The future of treatment for patients with relapsed/refractory multiple myeloma

Abstract

The ubiquitin-proteasome pathway was first validated as a target for cancer therapy with the demonstration of the activity of the boronic acid proteasome inhibitor bortezomib against relapsed and relapsed/refractory multiple myeloma. Another generation of proteasome inhibitors is now entering the clinical arena, including intravenous agents such as carfilzomib, CEP-18770, and marizomib, and oral drugs such as MLN9708 and ONX 0912. These novel agents will likely first be used for patients with disease that has either relapsed or been refractory to their prior therapy, including bortezomib-based regimens, through their ability to overcome drug resistance, or in patients who are intolerant of, or are not candidates for bortezomib. Preclinical studies also suggest that there is potential for proteasome inhibitors to act synergistically with other conventional and novel agents, and even with one another in rationally designed combination regimens. In addition, other inhibitors that selectively target only the immunoproteasome and not the constitutive proteasome, as well as drugs that bind to non-catalytic proteasome subunits, are emerging as potential drug candidates. Taken together, it seems likely that we have only begun to appreciate the full potential of inhibition of the proteasome, and this review will attempt to extrapolate our current knowledge in this area into a future algorithm for use of these inhibitors against multiple myeloma.

Introduction

Bortezomib (VELCADE®; PS-341), the first-in-class inhibitor of the proteasome [1], or multicatalytic proteinase complex [2], was originally found to be active against relapsed and relapsed/refractory multiple myeloma as a single agent in phase I through III clinical trials [3-6]. Its benefits included a robust overall (ORR) and complete response (CR) rate, strong response durability with improved time to progression (TTP) compared to prior therapies, and superior overall survival (OS) compared to dexamethasone, even in clinically and cytogenetically high-risk patients, as well as a predictable and manageable toxicity profile. Combinations based on bortezomib with other conventional and novel drugs extended the benefits for this patient population further, and a number of these regimens are now part of the standard of care [7], especially bortezomib with pegylated liposomal doxorubicin (DOXIL®), which is supported by phase III data [8]. Agents with activity against multiple myeloma in the relapsed and relapsed/refractory setting are often then incorporated into front-line therapy, and this has certainly been true for bortezomib. Indeed, in combination with either dexamethasone [9] or thalidomide and dexamethasone [10], and with melphalan and prednisone [11], bortezomib is now part of the standard of care for induction therapy in transplant-eligible and –ineligible patients, respectively. Moreover, its activity is being explored in many other combinations, such as with lenalidomide and dexamethasone [12, 13], and in other settings, such as with melphalan as a pre-transplant conditioning regimen [14-17], after transplant as part of consolidation [10, 18], and as a component of maintenance approaches [19-26]. While the increasing use of bortezomib in these early disease settings has been associated with substantial clinical benefits, it does raise the concern that further therapy targeting the proteasome in later lines may be less effective. Fortunately, retreatment with bortezomib in patients who have had a prior response to therapy, either in the front-line or relapsed and/or refractory setting, is one option that has been validated in a number of studies [27-32]. Even more encouraging is the development of emerging novel proteasome inhibitors for the relapsed and/or refractory setting that seem destined to play an important role in our armamentarium against myeloma, whose future will be examined in this contribution.

