Ubiquitin C-terminal hydrolase L1 (UCHL1) regulates post-myocardial infarction cardiac fibrosis through glucose-regulated protein of 78 kDa (GRP78)

Abnormal cardiac fibrosis indicates cardiac dysfunction and poor prognosis in myocardial infarction (MI) patients. Many studies have demonstrated that the ubiquitin proteasome system (UPS) plays a significant role in the pathogenesis of fibrosis. Ubiquitin C-terminal hydrolase L1 (UCHL1), a member of the UPS, is related to fibrosis in several heart diseases. However, whether UCHL1 regulates cardiac fibrosis following MI has yet to be determined. In the present study, we found that UCHL1 was dramatically increased in infarct hearts and TGF-β1-stimulated cardiac fibroblasts (CFs). Inhibition of UCHL1 with LDN57444 (LDN) reversed the myocardial fibrosis in post-MI heart and improved cardiac function. Treatment of LDN or UCHL1 siRNA abolished the TGF-β1-induced fibrotic response of CFs. We further identified GRP78 as an interactor of UCHL1 through screening using immunoprecipitation-mass spectrometer. We determined that UCHL1 interacted with glucose-regulated protein of 78 kDa (GRP78) and prompted GRP78 degradation via ubiquitination. Furthermore, we found that GRP78 was upregulated after UCHL1 knockdown and that the GRP78 inhibitor HA15 diminished the antifibrotic function exerted by UCHL1 knockdown in CFs stimulated with TGF-β1. This suggests that UCHL1 regulates cardiac fibrosis post MI through interactions with GRP78. This work identifies that the UCHL1-GRP78 axis is involved in cardiac fibrosis after MI.

www.nature.com/scientificreports/ Interference in the UPS leads to deficits in differentiation [11][12][13] . In addition, the UPS is fundamental in controlling cardiac fibrosis in cardiac hypertrophy 14,15 . To investigate how the UPS regulates CF activation under the conditions of MI, we focused on ubiquitin C-terminal hydrolase L1 (UCHL1), a member of UPS family. UCHL1, which was initially thought to be a neuronal marker, was found to be expressed in the other tissues, such as heart, liver and kidney, among others 16 . Previous studies have shown that the role of UCHL1 varies in a context-dependent manner. UCHL1 can function as an ubiquitin ligase in its dimer form other than hydrolase 17 , UCHL1 also stabilises proteins via targeting glutathione 18,19 . Therefore, there is a dual action of UCHL1 in a variety of diseases, including kidney diseases, neurodegenerative diseases and cancers, with UCHL1 affecting cellular proliferation, cell cycle, migration and invasion [20][21][22][23][24][25] . To date, only a few studies have assessed the role of UCHL1 in heart diseases. One study found that UCHL1 was markedly upregulated in post-MI heart 26 and another suggested that UCHL1 acts as a regulator of the Ang II-induced atrial fibrillation and inhibition of it attenuated atrial fibrosis in vivo 27 . Inspired by the finding that the UCHL1 exerts both endogenous and exogenous pro-activation effects in hepatic stellate cells, which are a type of fibroblast 28,29 , we hypothesised that UCHL1 may promote cardiac fibrosis following MI via induction of CF activation.
In this study, we investigated the role of UCHL1 in mouse MI models and primary CFs. Of note, we found that inhibition of UCHL1 improved cardiac function and attenuated cardiac fibrosis post MI. Furthermore, we found that inhibition and knockdown of UCHL1 hindered the cardiac fibrotic response via upregulation of glucose-regulated protein of 78 kDa (GRP78) in CFs. The underlying mechanism of this is largely attributed to the interaction between UCHL1 and GRP78, and the subsequent degradation of GRP78 by ubiquitination. Therefore, we developed a novel mechanism for maladaptive cardiac fibrosis post MI, providing insight into potential pathways to target for novel antifibrotic therapies.

UCHL1 is elevated in post MI hearts and CFs treated with TGF-β1.
