Carfilzomib Induces Cardiotoxicity Via ß5/ß2-Specific Proteasome Subunit Inhibition Pattern

BACKGROUND

All proteasome inhibitors (PI) approved for Multiple Myeloma (MM) treatment, Bortezomib (BTZ), Ixazomib (IXA) and Carfilzomib (CFZ), by design target the rate-limiting ß5 proteolytic proteasome subunit, but differ in their co-inhibition of the ß1 or ß2 proteolytic subunits at higher doses. CFZ is unique in inducing ß5/ß2 co-inhibition in a dose-dependent manner, which results in increased proteotoxic stress and higher cytotoxicity compared to the ß5/ß1 co-inhibition pattern mediated by BTZ or IXA. CFZ treatment, in particular at higher doses, is not only associated with increased clinical activity, but also with significantly higher cardiac toxicity compared to BTZ. The molecular basis for this difference is poorly understood, and mechanism-based pharmacological alleviation strategies are lacking. We hypothesized that the acute cardiotoxicity of CFZ is related to its unique ß5/ß2 inhibition pattern.

METHODS

Isolated primary murine cardiomyocytes showing spontaneous rhythmical contractions were used to assess the functional cardiotoxicity of PI in vitro. Cardiomyocytes were treated with the PI BTZ, CFZ or with highly-selective, subunit (ß1, ß2, ß5)-specific PI for 1h. Proteasome inhibition was directly verified using subunit-selective activity-based chemical probes. Contractility was assessed with motion vector image analysis, and calcium transients were examined by confocal microscopy after transduction of cardiomyocytes with a GCaMP vector. Differentially accumulated proteins after 1h PI treatment were quantitatively compared by liquid chromatography/mass spectrometry (LC/MS). In vivo, electrical changes and alterations in heart rate were monitored by electrocardiography (ECG). Calcium transients, ECG signals and LC/MS data were analyzed with R software version 3.5.1 (2018-07-02).

RESULTS

Treatment of cardiomyocytes with CFZ or with the combination of ß5/ß2-targeting subunit-specific PI impaired contractility in vitro in contrast to BTZ, or to co-inhibition of ß5/ß1 proteasome subunits. In vivo, the CFZ-type proteasome inhibition triggered acute bradycardia. Treatment of cardiomyocytes with CFZ induced a shift of intracellular calcium pools from the endoplasmic reticulum (ER) to the cytosol, consistent with ER-to-cytosol translocation of calcium described upon ER stress induction. Co-treatment with CFZ and cycloheximide, an inhibitor of protein synthesis, rescued cardiomyocytes from CFZ-induced functional impairment in vitro, suggesting that the accumulation of specific proteins is involved in CFZ-induced cardiomyocyte dysfunction. Quantitative proteomic comparison of primary cardiomyocytes treated by either CFZ-type or BTZ-type proteasome inhibition for 1h identified a selective accumulation of proteins of the retinoic acid pathway in CFZ-treated cardiomyocytes. Interestingly, co-treatment of cardiomyocytes with all-trans-retinoic acid (ATRA) prevented CFZ-induced acute cardiotoxicity in vitro.

CONCLUSION

Our data suggests that CFZ specifically impairs cardiac contractility through its unique ß5/ß2 proteasome subunit inhibition pattern, which results in more effective functional proteasome inhibition, protein accumulation and proteotoxic stress. The shift of intracellular calcium pools in cardiomyocytes upon CFZ treatment mirrors contractility impairment, and the specific changes identified in the retinoic acid pathway suggest ATRA as a potential drug candidate to alleviate CFZ-induced cardiac toxicity.