Next Generation Proteasome Inhibitors

Pre-clinical studies have validated a number of next-generation proteasome inhibitors in models of multiple myeloma that have already been translated to the clinic. These include agents that bind reversibly, such as the peptide boronates MLN9708 [33, 34] and CEP-18770 [35-37]. Other proteasome inhibitors under study, in contrast, bind irreversibly, including the epoxyketones ONX 0912 [38] and carfilzomib [39, 40], which was formerly known as PR-171, as well as the salinosporamide marizomib [41-43], which was previously known as NPI-0052. All of these drugs target at least the β5 catalytic subunit of the constitutively expressed 26S proteasome, as well as the corresponding β5_i, or low molecular weight polypeptide (LMP)-7 subunit of the immunoproteasome, both of which contain the chymotrypsin-like protease activity. In some cases, other constitutive proteasome and immunoproteasome subunits are bound as well, both in vitro and in vivo, though the contribution of these effects to their clinical anti-myeloma activity remains to be elucidated. An even larger number of proteasome inhibitors have been validated in laboratory studies that have not yet reached the clinic, but either show some promise in doing so themselves, or may serve as models for other agents that may do so. Attractive examples include agents that predominantly target proteasome catalytic sites other than β5, such as the β1 caspase-like [44] and β2 trypsin-like [45] activities. Other interesting inhibitors spare the constitutive proteasome while specifically targeting beta subunits of the immunoproteasome, including either LMP-2 [46-49], such as IPSI-001 [48], or LMP-7, such as ONX 0914 (formerly PR-924) [50]. Finally, still other proteasome inhibitors seem to exert their effects on protein turnover by interacting with alpha subunits of the proteasome, which have been classically felt to be structural in nature and not catalytic, or at the interface between the alpha and beta subunits. Examples of these include the cathelin-like peptide PR-39 [51, 52], 5-amino-8-hydroxyquinoline (5AHQ)[53], and chloroquine [54]. These may work in part by inducing allosteric changes in the structure of the proteasome, which could reduce or eliminate access by protein and peptide substrates to the proteolytically active β subunits. Importantly, agents with their mechanism of action could all ultimately be incorporated into our therapeutic algorithms against myeloma in the relapsed and relapsed/refractory settings.

Future Indications for Novel Proteasome Inhibitors

Next generation proteasome inhibitors could play an important role in therapy both alone, and in combination with other drugs, in patients with advanced multiple myeloma in several prominent settings, which are reviewed below.

Patients with Refractory Myeloma

Patients with multiple myeloma that is both relapsed and refractory to their prior line of treatment, and especially those who have already received regimens containing bortezomib and lenalidomide, have a poor prognosis [55], and need new treatment approaches [56]. Since such patients make up a group for whom there is an unmet medical need, they represent an attractive population for testing of any new drugs, and successful results in even a single arm phase II study could lead to regulatory approvals [57]. Next generation proteasome inhibitors in pre-clinical studies have been shown to have the ability to overcome drug resistance, including in some cases in models of bortezomib resistance [33, 35-39, 41-43], making them attractive candidates for this indication. Reports of initial phase I study data with carfilzomib [58, 59], marizomib [60, 61], and MLN9708 [62] all support the possibility that these agents will indeed be active for relapsed and refractory disease. Carfilzomib has advanced the furthest in clinical development, and as detailed earlier in this supplement, studies have now confirmed its substantial activity in this setting. In particular, the PX-171-003 phase IIb trial showed that 24% of 257 response-evaluable patients who had received a median of 5 prior lines of treatment, and were refractory to their prior line of therapy, achieved at least a partial response with carfilzomib, while 36% had a minimal response or better [63, 64]. Among patients who were refractory to bortezomib in a previous line of therapy, which was true for 65% of the population as a whole, up to 19% achieved a partial response or better, while 34% were able to obtain at least a minimal response. Response durability was also achieved, with durations of response of 7.4 months for those with a partial response or better, and 8.3 months for those with at least a minimal response.

