Molecular Targets and Mechanisms of Action of Specific Classes of PIs
In general, PIs demonstrate reversible or sustained binding to active sites in the 20S proteasome, primarily the ChT-L site.[1,9] However, different chemical classes have different β-subunit binding profiles that are thought to account for differences in pharmacokinetics and toxicity among PIs (Table).
It is presumed that peptide boronates form a stable, slowly reversible tetrahedral adduct with the active site threonine. Bortezomib reversibly binds to the 20S subunit, demonstrates high affinity for the proteasomal ChT-L active site and significant affinity for the C-L active site, and has no or minimal effect on T-L activity. Like other PIs, bortezomib targets numerous pathways by inhibiting the proteasome and controlling key transcription factors (eg, NFκB). In in vivo human tumor cell lines, bortezomib potently inhibited cell growth and proliferation. Biochemical studies indicate that bortezomib triggers pleiotropic signaling pathways, and a number of signaling events contribute to its antitumor activity. The apoptotic activity of bortezomib results from inhibition of NFκB activity, disruption of cyclin-dependent kinase activity, stabilization of c-Jun N-terminal kinases leading to Fas upregulation, stabilization of p53, and a shift of the proapoptotic and antiapoptotic balance in the Bcl-2 family of proteins.[1,9] The recovery of activity following bortezomib inhibition in vitro and in vivo probably takes place primarily by de novo synthesis of new proteasomal proteins rather than by reversibility of the boronate tetrahedral complex.[21,22]
The adverse event profile and intravenous administration of bortezomib, along with concerns regarding the potential for resistance, have underlined the need for next-generation PIs that have different molecular and chemical characteristics. These characteristics would include selectivity for various proteasome subunits, binding kinetics, and route of administration.
Both of the boronic acid PIs under development, MLN9708 and CEP-18770, are reversible inhibitors, primarily of proteasomal ChT-L activity, with potencies similar to that of bortezomib. In preclinical studies, CEP-18770 was a potent promoter of apoptosis in human MM cell lines and patient-derived cells while displaying a favorable cytotoxicity profile for normal epithelial cells, bone marrow progenitors, and bone marrow stromal cells. In MM cell lines, CEP-18770 exhibited concentration-dependent inhibition of NFκB activity, resulting in a reduction in proinflammatory cytokines (eg, tumor necrosis factor α, interleukin-1β), intracellular adhesion molecules, and the proangiogenic vascular endothelial growth factor (VEGF). CEP-18770 displayed similar potency to bortezomib in two hematologic tumor cell lines and in bone marrow aspirates from relapsed MM patients. Complete cell death was observed at 20 nanomolar (nM) with both agents.
Epoxomicin and its analogs are made up of a peptide that selectively binds with high affinity in the substrate binding pocket(s) of the proteasome and an epoxyketone pharmacophore that irreversibly interacts with the catalytic threonine residue to inhibit enzyme activity. Epoxyketones form a dual covalent morpholino adduct with the N-terminal threonine of the β5 subunit and react with both the hydroxyl and amino groups of the threonine residue at the active site. Carfilzomib demonstrates potent and sustained inhibition of proteasomal ChT-L activity (> 80% inhibition at ≥ 10 nM doses) in both in vivo and in vitro cellular models of MM by binding to the β5 subunits of the constitutive proteasomes and immunoproteasomes.[17,22,24] In ex vivo treated blood from normal human subjects and in samples derived from patients with MM, carfilzomib demonstrated greater selectivity for the β5 subunit compared with bortezomib. While bortezomib also displayed significant activity for the β1 subunit, carfilzomib had minimal affinity for the β1 and β2 subunits (at doses up to 100 nM). It is thought that the selectivity of carfilzomib for the β5 subunit contributes to its improved tolerability profile compared with bortezomib.
