomib. Preliminary results indicated that bortezomib consolidation was feasible; toxicity was low, with 5% grade 3 or 4 PN. The 6-month post-randomization CR/nCR rate was 35% versus 45% with no treatment versus bortezomib (P < .05). This translated into an improvement in median PFS, from 20 to 27 months (P = .04).

Bortezomib has also been combined with IMiDs and dexamethasone after ASCT. In the most striking report, Ladetto et al treated 39 patients who achieved at least VGPR after ASCT with four 28-day cycles of VTD.⁸⁰  Response was assessed by qualitative nested PCR and quantitative RT-PCR using tumor clone-specific primers. The CR rate increased from 15% after ASCT to 49% after VTD consolidation, and molecular remissions increased from 3% to 18%. With a median follow-up of 65 months after consolidation, only 2 patients in molecular remission had relapsed, and no patient had died from progressive disease.⁸¹  This study is the first to document persistent molecular remissions in MM patients receiving ASCT followed by novel agent-based consolidation therapy. Until now, such impressive results had only been reported in the context of myeloablative allogeneic stem cell transplantation. Results from another prospective randomized Italian trial incorporating TD or VTD as induction and as consolidation after tandem ASCT recently confirmed these findings.⁸²  The probability of upgrading from less than CR to CR after consolidation was significantly higher with VTD versus TD, and PFS was also significantly longer. These results support the use of consolidation, including bortezomib-based combinations; however, it should be noted that further response improvements may be seen in some patients after ASCT in the absence of any additional therapy. Thus, additional data from ongoing randomized trials comparing PI/novel agent-based consolidation versus no consolidation have to be awaited before this procedure can be considered a standard of care.

Bortezomib has also been investigated as maintenance post-ASCT in a large phase 3 trial of VAD versus PAD induction, with up to 2 years' maintenance using thalidomide on the VAD arm, or bortezomib on the PAD arm.⁷⁵  During maintenance with bortezomib, 38% of patients upgraded their post-ASCT response, and a landmark analysis showed a trend for longer duration of response in the maintenance phase using bortezomib versus thalidomide. Another phase 2 study showed upgraded responses with prolonged weekly bortezomib-dexamethasone followed by TD as maintenance therapy after single ASCT,⁸³  with few toxicities. In addition, results of a phase 3 Spanish myeloma group trial suggest a benefit with bortezomib-based maintenance. After induction with VTD, TD, or complex chemotherapy plus bortezomib, patients were randomized after ASCT to 3 years' maintenance with VT, thalidomide, or IFN-α.⁸⁴  With a median follow-up of 24 months, 2-year PFS was significantly greater with VT maintenance. Grade 1 to 3 PN was observed in 12% and 10% of patients receiving VT and thalidomide maintenance, respectively. Overall, the results indicate that bortezomib-based maintenance may upgrade responses and prolong PFS, but data from a prospective study comparing PI-based maintenance versus no maintenance are currently lacking.

Our understanding of the ubiquitin-proteasome pathway has increased considerably and, with it, the recognition of the important role that proteasome inhibition plays as a highly effective therapeutic strategy in MM. This has stimulated the development of new PIs with distinct efficacy and safety profiles.⁸  These second-generation PIs also provide potential new options for patients whose disease has become resistant to bortezomib.

Of the second-generation PIs, carfilzomib has progressed furthest in clinical development. Clinical studies have demonstrated durable antitumor activity in patients with relapsed/refractory MM⁹  and, notably, limited neurotoxicity (Table 4). In a large multicenter phase 2 study (PX-171-004) in patients who had received 1 to 3 prior regimens,⁸⁵  2 carfilzomib dosing regimens were investigated in bortezomib-naive patients and those previously treated with bortezomib. Patients received either carfilzomib 20 mg/m² for twelve 28-day cycles, or 20 mg/m² in cycle 1 and 27 mg/m² in cycles 2 to 12. In 129 bortezomib-naive patients, the ORR was 48% and was better in patients who received the 20/27- versus 20-mg/m² regimen.⁸⁵  In the 20/27-mg/m² group, responses included 2% CR and 27% VGPR. In the 20-mg/m² group, median TTP was 8.3 months. The most common AEs were fatigue and hematologic toxicity. The risk of PN was low with both regimens despite the fact that approximately 50% of patients had baseline neuropathy. In patients previously treated with bortezomib,⁸⁶  carfilzomib 20 mg/m² yielded 1 CR, 1 VGPR, and 4 PRs. Although the response rate was fairly low, median duration of response was 9 months and median TTP was 5.3 months.

An integrated safety analysis of 526 patients with relapsed/refractory MM who were treated in 3 phase 2 studies of carfilzomib 20/27 mg/m² was recently reported⁸⁷  and showed that the most common grade more than or equal to 3 AEs were thrombocytopenia (23%), anemia (22%), lymphopenia (18%), pneumonia (11%), and neutropenia (10%). PN was reported infrequently (14% overall) and was generally mild to moderate in severity. Although 72% of patients had grade more than or equal to 2 PN at study entry, only 13% reported treatment-emergent symptoms. Thus, the safety profile of carfilzomib is quite different from that of bortezomib, which is associated with a high risk of PN, albeit that SC administration of bortezomib is associated with a significantly lower risk of PN.⁴²  Preliminary results of another large multicenter phase 2 study of carfilzomib 20 mg/m² in relapsed/refractory MM (PX-171-003-A1) have recently been reported.^88,89  The ORR was 24%, and patients with unfavorable cytogenetics had a 28% ORR compared with 24% in patients with normal or favorable cytogenetics.