The ability of one proteasome inhibitor to overcome clinical resistance to another is not surprising, in that there are other examples, even in the myeloma field, of the successful sequential use of two different drugs that are nonetheless in the same class. Notably, patients who have been exposed to thalidomide can respond to lenalidomide-based regimens, even if their disease was previously refractory to the first generation immunomodulatory agent [65]. Optimal use of all the currently available and future proteasome inhibitors would be significantly advanced with the development of a better understanding of the mechanisms of resistance to this class of drugs. It seems clear based on pre-clinical studies that there are many pathways involved in de novo drug resistance which reduce the efficacy of bortezomib, and likely other proteasome inhibitors as well, in bortezomib-naïve models. Prominent examples include inducible drug resistance mechanisms such as activation of heat shock proteins (HSPs) like HSP-27, -70, and -90 [66-69]; stress response proteins like mitogen-activated protein kinase phosphatase [70-72]; and anti-apoptotic signal transduction pathways such as protein kinase B/Akt [67, 73-75], and their downstream targets including Bcl-2 family proteins [76]. Secondary or acquired resistance is also an important phenomenon, and exhibits its effects in patients who initially respond to bortezomib but then do not benefit from retreatment, or in some cases even progress while on proteasome inhibitor-based therapy. Laboratory models of drug-induced resistance suggest that overexpression of the β5 proteasome subunit [77, 78], or mutation of the β5 subunit to reduce bortezomib binding [77, 79, 80] may be involved. Other studies have implicated a role for multidrug resistance proteins [81, 82] that may act by promoting efflux of peptide-based drugs, like most proteasome inhibitors, from the myeloma cell. Which, if indeed any of these mechanisms, are important in vivo, however, is not known at this time. This leaves open the possibility that other changes in the myeloma cell itself, or in either the local microenvironment or the host as a whole, could contribute as well. Enhanced metabolism of bortezomib, which occurs through oxidative deboronation [83, 84], is one such potential mechanism. Another is induction of different proteases that may be able to substitute for some proteasome functions, such as tripeptidyl peptidase II, which has been observed in eukaryotes [85], or tricorn protease [86], which may normally work downstream of the proteasome [87]. Finally, mutations in the proteasome outside of the β5 subunit could be important as well, such as by activating other proteolytic activities, including the branched chain amino acid-preferring and small neutral amino acid activities [2], that may substitute for the chymotrypsin-like activity primarily targeted by bortezomib. Such knowledge would have a strong impact on the appropriate sequencing of the different proteasome inhibitors. For example, if studies of plasma cells from patients relapsed after bortezomib revealed enhanced deboronation activities, then epoxyketones like carfilzomib would be favored as the next line of therapy as opposed to other boronates, like CEP-18770 or MLN9708. In contrast, if peptide efflux were responsible for resistance, then use of a non–peptide inhibitor such as marizomib would be favored, or one could use approaches with bortezomib or carfilzomib in combination with an efflux pump inhibitor.

Patients with Relapsed Multiple Myeloma

Agents that show activity in relapsed and refractory myeloma are typically rapidly moved into earlier disease settings, including in patients with relapsed myeloma, and then eventually for newly diagnosed patients. Next-generation proteasome inhibitors are certainly following that path both as single agents [33, 35, 36, 40], and in rationally designed combination regimens based on strong pre-clinical data showing their ability to cooperate with other anti-myeloma agents [37-39, 41-43]. Data are particularly well developed for carfilzomib which, as detailed earlier in this supplement, was able to induce a 53% overall response rate in a bortezomib-naïve population of relapsed myeloma patients with a median of 2 prior lines of therapy, who received 20 mg/m² of carfilzomib during cycle 1, and then 27 mg/m² thereafter [88]. Carfilzomib was also able to be easily combined with lenalidomide and dexamethasone, as demonstrated in a phase Ib/II trial of this regimen for relapsed and/or refractory disease [89]. This study showed that full dose carfilzomib was tolerated well, along with the standard lenalidomide and low dose dexamethasone regimen [90]. Moreover, it achieved rapid responses and a 78% ORR [89], leading to the initiation of a phase III trial comparing the three-drug regimen to lenalidomide and low dose dexamethasone in patients with one to three prior lines of therapy [91].

