Abstract

Methods

Expanded methods are available in the on-line Data Supplement.

Results

Ubiquitin proteasome peptidase activity is markedly impaired in HCM and failing hearts compared to controls

Proteasome peptidase activity in whole heart protein extracts was measured over a range of ATP concentrations using synthetic fluorogenic peptides as substrates for chymotrypsin-like and caspase-like activities. Overall, ATP-dependent chymotrypsin-like activity was markedly decreased in failing and HCM hearts compared to non-failing control hearts (P<0.001, Figure 1A). In non-failing hearts, there was a small, non-statistically significant rise in chymotrypsin-like activity from 0 to 14 µM (p=0.37). In contrast, chymotrypsin-like activity decreased significantly in both HCM and failing hearts in response to ATP (P<0.001, range from 3.5 µM to 1 mM, data not shown for entire concentration curve). Basal and ATP-dependent caspase-like activities were similarly reduced in failing and HCM hearts compared to non-failing hearts (P<0.001, Figure 1B), with neither ATP-dependent activation nor inhibition in any of the 3 patient groups.

Open in a separate window

Figure 1

Cardiac 26S ubiquitin proteasome activity and ubiquinated protein accumulation in nonfailing, hypertrophic cardiomyopathy and end-stage failing hearts

NF=nonfailing, HCM=hypertrophic cardiomyopathy. All values are expressed as a mean ± SEM. A. Heart tissue homogenates (60 µg total cytosolic protein) were assayed for proteasome peptidase activity in the presence of ATP. Total chymotrypsin-like activity represents the difference in fluorescence in the presence and absence of the specific inhibitor, lactacystin (18 µmol/L). †P<0.05 for NF (n=6) vs HCM (n=17). *P<0.01 for NF vs HCM and NF vs Failing (n=14). B. Total caspase-like activity represents the difference in fluorescence in the presence and absence of the specific inhibitor, ZPNAC (10 µmol/L). *P<0.01 for NF (n=5) vs HCM (n=8) and NF vs Failing (n=8). C. Comparison of peak chymotrypsin-like activity in samples from the same patients (n=6) before and after implantation of a left ventricular assist device (LVAD). Each black square represents a single observation and red circle with line represents the mean % change in post-LVAD compared to pre-LVAD samples (*P<0.05). D. Comparison of chymotrypsin-like activity in samples from HCM patients with or without sarcomere gene mutations. *P<0.05 for mutation positive (n=6) vs mutation negative (n=7). E. Representative immunoblot of human total protein homogenates (50 µg) probed with an antibody against polyubiquinated proteins. F. Densitometric analysis for polyubiquinated proteins (>100KDa), standardized to GAPDH as a protein loading control. N=5 (NF), 11 (HCM), 10 (failing). *P<0.05 for NF vs failing.

Paired heart tissue samples collected prior to and after left ventricular unloading were analyzed from heart failure patients treated with left ventricular assist devices (LVADs) as a bridge to cardiac transplantation. The mean time from LVAD implantation to transplantation was 30.78 ± 11 (range 12.3–74) weeks and the mean increase in LV ejection fraction was 11 ± 4% (P<0.05). Proteasome activity increased significantly after mechanical unloading of the left ventricle (P<0.05 for [ATP] 0–28 µM, Figure 1C), consistent with recently published data²⁹, and indicating upregulation of the UPS machinery in response to unloading and partial functional recovery of the left ventricle. In the genotyped HCM subgroup, mean chymotrypsin-like activity was significantly lower in hearts from patients with vs those without pathogenic sarcomere gene mutations (P<0.05, Figure 1D) despite their younger age (37 ± 4 vs 54 ± 5 yrs, P<0.05)..

Proteasome dysfunction can result in intracellular accumulation of ubiquinated proteins^{12, 30}. Polyubiquinated proteins were increased in failing (P<0.05) compared to non-failing hearts (Figure 1E and F). However, ubiquinated protein content was not different between HCM and non-failing hearts despite a similar magnitude reduction in proteasome activity in HCM and failing hearts.

Akt and p53 levels are increased in HCM and failing hearts

In the next set of studies, potential pathways that could be affected by proteasome dysfunction were examined. Akt, a serine/threonine kinase, is the major mediator of mTOR activation and a key regulator of protein synthesis and cardiac hypertrophy. Akt is a direct target for degradation by the UPS, and Akt signaling is decreased in skeletal muscle atrophy^{31, 32}. We found Akt protein expression to be significantly increased in failing and HCM hearts compared to controls (Figure 2A), consistent with the observed decrease in proteasome activity in these 2 groups.

Open in a separate window

Figure 2

Akt and p53 expression in human whole heart homogenates

Representative immunoblots (left) and densitometric analysis (right) for Akt and p53. * P<0.05 for Akt and P<0.001 for p53 comparing both HCM and failing groups to non-failing.

Transcription factor p53 is a major mediator of apoptosis in response to various stimuli³³ including oxidative stress, and its stability is regulated by UPS degradation. p53 protein expression was markedly increased in both HCM and failing hearts compared to controls, with more variable levels seen in the HCM group (Figure 2). However, the expression of one of p53’s downstream targets, proapoptotic factor Bax, was not significantly different among groups (data not shown).