Carfilzomib has also been investigated in combination with lenalidomide and low-dose dexamethasone in patients with relapsed/refractory MM.⁹⁰  A phase 1b study combined carfilzomib 15 to 27 mg/m² with lenalidomide 10 to 25 mg plus weekly dexamethasone, yielding an ORR of 55%; the most common grade more than or equal to 3 toxicities were hematologic (thrombocytopenia, anemia, and neutropenia). Furthermore, 2 large, randomized, phase 3 trials are ongoing in patients with relapsed/refractory MM. The ASPIRE trial (N = 700) is comparing carfilzomib plus lenalidomide-dexamethasone with lenalidomide-dexamethasone alone in the relapsed setting (primary endpoint: PFS). The FOCUS trial (N = 302) is comparing carfilzomib monotherapy with best supportive care in the relapsed/refractory setting (primary end point: OS). In addition, early-phase studies are investigating carfilzomib in frontline MM in combination with either thalidomide (CARTHADEX)⁹¹  or lenalidomide (CRd)⁹²  and dexamethasone (Table 4).

Efficacy and safety data for marizomib are available from phase 1 trials. Results from 2 parallel, phase 1, dose-escalation studies conducted in Australia and the United States in patients with relapsed/refractory MM were recently reported together (Table 5).⁹³  Among 34 patients, 88% had been previously treated with bortezomib, and 71% were bortezomib-refractory. The MTD of marizomib was 0.4 mg/m² over a 60-minute infusion or 0.5 mg/m² over a 120-minute infusion. Dose-limiting toxicities included transient hallucinations, cognitive changes, and loss of balance, which were reversible. The most common drug-related AEs included fatigue, gastrointestinal AEs, dizziness, and headache. There was no evidence of PN or thrombocytopenia. Of 15 patients treated in the active dose range (0.4-0.6 mg/m²), 3 demonstrated a PR, all of whom were bortezomib-refractory. These early data suggest that marizomib has a safety profile that is not overlapping with that of other PIs and is active in bortezomib-refractory patients. A twice-weekly regimen of marizomib 0.5 mg/m² in combination with low-dose dexamethasone is being investigated further.

To date, phase 1 studies of MLN9708 have investigated the safety, tolerability, and preliminary antimyeloma activity of both oral and IV dosing in patients with relapsed/refractory MM and other malignancies (Table 5).^94-97  Preliminary data indicate that MLN9708 has promising activity and produces durable responses in heavily pretreated patients. A phase 1 dose-escalation study of twice-weekly oral MLN9708 determined the MTD to be 2.0 mg/m².⁹⁷  The study enrolled 56 patients, who had received a median of 4 prior therapies; all patients had received an IMiD, nearly all had prior bortezomib, and approximately 20% had either prior carfilzomib or marizomib. The most common grade more than or equal to 3 AEs were thrombocytopenia (34%), neutropenia (14%), fatigue (9%), and rash (9%). Only 11% of patients developed PN, which was grade 1 or 2. Among response-evaluable patients, the ORR was 13% (1 CR, 5 PR), and responses were durable for up to 16 months. A phase 1 dose-escalation study of weekly oral dosing has also been reported,⁹⁶  including 28 patients with relapsed/refractory MM who had received a median of 5 prior regimens. No dose-limiting toxicity had been reported at doses up to 3.95 mg/m²; thus, the MTD had not been reached. Similar to twice-weekly dosing, the most common AEs were fatigue and thrombocytopenia. Among 16 response-evaluable patients, 1 patient had a PR and 5 patients had stable disease for up to 10 months. These data suggest that weekly administration of this novel oral PI is well tolerated and has antitumor activity in heavily pretreated relapsed/refractory MM. In addition, as with bortezomib and carfilzomib, MLN9708 is being investigated in a phase 1 or 2 study in frontline MM in combination with lenalidomide and dexamethasone (Table 5).⁹⁵ 

In conclusion, over the past decade, we have moved from the era of proteasome inhibition as a novel therapeutic approach to one in which bortezomib has contributed greatly to improvements in the prognosis of MM. In the past 10 years, clinicians have substantially increased their knowledge of the effects of proteasome inhibition and learned how to use bortezomib highly effectively, including appropriate monitoring for known toxicities, extending its use across multiple areas of the treatment paradigm and successfully introducing a new administrative route. Initially used as single agent in the relapsed setting, this drug has become the backbone of frontline therapy in young and elderly patients, and ongoing studies are exploring new indications, such as consolidation and maintenance. In addition, increasing clinical experience in the use of bortezomib has resulted in the introduction of a new dosing regimen, the weekly schedule, which has demonstrated reduced toxicity while maintaining efficacy. Furthermore, the SC route of administration is now available, which is less neurotoxic, with efficacy comparable with that seen with IV bortezomib. Several second-generation PIs are now in clinical development; and thanks to the understanding we have gained through the development of bortezomib, it is probable that we will be able to move forward more quickly with developing combination regimens incorporating these novel PIs. Results of ongoing trials will prove important in determining their place in therapy.

The authors thank Steve Hill of FireKite for editorial assistance in the development of this manuscript.

This work was supported by the University Hospital Hotel-Dieu, Nantes, France.

Contribution: P.M. wrote the manuscript; and P.G.R., M.C., R.Z.O., J.F.S.M., A.P., and J.-L.H. corrected the paper and approved the final manuscript.

Conflict-of-interest disclosure: P.M., M.C., J.-F.S.M., and A.P. are on the advisory board of Millenium, Janssen, and Onyx; P.G.R. is on the advisory board of Millenium and Nereus; and R.Z.O. is on the advisory board of Millenium and Onyx. J.L.H. declares no competing financial interests.

Correspondence: Philippe Moreau, Hematology Department, University Hospital Hotel-Dieu, 44093 Nantes cedex 01, France; e-mail: philippe.moreau@chu-nantes.fr.