Irrespective of cell type, brief exposure (1 hour) to carfilzomib resulted in greater cytotoxicity (particularly to hematologic tumor cell lines) in terms of cell viability, apoptosis, and cell cycle progression, compared with bortezomib. Sustained inhibition of the proteasome with carfilzomib may explain the greater cytotoxic response. It is important to note that in vitro studies confirm that the sustained binding of carfilzomib to the proteasome does not significantly affect the rate of recovery of proteasomal activity (50% to 100% recovery within 24 hours in all tissues), which was only moderately slower than with bortezomib, suggesting that proteasomal recovery involves de novo protein synthesis. However, these results for carfilzomib cannot be generalized to whole blood; less than 50% recovery was observed after 1 week, compared with complete recovery after 48 hours in bortezomib-treated samples.
The binding profile of carfilzomib results in accumulation of ubiquitin-protein conjugates and proteasome substrates, as well as inhibition of myeloma cell proliferation via induction of apoptosis. Carfilzomib demonstrated potent antiproliferative and pro-apoptotic effects in myeloma cell lines and in patient-derived models of myeloma. Like bortezomib, carfilzomib-induced apoptosis occurs through intrinsic and extrinsic caspase pathways that converge on the effector caspase-3. However, in preclinical models of MM, carfilzomib demonstrated increased activity against MM cell lines, enhanced JNK phosphorylation, and greater potency in increasing caspase-3, caspase-8, and caspase-9 activity (1.5-fold, 1.8-fold, and 2.0-fold increases, respectively), compared with bortezomib. Of note, in both cell-line models and clinical samples, carfilzomib overcame primary and secondary resistance to bortezomib. The researchers suggest that this may be due to the sustained inhibition of the ChT-L subunit, which potentially requires the cell to synthesize and reassemble new proteasomes. This may overcome the cell’s attempt at resistance by overproduction of proteasome subunits and other ubiquitin-proteasome pathway proteins—a characteristic displayed by cells resistant to bortezomib.
The tripeptide epoxyketone ONX 0912 demonstrated sustained inhibition of ChT-L activity in vitro comparable to that observed with carfilzomib, and it induced tumor cell death across multiple tumor types with a half-maximal inhibitory concentration (IC50) of < 100 nM.
Salinosporamides bind to all three catalytic subunits of the proteasome and may also inhibit other cellular proteases, albeit at higher concentrations. NPI-0052, a non-peptide PI, provides sustained inhibition of the β5 subunit. In human erythrocyte 20S proteasomes and in MM cell lines, NPI-0052 inhibits ChT-L and T-L activities at lower concentrations than bortezomib; however, higher concentrations are needed to inhibit C-L activity. In contrast to bortezomib and carfilzomib, NPI-0052-induced apoptosis is predominantly mediated through caspase-8 (extrinsic) and not caspase-9 (intrinsic) pathways. NPI-0052 is a more potent inhibitor of NFκB and related cytokine secretion, and it is a significantly more potent inducer of apoptosis in MM cell lines compared with bortezomib (P < .005). It also significantly inhibits VEGF-induced migration of MM cells (P < .05). NPI-0052 appears to have a wider therapeutic index than bortezomib, with little reported effect on normal lymphocytes at therapeutic doses.
Rationales for Dosing Regimens
Bortezomib is limited by toxicities such as peripheral neuropathy (PN) and thrombocytopenia that restrict the clinical dosing regimen to a biweekly day 1/day 4 schedule, allowing recovery of proteasome activity between doses.[28,29] In an effort to improve tolerability in patients with relapsed/refractory MM, a recent analysis evaluated the feasibility of a single weekly dose of bortezomib in the setting of combination therapy. The majority of response parameters, including overall response rate and progression-free survival, were similar, while the safety profile was improved (with a PN incidence of 8% vs 28% in patients receiving twice-weekly dosing). However, long-term clinical experience with this dosing schedule has been limited thus far. A separate phase III study examined the effect of subcutaneous dosing of bortezomib on safety and efficacy parameters. The findings of this study were also promising and showed similar results for efficacy, with moderate improvements in the safety profile (PN ≥ grade 3 = 6% vs 16% in patients receiving intravenous dosing).
