Neurodegeneration Induced by Bortezomib Exposure in Vitro Occurs Via Proteasome Independent Mechanisms.

The dipeptide boronate proteasome inhibitor bortezomib (BTZ; Velcade®) is approved for the treatment of multiple myeloma and non-Hodgkin's lymphoma. Bortezomib-induced peripheral neuropathy (BIPN, Blood (2008)112:1593-1599) is seen in ∼30% of BTZ-treated patients and can result in dose reductions and discontinuations that may result in suboptimal levels of proteasome inhibition. Carfilzomib (CFZ), a tetrapeptide epoxyketone, is a selective and irreversible proteasome inhibitor that is structurally and mechanistically distinct from bortezomib. Single agent treatment with CFZ has demonstrated strong activity in relapsed and refractory myeloma and a favorable safety profile in Phase 2 trials (ASH2008:864 & 865). Importantly treatment-emergent PN was seen at low levels and did not result in dose modifications or discontinuations. The disparate safety data for these proteasome inhibitors suggest that non-proteasomal mechanisms may underlie BIPN. Using activity-based probes in peripheral blood mononuclear cell (PBMC) lysates, we previously demonstrated inhibition of non-proteasomal proteases by BTZ and other proteasome inhibitors with a boronate pharmacophore (EHA2009:0939). However, the involvement of the proteasome in the peripheral nerve degeneration and BIPN in BTZ-treated myeloma patients remains to be established.

To establish an in vitro model of peripheral nerve degeneration and to determine the effects of proteasome inhibition by BTZ and CFZ on neurite outgrowth and cell survival.

SH-SY5Y neuroblastoma cells were differentiated by long term culture in retinoic acid and brain derived nerve growth factor to induce neurite outgrowth. The effects of proteasome inhibitors were measured by high content image analysis of fluorescent images for cell survival (Hoechst nuclear counterstain) and neurite degeneration (FITC-mouse anti-beta-III-tubulin). Phase contrast images were also collected to observe morphological effects and gross cell death. Cell viability and proteasome inhibition was measured in undifferentiated and differentiated cells. The MEROPS (peptidase) database was mined for candidate serine proteases with a P1 selectivity of Leu/Phe/Tyr to identify candidate off-targets CFZ and BTZ and candidate proteases were validated by standard biochemical and cell biology techniques.

In differentiated SH-SY5Y cells, the average neurite length decreased by 33% following 24 hr exposure to 10nM BTZ but was unaffected by the same concentration of CFZ. Proteasome inhibition as determined by a fluorescent substrate for the chymotrypsin-like activity was equivalent (∼70%) after a 24 hr exposure for both compounds in differentiated cells, suggesting that neurodegeneration involves non-proteasomal pathways. With 72 hrs continuous exposure, BTZ was 10-fold more potent than CFZ at inducing neurodegeneration. Furthermore, in both undifferentiated and differentiated SH-SY5Y cells, BTZ was 5-fold more cytotoxic than CFZ. Database mining for serine proteases with a selectivity for Leu/Phe/Tyr at P1 was used to identify other potential BTZ targets that might underlie neurotoxicity. One candidate is HtrA2 (also called Omi), an inducible mitochondrial serine protease whose activity protects neurons from stress induced apoptosis (Hum Mol Genet (2005) 14(5):2099-2111). HtrA2 levels increased 2-fold in SH-SY5Y cells treated with either BTZ or CFZ for 6 hrs at 40 nM. Using a gel based assay and purified enzyme preparations, BTZ inhibited HtrA2 activity with an IC50 ∼ 4 nM, equivalent to its activity against the proteasome. In contrast, Carfilzomib did not inhibit HtrA2 at the highest concentration tested (10 mM).

These data demonstrate that BTZ induces neuronal cell death and neurite degeneration in vitro by proteasome-independent mechanisms. We propose that combined inhibition of the proteasome and HtrA2 by BTZ may underlie peripheral nerve toxicities in vitro and may be involved in BIPN in myeloma patients. In this model, CFZ, which mediates equivalent proteasome inhibition to BTZ in neurons, does not induce neurodegeneration due to inactivity against HtrA2. Future profiling of non-proteasomal targets of BTZ, including HtrA2 activity, in patient samples is merited. These results suggest that the favorable safety profile of CFZ in myeloma patients may be a result of its high selectivity for proteasomal proteases.

Arastu-Kapur:Proteolix, Inc: Employment. Ball:Millipore Corp: Employment. Anderl:Millipore Corp: Employment. Bennett:Proteolix: Employment. Kirk:Proteolix, Inc: Employment.