Genetic ablation of Cullin-RING E3 ubiquitin ligase 7 restrains pressure overload-induced myocardial fibrosis

Fibrosis is a pathognomonic feature of structural heart disease and counteracted by distinct cardioprotective mechanisms, e.g. activation of the phosphoinositide 3-kinase (PI3K) / AKT pro-survival pathway. The Cullin-RING E3 ubiquitin ligase 7 (CRL7) was identified as negative regulator of PI3K/AKT signalling in skeletal muscle, but its role in the heart remains to be elucidated. Here, we sought to determine whether CRL7 modulates to cardiac fibrosis following pressure overload and dissect its underlying mechanisms. For inactivation of CRL7, the Cullin 7 (Cul7) gene was deleted in cardiac myocytes (CM) by injection of adeno-associated virus subtype 9 (AAV9) vectors encoding codon improved Cre-recombinase (AAV9-CMV-iCre) in Cul7flox/flox mice. In addition, Myosin Heavy Chain 6 (Myh6; alpha-MHC)-MerCreMer transgenic mice with tamoxifen-induced CM-specific expression of iCre were used as alternate model. After transverse aortic constriction (TAC), causing chronic pressure overload and fibrosis, AAV9-CMV-iCre induced Cul7-/- mice displayed a ~50% reduction of interstitial cardiac fibrosis when compared to Cul7+/+ animals (6.7% vs. 3.4%, p<0.01). Similar results were obtained with Cul7flox/flox Myh6-Mer-Cre-MerTg(1/0) mice which displayed a ~30% reduction of cardiac fibrosis after TAC when compared to Cul7+/+ Myh6-Mer-Cre-MerTg(1/0) controls after TAC surgery (12.4% vs. 8.7%, p<0.05). No hemodynamic alterations were observed. AKTSer473 phosphorylation was increased 3-fold (p<0.01) in Cul7-/- vs. control mice, together with a ~78% (p<0.001) reduction of TUNEL-positive apoptotic cells three weeks after TAC. In addition, CM-specific expression of a dominant-negative CUL71152stop mutant resulted in a 16.3-fold decrease (p<0.001) of in situ end-labelling (ISEL) positive apoptotic cells. Collectively, our data demonstrate that CM-specific ablation of Cul7 restrains myocardial fibrosis and apoptosis upon pressure overload, and introduce CRL7 as a potential target for anti-fibrotic therapeutic strategies of the heart.


Introduction
The Ubiquitin-Proteasome System (UPS) is a selective protein degradation pathway that is involved in the pathogenesis of several cardiac disorders [1,2]. Central to the UPS is the recognition of a substrate by an E3 ubiquitin ligase, a step pivotal for the ubiquitin-mediated degradation of substrate proteins by the 26S proteasome [3]. Cullin-RING complexes (CRLs) constitute the largest group of E3 ubiquitin ligases [4]. Cullin 7 (Cul7; formerly known as p193) is a component of the Cullin-RING E3 ubiquitin ligase 7 (CRL7), a multimeric enzyme composed of the RING finger protein ROC1, SKP1-FBXW8 substrate targeting subunit and CUL7 as scaffold protein [5][6][7]. In vitro, CRL7 was shown to target insulin receptor substrate 1 (IRS-1), a component of the insulin and insulin-like growth factor 1 (IGF1) signalling pathways, for ubiquitin-mediated degradation by the 26S proteasome [6,8]. Mouse embryonic fibroblasts of Cul7-/-mice were found to accumulate IRS-1 and exhibit increased activation of IRS-1 downstream pathways phosphatidylinositol 3 (PI3)-kinase/AKT and MEK/ERK, respectively [6]. In vivo, heterozygosity of either Cul7 or Fbxw8 resulted in elevated PI3 kinase/AKT activation in skeletal muscle tissue upon insulin stimulation when compared to wild-type controls [9].
In the heart, Cul7 was initially identified as regulator of cardiac myocyte (CM) cell cycle activity in response to expression of DNA of tumorvirus oncoproteins [10][11][12]. Co-expression of dominant-interfering CUL7 and p53 mutants blocked E1A-induced apoptosis, resulting in robust CM proliferation [13]. In addition, CUL7 was shown to physically interact with Simian Virus 40 (SV40) Large T (LT) antigen oncoprotein and p53 [11,12]. Transgenic expression of a dominant-interfering Cul7 mutant resulted in CM cell cycle re-entry at the infarct border zone four weeks after permanent coronary artery occlusion and was associated with an induction of DNA synthesis in the interventricular septa of infarcted hearts [14].
It was hypothesized that in CM, expression of dominant-interfering CUL7 mutants mimicked the effects of LT-antigen binding to CUL7, thereby altering its function [14]. We previously reported that binding of LT-antigen to CUL7 impairs CRL7 ubiquitin ligase function, resulting in the accumulation of the CRL7 substrate protein IRS-1 and upregulation of its downstream signalling pathways PI3K/AKT and MEK/ERK [15]. In the heart, PI3K/AKT plays an important cardioprotective role by regulating various cellular functions, e.g. CM survival and tissue remodelling [16]. Experimental activation of PI3K/AKT signalling resulted in reduced CM apoptosis during ischemia-reperfusion injury [17,18] or pressure overload models [19]. In contrast, decreased PI3K/AKT signalling was linked to elevated apoptosis [20].
Based on these observations, we sought to investigate whether CRL7 contributes to the regulation of fibrosis and apoptosis in the heart.