To determine if UCHL1 regulates cardiac fibrosis, we measured UCHL1 expression in post-MI hearts. IHC showed that UCHL1 was highly expressed in the infarct area of MI hearts at 7 days and 14 days after MI as compared with corresponding areas of sham hearts (Fig. 1a). Importantly, the protein expression level of UCHL1 was increased at 7 and 14 days post MI. There were also increased levels of pro-fibrotic proteins, Col1 and α-SMA (Fig. 1b), indicating that UCHL1 was enhanced at 7 days post-MI, and that the increased levels of UCHL1 last at least 7 days.
Fibrotic models were established in CFs with TGF-β1 stimulation. UCHL1 expression in CFs was significantly increased at 24 and 36 h after TGF-β1 (10 ng/ml) stimulation, compared with cells without TGF-β1 stimulation (Fig. 1c). Col1 and α-SMA were also increased accordingly (Fig. 1c), indicating that TGF-β1 induces fibrotic response in CFs. Taken together, our data show that UCHL1 protein levels increased in both fibrotic post-MI hearts and TGF-β1-induced CFs, suggesting that UCHL1 may be involved in the process of cardiac fibrosis following MI.

Inhibition of UCHL1 by LDN mitigated post-MI fibrosis and enhanced cardiac function.
To assess whether UCHL1 mediates post-MI cardiac fibrosis in vivo, we used LDN to inhibit UCHL1. As LDN can also inhibit UCHL3, we determined the UCHL3 protein level in the sham group and MI group 14 days after operation and found that no significant differences between the two groups (Supplemental Fig. 1). Profibrogenic proteins, α-SMA and Col1, significantly increased in hearts 14 days after MI, as compared with the sham hearts (Fig. 2a). Fibrotic factors were significantly reduced by LDN in the hearts of the MI group at 14 days (Fig. 2a). Staining of collagen deposition by Masson's trichrome staining in sham/MI hearts and with/without LDN treatment showed that the UCHL1 inhibitor LDN dramatically prevented MI-associated infarct size and cardiac fibrosis (Fig. 2b). Our data indicate that UCHL1 inhibition prevents the heart from post-MI fibrosis. As cardiac fibrosis is a major factor in post-MI cardiac dysfunction, we next measured cardiac function in sham/ MI mice treated with/without LDN. Echocardiography results showed that cardiac ejection fraction (EF) and fractional shortening (FS) in MI mice was significantly reduced at 14 days after MI as compared with sham mice (Fig. 2c). However, UCHL1 inhibitor, LDN, significantly improved post-MI cardiac function. Our data supports the hypothesis that UCHL1 inhibition has a protective effect on cardiac function after MI.
Both inhibition of UCHL1 by LDN and knockdown UCHL1 by siRNA downregulated TGF-β1-mediated expression of pro-fibrotic proteins in CFs. We next suppressed activity of UCHL1 in CFs with LDN. While TGF-β1 stimulation increased profibrogenic proteins Col1 and α-SMA in CFs, inhibition of UCHL1 blocked TGF-β1-induced increases of these profibrogenic proteins (Fig. 3a). We next used siRNA to knock down UCHL1 in CFs (Fig. 3b). We found that UCHL1 siRNA decreased TGF-β1-induced increases of these profibrogenic proteins (Fig. 3c). The UCHL1 siRNA preventing CFs from fibrosis induced by TGF-β1 was also observed by immunofluorescence staining with α-SMA antibody (Fig. 3d). Collectively, our results show that UCHL1 regulates the fibrotic response of CFs. UCHL1 regulated fibrotic responses through GRP78. To identify the underlying mechanisms of UCHL1 on post-MI fibrosis, we performed immunoprecipitation-mass spectrometer (IP-MS) in TGF-β1induced CFs to find UCHL1 interactors (Fig. 4a). Through analysing the MS data, we list the proteins which number of PSMs > 20 as candidates that IP with UCHL1 (Fig. 4b). GRP78, also named endoplasmic reticulum chaperone BiP, was ranked at the top of the list (Fig. 4b). Therefore, we detected if there was crosstalk between UCHL1 and GRP78. We found that UCHL1 interacted with GRP78 in CFs with TGF-β1 stimulation using Co-IP (Fig. 4c). Immunofluorescence staining showed that UCHL1 colocalised with GRP78 in fibrotic areas of MI hearts (Fig. 4d) as well as in CFs induced with/without TGF-β1 (Fig. 4e). www.nature.com/scientificreports/ To find whether GRP78 is regulated by UCHL1, we examined the GRP78 protein level and found that knockdown of UCHL1 in CFs elevates the GRP78 protein levels (Fig. 5a). We next blocked protein synthesis in CFs with/without UCHL1 knockdown using CHX for 0-5 h and measured the subsequent GRP78 degradation. We found that the rate of GRP78 degradation was significantly decreased with UCHL1 knock down (Fig. 5b). Downregulation of UCHL1 by UCHL1 siRNA decreased poly-ubiquitination of GRP78 (Fig. 5c). Taken together, these results suggest that UCHL1 interacts with GRP78 and downregulates it via ubiquitination.
