All-Trans-Retinoic Acid Prevents Carfilzomib-Induced Cardiotoxicity By Decreasing Activation of the Renin-Angiotensin System

INTRODUCTION

Bortezomib (BTZ), Carfilzomib (CFZ) and Ixazomib (IXA) are proteasome inhibitors (PI) approved for Multiple Myeloma (MM) treatment. By design, they all target the rate-limiting proteasome ß5 subunit. CFZ treatment increases the survival of patients with relapsed/refractory MM compared to BTZ. However, CFZ treatment is associated with heart failure not commonly observed for BTZ. The molecular basis for CFZ-induced cardiotoxicity is poorly understood. We have shown before that CFZ co-inhibits the ß5 and ß2 proteasome subunits compared to co-inhibition of the ß5/ß1 subunits by BTZ. We hypothesized that the unique ß5/ß2 subunit inhibition pattern of CFZ likewise explains the CFZ related acute cardiotoxicity.

METHODS

Isolated primary murine cardiomyocytes treated with BTZ, CFZ or ß5, ß2 and ß1 subunit-selective PI were used as in vitro model. Selectivity of proteasome subunit inhibition was confirmed with activity based chemical probes. Cardiomyocyte contractility was assessed in vitro using motion vector image analysis. The accumulation of individual proteins in cardiomyocytes undergoing 1h treatment with different PI was analyzed by liquid chromatography/mass spectrometry (LC/MS). In vivo analysis of the effects of differential proteasome subunit inhibition patterns was performed by electrocardiography (ECG) monitoring and untargeted metabolomics analysis from murine heart and plasma using R v.3.5.1 (2018-07-02) and Matlab 2020a.

RESULTS

Co-inhibition of ß5/ß2 proteasome subunits by CFZ or by respective subunit-specific PI for 1h resulted in increased proteasome inhibition and proteotoxic stress in cardiomyocytes compared to ß5/ß1 inhibition induced by BTZ. Likewise, CFZ or the ß5/ß2 type co-inhibition of the proteasome impaired contractility of cardiomyocytes in vitro, in contrast to BTZ or to ß5/ß1-selective proteasome inhibition. LC/MS-based proteome analysis of murine primary cardiomyocytes treated with either CFZ-type or BTZ-type proteasome inhibition identified accumulation of Retinal Dehydrogenase 1 (ALDH1A1) in PI-treated cardiomyocytes, with CFZ showing a slightly stronger effect. ALDH1A1 is involved in retinoic acid metabolism. Co-treatment of cardiomyocytes with its enzymatic end-product, all-trans-Retinoic-Acid (atRA), restored contractility of cardiomyocytes undergoing CFZ treatment in vitro. Leucyl-Cistinyl Aminopeptidase (LCAP) was identified as the protein that accumulated in CFZ-versus BTZ-treated cardiomyocytes (Log2 CFZ = 1.1, Log2 BTZ = -0.49, p=0.002). LCAP is involved in the Renin-Angiotensin System (RAS) and its increase suggests selective activation of RAS with increased degradation of RAS metabolites via LCAP in cardiomyocytes undergoing CFZ treatment. Indeed, mRNA expression for angiotensin converting enzyme 2 (Ace2) and angiotensin receptor 1 (Agtr1) was increased in beating cardiomyocytes treated with CFZ. Conversely, treatment with aminopeptidase inhibitor Bestatin prevented CFZ-induced cardiomyocyte toxicity in vitro. Consistent with our model, co-treatment of cardiomyocytes with CFZ and Valsartan, an Agtr1 inhibitor, likewise prevented CFZ- induced toxicity.

In vivo, 1h CFZ treatment selectively led to acute bradycardia with local activation of the RAS and accumulation of angiotensin in the heart. CFZ/atRA co-treatment decreased local angiotensin levels and prevented CFZ-induced bradycardia. There were no differences in angiotensin plasma levels in between the treated mice, suggesting that the local RAS in the heart rather than systemic effects of the RAS mediates the cardiotoxicity of CFZ.

CONCLUSION

Our data suggest that CFZ specifically impairs cardiac contractility in a dose-dependent manner that is caused by its unique ß5/ß2 proteasome subunit co-inhibition pattern. CFZ leads to more efficient functional proteasome inhibition in cardiomyocytes, resulting in accumulation of LCAP that subsequently activates the local RAS and induce acute cardiotoxicity. Based on our findings, CFZ-induced cardiotoxicity may be prevented or mitigated by either blocking RAS activation at the angiotensin II receptor level using selective drugs like Valsartan, or by reducing angiotensin levels with atRA. This provides a rational strategy to prevent CFZ-induced acute cardiac toxicity by modulating the activity of the local RAS with approved drugs.