Generation of AAV vectors
Adeno-associated virus subtype 9 (AAV9) encoding iCre were prepared as described previously [23]. Briefly, infectious recombinant AAV vector particles were generated in HEK 293T cells by a cross-packaging approach whereby the vector genome was packaged into AAV9 capsid [25]. The rAAV titers, determined by measuring the copy number of viral genomes in pooled, dialyzed gradient fractions as described by Zentilin et al. [26], were in the range of 7x10 12 − 1x10 13 genome copies/ml. For AAV9-CMV-iCre production, HEK 293T cells were grown in triple flasks for 24 h (DMEM, 10% FCS). The AAV9-CMV-iCre backbone and helper (pDP9rs) plasmids were transfected into the HEK 293T cells using polyethylenimine (Sigma-Aldrich). The virus was purified after 72 hrs from benzonase-treated cell lysates over an iodixanol density gradient (Optiprep, Sigma-Aldrich). AAV titers were calculated by Real-time qPCR on vector genomes using FastStart Universal SYBR Green Master (Roche).

Animal disease models
Transverse aortic constriction (TAC) was performed as described previously [23]. In sham surgery, only the chest was opened, but no ligation of the aorta was carried out.
Cardiac dimensions and function were analyzed by pulse-wave Doppler echocardiography before TAC/sham surgery and before the animals were euthanized. All animal studies were performed in accordance with the relevant guidelines and regulations and with approval of the responsible authorities (Regierung von Oberbayern, Munich, Germany; permit number 55.2-1-54-2532-160-13).

Apoptosis assays
For terminal deoxynucleotidyltransferase-mediated dUTP nick-end labelling (TUNEL) analysis, paraffin sections were processed using the Situ Cell Death Detection Kit, TMR Red according to the procedures of the manufacturer for difficult tissue (Roche). DAPI was added for nuclear counterstaining. Images of whole-heart sections were acquired. Nuclei were automatically counted using an image analysis algorithm (MetaMorph). For in situ end labelling (ISEL) analysis, cryo-sections were processed using the KLENOW-FragEL DNA Fragmentation Detection Kit (Oncogene Research). Signal was developed with biotinylated nucleotides and horseradish peroxidase (HRP)-conjugated streptavidin, followed by incubation with diaminobenzidine.

Cell culture and conditioned medium experiments
Murine CM and CF were isolated from 8-week-old adult male Cul7+/+ and Cul7-/-animals as indicated and incubated in DMEM containing 10% FCS, 1% BrdU and 1% penicillin/ streptomycin.

Histochemical and immunohistological analyses
For analysis of collagen deposition, sections of left ventricular myocardium were fixed in 4% paraformaldehyde at room temperature (RT), embedded in paraffin and sectioned at 5 μm intervals. Hematoxylin/eosin (H&E) and Sirius Red/Fast Green staining were performed as described previously [23].

Immunoblot analysis
Protein extraction was performed in lysis buffer containing protease and phosphatase inhibitors (Roche) and immunoblots performed using standard procedures. Protein lysates were electrophoresed on 10% or 12% SDS-PAGE gels, transferred onto a PVDF membrane and blocked with BSA for two hours at RT.

Quantification of mRNA in isolated cells or tissue
RNA was extracted from primary CM, CF or whole heart tissue using TriFast (PeqLab) and oligo-dT-primed cDNA synthesis was performed with Superscript II (Invitrogen), according to the manufacturer's instructions. Quantitative Real-time PCR amplification of Cul7, Cyr61, Ctgf, Tgfb1, Mmp2, Mmp3, Mmp9, Pdgfa, Timp, Anp and Rpl32 was performed with primers listed in S1 Table using FastStart Universal SYBR Green Master (Roche). Data were normalized to Rpl32 as indicated. Gene expression was analysed using hearts from 8-week-old male Cul7-/-mice or respectively 12-week-old Cul7-/-mice after TAC. Material from AAV9-CMV-dsRed injected littermates served as control.

