Dear Editor, Bcl-2-associated athanogene 3 (BAG3) protein belongs to the family of co-chaperones that interact with the ATPase domain of the heat shock protein HSP70 through the BAG domain (amino acids 110–124). BAG3 contains also a WW domain, a proline-rich repeat (PXXP) and two conserved Ile-Pro-Val motifs that mediate binding to other proteins. 1 bag3 gene is constitutively expressed in a few normal cell types, including myocytes, and in several primary tumors, while its expression is inducible by stressors in other cell types. Many evidence indicate that BAG3 has a role in sustaining cell survival, through mechanisms that, depending on the cell context, rely on the ability of BAG3 to modulate levels or localization of apoptosis-regulating proteins, such as IKKу, Bax or BRAF, in either an Hsp70-dependent or-independent fashion. 1–3 BAG3 is expressed during cardiomyoblast differentiation and sustains myogenin expression. In cardiomyocytes, BAG3 protein localizes at the Z-disc and interacts with the actin capping protein, CapZβ1, stabilizing myofibril structure; BAG3 mutations can impair the Z-disc assembly and increase the sensitivity to stress-induced apoptosis. Mutations in bag3 gene have been associated with some forms of myofibrillar myopathy and dilated cardiomyopathy. 4–7 Interestingly, we detected BAG3 protein in supernatants of the rat cardiomyocyte cell line H9c2 and adult Human Cardiac Myocytes (HCMa) after 16 h of serum deprivation (Figure 1a). We then verified that BAG3 was present in extracellular vesicles isolated through a differential centrifugation procedure (Supplementary Figure S1A). To investigate the possibility that BAG3 could be released in vivo and become detectable in sera in conditions in which the heart is undergoing chronic stress, we analyzed sera from two patients affected by chronic heart failure (CHF). By western blot analysis, we detected a band recognized by anti-BAG3 antibody at the expected molecular weight in sera from patients but not from healthy donors (Figure 1b). We excised the band from a replica gel and subjected it to mass spectrometry, unmistakably identifying BAG3 (Supplementary Figure S1B). Furthermore, we found that sera from CHF patients recognized BAG3 protein in western blotting, using an anti-human IgG as secondary antibody (Supplementary Figure S1C). This result indicated the presence of anti-BAG3 antibodies in CHF patients’ sera. To confirm this finding, we developed an ELISA test using recombinant BAG3 to coat plates and anti-human IgG to reveal and analyzed sera from 52 CHF patients (EFo45%), compared with sera from 84 healthy donors. As shown in Figure 1c (and in Supplementary Figure S1D), we detected significantly higher values of anti-BAG3 antibodies in patients’ compared with controls’ sera. These data suggest that upon cardiac stress cardiomyocytes release BAG3 and this in turn results in production of auto-antibodies. There is no correlation with NYHA scores and antibody levels at this stage but screening of a larger number of patients in the future might be necessary to reveal potential correlations.

These results describe for the first time an extracellular BAG3 (eBAG3) released by stressed cardiomyocytes. As BAG3 lacks the consensus signal required for secretion via ER–Golgi pathway, it is likely to be released by the nonclassical secretory pathway. 8 eBAG3 release by stressed cardiomyocytes appears to result in production of autoantibodies that could potentially be used as a biomarker for CHF patients, in combination with other already established markers. The presence of anti-BAG3 antibodies in CHF patients’ sera indicates that released …