Bortezomib is now used in combination with a wide array of both conventional and novel agents against multiple myeloma, and it is likely that the same will be true for carfilzomib and other second-generation proteasome inhibitors. Indeed, a number of such trials are currently underway (Table 1), including combinations with other standards of care like lenalidomide and dexamethasone, as well as novel drugs like the histone deacetylase inhibitor panobinostat and the kinesin spindle protein inhibitor ARRY-520, while still others are in the planning stages. An especially interesting combination approach suggested by the different chemistries and targets of the various proteasome inhibitors is the possibility that two such agents could be combined to good effect. One such combination could include bortezomib, which can inhibit both the chymotrypsin-like and the post-glutamyl peptide hydrolyzing, or caspase-like activities [92, 93], with marizomib, which inhibits the chymotrypsin-like and trypsin-like activities, and the caspase-like activity as well, though to a lesser extent than bortezomib [41]. Such a regimen has already been validated using both in vitro [41] and in vivo [42] model systems. Similarly, pre-clinical studies have shown that inhibitors which predominantly target proteasome catalytic sites other than β5, such as the β1 caspase-like [44] and β2 trypsin-like [45] activities, can sensitize to bortezomib. It may also be possible to use bortezomib, which targets both the constitutive proteasome and the immunoproteasome, in combination with an immunoproteasome-specific inhibitor, though some data suggest that targeting both β5 and β5_i is needed to optimally induce cell death.[48] Finally, it is also possible that additive effects could be achieved by combining bortezomib or carfilzomib with some of the agents that target the alpha proteasome subunits, such as clinically relevant analogues of 5AHQ [53], and chloroquine [54]. Validation of some of these approaches would still be required, but one could envision randomized studies comparing, for example, bortezomib and dexamethasone to bortezomib with marizomib and dexamethasone, targeting either patients with previous bortezomib exposure, or even proteasome inhibitor-naïve patients. Were the synergy that has been demonstrated preclinically translate well to the clinic, it is possible that lower doses of each agent could be used, which would result in enhanced or at least similar anti-myeloma activity with decreased treatment-emergent toxicities, such as peripheral neuropathy.

Table 1. Ongoing Studies Incorporating Second Generation Proteasome Inhibitors for Relapsed and/or Refractory Multiple Myeloma^*.

Phase	Agent(s)	Proteasome Inhibitor Dosing	ClinicalTrials.gov Identifier
Carfilzomib Alone or in Combination
I/II	Carfilzomib + ARRY-520	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT01372540
II	Carfilzomib (for patients with varying renal insufficiency)	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT00721734
I/II	Carfilzomib (replace bortezomib in bortezomib-containing regimen to which patients are now refractory)	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT01365559
I/II	Carfilzomib (infusional)	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT01351623
I/II	Carfilzomib + Panobinostat	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT01301807
Ib	Carfilzomib + Lenalidomide/dexamethasone	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT00603447
I/II	Carfilzomib + Pegylated liposomal doxorubicin	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT01246063
II	Carfilzomib	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT00511238
III	Carfilzomib (vs. best supportive care)	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT01302392
III	Carfilzomib + Lenalidomide/ dexamethasone (vs. Lenalidomide/dexamethasone)	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT01080391
I/II	Carfilzomib + Vorinostat/ lenalidomide/dexamethasone	Not available	NCT01297764
II	Carfilzomib	Intravenously on days 1, 2, 8, 9, 15, 16 of every 28-day cycle	NCT00530816
CEP-18770 Alone or in Combination
I/II	CEP-18770	Intravenously on days 1, 8, 15 of every 28-day cycle	NCT01023880
I/II	CEP-18770 + Lenalidomide/dexamethasone	Intravenously on days 1, 8, 15 of every 28-day cycle	NCT01348919
Marizomib Alone or in Combination
I/II	Marizomib	Intravenously on days 1, 8, 15 of every 28-day cycle	NCT00461045
MLN9708 Alone or in Combination
I/II	MLN9708	Orally on days 1, 4, 8, 11 of every 21-day cycle	NCT00932698
I/II	MLN9708	Orally on days 1, 8, 15 of every 28-day cycle	NCT00963820

^*Based on a search of http://www.clinicaltrials.gov/ performed on July 23, 2011.