Wheat germ agglutinin staining
For determination of cardiomyocyte cross-sectional areas (CSA), 6 μm thick paraffin-embedded sections were prepared from murine hearts and stained with Alexa Fluor 647-conjugated wheat-germ agglutinin (WGA, Life Technologies, 1:100 dilution) for cell border determination and with SYTOX Green (Life Technologies, 1:1000 dilution) for nuclei detection. Images were taken from areas of transversely cut muscle fibers. The MetaMorph software (Molecular Devices) was programmed to recognize individual cells based on the WGA staining in an automated manner and proper thresholds were set for background and excessive fibrosis exclusion. MetaMorph's integrated morphometry analysis tool was used to calculate the average cell area of cardiomyocytes.

Statistical analysis
Data are shown as mean ± SEM. Statistical analysis was performed with Prism (GraphPad software, version 6). Differences between two means were assessed by unpaired t-test. Differences among multiple means were assessed by 1-way or 2-way ANOVA followed by the Bonferroni correction as indicated. A P-value of < 0.05 was considered significant.

Enhanced activation of PI3K/AKT signalling in cardiac myocytes depleted of Cul7
The CRL7 ubiquitin ligase complex was previously identified as regulator of the PI3K/AKT and MEK/ERK signalling pathways through ubiquitin-mediated degradation of IRS-1 [6]. Heterozygosity of either Cul7 or Fbxw8, both key components of CRL7, resulted in over-activation of PI3K/AKT signalling in skeletal muscle upon insulin stimulation [9]. To further delineate underlying molecular mechanisms of the anti-fibrotic phenotype observed in CM-specific Cul7 knockout mice, we investigated the abundance and activation of CRL7-regulated proteolytic target proteins and signalling pathways, respectively, in CM lysates by immunoblot analysis. Strikingly, CM lysates of Cul7 flox/flox AAV9-CMV-iCre mice showed an~3-fold (p<0.01) increase of AKT Ser473 phosphorylation when compared to Cul7 flox/flox AAV9-CMV-dsRed control mice (Fig 3A). IRS-1 protein level was increased~2-fold, albeit not statistically significant due to high variation between mice (p = 0.1705) (Fig 3B). Phosphorylation of ERK1/2 at Thr202/Tyr204 was moderately increased (~1.5-fold increase, p = 0.2555) (Fig 3C). Similar results were obtained with CM lysates of Collectively, these data confirm our previous findings of CRL7 as regulator of PI3K/AKT activation and provides first evidence for a physiological role of CRL7 in regulating PI3K/AKT signalling in the heart.

Loss of Cul7 results in reduced apoptosis of cardiac myocytes
Loss of CM, e.g. by apoptosis, is a hallmark of myocardial remodelling at conditions of sustained overload, leading to replacement fibrosis [30]. As PI3K/AKT signalling is a major prosurvival pathway to protect CM in states of cellular stress [16], we hypothesized that attenuated fibrosis in the CRL7 loss-of-function models are due to reduced CM apoptosis. To test this hypothesis, hearts of Cul7 flox/flox AAV9-CMV-iCre and Cul7 flox/flox AAV9-CMV-dsRed were subjected to TUNEL-staining three weeks after TAC surgery. Remarkably, myocardial tissue of Cul7 flox/flox AAV9-CMV-iCre mice showed a~4.5-fold reduction in TUNEL-positive (apoptotic) cells when compared to controls (0.6 ± 0.3% vs. 2.7 ± 0.3%; p<0.001) (Fig 4). To further substantiate these findings, we analysed heart samples from Isoprenaline (ISO)-treated transgenic mice with CM-specific expression of the dominant-negative CUL7 1152stop mutant [14]. For apoptosis quantification, heart sections of MHC-1152stop mice were subjected to in situ  Table). Of note, ISO treatment did not result in increased S-phase activity in transgenic mice expressing a dominant-interfering Cul7 mutant in CM (S2 Table). Collectively, the above data support a role for CRL7 in pro-  survival signalling of CMs and suggests that inhibition of CRL7-mediated PI3K/AKT signalling protects against pressure overload-induced apoptosis in the heart.

