Bortezomib-Associated Peripheral Neuropathy: Relationship Between Clinical Neurophysiologic Evidence in Previously Untreated Multiple Myeloma Patients and Preclinical Characterization in a Mouse Model.

Abstract 3860

Poster Board III-796

Bortezomib (Velcade^®) therapy for multiple myeloma (MM) results in high overall and complete response rates, but can also lead to peripheral neuropathy (PN). Bortezomib-associated PN has been shown to be a cumulative, dose-related, primarily sensory neuropathy that is reversible to baseline in the majority of cases (Richardson et al, Br J Haematol 2009). Further research is needed to determine the mechanisms by which PN arises and to determine potential neuroprotective strategies. A preclinical model that reflects the neurophysiologic changes seen in patients developing PN during bortezomib treatment would prove highly valuable. Here we assess the relationship between clinical neurophysiologic findings in untreated MM patients who received single-agent bortezomib and developed PN in a phase 2 study (Richardson et al, J Clin Oncol 2009) and preclinical neurophysiologic and histologic characterization of bortezomib-induced PN in a SwissOF1 mouse model (Bruna et al, J Peripher Nerv Syst 2009). By CTCAE grading, 41/64 (64%) patients in the clinical study developed sensory PN, and 7 (11%) had motor PN. In 35 of these patients, PN was also assessed using extensive neurophysiologic testing, including motor and sensory nerve conduction studies (NCS) and quantitative sudomotor axon reflex testing (QSART). Of these patients, 22/35 (63%) developed PN by modified consensus criteria (England et al, Neurology 2005), including 7 (20%) who had worsening of baseline MM-associated neuropathy, and 15 (43%) who developed new small-fiber (n=7) or both large- and small-fiber PN (n=8). Similarities between neurophysiologic changes seen in the 15 patients with new-onset PN and in 20 animals given bortezomib twice-weekly for 6 weeks in the mouse model are summarized in the Table. Histologic studies in mice showed significant reductions in myelinated fiber count (–9.9%) and myelin thickness (–17.5%), likely secondary to axonal damage, in bortezomib-treated vs control animals; however, PN with demyelinating features is uncommon clinically. Importantly, bortezomib-associated PN was reversible in both clinical and preclinical studies. By CTCAE grading, sensory PN resolved in 35 of the 41 (85%) patients who had PN in a median of 98 days (reversibility was not assessed by neurophysiologic testing). In the mouse model, after a 4-week wash-out period, complete normalization was seen in sensory NCS and histologic findings. Overall, there appears to be good correspondence between clinical manifestations of bortezomib-associated PN and evidence from the mouse model, with both showing a predominantly sensory PN that affects both large and small fibers and is reversible. Preclinical studies with bortezomib and other agents have suggested that this PN is a mechanism-based effect of proteasome inhibitors associated with cytoplasmic accumulations of ubiquitinated proteins in dorsal root ganglia neurons (Silverman et al, ASH 2008). Furthermore, histologic findings in mice of mild axonal loss and secondary loss of myelinated fibers are in accord with NCS findings, and may be due to the reversible disruption of proteasomal degradation of PMP22, a glycoprotein incorporated into myelin (Gilardini et al, Curr Med Chem 2008). Given the similarities with clinical findings, the SwissOF1 mouse model of bortezomib-induced PN appears to represent a promising vehicle for further exploration of this toxicity and the development of neuroprotective strategies.