Moreover, we wondered if the effect of UCHL1 on fibrosis was related to GRP78, so we determined the GRP78 protein levels in CFs with/without TGF-β1 and treated with/without UCHL1 knockdown. We found that GRP78 protein levels were upregulated by TGF-β1 and knockdown of UCHL1 led to greater increases in GRP78 levels ( Fig. 6a). We also used HA15, a GRP78 inhibitor, to inhibit GRP78, and found that inhibition of GRP78 abolished the downregulation of Col1 and α-SMA in response to UCHL1 siRNA in CFs with TGF-β1 stimulation (Fig. 6b). Together, our results suggest that UCHL1 regulates cardiac fibrosis by GRP78 (Fig. 7).

Discussion
Cardiac fibrosis is a critical pathologic process in ventricular remodelling after MI. Since the role of UPS in mediating cardiac fibrosis is unknown, we sought to determine the role of UCHL1 in cardiac fibrosis following MI. UCHL1 inhibition attenuated cardiac dysfunction and fibrosis following MI. LDN and UCHL1 siRNA inactivated CFs and notably, GRP78, a critical molecule involved in cellular homeostasis by handling unfolded www.nature.com/scientificreports/ proteins, was identified as an interactor of UCHL1. UCHL1 prompted GRP78 degradation by interacting with and ubiquitinating it. Knockdown of UCHL1 upregulated GRP78 in CFs with TGF-β1 stimulation and inhibition of GRP78 diminished the antifibrotic effect of UCHL1 knockdown. Thus, the UCHL1-GRP78 axis could be a novel target for cardiac fibrosis therapy. Dobaczewski et al. found that UCHL1 was dramatically elevated at 7 days and lasted for at least 21 days post but the underlying mechanism behind this process was unclear 26 . We hypothesised that UCHL1 may be a key mediator of post-MI remodelling. Since no studies focus on the role of UCHL1 in MI, we aimed to investigate the UCHL1 on mouse MI model. We inhibited UCHL1 using active site-directed inhibitor LDN 30 , and found that the treatment improved the cardiac function and attenuated cardiac fibrosis after MI. UCHL1 staining was observed in the area of fibrosis in the infarct heart using IHC. Therefore, we aimed to assess the negative effect of UCHL1 on the infarct heart. As expected, we verified the antifibrotic role of UCHL1 inhibition on CFs stimulated with TGF-β1 using LDN. Results from previous studies had very different results that suggest that UCHL1 is a promising repressor for CF activation 31 . The differences between these studies may be due to the source of CFs, as the CFs of our study are isolated from adult mice rather than neonatal rats; the neonatal heart but not the adult heart, possesses regeneration potential. Another potential difference between the studies is that our study stimulated CFs with TGF-β1, while previous studies used PDGF. The role of UCHL1 relies on the context of the cells. The antifibrotic role of LDN on the heart was also shown by another study that examined atrial fibrillation but did not use cell culture models 27 . In addition, the pro-activation effect of UCHL1 is observed in other types of fibroblasts, such as cancer-associated fibroblasts and hepatic stellate cells 29,32 . These findings suggest a novel potential target in CF activation.