Reduced myocardial expression of Tgfb1 upon genetic ablation of Cul7
To further delineated mechanisms of CRL7-mediated fibrotic remodelling of the heart, we studied the expression of a subset of genes encoding pro-and anti-fibrotic factors in CM upon genetic ablation of Cul7. Hearts of 12-week-old Cul7 flox/flox ; AAV9-CMV-iCre were harvested three weeks after TAC surgery and snap-frozen. Relative gene expression of transforming growth factor beta 1 (Tgfb1), cysteine rich protein 61 (Cyr61), connective tissue growth factor (Ctgf), matrix metallopeptidase 2 (Mmp2), matrix metallopeptidase 3 (Mmp3), matrix metallopeptidase 9 (Mmp9), platelet derived growth factor alpha (Pdgfa) and tissue inhibitor of metalloproteinase 1 (Timp1) was measured by quantitative RT-PCR. Surprisingly, of all candidate genes analysed, only Tgfb1 transcripts were markedly changed (-68 ± 0.2%, p<0.05) in TAC hearts when compared to Cul7 flox/flox AAV9-CMV-dsRed control animals (S10 Fig). These data suggest that altered Tgfb1 gene expression may contribute to the anti-fibrotic phenotype observed in Cul7 knockout hearts after TAC.

Discussion
This study characterized the role of CRL7 in cardiac remodelling and myocardial fibrosis. By using different genetic knockout animal models, we demonstrate that CM-specific ablation of Cul7 results in markedly reduced interstitial fibrosis upon pressure overload. Loss of Cul7 led to increased activation of the PI3K/AKT signalling pathway, that was accompanied with a significant reduction of myocardial cell apoptosis after TAC. Finally, we showed that targeted ablation of Cul7 in CM significantly downregulated mRNA level of fibrosis-modifying gene Tgfb1 under conditions of increased afterload. Collectively, our data indicate that Cul7 acts as a pro-fibrotic factor in the heart during conditions of increased myocardial stress, e.g. pressure overload.
Previous studies revealed a critical role for Cul7 in the heart in the context of cell cycle control [14]. Transgenic expression of a dominant-interfering Cul7 mutant (Cul7 MHC-1152 stop mice) resulted in a marked induction of CM cell cycle activity at the infarct border zone 4 weeks after permanent coronary artery occlusion and was associated with an induction of cardiomyocyte DNA synthesis in the interventricular septa of infarcted hearts [14]. Although cell cycle activity was not measured in the cardiac-restricted Cul7 null models employed in the present study, it is important to note that ISO treatment did not result in an increase in cardiomyocyte S-phase activity in transgenic mice expressing dominant-interfering Cul7 vs. WT mice. Thus, the marked reduction in cardiomyocyte apoptosis in these ISO-treated transgenic mice must occur independently of cardiomyocyte S-phase activity. Interestingly, expression of the MHC-1152stop mutant was also associated with a decreased infarct scar expansion in the septa of hearts, suggesting that CRL7 inactivation might partially counteract the adverse ventricular remodelling that occurs after infarction [14]. In a follow-up study, Hassink et al. corroborated these findings and showed that 4 weeks after myocardial infarction, ventricular isovolumic relaxation time constant (Tau) was decreased by 19% (p<0.05) and the slope of the dP/dtmax-EDV relationship was increased 99% (p<0.05) in infarcted MHC-1152stop mice when compared to infarcted non-transgenic littermates [31], suggesting a significant improvement in cardiac function. Our data further expands these findings by providing evidence for a pro-fibrotic role of Cul7 in the context of pressure overload-induced fibrosis. Interestingly, a recent study reported a pro-fibrotic role of Cul7 in patients with hepatocellular carcinoma and amplification of the 6p21.1 gene locus [32]. In these patients, elevated CUL7 protein concentrations were strongly correlated with advanced fibrosis in peritumoral liver areas. Collectively, these data suggest that CRL7 regulates general pro-fibrotic mechanisms that are effective in different organ systems.
What is the molecular mechanism of Cul7-mediated fibrogenesis? Pathological remodelling of the myocardium is typically characterized by loss of CM due to apoptosis, necrosis or phagocytosis [30]. We observed a marked reduction of apoptosis in Cul7-deficient CM, associated with a 3-fold over-activation of the PI3K/AKT signalling pathway. These observations are in concordance with previous reports using heterozygous conventional gene knockout models of Cul7 and Fbxw8 (the substrate targeting unit of the Cul7 ubiquitin ligase) [9]. In skeletal muscle tissue of Cul7 +/and Fbxw8 +/mice, AKT signalling was~2-fold increased under basal conditions and up to~6-fold under insulin stimulation when compared to control animals [9].