Oxidative protein damage and oxidative modification of 19S proteasome subunit Rpt 5 are increased in failing hearts

Lack of a difference in the absolute or relative content of proteasome species in heart tissue from patients with HCM or heart failure suggests a post-translational mechanism(s) to explain proteasome dysfunction. Oxidative injury to the proteasome is one such event that could interfere with protein substrate delivery to the 20S catalytic core. Accordingly, we assessed the degree of protein oxidation in human non-failing, HCM and failing hearts by measuring protein carbonyls and 4-hydroxynonenylated proteins. These oxidized derivatives were significantly increased in failing hearts (P<0.05), but not in HCM hearts, compared to non-failing hearts (Figure 3A and B). There was a non-significant reduction in total protein carbonyls after mechanical unloading with LVAD (Supplemental Figure 2), that correlates with improved functionality as shown above. To determine if the proteasomes themselves were oxidatively modified in failing hearts, we analyzed protein carbonylation of proteasome subunits in highly enriched 26S fractions prepared from non-failing and failing heart samples. Sufficient tissue was not attainable from any single HCM sample for this analysis. A single 53 KDa band was more reactive with the DNPH antibody in the failing compared to non-failing samples (Figure 3C). The identity of this band was confirmed to be Rpt5 by 2-D gel electrophoresis and immunodetection (Figure 3D). These data support the hypothesis that oxidative modification to the proteasome itself could account in part for the proteasome dysfunction observed in failing human hearts.

Open in a separate window

Figure 3

Oxidative protein damage in normal, hypertrophic and failing human hearts

Samples were analyzed for (A) protein carbonyls after exposure to 2,4-dinitrophenylhydrazine (DNPH) and (B) 4-hydroxynonelated proteins. The depicted membranes are representative of 5 to 6 different samples in each group. Densitometric analysis is presented in the graphs beneath the membranes. In the protein carbonyl samples, overall average density was calculated over the range of 30 to 115 kDa. *P<0.05 for NF vs failing hearts. C. (Left) Proteasome subunit proteins from highly enriched 26S proteasome fractions were reacted with DNPH and probed with a DNPH-specific antibody. (Right) The membrane was stripped and reprobed with antibodies specific for Rpt5 and Rpn2. D. (Left and middle) 2D-gel electrophoresis of proteasome subunits previously reacted with DNPH. These membranes were probed with an antibody specific for DNPH. (Right) The membranes were stripped and reacted with an antibody to Rpt 5.

Discussion

In this report, we show that two distinct human cardiac pathologies, HCM and heart failure, are associated with marked proteasome dysfunction compared to non-failing controls. One earlier study described highly variable levels of proteasome peptidase activities in human hearts, with no significant difference between failing and non-failing hearts²². The use of freshly prepared protein homogenates from hearts preserved with cardioplegia prior to freezing, and inclusion of ATP in our assays may explain the observed differences²³. An additional key finding in the present study is the identification of post-translational proteasome oxidation, suggesting that oxidation may be an important mechanism involved in proteasome dysfunction in human heart failure.

Our studies of proteasome peptidase activity yielded several unique and reproducible observations in human hearts. A small, but statistically insignificant rise in proteasome activity is observed in non-failing hearts in response to ATP, a magnitude lower than expected given the known dependence of the 26S proteasome on ATP for assembly and protein processing³⁴. This suggests a high level of basal activation in human heart tissue due to either residual activating ATP and/or a lower required activating concentration of ATP for the human compared to the rodent 26S proteasome. Additional possibilities include dissociation of the 19S from 20S during tissue processing³⁴ or diffusion of small fluorogenic substrates without the need for ATP-dependent processing by the 19S complex³⁵.

The most striking observation with regard to proteasome activity was the marked inhibition of activity at low concentrations of ATP in HCM and failing human hearts compared to non-failing hearts. A biphasic response of proteasome activity to ATP has been reported in rat hearts previously²³, but the reason for inhibition with increasing concentrations of ATP is currently unknown. Inhibition of activity by ATP in human HCM and failure is particularly notable as it distinguishes these disease states from the non-failing control hearts. We speculate that the addition of ATP could promote the assembly of a damaged or dysfunctional 19S to the 20S core, resulting in restricted substrate entry. The observation of oxidative damage to the 19S subunit Rpt5 supports this notion.

Mechanical circulatory support is a major advance in the treatment of refractory heart failure. LVAD support has many favorable effects on the myocardium, including increased myocyte contractile force³⁶, improved β-adrenergic responsiveness³⁶, normalization of Ca²⁺ cycling³⁷, and global changes in gene expression³⁸. We propose an additional benefit of mechanical unloading of the failing heart to be activation of the proteasome, analogous to findings with skeletal muscle unloading in which there is a shift in the balance from protein synthesis to degradation^{9, 39}. Activation of the proteasome in heart failure by mechanical unloading may reduce the toxic accumulation of oxidized or damaged proteins and contribute to left ventricular reverse remodeling.