To find the underlying mechanisms of UCHL1, we screened its interactor using IP-MS and identified GRP78 as candidate interactors. This is consistent with the finding that GRP78 is colocalised with UCHL1 in COS-7 cells 33 . Thus, there exists a possibility that UCHL1 interacts with GRP78 through the UCHL1-GRP78 complex. GRP78 is a molecular chaperone of the Hsp70 family with protective properties, such as stabilising the calcium concentration of endoplasmic reticulum as a calcium binding protein, transferring the misfolded protein out of the endoplasmic reticulum and helping to fold unfolded proteins 34 . To pinpoint if there is a direct interaction between UCHL1 and GRP78, we validated the interaction of UCHL1 and GRP78 via co-immunofluorescence www.nature.com/scientificreports/ and co-immunoprecipitation. We found that GRP78 was significantly increased in CFs treated with UCHL1 siRNA, consistent with an investigation in SK-N-SH cells 35 . The upregulation of GRP78 resulted from the reduction of ubiquitination by UCHL1 knockdown. Therefore, the effect of UCHL1 on cardiac fibrosis may be due to its control of GRP78. GRP78 is a master mediator of the unfolded protein response 34 . The effect of GRP78 on fibrosis is partly embodied in the 'two-edged sword' function of the unfolded protein response in fibrosis-related pulmonary diseases and diabetic nephropathy [36][37][38] . When it comes to fibrosis in MI, the role of GRP78 on ischaemic myocardium, either protective or harmful, lies on environment 39 . We found that GRP78 was upregulated in TGF-β1 stimulated CFs and a greater increase of GRP78 was observed in TGF-β1 stimulated CFs treated with UCHL1 siRNA. So GRP78 may play a protective role in TGF-β1 stimulated CFs. To find out whether UCHL1 exerts its pro-fibrosis effect through inhibition of the protective effect of GRP78 in the process of cardiac fibrosis, we used HA15 to inhibit the GRP78. HA15 specifically targets GRP78 and inhibits its ATPase activity 40 . HA15 diminishes the antifibrotic effect of UCHL1 siRNA with TGF-β1 stimulation. Collectively, our data suggests that the UCHL1-GRP78 complex plays a role in the regulation of cardiac fibrosis.
One limitation of this study is the gene manipulation. UCHL1 knockout mice should be used in future studies to confirm the effect of UCHL1 knockout. As UCHL1 was also overexpressed in cardiomyocytes, and the effect of UCHL1 on cardiomyocytes remains to be determined 26 . The signalling pathways upstream of UCHL1 remain www.nature.com/scientificreports/ unknown, which may relate to osteopontin which would link CFs and cardiomyocytes 41 . One potential challenge for clinical translation is that a clinical therapeutic would need to be tissue specific, and have an inhibitory effect on UCHL1 solely in the heart. This study demonstrates that UCHL1 enhances cardiac fibrosis after MI by interacting with and downregulating GRP78 by ubiquitination (Fig. 7). Interventions that target the UCHL1-GRP78 interaction may be a potential therapeutic strategy against cardiac fibrosis post-MI. Our findings deepen the understanding of maladaptive fibrogenesis post-MI with a focus on UPS.

Induction of MI and in vivo experimental design. MI was induced via permanent ligation of the left
anterior descending coronary artery (LAD). Briefly, mice were anaesthetised with sodium pentobarbital (intraperitoneal injection, 50 mg/kg; Merck, China) and mechanically ventilated by the HX-101E ventilator (Chengdu Taimeng Software Ltd., China). The thoracotomy was performed in the 4th left intercostal space, and then the LAD was occluded with 7-0 polyester sutures. The chest wall was closed using 5-0 polyester sutures. The success of the operation was confirmed by blanched heart apex and elevated ST segment on the ECG. Sham-operation was performed in the same manner just leaving the LAD without an occlusion.   Echocardiography. Echocardiography was performed 14 days after operation using a Vevo 1,100 High Resolution Ultrasonic Imaging System (Visualsonics, Toronto, Canada). The mice were slightly anaesthetised with pentobarbital sodium (20 mg/kg) and kept at a heart rate > 400 bpm under an ECG monitor. The hearts' long and short axis views were obtained in B-mode and M-mode, respectively. The M-mode photographs were acquired between the two papillary muscles. Left ventricular EF and FS were automatically calculated by the echocardiography software. All measurements were the mean of three successive cardiac cycles.