There is abundant evidence that activation of PI3K/AKT signalling has cardioprotective effects by limiting apoptotic cell death in the myocardium. Matsui et al. showed that adenoviral gene transfer of activated AKT protects CM from apoptosis in response to hypoxia in vitro [33]. Moreover, adenovirus-mediated AKT gene transfer in the heart diminishes CM apoptosis and limits infarct size following ischemia/reperfusion injury [34] and ameliorates doxorubicin-induced contractile dysfunction [35]. It is thus tempting to speculate that CRL7-mediated regulation of PI3K/AKT signalling contributes to CM survival, thereby preventing replacement fibrosis. However, we cannot rule out that CRL7 also affects other pro-fibrotic pathways independent of cell death.
It remains an open question whether CUL7 modulates apoptosis in a non-proteolytic manner, e.g. independent of CRL7's role in ubiquitin-mediated degradation of IRS-1. In a previous report, Tsai et al. identified a putative BH3 domain in the C-terminus of CUL7 [36] and postulated that CUL7 belongs to the BH3-only family of pro-apoptotic proteins. It was shown that forced expression of CUL7 in NIH-3T3 cells promoted apoptosis in a manner that was dependent on the integrity of the BH3 domain [36]. However, it should be noted that Kim et al. identified CUL7 in a functional screen for inhibitors of Myc-induced apoptosis, showing that expression of CUL7 prevented both c-Myc and N-Myc mediated apoptosis and promoted the transformation of neuroblastoma SHEP cells in a p53-dependent manner [37]. Further studies are therefore needed to dissect the different roles of CUL7 as a regulator of apoptosis in various cell types and tissues.
We observed a marked down-regulation of Tgfb1 gene transcripts in CM deleted of Cul7. Substantial evidence suggests that TGF-ß triggers fibrosis and apoptosis in different cell types [38]. Interestingly, expression of activated PI3K, AKT or IRS-1 was shown to antagonize TGFß-mediated apoptosis [39][40][41]. It was suggested that this pro-survival effect is dependent on PI3K/AKT-mediated suppression of Smad3, a central component of the TGF-ß signalling pathway [42]. It is tempting to speculate that over-activation of PI3K/AKT, as shown for Cul7-/-hearts, contributes to diminished Tgfb1 mRNA levels, thereby ameliorating fibrosis after TAC. In addition, PI3K/AKT signalling pathways regulate key events in the inflammatory response to damage [43]. It is therefore tempting to speculate that the observed effect of CRL7 ablation on cardiac fibrosis is partly mediated by PI3K/AKT-regulated inflammation and its downstream effects on fibrogenesis. Clearly, more studies are needed to further dissect the molecular mechanisms by which CRL7 restrains cardiac fibrosis. Whether pharmacological targeting of CRL7 is a useful strategy to mitigate excess fibrosis in the heart awaits future investigation.
It is generally believed that myocardial fibrosis promotes ventricular dysfunction, e.g. by increased diastolic stiffness and impaired relaxation (reviewed in [30]). However, an unexpected result of our study was the lack of functional cardiac improvement in Cul7-deficient mice despite a~50% reduction of interstitial myocardial fibrosis under conditions of pressure overload. A similar observation was made by Yang and colleagues who reported that the antifibrotic peptide Ac-SDKP markedly attenuated cardiac fibrosis without improving EF and E/A ratios [44]. It is tempting to speculate that CRL7-mediated modulation of fibrosis early in cardiac injury may affect cardiac repair and remodeling in an adverse manner, e.g. by suppression of protective collagen-mediated myocyte-to-myocyte interaction or alterations of the myocardial microenvironment. In addition, loss of regulated AKT activity upon genetic ablation of CRL7 may have affected cardiac function, as balanced AKT signaling is essential for contractility and the response to pathological stress [16]. Proof-of-concept studies have shown that CRLs are druggable targets. Wu et al. recently reported the identification of the small molecule drug suramin as inhibitor of CRLs that contain cullin 2, 3, and 4A [45]. Whether targeting CRL7 by small molecules will provide a novel strategy to impede excess fibrosis in the heart awaits future investigation. Further work is needed to dissect the molecular mechanisms of CRL7-mediated regulation of pressure overload-induced myocardial fibrosis and the potential of the CRL7 ubiquitin ligase as a pharmacological target for the treatment of cardiac fibrosis.