Myeloma Patients who are Intolerant of Bortezomib

Early studies of single-agent bortezomib in relapsed and/or refractory myeloma [3-6] identified four major categories of toxicities. These included constitutional symptoms such as fatigue and nausea, cytopenias such as neutropenia and thrombocytopenia, gastrointestinal effects such as diarrhea or constipation, and peripheral neuropathy. After more than a decade of experience with bortezomib [94], these remain the major notable toxicities, and of these, neuropathy, which was seen in 21-64% of patients treated on the phase III studies [95-97], has most influenced dosing. A number of approaches can be used to limit the incidence and severity of bortezomib-induced peripheral neuropathy (BIPN). Among these are the application of a dose-reduction algorithm [98], the use of bortezomib in the context of combination regimens [8], dosing of bortezomib once weekly [99-103], and, most recently, administration of bortezomib as a subcutaneous injection [104]. Despite these options, grade 2-4 peripheral neuropathy remains a significant complication of bortezomib in either the newly diagnosed or relapsed and/or refractory settings. For example, in a randomized study of bortezomib given either subcutaneously or intravenously with oral dexamethasone, the overall rate of neuropathy of any grade was 38% versus 53%, respectively, while the incidence of grade 3 or worse neuropathy was 6% versus 16%, respectively [104].

Fortunately, for patients who have previously developed BIPN, next-generation proteasome inhibitors seem to be emerging as a very rational option. Presentations of data from phase I studies of MLN9708 [62] and marizomib [60, 61] have reported a low rate of neuropathy, though one could argue that this may be due to the relatively limited exposure of patients to these agents on early phase clinical trials. A greater experience has been gained from studies of carfilzomib both with respect to the number of patients treated, and the duration of treatment. Phase I trials of single-agent carfilzomib showed relatively low levels of grade 3 and 4 peripheral neuropathy in all patients [58, 59], and this has been borne out in the two completed phase II trials as well [63, 105, 106]. Indeed, a pooled safety analysis of all four studies revealed that only 20 patients (3.9%) experienced any grade of neuropathy, with only 2 (0.4%) suffering grade 3 symptoms [107]. Moreover, the presence of baseline peripheral neuropathy did not seem to impact upon either the depth or durability of responses to carfilzomib, or on the tolerability of this agent [108]. One hypothesis that has been proposed to explain the different effect of these two drugs is that both bortezomib and carfilzomib induce the mitochondrial high temperature requirement protein A2 (HtrA2/Omi)[109]. However, this serine protease, which has been implicated in neuronal cell survival, is then inhibited by bortezomib but not by carfilzomib, possibly resulting in greater pro-apoptotic signaling in neurons of bortezomib-treated patients.

Trials targeting patients who previously could not tolerate bortezomib due to neuropathy-related adverse events will be needed in order for novel proteasome inhibitors to gain approval specifically for this indication. These could include patients who were intolerant of bortezomib due to other toxicities, including dermatologic effects such as epidermal necrolysis [110] or vasculitis [111], pulmonary effects such as interstitial pneumonitis [112, 113] or capillary leak syndrome [114], hepatitis [115], or cardiac toxicities [116], among others. The low incidence of these effects would probably dictate that the vast majority of patients eligible for such a study would indeed have neuropathy. Randomized studies in such patients comparing the novel inhibitor to bortezomib as a control may be difficult to perform for ethical reasons. However, it may in the future be possible to pursue such studies in bortezomib-naïve patients selected specifically for their risk of neuropathy. A recent trial of bortezomib with melphalan and prednisone as initial induction therapy found that the presence of neuropathy at baseline was the best predictor for the development of treatment-emergent symptoms [117]. If borne out by other findings, this would suggest that patients with preexisting neuropathy could benefit more from regimens based on a next-generation proteasome inhibitor such as carfilzomib. Moreover, molecular studies suggest that the use of gene expression profiling and single-nucleotide polymorphism analyses on patients at baseline may be able to identify those that are at increased risk of BIPN [118, 119]. Such patients would also be candidates for studies testing nerve-sparing proteasome inhibitors such as carfilzomib.