Histological analyses. While mice were under deep anaesthesia, mouse hearts were immediately extracted and fixed in 4% formaldehyde. Then the fixed hearts sequentially underwent dehydration and paraffin embedding. Subsequently, 4 μm cross-sectional slices were obtained. www.nature.com/scientificreports/ Immunohistochemistry (IHC). After blocking with 10% serum of the secondary antibody host for 1 h, the slices were incubated with a primary antibody against UCHL1 (1:500, Proteintech, 14730-1-AP) overnight at 4 °C. The slices were then incubated with streptavidin-HRP conjugated secondary antibody for 1 h at room temperature. A drop of 3,3′-diaminobenzidine (DAB) solution (Boster, AR1022, China) was applied until the brown positive stain appeared. Then the counterstain was performed with haematoxylin. All the slices were scanned with an Aperio VERSA System (Leica Biosystems, Germany). Three representative images from each sample were taken and analysed using the Image J software (National Institutes of Health, Bethesda, USA).
Masson's trichrome staining. To evaluate cardiac fibrosis, Masson's trichrome staining was performed in paraffin-embedded heart cross-sectional slices at the mid-papillary level. The area stained blue was the fibrosis area. The slices were captured using an Aperio VERSA System (Leica Biosystems, Germany). The areas were measured using the Image J software (National Institutes of Health, USA). The ratio of fibrosis to the whole left ventricle was used to compute infarct size expressed as a percentage of the left ventricle. And the ratio of fibrosis to viable myocardium in the peri-infarct area was used to compute interstitial fibrosis expressed as a percentage of surface area 42 .
Adult mouse CF isolation. CFs were isolated from 4-month-old male C57BL/6 mice that weighed 20-25 g by collagenase digestion. Briefly, after deep anaesthesia was induced in mice, the cardiac ventricles were rapidly removed and rinsed with cold sterile PBS under sterile conditions. Next, the ventricles were finely minced followed by digestion in 10 ml DMEM containing 0.2% type 2 collagenase (Worthington Biochemical, Lakewood, NJ, USA) at 37 °C for 90 min with gentle shaking. The supernatant was then centrifuged at 1000 g for 5 min and resuspended in culture medium (DMEM containing 10% FBS supplemented with 1% penicillin and 1% streptomycin) before plating into 10 cm culture dishes. Non-adherent cells were removed after culturing for 4 h and the adherent cells were further cultured in fresh culture medium. Afterwards, the isolated cells were detected by immunofluorescence using antibodies against vimentin. CFs, which are positive for vimentin, accounted for > 95% (Supplemental Fig. 2). The isolated cells were used for in vitro studies between passage 2 and 4. Cycloheximide (CHX) chase assay. After transfection, CFs were incubated with CHX (20 µM, HY-12320, MedChemExpress), then harvested at 0, 1, 2, 3, 4 and 5 h after CHX addition. The lysates were subjected to immunoblot for evaluation of GRP78 protein stability.
Western blot. Protein was extracted from heart tissue or CFs using RIPA buffer (Beyotime, China) containing protease inhibitor (Complete Mini EDTA-free, Roche) and phosphatase inhibitor (PhosSTOP, Roche www.nature.com/scientificreports/ Immunoprecipitation-mass spectrometer (IP-MS). The immunoprecipitants of eluted proteins were subjected SDS-PAGE. Gel lanes were excised following coomassie blue staining and submitted for mass spectrometry with Orbitrap Elite mass spectrometers (Thermo Fisher Scientific).
Ubiquitination assay. CFs were treated with Scrambled siRNA or UCHL1 siRNA for 48 h. MG132 (10 μM) was added 12 h before the cells were harvested. Whole cell lysates were immunoprecipitated with anti-GRP78 antibody and then immunoblotted with anti-ubiquitin (Ub) antibody to evaluate the ubiquitination level of GRP78 protein.
Statistical analysis. Statistical analysis was performed using GraphPad Prism 6 (GraphPad Software Inc, USA). Differences between two experimental groups was analysed using the Student's t-test and multiple comparison groups were analysed using one-way ANOVA with Tukey test. All data were presented as the means ± SD. P < 0.05 was considered statistically significant.

Study approval. All animal experiments and surgical procedures were conducted in line with the Animal
Care and Use Committee Guide of Wuhan University, which conforms to the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (NIH Publication No. 85-23, revised 1996). All methods were performed in accordance with relevant guidelines and regulations.

Data availability
All relevant data supporting the conclusions are included in this published article and its supplementary information files.