Prior studies of protein degradation, aggregation, and autophagic cell death in human heart tissue have been largely limited to end-stage heart failure^{13, 20, 22}, with little direct evidence for proteasome dysfunction. Data on human cardiac hypertrophy is limited to one previous report demonstrating an increase in ubiquitin-related autophagy with aortic stenosis, only when concomitant systolic dysfunction was present²¹. Our study is the first to examine proteasome function in human HCM. In contrast to the present work, proteasome activation occurs in response to hypertrophic stimuli in animal and cellular models^17–19 Differences in the nature of the inciting stimulus, rate of hypertrophy development, and disease duration may not adequately reflect the chronic and insidious nature of human hypertrophic heart disease. Over half the cases of familial HCM are linked to mutations in sarcomere genes^{27, 40}, and this held true in the patients in this study. Proteasome activity was lower in patients with sarcomere gene mutations, compared to those without mutations. Along with previous observations of the effects of myosin binding protein C truncation mutant expression on UPS function in neonatal cardiac myocytes⁴¹, our results support the hypothesis that sarcomere mutant protein expression contributes to proteasome dysfunction in HCM. Decreased proteasome activity in human HCM also challenges the applicability of proteasome inhibition as a therapeutic strategy for suppressing cardiac hypertrophic growth.

Consistent with previous reports, polyubiquitinated proteins were increased in failing hearts compared to non-failing controls. Somewhat surprisingly, this was not the case for the HCM hearts, despite a comparable magnitude of reduction in proteasome activity to the failing hearts. The reason for this finding is not clear, but possible explanations include differences in the duration of proteasome dysfunction or in the rate or magnitude of accumulation of damaged or misfolded proteins in HCM compared to failing hearts.

Increased expression of key mediators of hypertrophy and apoptosis pathways, Akt and p53 respectively, is notable since both proteins are direct targets of the UPS and proteasome dysfunction would therefore be expected to increase their stability and steady state expression. Elevated p53 expression has been observed previously in failing human hearts⁴², but has not been assessed in human hypertrophy. The importance of the finding of increased p53 expression in human HCM patients is highlighted by previous work showing that p53 is essential for the transition from hypertrophy to heart failure in an animal model of pressure overload by inhibition of Hif-1 and subsequent impairment of angiogenesis⁴³.

What is the mechanism for proteasome dysfunction in human heart failure and HCM? Studies on the cardiac proteasome reveal significant heterogeneity in expression of different subunits and isoforms²⁶. Different proteasome subpopulations confer proteolytic substrate specificity, and may fluctuate in response to environmental stresses. For example, upregulation of the 11S activated proteasome was recently described in an experimental rat model of diabetic cardiomyopathy⁴⁴. However, in our study, lack of a change in proteasome subunit expression or in the availability of proteolytic active sites invokes a post-translational mechanism to explain proteasome dysfunction in the human diseased heart. One attractive hypothesis is that excess production of oxidized proteins could inhibit proteasome function by interfering with protein substrate delivery to the 20S catalytic core^{14, 15, 45}, and/or by oxidative modification of the proteasome itself¹⁵. Increases in total protein oxidation and oxidative modification of 19S subunit, Rpt 5, with heart failure support this idea. These data are particularly intriguing, given the recent observation that Rpt 5 plays a critical role in assembly and activation of the 26S proteasome⁴⁶. The present observations suggest that strategies to reduce proteasome oxidation may hold therapeutic value for the treatment of heart failure.

Unlike failing hearts, total protein oxidation was not significantly increased in HCM hearts, although a trend was observed. Other post-translational modifications, such as phosphorylation, may play a more important role in this subgroup of patients. Phosphorylation of proteasome subunits in vitro promotes assembly of the 26S and increases proteolytic activity^3–6. However, the functional significance of proteasome phosphorylation in vivo is largely unknown, and will be important to address in future studies on HCM and failing hearts.

The use of human heart tissue to identify potential disease mechanisms has obvious scientific value, although certain limitations should be acknowledged. In most cases, data is obtained at a single time point at an advanced stage of disease, which precludes tracking of disease pathways or determination of causality. In addition, statistical power is reduced due to limited availability of non-failing donor heart tissue and heterogeneity of clinical disease.

In conclusion, a marked and consistent decrease in proteasome activity in HCM and failing human hearts is observed in the absence of changes in proteasome protein content or availability of proteolytic active sites. A post-translational defect in regulation of protein substrate processing, and/or oxidative damage to accessible proteasome subunits, appears to be responsible for impaired protein degradation in these human disease states. Defective protein quality control has significant implications for disease pathogenesis in promoting the toxic accumulation of protein aggregates⁴⁷ and increasing steady state levels of pro-hypertrophic and pro-apoptotic factors. Future work in human tissue, complemented by animal models, will be necessary to define the precise mechanisms for proteasome dysfunction, and to establish a causal link to cardiomyopathy progression.

Supplementary Material

Acknowledgements

Footnotes

Reference list