Conclusions

Proteasome inhibition has been firmly established as part of the standard of care for patients with newly diagnosed, and relapsed and/or refractory multiple myeloma during the first decade of experience with bortezomib, which was the first clinically relevant agent in this class. The next generation of proteasome inhibitors with novel pharmacologic, pharmacokinetic, and pharmacodynamic properties have now entered the clinical arena, including carfilzomib, CEP-18770, marizomib, MLN9708, and ONX 0912, and others will likely soon do so. Among the attractive molecular properties of some of these agents are their ability to inhibit the proteasome more specifically, to bind more proteasome protease activities, to target the immunoproteasome, and to induce proteasome inhibition by binding to alpha subunits, which may induce allosteric changes in the proteasome itself. From a clinical perspective, these agents are attractive due to the possibility to dose some through the oral route, the likelihood that they can overcome drug resistance, including to bortezomib, the ability to easily combine them with other anti-myeloma agents, and the possibility to improve upon the safety profile with a lower incidence of peripheral neuropathy. Regulatory approvals in the relapsed and/or refractory setting will likely first be achieved for these agents as standalone approaches in patients whose myeloma has progressed on their most recent line of therapy. They will then be good candidates for use in patients with fewer lines of therapy who have relapsed after a prior treatment either as single agents, or in combination regimens with standard or novel drugs, including other proteasome inhibitors. Randomized studies comparing bortezomib with next-generation agents will be needed to determine which drugs will be preferred in proteasome inhibitor-naïve patients, and may be able to identify subgroups who would benefit more from one agent or another. For patients who have previously received bortezomib, optimal use and sequencing of other proteasome inhibitors will be aided by an understanding of the mechanisms of de novo and acquired drug resistance in each patient’s plasma cells, as well as the potential contributions to resistance from the marrow microenvironment and the macroenvironment (Figure 1). Over the next decade, the validation of next-generation agents and their use in rationally designed combination regimens and sequences seems destined to further revolutionize the care of myeloma patients, and bring us closer to a cure for this disease.

Fig 1.

A possible future decision tree for use of novel next-generation proteasome inhibitors in patients who have previously received bortezomib-based therapy either as part of induction, or as part of a salvage regimen, and are now considered to be bortezomib refractory. Mechanisms that may be responsible for this resistance are indicated, as are some approaches that could overcome these mechanisms through the use of novel next-generation proteasome inhibitors. Increased metabolism of bortezomib through deboronation could be overcome with the use of non-boronate inhibitors such as carfilzomib, marizomib, or ONX 0912. Enhanced efflux of peptide-based drugs could be overcome with the use of non-peptide proteasome inhibitors like marizomib, or clinically relevant derivatives of 5AHQ and chloroquine. Activation of anti-apoptotic pathways such as Akt or HSPs could be overcome with the use of specific pathway inhibitors with bortezomib or carfilzomib. Mutation of the β5 proteasome subunit could be overcome with the use of inhibitors that better target the new binding pocket through the amino acids that are adjacent to the moiety that binds the threonine active site of β5. Overexpression of the β5 proteasome subunit could be overcome with the use of irreversible inhibitors that can be safely given at high doses, like carfilzomib.[120] Induction of new proteases that could substitute for the proteasome could be overcome with the use of specific agents targeting these new proteases either alone, or in combination with proteasome inhibitors. Sequestration of toxic ubiquitinated proteins in aggresomes could be overcome with the use of agents such as histone deacetylase inhibitors, which appear to block aggresome formation and therefore enhance proteasome inhibitor-mediated cell death.[121-123] Finally, deubiquitination of ubiquitinated proteins to reduce their cytotoxic effects could be overcome with the use of clinically relevant inhibitors of deubiquitinating enzymes [124], which also could be used alone or in combination with proteasome inhibitors. Please note that many of these mechanisms should be considered hypothetical at this time even in the pre-clinical setting, and none have been fully validated in the clinical arena. This leaves very open the possibility that other, as yet uncharacterized mechanisms may contribute, or even play predominant roles in resistance to proteasome inhibitors.

Abbreviations

5AHQ

5-amino-8-hydroxyquinoline

BIPN

bortezomib-induced peripheral neuropathy

CR

complete response

HSP

heat shock protein

HtrA2

high temperature requirement protein A2

LMP

low molecular weight polypeptide

ORR

overall response rate

OS

overall survival

TTP

time to progression