In human tumor xenograft models, carfilzomib administered on either 2 or 5 consecutive days at doses resulting in > 80% proteasome inhibition demonstrated increased antitumor activity with acceptable toxicity compared with the dosing regimen of bortezomib. Phase I dose-escalation studies of two dose-intensive schedules of carfilzomib (a 2-week cycle with dosing on 5 consecutive days followed by 9 days rest in one study, and a 4-week cycle with dosing on 2 consecutive days weekly for 3 weeks followed by 12 days rest in another) confirmed proteasome inhibition levels exceeding 75% in whole blood 1 hour after the first dose, an absence of dose-limiting toxicities, and no PN ≥ grade 3.
The development of an oral PI will improve both flexibility of administration and patient convenience. Orally administered ONX 0912 demonstrated proteasome inhibition > 80% at doses four to ten times less than the maximum tolerated dose. In human tumor xenograft models of MM, ONX 0912 displayed antitumor activity at least equivalent to carfilzomib at well-tolerated doses.
Although both carfilzomib and bortezomib potently inhibit the ChT-L activity of the proteasome, bortezomib exhibits off-target activity directed at several additional serine proteases. This off-target activity is presumably due to the fact that bortezomib was originally identified as a potential inhibitor of serine proteases and still retains some activity against those targets. However, the extent to which off-target inhibition contributes to the toxicities of bortezomib, especially PN, remains to be determined. In preclinical studies, carfilzomib did not exhibit the off-target activity against non-proteasomal proteases that is observed with bortezomib.[34,35] Comparative data from an in vivo model of peripheral nerve degeneration demonstrated that bortezomib is ten times more potent at inducing neurodegeneration and five times more cytotoxic to neuroblastoma cells than carfilzomib. While equivalent proteasome inhibition was reported for the two agents, only bortezomib inhibited nonproteasomal targets, including the mitochondrial serine protease HtrA2 (which protects neurons from stress-induced apoptosis) and such serine hydrolases as cathepsin G, cathepsin A, dipeptidyl peptidase-4, and chymase. In contrast, carfilzomib displays minimal off-target activity, primarily because it is substantially more selective in its binding to N-terminal threonine proteases. When given to animals or humans in daily intravenous doses for up to 5 days (the maximum number of daily doses tested), carfilzomib produced extremely high levels (> 80%) of prolonged proteasome inhibition without dose-limiting PN.[22,36] Bortezomib-induced grade 3 to 4 PN occurs in up to 30% of patients with recurrent disease, potentially necessitating dose modifications that may lead to suboptimal levels of proteasome inhibition. Research is currently focused on further characterization of the nonproteasomal targets of bortezomib and the clinical consequences of their inhibition.
One recent psychophysical study of PN in 11 patients with MM treated with bortezomib in the front-line setting and 26 control subjects found persistent neurophysiological changes associated with bortezomib treatment. These changes were characterized by decreased skin temperature in the area of pain, diminished touch and sharpness detection, and decreased sensitivity to skin heating. Tandem immunohistochemical analyses of patient skin biopsies revealed significant decreases in the density of epidermal nerve fibers and a complete loss of Meissner corpuscles. While such PN can be a painful and debilitating consequence of therapy with bortezomib, pilot studies of complementary medical approaches, including acupuncture, have shown some promise in reducing the intensity of these effects. A recent study involving 20 patients with bortezomib-induced or thalidomide(Drug information on thalidomide)-induced PN evaluated the effect of electroacupuncture over a period of 9 weeks. Improvements were seen in Functional Assessment of Cancer Therapy/Gynecologic Oncology Group (FACT/GOG)-neurotoxicity scores, pain severity, self-reported physical well-being, and coin test scores. Larger controlled studies are needed to better characterize the nature of bortezomib-induced PN and the effects of complementary therapies.
Over the past decade, proteasomal inhibition has emerged as an important therapeutic strategy, and PIs represent a promising treatment option for the management of MM. Variations in the binding profiles of different PIs may translate into key differences in pharmacokinetic and toxicity profiles that may prove clinically relevant in the treatment of MM. Further clinical data from emerging classes of PIs are awaited with interest.
Acknowledgments: The author would like to thank Anna Battershill, MSc, and Brian E. Szente, PhD, of Fishawack Communications for their assistance with manuscript development. Their editorial support was funded by Onyx Pharmaceuticals.